Elsevier

Journal of Hydrology

Volume 393, Issues 3–4, 8 November 2010, Pages 233-244
Journal of Hydrology

Hydrological modelling of a closed lake (Laguna Mar Chiquita, Argentina) in the context of 20th century climatic changes

https://doi.org/10.1016/j.jhydrol.2010.08.019Get rights and content

Summary

A major hydroclimatic change occured in southeastern South America at the beginning of the 1970s. This change was recorded in Laguna Mar Chiquita (central Argentina), the terminal saline lake of a 127,000 km2 catchment as a dramatic rise in lake level larger than any observed over the past 230 years. Based on available continuous lake level monitoring since 1967, our study aimed to develop a lake water balance model for investigating the link between climate and lake level variations. Since un-gauged downstream surfaces represented approximately 80% of the catchment, the main challenge of the model development and implementation came from estimating the magnitude of catchment inputs from sparsely available gauge data. We determined a strongly negative water balance in the un-gauged part of the catchment that can be attributed to evapotranspiration in two large surface water hydrosystems. The chloride balance indicated that the lake is hydrologically closed, without significant groundwater outflows. Using contrasted hydroclimatic conditions, the robustness of the model calibration was evaluated with the model residual, and a short validation proposed for the 1998–2006 time period. Sensitivity analyses were performed in order to identify the main forcing factors of lake variations. We determined that the abrupt lake level rise in the early 1970s could be attributed to increased runoff in the upper northern sub-basin, suggesting a tropical climatic influence. Based on available hydroclimatic data, we propose a continuous lake level simulation for the 1926–2006 time period which could be used as a reference curve for better constraining paleohydrological reconstructions from sedimentary proxies.

Introduction

In southeastern South America (SESA), the hydrologic cycle is known to be characterized by substantial variability over the 20th century (Garcia and Mechoso, 2005, Pasquini et al., 2006). A wet period resulting from a well-documented significant increase in precipitation and streamflows in both the Paraná-Plata Basin and central Argentina occurred from the early 1970s until the beginning of the 21st century (Garcia and Vargas, 1998, Genta et al., 1998, Planchon and Rosier, 2005). This substantial increase in precipitation is associated to higher frequency and severity of extreme hydrologic events (Camilloni and Barros, 2003, Berbery and Barros, 2003). Since this region is highly dependent on agricultural and hydroelectricity, changing conditions have had important and immediate impacts on the local economy and society. Currently, a debate concerns the processes that govern precipitation variability in SESA which induces increasing hydroclimatic trends, as well as the understanding of their link to anthropogenic activities (CLARIS-LPB project, http://www.claris-eu.org/).

Laguna Mar Chiquita (30°54′S–62°51′W) is an extensive saline lake located in central Argentina, west of the Paraná-Plata Basin (Fig. 1a), and has clearly undergone 20th century hydrologic changes through abrupt lake level fluctuations. The lake is a unique site in South America because continuous lake level measurements, collected since 1967, record significant increases in water level since the early 1970s. Historical information, based on limnological studies which began at the end of the 19th century (Harperath, 1887, Von Grumbkow, 1890, Frank, 1915, Kanter, 1935, Bertoldi de Pomar, 1953, Reati et al., 1997), had allowed a semi-quantitative reconstruction of lake evolution prior to the instrumental period (Fig. 1b). Additionally, based on lake sediment proxy data, lake variations have been reconstructed for the last 230 years (Piovano et al., 2002, Piovano et al., 2004, Piovano et al., 2006, Piovano et al., 2009), and have shown that the lake was characterized by dominant dry conditions until the beginning of the last quarter of the 20th century, with the highstand from the early 1970s the most important in length and magnitude. During the dry period, the lake surface was reduced to less than 2000 km2, whereas in 2003 it reached an area of approximately 6000 km2, making it not only the largest saline lake in South America but also one of the largest saline lakes in the world.

The link between climate variability and the hydrologic changes that occurred in the early 1970s in Laguna Mar Chiquita has never been investigated. At the regional scale and from comparisons with data obtained from the La Plata Basin, it has been suggested that increased lake levels could be attributed to an increase in precipitation and river discharge (Piovano et al., 2002). Questions still remain as to whether lake level changes in Mar Chiquita are mainly the result of climatic forcing or whether they are due to an influence from anthropogenic activities.

In this study, a mass balance model is used in order to investigate the link between climate and lake level variations. Calibration of this model, based on lake level measurements, allowed us to investigate lake water balance and its variations within the last quarter of the 20th century. Sensitivity analyses were then developed to identify the main factors responsible for the forcing of lake level variations. Finally, based on available river discharge and precipitation data, the lake model was used to simulate monthly lake levels for the period prior to lake monitoring, and to provide an extended lake level curve that could be utilized to constrain sedimentary proxy data reconstructions.

