INTERNATIONAL COURT OF JUSTICE
DISPUTE OVER THE STATUS AND USE OF THE
WATERS OF THE SILALA
(CHILE v. BOLIVIA)
REJOINDER OF THE
PLURINATIONAL STATE OF BOLIVIA
ANNEXES 24-28
VOLUME 5 OF 6
15 MAY 2019
LIST OF ANNEXES TO THE REJOINDER OF THE
PLURINATIONAL STATE OF BOLIVIA
VOLUME 5 OF 6
ANNEX N°
TITLE PAGE N°
(ANNEXES 24-28)
TECHNICAL DOCUMENTS (ANNEX 24-26)
Annex 24 DHI, “Analysis and assessment of Chile’s reply to
Bolivia’s counter claims on the Silala Case”, March 2019
(Original in English)
5
Annex 25 DHI, “Updating of the mathematical hydrological model
scenarios of the Silala spring waters with: Sensitivity
analysis of the model boundaries”, April 2019
(Original in English)
47
Annex 26 FUNDECO, “Study of the Water Requirements of the
Silala Wetlands”, April 2019
(English Translation)
89
OTHER DOCUMENTS (ANNEXES 27-28)
Annex 27 Note S/N of The Antofagasta (Chili) and Bolivia Railway
P.L.C addressed to the Company DUCTEC S.R.L.,
Antofagasta, 23 August 2000
(Original in Spanish, English Translation)
155
Annex 28 1906 Chilean Concession to THE ANTOFAGASTACHILI
AND BOLIVIA RAILWAY P.L.C. Obtained from
the data base of Chile’s Direction-General of Water, 2019
http://www.dga.cl/Paginas/default.aspx
(Original in Spanish, English Translation)
159
Annex 24
DHI, “Analysis and assessment of Chile’s reply to Bolivia’s
counter claims on the Silala Case”, March 2019
(Original in English)
7
D~
Single Product
Analysis and assessment of Chile's
reply to Bolivia's counter claims on the
Silala Case
The expert in WATER ENVIRONMENTS
Pluri-national State of Bolivia,
DIREMAR
18 March 2019
8
D~
This report has been prepared under the OHi Business Management System
certified by Bureau Veritas to comply with ISO 9001 (Quality Management)
Approved by
Oluf Zeilund Jessen
Head of Pro·ect, Water Resource De artment
9
Single Product -
Analysis and assessment of Chile's
reply to Bolivia's counter claims on the
Silala Case
Prepared for
Represented by
Proiect manaaer
Quality supervisor
Project number
Approval date
Revision
Classification
Pluri-national State of Bolivia, DIREMAR
Dr. Emerson Calderon
Roar A. Jensen
Michael B. Butts
11823606
18 March 2019
2.0
OHi A/S • Agern Alie 5 • DK-2970 H0rsholm • Denmark
D~
Telephone: +45 4516 9200 • Telefax: +45 4516 9292 • [email protected] • www.dhigroup.com
10
11
D~
CONTENTS
1 Executive Summary .................................................................................................... 7
2 Introduction ................................................................................................................. 9
2.1 Background for this report ........................................... .................................................. 9
2.2 The structure of Chile's reply and this response ........... ................. .... .. .......................... 9
3 The technical arguments and evidence in Chile's reply ......................................... 11
3.1 Comments to Chile's assessment of the technical agreements on the hydrology of
Silala ........................................................................................................................... 11
3.2 Evaluation of the technical disagreements on the Silala hydrology .............................. 12
3.2.1 The internationally agreed impacts of artificial wetland drainage ................................. 13
3.2.2 Regarding the Far Field model and basin water balance ............................................. 15
3.2.3 The underlying interpretation of the geology and hydro-geology .................................. 17
3.2.4 Disputes regarding the Near Field numerical model ............................................. ....... 22
3.2.5 Analysis of Chile's comparative studies of the Silala wetlands ................. .. ................. 26
4 Assessment ............................................................................................................... 29
5 Technical conclusions .............................................................................................. 37
6 References ................................................................................................................. 39
The expert in WATER ENVIRONMENTS 5
12
6
13
1 Executive Summary
DHI has been requested by DIREMAR to provide a technical response to a subset of
technical issues in of the document "International Court of Justice (Feb 2019): Dispute
over the status and use of the waters of the Silala (Chile v. Bolivia). Reply of The
Republic of Chile (Volumes 1-3)".
DHI has only commented on these issues. The absence of comments to other parts of
the above-mentioned documents does not in any way imply that DHI agrees with them.
The structure the report is as follows:
Chapter 2: Background and structure
Chapter 3: The technical arguments and evidence in Chile's reply
Chapter 4: Assessment of the strengths and weaknesses
Chapter 5: Technical conclusions and recommendations
Under current conditions the water from the Silala Basin flows as both surface water
and groundwater, however in the responses provided by Chile, there is an underlying
presumption that the water resource supplying the Silala River basin is known.
Groundwater is known as the "hidden" resource, which is generally more challenging to
quantify. The spatial extent and volume of the groundwater aquifer discharging
groundwater to the Silala springs remain unknown.
Radiometric data and estimated groundwater travel times for the inferred groundwater
catchment indicate that a large part of present-day groundwater discharge at the Silala
springs, was likely recharged under climatic conditions that prevailed thousands of
years ago. Long-term declines in aquifer storage associated with reduced modern
recharge, as is common throughout the Andes, cannot be eliminated with available
data. Therefore, it cannot be claimed conclusively that the current inflows to and
outflows to the aquifer supplying discharge to the Silala springs are balanced. As a
result, the discharge to the Silala springs may well be decreasing with time, in which
case it is a non-renewable groundwater resource.
The claim that the "works do not have a significant impact on surface water flows" is
contradicted by the on-site field evidence, the scientific literature and by Chile's own
experts.
OHi has emphasised that the impact assessment of removal of the canals is uncertain,
as is also acknowledged by Chile's expert. We also agree that, under circumstances
where the boundary is located close to the interventions and therefore may be affected
by them, sensitivity analyses should be considered.
The validity of Chile's simplified impact calculations is questionable and therefore do
not support the claim that DHl's impacts are exaggerated. The analysis is based on the
one-dimensional Darcy equation, which is only valid under idealized conditions not
satisfied at Silala. The groundwater aquifer is not homogeneous. The aquifer is both
confined and unconfined. The groundwater flow is not one-dimensional but rather
highly three-dimensional. In particular, the one-dimensional Darcy approach does not
represent correctly the observed changes in groundwater gradients and therefore the
flows towards the spring discharge zone and lacks reference to field data.
7
14
8
Chile emphasizes the importance of the highly complex and three-dimensional geology,
yet they then ignore this complex geology in their simplified analysis. This is a clear
inconsistency, which brings into question the validity of their assessments of the
canalization impacts.
Regarding geology and the 3-D hydrogeological model developed by DHI , Chile finds
no basis for incorporating a fault zone with high hydraulic conductivities. Chile claims
that DH l's hydrogeological model does not account for vertical variability in permeability
based on flowing conditions at a Chilean borehole. Chile also claims that DHl's model
does not reflect Chile's interpretations and that differences in chemistry and isotopic
compositions of groundwater are not reflected in the hydrogeological model. OHi
agrees that geological and hydrogeological models involving many data sources are
subject to interpretation. Nevertheless, the field data do support DHl's hydrogeological
model and it is consistent with Chile's borehole information. Chile's claim that there is
no basis for incorporating a fault or a highly fractured zone is not correct. Chile's
statement that DHl's does not account for vertical variability is also incorrect since both
the conceptual and numerical models account for both vertical anisotropy and changes
in horizontal hydraulic conductivity with depth.
We do not oppose the statements made by Chile on the preference for water balance
preserving groundwater models with fixed upstream boundaries and "known" recharge
input. This concept was considered and rejected due to the lack of hydrogeological
data in the Far Field and the inability, based on available data, to define correct
catchment and aquifer boundaries. A groundwater calibration of the whole Far Field
area to match the few data in the Near Field would have to be based on a lot of
assumptions about the presently uncharacterised areas of the aquifer.
The technical approach employed was therefore to collect hydrogeological information
within and in the vicinity of the Near Field. This allowed for the development of a
numerical model that was calibrated to field characterisation data including hydraulic
parameters and head distributions at various depths. The calibrated Near Field model
reproduces the field data, which improves the reliability of the predictive capability of
the model. In contrast, Chile has, repeatedly and strenuously, highlighted the
importance of the geologic framework but then choose to ignore what they themselves
consider important. Instead they chose to use a highly simplified, uncalibrated
analytical solution that they agree does not capture the site hydrogeological conditions
to critique model results.
A sensitivity and uncertainty analysis of the Near Field boundary conditions could be
considered.
15
2 Introduction
2.1 Background for this report
Introduction to the contract and its objectives
On 6 June 2016, the Republic of Chile filed an application with the International Court
of Justice against the Pluri-national State of Bolivia, concerning the dispute over the
Status and Use of the Waters of the Silala (Chile v. Bolivia). In the framework of this
legal case, Chile presented its Memorial on 3 July 2017 and Bolivia presented its
Counter-memorial and Counterclaim on 31 August 2018. OHi has under a former
contract for Bolivia's Strategic Office for the Maritime Claim, Silala and International
Water Resources - DIREMAR carried out technical analyses and assessments of the
surface water and groundwater flows in the Silala Springs System, analyses that were
referenced in Bolivia's Counter Claim to Chile.
On 15 February 2019, Chile presented its response to Bolivia's counter claims1 . OHi
has been contracted by DIREMAR (through the product-based consultancy contract:
CDP-IN° 0712019) to make a technical assessment of the Chilean reply.
The present report describes the output of this assessment.
OHi is a technical water institution and our assessments reported in this document
concerns only the technical arguments put forward in Chile's reply.
2.2 The structure of Chile's reply and this response
In this section, we introduce Chile 's reply and its structure highlighting the parts we
have analysed and evaluated in this report. The section also gives a short description
of how our report is structured.
Chile's reply consists of three main chapters.
• Chapter 1 introduces the process and the structure of the reply. This
chapter will not be commented on in this report.
• Chapter 2 presents Chile's legal arguments against Bolivia's second
counter claim: "Bolivia has sovereignty over the artificial flow of Silala
waters engineered, enhanced, or produced in its territory and Chile has no
right to any part of that artificial flow". The arguments put forward in this
chapter are of legal rather than technicalnature and OHi does not have the
expertise to comment on these points .. OHi has, in previous reports,
referred to a modified hydrological system affecting flow rates in surface
water and groundwater. This terminology is understood to correspond to
the 'artificially enhanced flows' described in Bolivia's counter claim.
1 International Court of Justice (Feb 2019): Dispute over the status and use of the waters of the Silala (Chile v. Bolivia). Reply of The
Republic of Chile
9
16
Chapter 3 describes Chile's comments to the technical arguments
presented in Bolivia's Counter memorial and is the principal focus of this
report.
In Subsection A of Chapter 3, Chile lists the points they find the two
parties agree upon. We have commented on these points in Section 3.1
below.
In subsection B of Chapter 3, Chile explains the technical assessments
and analyses to which they do not agree with Bolivia. The points made in
the subsection is based on analyses and views as stated in two attached
reports prepared by Chile's technical experts2 3(Drs. Wheater and Pearch).
We analyse these points, in detail, in Section 3.2 of this report.
Furthermore, Section 4 of this report summarises Chile's arguments with our evaluation
2 Wheater, H.S. and Peach, D. W. (2019) Impacts of Channelization of the Silala River in Bolivia on the Hydrology of the
Silala River Basin, Attachments to Reply of the Republic of Chile to ICJ on Dispute over the status and use of the waters
of the Silala.
3 Peach, D. W. and Wheater, H.S. (2019) Concerning the Geology, Hydrogeology and Hydrochemistry of the Silala River
Basin, Attachments to Reply of the Republic of Chile to ICJ on Dispute over the status and use of the waters of the Silala
10
17
D~
3 The technical arguments and evidence in Chile's reply
In this section, we comment on and discuss Chile's technical arguments regarding
impacts of the canalisation on cross-border surface water flow.
3.1 Comments to Chile's assessment of the technical agreements on the
hydrology of Silala
In Section A, Chile and Bolivia largely agree on the nature and functioning of Si/ala as
an international watercourse.
In Section A4 of the Chilean response, which is intended to summarise the points on
which the Parties agree, Chile has introduced several claims to which OHi does not
fully agree.
OHi can agree that water from the Silala System basin will flow from Bolivia, as surface
water or groundwater. In many of the responses provided by Chile, there is an
underlying presumption that the spatial extent, hydrogeologic properties and storage of
this aquifer are precisely known and constant in time. The aquifer's extent, thickness,
hydraulic properties and recharge rates to the aquifer are not known and thus, the
volume of water stored in the aquifer supplying the Silala springs is not precisely
known, (section 3.2.2, below). In addition, analyses of the groundwater residence times
indicate the predominant source of groundwater discharging at the Silala springs,
originated under climatic conditions that prevailed on the order of thousands of years
ago. It cannot be claimed conclusively that the present-day conditions in the aquifer
supply to the Silala springs are steady-state, as draining of the aquifer from periods of
higher recharge and higher groundwater elevations may still be occurring.
OHi can agree "that the Silala River is a perennial flow that rises at two sets of springs
in Bolivia and flows along the natural topographic gradient from Bolivia into Chile"5
under current conditions. This, however, provides no evidence to the claim that "the
direction of flow has been the same for thousands of years". In practice, it is likely that
recharge rates, groundwater flow patterns and Silala spring discharge have varied
through time in response to variations in climate that may have changed substantially at
the scale of millennia.
Draining of the Silala wetlands has led to a reduction of the spatial extent of the
inundated areas, not a "possible" reduction. This will result in , not "may" result in, a
reduction in open water evaporation . The reduction in extent of inundated areas is also
reflected in the presence of invasive grasses in soils that are now drained. Both sides
can agree that evapotranspiration is a small component of the total water balance of
the Silala River System.
While we agree that the influence of the wetland evaporation on the cross-border flows
is minor, we do not agree with quantitative conclusions drawn by recent studies in
Chile6 because of several shortcomings in the study.
4 Vol 1. Section A Chile and Bolivia largely agree on the nature and functioning of the Silala River as an international
watercourse
5 Vol 1. Para 3.7
6 Vol 1, Para 3.10
11
18
a) The study compares different wetlands with differing hydrogeologic properties,
gradients and discharge relationships.
b) The study does not cover the pre-drainage period of the Bolivian Wetlands.
c) The evapotranspiration calculation is, unproven on the site, neglects
evaporation from open water, that would be prevalent in an undisturbed
wetland, and gives evapotranspiration values that are only 50% of (or lower) the
potential or reference values, which are considered implausible for natural
wetlands with a constant supply of water (i.e. not rate limited by water
availability.)
The study shows changed hydrological behaviour of the drained and undrained
wetlands, contradicting Chile's own claims elsewhere, section 3.2.1.
3.2 Evaluation of the technical disagreements on the Silala hydrology
12
Chile claims the Bolivia 's contentions on the impacts of the canalization are untenable
as a matter of fact. We will in this section demonstrate that although our previous
assessments are inherently uncertain, the impact assessment is valid, and the "facts"
presented by Chile are disputable as described in Section 3.2.2.
Chile claims in their reply that DH l 's impact assessment of the 1928 Canalization in
Bolivia being up to 30-40% higher cross border surface flows is unrealistic and based
on an unsuitable and fundamentally flawed numerical model. Chile 's argumentation
consists of two parts; the conceptualisation of the Near Field model together with the
water balance and the interpretation of the local hydrogeology.
In section 3.2.1, and as a first step, we present international evidence that in general
artificial drainage of wetlands is known for having the impacts shown by our model in
Si/ala and some are also detected visually in the field. So, the model calculates the
right types of impacts.
In section 3. 2. 2, we review the Far Field analyses and implications for the water
balance.
In section 3.2.3, we address the interpretations of the local hydrogeology and how this
is reflected in the model.
Finally, in section 3.2.4, we address Chile 's criticism of the conceptualisation of the
Near Field integrated groundwater surface water model.
19
3.2.1 The internationally agreed impacts of artificial wetland drainage
The section summarises and references through other cases generally accepted
impacts of drainage networks. These include:
• Lowering of groundwater table
• Drying out the top soils
• Increase of superficial discharge.
There can be no doubt that the creation of artificial drainage canals in the Silala
wetlands has created a change in the hydrological regime of Silala area. These
changes have resulted in enhanced drainage or discharge from the aquifer resulting
from a lowering of the groundwater table within the wetland areas. In turn, this has
increased the surface water flows emerging from these wetlands. The claim7 that the
"works do not have a significant impact on surface water flows" is contradicted in the
scientific literature as described below. Furthermore, it contradicts Chile's own experts
who agree that the ": Construction of drainage channels and river channelization in the
1920s will have had some effect on the flow. An increase in flow due to these works is
expected" 8•
Inspection of the current drained system and the remnants of the original wetland
vegetation provide direct evidence of significant impact by the artificial drainage in the
Silala wetlands (CR, p135, Figure 16).
General facts on artificial drainage
There is considerable evidence that drainage has a detrimental effect on wetlands.
Globally, some 50% of wetlands have been lost since the 19th century as a result of
drainage (Gibbons et al, 2006).
There is also considerable direct evidence that drainage changes the hydrological
regime of wetlands, which is consistent with the physics of the wetland flow processes.
Outflows from wetlands are increased by the construction of drainage ditches,
channels, and canals, or the removal of natural barriers such as vegetation and by
straightening streams (US EPA, 2008). Artificial drainage networks remove water more
rapidly, reducing the flooded area, as well as the duration and frequency of flooding
(Erwin, 2009; Blann et al., 2009). The effect of the alteration of channels and canals
can be two-fold-not only do the new excavations convey more water out of the
wetland, the spoils may concentrate or otherwise alter the natural drainage through the
wetland (Chabreck 1988). Ditches and tile drains increase the discharge of shallow
groundwater, thus lowering water tables in the vicinity of the drains (Whiteley, 1979; US
EPA, 2008). This produces drier conditions within the wetland and in turn can change
the land cover (vegetation and open water) (South et al. , 1998; Finlaysen et al., 2005;
Kadlec and Wallace, 2009) and the evapotranspiration (ET). In natural undisturbed
wetlands, water tables tend to remain close to the surface and water table fluctuations
7 Vol 1 para 1.7
8 Wheater, H.S. and Peach, D. W. (2019) Impacts of Channelization of the Silala River in Bolivia on the Hydrology of the
Silala River Basin, Attachments to Reply of the Republic of Chile to ICJ on Dispute over the status and use of the waters
of the Silala., p101.
13
20
are generally limited (Evans et al., 1999; Lapen et al. , 2000, Holden and Burt 2003;
Holden et al., 2011 ).
In drained wetlands, the groundwater table is lowered at the drains and local
groundwater gradients near the drains are altered . Groundwater discharge from the
wetland is enhanced resulting in a lower average groundwater table in the wetland.
This is consistent with the physics of hydrological processes in wetlands and with direct
field evidence in monitored wetlands, (Luscombe et al, 2016, Holden et al., 2011 ).
Holden et al (2006) observe significantly lower overland flow in the drained catchments
and throughflow was more dominant. Chile's own experts agree "that there may have
been some changes to river-groundwater interactions"9•
In summary, the discharge will be increased by artificial drainage of a wetland area
and the ratio between surface water and groundwater discharges will be changed.
Where surface water and groundwater are closely linked (such as in the Silala Springs
system) over-exploitation of surface and/or groundwater would be expected to result in
declines in the spatial extent and health of the original type of wetland and lead to
changed vegetation or degradation of the original ecosystem.
The above general effects of drainage are also reflected in the on-site physical
evidence in Silala including:
• efficient drainage and conveyance via Chilean made canals that dissect and
disturb the Silala wetlands;
• diffuse and scattered inflows concentrated by excavations and drain collection
systems;
• dried out former wetland areas;
• declines in water table elevations;
• declines in soil water content;
• decrease in the flooded area when compared to natural bofedales; and,
• the presence of invasive grasses in the drained soils
The changes in Silala are consistent with the effects of drainage described in the
above-mentioned literature and reports. Changes in flow regimes and flow rates from
drainage must thus be expected from general experience.
9 Wheater, H.S. and Peach, D. W. (2019) Impacts of Channelization of the Silala River in Bolivia on the Hydrology of the
Silala River Basin, Attachments to Reply of the Republic of Chile to ICJ on Dispute over the status and use of the waters
of the Silala ,, p 4
14
21
3.2.2 Regarding the Far Field model and basin water balance
In this section, we discuss the problem of quantifying the water resource in the Si/ala
Basin and why this is important. We also explain our approach to the Far Field model,
its rationale and limitations.
Chile states that: "independently of any modelling efforts, all water in the Si/ala River
basin will flow from Bolivia into Chile, whether as surface water or groundwater
"(CR.3.5 a).
Although this is true in a general sense (as confirmed by terrain and groundwater
gradient in the wetlands and ravines10) , the statement raises some critically important
questions:
1) What is the correct Silala catchment and its area?
2) What is the quantity and sustainability of the basin's water resources?
In wetter basins where the discharge is dominated by surface water, these questions
may be determined through surface flow observations, climate data and topography.
This is not the case in Silala where the springs and canals are fed principally by
groundwater from a Far Field much larger than the topographical catchment 11 •
With the purpose of quantifying and understanding the basin's water resources, a water
balance model of the plausible Far Field (also based on topographic divides) was
established. The Far Field model is a distributed water balance model with a detailed
infiltration model designed to assess the recharge to groundwater. The model is
integrated with a coarse and simplified groundwater model based on the very sparse
hydrogeological mapping information available on the Far Field area.
The model analysis of the Far Field led to an understanding of which components of
the water balance can be estimated, the reliability of these estimates and the
implications for making quantitative assessments in Silala. The method adopted also
provided an independent estimate of groundwater residence time.
The Far Field analysis showed:
• Due to unknown hydrogeological aspects, it was not possible to delineate
correctly the spatial extent of the aquifer supplying groundwater to the Silala
springs. Therefore, the basin area and boundary; the hydraulic properties and
thicknesses of aquifers; as well as the transient changes in groundwater
volumes and heads are all unknown.
• A first order estimate of the recharge, using a plausible hydrological (Far Field)
catchment was made (i.e. aquifer boundaries are equivalent to the hydrological
catchment). The resulting recharge from this area was on the lower end of the
estimated cross border flows (groundwater and surface water combined).
• Available data suggests that recharge rates in the Far Field range between 21
mm/year and 374 mm/year, which if the aquifer is steady-state would result in
discharge rates of 151-37 4 1/s. Application of available hydro-geologic data to
10 Bolivian Contra Memorial (BCM), OHi 2018
11 Bolivian Contra Memorial (BCM), OHi 2018, Annex A.
15
22
D~
estimate combined cross border groundwater and surface flow results in
calculated values of at least 260-440 1/s.
• The estimated recharge may explain a large proportion of the water supply to
the Silala wetland but the uncertainty in the climatic conditions, the groundwater
flows in the Far Field and transboundary groundwater flows mean we do not
know with precision the magnitude of the groundwater resource providing water
to Silala springs.
• It is also plausible that either the contributing area may be substantially different
or that other sources of water, such as non-renewable water may contribute to
the Silala springs.
• The underlying assumption in the first order recharge estimate is that there is a
balance between inflows and outflows. The recharge estimates are based on
recent climate conditions, whereas the parts of the water emerging in Silala now
was recharged under the climatic conditions prevailing thousands of years ago,
which were considerably different.
In their response, Chile refers to "all the water"12 but do not recognise that this is, in
fact, an unknown quantity.
Regarding the modelling approach taken by DHI
A number of idealisations and assumptions are made as the starting point for the later
conclusions by Chile that DH l's Near Field model produces erroneous results.
We do not oppose the statements made by Chile on the preference for water balance
preserving groundwater models with fixed upstream boundaries and "known" recharge
input (similar to the Far Field model). This concept was considered -and rejected due
to the lack of hydrogeological data in the Far Field and the inability to define the vertical
and lateral boundaries of the aquifer. Development of a groundwater model of the Far
Field would have required many assumptions regarding the geologic framework,
hydraulic conductivity, specific yield, specific storage, effects of faults on groundwater
flow patterns, and recharge rates in areas that remain completely uncharacterized.
OHi determined that the available field data (none) did not justify detailed groundwater
flow modelling in the Far Field. And that attempt to do so would have to be based on
several assumptions that would be difficult to verify and would be highly uncertain.
Estimates of groundwater recharge using an aquifer extent equivalent to the
topographic catchment results in aquifer discharge rates that are lower than those
estimated from field data. As argued above it is also plausible that either the
contributing area may be substantially different or that other sources of water, such as
non-renewable water may contribute to the Silala springs. This introduces considerable
uncertainty into the concept that the aquifer is in an equilibrium conditions within a
catchment with closed upstream boundaries.
12 CR Vol 1, Para 1.10, Para 3.33
16
23
If, in theory, the surface water and aquifer are accurately mapped, and the recharge
rate is constant in time, then the groundwater system will reach a steady-state condition
in which the total recharge is equal to the total discharge.
However, the potential exists for inter-aquifer inflows into the Silala aquifer, as
evidenced by high relative groundwater ages and discharge rates that are impossible to
explain based on modern recharge. This suggests that either there are exchanges of
groundwater with aquifers outside the assumed groundwater basin or there are longterms
trends in the subsurface storage associated with long-term climate trends that
may contribute to discharge rates to the Silala.
The vast majority of the groundwater discharging to the springs and collected in the
canals is much older than the canals themselves. The groundwater flow regime is
therefore not necessarily in equilibrium or steady-state conditions. Chile provides no
evidence that the aquifer is in a steady-state condition (aquifer inflows equals the
outflows) and their assertion that the aquifer is in a steady-state condition is not
supported by available site data or mass balance calculations.
Since the impacts to be assessed stem from the canals implemented in the wetlands
and ravines, the largest impacts are assumed to be found closest to the wetlands.
Therefore, the approach taken was to collect hydrogeological data within and in the
vicinity of the Near Field and confine the Near Field model to the area over which the
available hydrogeological data may reasonably be extrapolated. The downside of this
approach is that the flow through the system is less well defined due to the open
boundaries.
Comments regarding DHI table of results
OHi has emphasised that the impact assessment of removal of the canals is uncertain,
as also acknowledged by Chile's expert. We also agree that, under circumstances
where the model boundary conditions are located close to the hydraulic stress being
simulated (i.e. draining of the system by canals), the results may be sensitive to
boundary effects. A sensitivity analysis would improve our understanding of these
effects and provide greater certainty in the range of potential wetland impacts relative to
an undisturbed state.
3.2.3 The underlying interpretation of the geology and hydro-geology
Chile argues that: "the model is built on an incorrect interpretation of the geology and
hydrogeology"
More specifically they argue "DH/ assume in their model a distribution of high hydraulic
conductivity in the region of this assumed fault system that has no basis. " Pg. 201
Chile further claims that the model "does not represent the geology correctly either
stratigraphically or structurally and invokes a fault system that is both unmapped and
geometrically unlikely. " Pg. 201.
This is not correct. The conceptual hydrogeological model implemented in the
integrated Near Field model is based on detailed on-site mapping by Bolivian
geologists who have identified a highly fractured zone along the ravines the Carones
17
24
D~
18
and Orientales wetlands and above the latter. Sergeomin (2017) geologic mapping
indicates a relatively small displacement of 5 m at the border (Figure 1 ). This is not
visible at the scale of the conceptual hydrogeologic cross section (Figure 36, Annex F),
but a small displacement can be added to an updated figure to improve the conceptual
aspects of the figure.
]__
CORTEG-G"
- s
Figure 1 Geologic cross section near the Chilean border (Sergeomin, 2017)
As discussed in the report and as is routinely done in hydrogeology, the geologic units
were lumped into hydrogeologic units based on hydraulic properties. The inclusion of
HGU7 (Silala fault zone) as a unique hydrogeologic unit within the hydrogeologic
framework model is supported by:
1. Geologic maps (Sergeomin, 2017) that includes a mapped fault along the main
ravine near the border that splays to form the ravines that extend into the
northern and southern wetland areas (Figure 1 ).
2. Diamond core drilling with highly fractured and brecciated ignimbrite in the
areas near the mapped faults.
3. Pumping test results from Bolivia (DS-4P), located within the area of the
mapped fault, that yielded hydraulic conductivity estimates between 14 mid and
138 mid. Late-time data results from this test which have hydraulic conductivity
results that are 50% lower than early-time, suggesting decreases in hydraulic
conductivity at greater lateral distances from the pumping well.
4. Pumping test results from Chilean tests near the border with an estimated
hydraulic conductivity of 6.5 mid.
5. Geometric mean value for hydraulic conductivity from 19 slug tests within
the fault zone (HGU7) of about 7.5 mid.
6. Geophysical surveys (COFADENA, 2017) measure low resistivity bedrock,
interpreted as a zone of brecciated rock associated with faulting along the
ravines and Sergeomin (2017) mapped faults.
The scientific evidence provided by points 2 - 5 above prove the existence of highly
fractured ignimbrite with high hydraulic conductivity. This point cannot be credibly
disputed.
