Submitted to: Transactions of the ASAE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 10/1/1999
Publication Date: N/A
Citation: Grunwald, S., Norton, L.D. 1999. An AGNPS-based runoff and sediment yield model for two small watersheds in germany. Transactions of the American Society for Agricultural Engineers. 42(6):1723-1731. Interpretive Summary: Non-point source pollution is a major concern for water quality. We studied the event-based Agricultural Non-Point Source (AGNPS) model to predict surface runoff and sediment yield for two small watersheds with measured data in Bavaria. A modified version of AGNPS that included using stream power instead of the slope length and gradient factor (LS) of the Universal Soil Loss Equation (USLE) and using sediment particle size related to runoff velocity for channel erosion produced better predictions than the unmodified model. The modified model had the greatest model efficiency and more accurately predicted runoff and sediment yield at the watershed outlet than the unmodified AGNPS model. The impact of this work is that more accurate non-point source predictions can be made with the modified model with a minimum of additional inputs required by the user.
Technical Abstract: The AGNPS model is used to simulate runoff, sediment and nutrient transport in watersheds. This study was to compare runoff and sediment predictions from the AGNPS model and modified versions to measured data. Shortcomings of the AGNPS model were examined. The study was conducted with 52 events, 22 for calibration and 30 for validation. Evaluation of outputs was based on statistical comparisons between measured and predicted values for each runoff event. Three surface runoff prediction methods were compared: uncalibrated curve number (CN)(Q1), calibrated CN (Q2), and Lutz (Q3). Modifications to sediment calculations encompassed; replacement of the USLE LS factor (S1) by one based on stream power theory (S2), and linkage of channel erosion by categories of particle size to runoff velocity (S3). Median runoff was under- predicted by 55.5% using Q1, over-predicted by 36.8% using Q2 and by 13.1% using Q3 in watershed G1. Coefficient of efficiency (E) was 0.96 followed by 0.93 for Q2 and 0.25 for Q1 in G1. In watershed G2 median runoff was under-predicted by 80.0% using Q1, and over predicted by 45.0% using Q2 and 35.0% using Q3. Best E was for Q3 (0.83) followed by 0.76 for Q2 and 0.24 for Q1 in G2. Sediment yield was under-predicted by 57.2% using S1, 47.6% using S2 and by 4.8% using S3 in G1. The largest E was 0.90 for S3 followed by 0.57 for S2 and 0.26 for S1 in G1. Median sediment yield was under-predicted by 53.9% using S1, and by 38.5% using S2 and over-predicted by 3.3 % using S3 in G2. E was 0.72 (S3) followed by 0.60 (S2) and 0.57 (S1) in G2. Results showed calibration of CN and Lutz method for runoff and variant S3 for sediments with the AGNPS model gave the best actual vs. predicted results.