HYDROLOGIC STUDY OF MAR CHIQUITA SYSTEM, R. ARGENTINA

Weber, J. and Menajovsky, S.

Hydraulics Laboratory, FCEFYN, National Univ. of Córdoba,

V. Sarsfield 1601, 5000, Córdoba, Argentina, Tel./Fax 54 351 4334148,

E-mail: rodrig@com.uncor.edu

Abstract: The hydrological system of Mar Chiquita and the Wetlands of Dulce River is a unique ecosystem in the Latin-American region. It is constituted by the biggest salt lake of South America - Mar Chiquita or Mar de Ansenusa - and Wetlands of Dulce River – with an extension of around 10,000 km2. This system is complex and its hydrologic knowledge is still insufficient. A wide quantity of hydro-meteorological information have been gathered, previously depurated and used to simulate Mar Chiquita evolution during a period of thirty years between 1967 and 1997. It has been used a model of hydrologic balance specifically developed for the study of the lake conduct before hypothetical flow extractions in the main tributary. This model has been used with six months and annual intervals, involving recent details such as taking into account the variability of the losses (evaporation y evapotranspiration) in function of the salinity, considering separately the lake and the wetlands, and the precipitation contributions by the levels of the lake. The analysis of the results shows the importance of the regimen of Dulce River as the main tributary and the effective evaporation. The impact in the lake of flows extractions upstream and the natural flow reduction of the Dulce River through the wetlands have been modelled.

 

Keywords: MAR CHIQUITA system, model of hydrologic balance

Mar Chiquita and its wetlands are located at the center of Argentina. It is a large and complex eco-hydrologic system mainly formed by a main salt lake, lagoons, wetlands of Dulce River in the Northen area, and by the semi-ephemeral Suquía and Xanaes rivers (see figure 1). It is the biggest salt lake in South America, with an area near 6.000 km², characterized by its little depth (approximately 10 m.) and large wetlands of 10,000 km² with a high biodiversity (e.g. Bucher, 1992, Reati et al, 1997 and UNC, 1998).

Fig. 1  Localization Area of the Mar Chiquita lake
and wetlands of Dulce River.

Numerous hydro-meteorological information have been compiled, which consist of series on precipitation, evaporation, temperature and wind in Córdoba and Santiago del Estero provinces. Series of levels and flows in the main tributaries have been obtained, which were also filled for the simulation period (30 years between 1967/97). The water level series of Mar Chiquita lake were obtained at Miramar City (south coast). The geometric functions of the lake (curves height-area-volume) were computed using previous bathymetric data up to 66 m over sea level and a set of 8 satellite images over such level. In this way, a set of consistent information has been completed, allowing a six-month step hydrologic balance within the last 30 years covering three types of climatic periods (dry, medium and wet). The results of the numeric simulation allow a better knowledge of this complex system supporting a solid basis for the water resources management in this region. Being of special utility to analyze the effects that can generate the increase of irrigation intakes from Dulce River upstream wetlands and lake, in order to prevent reduction effects such as the Aral sea case (Ellis, 1990 and Micklin, 1992).

2  HYDROLOGIC SIMULATION OF MAR CHIQUITA SYSTEM

The conceptual model of Mar Chiquita System updates previous studies (CIHRSA,1990) including the following hydrologic components: Mar Chiquita lake body, Dulce river wetlands area, three tributaries: Dulce, Suquía and Xanaes rivers from field measurements, Precipitation, Evaporation and flow extractions in the tributaries. The numerical model of the hydrologic balance solve the conservation of mass equation applied the system,

D V = S Q_{i D} T + A_L (P-E) D T – Q_ext D T                              (1)

Being D V volume variation of the lake, Q_i flow contribution of the tributaries, P media areal precipitation over the lake, E evaporation, A_L lake area, Q_ext flow extraction to the system and D T time step. The model was initially calibrated by adjusting the series of the tributary contributions, measured upstream the lake to the balance results assuming as data the series of precipitation, evaporation and lake levels. The extraction flows to the system were assumed to be void during the process of model calibration, and variables between 4 y 60 m³/s during the simulations. The time step was, according to the simulated case, 6 or 12 months. The results shown herein are with a 6 months interval. The calibration was performed for the 1967–1997 period, due to the availability of 30 years with all basic data for the calibration of model (mainly mean water levels).