The mapped fault is consistent with the observed displacement between units on either
side of the ravine as shown in Figure 2 and the findings of points 2 - 6. However, OHi
25
concedes that the precise geometry, transect and width of brecciated rock has not been
perfectly determined. The width was approximated based on the geophysics and the
ravine geometry, which is reasonable to capture groundwater flow patterns and fluxes
through the Near Field.
Figure 2 Geologic map of the Near Field area (Sergeomin, 2017)
Assuming (contrary to Bolivia's assessment) that these high hydraulic conductivity
values should not be confined to a relative narrow highly fractured zone along the
ravines would imply that they are representative for the ignimbrite layer as a whole. The
implication of which, would be significantly greater transborder groundwater flow than
estimated by OHi. Such transborder flows would be difficult to justify from the water
balances using the delineated hydrological catchment as the aquifer boundaries.
Higher transborder flows would require either a much larger aquifer extent, inter-aquifer
flows and/or that the flow system is experiencing changes in subsurface storage
associated with long-term declines in groundwater recharge.
Hence, there is strong technical basis for the hydraulic parameters used in this zone
from both Chilean and Bolivian tests. To claim that its characteristics is unsupported by
evidence is not accurate.
Chile further claims that the model "does not represent the geology correctly either
stratigraphically or structurally .. . " Pg. 201
Regarding the stratigraphy, hydraulic testing has demonstrated that the hydrostratigraphic
layers of various ignimbrites have similar hydraulic conductivity values,
varying over less than an order of magnitude, except when within the Silala fault zone.
At the scale of the Near Field analysis and considering the objective of assessing
groundwater flow patterns and flow rates, as well as groundwater - surface water
interactions, the small-scale stratigraphy arguments are not relevant.
The concept of a Representative Elementary Volume (REV) is applied to capture the
hydraulic behaviour of the rock mass. The REV is the smallest volume over which a
measurement can be made that will yield a value representative of the whole and has
to do with the well-known scale effects of hydraulic measurements in hydrogeology.
Extensive hydraulic testing suggests that the ignimbrite REV is on the order of 100's of
19
26
D~
20
meters in the horizontal dimension or volumes greater than at least 10,000 m3
• As a
result, small scale, laterally discontinuous stratigraphic features are unlikely to exert
strong controls on groundwater flow patterns, or groundwater discharge relationships at
the scale of the Silala Near Field analysis. The calibrated Near Field Model is evidence
of this, in that it utilises an Equivalent Porous Media (EPM) approach and effectively
captures the groundwater head distributions and discharges to various sectors of the
Silala springs.
Chile claims that "DH l's modelling takes no account of the vertical variability of
permeability, as demonstrated by the artesian flowing conditions at Chilean borehole
SPW-DQN". Pg. 201.
Chile's statement is incorrect since both the conceptual and numerical models account
for both vertical anisotropy and changes in horizontal hydraulic conductivity with depth.
Interpretation of the hydraulic testing data led to vertical anisotropy estimates between
1:10 and 1 :50 (Kv:Kh), as reported in the conceptual model. These values were initially
used in the numerical model and calibrated vertical anisotropy was included for some
units. Furthermore, the horizontal hydraulic conductivity decreases with depth by at
least one order of magnitude.
Chile claims that "DH/, in their models, do not consider the downstream Chilean
geology, including the controlling influence of the faulting or the presence of low
permeability Pliocene dacitic lavas overlying the Cabana lgnimbrite and underlying the
Chilean Si/ala lgnimbrite. " Pg. 201
This may be a reasonable comment but the mandate given at the time of constructing
the hydrogeological model was to stop it at the border. Although not proven with
sufficient hydrogeologic data, it is plausible that both decreasing ignimbrite thickness
and underlying low permeability lavas, as well as the hydrogeologic controls associated
with the Cabana Fault, might serve to reduce down-gradient transmissivity of the
ignimbrite aquifer.
However, the transborder flows, groundwater flow patterns and groundwater-surface
interactions in the Near Field are all calibrated to actual measurements of:
• spring flows by spring sector;
• canal discharge measurements;
• groundwater heads; and,
• hydraulic conductivity and storage parameters.
This includes measurements made near the border with Chile. Because the calibrated
values match measured values near the border and Near Field more generally,
downgradient effects are implicitly incorporated in the Near Field Model. For example,
examination of Figure 40 (Annex F) indicates that measured horizontal hydraulic
gradients increase towards the border. This increase may be related to increases in
ground surface slope or it may be associated with downgradient changes in
transmissivity (as proposed by Chile), or both.
27
D~
Therefore, downgradient hydrogeologic conditions that affect the border area would be
reflected in the gradient boundary condition applied at the border.
Chile claims that the Bolivian stratigraphic column is incorrect based on their age dating
and observed stratigraphic relationships. The principal implication of this disagreement
is whether the Lavas Si/ala Chico and lnacaliri 1 are intruded ignimbrites or whether
they intruded older dacitic and andesitic lavas. These relationships will exert controls
over the spatial extent and anticipated depths of the ignimbrite aquifer.
OHi has not acted as Bolivia's geological expert on Silala and cannot comment on
Chile's stratigraphic arguments. However, this discussion is not pertinent to the
hydrogeological model as this narrow domain would have ignimbrite in both
interpretations.
Chile states that "The thickness of these ignimbrite deposits is unknown but likely to be
large, perhaps over 200 metres, but this has not been proven by drilling."
It remains true that the vertical extent of the ignimbrite aquifer remains unknown.
However, if the Bolivian stratigraphic interpretation is correct it is reasonable to assume
that the thickness of ignimbrites may be quite large and on the scale of the Near Field
Model thickness of 400 m. If the thickness is less than assumed and the ignimbrite
aquifer is underlain by lower permeability lavas, the implication of this would be smaller
transborder groundwater flow. However, the hydraulic conductivity of these layers is
already significantly reduced in the model (i.e. below 200 m) and the effects on
simulated shallow groundwater-surface interactions would be negligible. Furthermore,
analytical estimates of transborder flow are based on a very conservative ignimbrite
aquifer thickness of 117 m (proven depth from drilling).
In summary, the stratigraphic discussion is not pertinent to the hydrogeological model
where both interpretations coincide. As the conductivity of the model ·s ignimbrite layers
decreases with depth, the impact of a possible change is deemed not to be important
for the canal impact assessment.
Chile finally claims that "the differences between the chemistry and isotopic
compositions of Bolivia 's Cajones and Orienta/es spring waters indicate different
origins, which, though accepted by OHi in their report (BCM, Vol. 4, p.94), have not
been represented in their modelling" Pg. 202
OHi acknowledges that local inflow from more local flow regime, including perched
aquifers in the Cajones wetland is a possibility. However, hydrogeological
characterization activities close to the wetlands have not detected perched aquifers.
Furthermore, OHi clearly concluded that it is likely that there are two primary and
distinct sources of groundwater discharging to the Silala springs.
However, the objective of the integrated Near Field Model is not to determine the
source of the discharging water but the change in surface water / groundwater
discharge from the wetlands due to the canalization (a near surface intervention).
Hence, it has not been the intention to represent the various sources of groundwater
explicitly in the model as it would not reflect on the split between surface water and
groundwater discharge from the Si/ala Wetlands.
21
28
22
3.2.4 Disputes regarding the Near Field numerical model
Chile claims that "DHl's modelling results are entirely dependent on magnifying the
impact of the channels by modelling just 1 % of the relevant area (the Near Field)"
(3.5.c).
However, it must be recognised that the effects of canals occur primarily in the 1 %
area and not the remaining 99 %. Furthermore, the information on hydrogeological
characteristics in the remaining 99%, is almost non-existent, and inclusion of this large
area would introduce other uncertainties as described previously. Hence, we maintain
that a local scale model of the Near Field is the most appropriate and reliable method to
assess the impacts of channelisation on the wetlands and local groundwater-surface
water interactions.
The Near Field model was built to focus on the impact assessment of the canals with
reference to available data. The Near Field model includes both a groundwater model
based on field measurements and a detailed description of surface features including
springs, wetland and canals. The integrated model considers the interaction between
surface water and groundwater. The Near Field model boundaries are located in the
area where the groundwater heads can be assessed from the borehole and spring
observations.
• The model includes the entire drainage network and the springs which is the
primary impact area with respect to removing canals
• Along the upstream boundaries of the model, head boundary conditions based
on measured water table data were applied. While a constant gradient was
applied along the model's downstream boundary (the international border).
• Applying boundaries located far from hydrogeological observations in the
ungauged Far Field would also be inherently uncertain -if not speculative.
The Near Field model is, as opposed to the Far Field model, calibrated to measured
groundwater heads and groundwater discharge to the various spring areas. Calibrating
a Far Field model would require extensive speculation on hydrogeological properties,
aquifer extent, aquifer thickness, inter-aquifer flow etc., for which we have almost no
data. Instead, measured head and flow values were relied upon and the calibrated
Near Field model robustly captures these measured values.
DH l's analysis has focused on impacts of the canals in an integrated hydrological
context based on data collection in the Near Field. The catchment water balance
remains poorly defined and may be affected by transient changes in head and storage
associated with changes in long-term trends in climate over millennia.
The Near Field model and the Far field model are two separate models developed with
different and distinct goals. Hence, the Near Field surface water- groundwater model is
not linked to the Far Field water balance model and the Near Field model is not meant
to close a larger scale catchment water balance.
We do not claim that the Near Field model includes the total groundwater flow across
the border. The aquifer(s) are likely to extend into Chile in a cross section wider than
the downstream model boundary and it does not include the aquifer conditions on the
Chilean side of the border.
29
The groundwater table and flows at the near field , control the flow to the wetlands and
the spring discharge. A closed water balance approach would be an idealisation that
cannot be supported with respect to groundwater as groundwater flow divides cannot
be delineated without more information on aquifers and water tables in the Far Field.
It is evident that groundwater inflows feed and maintain the spring and canal flows. In
the Near Field model groundwater flows are closely tied to groundwater boundary
conditions. OHi acknowledges that the model results, including the impacts of the
drainage network, are most likely sensitive to boundary conditions. Sensitivity and
uncertainty analysis have not been conducted as part of the Near Field modelling work,
but such analyses could be beneficial in improving our understanding of these effects
and provide greater certainty in the range of potential wetland impacts relative to an
undisturbed state.
Any groundwater model is an approximation both with respect to its resolution, its
process description, its parameters and its boundary conditions. This also applies to the
Near Field model. The model area extent, the boundary type and the data used to
describe the horizontal and vertical distribution are based on interpolation, assumptions
and generalisations which introduces uncertainty. With the boundary conditions applied
the hydrogeological parameters are adjusted with the objective of simulating the canal
flows collecting upstream spring discharges and diffuse seepage inflows. Expanding
the model area implies formulating another boundary condition approximation and
covering an area for which no data exists. This does not resolve the uncertainty issue.
Chile's comments on DHl's table of model scenario results.
Chile says that the scenario comparison shows nothing.
We will maintain that it does but do also emphasise that the results are uncertain and
that a sensitivity analysis should be used to quantify the uncertainty and may help to
reduce it.
The major water balance components of the Near Field model runs are presented. The
baseline model (with canals), the 'no canal' scenario and the 'restored wetlands
scenario' have different inflows.
Chile argues that with a closed water balance and constant recharge in equilibrium with
groundwater the inflow to the Near Field area should be the same in all three scenarios.
Based on field data OHi does not find basis for assuming that the recharge and inflow
should necessarily be the same in all scenarios. On the contrary, feedback from
removing canals on the groundwater flow changes the total inflow and rate of
groundwater discharge to surface water. Only under the assumption that all upstream
water must run through the Near Field Model will this be the case but this assumption is
unlikely and is not verified from by the available field data. However, the same inflow in
all scenarios would definitely simplify the scenario comparison and this could be further
analysed through a sensitivity analysis.
While the physical aquifer and surface water system represented in the Near Field
Model cannot be considered a closed system, for each of the model runs , conservation
of mass, in the model itself, should apply. The Near Field model is run as a dynamic
23
30
24
model but with steady state boundaries. Storage end error terms are negligible (0 - 5 %
of the total inflow) for the model runs. Consequently, the effects predicted by the
numerical model are not to any appreciable degree influenced by the numerical
imbalances or errors. Adding the water balance components applying the correct sign
(inflows versus outflows or losses) shows that the water balance adds up within a
deviation of 3 % of inflow on the Baseline model and 1 % on the scenarios without
canals. Unaccounted flow is considered well within uncertainty bounds and does not to
any significant degree change the impact assessment.
DHI does not claim that the Near Field model represents a closed water balance
system and flow conditions are not transferred from the coarser Far Field model to the
Near Field model. That would be incorrect, given the uncertainties, especially in
subsurface conditions. A head boundary is an open boundary with inflows depending
on internal model groundwater heads and it does not close the flow system. With higher
groundwater heads internally in the Near Field model less water enters the model
domain through the fixed head boundaries as the gradient changes.
With no canals, less water enters the surface water system and more enters the
groundwater inside or outside the Near Field model domain. The differences in inflow
between the scenarios reflect changes in total inflow caused by changed groundwater
gradients and flows. Groundwater entering the Near Field or flowing past the Near Field
domain will likely flow into Chile.
DHI has emphasised that the impact assessment of removal of the canals is uncertain,
as also acknowledged by Chile's expert. We also agree that under circumstances
where the boundary is located close to the interventions and therefore may be affected
by them. Sensitivity analyses may improve the uncertainty assessment of the impacts,
and they may be reduced.
Why Chile's simplified impact calculations are not valid
When Chile claims to have proven the surface flow impacts from the canals to be small
- it is simply not correct.
Differences in flow between two 1-D Darcy profiles located within the Near Field and
Far Field model area are used to suggest that effects of removing canals are
exaggerated. A ratio between Darcy profiles flow estimates is carried forward by Chile
as a quantitative measure of overestimation by the Near Field model.
For a number of reasons this approach is not valid:
On page 51 Chile writes, "Hence, both Chile's and Bolivia's experts confirm the
complex nature of the Silala River and groundwater flow systems." and yet an idealised
hillslope element is considered suitable for flow calculation. This is inconsistent.
The one-dimensional Darcy equation is valid only under idealised conditions which are
not satisfied at Silala:
1) The groundwater aquifer is not homogeneous.
2) The groundwater flow is not one-dimensional rather it is highly threedimensional.
3) The aquifer is both confined and unconfined.
31
To calculate plausible water levels and water level gradients under the conditions in
Silala, the Darcy equation has to be solved in three dimensions for a large number of
elements each of which can be assumed to be homogeneous and to which realistic
properties can be assigned. This cannot be done in a simplified hand calculation. This
is, however, exactly what is done in the mathematical models, used by OHi.
Furthermore,
Chile's Darcy profiles lack reference to observed field data.
Chile has used the Far Field model for extracting groundwater table
gradients in comparison to the Near Field model which is a
misunderstanding and incorrect. The Far Field model, as described by
OHi, is developed and used for overall water balance and recharge
calculations only and it is based on very limited field data, no
hydrogeological data, uncertain catchment boundaries. Since it has not
been calibrated against any groundwater data, the simulated groundwater
heads cannot be used for quantitative assessments and were only
included in the report to illustrate possible gradient and flow directions.
Consequently, the groundwater levels extracted from a figure are not a
reliable measure to be used in this context.
It is also clear that profiles could have been picked in other locations
which each would have different gradients, hydraulic conductivities, flows
and impacts of canals.
Comparing two single, random transects based on an overly simplified, idealised
method can under no circumstances be translated, directly nor relatively, to outputs of
the 3-D integrated Near Field model results.
Figure A5 (CR Vol.1. page 149), although overly simplified, illustrates the inconsistency
in Chile's proposed Darcy profile approach. At the Near Field multiple layers and
combined horizontal and vertical gradients and flow directions are depicted. They are
not considered by a Darcy hillslope transect.
It is noted that Chile, despite the inadequacies of this approach, calculate an effect of
the canals on groundwater inflow. The effect is, however, much lower than the
estimates derived from DHl's Near Field model. The Darcy profile flow estimates are
presented in the unit (discharge per unit width, (m3/day/m). The width along the Far
Field outer boundary is much larger than the Near Field model boundary meaning that
the 1-D profile results cannot be carried forward as a measure to explain the responses
of a 3-D model. This important fact is left uncommented.
Highly calibrated fully integrated three-dimensional groundwater models cannot be
replaced by idealised hill slope estimates, and we can for the many reasons described
above not accept Chile's estimate of changes in groundwater discharge.
The integrated Near Field model is based on field data and includes both groundwater
and surface water. Despite data limitations, it is the most comprehensive and calibrated
tool for assessing canal impacts, and it is considered the best tool for providing
technically sounds estimates to changes in groundwater discharge. As with all models,
its predictions are subject to uncertainty associated with model construction and
parameterisation. However, the limitations associated with the vastly simplified cross
section Darcy flux calculations (as proposed by Chile) are unquestionably greater.
25
32
Additional model sensitivity analysis will improve our understanding about the range of
likely impacts, as constrained by the measured field values.
3.2.5 Analysis of Chile's comparative studies of the Silala wetlands
Chile questions some conclusions of the 2018 report from the Ramsar Convention:
• In 3.37, it is questioned: "That the groundwater system is classified as a nonrenewable
aquifer".
We agree that such statement cannot be based solely on the high age of water
detected in the springs and that DH l's assessment is that a large part of the
cross-border water flow may currently be recharged under the present climate.
Although our sensitivity analyses indicate an aquifer recharge rate of the same
order of magnitude as surface water and groundwater discharge at the border,
there nevertheless remains a discrepancy between the two values.
In other words, it is not proven that the aquifer is in a steady-state condition.
Given the long residence times, it may well be experiencing transient declines in
aquifer storage associated with millennial scale climate changes.
• In 3. 39, it is questioned: "That there are only vestiges of the original wetlands
that used to cover 14.1 Ha The current surface area covers only 0.6 Ha".
We agree that the latter area seems small, but it is not clear to us if it refers to
the remaining undisturbed parts of the wetland, which as indicated by Chile's
Figure 12 (from BCM, Vol 2 p. 333) is much smaller than the canalised area.
Chile, in their argumentation (3.34), seems to use the vegetated area as a measure of
the wetlands not being degraded by the canalisation. The newer satellite studies 13, 14 (as
also recognised by Chile) cannot be expected to capture a deterioration originating
from the artificial drainage implemented around 50 years before the start of the analysis
period. Both studies show a large seasonal variation in the evaporating surface of the
Bolivian drained wetlands while this seasonality was not found in the undisturbed
Quebrada Negra wetland in Chile. Castel correctly associated the seasonality with the
fact that large parts of the drained wetlands are now dominated by invasive grasses
growing on the drained soils and that this vegetation depend on the soil moisture
storage after the wet winter season and dries out during summer. This indicates
another biological and hydrological regime than that the original undrained wetland with
ponding stagnant water all year round.
Hence, Chile's quoted satellite studies support that the drained wetlands have changed
(and potentially biologically deteriorated) as a consequence of the drainage.
We cannot agree with Munos and Suarez' conclusion that" the canalisation activities in
Bolivia 's wetlands, has not significantly affected the area of the active wetlands in the
valley floors ".
On the contrary, they detect a larger evapotranspiration seasonality in the Bolivian
wetlands, grasses occupy a larger proportion of the former fully saturated and flooded
13 Castel A. (DIREMAR 2017), CR Vol 2 Annex 98
14 CR Vol 3 Annex XIII Munoz and Suarez, 2019, Quebrada Negra Wetland Study
26
33
wetland. Furthermore, in the Orientales (Southern Wetland), photographs e.g. (CR,
p135, Figure 16) shows signs of old dead wetland vegetation indicating a formerly more
extensive wetland.
Chile's Experts conclude with reference to Munoz and Suarez that the evaporation
rates from the two Bolivian wetlands should be higher than those of the undisturbed
Quebrada Negra and that the drainage should therefore not have any impact on the
evaporation (CR p. 141 ).
We do not agree with this conclusion. The methodology leading to it is based on an
evaporation formula (unproven on the site) that neglects open water evaporation, which
will not be correct for the undisturbed wetland, and which gives actual
evapotranspiration rates only 50% or lower than the potential ones (CR p137 Table 3).
This is considered implausible for wetlands with abundant supply of water and shallow
depth-to-groundwater.
In spite of our disagreements with Chile's conclusions on the wetland areas, state of
degradation and evaporation rates, we do emphasise that basically the two parties
agree on the influence of the wetland evaporation on the cross-border flows being
minor.
27
34
This page is intentionally left blank.
28
35
4 Assessment
The following table contains a long list of arguments from Chile.
It is divided into two parts: first a part on the points Chile states that the two parties
agree upon, which has been included since we feel that this contains some statements
from Chile that are open to dispute.
The second part concerns the argument that the parties (according to Chile) does not
agree upon. In addition to DHl's comments, this part also includes our evaluated
strengths and weaknesses in Chile's argumentation.
In both parts, the first column is a reference to the statement in the main part of CR
while the second column references statement in the Experts reports. The following
table summarises the key arguments presented by Chile. For each subject or item, we
have highlighted if it can be viewed as an agreement or a disagreement to be
contested. DH l's evaluation of items on which we disagree is evaluated shown in the
right two columns.
29
36
This page is intentionally left blank.
30
37
D~ D~
Reference Expert's Chile's response DHI Comments
report
page
number
3.7 Chile and Bolivia agree that the Silala River is a Agreed
perennial flow that rises at two sets of springs in
Bolivia and flows along the natural topographic
gradient from Bolivia into Chile
3.7 The Parties also agree that the channelization in Agreed
Bolivian territory did not alter or divert the natural
direction of the flow of the water from Bolivia
towards Chile
3.7 Thus, the direction of the flow of Silala waters has Perhaps, but this conclusion cannot be derived from the former two
been the same for thousands of years. statements. Flows and rivers change in time over thousands of
years and the drainage directions may well have changed in parts
of the catchment. What matters is the natural and affected flow
situation at present.
3.8 The Parties agree that surface water runoff Correct. This is about the surface runoff from the catchment into
contributes a very minor proportion of the average the Silala wetlands. NOT to be mistaken for the trans-border
daily flow of the Silala River, which is groundwater surface water flow in the canal.
dominated.
3.8 They also agree that the Silala stream interacts Agreed.
with groundwater throughout its course and that
the direction of the subsurface water (as of the
surface water) is westward towards and into Chile.
Bolivia's expert OHi estimates that the groundwater
flow is at least of the same order of magnitude as
the surface flow.
3.9 Chile and Bolivia agree that the channelization Correct.
undertaken in 1928 may have resulted in reduced
direct loss of water to evaporation, due to a
possible reduction of the extent of surface water in
the Bolivian wetlands
31
38
D~
Reference Expert's Chile's response DHI Comments
report
page
number
3.9 Both sides agree that this reduction of evaporation Correct.
is a small component of the total water balance of
the Silala River system.
3.10 The conclusions as to evaporation have been Disagree.
reinforced by recent studies by Chile, in which Any changes in evapotranspiration are important for documentation
estimates of evaporation from Bolivia's wetlands of changes in the wetland and for preservation of the wetland and
(with channelization) are very similar to wetland habitats.
evaporation from a similar wetland in the Silala In terms of the overall water balance and water sharing the
River basin in Chile (which has no channelization). changes in evapotranspiration are small. However it cannot be
This suggests that the effects of channelization on concluded from this, that the effects of channelization on the water
the water balance, if any, are very limited. balance are limited, .
Finally, we do not agree to the conclusions drawn from the studies.
We do not agree that quantitative conclusions can be drawn from
this since:
A. The study compares different wetlands
B. The study does not cover the pre-drainage period of The
Bolivian wetlands.
C. The formula used disregards open water evaporation and
gives evapotranspiration values far below the potential ones,
which is not to be expected for localised areas well supplied
with water.
The study actually shows changed hydrological behaviour of the
drained and undrained wetlands
3.11 Chile's and Bolivia's experts agree that the springs Correct.
in the Orientales and Cajones wetlands in Bolivia
have different isotopic and chemical compositions,
implying different origins and different recharge
areas
32
39
D~ D~
Reference Expert's Chile's response DHI Comments
report
page
number
3.11 Chile's and Bolivia's experts confirm the complex Correct
nature of the Silala River and groundwater flow
systems
3.12 Despite these important convergences between This is not surprising given the sparse data availability.
Chile's and Bolivia's experts, they maintain DHl's assessment of hydrogeology is based on SERGEOMIN's
different interpretations of the geology and field surveys, and extensive drilling and testing in crucial parts of
hydrogeology of the Silala River basin the basin
3.13 Bolivia's proposed succession and dates of The layers used in Bolivia's hydrogeological model are based on
(permeable) ignimbrite and lava deposits in the Bolivia's geological field surveys on site, Chile's own geological
Silala River valley cannot be reconciled with Chile's interpretation from (CM) and extensive hydrogeological drilling and
recent geological mapping, radiometric dating pump test results including Chilean one. Obviously new
results, drilling evidence and pumping test results interpretations made by Chile after the model establishment is not
included.
3.13 This means that the aquifer system in the This means that the recent Chilean interpretation of the
ignimbrites identified in Chile has not been hydrogeology may be subject to dispute between the parties
recognized by Bolivia
3.13 Bolivia infers a massive geological fault system The intensively fractured zones identified in the Bolivian parts of the
that would run from the Orientales wetland to the basin is supported by:
Cajones wetland in Bolivia, bending around and • Intensive geological field surveys
following the line of the Silala River into Chile, that • High hydraulic conductivities by field testing both in Bolivia
Chile's experts consider highly implausible. and in Chile
• The lone existence of the ravines with distinctly different
weathering and fracturing that the surroundings.
3.13 This inferred fault is not evidenced by any This statement is not sufficiently documented by Chile.
displacement of rocks on either side of the river
valley, as would necessarily occur in a major fault
33
40
D~
Reference Expert's Chile's response DHI Comments
report
page
number
These differences in interpretation do not affect Uncertainty does not mean unreliability.
Chile's and Bolivia's common understanding of the A second source of support is also clearly cited in DHl's analysis.
Silala River as an international watercourse. The discharge measurement by Fox before canalisation compared
However, they do affect the reliability of the OHi to the current flow indicate an effect in the same order as the
Near Field model, which is the only source of model.
support for Bolivia's claims for the large effects of
channelization, as discussed in more detail in
section B below. The differences also affect the reliability of Chile's interpretation.
3.5.a Independently of any modelling efforts, all water in Yes. The questions are: what is the basin extent? How much water
the Silala river basin will flow from Bolivia into is there and when will it flow into Chile.
Chile, whether as surface water or groundwater;
RC,p101,1 The Silala River flows naturally from Bolivia to Agreed. However, in spite of serious efforts, the larger hydrological
Chile. The river rises in two sets of springs in basin still remains poorly defined and there is uncertainty about the
Bolivia, which maintain the Cajones and Orientales boundaries, parts of the assumed basin draining to other areas or
wetlands. other areas draining to Silala. Since the water balance (if partly
fossil /or fully renewable) has still not be finally determined any
increase of the discharge may be unsustainable and irreversible.
RC, The river is primarily fed by groundwater and Correct.
p101 ,2 interacts with groundwater along its course to the
border and beyond.
RC, In addition, there are substantial groundwater flows Yes, but the total cross border groundwater flow has not been
p101 ,3 from Bolivia to Chile, likely of an equivalent quantified.
magnitude to the surface water flows.
RC, Construction of drainage channels and river This statement by Chile recognizes that the effect of drainage is an
p101,4 channelization in the 1920s will have had some increase in flow and therefore can be neglected.
effect on the flow. An increase in flow due to these
works is expected.
RC, The impact of drainage on evaporation from the Changes in ET losses should not be ignored. See previous
p101 ,4 wetlands is small. comments for reference 3.10. 'Small' as compared to what?
34
41
D~ D~
Reference Expert's Chile's response OHi Comments
report
page
number
RC, The Silala is an
p101 ,5 international watercourse
p. 102 We have shown that the model is based on The Near Field model is based on geological and hydrogeological
incorrect geology, interpretations of the best information available at the time and
combines the geological mapping with facts from hydrogeological
boreholes and testing. I.e. evidence from the field .
p. 140 Bolivian estimates of a 30-40% effect on flows are Uncertainties in general and in the boundary conditions in
implausible. particular affect the estimates. The 30-40 % range is range of the
results but not indicative of the full uncertainty range.
Calculations show that incorrect assumptions of
the model's boundary conditions lead to an The factor referred to is derived from idealised Darcy profile
overestimate of the impacts, by a factor of approximations cannot be used for any absolute or relative
approximately 20. measure of impact or accuracy of the Near Field model.
OHi has stated the large uncertainty in the results
3.5.b The inflow in each scenario modelled by OHi is The attempt to argue that a closed water balance can be adopted is
different, causing the outflow in each scenario to not valid for the Silala catchment. Such generalised and invalid
be different as well, proving nothing about the assumptions will conceal the actual uncertainties with respect to
impact of channelization; catchment delineation, groundwater flow and recharge.
3.5.c DHl's modelling results are entirely dependent on The effects of canals occur primarily in the 1 % area and not the
magnifying the impact of the channels by modelling remaining 99 %. Including larger areas introduces other
just 1 % of the relevant area (the Near Field); uncertainties.
