Modelagem do escoamento, da produção de sedimentos e da transferência de fósforo em bacia rural no sul do Brasil

Abstract

Areas under intensive tobacco crop cultivation have been impacting the water balance and increasing soil erosion in Southern Brazil. Application of large amounts of mineral fertilizers, higher than the tobacco requirement, contributes to excessive phosphorus loads in soil and water bodies. The assessment of potential environmental impacts due to changes in land use and management practices is necessary to achieve the sustainable management of natural resources such as soil and water. In this study the hydrological processes, the sediment yield and the phosphorus transfer in the Arroio Lino watershed were evaluated by using the Soil and Water Assessment Tool (SWAT) model. The small watershed (4.18 km2 ) is located in Southern Brazil and its main land use is tobacco crop under conventional tillage. Measured streamflow and sediment yield at the watershed outlet were used for model streamflow sensitivity analysis, calibration and validation. Phosphorus loads at four sub-watersheds (A1, A2, B, C) and at the watershed outlet were used for model sensitivity analysis and calibration. A Latin Hypercube (LH) and One-factor- At-a-Time (OAT) sensitivity analysis was performed on input variables. Model calibration was performed with the Shuffled Complex Evolution Algorithm-Uncertainty Analysis (SCE-UA). Time series plots and standard statistical measures were used to verify model predictions. Adaptations of the model parameters for the reality of the region in which the watershed is located were made during the calibration and validation of the model. The predicted monthly streamflow matched the observed values, with a Nash Sutcliffe coefficient (NSE) of 0.87 for calibration and 0.76 for validation. In an attempt to account for sediment transport and deposition processes across the landscape, the SWAT model was modified to simulate landscape sediment transport capacity. The new deposition routine performed better during calibration (NSE of 0.70) than SWAT standard version (NSE of -0.14) in the studied watershed. The modified model provided reasonable simulations of sediment transport across the landscape positions. The results suggest that the integration of the sediment deposition routine in SWAT increases model predictions accuracy in steeper areas, while at the same time significantly improves the ability to predict spatial distribution of sediment deposition areas. The predicted P loads are in the order of magnitude of the measured ones, however, the model failed to predict the P loads in three sub-watersheds (A1, A2 and B). Although occupying only 29% of the total land cover, cropland is the primary source of nutrients in the watershed (80%). After calibration of streamflow, sediment yield and phosphorus loads, the simulation of distinct management scenarios was done for the Arroio Lino watershed. Three scenarios of management practices: conventional tillage (CT), minimum tillage (MT) and no-tillage cultivation (NT) with reduction of 50% of fertilizer rate application were tested over a period of 30 years. No-tillage practices did not significantly affect water yield, but greatly affected sediment yield due to reduction of soil erosion. The soluble phosphorus losses decreased mainly when the fertilizer doses decreased. The simulation results suggest that conventional tillage practices should be replaced by less intensive tillage practices in order to minimize environmental impacts caused by a single land use.