As evaporation estimators, Lungeon, Meyer and Rohwer (Custodio, 1976) were compared with measurements, selecting the first one (ec. 2), to which a correction for the occurrence of variable salinity was added:

Em = 0.398 . D . (es-ed) . ( (273-T) / 273 ) . 760 / (Pa-es)                   (2)

Being: Em monthly evaporation, D number of days per month, es and ed: monthly steam saturation pressure average and monthly steam pressure average respectively (mmHg), T monthly average value of the maximum daily temperatures (oC), Pa mean atmospheric pressure (mmHg). The saturation steam pressure was calculated as es = ew-0.00066.Pa.(Ta-Tw). (1+0.00115 .Tw), with ew, steam pressure corresponding to the average temperature with humid bulb thermometer, Ta air temperature and Tw average temperature with humid bulb thermometer. The monthly average steam pressure was estimated as ed= HR .es/100, being HR the relative humidity (%). The results of vertical balance of mass in the lake and wetlands have been published by Pagot el al (2000). The volume variations have been obtained from the measured series of water levels and the corresponding geometric functions Area (level) and Volume (level), see figure 2.

The water levels were measured in the south coast of the lake in Miramar City. These series were previously corrected by datum changes. In this research monthly levels were adopted to avoid the influence of the wind effects (“wind set up” that can affect in more than 0.5 m). As it was mentioned before, the availability of measured levels from 1967 to 1997 has allowed the pre-calibration of the model in this period, when closing the equation for the volume balance because all its terms were known.

They were obtained from the bathymetric campaigns carried out by CIRSA (1979) completed with 6 satellite images in the periods corresponding to different water levels of the lake. The satellite images used correspond to the following years: 1972, 1976, 1981, 1986, 1993 and 1997, covering the periods of low, medium and high water levels. Images corresponding to the years: 1976 and 1981 are shown in figure 2. More details were published by Hillman et al (2000).

Fig. 2  Satellite images of Mar Chiquita lake
for medium -1976- and high levels -1981-.

The contributions corresponding to the precipitation have been obtained from the pluviometric series, generated in eight sub-series, which correspond to each side of the lake (N, S, E, W, SE, NE, NW and SW). These ones were formed in turn from series measured by DIPAS in 23 pluviometric stations (see localization in figure 4). The calculus of average areal precipitation considers the relative influence of each sub-series due to the lake area through an expression of the form,

P=a _N (h)P_N +a _NW (h)P_NW +a _NW (h)P_NW a _S (h)P_S + _E(h)P_E+a _O (h)P_W+a _SE (h)P_SE+a _SW (h)P_SW      (3)

With S a _i = 1, where the coefficients obtained by means of Thiessen polygons. In figure 4 the areal distribution of the annual average precipitation is shown, which corresponds in this example to the period 1970-1973.

Fig. 3  Geometry of Mar Chiquita lake (left)

 

Fig. 4  Isohyets for the lake and wetlands region, with the pluviometric stations (right).

As mentioned before, the evaporation was obtained from direct measures taken in Miramar tank (with a correction factor = 0.7), and the calculus by Lungeon expression using meteorological data measured in the periods without direct evaporation measures. In the intervals without meteorological measures equivalent series were used, generated from the monthly average values measured in Miramar. The correction of the evaporation due to salinity variations was determined by,

EVPc=EVP (2 - r )                                        (4)

where r is the lake water density, which is calculated by the following empiric expression obtained from the measured data:

r = 0.0007 S + 1                                        (5)

where S is the total content of salt dissolved in the water of the lake expressed in gr/l. This content of total dissolved salt S was obtained from a experimental relation that combines direct measures of S and the volumes V corresponding to the sampling dates. These functions were included in the model to automatically carry out the direct and inverse transformations.

The monthly flows near Mar Chiquita lake for Xanaes and Suquía rivers have been obtained from series of water levels and some discharge measurements. The rating curves were determined near the lake, extrapolating experimental values with the Manning formula, assuming steady and uniform flows. Dulce river discharges into the lake are one of the main results of the model, due to they can not be measured near the Mar Chiquita because of the wetlands existence and the very difficult access conditions. Then the closer data have been measured in Los Quiroga dam, near 300 km upstream the lake, where the main present and future extractions for irrigation are located.

The numerical model is a simple dynamic one, developed in FORTRAN to solve equation 1, that applies a time discretization in finite differences and solves the non linear terms by means of an interactive standard procedure until reaching the convergence (e.g. the evaporation due to salinity, which is dependent of the volume that is unknown in each temporal step). The empiric functions of direct geometry A(H), V(H) and the inverses H(A) and H(V), and those of salinity S(V) and V(S) were conveniently adjusted as variable degree polynomials by sections, being evaluated in high precision in order to reduce the numeric error. The simulations were carried out with the model (previously pre-calibrated for the period between 1967 and 1997), assuming constant flow extraction in each one. The simulations included several possible extraction flows, which were adopted as 4, 10, 16, 20, 40 and 60 m³/s. The initial conditions and the input variables for the simulated cases correspond to the series of reliable hydrological data: measured or generated with previously selected predictors (calibrated with measured data when it was possible). The results of the model are the volume evolution V(t), including those measured as well as the ones resulting from the different hypothetical extractions. Also the corresponding series of levels H(t), areas A(t) and salinities S(t) were obtained. These results are presented in graphic form in Figure 5 (continue dark line represents the evolution measured corresponding to a null flow extraction).