On the wetland changes and deterioration
3.5.c Bolivia relies on a 2018 Report of the Ramsar It seems that the areas in the Ramsar report are not reflecting the
Convention Secretariat and its contentions that the full wetland - can it be the undisturbed wetland?
wetlands in Bolivia are severelv deteriorated,
35
42
D~
Reference Expert's Chile's response DHI Comments
report
page
number
however this is contradicted by Bolivia's own 2017
Castel study and a Ministerial report
p. 102 New studies based on detailed monitoring of an The satellite remote sensing studies referred to commence around
undisturbed Chilean wetland within the Silala 50 years after the implementation of drains, hence they do not
basin, coupled with high resolution remote sensing reflect the change in wetland area due to drainage.
data, show that Bolivian and Chilean wetlands
continue to fully occupy the valley floor, and
seasonally extend up the base of adjacent
hillslopes.
p. 102 The condition of the wetland vegetation, as The evaporation formula used
indicated by remote sensing, is similar in all three a) Neglects open water evaporation
wetlands. Associated estimates of actual b) Results in actual evaporation around 50% of the potential
evaporation suggest that the higher evaporation values. Not to be expected in wetlands fully supplied with
rates are observed from the Cajones and water.
Orientales wetlands, some 10% greater than that c) Local wind conditions are important
of the undisturbed Quebrada Negra wetland.
rn-- · 1,.n.,.?_ IA H t ,I,.P_,.~.,.c..:_:,.t_ f,,r,.n_,m,,, t._h,,,P._, .c.,:.,:_~. ._t.P.,..l,l,.i._t.P_, n....~....t..~.... , iH t .~..n. tr"n't":.l.._~. r...c. :,:._:., t._h,,.~_.t._ t._h,,.P,_,r,..P_, .n....i c, ..:..:......~ ::n:,,r ..P....P.... .
has been no significant change in evaporation Impact from evapotranspiration approximately 2% of the crossassociated
with the channelization of the Bolivian boundary flows.
wetlands, and hence no effect of evaporation
changes on river flows .
There are effects in the evapotranspiration shown in the satellite
data. The seasonality of the plant cover is higher (factor 3) in the
drained wetlands than in the undisturbed wetland (factor 1.6).
Invasive grasses dominate the drained not the undisturbed
wetland.
36
43
D~
5 Technical conclusions
The groundwater discharging at the Silala Near Field (covering springs and
canals) is associated with an ignimbrite aquifer of unknown spatial and vertical
extent, with uncertain hydraulic properties and an unquantified volume of water in
storage.
The groundwater water emerging in the Silala springs has high relative ages
indicative of long residence times in the aquifer, potentially up to many millennia.
As a result, changes in aquifer storage over time cannot be ruled out. By
extension, this means that the inflows and outflows to the aquifer may not be
equivalent and that a portion of the discharging waters may be non-renewable or
fossil water.
Chile recognises that the drainage network and canals have increased surface
flows but downplay their significance as 'negligible'. The claim that the "works do
not have a significant impact on surface water flows" is contradicted by the onsite
field evidence, the scientific literature and by Chile's own experts. The exact
magnitude of impacts remains an unresolved issue, but it is clearly greater than
the 1.2% Chile purports.
OHi acknowledges that the Near Field Model lateral boundaries are close to the
interventions and therefore may be affected by them. Sensitivity analyses could
be considered to address these uncertainties.
The validity of Chile's simplified impact calculations is questionable and therefore
does not support the claim that DHl's impacts are exaggerated. The analysis is
based on the one-dimensional Darcy equation, which is valid only under idealised
conditions not satisfied at Silala. The groundwater aquifer is not homogeneous
but heterogeneous and anisotropic, as suggested by Chile in other sections of
the response. Groundwater flow is not one-dimensional but highly threedimensional
and the aquifer is both confined and unconfined. The onedimensional
Darcy approach, in particular, fails to account for the threedimensional
groundwater flow pattern towards the Silala springs and the change
in groundwater gradients towards the discharge zone.
Regarding the geology and the three-dimensional hydrogeological framework
model developed by OHi , Chile finds no basis for incorporating a fault zone and
further claims that the model does not account for vertical anisotropy. This is not
correct. Chile claims DH l's hydrogeological framework model does not reflect
Chile's latest geological interpretations. However, these are generally distant
from the area of interest of the model and unlikely to affect the hydraulic
properties at the REV15 scale or model results regarding groundwater discharge
to the springs and canals or transborder flows.
DHl's conceptual model proposes that groundwater discharging to the Southern
and Northern wetlands have differing residence times, general chemistry and
isotopic compositions associated with different flow paths within the aquifer. Both
15 Representative Elementary Volume, the smallest volume of a porous media over which a measurement can be
made that will yield a value representative of the whole
37
44
D~
J-38
a shorter more local flow path and a longer, more regional flow path were
conceptualised. Chile's interpretation is largely similar, with the exception that
they conclude the shorter flow path is associated with a perched aquifer system
that has not been encountered in the Bolivian territory.
Chile's assertion that the numerical model does not include the two differing
sources of water, stems from the assertion that both the local and regional flow
regimes are conceptualised to have origins outside of the Near Field.
Furthermore, the intent of the Near Field model was to simulate effects
associated with canalization and Chile's modifications to the natural system - not
the origin of the spring waters.
We do not oppose the statements made by Chile on the preference for water
balance preserving groundwater models with fixed upstream boundaries and
"known" recharge input. This concept was considered and rejected due to the
lack of hydrogeological data in the Far Field and the inability, based on available
data, to define correct catchment and aquifer boundaries. A groundwater
calibration of the whole Far Field area to match the few data in the Near Field
would have to be based on a large number of assumptions about the presently
uncharacterised areas of the aquifer.
The approach taken was therefore to collect hydrogeological information within
and in the vicinity of the Near Field and confine the Near Field model to the area
where hydrogeological information is available. The downside of this approach is
that the model area does not include groundwater in the full width of the aquifer
and the inflow to the system is not fixed but a function of the boundary condition.
review of chiles reply - v3.docx I BG/ 2019-02-07
45
6 References
/1/ Blann , K. L., Anderson, J. L., Sands, G. R. , & Vondracek, B. (2009). Effects of
agricultural drainage on aquatic ecosystems: a review. Critical reviews in
environmental science and technology, 39(11 ), 909-1001 .
/2/ Chabreck, R. H. (1988). Coastal marshes: ecology and wildlife management
(Vol. 2): U of Minnesota Press.
/3/ Erwin, K. L. (2009). Wetlands and global climate change: the role of wetland
restoration in a changing world. Wetlands Ecology and management, 17(1 ),
71.
/4/ Evans, M., Burt, T., Holden, J., & Adamson, J. (1999). Runoff generation and
water table fluctuations in blanket peat: evidence from UK data spanning the
dry summer of 1995. Journal of Hydrology, 221(3-4), 141-160.
/5/ Finlayson, C., D'Cruz, R., & Davidson, N. (2005). Millenium ecosystem
assessment. Ecosystems and human well-being : wetlands and water.
Synthesis. World Resources Institute, Washington, DC.
/6/ Foley, J. A. , Defries, R. , Asner, G. P. , Barford, C., Bonan, G., Carpenter, S.
R. , ... Gibbs, H. K. (2005). Global consequences of land use. Science,
309(5734), 570-574.
/7/ Gibbons, J. W. , Winne, C. T., Scott, D. E., Willson, J. D., Glaudas, X. ,
Andrews, K. M., ... Tsaliagos, R. N. (2006). Remarkable amphibian biomass
and abundance in an isolated wetland: implications for wetland conservation.
Conservation Biology, 20(5), 1457-1465.
/8/ Holden, J., & Burt, T. P. (2003). Runoff production in blanket peat covered
catchments. Water resources research, 39(7).
/9/ Holden, J., Evans, M., Burt, T., & Horton, M. (2006). Impact of land drainage
on peatland hydrology. Journal of Environmental Quality, 35(5), 1764-1778
/10/ Holden, J., Wallage, Z., Lane, S., & McDonald, A. (2011). Water table
dynamics in undisturbed, drained and restored blanket peat. Journal of
Hydrology, 402(1-2), 103-114.
/11 / Kadlec, R.H. , & Wallace, S. (2008). Treatment wetlands: CRC press.
/12/ Lapen, D. R., Price, J. S., & Gilbert, R. (2000). Soil water storage dynamics in
peatlands with shallow water tables. Canadian Journal of Soil Science, 80(1 ),
43-52.
/13/ Luscombe, D. J., Anderson , K., Grand-Clement, E., Gatis, N., Ashe, J.,
Benaud, P., Brazier, R. E. (2016). How does drainage alter the hydrology of
shallow degraded peatlands across multiple spatial scales? Journal of
Hydrology, 541 , 1329-1339
/14/ Phillips, B., Skaggs, R. , & Chescheir, G. (2010). A method to determine
lateral effect of a drainage ditch on wetland hydrology: field testing.
Transactions of the ASABE, 53(4), 1087-1096.
39
46
/15/ South, C., Susan, C., Grimmond, B., & Wolfe, C. P. (1998).
Evapotranspiration rates from wetlands with different disturbance histories:
Indiana Dunes National Lakeshore. Wetlands, 18(2), 216-229.
/16/ Sun, G. , Noormets, A., Chen, J. , & McNulty, S. (2008). Evapotranspiration
estimates from eddy covariance towers and hydrologic modeling in managed
forests in Northern Wisconsin, USA. Agricultural and forest meteorology,
148(2), 257-267.
/17/ U.S. EPA. 2008. Methods for Evaluating Wetland Condition: Wetland
Hydrology. Office of Water, U.S. Environmental Protection Agency,
Washington, DC. EPA-822-R-08-024.
/18/ Whiteley, H. (1979). Hydrologic implications of land drainage. Canadian Water Resources
Journal, 4(2), 12-
J-40 review of chiles reply - v3.docx I BG! 2019-02-07
Annex 25
DHI, “Updating of the mathematical hydrological model
scenarios of the Silala spring waters with: Sensitivity analysis
of the model boundaries”, April 2019
(Original in English)
49
Single Product
Updating of the mathematical hydrological
model scenarios of the Silala spring waters with:
Sensitivity analysis of the model boundaries
Pluri-national State of Bolivia, DIREMAR
April 26, 2019
OHi A/S • Agern Alie 5 • • DK-2970 H0rsholm • Denmark
Telephone: +45 4516 9200 • Telefax: +45 4516 9292 • [email protected] • www.dhigroup.com
50
D~
This report has been prepared under the DHI Business Management System
certified by Bureau Veritas to comply with ISO 9001 (Quality Management)
Approved by
Oluf Zeilund Jessen
Head of Project, Water Resources Department
sensitivity analyse_complemented_v2.docx / RAJ / 2018-04-26
51
Single Product
Updating of the mathematical
hydrological model scenarios of the
Silala Spring waters with:
Sensitivity analysis of the model
boundaries
Prepared for
Represented by
Proiect manaaer
Qualitv supervisor
Project number
Approval date
Revision
Classification
Pluri-national State of Bolivia, DIREMAR
Dr. Emerson Calderon
Roar A. Jensen
Torsten Vammen Jacobsen
11823606
26 April 2019
1 (amendment)
Confidential
DHI A/S • Agern Alie 5 • • DK-2970 H0rsholm • Denmark
' I
I
I
Telephone: +45 4516 9200 • Telefax: +45 4516 9292 • [email protected] • www.dhigroup.com
~
52
D~
4 sensitivity analyse_complemented_v2.docx I RAJ I 2018-04-26
53
CONTENTS
1 Executive Summary .................................................................................................... 7
2 Introduction ................................................................................................................. 9
2.1 Background for this report ............................................................................................. 9
2.2 Scope of this study ............................ .. ............................. ... .......................................... 9
2.3 About the content of this report ...................................................................................... 9
2.4 The overall hydrology of Silala Springs ........................................................................ 10
3 The modelling approach and the applied models ................................................... 15
3.1 The conceptual hydrogeological model ....................................................................... 15
3.2 Components and configuration of the applied numerical model ................................... 19
3.2.1 Modelling approach ..................................................................................................... 19
3.2.2 Implementation of the conceptual models ................................................................... 19
3.2.3 Model calibration and performance ........................................ ..................................... 20
4 The boundary conditions in the Near Field groundwater model ............................ 22
4.1 Fixed head boundaries (along upper inflow boundaries) .............................................. 22
4.2 Closed boundaries (where groundwater flows parallel with the boundary) ................... 22
4.3 Fixed groundwater table gradient (along the lower model outflow boundary) ............... 22
4.4 The observed groundwater levels in the Silala Springs ............................................... 23
5 The scenarios previously analysed ......................................................................... 25
6 The approach to the sensitivity analyses of the boundary conditions .................. 27
6.1 The extent of the Near Field model ............................................................................. 27
6.2 Fixed u/s head boundary (the upper limit to the flow changes) .................................... 28
6.3 Unchanged u/s flux boundary (the lower limit to the flow changes) .............................. 28
6.3.1 The sensitivity runs ...................................................................................................... 29
6.4 The central estimate .................................................................................................... 29
6.5 The influence of the downstream flux boundary .......................................................... 29
7 The results of the sensitivity analyses .................................................................... 31
7.1 Fixed head boundary - the upper bound of the impact range ....................................... 31
7.2 Fixed flux into the model - the lower bound of the impact range ................................. 31
7.3 The assessed sensitivity range and the pre-canalisation flow observations ................. 32
7.4 Sensitivity to the downstream boundary condition ....................................................... 34
8 Summary and conclusions ....................................................................................... 35
8.1 Background ................................................................................................................. 35
8.2 The approach to the sensitivity analyses ................................................. .. ... .. ............. 36
8.3 Results of the sensitivity of the upstream boundary ..................................................... 37
8.4 Results of the sensitivity analysis of the Downstream Boundary .................................. 37
9 References ................................................................................................................. 39
The expert in WATER ENVIRONMENTS 5
54
FIGURES
Figure 2-1 Location of the Silala Springs System ............................ ............................................. 10
Figure 2-2 Approximate extent of the Silala Near Field. (Mulligan & Eckstein, 2011) ................... 11
Figure 2-3 Approximate extents of the Silala Far Field (the groundwater catchments likely to
contribute to Silala Springs System) .............................. ............................................. 11
Figure 2-4 Borehole locations and groundwater level contours in the Silala Near field,
interpolated form Piezometer wells spring elevations and wetland excavations for
soil sampling. N.B. the contouring away from the wetlands and the boreholes are
uncertain (OHi, 2018) .................................................................................................. 13
Figure 3-1 Hydrogeologic Framework Model rendered in 3D. The Silala Fault (HGU7) is
highlighted in red. Remaining units are displayed with transparency for easier
viewing of modelled subsurface ......................... ...... ...... ..... ............. ........... .. .......... .. .. 17
Figure 4-1 Groundwater level maps used in definition of groundwater component boundary
conditions ...................................................................... ..... ................... ..... .. .............. 24
Figure 4-2 Illustration of groundwater boundary condition ............................................................ 24
Figure 7-1 Principle sketch of the range of changes in surface water as generated in the
sensitivity analysis of the upstream head boundaries .................................................. 33
TABLES
Table 7-1 Results of the outer bounds of the sensitivity analyses of the upper head
boundary conditions in % of the flow components in the baseline simulation with
the canals .............. .................... ....... ............................. ... ....................... ... ................ 32
Table 7-2 Results of the outer bounds of the sensitivity analyses of the upper head
boundary conditions as changes in I/s from the flow components in the baseline
simulation with the canals ........................................................................................... 32
Table 7-3 Results of the sensitivity analyses of the downstream groundwater gradient. ............. 34
6 sensitivity analyse_complemented_v2.docx I RAJ/ 2018-04-26
55
1 Executive Summary
This report documents the sensitivity analyses of the boundary conditions of the
numerical integrated surface water - groundwater model of the Silala Springs focusing
on the impacts of the model boundaries on the model results.
Due to the absence of surveyed hydrogeological data further away from the wetlands
and the specific purpose of assessing canal impacts, the model is confined to an area
quite close to the Silala springs. Observed groundwater levels have been used as
boundary conditions to the model. It has been found , however, that the model
boundaries are affected by the changes introduced by the removal of the canals and
that the chosen boundary conditions will therefore also have a bearing on the produced
results for a situation in which the canals have been removed. When considering the
baseline model and the 'no canal"/"Undisturbed" scenario results, the sensitivity and
uncertainty should therefore be taken into account.
This project has produced sensitivity analyses of the model boundaries focusing on
their impacts on the distribution between groundwater and surface water in a situation
without the canals. Through a number of boundary condition sensitivity simulations, the
study has identified ranges of canal impact, which are measured by changes in surface
water and groundwater flows relative to the present canalised conditions.
The analyses have been carried out as two separate parts.
• A sensitivity analysis of the upstream boundaries and
• An analysis of the sensitivity of the results to changes in the downstream
outflow boundary along the border to Chile.
Regarding the upstream boundaries, our approach has been to frame the possible
range of model results between two sets of assumptions - each considered to
represent either the lower or the upper bound of the flow changes, relative to the
present baseline situation, assuming that the canals are dismantled and wetlands are
restored.
Assuming that no changes will occur on the boundary will lead to the largest impacts on
the surface water flows and, hence, such analysis will represent the upper bound. This
corresponds to the simulations already reported in the original study (DHI , 2018).
To assess the boundary conditions that would represent the smallest impact of
removing the canals (the lower bound), ii is assumed that the discharge through the
Near Field without the canals will be the same as with the canals.
The downstream model boundary controls the flow across the border and affects the
ratio between surface water and groundwater flow. This boundary assumes a fixed
groundwater gradient (a fixed slope of the groundwater heads) The sensitivity of model
results to this boundary was analysed by varying the gradient with +/- 20% of the
calibrated value of 0.05.
The expert in WATER ENVIRONMENTS 7
56
8
Findings
The sensitivity analysis of the upstream boundary conditions resulted in the following
ranges of the transborder flow in the scenario without the canals:
• For the boundary sensitivity cases tested the simulated range of decrease
in transborder surface flow when removing the canals is 11 % - 33 %.
• The groundwater flow will increase between 4% and 10% of the modelled
groundwater flow in the present situation.
• The evapotranspiration from the wetlands is much better defined and will
increase between 28% and 24% of the baseline values or between 3 and
3.4I/s.
The values do not add up to zero as only one of the analysed boundary conditions
assume the same amount of flow through the Near Field with and without the canals
while the other assumes that a part of the original flow bypasses the Near Field in the
surrounding groundwater layers.
The sensitivity range of assessed surface water flow crossing the border in the "No
canal" scenario is quite large and it is not possible to quantify the most likely surface
water flow value within this interval from this analysis.
However, the only field observations of the pre canal situation (Fox, 1922) is 131 I/s or
18% lower than the 160 I/s measured at the de-siltation tank (the assumed same
location) during 2017. A reduction of 18% is quite close to the centre of the sensitivity
range.
The downstream gradient does influence the ratio of groundwater versus surface water
out flow.
In the "No canal" situation, applying a 20% higher head gradient along the lower
boundary increases the groundwater fraction across the border at the expense of the
surface water fraction . More specifically, it results in 7% less overland flow and 8%
more groundwater flow across the border. Evapotranspiration is almost unchanged
(decreasing less than 1 percent). Application of a 20% lower downstream gradient has
the opposite effects.
While the flow through the model increases with the downstream gradient, both in the
present baseline situation (with the canals) and in the situation without the canals, the
impact from the gradient on the model inflow in these two situations is small (less than
1 percent).
sensitivity analyse_complemented_v2.docx I RAJ/ 2018-04-26
57
2 Introduction
2.1 Background for this report
Introduction to the contract and its objectives
On 6 June 2016, the Republic of Chile filed an application with the International Court
of Justice against the Pluri-national State of Bolivia, concerning the dispute over the
status and use of the waters of the Silala (Chile v. Bolivia). In the framework of this
legal case, Chile presented its Memorial on 3 July 2017 and Bolivia presented its
Counter-memorial and Counterclaim on 31 August 2018. Under a former contract for
Bolivia's Strategic Office for the Maritime Claim, Silala and International Water
Resources - DIREMAR, OHi carried out technical analyses and assessments of the
surface water and groundwater flows in the Silala Springs System with and without the
implemented drainage canals, analyses that were referenced in Bolivia's Counter Claim
to Chile.
On 15 February 2019, Chile presented its response to Bolivia's counter claims' . in
which Chile questions the findings of OHi, particularly the boundary conditions of DH l's
numerical model used for the assessment of the possible hydrological impacts of the
drainage canals implemented in the wetlands early last century. OHi supports the need
for a sensitivity analysis considering that the database for the modelling is quite sparse.
OHi has been contracted by DIREMAR (through the product-based consultancy
contract: CDP-IN° 16/2019) to carry out a sensitivity analysis of how these boundaries
affects the assessed hydrological impacts of the canalisation.
2.2 Scope of this study
According to the Terms of Reference, the scope of the study is:
a) Sensitivity analysis of the established Near Field model boundary condition's
influence on the modelled canal impacts, with the aim of strengthening the
impact assessments made in 2018
b) Description of the analyses made and their rationale
c) Discussion of the results
2.3 About the content of this report
The overall contents of the report are given in the contract and the report therefore
follows the structure given.
This report is to be seen as an extension to the former modelling reports by OHi (OHi,
2018) which, in its annexes, includes detailed descriptions of the applied modelling
1 International Court of Justice (Feb 2019): Dispute over the status and use of the waters of the Silala (Chile v. Bolivia). Reply of The
Republic of Chile.
The expert in WATER ENVIRONMENTS 9
58
approach, its calibration and its results). To make it easier to read, the introductions to
the applied models are repeated in section 3 of this report as follows:
- Section 3.1 describes the hydrogeological conceptual model.
- Section 3.2 introduces the configuration and components of the integrated model.
Section 4 describes and discusses the model boundary conditions.
Section 5 describes and discusses the scenarios previously analysed
Section 6 describes the approach to the sensitivity analyses carried out and
Section 7 describes the results and conclusions of these analyses.
Summary and conclusions are included as section 8.
2.4 The overall hydrology of Silala Springs
10
This sub-section gives a brief introduction of the hydrology of Si/ala Springs System.
Please refer to (DH/ 2018) for more detailed information.
The Silala Springs are located at
altitudes from 4300 to 4400 m above sea
level in the arid Western part of the
Potosi Department of Bolivia, a few
kilometres from the border with Chile
(see Figure 2-1 ).
The Silala Springs System is fed by
groundwater from sources further inside
Bolivia and constitutes the only flowing
surface water resource on the Bolivian
side of the border within a distance of 20
kilometres.
Today, the Silala Springs System in
Bolivia is a modified flow system in
which a fine network of pipes and stonelined
canals drains the Silala wetlands
and conveys the water efficiently from Figure 2-1 Location of the Silala Springs System
the large number of individual springs in
the Northern and Southern wetlands in
Bolivia to a water intake on the Chilean
side of the international border around 4
km downstream (see Figure 2-2).
Contributions from superficial catchment runoff are small in comparison to the
stationary or slowly varying groundwater inflow to the wetlands which is in the order of
160-210 1/s.
sensitivity analyse_complemented_v2.docx f RAJ/ 2018-04-26
59
' '
' ' ' '
' ' '
FCAB J
tntakt ', ~
' 0
C;.\?.,
~ -- ';;,
(<!' \.,,
u 1.5 km
Figure 2-2 Approximate extent of the Silala Near Field. (Mulligan & Eckstein , 2011)
Figure 2-3 Approximate extents of the Silala Far Field (the groundwater catchments likely to contribute
to Silala Springs System)
Silala has a desert climate with low precipitation, low temperatures but high potential
evaporation. Outside the wetlands, the vegetation is very sparse and top soils are
coarse and sandy, originating from weathered or glacier-eroded lava and ignimbrite
formations.
The base rock formation consists of ignimbrite layers with a general inclination towards
the West and the valley in which the Silala springs, wetlands and canals coincides with
major faults in the ignimbrites (SERGEOMIN, 2001 ).
The ignimbrites are porous and fractured and have been found to have significant
hydraulic conductivities. In some areas, the ignimbrites are found directly under the top
soils, while in other parts of the area, they are superimposed by layers of lavas which
have been deposited during later eruptions.
The expert in WATER ENVIRONMENTS 11
60
D~
12
The potential groundwater heads observed in the piezometer wells, established by
DIREMAR during 2017 in the area around the Silala springs, indicate a groundwater
flow from higher Eastern grounds towards the Silala Springs and further on towards the
international border.
Although the spatial extent and volume of the groundwater aquifer discharging
groundwater to the Silala springs remain unknown (as indicated in Figure 2-3), both the
observed groundwater levels (sloping towards the Chilean border) and the high
hydraulic conductivities of the aquifer show that water from the Silala River basin will
flow from Bolivia as surface water or groundwater (DIREMAR's Hydrogeological field
survey program, 2017, Figure 2-4 ).
It is noted that the terrain gradients are very small along the road to Laguna Colorada
crossing the upper parts of the Far Field. It is therefore likely that the underlying
groundwater head gradients along the road are also small, and that modest changes in
the groundwater heads along the road might influence the location of the groundwater
divide towards the Laguna Colorada Basin, potentially affecting the flow to that basin. It
is however not deemed likely that the back water impacts from removal of the canal
should travel so far upstream.
sensitivity analyse_complemented_v2.docx I RAJ/ 2018-04-26
61
Figure 2-4
n ~ ...
A Spring
• Plezometers (Groundwater Clevetlon. masl)
-- Oocombor 2017 Wotcr Lovols
... ..... lnterna1lonal Border
~ · , \ 4,fe'.2;11'1A, AU' \ ,;
\~
Borehole locations and groundwater level contours in the Silala Near field, interpolated form Piezometer wells spring elevations and wetland
excavations for soil sampling. N.B. the contouring away from the wetlands and the boreholes are uncertain (DHI, 2018)
The expert in WATER ENVIRONMENTS 13
62
D~
This page was intentionally left blank.
14 sensitivity analyse_complemented_v2.docx /RAJ/ 2018-04-26
63
D~
3 The modelling approach and the applied models
The section describes the components and configuration of the applied model
3.1 The conceptual hydrogeological model
A conceptual hydrogeological model has been used to establish the integrated
numerical model and ensures that the tatter corresponds reasonably well to the
hydrogeological conditions observed in the field. For more detailed information on the
conceptual hydrogeological model, please refer to (DHI, 2018)
The conceptual hydrogeological model was established by combining the surface
geological mapping with topographies and the extensive hydrogeological borehole
program (launched by DIREMAR during 2017).
The data from hydrogeological characterisation program (the borehole data, the
hydrogeological borehole test results, and electro resistivity transects) has been
combined with previous Bolivian data (surface geological mapping, water quality and
surface water flow rates) along with Chilean borehole data and pumping test results
(Arcadis 2017). The combined data has been used to develop a Hydrogeological Model
(HCM) of the Silala Near Field and, to a lesser extent, the Silala Far Field areas.
The conclusions of the combined data analyses regarding the conceptual groundwater
flow system of the Silala Near Field include:
• Groundwater discharge is the principal source of water to the Silala Spring
System. Dominant sources of groundwater to the springs are:
- Northeast trending structures including several large faults. These
fault zones are brecciated and have elevated hydraulic conductivity
relative to the surrounding materials and are interpreted to be
transmitting groundwater over large distances (i.e. Silala Far Field or
beyond).
- A network of small apertures, Northwest trending fractures act as
conduits transmitting groundwater along strike.
• Pumping tests completed in the Southern Wetland indicate a transmissive
ignimbrite aquifer with large-scale hydraulic conductivity estimated to be
about 18 mid and locally higher conductivity within the Silala Fault Zone
(up to 54 m/d). These are higher than the 6.5 m/d estimated from the
pumping tests in Chile near the border;
• Hydraulic test data indicates that:
- Fractures in the ignimbrites are well connected over a large scale and
appear to control the flow characteristics of the aquifer.
- The aquifers approximate a porous media.
• Groundwater head measurements indicate that groundwater is
discharging to the Southern and Northern wetlands (gaining) but much
The expert in WATER ENVIRONMENTS 15
64
16
further downstream the groundwater may be hydraulically disconnected
from the Silala Canal at the Chilean-Bolivian border (disconnected losing
stream);
• The hydrochemistry and age of the groundwater discharging into the
Northern wetlands is significantly different from that of the Southern
wetland. Water in the Southern wetland was found to be considerably
older than water in the Northern wetland. Isotope analyses indicate the
apparent average age to be up to 1,000 years in the Northern and 11 ,000
years in the Southern wetland, respectively. Although such analyses may
over-estimate the real water age (OHi, 2018, Annex F), the age of the
spring water is indeed very old. A likely interpretation of the difference in
water chemistry and age is that this older water is derived from flow within
the Silala Fault Zone from a sub-regional to regional flow regime (i.e. the
Silala Far Field), while the younger water in the Northern wetland is more
likely to be derived from localised flow closer to the Silala Near Field.