Section snippets

The lake watershed

Laguna Mar Chiquita (30°54′S–62°51′W) is a terminal lake occupying a tectonic depression that formed during the middle Pleistocene (Kröhling and Iriondo, 1999). The lake is limited at its eastern shore by a fault (Bordo de Los Altos) that forms a small cliff that prevents surface outflow towards the Parána-Plata Basin. The lake catchment is estimated (based on the digital elevation model (DEM)) to comprise roughly 127,000 km2 from 26°S to 32°S and from 62°W to 66°W (Fig. 1a), with a relatively

Lake level measurements

Monthly measurements of water lake levels spanning the 1967–2006 time period were obtained from the Laboratorio de Hidráulica at the Universidad Nacional de Córdoba (the location of lake level measurements in Fig. 1a). Lake level records have been conducted continuously since 1967, except for a gap between November 1997 and January 2001 for which only three monthly measurements of water lake levels were available. Large hydrologic changes occurred during the 1967–2006 time period with a

Basic equation

Monthly changes in lake level were calculated using the water balance equation combined with quantitated area–volume-level relationships. The dynamic lake water balance equation can be expressed as follows:ΔVΔt=A(V)(P-E)+Qin-Goutwhere for the monthly time step Δt, ΔV is the lake volume variation (m3), A is the lake area (m2), as a function of lake volume V; P is the on-lake precipitation estimated from the six rainfall stations (m), E is the evaporation from the lake’s surface (m), Qin is the

Evidence and quantification of a water loss

First, the lake level simulation was performed by neglecting groundwater outflows Gout, and estimating Qin from the sum of measured river discharges (Qin = QR1 + QR2 + QR3a). The results showed a significant lake level overestimation, on average, ∼1.3 m above the observed lake level (Fig. 5; EF = 0.70). Then, in order to estimate this discrepancy in terms of water volume, we introduced a constant calibration parameter γ in the lake model as follows:ΔVΔt=A(V)(P-E)+Qin-γ

The adjustment was determined

Dependence of the model residue on climatic conditions

Despite the statistically significant lake level simulation performed with a constant γ value, discrepancies remained between the simulated and observed lake level curves (Fig. 5). These discrepancies could be related to the uncertainties in input variables (discharge, precipitation, and evaporation times series), measured lake levels, or model structures (the morphometric relation A(v), and the calibration parameter). We analyzed the model residue in order to detect possible trends or biases.

Lake water balance variability and sensitivity analyses

The model allowed us to quantify the lake water balance and the distribution of lake water inflows between local influences (i.e. direct precipitation), and remote influences (i.e. catchment contribution) mainly due to upstream river flows. We determined that the catchment contributed, on average, 33% to total inflows during the simulation period, and its proportion was highly variable, from a zero value (e.g. dry years of 1972 or 1989) to almost 71%.

Sensitivity analyses were performed in order

Extrapolated lake level simulation: 1926–2006

Using the available precipitation and river discharge data, we were able to begin the simulation in 1926, i.e. prior to the lake level monitoring. As in Section 6.2, we used the correlation established between R3a and R3b as an input in the lake model.

The correlation determined between water level and TDS (H = 82.3 * TDS−0.0446; r2 = 0.95) can be used to provide an indication of past lake levels. We tested different initial lake levels (±1 m) based on the value determined using the 1926 TDS

Conclusion

A monthly water balance model was built in order to investigate the link between climate and lake level variations. First, the lake model calibration showed a significant water loss, which was not induced by groundwater seepage, as indicated by the chloride balance. This water loss was attributed to evapotranspiration in the un-gauged part of the catchment. Second, no model dependence on hydroclimatic conditions was evidenced, supporting use of the model under contrasting climatic conditions

Acknowledgements

Our work was supported by a PhD Grant from the French Ministry (Magali Troin), and was benefited from funding from the French CNRS-INSU (PNEDC and LEFE-EVE programs, AMANCAY Project), and from the French National Research Agency (program VMC, Project ANR-06-VULN-010, ESCARSEL Project). The work also benefited from funding that supported the CLARIS-LBP Project (EC-FP7), ECOS-MINCYT (PA08U02, France-Argentina cooperation), and PIP 112-200801-00808 (CONICET). We also thank the Laboratorio de

References (38)

  • E.R. Berbery et al.

    The hydrologic cycle of the La Plata basin in South America

    J. Hydromet.

    (2003)
  • Bertoldi de Pomar, H., 1953. Contribucion al Conocimiento del Origen de la Laguna Mar Chiquita de la Provincia de...
  • Bucher, E.H., 2006. Banados del Rio Dulce y Laguna Mar Chiquita (Cordoba, Argentina), Academia National de Ciencias,...
  • Durigneux, J., 1978. Composición química de las aguas y barros de la laguna Mar Chiquita en la Provincia de Córdoba....
  • H. Frank

    Contribucion al conocimiento de las Salinas Grandes y la Mar Chiquita de la Provincia de Cordoba

    Revista del Centro de Estudiantes de Ingeniera

    (1915)
  • N.O. Garcia et al.

    The temporal climatic variability in the Rio de la Plata basin displayed by the river discharge

    Clim. Change

    (1998)
  • N.O. Garcia et al.

    Variability in the discharge of South American rivers and in climate

    Hydrol. Sci.

    (2005)
  • G.I. Gavier et al.

    Deforestacion de las Sierras chicas de Cordoba (Argentina) en el periodo 1970–1997

    Academia Nacional de Ciencas, Cordoba, Argentina, Miscelaneas

    (2004)
  • J.L. Genta et al.

    A recent increasing trend in the streamflow of rivers in southeastern South America

    J. Clim.

    (1998)
  • Cited by (54)

    • Hydrogeochemistry of a small saline lake: Assessing the groundwater inflow using environmental isotopic tracers (Laguna del Plata, Mar Chiquita system, Argentina)

      2019, Journal of South American Earth Sciences
      Citation Excerpt :

      Rodríguez et al. (2006) carried out a hydrological digital simulation mass balance in Laguna Mar Chiquita for the record period 1967–1997. In the same sense, Troin et al. (2010) performed a hydrological mass balance model to investigate the link between climate and lake-level variations in the Laguna Mar Chiquita. The chloride mass balance model performed by these authors also suggested that the groundwater outflow was not significant and water loss was attributed mainly to evaporation.

    View all citing articles on Scopus
    View full text