Analysis of shallow lakes dynamics as related to precipitation in the Pampas region based on Landsat imagery

Abstract

Shallow lakes in the Pampas are characterized by their high number as well as by frequent variations in their hydrometric level and their surface area, depending on regional rainfall and underground water balance [1],[2]. More than 150,000 shallow lakes of various sizes have been counted in the province of Buenos Aires, with at least 10,500 of them having a surface area above 10 ha [2]. They exhibit changing dynamics (seasonally and annually) without a definite pattern. Preliminary work indicates an inverse relationship between the volume of water retained in a shallow lake and trophic state indicators, linking rainfall and trophic level [3]. Thus, Pampean shallow lakes are not only conditioned by the activities taking place in their surroundings, but they also respond to climatic forces whose dynamic effect has not yet been sufficiently studied. Understanding the dynamics of interactions between climate and shallow lake ecosystems in the Pampas is essential for their conservation and resource utilization, as well as for guaranteeing the provision of ecological services. Hence the interest in linking the variation in water volume in one or more shallow lakes to precipitations over the region since the said variation is linked to nutrient concentrations and plankton abundance conditioning the ecosystem’ s food web. The purpose of this study is to analyze the temporal dynamics of the area of five shallow lakes in different basins of the Buenos Aires province over the period 1984-2015 through the application of satellite image processing tools [4]. Two images per year were selected (Landsat 4 TM, Landsat 5 TM, Landsat 7 ETM+ and Landsat 8 OLI), one corresponding to summer and the other to winter. The evolution of the free water surface of these shallow lakes could be explained using a generalized linear model (GLM) and taking as variables the accumulated precipitation and the number of rainy days over a fixed-length period prior to the capture date of the satellite image. Accumulated precipitation and frequency of rainy days were calculated for three, seven, thirty, ninety, and one hundred and eighty days, as well as for one, two, three, and four years prior to the capture date, respectively. When analyzing the interdecadal precipitation medians, a decrease was observed in all meteorological stations involved (Azul, Benito Juárez, Olavarría and Tres Arroyos). This led to splitting the series of computed lake areas into two consecutive periods: 1984-2000 and 2001-2015. The areas reconstruction for the five shallow lakes varied in regard to both the explanatory variables and in the reconstruction coefficients when the series were divided into two periods. In all cases except one, better reconstructions were obtained for the second half. The results regarding the behavior of the shallow lakes in response to antecedent rainfall differed in each of the basins, reflecting the systems characteristics [5]. It may be interpreted that when the area responds to precipitation accumulated over longer periods the underground contribution is important, meaning that the memory of the system prevails, while a significant superficial contribution would explain the shallow lake’ s response to precipitation accumulated over short periods, pointing to the importance of runoff.