It is found that water from the Silala River basin will flow from Bolivia, as surface water
or groundwater. However, the spatial extent and volume of the groundwater aquifer
discharging groundwater to the Silala Springs remain unknown.
sensitivity analyse_complemented_v2.docx I RAJ/ 2018-04-26
65
Figure 3-1
D~
' ¾-?;
~~
Hydrogeologic Framework Model rendered in 3D. The Silala Fault (HGU7) is highlighted in red . Remaining units are displayed with transparency for easier
viewing of modelled subsurface
D~
The expert in WATER ENVIRONMENTS 17
66
D~
This page was intentionally left blank.
18 sensitivity analyse_complemented_v2.docx /RAJ/ 2018-04-26
67
3.2 Components and configuration of the applied numerical model
The section describes the modelling approach taken, why it was chosen along with the
strengths and weaknesses and alternatives. Furthermore, it contains a brief description
of the implementation of the field data and the hydrogeological conceptual model into
the numerical integrated model and introduction to the calibration of the model and the
results on the baseline simulation of the present situation. The descriptions are
included here to comply with the Terms of Reference and ease the reading. For more
details please refer to the original modelling reports ((DHI , 2018), Annex F, G and H).
3.2.1 Modelling approach
Although we acknowledge the preference for closed water balance groundwater
models with fixed upstream boundaries and "known" recharge input, we rejected the
approach due to the lack of hydrogeological data in the Far Field, uncertainty in climate
data and recharge rates and the inability based on available data, to accurately define
catchment and aquifer boundaries. A groundwater calibration of the whole Far Field
area to match the few data in the Near Field would inevitably be based on a lot of
assumptions about the presently un-characterised areas of the aquifer.
The technical approach employed was therefore to collect hydrogeological information
within and in the vicinity of the Near Field . This allowed for the development of a
numerical model that was calibrated to an extensive field characterisation database
including hydraulic parameters and head distributions at various depths.
We do also acknowledge that the downside of the chosen approach is that the model
boundaries may affect the model results. The present report aims at quantifying model
sensitivities in this respect.
To cope with all the important hydrological processes playing a role in the Silala
wetlands, a proven integrated physically based hydrological modelling system (MIKE
SHE) has been applied.
3.2.2 Implementation of the conceptual models
The total model area of the Silala Near Field is 2.7 km2 and the main elements of the
model were established in accordance with the conceptual models described in the
previous sections.
All canals are represented in a hydrodynamic one-dimensional model in terms of cross
sections and levels as surveyed in the field. Canal modifications as observed in the
field have been included.
Flow and water ponding on terrain are described in the two-dimensional overland flow
component in a 10 m by 10 m grid of the whole Near Field. The terrain levels of the
overland flow model have been interpolated from the detailed drone survey of the area.
The unsaturated zone model calculates the evapotranspiration from the wetlands and
all upland areas and have been established using standard parameters for the soil
types found in the wetlands during the soil survey. This model uses the same grid
resolution as the overland flow model.
The expert in WATER ENVIRONMENTS 19
68
20
The 3-D hydrogeological conceptual model (described in Section 3.1) is implemented in
the numerical groundwater model. The hydrogeological units and their spatial extents
defined in the hydrogeological conceptual model are represented. The numerical
groundwater model applies three layers. The top layer has varying thickness and
hydrogeological properties as it incorporates all surficial deposits. The second layer
includes the upper Silala ignimbrite and the third layer represents the deep ignimbrite
layer. The fault line defined from the surface to a depth of 400 m cuts across the layers
and introduces a high permeable flow zone along the canals. The same horizontal grid
resolution of 10 m by 10 mas in the above models is used.
The input data to the model are data in terms of precipitation, potential evaporation and
temperature. The boundary conditions for the groundwater model are based on ground
water levels (as determined by the field observations) and assigned along the upper
groundwater boundaries controlling inflow and a constant groundwater gradient along
the downstream boundary controlling outflow.
3.2.3 Model calibration and performance
Model parameters have been adjusted iteratively in a calibration process to
demonstrate that the model qualitatively describes the Silala Near Field hydrology in
accordance with conceptual understanding and that the model results produced
quantitatively match the measured values, in particular the canal flows at the gauging
sites.
II has been found that the overall results of the integrated model reproduce the
important characteristics from the field observations in terms of:
• Significant groundwater inflows to the Silala Near Field area through the
high permeable fault zone and upper Silala ignimbrite
• Overall groundwater flow towards the low-lying wetlands, the canals and
the deep cut ravine sections
• Groundwater feeding surface water by discharges to the springs, canal
and drainage network
• Upstream gaining canal reaches versus the downstream neutral or loosing
reach from the confluence to the border.
• Outflow of the Silala Near Field area as combined canal and groundwater
flow at the border.
sensitivity analyse_complemented_v2.docx I RAJ/ 2018-04-26
69
D~
The calibration against field data shows that:
• The model simulates groundwater discharge to the canal system in terms
of measured mean canal flow (C1-C7) reasonably well, i.e. 0 -18 %
deviation.
• The largest relative difference is found at upstream Southern canal (C1-
C3). From C4 to the downstream confluence and border area including
the Northern branch (C6), the model performs well with differences to the
observations which are within the canal flow measurements uncertainty.
• The calibrated model water balance shows groundwater flow across the
downstream model boundary in the order of 106 I/s compared to surface
water flow of 150 I/s. The width of the downstream model boundary is 450
m (with the ravine in the centre). For comparison, the rough hand
calculation in (OHi, 2018) Annex F assessed 230 I/s (or more), but over a
much larger cross section and using less information.
• Evapotranspiration mainly occurs in the wetlands and along the canal
riparian corridor. Due to the restricted total area the total ET losses
correspond to only 10 I/s under current conditions.
In summary:
The numerical model is developed from the conceptual understanding and the field
data collected. The calibrated model is able to simulate the canal flows (C1-C7)
reaching approximately 150 I/s at the border.
The model results suggest a considerable groundwater flow component but this cannot
be confirmed by measurements and is therefore more uncertain than surface water
flows. However, the model results confirm a coupled groundwater - surface water
system within the Silala Near Field area. This coupled system extends across the
border.
The calibrated model is in reasonable agreement with the current conditions and
therefore a sound basis for estimating the impacts of the canals.
The expert in WATER ENVIRONMENTS 21
70
4 The boundary conditions in the Near Field groundwater
model
This section describes the various types of boundary conditions applied in the Near
Field model, their characteristics strengths and weaknesses. It also describes the
observed groundwater conditions in the area.
4.1 Fixed head boundaries (along upper inflow boundaries)
In order to compute the groundwater flow and water level conditions at the outer model
boundaries, boundary conditions must be assigned. Different types of groundwater
boundary conditions are available and they are all approximations based on
assumptions on the prevailing flow or water table condition. The Silala Near Field area
receives groundwater inflows from a larger catchment along parts of the model
boundaries. A prescribed head boundary is used along those sections of the boundary
where groundwater flows from outside the model area enter the Silala Near Field model
area. The fixed head boundary implies that flow into the model area may change if the
groundwater tables changes, e.g. due to changes in the surface water system.
Increasing groundwater table inside the model will propagate to the boundary
decreasing the flow gradients and the inflow.
This fact complicates the comparison of scenarios. In (Chile R. , 2019), there were
questions about the boundaries of the model changing its inflow in the modelled
scenario.
The changes in inflow between scenarios is an indication of the boundaries being
affected by the intervention. In the next section, we analyse the sensitivity of the main
model results to variations in the groundwater heads along the model boundaries.
4.2 Closed boundaries (where groundwater flows parallel with the
boundary)
Where the model boundary runs perpendicular to the head contour lines, there is no
water table gradient to drive inflows and the sections are subsequently assumed closed
(no flow). Model report reference ((DHI , 2018) Annex G).
4.3 Fixed groundwater table gradient (along the lower model outflow
boundary)
22
At the downstream model boundary, at the border, a groundwater table gradient
boundary condition is applied. ARCADIS 2017 calculated groundwater table gradients
of approximately 0.05 (mlm) between boreholes at the border and upstream of
Quebrada Negra. The gradient boundary condition implies that groundwater flow
across the boundary is adjusted to maintain the specified water table gradient.
sensitivity analyse_complemenled_v2.docx / RAJ f 2018-04-26
71
D~
4.4 The observed groundwater levels in the Silala Springs
Water level data available for the Silala Near Field area includes piezometric levels
from boreholes, spring water level and water levels recorded as part of the soil survey.
The water level information has been processed to derive a piezometric contour map
(Figure 2-4). The highest density of observed water levels is found relatively close to
the canals and wetlands, i.e. the central parts of the model area. The contour lines
have been extrapolated away from the observation points which means that the
uncertainty on the contours is higher at the model boundaries than the internal areas
along the wetlands and canals.
The contour data has been used as direct input for describing both the initial
groundwater head map for the integrated model simulation and the boundary
groundwater head values at the model boundaries of all groundwater layers (Figure
4-1 ). The groundwater table elevations range from 4420 m at the model boundary
upstream of the Southern wetland (East) to approximately 4290 m close to the
downstream model boundary at the Bolivian-Chilean border (West). The groundwater
contours indicate significant groundwater head gradients and inflows upstream of both
the Southern and Northern wetland,
In the model, the piezometric contour map has been used to define boundary
conditions for all groundwater layers, which implies that any upward pressure gradients
from deeper confined layers are not represented in the model boundary conditions.
Pressure transducers installed in a number of boreholes were used for water level
monitoring. Groundwater table elevations were collected for a relatively short period.
The water tables were relatively stable with an average temporal variation of less than
0.5 m in all boreholes except two (Annex F). The groundwater head values assigned as
boundary conditions in the integrated model have thus been considered constant in
time. The inflow to the model area is a function of the assigned boundary head values,
the hydraulic conductivity and thickness of the geological layers and the groundwater
heads inside the model domain. With a fixed head upstream boundary and the
downstream gradient controlled outflow boundary condition, the stationary integrated
model will gradually approach a steady-state equilibrium balancing upstream
groundwater inflows versus downstream surface and subsurface outflows.
n al polC!lllal head ,n the sattered zone
,,-....... ~~,.......~-,.......~-........ ~-,.......~-........ ~-........ ~-,-.....~ ..........
600000 11<>1000 601500 602000 602SOO IJO]l)()CI 603500
The expert in WATER ENVIRONMENTS
-·..,-._..,.
Wtl U20
UOO MIO
.&)tO.~
~ -•ltO
Q.11 - "31a
..... ,,10
o.sa-•liG
,t,l'D -'1.'MI
,1,31 . 040
• 3 . o,a
011 , &!2(1
Q)Q , aJtO
-'1-N~•.•,J.O.O f'r.«1\1.Mie
23
72
Figure 4-1 Groundwater level maps used in definition of groundwater component boundary conditions
Figure 4-2 below illustrates groundwater tables, cross sectional flow and discharge
from an upstream head boundary towards a downstream surface water body.
Ground sudace
Figure 4-2 Illustration of groundwater boundary condition
24 sensitivity analyse_complemented_v2.docx I RAJ / 2018-04-26
73
D~
5 The scenarios previously analysed
In this section, we describe the scenarios investigated in the previous assessment
(OHi, 2018) and the main findings.
To address the main objective of this project, the following baseline and scenario
models have been run as follows:
1) Baseline model. Represents the (2018) Silala Near Field area with the canal and
drainage network as it is today. The surface water canal model includes both
reaches which are more or less unchanged compared to the original canal
construction but also reaches where the canal has been removed or blocked. The
baseline scenario is used as a reference to estimate the magnitude of changes.
2) No canal scenario. The entire canal and drainage network included in the baseline
model is removed. As a result, the surface flows are no longer concentrated in
narrow defined channels but can appear and flow across the entire surface area
and the direction of flow is largely controlled by the surface topographical slope.
3) Wetland restoration (undisturbed) scenario. By removing the canal and drainage
network, the basis is created for restoration of the degraded wetlands and riparian
corridors. The scenario considers the effects of long-term peat accumulation in
wetlands. The scenario was in our former analyses (OHi, 2018) referred to as the
"Undisturbed" wetland scenario. We have in this report maintained this title to avoid
confusion although "wetland restoration" might be a more appropriate title.
According to the integrated model scenario results removing the canals and restoring
wetlands will affect both groundwater and surface water and both inflows and outflows
of the Silala Near Field area.
1. The simulated surface water flow at the downstream model boundary (located
at the Bolivian-Chilean border) reduces by 31-40 % relative to the present
situation.
2. The simulated groundwater flow at the downstream model boundary (located at
the Bolivian-Chilean border) increases by 7-11 % relative to the present
canalized situation
3. The total model boundary inflow at the upstream model boundary decreases by
10-15%.
4. The evapotranspiration increases by 20-30 % by removing the canals and
restoring wetlands. This increase amounts to 2-3 I/s in the situation without the
canals and is included in the cross-border flow changes mentioned under point
1 and 2
5. For the confluence to border section, a maximum of 25 % of surface water may
be lost to subsurface flows. Infiltration loss in this section is included in the
cross-border flow changes mentioned under point 1 and 2.
6. All of the scenario results and local model analysis suggest that both surface
water flow and groundwater flow cross the border.
The expert in WATER ENVIRONMENTS 25
74
26
The flow impact percentages describe the model results ranges but not the uncertainty
on model results. Model predictive uncertainty depends on a number of factors and
uncertainty sources, e.g. limitations in input data, model structure, parametrisation and
measurement errors. A strictly quantitative uncertainty analysis is not feasible and has
not been attempted but model uncertainty should not be ignored in the interpretation of
results.
sensitivity analyse_complemented_v2.docx I RAJ 12018-04-26
75
D~
6 The approach to the sensitivity analyses of the boundary
conditions
The differences in model inflow in the above scenarios (Section 7) indicate that model
boundaries are affected by the analysed interventions (the removal of the canals). This
section describes the approach to the sensitivity analyses aiming at quantifying this
effect. Both the upstream and the downstream boundaries influence the results and are
analysed separately. The approach taken to the upstream boundaries is to describe the
upper and lower limits of the possible flow variations.
6 .1 The extent of the Near Field model
Due to the absence of surveyed hydrogeological data further away from the wetlands,
the model is confined to an area quite close to the Silala Springs (The Near Field)
where the observed groundwater levels have been used as boundary conditions to the
model. The rationale of this modelling approach was further described in section 3.2.1
above.
It has been found, however, that the model boundaries are affected by the changes
introduced by the removal of the canals and that the choice of boundary conditions will
therefore also have a bearing on the produced results for a situation in which the canals
have been removed .
When considering the baseline model and the 'no canal' scenario results, the sensitivity
and uncertainty should be taken into account. This sensitivity analysis is specifically
looking at the model boundary sensitivity and does not represent a full model sensitivity
or uncertainty analysis. The reason for focusing on the groundwater boundary
conditions is that the Silala springs and surface water system are fed entirely by
groundwater from a large poorly delineated upstream catchment. The Near Field model
include the springs, wetlands and canal features and extends to the Bolivian-Chilean
border. It does, however, not represent a closed water balance unit but uses head
dependent groundwater inflow boundaries in a restricted area which introduces a
potential sensitivity with respect to key model results.
The existing channels provide a network that drain the groundwater and conveys this
water rapidly away from the Silala springs. By removing the channels, the groundwater
is drained less efficiently, the resistance to flow emerging on the surface is increased
and the groundwater levels will increase.
With higher groundwater levels, closer to the surface the evapotranspiration will
increase.
The increase in groundwater level within Silala Springs will propagate both upstream
and downstream of Silala (although the increase will reduce with distance from the
springs). The increased groundwater levels upstream of Silala can reduce groundwater
flow into the wetlands, in which case a fraction of the ground water will flow around the
model area and the groundwater inflow to the model will then be reduced. Likewise the
raised heads inside the model area will lead to higher cross border groundwater
discharge and also higher heads on the Chilean side of the border.
The expert in WATER ENVIRONMENTS 27
76
6.2 Fixed u/s head boundary (the upper limit to the flow changes)
The flow in surface water and groundwater is linked to the boundary inflow which again
depends on the water level differences at the outer groundwater model boundary
Application of a fixed upstream (u/s) head boundary is a good and an appropriate
approach for the baseline (present) situation. The groundwater heads can be assessed
from borehole observations representing the de-facto conditions in the field.
The Near Field area does not represent a closed catchment unit and the groundwater
flow towards Chile extends beyond and also partly bypasses the Near Field area. With
extra resistance to overland flow in the Near Field area, the fraction of the total Silala
discharge going to superficial flow is expected to decrease and the rest (except the
increased loss to evapotranspiration) will discharge as groundwater through the Near
Field or the adjacent aquifer sections (assuming the ground water is in balance2).
Since the boundary heads are fixed , the gradients near the boundaries will decrease
and lead to smaller total flow through the model in the "No canal" and the "Undisturbed
wetland" scenarios as compared to the present baseline. Therefore, the change in
surface discharge as derived from the scenario are considered to represent the upper
limit to the flow changes.
6.3 Unchanged u/s flux boundary (the lower limit to the flow changes)
In the scenario, using a fixed upstream flux boundary, we will force the same
groundwater flow into the nearfield model area as in the present situation.
If the groundwater drainage capacity is reached, the water levels (heads) will rise
above terrain level and create surface flow. This excess flow, that cannot be drained
through groundwater, will run off superficially from the wetlands and the ravines as is
also observed in other undisturbed bofedales.
The fixed flux boundary condition will exaggerate the total flow through the Near Field
since increased heads along the boundary will - in reality - divert part of the
groundwater flow around the model area through adjacent aquifer areas.
Hence, results with a fixed flux boundary are likely to give too small changes in the
surface flow components and the real situation lies somewhere between the two
extremes (a fixed flux boundary representing the smallest changes and a fixed head
boundary representing the largest changes).
Since the results originally reported were generated using a head boundary, a
sensitivity analysis is made by varying the heads along the boundary.
The groundwater flux into the calibrated baseline model (through the observed fixed
head boundaries) have been extracted and introduced in as a fixed flux boundary in the
"No canal" and "Undisturbed" scenario, respectively. A simulation of the scenario with
this boundary condition produces corresponding groundwater heads along the model
2 It has not been proven that the groundwater system is in balance (that recharge to the aquifers matches the outflow), It
has been demonstrated , however that the recharge from a plausible groundwater catchment cover a large part of the
assessed combined groundwater and surface water discharge trough the Near Field , (DHI , 2018).
28 sensitivity analyse_complemented_v2.docx /RAJ/ 2018-04-26
77
D~
boundaries. The differences between these scenario heads and the fixed heads from
the simulation have been used to assess the sensitivity of the scenario to the boundary
conditions.
6.3.1 The sensitivity runs
Several fractions (0, 0.25, 0.50, 0.75 and 1.0) of the head differences between the
baseline and the scenario with unchanged inflow to the model have been added to the
original baseline boundary and used to simulate the boundaries influence on the
results. Of these simulations, the 0- fraction represent the unchanged original head
boundary (head approximation) while the 1.0- fraction represent scenario where the
inflow to the model is kept constant (flux approximation).
6.4 The central estimate
The range of surface water flow crossing the international border without the canals (as
assessed by this sensitivity analysis) is wide. However, it is not possible from this
analysis to determine a 'best estimate' within this range.
The best assessment of the canal impact is still considered to the field observations of
the pre-canal situation by (Fox, 1922) of 131 I/s or 18% lower than the 160 I/s
measured at the de-siltation tank during 2017.
6.5 The influence of the downstream flux boundary.
The removal of the canals may redistribute the groundwater levels in the Near Field
with a tendency of having higher groundwater heads and potentially large contribution
of overland flow in the downstream part of the model close to the downstream
boundary. Therefore, it is also relevant to analyse how the downstream boundary
affects the model.
The calibrated baseline scenario uses a fixed groundwater gradient of 0.05 m/m which
is close to the overall gradient calculated from borehole observations to the overall
slope of the terrain. Local differences in groundwater gradients may, however, easily
occur and the sensitivity of the results to smaller changes in this gradient(+/- 20%) has
been investigated.
The expert in WATER ENVIRONMENTS 29
78
30 sensitivity analyse_complemented_v2.docx /RAJ/ 2018-04-26
79
7 The results of the sensitivity analyses
The results described here concern the sensitivity to the model boundary conditions of
the "No Canal" and "Undisturbed" scenarios as considered in the original modelling
report (DH/, 2018).
7 .1 Fixed head boundary - the upper bound of the impact range.
The sensitivity analysis has resulted in a range of possible changes in which the upper
bound (the result leading to the largest flow changes from the baseline simulation) is
considered to correspond to the changes originally simulated and reported in (OHi,
2018).
The simulated changes in the main flow components as percentages of the baseline
flows are listed in Table 7-1 while the specific changes in I/s are included in Table 7-2.
It is noted that in this upper bound impact case, the transborder surface flow is
assessed to be around 33% lower than the simulated flows of the baseline with the
canals. This corresponds to a drop-in surface water flows in the order of 50 I/s in the
"Undisturbed" scenario.
Groundwater flow through the model is assessed to increase by 10% and
evapotranspiration by 34% (3.4 I/s)- also in the "Undisturbed" scenario.
7.2 Fixed flux into the model - the lower bound of the impact range.
The constant flux boundary is considered to represent a lower bound for the magnitude
of the flow changes in the "No canal" and "Undisturbed" scenarios, as compared to the
baseline with the canals. We consider this a lower bound case since higher heads
along the upper boundaries of the model area will inevitably generate higher flows bypassing
the model area, in turn, reducing the head increases along the boundary and
boundary flows into the Silala system. A quantification of this by-pass groundwater flow
is however difficult due to the lack of hydrogeological characteristics of the ignimbrite
aquifer further away from the Silala ravine.
The simulated changes in the main flow components as percentages of the baseline
flows are listed in Table 7-1, while the specific changes in I/s are included in Table 7-2.
It is noted that, in this lower bound case, the transborder surface flow is assessed to be
around 11 % lower than the simulated baseline flows (with the canals). This
corresponds to a drop-in surface water flows in the order of 16 I/s.
Groundwater flow through the model is assessed to increase by 4 % and
evapotranspiration by 28% (2.8 I/s, only).
The range in assessed surface transborder flow in the "No canal" and "Undisturbed"
scenarios is quite large and it is not possible from this analysis to determine where in
this range the most likely superficial flow value will be.
The expert in WATER ENVIRONMENTS 31
80
Table 7-1
However, the field observations of the pre canal situation by (Fox, 1922) 131 I/s or 18%
lower than the 160 I/s measured at the de-siltation tank (the assumed same location)
during 2017.
Results of the outer bounds of the sensitivity analyses of the upper head boundary conditions in % of the
flow components in the baseline simulation with the canals3
Canalised Change from canalised conditions
situation (1/s) (% of baseline)
Baseline Lower Bound Upper Bound
Inflow to model 253.6 0 -11
Surface outflow 149.0 -11 -33
Groundwater outflow 106.3 4 10
Evapotranspi ration 10.0 28 34
Storage and num. inacuracy -11.7
Table 7-2 Results of the outer bounds of the sensitivity analyses of the upper head boundary conditions as changes
in 1/s from the flow components in the baseline simulation with the canals
Canalised
Changes from canalised conditions (1/s)
situation (I/s)
Baseline Lower Bound Upper Bound
Inflow to model 253.6 -1 -27.9
Surface outflow 149.0 -16 -48.6
Groundwater outflow 106.3 4 10.8
Evapotranspi ration 10.0 3 3.4
Storage and num. inacuracy -11.7 8.4 6.6
7.3 The assessed sensitivity range and the pre-canalisation flow
observations
The sensitivity simulation results in a rather wide range of possible surface water flow
decreases ( 11 % - 33% of the present canal flows) if the canals are removed. It is not
possible to identify a central estimate of the surface water changes through the
sensitivity analysis.
It is noteworthy though, that the only field observation of the non-canalised flow
situation in Silala (Fox, 1922) falls close to the middle of the sensitivity impact range.
Fox measured 131I/s (constantly flowing) at a location which from his description must
be pretty close to the present de-siltation chamber. This corresponds to 18% of the
canalised flow (160 I/s) measured at this site during 2017.
3 Note: The "Balance" (4% of the infiow) in the baseline situation cover smaller storage changes and smaller numerical
inaccuracies in the hydrodynamic model of the canal flows. The "balance" is not valid for the percent wise changes, which
are calculated from different flow components.
32 sensitivity analyse_complemented_v2.docx I RAJ/ 2018-04-26
81
D~
Lower Bound
(flux approx.)
Sensitivity range
Pre-canalization
observations
.. l ..
1 0 15 20 25
Upper Bound
(head approx.)
..
30 3 5
Decrease on surface water flow
(% of canalised situation)
Figure 7-1 Principle sketch of the range of changes in surface water as generated in the sensitivity
analysis of the upstream head boundaries.
The expert in WATER ENVIRONMENTS 33
82
7.4 Sensitivity to the downstream boundary condition
At the downstream end of the model, a fixed gradient boundary was used. The
downstream gradient has an impact on groundwater at the boundary and therefore on
the groundwater heads upstream of the boundary. Hence, the gradient influences the
fractions of the total transborder flow that will pass as groundwater and as surface
water, respectively.
If groundwater heads rise above ground, it will create surface flows.
The sensitivity of model results to this boundary was analysed by varying the gradient
with +/- 20% of the calibrated value of 0.05.
The results that are shown in Table 7-3 shows small impacts from this gradient on the
total inflow to the model and on the evapotranspiration while the fractions of
groundwater and surface water outflows are more affected around +/- 7% and +/- 8%
of the values in the baseline simulation with the canals.
It is also noted that the two "No canal" and the "Undisturbed" scenarios respond almost
identically to these changes in downstream boundary conditions.
Table 7-3 Results of the sensitivity analyses of the downstream groundwater gradient.
Downstream boundary
Gradient 0.04 Gradient 0.06 Gradient 0.04 Gradient 0.06
(-20"/o) (+20%) (-20%) (+20%)
Scenario
No canal No canal Undisturbed Undisturbed
situation situation situation situation
Unit
% change from % change from % change from % change from
org. scenario org. scenario org. scenario org. scenario
Inflow to model -0.8 0.8 -0.8 0.8
Superficial outflow 7.0 -7.1 7.4 -7.2
Groundwater outflow -8.2 7.9 -8.0 7.8
Evapotranspi ration -0.1 -0.8 0.3 -0.2
34 sensitivity analyse_complemented_v2.docx /RAJ/ 2018-04-26
83
D~
8 Summary and conclusions
8.1 Background
A numerical integrated surface water - groundwater model has been established for
the Near Field of the Silala springs and successfully calibrated against observations of
surface water flows along the canals and groundwater levels in the area.
The model parameters and configuration are based on a three-dimensional
hydrogeological conceptual model, which combine the information collected through an
intensive hydrogeological field survey program executed by DIREMAR during 2017 with
geo resistivity measurements and detailed surface geological mapping of the Silala
area.
Hence, the existing model represents the present conditions of the Near Field of the
Silala springs according to the conceptual model. Furthermore, the Near Field model
and the parameters are based on the measured parameters from the field and it is
considered applicable for simulating the hydrological conditions in a situation without
the implemented canalisation works.
Due to the absence of surveyed hydrogeological data further away from the wetlands,
the model is confined to an area quite close to the Silala springs where the observed
groundwater levels have been used as boundary conditions to the model. It has been
found, however, that the model boundaries are affected by the changes introduced by
the removal of the canals. The model results are thus sensitive to the boundary
conditions.
When considering the baseline model and the 'no canal' scenario results, the sensitivity
and uncertainty should be taken into account. This sensitivity analysis is specifically
looking at the model boundary sensitivity and does not represent a full model sensitivity
or uncertainty analysis. The reason for focusing on the groundwater boundary
conditions is that the Silala springs and surface water system are fed entirely by
groundwater from a large poorly delineated upstream catchment. The Near Field model
include the springs, wetlands and canal features and extends to the Bolivian-Chilean
border. It does, however, not represent a closed water balance unit but uses head
dependent groundwater inflow boundaries in a restricted area which introduces a
potential sensitivity with respect to key model results.
This project has produced a sensitivity analyses of the model boundaries focusing on
their bearing on groundwater and surface water in a situation without the canals.
Through a number of boundary condition sensitivity simulations, the study has identified
ranges of canal impact, which are measured by changes in surface water and
groundwater flows relative to the present canalised conditions.
The expert in WATER ENVIRONMENTS 35
84
8.2 The approach to the sensitivity analyses
36
The analyses are carried out as two separate parts.
• A sensitivity analysis of the upstream boundaries and
• An analysis of the sensitivity of the results to changes in the downstream
outflow boundary along the border to Chile.
Upstream Boundaries
A groundwater model covering a substantially larger area Than the Near Field could
have been useful to estimate of how far away from the Silala wetlands the groundwater
heads will be significantly affected by the canal removal. Due to the data and time
constraints of this project, it has not been possible to establish and properly calibrate
such a model.
Instead, our approach has been to frame the possible range of model results between
two sets of assumptions - each considered to represent either the lower or the upper
bound of the changes relative to the present baseline situation.
Using fixed upstream groundwater heads between the baseline and the scenarios will
lead to the largest flow impacts, hence, it will represent the upper bound. This is the
simulation already reported in the original study.
To assess the boundary conditions that would represent smallest impact of removing
the canals compared to the present baseline conditions, the following approach is used:
• It is assumed that the discharge through the Near Field without the canals
will be the same as without the canals.
