farshad kiani; Behroz Behtari nejad; Ali Najafi nejad; Abdolreza Kaboli4
Abstract
Introduction: population growth, urbanization and land use changes cause negative effects in natural ecosystems and water resources. Soil erosion is one of the most important problems in agriculture and natural resources of Golestan province. Using low cost and accurate methods for planning and proper ...
Read More
Introduction: population growth, urbanization and land use changes cause negative effects in natural ecosystems and water resources. Soil erosion is one of the most important problems in agriculture and natural resources of Golestan province. Using low cost and accurate methods for planning and proper management of land and water resources are essential for estimating consequences of soil erosion and providing appropriate solutions to reduce soil losses.
Materials and Methods: The study area is located in eastern part of Golestan province with an area of 1524 square Kilometers. The average annual precipitation of the region is 496 millimeters. In this watershed, rainfall decreases from south and south west to north and north east (due to the remoteness from the Caspian Sea), while evapotranspiration, temperature and the number of dry months increase in the same direction. Also the average annual temperature of the watershed and its relative humidity and evaporation are 17.8°C, 68.5 % and 1398.34 millimeters, respectively. Tamer watershed was divided into 15 sub-watersheds by adding an outlet in the site of Tamar gauging station. In this study, the SWAT model was used to simulate erosion and sedimentation. To compare the measured and simulated data and evaluation of the SWAT performance in terms of simulating flow and sediments, daily flow (cubic meters per second) and sediment (tons per day) data at the Tamar gauging station located in Tamar’s watershed outlet was collected from the studies of water resources organization (Tamab). Simulated values were generally consistent with the data observed during calibration and validation period. At this stage of calibration, the SUFI-2 model was used to optimize the parameter values. In this study, daily rainfall and temperature data recorded during an 8-year period by the stations within the watershed were imported into the model. The daily discharge data and daily sediment data of Tamar station recorded during 1999- 2006 were selected. Then model was run using runoff and sediment parameters, and ranges of parameters were adjusted at each iterations, and therefore SWAT model was calibrated using SUFI-2 model. After calibration, model must be validated and its ability to predict future events must be determined. Validation was performed using the runoff and sediment data recorded in Tamar gauging station from 2007 to 2010.
Results and Discussion: NS, R2, R-factor and P-factor were estimated for runoff calibration about 0.76, 0.77, 0.06 and 69 and for runoff evaluation 0.72, 0.75, 0.05 and 69 respectively. The same parameters were also measured for sediment calibration about 0.54, 0.62, 0.15, and 16 and sediment evaluation 0.55, 0.61, 0.35, and 12 respectively. The results showed that irrigated agriculture 24.95 and 15.56 t ha -1y-1 respectively, with average erosion and sediment ha of agriculture by an average of 20.23 and 12.33 t ha -1y-1 respectively erosion and sediment erosion and deposition are tons per hectare maximum value. Results also showed that the soil loss caused by erosion in this watershed is average 6.49 t ha -1y-1 in sediment and 10.28 t ha -1y-1 in erosion.
Conclusion: The assessment factors showed that model has successfully simulated the daily runoff discharge during calibration and validation phases with a Nash-Sutcliffe coefficient of 0.76 and 0.72. A Nash-Sutcliffe coefficient above 0.5 could be acceptable for sediment simulation. However, sediment load simulated for rainy seasons has been lower than actual value while this value has been higher than actual value during dry seasons. In most months of the year, model results are higher than measured values and this issue is more pronounced in the peak runoffs. This issue is due to limitations in spatial distribution of rainfall, so when a small area in watershed experience a severe rainfall, model considers the impact for the entire watershed and therefore overestimates the total runoff. The results showed that SWAT model can be a useful tool for the simulation of flow and sediment basins in the loess land.
Simulation results showed that land use changes have resulted in corresponding increases in surface runoff and sediment. Rates were highly variable both spatially and temporally, and the agricultural lands were most significantly affected. These land use changes have negative implications for the ecological health of the river system as and local communities.