Fig. 5  (a) Volumes Evolution and (b) salinities due to different extractions.

Fig. 6  Dulce river discharge: Measured in Los Quiroga Dam (300 km upstream) and modeled input in Mar Chiquita lake.

3  CONCLUSIONS

The results show that the hydrological system of Mar Chiquita, in spite of its high complexity, can be numerically modeled in a way that quantitative results allow the taking of decisions based on rational and solid elements. Planning politics about the use of these resources should be careful and be based on studies that consider all the aspects of the involved watershed (see UNC, 1998). It is evidenced that still with small flow extractions (smaller than 5 m³/s) the hydrological system is potentially fragile during dry periods as those happened between 1967 and 1973. This can be seen in the reduction of lake volumes and the increase of salinity levels. However, for these small extractions, the lake–wetlands system demonstrates the capacity to quickly increase its dimensions if the hydrological conditions change, when increasing the precipitations and surface contributions, like it happened from 1976 when the lake overcame the average levels (66 m o.s.l.). For extractions of major flows -e.g. 10 m³/s - the critical interval of the lake increases in almost one year. This critical period occurs when salinity level is greater than 50 gr/l -considered the first threshold of salinity for fish reproduction-.

It is also observed that with these extraction flows, at the beginning of the 90s and during more than 3 years the levels of salinity overcome the 50 gr/l. For extractions of 20 m³/s the recovery of the lake would take 3 more years, up to 1981. Also, from 1988 salinity values would severely increase, overcoming a second threshold of salinity of 100 gr/l (a limit for fish survival). At the end of the period the level of salinity would stay above of the critical value without never ending up recovering to smaller values or similar to 50 gr/l. For major extractions (e.g. 40 or 60 m³/s) the effects would be extremely severe. The lake would dry off in more than half of the period simulated of 30 years, and salinity, would always stay above the first threshold and in approximately 80 % of the time above the second threshold, (see Figures 5a-b).

The modelling of Dulce river discharge showed a flow reduction near 40 % from Los Quiroga dam to the lake, 300 Km downstream, due to the effects of the wetlands which act as a natural reservoir increasing evapo-transpiration (see Figure 6).

The simulation made should be improved with the availability of new data, reason why it is recommended to invest major efforts in the measurement of the several parameters involved in the hydrological phenomena of the largest salt lake of South America.

Bucher, E.H., 1992. “Population and conservation status of flamingoes in Mar Chiquita, Córdoba, Argentina”. Coloquial Waterbirds 15(2): 179-184.

CIHRSA, 1979. “Estudio Batimétrico de la Laguna de Mar Chiquita”, Centro de Investigaciones Hídricas de la Región Semiárida, INCYTH CONICET, Va. Carlos Paz.

CIHRSA, 1990. “Estudio del régimen Hidrológico del río Dulce y Laguna de Mar Chiquita”, 73 pgs, Centro de Inv. Hídricas de la Región Semiárida, INCYTH CONICET.

Custodio,E. and Llamas,M. 1976, Hidrología Subterránea, Tomo I, Edic. Omega, Bcna.

Ellis, W.S. 1990. A Soviet Sea Lies Dying. National Geographic 177, February, 72-93.

Hillman,G.; Rodriguez,A.; Pagot., Menajovsky,S.; Bernasconi, I. and Caama?o,G. 2000. Funciones geométricas y de salinidad de la Laguna de Mar Chiquita, Córdoba, R. Argentina, XIX Congreso Latinoamericano de Hidráulica, IAHR, 2, 477-486, Arg., ISBN 950-33-0269-2.

Micklin,P. 1992. The Aral Crisis: Introduction to the Special Issue. Post-Soviet Geography 33, May, 269-282.

pagot., Caama?o,G.; Rodriguez,A.; Hillman, G., Bernasconi, I. and Menajovsky,S. 2000. Flujo hidrometeorológico en el sistema Ba?ados del río Dulce-laguna de Mar Chiquita, R.. Argentina, XIX Congreso Latinoamericano de Hidráulica, IAHR, 467-466, 2, ISBN 950-33-0269-2, Arg.

Reati, G.J., Florin, M., Fernadez, G.J. and Montes, C., 1997. “The Laguna de Mar Chiquita (Córdoba, Argentina): a little know, seculary fluctuating, saline lake”. International Journal of Lake Research 5: pp 187-219.

UNC (1998). “Valoración del Impacto Ambiental en la Región de Mar Chiquita y la Cuenca Afectada por el Canal Federal”, Informe Final al CFI, 650 pgs., Cba., Argentina.