• This situation can be obtained by introducing a head increase along the
boundary that will maintain the same distribution of ground water fluxes
into the model as in the baseline scenario
Downstream Boundary
In the downstream end of the model, fixed gradient boundary was used. The
downstream boundary has an impact on the groundwater levels upstream of the
boundary and therefore, on the fractions of the total transborder flow that will pass as
groundwater and as surface water. The sensitivity of model results to this boundary
was analysed by varying the gradient with +/- 20% of the calibrated value of 0.05.
sensitivity analyse_complemented_v2.docx I RAJ/ 2018-04-26
85
D~
8.3 Results of the sensitivity of the upstream boundary
The sensitivity analysis of the upstream boundary conditions resulted in the following
ranges of the transborder flow in the scenario without the canals:
• For the boundary sensitivity cases tested, the simulated range of
decrease in transborder surface flow when removing the canals is 11 % -
33 %.
• The groundwater flow through the modelled area will increase between
4% and 10 % of the corresponding flow in the present situation
• The evapotranspiration from the wetlands is much better defined and will
increase between 28% and 34 % of the baseline values or between 2.8
and 3.4 I/s.
The values do not add up to zero as only one of the analysed boundary conditions
assume the same amount of flow through the Near Field with and without the canals,
while the other assumes that a part of the original flow bypasses the Near Field in the
surrounding groundwater layers.
The range in the assessed surface water flow crossing the border in the scenarios is
quite large and it is not possible to quantify the most likely surface water flow value
within this interval from this analysis.
The field observations of the pre canal situation made by (Fox, 1922) of 131 I/s or 18%
lower than the 160 I/s measured during 2017 (DHI, 2018) is located close to the centre
of the range and is regarded to be a valid measure of the impact from dismantling the
canals
8.4 Results of the sensitivity analysis of the Downstream Boundary
The groundwater flow across the downstream model boundary located at the BolivianChilean
border depends on the hydrogeological model, the hydrogeological unit
properties, the water table gradient, the width of groundwater flow section and the
downstream hydrogeological conditions in Chile. The boundary is described by a water
table gradient and the sensitivity to variations in the gradient is tested.
The downstream gradient does influence the ratio of groundwater versus surface water
outflow in the no canal situation. Applying a 20% higher head gradient along the lower
boundary increases the groundwater fraction across the border at the expense of the
surface water fraction. More specifically, it results in 7 % less overland flow and 8%
more groundwater flow across the border. Evapotranspiration is almost unchanged
(falling with less than 1 percent).
Application of a 20% lower downstream gradient has the opposite effects.
While the flow through the model increases with the downstream gradient, both in the
present baseline situation (with the canals) and in the situation without the canals, the
impact from the gradient on the model inflow in these two situations is small (less than
1 percent).
The expert in WATER ENVIRONMENTS 37
86
This page is intentionally left blank.
38 sensitivity analyse_complemented_v2.docx /RAJ/ 2018-04-26
87
D~
9 References
Alcayaga, H. (2017). CHARACTERIZA TJON OF THE DRAINAGE PATTERNS ANO RIVER
NETWORK OF THE SILALA RIVER ANO PRELIMINARY ASSESSMENT. Universidad
Diego Portales.
Arcadis. (2017). Detailed Hydrogeological Study of the Si/ala River.
Arcadis. (2017). Detailed Hydrogeological Study of the Si/ala River. International Court of Justice
Dispute over the status and use of the waters of Si/ala (Chile vs. Bolivia). Memorial of the
Republic of Chile, Volume IV, Annex 2.
Chile, M. o. (2017). Volume I, Memorial and experts report,. International Court of Justice.
COFADENA. (2017). Proyecto Geofisico 28 Lineas Tomograficas Zona Si/ala. La Paz, Bolivia:
Corporacion de las Fuerzas Armadas para el Desarrollo Nacional (COFADENA).
DHI. (2018). Study of the Flows in the Si/ala Wetlands and Springs System, Final Report.
Fox, R. H. (1922). Engineering hydraulic works to capture and analyse the water of the Siloli Plains
. South African Journal of Science. Vol. 19, p: 120-131 .
G.Skrzypek, Z. E. (2011 ). Distichia peat-A new stable isotope paleoclimate proxy for the. Earth
and Planetary Science Letters vol 307, 298-308.
Garraud, R., Vuille, M. , & Clement, A. (2003). The climate of the Altiplano: observed current
conditions and mechanisms of past changes. Palaeography, Palaeoclimatology,
Palaeecology, 194, 5-22.
Houston, J. (2007). Recharge to groundwater in the Turi Basin, northern Chile: an evaluation based
on tritium and chloride mass balance techniques (vol 334: 534-544 ed.). Journal of
Hydrology.
Ministerio de Energias. (2017). Analisis Fisico Quimico de Aguas. La Paz, Bolivia: INSTITUTO
BOLIVIANO DE CIENCIA Y TECNOLOGIA NUCLEAR CENTRO DE INVESTIGACIONES
Y APLICACIONES NUCLEARES UNIDAD DE ANALISIS Y CALI DAD AMBIENT AL.
Mulligan, B., & Eckstein, G. (2011 ). The Silala/Siloli Watershed: Dispute over the Vulnerable Basin
in South America. Water Resources Development Vol 27, no 3., 27(no.3).
Munoz, J., Suarez, F., Fernandez, B., & Maas, T. (2017). Hydrology of the Sita/a River Basin,
International Court of Justice over the status and use of the waters of Sita/a. Memorial of
the Republic of Chile, Volume 5, Annex VII.
SERGEOMIN. (2001 ). Studies of Hydrographic Catchments, Catchment of the Sita/a Springs,
Catchment 20. DEpartment of Potosi.
SERGEOMIN. (2003). Estudio de Cuenca Hidrograficas: Cuenca manantiales def Si/ala. La Paz,
Bolivia: SERGEOMIN Servicio Nacional de Geologia y Mineria Bolivia.
SERGEOMIN. (2003). Studies of Hydrographic Catchments, Catchment of the Sita/a Springs,
Catchment 20. DEpartment of Potosi.
SERGEOMIN, 2. (2001 ). Mapa no 2 Geologia, Hidrologia y hidrogeologia de los manantiales def
Si/ala. SERGEOMIN.
SERGEOTECMIN. (2004). lnvestigaciones en los Manantiales def Sita/a - Presentacion Sita/a
FISICO-QUIMICO ISOTOPOS, La Paz: Internal investigation report.
V. Orzag, S. E., & Cespedes. (2017 (a)). Caraterisaci6n def los suelosdel los bofedales def Si/ala y
areas aledaiias (informe preliminar). LA Paz: DIREMAR.
Viceministerio de Recursos Hidricos y Riego del Ministerio de Medio Ambiente y Agua. (2017).
lnforme Tecnico: 2OA Campana de Muestreo de Agua para el Analisis Hidroquimico e
/sotopico y Oiagnostico de Canales en /os Manantiales def Si/ala. Bolivia.
The expert in WATER ENVIRONMENTS 39
88
D~
J-40 sensitivity analyse_complemented_ v2.docx I BG! 2019-02-07
Annex 26
FUNDECO, “Study of the Water Requirements of the Silala
Wetlands”, April 2019
(English Translation)
90
STUDY OF THE WATER REQUIREMENTS OF
THE SILALA WETLANDS
[TRANSLATION]
By:
FUND-ECO - FOUNDATION FOR THE DEVELOPMENT OF ECOLOGY
La Paz – APRIL 2019
91
EXECUTIVE SUMMARY
The Silala bofedals are unique RAMSAR sites, located in an arid region of the
Bolivian south-west, near the border with Chile, and constitute insular
ecosystems that are naturally vulnerable to fragmentation. The canalization
installed at the beginning of the 20th century severely affected their structure
and, therefore, their ecosystem functions. The objective of the present study
was to evaluate the water requirements of bofedals for their conservation in
original conditions before their intervention. To achieve this, the work comprised
three stages: 1) Bofedal characterization (abiotic and biotic aspects); 2) Analysis
of the relationship between water variables with the plant community and
macroinvertebrates; and 3) Bofedals water demand estimation.
Our results show that bofedals are in a high state of degradation: we have
confirmed the loss of peat depth, with current values of 0.1-1.5 m in the North
Bofedal and 0.25-0.8 m in the South Bofedal, compared to 4.4 m depth average
in the Quebrada Negra wetland. Degradation is particularly greater in the South
Bofedal with bare soil cover, salt crust, dead vegetation that exceeds 50% and
an area covered by cushions of less than 10%; however, bofedals, in order to be
considered in a conserved state, must have a cushion plant cover of more than
the 40%. Currently, the [vegetation] cover is dominated by species that are
indications of disturbance (Carex cf. maritima, Puccinella frigida and different
grasses) and that are explained by the depth of the water table. The canalization
works also affected macroinvertebrate fauna, having found taxa representative
of bofedales (Dorylaimus, Hydrozetes, Homochaeta) in less than five sites of
the North bofedal, while in affected places the macrofauna is more typical of
environments with high flow rates (Andesiops peruvians, Austrelmis and
Hydroptilidae), due to the high water speed resulting from the canalization
works.
As a RAMSAR site, the bofedals should be part of the national strategy for a
restoration plan that responds to international, bilateral and national
commitments; in consequence, the main objective of this work was to estimate
the water requirements of this ecosystem. Thus, we use three criteria to estimate
bofedal water requirements: 1) the calculation of evapotranspiration; 2) the
calculation of the volume of water necessary to saturate them; and the 3)
estimation of the minimum “ecological” flow as per Chilean legislation
92
which has several ecological limitations, but is useful to estimate the initial
flow values needed to recover the Silala bofedals. 1) The potential evapotranspiration
of a bofedal under suitable conditions varies between 1523-1653 mm/
year, which equates to a flow of 5.9 l/s on an annual average for a potential area
of 11.79 ha. 2) The volume of water currently retained in the peat of the Silala
bofedals is 48.4 thousand m3, considering the potential area (11.79 ha) this
value reaches 353.8 thousand m3 and considering the peat depth of the Quebrada
Negra Ravine will reach 443, a thousand m3. 3) The minimum ecological
flow calculated according to Chilean legislation, corresponds to 33.4 l/s annual
average.
All these calculations are estimations of the water necessary to restore the
bofedals to the potential conditions of their state of conservation and extension,
ensuring adequate hydraulic conditions for the development of peat, vegetation
and other ecosystem functions. The minimum flow value calculated should
only be taken as an initial reference as it will most likely increase, considering
that the bofedals are found in an arid region; thus, it is advisable to carry out
continuous monitoring of the water table, vegetation and macroinvertebrates,
to observe their behavior over time.
93
INDEX
1. INTRODUCTION
1.1. Optimal conditions of a bofedal
1.2. Water requirements of the vegetation
2. OBJECTIVES OF THE STUDY
3. METHODOLOGY
3.1. Area studied
3.2. Relation between the bofedal’s plant and macroinvertebrate communities
with
environmental (hydrological and physical-chemical) variables
3.2.1. Vegetal and biomass cover
3.2.2. Macroinvertebrate abundance and wealth
3.2.3. Physical-chemical variables
3.2.4. Laboratory sample processing
3.2.5. Data analysis
3.3. Water requirement estimation
3.3.1. The effect of the phreatic level on the vegetation
3.3.2. Potential evapotranspiration calculation
3.3.3. Vertical overland water flow calculation
3.3.4. Bofedal water volume calculation
3.3.5. Ecological flow
4. RESULTS
4.1. Characterization of the North and South Silala wetlands
4.1.1. Climate
4.1.2. Hydrology
4.1.3. Soils
4.1.4. Physical-chemical variables
4.1.5. Vegetation characterization
4.1.6. Macroinvertebrates associated with the bofedals
4.2. Determination of the most determinant hydrological variables for the
vegetation and macroinvertebrates of the North and South Silala bofedals
4.2.1. Vegetation and hydric variables ratio
4.2.2. Macroinvertebrate and hydric variables ratio
4.3. Vegetation and macroinvertebrate fauna ratio
94
4.4. Estimation of the water requirements of the North and South Silala bofedals
4.4.1. Relation of ideal bofedal cover O. Andina (characteristic of a
bofedal) vs. environmental variables
4.4.2. Potential evapotranspiration (PET) calculation. Value of the water
requirements for the preservation of the Silala bofedals
4.4.3. Calculation of the water value for a bofedal
4.4.4. Vertical flow calculation
4.4.5. Minimum ecological flow calculation
5. DISCUSSION
5.1. Bofedal characterization
5.2. Relation between the hydric conditions and the macroinvertebrate and
plant
community
5.3. Water requirement of the Silala bofedals
6. CONCLUSIONS
7. REFERENCES
95
“STUDY ON THE WATER REQUIREMENTS OF
THE SILALA WETLANDS”
1. INTRODUCTION
The maintenance of a natural flow in aquatic environments is important for the
life cycle of aquatic and terrestrial organisms that benefit from the ecosystem
services that these environments provide (Lamouroux and Capra 2002).
Attaining knowledge on the water requirement of an aquatic ecosystem helps
establishing bases to determine: 1) its state in relation to the characteristics of
its environment and, once the characteristics of the aquatic environment have
been established, that is, a continuous record of hydraulic parameters and a
continuous monitoring of the bio-indicator fauna and flora, 2) tools can be built
to objectively determine the use and utilization of water in the ecosystem. This
water requirement is called ecological flow.
The concept of ecological flow has changed over time. In the 70s, it evolved
from a proposal of a simple system of minimum flow with a fixed numerical
value based on hydrological data (daily and monthly average flows) to
holistic approaches designed to evaluate the process requirements of an
entire undisturbed river system (Arthington et al., 2004). The ecological flow
is defined as the water regime of a fluvial system necessary to maintain the
ecosystems and their benefits, in a condition close to the pristine condition and
that could eventually be assigned for activities that require the use of water
from a river system, through environmental, social and economic evaluation
processes (Dyson et al., 2003).
1
96
The methods developed to determine the ecological flow are based on qualitative
and quantitative indirect criteria of the functioning of the water body. These
indicators are necessary to estimate variations of the status or value by a range
of flow and, in some cases, by different hydrological periods. Taking into
account these criteria and the associated methods will allow the development
of strategies to which the ecosystem is expected to reach adequately (Pouilly
et al., 2014).
Within the application of the ecological flow, it is contemplated that the
ecosystem to be used has enough natural characteristics to be able to maintain
itself in terms of conservation. In other words, when the ecosystem suffers some
type of intervention (natural or external), the holistic method of ecological flow
cannot be applied. In this case, it is better to carry out restoration and purification
programs for the intervention before trying to use the water resource, taking into
consideration an alternative to use hydrological method of ecological flow, as
a first approach to the concept of Dyson et al. (2003), also given the conditions
of intervention of the system
According to the RAMSAR convention, wetlands require an adequate quantity
and quality of water to maintain their ecological characteristics and provide
ecosystem water-related services and benefits to human beings (Barchiesi, et.
al, 2018). In this connection, the Ramsar Wetland Convention recognizes that
water, wetlands and people are intrinsically connected and that all wetlands
have water requirements that are fundamental to regulate their water cycle.
Specifically, High Andean wetlands, within the Ramsar Convention, are
recognized as vulnerable, highly fragile ecosystems in which a flora and fauna
characteristic of these environments are recognized, reason why they should be
considered within the national strategy (RAMSAR 2005).
In the case of Bolivia, the South Lipez Bofedals are recognized as Ramsar sites
since 27 June 1990 with an area of 1,427,712 ha (RAMSAR N. 0489). The
conservation of our wetlands is of high relevance since Bolivia is recognized
for having the largest wetland area, for this reason the Ministry of Environment
and Water (2017) has established the: Strategy for the integral management
of wetlands and RAMSAR sites, promoting management principles, strategic
guidelines and strategic alliances for their conservation.
2
97
Depending on their nature and hydrogeological, physical, chemical and
biological characteristics, wetlands have different water requirements. For
example, it is known that bofedals require a high phreatic level to maintain their
functions of moisture retention, peat storage and flow regulation (Cooper et al.,
2015). In high Andean bofedals dominated by Distichia muscoides (Cordillera
Real and Occidental of Bolivia) a positive relationship was found between leaf
elevation and the mean phreatic level of bofedals. In fact, a site with a phreatic
level depth of nearly constant 100-cm for a year seems to have disconnected
the plants from their groundwater source, causing the surrounding plant-life to
wither.
In altered systems, the water regime varies in the time scale of hours, days, seasons
and years, controlling the ecology (its diversity and functioning). Therefore,
the following aspects must be taken into account for the water/flow regime:
the minimum and maximum flow, the occurrence-frequency of each flow, the
duration, the period with the specific flow, the regularity – predictability and
the rate/velocity of change. All these aspects produce alterations and therefore
the degradation of the ecosystem (Poff et al. 1997).
1.1. Optimal bofedal conditions
The natural characteristics of a stable bofedal are generally: peat depth, which is
formed in anoxic situations in cold environments, where the decomposition of
plants does not exceed production (Aerts et al., 2001). Bofedals are ecosystems
with vegetation that has a high organic matter content and, particular, peat that
relies on water. Peat has the capacity to retain water for longer times, functioning
as a sponge and ensuring the availability of water for plants—especially during
the dry season.
In the central zone of the Andes (between Bolivia and Peru), preserved
bofedals are formed mainly by plants that form cushions of the juncaceae
family (Distichia muscoides and Oxychloe andina) on which other species are
develop (Ruthsatz 2012, Salvador et al. 2014). These cushion plants retain their
withered leaves, are the basis for peat formation and best retain water compared
to other species (Lorini 2013, Benavides et al. 2015). Bofedal vegetation is
3
98
directly related to land and aquatic fauna, forming the fauna community
structure. Fauna inhabits, consumes and establishes its biological cycle as a
function of the resources that surround the bofedal.
The aquatic fauna that lives in the bofedal waterbodies is directly related
to vegetation, since the water dilutes what is stored in the riverbanks. For
example, the diversity of macroinvertebrates depends on the physical-chemical
characteristics of the water and microhabitat structures. These conditions
will be modeled on basis of the water characteristics, the soil and the type
of vegetation where the bofedal develops. According to the study by Oyague
and Maldonado-Fonkén (2015), some of the most important parameters for
the development of macroinvertebrate communities in the bofedal waterbodies
are vegetation cover, aquatic vegetation and water level. The availability
of water and, in particular, its depth level, is a key element in determining
the composition and abundance of macroinvertebrates in the community,
because these characteristics give way to different types of microhabitats for
the colonization of these organisms. Conversely, although bofedal vegetation
requires permanent water, the maintenance of the phreatic level is not necessarily
related to water depth (Oyague and Maldonado-Fonkén 2015). It is for this
reason that both the vegetation and the macroinvertebrate community should
be considered as bio-indicators to study the status of a bofedal.
1.2. Water requirement of the vegetation
Plant productivity is determined by the amount of water available in the
soil. Thus, to estimate how much water a plant needs to grow (produce new
biomass), it is necessary to have information on its water requirement (Medrano
et al., 2007). For instance, in a crop field, this relationship is estimated through
accumulated biomass (g/ha) and water used (m3/h). If this comparison is made
for a single species, it is possible to determine the water availability regimes as
a function of the water to be used.
A plant’s biomass production is the result of photosynthetic activity and
water expenditure (transpiration). This process involves the absorption of
water from the soil through the roots, the conduction of the water to leaves,
4
99
the entry of water into stomata wall, and water evaporation from stomata into
the surrounding air (Dingman 2002). The more open the stomata, the higher
the entry of CO2 necessary to produce photosynthesis (which is essential to
produce biomass), but, at the same time, the greater the loss by transpiration.
This water spent in the form of transpiration depends on the plant type and
environmental conditions.
Biomass production depends on the efficiency of transpiration. Therefore, to
understand the efficiency of the use of water, it is necessary to comprehend the
physiological processes that exist in the water flow of plants. Consideration
must be given to the availability of water in the soil (precipitation, groundwater,
etc.), root type, plant surface, and climate conditions. Currently, there is various
literature on the indexes that indicate the efficiency of water used by a certain
species; nevertheless, this type of studies was focused on cultivated species of
global importance.
Under the agronomic-hydrological perspective, water requirement can be
understood as the potential evapotranspiration (PET), that is, the amount of
water that evaporates from the soil and vegetation under optimal conditions,
without limitations to the availability of water (phreatic level, flow speed, etc.).
The PET in wetlands can be calculated using different methodologies such as:
Renewal of the surface (Paw and Brunet 1991), LIDAR (Eichinger et al., 2000)
both of which require the use of specialized equipment. Empirical formulas
such as Blaney and Criddle, Priestley and Taylor (1972) and others have been
determined in environments other than bofedales and their applicability is
questionable unless a calibration is performed (Drexler et al., 2004). Another
formula that has been successfully tested in wetlands is the Penman-Montheith
equation (Drexler et al., 2004). Despite having some limitations such as not
knowing all the parameters accurately (vegetation stomatal conductivity, leaf
area index), this is the most rigorous approach to evapotranspiration from the
physics of the process (Dingman 2002). The calculation of the PET requires the
data of the leaf index (Leaf Area Index), the stomatal conductivity (Federer et
al., 1996), and the climatic conditions of the area.
5
100
2. OBJECTIVES OF THE STUDY
General objective
To evaluate the water requirements of the Silala bofedals on basis of a
characterization and conceptualization of their hydrological and ecological
functioning, and the estimation of the water quantity and hydraulic convictions
necessary for their preservation in an ideal state.
Specific objectives
• To determine characteristics of the Silala Bofedals, taking into
account the studies carried out by DIREMAR;
• To deduce the alterations and impacts that the Silala wetlands have
endured, taking into account the studies carried out by DIREMAR to
date;
• The establish the relationships between the hydrological and
ecological variables (vegetation and macroinvertebrates) of the Silala
wetlands, identifying the most important variables for their conservation;
• To estimate the water demand required by the Silala bofedals in
Bolivian territory.
3. METHODOLOGY
3.1. Area studied
The Silala is located in Canton Quetena, in South Lipez municipality, within
the Potosi Department, between the 4,200 and 5,400 meters above sea level.
Geographically, the area studied is found at a mean south latitude of 22° 0’ and a
mean west longitude of 68°0’. The North Bofedal, with a surface area of 20,373
m2 and the South Bofedal, with a surface area of 90,503 m2 are both found
within this area (Figure 1). These bofedals constitute an important RAMSAR
site and are part of the “Eduardo Abaroa Andean Fauna National Reserve.”
The climate of the region is predominantly arid, characteristic of high mountain
desert areas. Throughout the year, it presents a high thermal amplitude, with
minimal temperatures of -19°C and maximum temperatures of 21°C, from day
to nighttime, respectively, and an annual average of 1.6ºC.
6
101
The lowest temperatures are recorded between May to August and the highest
ones from December to March (Figure 2). Precipitation is unimodal, with a
marked rainy season from December to March and an annual average of 125
mm/year. According to the Koppen climate classification, the area’s climate
falls within the category of cold high mountain climates (Montes de Oca, 1997).
Figure 1. Location of the Silala Bofedals, Potosi, Bolivia. The red dots represent
the sampling sites of the North bofedal, and the blue ones those of the South
Bofedal.
The region is characterized by a predominance of xerophytic puna systems
(Ministry of Foreign Affairs 2014, Rivas-Martinez et al., 2011), with a record of
86 species, subdivided into 65 genera of 35 families, where the Poaceae family
is the most diverse (FUNDECO, 2018a). An amount of sixty-one and seventyseven
springs have been found in the South and North bofedal, respectively
(TECHNICAL CONSULTANTS, 2018).
7
"'
~ C <>
~ 7 8 601000 601500 602000
0 600000 1200000 '!l')OOOO
Coordenadas Este (\)TM)
eooeoo 600900 &01000
Coo<denodo e,., <VTM)
E•r.ae!ont• cit ffl.,._.treo
(Bolt<l.a "<lf1•} •
... C-Ew(VtV):
rti
"'"" """ _.,E,.(UTM)
Est.acion.s d• -•1r.o
(8-ofMalau,) •
...,.
102
Figure 2. Climate diagram for 2018 (above) and 2017 (below) of the Silala
Weather Station. Prepared on basis of data from the National Meteorology and
Hydrology Service (SENAHMI, for its Spanish acronyms) Bolivia.
8
01.1 ·c
36
30
2S
20
,s
10
5
0
Ene Feb Mar
009'C
"
"
Silala
22'01' S I 68' 03' W, 4400 m
Ene,Ofc 2018
Apr May ..., Jul P,;o
22-01' $ I 61-03' W
4400m
2017
Sep
i 167.8mm
OC1 -0~
!: 10,nwn
70
60
m T p
·c mm
50
Ene 5.9 36
Feb 6.0 74
40 Mtlr 5,5 28
.A,.p,r 4.0 22 ·1.8 0
30 Jun -3.3 Jul •2.4 2'
Ago -2.4 2
20 Sep -0.0 0
-001 2.0 0 3.6 0
10 Cl,c 3.7 0
0
"
T • "' - '"' 00 0
" , .. 00 0
,".". '. "
....... ..0",8, •2 ,. .,e '0
,q, ... ... 0 11 0
.0.,0, " 0 32 0
"" ., "
103
3.2. Relation between the bofedal’s plant and macroinvertebrate
communities with environmental (hydrological and physical-chemical)
variables
A 1 x 1 m2 area was systematically selected in 20 sites for the North bofedal and
22 sites for the South bofedal, trying to take the representative microhabitats of
the bofedal in a representative manner (Figure 3). The coordinates of the sites
selected are presented in Annex 1. The work was carried out from 1 April to
5 April 2019. Biotic and abiotic data (hydric variables) were both taken from
each quadrant. At the same site, when there was an associated pool or the closest
waterbody, the macroinvertebrates were recorded and the physicochemical
variables were measured.
9
r-- ~~:-:-
i
!
~
I
L
--------------------------------Bofedai -s-.;,----------------
@
@ • 22
@
@
@ @
@
@
@ 6)
104
Figure 3. Sampling design to relate macroinvertebrate and plant communities
with relation to the environmental variables (hydric and physical-chemical). PN
indicates the piezometers installed in the North Bofedal and PS those installed
in the South Bofedal. Photographs by A.C. Simon Pfanzelt and Loly Vargas
Callisaya.
3.2.1. Vegetal and biomass cover
At each sampling site (see coordinates in Annex 1), the vegetation composition
was evaluated using a 1m2 quadrant—one of the most common sampling
methods for herbaceous vegetation (Stohlgren 2007). This method allows
recording almost all species and their cover in bofedals (Garcia 2004). 10cm2
of vegetal biomass was then extracted from the central part of each of the initial
quadrants. Each sample was labeled and stored in Ziploc® bags and deposited
in a container to then register their wet and dry weight at the National Herbarium
of Bolivia, Ecology Institute of the Higher University of San Andres.
3.2.2. Macroinvertebrate abundance and wealth
Samples of macroinvertebrates were collected in each of the determined
sampling sites or otherwise in the nearest water bodies. A Surber trawl net (15
x 15 cm, 0.02 cm2, with a 250 μm mesh opening) was used for the collection.
In total, 38 samples were collected, 18 and 20 in the South and North bofedal.
The samples were preserved in alcohol at 96% and evaluated in the Limnology
Unit of the Institute of Ecology of the Higher University of San Andres.
3.2.3. Physical-chemical variables
a) Soil moisture: At each sampling site, a soil sample was taken from the
central part of each quadrant. The soil samples were of 4 cm in diameter by 10
cm in length (Figure 4) and were collected at 3 different depths: 1) 10 cm, 30 cm
and 50 cm to determine moisture flow (Darcy 1856, Campbell 1974, Dingman
2002). The moisture flow calculation procedure is described in section 3.3.3.)
The soil samples from depths B and C were taken when it was possible to
extract the material. The soil samples were labelled and stored in Ziploc® bags
to preserve their moisture. At the end of the day, the wet samples were weighed
with a balance of 0.01 g. of precision (Denver Instrument, Canada).
10
105
Figure 4. From left to right: drilling, extraction and measurement of samples
to determine soil moisture.
b) Peat depth: Three peat depth measures (m) were randomly taken from
each quadrant, using a depth gauge (Aquatic Research Instruments brand).
c) Physical-chemical parameters: In the waterbodies associated with the
sampling sites the pH, electrical conductivity, and dissolved oxygen (mg/l and
%) were measured using Hach equipment (HQ 40D, USA, multi-parameter).
The distance to the sampling site and habitat characteristics (water body depth
(m), point width (m), velocity (m/s, (Global Water BA1100 model FP111 Flow
Probe, 3.7-6 Handle) were also measured as part of this study (Figure 5).
11
106
Figure 5. Field sampling at the waterbodies: a) physical-chemical variable
measurement, b) macroinvertebrate sample collection with the Surber net; and
c) hydric variables.
a) Piezometer installation: Piezometers were installed in the central part of
each quadrant, where the plants were evaluated to relate the soil moisture data
and phreatic levels with the vegetal cover characteristics. The present study is
not intended to monitor the regional flow of the groundwater in the bofedals,
but rather the local vertical flows.
