Skip to main content
ARS Home » Southeast Area » Oxford, Mississippi » National Sedimentation Laboratory » Watershed Physical Processes Research » Research » Publications at this Location » Publication #285270

Title: Estimation of runoff, peak discharge and sediment load at the event scale in a medium-size Mediterranean watershed using the AnnAGNPS model

Author
item BISANTINO, TIZIANA - University Of Bari
item Bingner, Ronald - Ron
item CHOUAIB, WAFA - National Research Institute For Rural Engineering, Water And Forestry (NRIREWF)
item GENTILE, FRANCESCO - University Of Bari
item LIUZZI, GIULIAN - University Of Bari

Submitted to: Land Degradation and Development
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 2/26/2013
Publication Date: 4/3/2013
Citation: Bisantino, T., Bingner, R.L., Chouaib, W., Gentile, F., Liuzzi, G.T. 2013. Estimation of runoff, peak discharge and sediment load at the event scale in a medium-size Mediterranean watershed using the AnnAGNPS model. Land Degradation and Development. Wiley Online Library. DOI: 10.1002/ldr.2213.

Interpretive Summary: Evaluation and assessment of watershed agricultural management practices are critical in applying the best practices possible to control erosion and improve the downstream water quality. Watershed models have been extensively used within the U.S. for this purpose, but evaluations outside of the U.S. have been limited. An evaluation of the USDA watershed model AnnAGNPS was performed within a southern Italy experimental watershed for 2007-2009. Calibration and validation were performed for separate two and one year periods, respectively. A sensitivity analysis for the most meaningful parameters was performed to evaluate the weight of each factor and to assign the correct parameterization avoiding a long sequence of simulations to calibrate the model. The Soil Conservation Service (SCS) Type I rainfall distribution was found to be suitable for these Mediterranean conditions as it provided the best values of the statistical parameters used to evaluate the performance of the model. The model predictions have proven to be good for runoff and peak discharge and satisfactory for sediment yield. The relative error is lower for high events, this result being quite interesting in semi-arid environments, where most of the annual sediment yield is concentrated in few, severe events. Understanding how the model performs under unique conditions with agricultural practices that are not common in the U.S. can provide insight into enhancements that may be needed to improve the model. These model assessments are critical since they are often the only means to evaluate the effects of conservation practices on water quality within a watershed.

Technical Abstract: Sediment transport in rivers is an indicator of soil eroded from various sediment sources, of which agricultural land can be a significant one, and the intensity of the phenomenon provides a measure of land degradation at a watershed level. The use of distributed models to solve problems in water resources applications including environmental impacts of land-use changes, effects of climate change on water resources, and water planning and management in a catchment, is quite complex in large Mediterranean catchments for the lack of sufficient data to characterize spatial variability, for the integration of field measurements and model parameter element, and for the imperfect representations of the real processes. In this paper the AnnAGNPS model has been used to estimate runoff, peak discharge and sediment load at the event scale in a semi-arid, medium-size watershed. The study area is the Carapelle torrent (Southern Italy), where continuous rainfall, streamflow and sediment load data are available. Nineteen flood events have been registered in the period 2007-2009 and selected for the application of the model. The aim of the paper is the evaluate the predictive accuracy of the model, given the specific conditions of the semi-arid environments. Particular attention has been focused on the definition of the local crop parameters, of the management scheduling features and to the influence of the synthetic storm distributions on peak discharge predictions. The sensitivity analysis for the most meaningful parameters has been carried out to evaluate the weight of each factor and to assign the correct parameterization avoiding a long sequence of proves to calibrate the model. The SCS Type I was found to be the suitable storm distribution as it provided the highest values of the statistical parameters used to evaluate the performance of the model. The model predictions have proven to be good for runoff and peak discharge and satisfactory for sediment yield. The relative error is lower for high events, this result being quite interesting in semi-arid environments, where most of the annual sediment yield is concentrated in few, severe events.