A modeling approach to explore the influence of different crop rotations on water-table depths and crop yields in the Pampas

Abstract

In humid plains with shallow water-tables, vegetation and groundwater influence each other through different mechanisms. Therefore, agricultural management decisions, such as the selection of the crop rotation, can affect water-table levels and these, in turn, can affect the crop performance. In this work, we evaluated the effect of different crop rotations on water-table levels and crop yields in the Argentine Pampas, a region that shows rising water-tables in the last decades that challenge agricultural activities. For this purpose, we modeled the dynamic of water-table depth in different topographic positions with Hydrus-1D and we used empirical models to estimate the grain yield. Crop rotations evaluated were wheat/soybean (W/S), wheat/soybean – maize (W/S – M), cover crop/maize – soybean (CC/M – S) and soybean – maize (S – M), and the simulation period extended throughout 41 growing seasons. Simulated water-table levels and crop yields showed a satisfactory matching with field observations, providing enough confidence for modeling exercises. According to modeling results, the crop rotation clearly affected the water-table depth, with the most intense rotation (i.e. W/S) showing the deepest water-table (e.g. W/S = 2.40 m vs. S–M = 1.90 m in the backslope). In the shoulder, the W/S rotation deepened the water-table in such an extent that more than 76 % of the time it was located below 3.5 m. However, with the least intense crop rotation (S – M) the proportion of time that the water-table remained deeper than 3.5 m depth was reduced to only 15 %. In the footslope, the least intense rotation showed the water-table level riskily shallow (< 1 m depth), negatively affecting most crops, almost half the time (48 %), while with the W/S rotation this percentage was reduced to 14 %. Through their influence on water-table depth, crop rotations also affected grain yields as suggested by empirical models. In maize, the predicted yields were maximum in the backslope for all rotations (~11 Mg ha-1), while they fell more strongly in the footslope with the least intense rotation (−28 %) but in the shoulder with the most intense one (−31 %). In soybean, the topographic position for which the highest yields were predicted depended on the crop rotation, being the footslope for W/S and the backslope for the other rotations. Our work provides novel insights on the influence of agricultural management on water-table levels and crop yields, as well as new tools to help in decision-making aimed at a more efficient management of water resources.