The installation of the piezometers was carried out with the help of a drilling
machine - linear sampler and an auger (Figure 6) following the following
procedure: 1) the drilling-sampler was introduced, in some cases with the help
of manual percussion up to the depth of the sampler (30 cm); 2) sample is
collected; 3) the drilling-sampler is introduced again, and so on. Normally, it
was possible to recover 30 to 45 cm of soil, after which the introduction of
the drill was no longer possible due to the presence of clasts in the ground or,
in some cases, due to the resistance of the ground; 4) the drilling depth was
extended with the use of an augur to a depth sufficient to measure the water
table, and up to a maximum of 1.50 m.
The piezometers are constructed with a PVC pipe of 1.5 inches in diameter.
They were lined with micrometric silk (100 μm approx.) near the filter. In the
North bofedal, an additional piezometer was installed in the PN17 and PN29
sites, at different heights to observe pressure differences and to calculate the
vertical flow of water in porous material (Darcy 1856).
12
107
Figure 6. Piezometer installation in the Silala bofedals. Right side: expansion
of the hole made with augers. Center: piezometer installation. Left side: site
with the two piezometers installed (PN17).
The phreatic level was measured in each of the piezometers installed using a
manual probe. The measurements were made 24 hours or more following the
piezometer installation to ensure the representative of the measured values.
b) Piezometer calibration: Topographic measurements were made to
determine the location, terrain altitudes and the altitude of the upper part of
each piezometer in order to calculate the altitude and depth of the phreatic level
at each point (Figure 7). The measurements are linked to the polygonal base
installed by DIREMAR in 2017.
Figure 7. Topographical piezometer location and altitude measurement.
13
108
3.2.4. Laboratory data processing
The vegetable biomass samples were cleaned of excess organic matter and
manually separated at the level of genus or species. These were then weighed
by species using a precision balance Sartorius model Acculab in the facilities of
the Limnology Laboratory of the IE-UMSA. The samples were dried to reach a
constant weight in the facilities of the National Herbarium of Bolivia, IE-UMSA.
The macroinvertebrate samples were cleaned, separated and identified at
the highest taxonomic resolution possible in the Limnology laboratory
(Figure 8). The dissection plate assessment was used to determine the
structures of taxonomic importance, based on each taxon and with the help
of specialized bibliography (Dominguez and Fernandez 2009, Epler 2001).
3.2.5. Data analysis
a) Diversity indexes calculation: With the vegetation (subheading 3.1.2.)
and macroinvertebrate (subheading 3.1.3.) data, the Shannon diversity
index (H), Shannon Equality (EH), and inverse diversity of Simpson (Cinv)
were all calculated. The Inverse diversity of Simpson was calculated for
macroinvertebrates (Annex 7).
For macroinvertebrates, an index of bofedal health (IBH) was calculated from
the flow values (with the measures of velocity, depth and width) as a function
of the similarity to the communities with a composition reference of the sites
that were better preserved in the North and South Bofedals using the UPGMA
arithmetic grouping and Sorensen linkage methods.
14
109
Figure 8. Evaluation of macroinvertebrate samples in the Limnology-IE
laboratories.
b) Statistical analysis: A descriptive statistic was applied to explore biotic
vegetation and macroinvertebrate data, together with the hydrological abiotic
(phreatic level, electrical conductivity) and soil (soil moisture and peat depth)
data. These analyses were performed to characterize both bofedals.
A DCA (Discriminant Correspondence Analysis) and CA (correspondence
analysis) were used respectively to determine if there is any grouping in
the vegetation composition and in the composition of macroinvertebrates,
respectively. An ANOSIM Similarity Analysis was used to observe the
differences between macroinvertebrate communities of the North and South
bofedals.
To determine the relationship between the hydrological and biological variables,
the Canonical Correspondence Analysis (CCA) was used. This analysis allows
establishing the relationship between 1) the composition and biomass of the plants
and/or 2) macroinvertebrate composition of the North and South Bofedals with
respect to the environmental variables (phreatic level, soil moisture, peat depth,
distance from the canals, pH, electrical conductivity, dissolved oxygen, and
total dissolved solids). In the analysis of macroinvertebrates, the environmental
variables included the width and depth of the waterbody. With the variables that
showed some type of relationship, tests of significance were performed through
permutations to determine the significance of these relationships.
15
110
Analyzes were then run separately evaluating the influence of each environmental
variable with respect to the plant composition or macroinvertebrate composition.
The forward selection method was used to discard the variables that were not
relevant for vegetation and macro-invertebrates.
The multivariate analyzes were completed with the statistical analysis software
R 3.5.3. and the Correspondence Analysis and boxplot graphs to evaluate the
[effects of the] canalization on macroinvertebrates were completed in PAST (V.
316)
3.3. Estimation of water demand
3.3.1. Effect of the phreatic level on the vegetation
A preserved bofedal is mainly formed by species that compose its structure and
over which other species develop (Ruthsatz, 2012). We have chosen a cover
of O. Andina as a reference for a preserved bofedal patch. The effect of the
phreatic level on the coverage of O. andina was evaluated using a generalized
linear model (GLM) and assuming a binomial distribution. The independent
variables were groundwater level, peat depth, pH and electrical conductivity.
GLM analyzes and graphs were performed in R 3.5.3.
3.3.2. Potential evapotranspiration calculation (PET)
Evapotranspiration is a collective term for all the processes through which water
on the surface of the soil, or close to it, becomes part of the water contained
in the atmosphere in the form of vapor. The term includes direct evaporation
from the soil surface, transpiration through the leaves of plants, evaporation
from water bodies and sublimation from ice or snow surfaces (Dingman,
2002); In bofedals, it is assumed that the amount of evaporated water depends
on the amount of water available in the soil and environmental factors such
as temperature, radiation and relative humidity. Potential evapotranspiration
(PET) is the amount of water that enters the atmosphere assuming that water
availability is not a limiting factor (Medrano et al., 2007). From another
perspective, it is the amount of water that must be added to the bofedal to
maintain a constant water content.
In the present study we have used the Penman-Montheith equation to estimate
the evapotranspiration rate from a vegetated cover surface
16
111
incorporating the hydraulic conductivity of the foliage (Monteith 1965). This
equation has been successfully tested in wetlands (Drexler et al., 2004) and in
different environments (Calder 1977, Berkowicz and Prahm 1982, Lindroth
1985, Allen et al., 1998).
The Penman-Monteith formula is described below.
The atmospheric conductance (Cat) can be calculated as:
Foliage conductivity (Ccan) can be abstracted as the equivalent
conductivity of the set of leaves, with their respective
conductivity, working in parallel. It can be calculated as:
17
Where: I),. is the slope of the saturation vapor curve as a function of temperature; K + L = incident
net radiation; Pa= air density; Ca= specific heat of the air; e.* = air vapor saturation pressure; Wa =
degree of air saturation; Pw = water density; Av= latent heat of water evaporation; y = psychrometric
constant air; Ca1 = atmospheric conductance; Ccan = leaf conductance.
Where: Ya is the average wind speed at a height of 2 meters above ground level; Zd is the height of
the plane of zero displacement, approximately equal to 0.7 of the height of the vegetation; z0 is the
roughness of the terrain; it is calculated as 0.1 of the height of the vegetation; Z"' is the height above
the vegetation to which the wind speed is measured; 2 meters by default.
Ccan = fs X LA I X C,eaf
Where: f, = factor of shelter that takes into account that the leaves cover one another from the wind
and the sun. The values off, vary between 0.5 and I. The Leaf Area lndex is the relation between the
total surface of the leaves of the plants and the horizontal surface covered by vegetation. C1ear =
stomata( conductance of the leaves, is calculated from the maximum stomata( conductance C*1e,r
applying a series of reduction factors that allow taking into account the effect of different
environmental variables on the opening of the stomata: incident radiation (illumination), the deficit
of vapor pressure in the air, the temperature and the water deficit in the soil (Tobin et al., 1988).
112
The PET calculations were made with the climatic data available at the time
level from March 2017 to May 2018 obtained at the Silala weather station. Two
approaches or methodologies were used: 1) that of the crop coefficient and 2)
that of the simulation of the evapotranspiration process. To calculate the foliage
conductivity, the approximations were the following: To calculate the shelter
factor we used the value 0.9 since the vegetation does not produce much shade.
The LAI is calculated based on the amount of vegetation coverage over the
total coverage. To calculate the stomatal conductance, we calculate the model
using two approaches:
1) The crop coefficient (kc) for a surface covered with grass. Zea Mamani (2015)
based on experimental data with lysimeters established culture coefficients for
each month adapted to High Andean bofedals in Puno, Peru (Table 1). These
coefficients can be considered representative for the bofedals of Silala.
Table 1. Crop coefficient Kc for dry puna bofedals, based on Zea-Mamani
(2015)
2) Simulation of the evapotranspiration process (st); we have adopted a value
of LAI = 4 and a maximum stomatal conductance C *leaf = 6.6 mm / s according
to recommendations for fully covered wetland surfaces (Federer et al., 1996).
We propose the calculation of two PET scenarios (Table 2), according to their
potential (ideal conservation conditions) and current surface areas. In turn, each
of these scenarios was calculated using the crop coefficient adapted by Zea
Mamani (2015) and the simulation of the process of evapotranspiration (St).
Table 2. Proposed scenarios to determine the ETP of the Silala bofedales. The
symbol St stands for transpiration simulation
18
Site
~ Bofedal
~ Bofedal
Average Kc
Scenario
Pot.ential area
Current area
Jan
1.66
1.23
1.45
Feb Mar AEr May Jun Jul Aug
1.43 1.31 0.68 0.71 0.83 0.74 0.72
1.31 1.16 1.15 1.05 0.13 1.22 1.15
1.37 1.24 0.92 0.88 048 0.98 0.94
Bofedal Surface Vegetation cover
a
11.7933
0.7679
0.920
0.756
SeE Oct Nov
0.59 0.81 0.95
1.24 1.23 1.43
0.92 1.02 1.19
Methods used
Kc, St
Kc, St
Dec
1.36
1.53
1.45
113
3.3.3. Calculation of the vertical flow of water in the soil
This calculation was completed on basis of the soil moisture content and
phreatic level. Just as the recharge of water in the soil by infiltration produces a
downward movement in unsaturated flow, the water consumption of the plants
also produces an upward movement of water in the soil. This flow is reflected
in a depression of the capillary fringe, together with the establishment of a
characteristic humidity profile that can be established theoretically through the
solution of Darcy’s differential equation for unsaturated vertical flow:
Where φ is the soil porosity, φae is the input voltage of air in the soil, K*h is
the saturated hydraulic conductivity of the soil and b is a constant that depends
on the type of soil. The limitation in the application of these formulas is in: (i)
the uncertainties that exist in relation to the porosity, hydraulic conductivity
and the (ii) coefficient b of the soils, which present in general, large variations
from one point to another and in depth (e.g. Dingman, 2002). The impossibility
of determining the pore tension when the material above the water table is
saturated (capillary fringe), since in this zone the pore tension varies as long as
the humidity is constant. We have used the porosity and hydraulic conductivity
vales (Table 3) obtained from the work carried out by Orsag et al., (2017).
19
Where Ki,(0) is the hydraulic conductivity of the soil as a function of humidity 0 and p is the pore
pressure in the soil. Denominating as <p to the hydraulic head p/Yw, we have:
Knowing the volumetric moishlfe of the soil 0, the hydraulic pore head <p and the hydraulic
conductivity Ki, can be estimated using the empirical formulas of Campbell (1974):
114
Once the pore stress φ and the hydraulic conductivity Kh have been estimated,
the flow is estimated using the finite difference method applied to the Darcy
equation:
In some cases, the depth of the water table can be used as data, where it is
known that the pore stress φ=0 and that the hydraulic conductivity Kh is equal
to the hydraulic conductivity in saturated soil k*h.
At points where two piezometers are installed at different depths (PN17 and
PN29), the Darcy equation is also applied considering saturated conditions.
3.3.4. Calculation of the bofedal water level
To calculate the volume of water, we use the following formula:
V = a * h * P
Where, a = area of the bofedal, h = average height (depth of the peat); p =
porosity of the peat has a value of 0.57 based on the work of Orsag et al.,
(2017). We propose three scenarios of water volume for both bofedals based
on the potential and current area of both bofedals calculated by FUNDECO
(2018a) and the depth of the peat.
Scenarios:
a) Potential volume 1: It includes the potential area of both bofedals (11.79 ha)
and the maximum depth of peat reported (6.6 m) in the Negra Ravine (Muñoz
and Suarez 2019).
b) Potential volume 2: It includes the potential area (11.79 ha) and the
potential depth of the topographic relief measured from the difference between
the maximum and minimum levels in 40 transects. This was done based on
the potential area and the average depth of the bofedals measured in this work
(Figure 9).
20
Table 3. Soil variables used for the calculation of vertical flows. Source(~ et al., 2017).
Substratum Porosity cp [ m• / m'] Saturated hydraulic Dry ooit weight
conductivity K*~ [gr/cm•]
(cm/d]
Organic matter (peat) 0.57 2.29 0.87
Sand - Loamy Sand 0.47 0.99 1.46
115
a) Current volume: Includes the current area (0.76 ha) and the average depth
of the bofedals measured in the present study (Figure 9).
Figure 9. Profile of the bofedal showing the potential and current area. On basis
of the potential area, the dimensions were calculated to determine the potential
depth of the bofedal. The arrows indicate the maximum and minimum levels.
3.3.5. Ecological flow
The minimum ecological flow is that which is imposed on the new rights of
use of waters that are constituted in natural water channels. Its objective is to
avoid the abiotic effects, such as the decrease of the wet perimeter, the depth,
the velocity of the current and the increases in the concentration of nutrients
produced by the reduction of flow, significantly alter the natural conditions of
the channel (Boettiger, 2013).
In Bolivia we do not have a regulation that specifies the ecological flow that
ecosystems need. That is why in this study we carried out this calculation
according to our neighboring country (Chile). Its regulation for the determination
of the minimum ecological flow (MMA 2012), approved by Decree N° 14 of
22 May 2012, Article 3 establishes: For each month of the year, the minimum
ecological flow at the requested collection point will be determined considering
the flow rate equivalent to twenty percent of the average monthly flow (QMM)
of the respective surface water source. This value will have a maximum limit
of twenty percent of the annual average flow (QMA) and using
21
Profundidad de la turba en base a FUN DECO (2018)
···r ···r ·····1· ·····1·,t;~, .. ···r ·· ···r ·····
Profundidad de la turba actual
!
116
hydrological statistics of the last 25 years. Additionally, in Articles 6 and 7,
indicates that, in qualified cases, such as those in which risks are identified for
the habitat of such magnitude that compromise the survival of the species, a
higher value of minimum ecological flow may be set, which may not exceed
forty percent of the annual average flow (QMA).
For the Silala bofedals we use the output flow from 2013 to 2016. Based on
the data from the Silala station, to this average monthly flow we calculate
its monthly average flow of 20% (QMM). Then we calculate the difference
between the QMM and the QMA to calculate the minimum ecological flow.
Finally, given that seasonality does not influence the flow, we obtained the
average for the year. The details of the calculation are presented in results in
subtitle 4.4.5.
22
117
4. RESULTS
4.1. Characterization of the North and South Silala wetlands
4.1.1. Climate
The Silala area has an automatic meteorological station located at the PMASilala
Advanced Military Post (UTM WGS84, East 601944 m, North 7566064
m). The variables monitored by this station are precipitation, maximum,
minimum and average temperature, maximum relative humidity, minimum
and average, maximum and average wind speed, wind direction, atmospheric
pressure, insolation and average solar radiation. However, given the few data
(since March 2017) they do not allow a direct characterization of the climate in
the bofedals, so the Laguna Colorada and Sol de Mañana stations were taken
as support.
In Silala, rainfall has an annual average of 125 mm/year (Laguna Colorada
Station, 1980-1998). Most of the precipitation takes place between December
and March. The average monthly maximum temperatures are greater than zero
and show a low year-on-year variation between 12 and 20 °C depending on the
time of year and the minimum is less than zero all year round, with an interannual
variation between -5 and -20 °C. The average monthly relative humidity
has values lower than 50% most of the year. The low contribution of humidity
along with the high convective capacity of the wind is responsible for the low
levels of precipitation.
4.1.2. Hydrology
Flow Rates: The Silala bofedals are crossed by a series of drainage canals that
were built at the beginning of the 20th century. These have changed the flow
patterns, favoring the collection of a greater amount of water, to the detriment
of the bofedals. The North Bofedal is crossed by a network of 688 meters of
main canals and 1,112 meters of secondary canals, with a drainage density of
887 m/ha (IHH, 2018). The South Bofedal is crossed by a network of 2,021
meters of main canals and 814 meters of secondary canals with a drainage
density of 789 m/ha.
Based on data from SENAMHI (2017) and the DGA (2019) we have plotted
the outflow of Bolivia and the inflow for Chile (Figure 10) for the period from
2013 to 2016.
23
118
Both flows show little variation throughout the year, suggesting that there
is no marked effect of seasonality. On the other hand, the registered flows
could not be simulated through the hydrological modeling of the topographic
basin leading to the conclusion that the hydrogeological basin covers a larger
territory (TECHNICAL CONSULTANTS 2018). Both indicators indicate that
the bofedals are fed by deep groundwater sources.
Figure 10: Average monthly outflow of Bolivia and inflow to Chile. Both
calculations were made for the period 2013 to 2016 based on data from
SENAMHI (2017) and data from the DGA (Directorate General of Water,
Chile).
Water table: The depth of the water table varied between 0.1 to 0.5 meters
during the first week of April. The lowest values indicate that the water is close
to the surface. The North Bofedal presented the level of water closest to the
surface in comparison with the South Bofedal. However, in both bofedals the
values have a wide variation indicating that there is a lot of heterogeneity in the
amount of groundwater present in the bofedals (Figure 11).
24
200
150
'iii"
~
ii 100
".:C:.,l u
50
0
Caudales promedio mensuales (2013-2016)
Silala (Bolivia) y Siloli (Chile)
... e e I e e e e e e
•-------•--------►--♦-----------... ......... ,,"1 - --:I ..
ENE FEB MAR ABR MAY JUN JUL AGO SEP OCT NOV DIC
~ Silala (BOL) -◄-- Siloli (CHI)
119
Figure 11. Histogram of the water table of the first week of April 2019 for the
North and South bofedals. The red line indicates the average in each site. It
is appreciated that the median of the water table is 0.19 meters for the North
Bofedal and 0.22 meters for the South Bofedal.
4.1.3. Soils
The soils of the Silala bofedals are of alluvial origin with a superficial horizon
formed by organic matter (Alzerreca et al., 2001). Because this material is in
proportions greater than 20%, allows the retention of high amounts of water
during the rainy season (Orsag et al., 2017).
Depth of the peat: The North Bofedal presented a greater depth of peat compared
to the South Bofedal (average: 0.88 and 0.56 meters) respectively. The variation
in the North Bofedal was from 0.1-1.95 meters and in the South Bofedal from
0.25-0.8 meters (Figure 12). Annex 4 details the depths and coordinates of each
of the evaluated sites (Annex 6).
25
•·
3·
2· z
1 -
.!!
!! o-
! .!".!. •
3·
2· .,
1.
O·
00 01 02 OJ OA
Nrm lr~atJCo (m)
120
Figure 12: Depth of peat in the North and South bofedals. Number of samples
n = 22 north and 20 south
4.1.4. Physicochemical variables
In general, there are differences between the physicochemical variables evaluated
in the North and South bofedals (Table 5). The water in the South Bofedal is
more basic, with greater electrical conductivity and a greater concentration of
dissolved oxygen. In contrast, the North Bofedal has slightly acidic to neutral
conditions. The values of the physicochemical variables of each of the sites are
presented in Annex 6.
26
2.0
[
I 1.5 ;
~
i!
f 1.0
05 + 00
none our
Table 5: Values of each physical-chemical parameter of water for both bofedals
Variable
North Bofedal South Bofedal
Average (Range) Average (Range)
pH 7.48 (6.56-8.35) 8.70 (6.99-9.84)
Temperature (°C) 11.89 (8.7-16) 12.65 ( 4-20.6)
ORP• (mV) -15.99 (-68-3 1.8) -86.79 (-152.9-8.1)
Dissolved Oxygen (mg/l) 5.28 (1.82-8.07) 9.25 (3.61-16.6)
Electric Conductivity
0.15 (0.09-0.34) 0.56 (0.12-1.74)
(~cm)
TDS•(mg/1) 0.o7 (0.05-0 .05) 0.29 (0.06-0.87)
121
4.1.5. Characterization of the vegetation
The vegetation cover had an average of 84.68 and 48.9% in the North and
South bofedals respectively. Both the productivity (g) and the water content in
the plants (g) were higher in the North Bofedal compared to the South Bofedal
(Table 4).
Table 4: Averages and standard deviation of the coverage (%), productivity (g)
and water content (g) of the North and South bofedals.
Regarding the composition, the vegetation in the South Bofedal is quite
fragmented and even the deterioration seems to be greater compared to the
composition evaluated by FUNDECO (2018). Specifically, the southernmost
and middle section of the bofedal is dominated by Carex cf. maritima, data
that agrees with the evaluation of this year. In the evaluation of March
2018 it was observed that the central part of the bofedal was composed of
small patches of vegetation with less degradation of Festuca potosiana and
Oxychloe andina (this last cushion species is typical of bofedals). However,
this year the coverage of O. andina was smaller and in general the patches
were dry although this year was particularly rainier. The southernmost section
(reaching the water tank) was dominated by F. rigescens, followed by O.
andina (FUNDECO 2018), data that agrees with what was found this year.
In the North Bofedal, we recorded that the cover is mainly formed by Oxychloe
andina, Zameioscirpus muticus, Eleocharis atacamensis, Phylloscirpus
deserticola. According to the Principal Component Analysis (PCA) carried out
on the year 2018, the Festuca potosiana grass was shown as the dominant species
followed by O. andina, which reached a total coverage of 15%. Conversely, this
year there seems to be a change in the dominance of species. Our evaluation
sites were dominated by Zameioscirpus sp. and O. andina and thirdly by F.
potosiana (Figure 13).
27
Bofedal
North
South
Vegetable cover
(%)
84.68 ± 6.7
48.9 ± 11.42
Dry biomass
(g)
0.14± 0.6
0.11± 0.59
\Vater content in
plants (g)
0.4 ± 1.94
0.37 ± 1.91
122
Figure 13. Presence of some individuals of Festuca sp. of medium height on
the cushions of O. andina. North Bofedal
4.1.6. Macro-invertebrates associated with the bofedals
A total of 9,164 individuals of macro-invertebrates were obtained. Which are
grouped into 35 taxa, of these 31% corresponds to the group of non-insects
(Acari, Tricladida, Tardigrada, Crustacea, Nematoda and Oligochaeta) and
69% to the group of insects (Plecoptera, Trichoptera, Ephemeroptera, Diptera,
Coleoptera, Hemiptera). Compared to the previous report of macro-invertebrates
(FUNDECO 2018a), the information was supplemented and new taxa specific
to bofedals were added.
Diversity: The diversity and equitability indexes of each evaluated site are
presented in Annex 7. There are significant differences between the communities
of the North Bofedal and the South Bofedal (Figure 14), these differences are
mainly due to the presence of a higher density of Nematodes (Dorylalmus
sp.), Oligochaetes (Homochaeta sp.) and mites (Hydrozetes sp.). However, no
significant differences were found in relation to the indexes of diversity and
evenness (Cinv, H and EH).
The health index of the bofedal (ISB) was calculated, which refers to
the similarity using the Sorensen beta diversity index on an average of the
composition of the stations in the best state of conservation in this case the
PN7, PN21, PN23, PN24 and PN27 points (Figure 15). In this sense, the North
Bofedal presents a macro-invertebrate fauna more similar to the bofedal water
wells.
28
123
Figure 14: Correspondence Analysis (CA) and ANOSIM test of macroinvertebrate
communities.
4.2. Determination of the most determinant hydrological variables for the
vegetation and macroinvertebrates in the North and South Silala bofedals
4.2.1 Vegetation relationship with environmental variables
Vegetation composition: For the North Bofedal the global composition of plants
was not explained by any of the environmental variables (water table and depth
of the peat) according to the Canonical Correspondence Analysis (Trace = 0.18,
P = 0.26).
Despite the absence of a relationship, we observed a trend: O. andina,
Zameioscirpus sp. and Philloscirpus desertícola (species of bofedals) prefer
sites of greater depth of peat and low water table (close to the surface).
However, these conditions are not decisive for characterizing the vegetation at
least during the month of sampling (April).
29
ANOSIM; 9999 permutaciones, UPGMA-Bray Curtis; P<0,01
1,6
1,2
0,8
0,4
N
.; 0,0
<(
-0,4
-0,8
-1 ,2
-1,6
-2,0~---,----~-----,,----,-----,----,----r----.-----,-----~
-1,8 -1,S -1,2 -0,9 -0,6
Axis1
-0,3 0,0 0,3 0,6
124
Figure 15. Relationship between environmental variables and the plant
community of the North Bofedal. NivFrea = water table, Proft = peat depth. In
brown are the abiotic components, in blue the plants and in green the typical
species of bofedal. The acronyms of species are in Annex 2. The Canonical
Correspondence Analysis was applied.
The sites with the deepest water table, that is, with the lowest water availability,
are located at points N17 and N11. These sites correspond to places with greater
disturbance (with a higher proportion of Festuca rigescens and peat without
vegetation) and visually drier sites. Conversely, places with a lower depth of
water table are related to greater coverage of O. andina. Some points that are
grouped are sites N19, N23, N24, N27, which in turn have presence of species
associated with high water levels.
In the South Bofedal the composition of the vegetation was significantly
explained by the groundwater level and depth of the peat (Trace = 0.46, P <
0.05) (Figure 16).
30
Werspa
Deyvio
tur
Gensp Fesrig N12
Deys e
-1 ,2
N30
N3 h_cam N15 agua
N1 mu
roca carmar
Aresploboli
De~t~sti Lemsp
Ulmai uelo
-2 0 2
Eje 1(98%)
125
Figure 16: Relationship between environmental variables and the plant
community of the South Bofedal. NivFrea = water table, Proft = peat depth. In
brown are the abiotic components, in blue the plants and in green the typical
bofedal species. The acronyms of species are in Annex 2. The Canonical
Correspondence Analysis was applied.
Oxychloe andina has high coverage at sites S11, S31 where the peat depth is
high. Given the degradation process in place in the same sites, high coverage
of dead vegetation was also found.
Water content in plants and productivity Both the water content in the plant
(Annex 8) (g) and productivity (dry biomass g) (Annex 9) presented a similar
response to what was found with the composition of the species. In the North
Bofedal, the CCA with the water content did not generate a significant model
(Trace = 0.05, P = 0.98). The water content of O. andina does not respond to the
water table and the peat depth (Figure 17).
31
1,
nostoc
#
s
h t>urro
Liimac
Loboli
Deyspe
;.:.: ;o,0t--------------------=-===..,.,..=..:..=-----1.-afll!SIDL_ _____J cU_ ___ j iii'
agua S14 h_ave 83fegM Ni...Frea
-1,6 suelo
S11
r a proft Oxyand
RarlSp
-3,2 Deychr
.3 -2 -1 0
Eje 1 (99%)
126
Figure 17: Relationship between environmental variables and the water content
of the North Bofedal plants. NivFrea = water table, Proft = depth of the peat. In
green the typical bofedal species. The acronyms of species are in Annex 2. The
Canonical Correspondence Analysis was applied.
The water content in the plants (Annex 8) is positively associated with the
depth of the peat and intermediate conditions of the water table (Trace = 0.46,
P <0.05). While species such as Werneria spathulata and Eleocharis do not
require high depth of peat and are indifferent to the conditions of the water table
(Figure 18).
32
Lilmac
- 1,6
-3,2
-6
Fesrig
N3
Loboli N2
mu
ucfri
Carmar
N30
Dey pi
N1
proft
11
Phyde
Eie 1(90%)
NivFrea
Aresp.
Gensp
werspa
127
Figure 18: Relationship between environmental variables and the water
content of the plants of the South Bofedal. NivFrea = water table, Proft = depth
of the peat. In green the typical bofedal species. The acronyms of species are in
Annex 2. The Canonical Correspondence Analysis was applied.
4.4.2 Macro-invertebrate relationship with environmental variables
The variation between the environmental variables (physical-chemical water
and morphometric) can be explained by the first two main components by 68%
(Figure 19). The first axis separates the sites according to physical-chemical
variables pH, Dissolved Oxygen (DO) and electrical conductivity (EC) and
physical variables such as bed width. The second axis orders the sites according
to the variables such as average speed, average depth and flow (Q). At first
glance it is observed that groups are formed according to the North and South
bofedals, which differ according to the environmental variables. The sampling
points of the South Bofedal are more dispersed in relation to the sampling
points of the North Bofedal, indicating that these are more homogeneous with
respect to the mentioned variables. The scatter-grams of correlations between
all the variables measured in water are in Annex 4.l.
33
gra
1,2-
Nr/Frea
ii Werpyg N
;;;o,O- S16 7 ill- Pucfri Phydfl1
Fesrig
Platub
Aresp.
S8 S14 S23
-1 ,2-
Gensp
e
pa
Oeys
oboli
-2.4· L~~----:±.~-;------
- 1,6 -0,8 Eje 1ci(o98 %) o,8
128
Figure 19: Sorting of the sites according to the Principal Component Analysis
(PCA) according to environmental variables (physical-chemical water and
morphometric). The light blue points correspond to the sites of the South
Bofedal (PS) and the red points to the sites of the North Bofedal (PN).
In order to evaluate that the bofedals have different characteristics in terms of
environmental parameters, a boxplot was carried out (Figure 20). In the figure
it can be seen that there are significant differences (Kruskal-Wallis, Table 6)
between both bofedals, in terms of pH, Dissolved Oxygen (DO) and Electrical
Conductivity (EC).
34
*gi
N u
Q.
0
-3
0
pH
Profundidad
Caudal
0
0
-2 -1
• VelOGidad
0
PC1 39%
2
129
The cumulative variability of the relationship between environmental and
biological (macro-invertebrates) variables can be explained by the first two
axes by 53% significantly according to the Canonical Correspondence Analysis
(CCA) (Figure 21). In axis 1 the variables: width, depth of the water column and
flow (Q) (Table 7), in these sites predominate organisms such as Simulidae and
two taxa belonging to the order Trichoptera; that is, these organisms are present
in environments where water conditions provide speed and also maintain a
considerable depth to house coarse substrate. These taxa inhabit and develop
their metabolic activities on this type of material.
Axis 2 explains the variables pH, dissolved oxygen and inversely the depth
of the peat (Table 7), in these environments taxa such as Glossiphoniidae
predominate. These taxa are related to environmental conditions; in the case
of taxa associated with peat depth, they will be present in greater abundance
when the amount of peat is shallower, which is interpreted as a requirement
associated
35
pH DO EC
.
f---------i . I
- I ·l L
---- ---- c:::;::
- - . - - . - ~ .
Figure 21. Boxplot of each phys1cal-chem1cal vanable of both the North (Red) and South (Sky blue)
Bofedal.
Table 6. ~ -Wallis analysis of each physical-chemical variable as a function of the South and
North Bofedals. Value of p :S 0.001 • • •
Variable c1,;2 w;_ p - value
pH 16.60 I 4.6 e-5 •••
DO 16.71 I 4.4 e-5 •••
EC (~ cm) 19.23 1 1.2 e-5 • • •
130
with the dynamics of the water body’s bank. This variable is what explains the
tendency of separation of the North Bofedal, in conjunction with other variables
with less impact such as conductivity.
Figure. 22. Canonical Correspondence Analysis (CCA) show the sorting of
sites sampled in the area of the bofedals studied, according to the abundance
of the macro-invertebrate community and the most relevant environmental
variables. a) Ordering by visualizing the macro-invertebrate taxa and b)
ordering by visualizing the sampling sites in red in the North Bofedal sites and
in light blue ones in the South Bofedal, the vectors indicate the environmental
variables.
36
a
b
3 <.)
q
.,,
0
~
-3
0
0
0
0
Prom Hume
ofr
0
.... -- - . ·· a
-2
Cyp
Lan
EPH
Pod Cr.01
0
o •
proft
•
o,
• • • i••
•
_,
CCA1 31%
Mel
........................ Pf'l!ro_H~ro~.
TAB
-2
Co~dex Dist_pMI
MUS Ho : HCNJn~
Pl@~ q,..gr... H.liw ca,
~lau :corHCN2
SYR Alo EMP
-1
Ect
Hzt
TAR
Dor
[ ISO
0
CCA1 31%
□
□
131
Table 7. Contribution scores of each contrast variable, of greater weight in each
component of the CCA and statistics of significance with 999 permutations.
4.3. Relationship between vegetation and macro-invertebrate fauna
In the North Bofedal the sites PN7, PN21, PN23, PN24, PN25 and PN27
are related to the species: O. andina, Zameioscirpus, mosses and macroinvertebrates
such as: Dorylaimus (Nematoda), Homochaeta (Oligochaeta)
and Hydrozetes (Acari). Thus, this group of taxa is part of the biodiversity
of conserved bofedals. The sites PS03, PS06, PS17, PS20, PS26 and PN15
are associated with bodies of water with higher flow rates, which is related
to plant species such as: Lilaeopsis macloviana, Lemna sp., Eleocharis sp.
and macro-invertebrates such as: Simulium sp. Neotrichia sp., Andesiops
peruvianus, Hexatoma sp., Hyallela tiwanacu and Austrelmis sp. Finally, sites
PS08, PS23 and PS24 are shallow bodies of water that do not exceed 5 cm and
are related to Nostoc sp. (Cyanophyta), Cricotopus-Oliveiriella, Metrichia sp.
and Cyprinotus sp. (Fig. 24).
Based on the characteristics found in the sites PN7, PN27, PN25, PN27, based
on the macro-invertebrate fauna, the bofedal health index (ISB) was generated.
The index values vary from 0 to 0.2, from 02 to 0.4 and from 0.4 to 0.9. This
index compares the ideal composition of the fauna of macro-invertebrates in
conserved bofedals and the Silala bofedals. When comparing this index with
the abundance of O. andina, a good relationship between these variables could
be seen in the North Bofedal (R2 = 0.6, p <0.05) because it has some patches
with a better state of conservation (Figure 24).
37
Variable CCAl CCA2 F llL p - ~
~ (m) 0.87 0.26 6.56 0.001 ***
~Rsl2fu(m) 0.74 0.38 1.74 0.04 •
Q (m3/s) 0.61 0.35 1.24 0.28
pH 0.008 0.94 5.10 0.001 ***
OD (mg/I) -0.06 0.86 2. 15 0.004 **
T Rsl2fu -0.26 -0.52 1.44 0.11
132
2,0
1,6
1,2
0,6
~ 0 4
~ .
0,0
-0,4
-0.8
-1,2
~26 .... 11 H ..,.. • ft,11s
•._ · L- •P PS20
h burro .P06 ~17 •Au$ . ~ •~• PLSil0m3a t A pet 'HCJli
s..
-1,2 -0.9 -0,6
CYP
·=-·,, .,.,
no.toe
•Lan •cr-O .....
.,,_ ·e,,..-,,=
.TAS
bey.pl
• °"16
' Vague •G.n•p
·c"']M' 3
~ t~ GLO • 'PSt
•ooiwa We,pyg h_av~~ •eotmu
•Clau a.IP .-PN11••,.. Eel °l'N3o
wencpa-.»uc
&5o1 .,.~ o, "pN29A "514 -~
•.Jun,111 ·- 03 'h_c.am i:"e•rig A..106 ~ Ff.:e ••
-0,3 0,0
Ax~ 1
,, _......, 11' •.__ .,.PNt
Uo.lUUII • · - · l'N t 0
HQ<2 co:.,._~ TAR
Ptf,'l •
0,3 0,6
Vegetaci6n y
macroinvertebrados
de bofedal
·PseMUS
·ru'l>sn
•PS01 'R.ansp
""'26
~ 32 •~ ½JW10 SYR
~ "'°~'r~ t<Zl
o.•· 1,2
Figure 24: Con-espondence Analysis (CA) between the taxonomic proportion of vegetation
composition and abiotic coverage and macro-invertebrate fauna.
Figure 25 shows the spatial relationship between the ISB bofedal health index and the cushion
coverage of 0. andina
133
Figure 25. Vegetal cover of Oxychloe andina (%) in relation to the Bofedal
Health Index (ISB) in the South (above) and North (below) bofedals.
39
- 7566000
~
1-
2-
L.
:::,
(I) 7565900
(J)
(1J
"O
(1J
C
<l)
~ 7565800
0
8 0
ISB
0 to 0.2
• 0 .2 to 0 .4 e 0.4 to 0 .9
i' 7566400
1-
2-
L.
:::,
(/)
(J)
(1J -g 7566300
C
<l)
~
0
0
0
7566200
ISB
Canal principal
602800 603000 603200
Coordenadas Este (UTM)
Oxychloe andina (cobertura %)
0 to 1.5
o 1.5 to 30
0 30 to 92.01
Canal principal
600800 600900 601000
Coordenada Este (UTM)
• 0.1 to 0.4
Oxychloe andina (cobertura %)
0 to 5
• 0.4 to 0.6 e 0.6 to 0.9
o 5 to 40
0 40 to 65.01
134
4.1. Estimation of the water demand of the Silala bofedals
The water demand for the maintenance of bofedals has been determined by
evaluating different aspects:
1) O. andina relationship with respect to environmental variables.
Since O. andina is a typical species and forms the bofedal structure,
knowing how environmental variables and especially water influence
their coverage and biomass gives us guidelines to understand the
minimum requirements needed by plants that form the bofedal structure.
2) Calculation of the amount of water needed to maintain both
bofedals (ETP) from the potential and current condition. We propose two
values from the potential and current area, see scenarios subtitle 3.3.2).
3) Estimation of the water volume of the bofedals. The volume that the
bofedals can store will be the result of their area and depth. We calculate
this volume based on its current and potential area, see scenarios subtitle
3.3.3)
4) Calculation of vertical flow. It is related to the amount of rising
water that is the reflection of the evapotranspiration process.
5) Calculation of the minimum ecological flow. We present a
minimum flow value based on Chilean regulations. However, this is still
approximate data that should be studied in greater detail in future work.
4.4.1. Coverage ratio of O. andina (typical species of bofedals) vs.
environmental variables
For the North Bofedal the models that better explain the coverage of O. andina
are the depth of the peat, pH, electrical conductivity and flow (Table 8).
Table 8: Models that affect the coverage of O. andina.
The peat depth (above 0.5 m) has a positive effect on the coverage of
O. andina (Figure 23 a). O. andina coverage reduces when the pH level
increases. A greater coverage of O. andina is related to low levels
of electrical conductivity that does not exceed 0.15 mS/cm. Contrary
to our expectations, the coverage of O. andina was not related to the
water table. However, it seems that the water table fluctuates between
40
Variables □~ ~ %)
Peat depth + pH + electrical conductivity 232 0 80
135
between 0.1 and 0.3 meters for O. andina. The South Bofedal data had high
dispersion (chat value = 1291.6) and did not comply with the assumption
of variance homogeneity, so the GLM analysis with binomial and/or quasibinomial
distribution was not performed.
Figure 23: Coverage of Oxychloe andina with respect to environmental
variables in the North Bofedal. Dispersograms elaborated based on the variables
that best explain the coverage according to the binomial GLM and logistic link
function.
41
a) b)
100·
!O· . j . . "' I .,. ~ ! ~ ,..
g ii ii'
0
~ e
J e •· ~
-25·
o'o o's 1'0 ,'s i'o o'o o", 0'2 o', •• Profllnd~ed cit turl>• (m)
Nlvol fro611co (ml
c) d)
... ,.. .
i
...
i
1, ... I"· !
! ,.. j i (J ,. I •·
0
1
•20·, 11) o',, ,lo o2, ,lo ,'o ,', ,', ConducCMdad ol6c:Cr!Q (m$Jc:m) .. pH
136
Spatial representation
The representation of some environmental variables in the South Bofedal
suggests the low correspondence with the increase in O. andina coverage and
the depth of the peat, water table, pH and electrical conductivity (Figure 24).
Figure 24: Vegetal cover of Oxychloe andina (%) in relation to the depth of the
peat (m), water table (m), pH and electrical conductivity (mS/cm) in the South
Bofedal.
In the North Bofedal (Figure 25), the peat depth where the coverage
of O. andina varies from 30-90% corresponds to sites from 0.15 m of
peat. The water table closest to ground level (between 0.15-0.31 m) is
associated with sites with greater coverage of O. andina. In sites with low
42
~ Profundidad de la turba (m) ~ Nivel freatico (m)
,,.... 7566000 o to 0.3 0 to 0.17
~ I- • 0.3 to 0.6 • 0.17 to 0.26
-J I 0.6 to 0.9 • 0.26 to 0.49 I.. :,
(J) 7565900
r/J
111
'O
111
C:
Q)
'E 7565800
0
0 ( 0 ()
1.5 to 30
Q 30 lo 92.01 Canal principal Canal principal
~ pH ~ Conductividad (mS/cm)
,,.... 7566000 6 to 7 0 10 0.28
~ • 7 to 9.01 • 0.28 to 0.62
I- • 9.01 lo 9.85 e 0.62 to 40.27
2.
I.. :,
IJl 7565900
r/J
ro
'O
ro
C:
ID
'E 7565800
0
0 () 0 0
Canal principal Canal principal
602800 603000 60320: 602800 603000 60321)(
Coordenadas Este (UTM) Coordenadas Este (UTM)
137
pH (slightly acid to neutral), the sites with the highest coverage of O. andina are
found. While the correspondence is not so clear with the electrical conductivity.
Figure 25: Vegetal cover of Oxychloe andina (%) in relation to the depth of the
peat (m), water table (m), pH and electrical conductivity (mS/cm) in the North
Bofedal.
4.4.2 Potential evapotranspiration (PET) calculation. Value of the
water requirements for the preservation of the Silala bofedal.
43
~ Profundldad de la turba (m) ~ Conductividad (mS/cm)
~ 7566400 • 0 lo 15.01 • 0 to 0.12
f- • 15.01 lo 30.01 • 0.12 to 0.15 J I 30.01 lo 46.01 I 0.15 to 0.34
'-' ...
::,
r/)
1/J
t1l
~ 7566300 Oto 5
C S to 40
Q) 0 40 lo 65.01 u...
0
0
0 = = Canal principal Canal principal
1i;i:~?nn
~ pH ~ Nivel freatico (m)
~ 7566400 6 to 7 • 0 to 0.15
f- • 7 to 8.1 • 0.15 to 0.31
J e 8.1 to 8.4 I 0.31 to 0.45
'-' ...
::,
r/)
1/J
t1l
~ 7566300
C
Q)
u...
0
0
0 = = Canal principal Canal principal
7566200
600800 600900 601000 600800 600900 601000
Coordenada Este (UTM) Coordenada Este (UTM)
138
ETP for the current area: For the scenario of extension and current state of
bofedals (0.76 ha, vegetation cover of 76%), the average annual flow required for
maintenance is 0.34 l/s (Figure 26). Using the crop coefficient or the simulation
of evapotranspiration and crop coefficient, they show similar results. This ETP
value differs over the months.
ETP for the potential area: In the scenario of potential extension of the bofedals
(11.79 ha, vegetation cover of 92%), the flow rate for the conservation of the
Silala bofedals is of 5.9 l/s on average annually, with a fluctuation between 3
and 9 liters per second between the dry and wet seasons respectively (Figure
27). Using the Crop or evapotranspiration simulation coefficients, both indicate
comparable values, which reinforce the reliability of our results. Highlighting
that the months that need a greater flow are between November and February.
More details can be found in Annex 10.
Figure 26: Water requirement for maintenance of the Silala bofedals in liters
per second. In black the results are shown for the current area and in red for the
potential area.
Annual evapotranspiration (ETP) is 1256 to 1563 mm/year in the current scenario
and 1523 to 1623 mm/year in the potential area scenario. In both cases the lowest
44
Requerimiento de agua para mantenimiento bofedales
12
10
vi' 8
:=:
ii 6
-a ::, a 4
2
0 ------
--Actu<1I con simulaci6n --Potenci<1I con simulaci6n
- - - - Actu<1I con Kc - - - Potencial con Kc
139
value corresponds to the cultivation coefficient method and the highest to
the evapotranspiration simulation. The monthly variation of the monthly
evapotranspiration in the bofedals, throughout the study period, is shown in
Figure 27. The summary of the ETP calculation for the four developed scenarios
is presented in Annex 7.
Figure 27. Evapotranspiration in the Silala bofedals in millimeters/month. The
results are shown in black for the current area and in red for the potential area.
4.4.3 Calculation of the water volume of the bofedal
Based on our formula and proposed scenarios we obtained three water values
(Table 9) that the bofedals need. Under the first scenario with a potential area
and an average peat depth of 6.6 m, we would expect the water level to have
a minimum of 444 m3. A second potential volume but adjusting to an average
depth of 4.7 m, the volume required for the bofedals is 354 m3. Finally, the
current value at which the bofedals are likely to be found is 48.4 m3.
45
250
200
E 150 ..s
~ 100
50
0
Requerimiento de agua para mantenimiento bofedales
---Actual con simulaci6n --Potencial con simulaci6n
- - - - Actual con Kc - - - Potencial con Kc
140
Table 9. Potential water volume values in the three scenarios.
4.4.2 Vertical flow calculation
The calculation of vertical flow in the soil from soil moisture and groundwater
data shows a predominance of upward flows, which are a reflection of the evapotranspiration
process. In the North Bofedal, calculated flows range from 0.8
mm/day to 9.9 mm/day, averaging 3.4 mm/day. In the South Bofedal the calculated
flows vary between 0.1mm/d and 8.1 mm/d, averaging 2.2 mm/d. Table
10 shows the results obtained. The evapotranspiration (ETP) for the month of
April in the current situation is 131 mm/month, which would be equivalent to
a vertical flow of 4.37 mm/day for both Silala bofedals. The calculation of the
vertical flow is presented in Annex 8.
Table 10. Vertical flows estimated from humidity and water table measurements.
46
Scenar ios
Average Volume
depth (m) (m3)
Potential volume 1 6,6 443,664
Potential volume 2 4,7 353,811
CmTent volume 0,7 48,399
North Bofedal South Bofedal
Point Estimated flow Point Estimated flow
[mm/dl [mm/dl
PN-5 0.8 PS-I 2.0
PN-9 2.1 PS-3 3.7
PN-11 1.5 PS-6 3.7
PN-12 5.5 PS-14 1.7
PN-13 -0.3 PS-16 2.2
PN-17 6.8 PS-17 0.1
PN-27 9.9 PS-18 1.4
PN-29 1.4 PS-20 1.3
PN-30 2.4 PS-21 0.0
Avera2e 3.4 PS-31 0.1
PS-32 8.1
PS-33 2.1
Averae:e 2.2
141
4.4.5 Calculation of the minimum ecological flow
In the study area, there are flows with little or no seasonal and inter-annual variation
(TECHNICAL CONSULTANTS 2018). That is why the available 2013-
2016 flow series are representative and can be used in the calculation of the
minimum ecological flow. Table 11 shows the calculation and results obtained.
Table 11. Calculation of the minimum ecological flow according to regulations
to determine the minimum ecological flow. Decree 14 (MMA, 2012).
The calculated minimum ecological flow has little seasonal variation, so it can
be reduced to an average value of Qecol = 33.4 l/s.
47
Month Averqe 10%QMM 10%QMA Q..a
monthly flow [Vs] [Vs] [Vs]
OMMWsl
JAN 168.9 33.8 34.9 33.8
FEB 173.2 34.6 34.9 34.6
MAR 172.3 34.5 34.9 34.5
APR 177.4 35.5 34.9 34.9
MAY 178.1 35.6 34.9 34.9
JUNE 179.1 35.8 34.9 34.9
JULY 177.7 35.5 34.9 34.9
AUG 181.9 36.4 34.9 34.9
SEPT 178.7 35.7 34.9 34.9
OCT 171.3 34.3 34.9 34.3
NOV 167.2 33.4 34.9 33 .4
DEC 167.1 33.4 34.9 33 .4
174.4 <--- Annual Average Flow QMA
142
5. DISCUSSION
5.1. Bofedal characterization
The depth of the water table was greater in the South Bofedal compared to the
North Bofedal, suggesting lower water availability in the South Bofedal. In
spite of having a punctual data for April, our data are similar to those obtained
by the study by Orsag et al. (2017). Our evaluation gives us a general idea
about water availability and how it varies in different points of the bofedal.
However, the water table must be monitored at least during one hydrological
cycle (Table 12). This is because the water level can be affected by the annual
and eventual hail of the zone.
The variability of the water table depth was high in the Silala bofedals. This
suggests that water availability is not the same for the entire plant community.
Wetlands and bofedals in good conservation status are closely related to nearsurface
water levels and these values are almost constant throughout the bofedal
(Lorini 2013, Cooper et al. 2019). Therefore, our data point to unfavorable
water conditions, which affects peat formation, plant species composition and
macroinvertebrates.
Average peat depth data were less than 1 m, values similar to those found by
Orsag et al. (2017) also for the Silala (Table 12). However, our values are
somewhat lower compared to the UERH-COFADENA study (2017) which
suggests an average thickness of 1.3 m for the South Bofedal and up to 4 m in
the deepest parts of the North Bofedal. We take several sampling points trying
to cover the whole bofedal so we consider that we have high reliability of these
data.
An average peat of less than 1 m indicates degradation of this ecosystem and its
functionality. This is because the conserved bofedals have a much higher level
of peat. For example, the peat depth reaches between 3 and 4 m in Aychuta,
Sajama (Meneses et al. 2014) or between 4.4 and 6.6 m in Quebrada Negra,
Chile (Muñoz and Suárez 2019). Peat production is the result of the balance
between plant production and its decomposition (Hribljan et al. 2015, Cooper
et al. 2015). Thus, alterations in vegetation and reduced water availability can
alter the carbon sink in peat and function more as a resource (Limpens et al.
2008).
48
143
That is to say, to cause an increase of CO2 gases in the atmosphere. In addition,
it causes a decrease in water retention capacity.
Table 12. Average value and standard deviation of the water table in the Silala
bofedals *Data extracted from Orzag et al. (2017)
5.2. Relation between the hydric conditions and the macroinvertebrate
and plant community
In the spatial analysis taking into account the important variables it was
corroborated that pH and dissolved oxygen are the most important variables
to determine not only the key characteristics of each bofedal but also the
association with the communities of macroinvertebrates. As in many other
aquatic systems, oxygen availability is an important factor in defining the trophic
status of habitats that influences the aquatic composition of macroinvertebrates
(Wetzel 2001, Domínguez and Fernández 2009).
The depth of the water column is also a determining factor for the
macroinvertebrate community. In bofedals it is a critical factor because it affects
microhabitat availability, water quality, connectivity and water exchange rate.
It therefore influences the composition and richness of macroinvertebrates. It is
therefore essential to maintain a depth appropriate to the natural characteristics
of the bofedal.
Each of the bofedals that we evaluated had a particular environmental
heterogeneity, given by the physicochemical conditions of the water as well
as its diversity. Like the Peruvian bofedales (Oyague and Maldonado-Fonkén
2015), the Silala bofedales are important for the physical characteristics of
the habitat before the chemical ones. This shows that the conditions of the
riverbank and its relationship with the vegetation of the bofedal have a great
weight on the community structure.
49
Bofedal
North
South
Water table (m)
2017 (Average± D.E.)* 2019 (Average± D.E.)
0.26±0.14 0.19±0.12
0.34 ± 0.17 0.23 ± 0.12
144
The peat depth showed high association with a large group of macroinvertebrates
including Dugesia, Oligochaeta and several Diptera. This result suggests
the existence of preserved patches in the North Bofedal. The relationship
with the aquatic fauna is direct and has to do with the degree of vegetation
cover as mentioned by Oyague and Maldonado-Fonkén (2015). The bofedal
dynamics is very complex, but the relationship between the vegetation and
the macroinvertebrate community goes hand in hand, so that the degree of
conservation in which the bofedal is found will also be reflected in the structure
of the aquatic communities. Where those taxa that are more adapted to this
type of environment will be present in greater abundance. Coincidentally, the
results with the vegetation suggest that peat depth is an important factor in the
structuring of the communities in the North Bofedal, although we did not find
that this variable is important to explain the variation in species cover in the
South Bofedal.
The water table did not show a significant effect on the composition of plants,
both in the North and South bofedal, probably because the point measurements
we made are not yet sufficient to explain the coverage of plants, in addition the
degraded state of many sites evaluated may be masking the role of this variable
in the structuring of plant communities.
Contrary to multivariate results (which take into account all species and various
environmental variables), analyses focused on understanding the coverage
of Oxychloe andina according to the environmental variables measured, we
determined that the peat depth and the physical-chemical conditions of the
water (pH and electrical conductivity) are important factors that explain the
coverage of this species in the North Bofedal.
5.3. Water requirement of the Silala bofedals
The species O. andina forms the bofedal structure, so its cover and biomass
can be used as indicators of representative bofedal sites. Our results show a
positive relationship between the cover of this species and the peat depth. This
suggests the importance of the productivity of this species for the formation and
accumulation of peat. Our analyses also indicate that basic pH conditions and high
electrical conductivity are limiting factors for O. andina. This suggests that O.
50
145
andina in the Silala bofedal prefers slightly acidic to neutral conditions although
it has a wide tolerance range (Meneses 1997).
In our study we did not find a direct relationship between the water table and soil
moisture with the cover and biomass of O. andina. It is possible that a specific
data as in our case is not enough to understand the relationship of this plant in
cushion with the water level. However, Lorini (2013) showed that places with
greater coverage of Andean O. are closely related to water table levels between
0 and 0.8 m in a preserved bofedal (Aychuta, PN Sajama). While in a slightly
disturbed bofedal (Lagunas, PN Sajama) O. andina is generally found between
0-0.5 m. In addition, according to Lorini (2013) throughout the year, sites with
O. andina cover remain more constant and homogeneous compared to sites that
are composed of other species. This shows the water retention capacity of this
species, which is similar throughout the year. In Silala, the interventions of the
canals reduce the water flow even in places where the cushions are, which is
why there is no clear pattern. It is hoped that future data from the Silala on the
water table will help us to better understand this relationship.
The amount of water required by the Silala bofedals to maintain their condition
(ETP) is: 1523 to 1653 mm/year, equivalent to an average annual flow of 5.9
l/s. This ETP value only shows the volume needed to maintain the system once
an adequate volume has been reached. The values we obtained were almost
double the one calculated by Muñoz and Suarez (2019) for the Quebrada Negra
(1653 versus 698 mm/year), but their data were based on indirect methods
(NDVI). However, our calculation is comparable to the work of Zea Mamani
(2015), who used the direct method (lysimeter) to calculate ETP in dry puna
bofedals in Peru. Thus, our results are very close to the reality of the system.
But to have a greater approximation, future studies must evaluate physiological
variables of the bofedal vegetation as estomatic conductance of the species and
its Index of foliar area.
The evapotranspiration values calculated by DHI (2018) for the current and
restored bofedals conditions reach 125 mm/year and 164 mm/year respectively.
These values are not comparable with our results because they cover the entire area
known as the Near Field, with a total surface area of 2.5 km² or 250 ha, including
51
146
a large expanse of desert, while our results refer to bofedals only up to a
potential surface area of 11.79 ha. On the other hand, according to DHI (2018),
the evapotranspiration flow between the situation of restored bofedals and the
current (reference) situation increases by 3 l/s (from 10 l/s to 13 l/s). In our
study, the increase in evapotranspiration flow from the bofedal between the
same scenarios is 5.56 l/s (5.9 l/s - 0.34 l/s). If we introduce our result in the
water balance calculated by DHI, the flow to Chile will be reduced by 2.5 l/s
more than predicted by DHI (2018).
Under current conditions, the Silala bofedals have barely 10% water volume
with respect to their potential capacity. In other words, they need to reach an
adequate volume (via recovery and conservation) in order to allow the recovery
of the peat and the plant community. In time scale, reaching an adequate water
level does not mean the same time scale for the vegetation and peat recovery.
The plants that form the bofedal structure grow at a rate of less than 10 cm per
year as observed for Distichia muscoides (Cooper et al. 2015). And in the peat
case, the production of plant matter and favorable (anoxic) and water conditions
are needed to achieve proper functioning. This process could take more than
ten years until both bofedals are restored.
The vertical flow of water in the soil, estimated from the moisture content
in April 2019, varies from 2.2 to 3.4 mm/day; these values reflect the actual
evapotranspiration in the Silala bofedals and can be compared with the value
of the potential evapotranspiration calculated for the same month, of 4.37 mm/
day, for an ideal condition without hydric stress. The lower values obtained
in the field show that the bofedals are currently in a state of hydric stress due
to the water table depression as a consequence of the drainage of the bofedals
through the canals.
We calculate the ecological flow according to Chilean regulations. This
estimated value for the Silala bofedals corresponds to 33.4 l/s. According
to this regulation, it is the minimum needed to restore potential conditions,
guaranteeing adequate hydraulic conditions for the development of peat and
vegetation. However, this flow value could increase considering that we are in
an arid region and the bofedals need to be restored first.
52
147
6. CONCLUSIONS
Based on our results (abiotic and biotic variables) we conclude that both Silala
bofedals are in a state of degradation, particularly the South Bofedal. The peat
depth of both bofedals varies between 0.1-1.5 m in the North Bofedal and
between 0.25-0.8 in the South Bofedal. This value is much lower compared
to neighboring and better preserved bofedals that have an approximate of 4
m. The water table depth is high and variable in most of the sites evaluated,
indicating heterogeneity in the habitat and non-ideal conditions regarding the
water level for the plants. In accordance with these degradation conditions,
abiotic cover (bare soil, saline crust, dead vegetation) in the South Bofedal
exceeds 50% and 15% in the North Bofedal. The South Bofedal is dominated by
species indicating disturbance (Carex cf. maritime) and the North is dominated
by Zameiocirpus sp. with some patches of Oxychloe andina. This last species,
in spite of being typical of bofedal, is in the middle of the high presence of
graminoids.
Preserved bofedal sites are related to a composition of macroinvertebrates
associated with bofedales (Dorylaimus, Hydrozetes, Homochaeta), in addition
to the Silala bofedals, are positively related to the well depth and peat, providing
data on their indicator value of bofedals in good condition.
The vegetal composition of the North Bofedal was not related to the
environmental variables (hydric). While the South Bofedal were related to the
water table. Since there is no relationship with the environmental variables in
the North Bofedal, this does not mean that water availability is not relevant. It
should be noted that the water table is a dynamic variable and the water value
can be influenced by local events such as hail and it is necessary to know its
fortnightly fluctuation ideally. The lowering of the water table at the site also
caused the greater gradation of organic matter that reduced the peat depth at
the site. Peat depth is an important factor according to our results, we found
positive relationship between O. andina cover and peat depth, supporting the
role of peat in water retention to maintain healthy O. andina cushions.
It was determined that the vertical upward flow of water in the soil
predominates in the Silala bofedals, showing that the evapotranspiration
53
148
process predominates and not the water infiltration process, supporting the rest
of the results on the degradation of the site.
The water requirement to maintain bofedals in potential conditions is 1523
to 1653 mm/year and in current conditions 1256 to 1563 mm/year. These
values are comparable between methodologies used (culture coefficient and
evapotranspiration simulation) and also according to direct measurement works
in similar bofedals.
The volume of water currently retained in the peat of the Silala bofedals is 48.4
thousand m3, only 10% of its capacity. However, this level should reach a value
between 353,811 - 443,664 m3 to reach its potential volume. The time in which
both bofedals can reach their ideal condition is unknown.
The minimum ecological flow rate for the Silala bofedals is 33.4 l/s. This value
is based on Chilean regulations. However, the best adaptation must be sought
for an arid system, which first needs specific hydric conditions for its recovery.
Considering the area estimated by FUNDECO (2018a), the volume that the
bofedal should reach is approximately 353.8 thousand m3 and up to 443.7
thousand m3 considering the peat depth of Quebrada Negra, values that should
be reached before establishing the real ecological flow for the Silala bofedals.
The water requirement to maintain the Silala bofedals in the potential conditions
(11.79 ha) is 1523 to 1653 mm/year, which is equivalent to a flow of 5.9 l/s in
annual average, with a fluctuation between 3 and 9 liters per second between
the dry and wet seasons respectively.
54
149
7. REFERENCES
Aerts, R; Wallen, B; Malmer, N; De Caluwe, H. 2001. Nutritional constraints on
Sphagnum-growth and potential decay in northern peatlands. Journal of Ecology
89(2):292-299. DOI: https://doi.org/10.1046/j.1365-2745.2001.00539.x.
Allen, RG; Pereira, LS; Raes, D; Smith, M. 1998. Crop evapotranspiration-
Guidelines for computing crop water requirements-FAO Irrigation and drainage
paper 56. Fao, Rome 300(9):D05109.
Alzerreca, H; Prieto, G; Laura, J; Luna, D; Laguna, S. (2001). Characteristics
and distribution of bofedals in Bolivia. La Paz, Bolivia, Integral Association of
Camelid Cattle Breeders in the High Andes (AIGACAA).189 p.
Arthington, AH; Tharme, RE; Brizga, SO; Pusey, BJ; Kennard, MJ. 2004.
Environmental flow assessment with emphasis on holistic methodologies. In:
Proceedings of the Second International Symposium on the Management of
Large Rivers for Fisheries. 2:37-65.
Barchiesi, S; Davies, PE; Kulindwa, KAA; Lei, G; Martinez Ríos del Río, L.
2018. Application of ecological flows with benefits for society and different
wetland ecosystems in river systems. s.l., Ramsar Policy Note N°. 4. Gland,
Switzerland: Secretariat of the Ramsar Convention.
Benavides, J.C. 2014. The effect of drainage on organic matter accumulation
and plant communities of high-altitude peatlands in the Colombian tropical
Andes. Mires and Peat, 15(01): 1-15
Berkowicz, R; Prahm, LP. 1982. Evaluation of the profile method for estimation
of surface fluxes of momentum and heat. Atmospheric Environment (1967)
16(12):2809-2819. DOI: https://doi.org/10.1016/0004-6981(82)90032-4.
Blaney, HF; Criddle, WD. 1950. Determining water requirements in irrigated
areas from climatological and irrigation data, USDA SCSTP-96. US Department
of Agriculture, Washington, DC.
Boettiger, C. 2013. Ecological or minimum flow: regulation, criticism and
challenges. Minutes of Water Law No 3:1-12.
Calder, IR. 1977. A model of transpiration and interception loss from a spruce
forest in Plynlimon, central Wales. Journal of Hydrology 33(3-4):247-265.
DOI: https://doi.org/10.1016/0022-1694 (77)90038-5.
55
150
Calder, IR. 1978. Transpiration observations from a spruce forest and comparisons
with predictions from an evaporation model. Journal of Hydrology
38(1-2):33–47.
Campbell, GS. 1974. A simple method for determinig unsaturated conductivity
from moisture retention data. Soil Science 117(6):311-314. DOI: https://doi.
org/10.1097/00010694-197406000-00001.
TECHNICAL CONSULTANTS. (2018). Study of georeferencing, topographic
surveys and determination of the infiltration capacity in the event of possible
surface runoff in the area of the Silala springs. La Paz, Bolivia, DIREMAR.
688 p.
Cooper, D. J., Kaczynski, K., Slayback, D., & Yager, K. (2015). Growth and
organic carbon production in peatlands dominated by Distichia muscoides, Bolivia,
South America. Arctic, Antarctic, and Alpine Research, 47(3), 505-510.
Cooper, D. J., Sueltenfuss, J., Oyague, E., Yager, K., Slayback, D., Caballero,
E. M. C.,Argollo, J. & Mark, B. G. 2019. Drivers of Peatland Water Table Dynamics
in the Central Andes, Bolivia and Peru. Hydrological Processes. https://
doi.org/10.1002/hyp.13446
Darcy, H. 1856. The public fountains of the city of Dijon. Explanation and
application of the principles and formulas to be followed in water distribution
matters: work completed by an appendix relating to the water supplies of
several cities for water filtering and the manufacture of cast iron, lead, sheet
metal and bitumen pipes. s.l., Dalmont. 664 p. (Darcy, H. 1856. Les fontaines
publiques de la ville de Dijon. Exposition et application des principes à suivre
et des formules à employer dans les questions de distribution d’eau: ouvrage
terminé par un appendice relatif aux fournitures d’eau de plusieurs villes au
filtrage des eaux et à la fabrication des tuyaux de fonte, de plomb, de tole et de
bitume. s.l., Dalmont. 664 p.)
DGA. 2019. General Directorate of Water, Ministry of Public Works MOP -
Chile (online, website). Accessed 21 Mar. 2019. Available at http://snia.dga.cl/
BNAConsultas/reportes.
Dingman, SL. 2002. Physical hydrology. 2. ed. Upper Saddle River, NJ, Prentice
Hall. 646 p.
Domínguez, E; Fernández, HR. 2009. South American benthic macroinvertebrates:
systematics and biology. s.l., Fundación Miguel Lillo. University of
Tucumán.
Drexler, JZ; Snyder, RL; Spano, D; Paw U, KT. 2004. A review of models
and micrometeorological methods used to estimate wetland evapotranspiration.
Hydrological Processes 18(11):2071-2101. DOI: https://doi.org/10.1002/
hyp.1462.
56
151
Dyson, M; Bergkamp, G; Scanlon, J; Initiative, IW and N. 2003. Flow: the
essentials of environmental flows. Edición: 2. Gland (Suisse); Cambridge
(Royaume-Uni), World Conservation Union. 118 p.
Dyson, M., Bergkamp, G., & Scanlon, J. 2003. Flow: the essentials of
environmental flows. Edición: 2. Gland (Suisse); Cambridge (Royaume-Uni),
World Conservation Union. 118 p.
Eichinger, W; Cooper, D; Kao, J; Chen, LC; Hipps, L; Prueger, J. 2000.
Estimation of spatially distributed latent heat flux over complex terrain from a
Raman lidar. Agricultural and Forest Meteorology 105(1-3):145–159.
Epler, J. 2001. Identification manual for the larval Chironomidae (Diptera) of
North and South Carolina.
Federer, CA; Vörösmarty, C; Fekete, B. 1996. Intercomparison of Methods
for Calculating Potential Evaporation in Regional and Global Water Balance
Models. Water Resources Research 32(7):2315-2321. DOI: https://doi.
org/10.1029/96WR00801.
FUNDECO. (2018a). Environmental impact assessment study in the Silala. La
Paz, Bolivia, DIREMAR. 128 p.
Garcia, MC; Meneses, RI; Naoki, K; Anthelme, F. 2014. Methods for evaluating
the effect of grazing on bofedal plant communities. Ecology in Bolivia 49(3):91-
103.
Hribljan, J. A., Cooper, D. J., Sueltenfuss, J., Wolf, E. C., Heckman, K. A.,
Lilleskov, E., & Chimner, R. A. (2015). Carbon storage and long-term rate of
accumulation in high-altitude Andean peatlands of Bolivia. Mires and Peat.
15:12-14
IHH. (2018). Characterization and efficiency of the hydraulic works built and
installed in the Silala sector. La Paz, Bolivia, DIREMAR. 352 p.
Lamouroux, N; Capra, H. 2002. Simple predictions of instream habitat model
outputs for target fish populations. Freshwater Biology 47:1543-1556.
Lindroth, A. 1985. Canopy conductance of coniferous forests related to climate.
Water Resources Research 21(3):297–304.
Lorini, H. (2014). Influence of glacial melting and floristic composition on
subsurface runoff patterns in bofedals and grasslands of two micro-basins. s.l.,
Sustainable Water. 27 p.
57
152
Limpens, J., Berendse, F., Blodau, C., Canadell, J. G., Freeman, C., Holden, J.
Roulet, N., Rydin H. & Schaepman-Strub, G. (2008). Peatlands and the carbon
cycle: from local processes to global implications-a synthesis. Biogeosciences,
5, 1475-1491.
Medrano, H; Bota, J; Cifre, J; Flexas, J; Ribas-Carbó, M; Gulías, J. 2007.
Efficiency in the use of water by plants. Geographic Investigations (Esp)
(43):63-84.
Meneses, R. I. 1997. Study of the vegetation in the mining area of Milluni
(Province of Murillo, Department of La Paz). Degree Thesis. Higher University
of San Andrés, La Paz. 107 p.
Meneses, RI; Palabral, A; Molina, J; Acho, J; Plata, O; Loza, S. (2014). Floristic
and limnological characterization of bofedals in the municipalities of Turco,
Pelechuco, Curva, Charazani and Curahuara de Carangas for the second phase
of the National Bioculture Program (PNB). La Paz, Bolivia, Sustainable Water.
167 p. Technical document not published.
Ministry of Foreign Affairs. 2014. First diagnosis of the mountainous area of
the Plurinational State of Bolivia. La Paz, Bolivia.
Ministry of the Environment, 2012, Regulation for the determination of the
minimum ecological flow, Decree No. 014, Republic of Chile - Santiago,
Articles 1 - 12.
Ministry of Environment and Water, 2017. Strategy for the Integrated
Management of Wetlands and RAMSAR sites in Bolivia, La Paz - Bolivia.
Monteith, JL. 1965. Evaporation and environment. Symposia of the Society for
Experimental Biology 19:205-234.
Montes de Oca, I. 1997. Geography and natural resources of Bolivia. s.l., s.e.
Muñoz, JF; Suárez, F. (2019). Study of the Quebrada Negra Wetland. Chile,
s.e. 95 p. (Annex XIII).
Orsag, V; Miranda, R; Torrez, E. (2017). Characterization of soils in the Silala
Bofedals and surrounding areas. La Paz, Bolivia, DIREMAR. 241 p.
Oyague, E; Maldonado-Fonkén, MS. 2015. Relationships between aquatic
invertebrates, water quiality and vegetation in an Andean peatland system.
Mires and Peat, 15:1-21.
Paw, UKT; Brunet, Y. 1991. A surface renewal measure of sensible heat flux
density. In Proc. of the 20th Conference on Agriculture and Forest Meteorology,
Salt Lake City. s.l., s.e. p. 52–53.
58
153
Poff, NL; Allan, JD; Bain, MB; Karr, JR; Prestegaard, KL; Richter, BD; Sparks,
RE; Stromberg, JC. 1997. The Natural Flow Regime. BioScience 47(11):769-
784. DOI: https://doi.org/10.2307/1313099.
Puoilly, M., Ibñez, C. Zepita, C. Gutierrez, R. Aguilera, G. & Aguirre, M. 2014.
C5- Environmental flow: Perspectives of evaluation in the TDPS system. In
M. Pouilly; X. Lazzaro; D. Point; M. Aguirre (2014). Knowledge base line on
hydrological resources in the TDPS system with a focus on the Titicaca Lake
basin. IRD - IUCN, Quito, Ecuador. 320 pp.
Priestley, CHB; Taylor, RJ. 1972. On the assessment of surface heat flux and
evaporation using large-scale parameters. Monthly weather review 100(2):81–
92.
Ramsar. 2002. “Wetlands: water, life and culture. 8th Meeting of the Contracting
Parties to the Convention on Wetlands (Ramsar, Iran, 1971). s.l., s.e.
RAMSAR N. 0489. 1990. List of RAMSAR - Los Lipez. Plurinational State
of Bolivia.
Ramsar, C. 2005. 9th meeting of the Contracting Parties to the Ramsar
Convention on Wetlands. Retrieved from http://www. CONDESAN. org/
docinfonotas/ramsar, 202.
Rivas-Martínez, S; Navarro, G; Penas, A; Costa, M. 2011. Biogeographic
Map of South America. A preliminary survey. (Serie International Journal of
Geobotanical Research) 1:21-40. DOI: https://doi.org/10.5616/ijgr110002.
Ruthsatz, B. 2012. Vegetation and ecology of the high Andean bofedals of
Bolivia. Phytocoenology 42:3-4.
Salvador, F; Monerris, J; Rochefort, L. 2014. Peatlands of the Peruvian Puna
ecoregion: types, characteristics and disturbance (Peatlands of the Peruvian
ecoregion of Puna: types, characteristics and disturbances). (Serie 3) 15:1-17.
Squeo, FA; Warner, BG; Aravena, R; Espinoza, D. 2006. Bofedals: high altitude
peatlands of the central Andes (en línea). Revista chilena de historia natural
79(2). DOI: https://doi.org/10.4067/S0716-078X2006000200010.
59
154
Stewart, JB. 1989. On the use of the Penman-Monteith equation for determining
areal evapotranspiration. Washington, s.e., (IAHS publication). p. 3-10.
Stohlgren, TJ. 2006. Measuring Plant Diversity (on line). s.l., Oxford University
Press. DOI: https://doi.org/10.1093/acprof:oso/9780195172331.001.0001.
Tobin, AK; Sumar, N; Patel, M; Moore, AL; Stewart, GR. 1988. Development
of photorespiration during chloroplast biogenesis in wheat leaves. Journal of
experimental botany 39(7):833–843.
UERH-COFADENA. (2017). Geophysical study of the surrounding area of the
Silala bofedals using resistive electrical tomography (RET). La Paz, Bolivia,
DIREMAR. 176 p.
Wetzel, RG. 2001. Limnology: lake and river ecosystems. s.l., gulf professional
publishing.
Zea Mamani, R. 2015. Experimental determination of the water needs of the
bofedal in dry and wet puna in the department of Puno.
60
Annex 27
Note S/N of The Antofagasta (Chili) and Bolivia Railway P.L.C
addressed to the Company DUCTEC S.R.L., Antofagasta, 23
August 2000
(Original in Spanish, English Translation)
156
ANTOFAGASTA, 2 3 AGO. 20G
ANTOFAGASTA
(CHILI) ANO BOLIVIA
RAILWAY P.L.C.
{Ferrocarril
de Antofagasta a Bolivia)
Senores
DUCTEC S.R.L.,
Fono: 206700 .88~~ff1~ 2Jf
Fax(OSS) 206220
Antot.agasta • CHILE
Guerrilleros Lanza No. 1437, M·
La Paz - Bolivia
De nuestra consideraci6n:
Persisten ustedes en enviarnos factu ras emitidas en
contra nuestra por una supuesta captaci6n y entrega de agua en una
cuenca ubicada en Bolivia.
Ta! como le senalamos en nuestra anterior carta, les
reiteramos que rechazamos cualquiera pretension de obtener ta! pago.
Vuestra aspiraci6n es totalmente improcedente y carece de todo
fundamento, tanto de hecho como de derecho.
En atenci6n a que vuest.ra intenci6n de cob.ro la
presentan avalada por la ca.rta de un organismo denominado
Superintendencia de Saneamiento Basico, nos hemos contactado con el
Ministerio de Relaciones Exterio.res de Chile, quien junto con recibir los
antecedentes, nos ha reiterado la instrucci6n de abstenernos de discutir
este tema internacional.
Esperando no tener que reiterar los terminos de la
ca.rta de fecha 8 de mayo ultimo y los de la presente, nos despedimos
a tentamente,
_______ ...r /) -------:r::----
C w--~~3 Q'ome~z.-Sa-.c! - - - · ··
Sec;.ci6n Control rs:, eedores
Ferrocarril de Ant fagasta a Bolivia
157
Letterhead on this page: FCAB’s Logo, address and contact information]
ANTOFAGASTA, [Affixed with a date-stamp reading 23 August 2000]
To
DUCTEC S.R.L.,
# 1437, M Guerrilleros Lanza Street
La Paz-Bolivia
Of our consideration,
You persist in sending invoices issued against us for an alleged abstraction and
delivery of water from a basin found in a Bolivia.
As we had explained in our preceding letter, we hereby reiterate that we
reject any intention to receive such payment. Your aspiration is completely
inadmissible and unfounded, in the facts and law.
Given that your intention to collect charges is underpinned by the letter
addressed by an organ named the Superintendence of Basic Sanitation, we have
gotten in touch with the Ministry of Foreign Affairs of Chile which, along with
receiving the background [data], has reiterated the instruction that we are to
abstain from discussing this international matter.
Hoping we will not have to repeat the terms of the letter addressed to you last 8
May, or those of the present letter, we respectfully bid you farewell,
[Signed by]
Luis Gomez Saldana Supplier
Control Area
Antofagasta-Bolivia Railway
159
Annex 28
1906 Chilean Concession to
THE ANTOFAGASTA-CHILI AND BOLIVIA RAILWAY P.L.C.
Obtained from the data base of Chile’s Direction-General of Water, 2019
http://www.dga.cl/Paginas/default.aspx
(Original in Spanish, English Translation)
160
IWIW.dgul/i,oox:t~_rns!Ol"(OS/l'qw/del
I l,ctc1:1hX,I Derechosde~ietlrode~registradosen«.4
tt;,l'.Klli'l ~r..n,111 So,.:,_,,,,.o,:,;r:-,'lmsir~1,1e1-oV"t011io-
~1""'8 .a«,J~CH_~e.>.<,.1,1. ..... l .l. ~.,Grmn-,o.l,o,,o._~~,,t <_"_"'° .~_·:-c9,1-o..,.,....,.n.,,owo,i.
~IUVlelilt;ir~:~y '°'"""'""~.,......,,,~,-,,, o:.oectr..<.il!~dllCo:;ooo~cH
lt:rolnCr.r:,:i~
'-'""""""""r.=0-,,,..,., .. ,,,,..._ .... ..,,,, .... ,.,..,. .>S«<Oifl"
~-,m:,'fM-:15 ,., ,.,,,,~,.,,.,11.,.,.,..c.,
11<41lolou~-.-""°"",.....,....o:<doo:,oo,.,,,,boo,_.,,,.w.,
.. .,,.,,....,s,-'""°""""ct1Co:tn•:dcooo~""""'"'!:a1100.0.1u"°""""' o_ .. ,...,__, 'F".,.,, . .-""""°'"°""'""lb-',m
~ . r<r-lO::i~ool«ii"..t-'I OOICl..-ool(tOlt!to-.....,,..,o;.n ,.,m,:« .. ~;JGA.
Utilitarios
.'n--.:,rtl
~11iem1,n ,,,_, --__
.k..r..o,..b .at~ __ -~e1,
"°-"'~ dl-Y.IFul'<Gllln
l>llm'ol-P':Jt
l.\l.Office
"~"~"".-' -,.•-
di·~-"'·"-"' .. ... _ ...... ==- 9,ll"ln. .... ,,_
rJ.Office
I
i
"J'
j
• ---=i.=.• ~========1-
161
Date of
Resolution/
Sent to .Judge/
Annual
N' File Code
Application
Region Province Community
Applicant \Vater Source Water
Basin Sub-Basin Sub-Sub-Basin Source
Exer cise of
Average Flow Unit
N' Name Natu r e Classifica tion Use Right
Registration in
Flow
C.B.R.•
ANTOFAGAS
TA CHILI Superficia Loa Alto River
Siloli
Pem1anen
122584 UA-0202-809028 I Antofagasta The Loa Calama AND BOLIVIA 11/6/1901 and River/Estuary Loa River ( under the Salado Salado River
River
and 237 u,
RAILWAY Flowing River) continuom
PLC
162
Caudal 1Caudal Acoiones Acciones
UTM UTM UIM UlM Referenciaa
C011igo
1Pos~
N'
Ecologico Ecologico enel enla
Norte Este
Huso Darum
Latirud Longitud
Datum
Norte E~e Latitud Longlud
Referencia a puntos conocidos de captacion
puntos
Distancia
Unidad Desnivel
C.B.R. Fojas N'CBR Aio Expediente
misdeun
Certfficadol
11/s) Promedio? Cauce Fuente
Captacion Captacion Captacion Captacion Restlucion Restitucion Re~itucion Restitucioo conocidosde Distancia 1ml Antiguo
puntode
Aio
Im) Im) Im) Im) rastitucioo captacion?
F.CAB. ES DU ENA DE UN DE RE CHO DE 20.500
M3/D~ EQUIVALENTE A 237 US. QUE SE CAPT AN OE
DOS REPRESAS.
REP RE SA I UBI CADA EN EL GAUGE NA !URAL OEL
RIO SILOLI. EN TERRITORIO DER. BOLIVIA. A 575
MTS. /J. ORIENTE DEL LIMITE INTERNACIONAL CON
lA R. DEC HILE. COOR. UTM 7566250 NY 600925 E
REP RE SA 2 EN El CAUCE SILOLI, EN TERRITORIO
DE CHILE, A 36 MTS. Al OE STE DEL LNIMITE
INTERNACION/J. CHILE-BOLIVIA, COO. UTM 7565750
NY600925E.
lA FUENTE OEL RIO SILOLI ESTA SITUADA EN UNA
ZONA DENOMINADA V ERTIENTE DEL CAJON Y
PARTE DElA S VERTIENTES ORIENTALES DEL
DEPTO. DE POTOSI, PROV. DE SAN ANTONIO LOPEZ,
VICECANTON QUETENE, BOLIVIA, A 3,5 KMS. /J. C.BR. PAG.26
0 0 7566250,000 600925,000 19 1956 0 0 ORIENTE DE lA FRONTERA ENTRE CHILE Y BOLIV~ 0 OCalama 2VTA. 2 1990 LIBROII 2778QOII
163
UTM UTM
North East
Collection Works Collection Works
F.C.A.B. owns a right of 20,500 m3/day equivalent to 237 l/s that are collected from two
dams.
Dam 1: Located in the natural course of the Siloli River, in the territory of the Republic of
Bolivia, 575 meters east of the international boundary with the Republic of Chile, UTM
Dam 2: In the Siloli channel, in the territory of Chile, 36 meters west of the Chile-Bolivia
international boundary, UTM Coordinates: 7565750 N and 600925 E.
The source of the Siloli River is located in an area called the Cajon spring and part of the
eastern springs of the Department of Potosi, Province of San Antonio Lopez, Quetene Vice-
Canton, Bolivia, 35.5 kilometers east of the border between Chile and Bolivia.
Certificate N°
/ Year
7566250 600925 Calama
C.B.R. 2 VTA. 2 1990 Page 26,
Book II 2778/2011
Reference to known points of collection C.B.R. Pages C.B.R. N° Year Old File
Code
164
derechosdeagua.dga.cl/index.php
Antofaga,t, El lo, Cal,m,
Dates derecho concedido
Superficial y Corriente
Rio Loa Alo (b,~ junta Rio Salido) Rio Salado
Rio Siloli
165
IFile Code/ Appllcatton Nnmbe1·
I UA-0202-809028/ 1 r lAr--p_p_lic-an-t -Na_m_e ________, IA- p-pl-ic-an-t t-yp-e1
181148200-5 IANTOFAGASTA CHI LI AND BOLIVIA RAILWAY P.LC. ILegal F IProvince ICommnnity
IAntofagasta IThe Loa ICalama
Data of the concession right
r ater nature !Source classification IRighttype
1superficial and Flowing 1River/estu3l)' !Consumptive
IResolotion Unit/Reconl/C.B.R. 1»ate o!Resolution/Sent to Judge/Registration in C.B.R. IResolotion N'/Reconl/CBR N'
!Supreme Decree 111/0611906 1194
1»ate on which the Comptroller is informed ITra nsliory Article r ater Use
I
INULL r11
IBasin ISob-Basin ISub-Sub-Basin
ILoaRiver ILoa Alto River (under the Salado River) !Salado River
!Source ISiloli River
flowaata
Janua1y Febm1y March April 1~1ay IJune IJutr !August !September October November ecember
1211 1211 1211 1211 1211 1211 1211 1211 1211 211 211 211
166
I derechosdeagua.dga.cl/index.php
Ud lU~ ue Ci:tllUi:tl
lb lfflrm l'1mi) ., ~ • • "H I '' I' I I' ' ' I' I 1· '
m 237 m 237 237 m m m m m m m
·-- - - ·- - - I ll"'Jlt ' ' ' 'I • 11111:::1 1r1 IIJ11 :1t:l1 I , tt:11 • :11::1111111111:.I Ir. rnr.1wi - - - -
Ptrm,nente y Cantin uo m Lt,,
- - - - - - - - 'I. 11 I lli.'\I . I. ' 11:1 l lfCI I 'I HI' I ' I• - - -
0 0
lllll •uo1 • , ra1-Lt 1111 r.tJr.·il-l - CD l!?!m
7S661IO 600925 19 1956
11r. t111 l ■ -F. t r.t l ' I ■ ■■♦ 1 1 : 1 1 111 , l ■ -F. t 1 r;1 ■ 1 t 1 1 ■ I. - - -
--- --- -- -- - - -- 11111 I · 1111.-.:111,Pll 1111 • ' II ■ r:11111 :111:1,1 1111 I IIIIP :1 I 1111 -
0 0
- F.cAB. ES DUOO. OE UN DERECHO DE 10.100 Ml DIA EQUIVALENTE A 237 L, S. QUE SE CAPTAN DE OOS REPRESA'i. REPRESA I UBICADA
. ' lllfllft1.'I - EN EL CAUCE NATURAL DEL RIO SILOU, EN mm IOO DER. BOUVIA, A 575 MTS. Al ORIENTE DEL Lil,IITE INTERNAWNAL CON LA R. DE
11111 • 11 f,'l'H=l'I: llt: '.I fll - CHILE
-. ' lllllllfL, - - -- 1,11111>'1.IL:.it:. 1;11
llll!a - mII:Ir. II1II1.,r. ~
0 0
g ,.
c.B.R. Cal,ma 2\/TA. 1
lb - - .n.- - ..,_
"" - 11"1 I I 1I:n1 ~t ' "'" JI~ • .. I' 11m11r111 , 1 r. r11r,1Itr ,ti ':li'il - - - - - -
mo PAG. 16 LI BRO 1110-0-0) 2778/2011
167
r[1~,IRiill r ~IAvm~llm ruli r ~IAWl!~lliw!
~mnruirul ~~ ~@lrni~ui m Ui
~ilj1Ullnv~) ~vm~«l(!lilfi!I! ~dimklluuul ~dim1lltrn1111
~ ~
Uf MNm 'mrn ~1.ctfo~ Wijl'~ ~~ UTM ~~iir n~1.ctfo~ Wm'~~~ Hn!ij D~fa
111~~1 rl 111 ll!ll
1~iuililliml1 1~iidinll1~!1 ~ra
F~11tmlilllro1(ll) ~ThM 1!ililiili(ll) ~llliluililllilll1 ~llliiiliL1'1!1
I I I I
R1Jmm fo ~ijW~rn ~fffiij~ f.CAB. IT,~ Tirn O\\I RO f Afil GHT Ofl~ll~~ Ml IDm~UIVAWITTO 1Jn I~. THAT AIIB mmmDfR OM TWO DAM~.D AM l:W ~AIBD INTH E NATill\i\1 mlJ1l~t Of Tilt ~11011 RMR INT ilt
~iji~I! IBAAITORY Of R. ~01MA m MT~. TO TilHA~T Of Tilt INIBRNATIONA1 ~OUNDARY Willi THE R. Of ~IB1t
R1Jmm fo ~ijW~ r~nlnliij~
~iji~I!
rum ru,reu,i ~11(11)
I I I
Im ~~I ~w
~Imm IVIA I
I
~m fi!lk~il ~11!i1n11111lliu iulrdin11Rfl ~ llKll1~/Y1u
1~1 rAGU BOOKll(III)
I
rn~IDll
Volume 5 - Annexes 24-28