Skip to main content
ARS Home » Southeast Area » Tifton, Georgia » Southeast Watershed Research » Research » Publications at this Location » Publication #331016
Title: Assessing applicability of SWAT calibrated at multiple spatial scales from field to stream

Author
item CHO, JAEPIL - Apec Climate Center (APCC)
item HER, YOUNGGU - Florida Department Of Agriculture
item Bosch, David - Dave

Submitted to: Trade Journal Publication
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 3/25/2015
Publication Date: 5/1/2015
Citation: Cho, J., Her, Y., Bosch, D.D. 2015. Assessing applicability of SWAT calibrated at multiple spatial scales from field to stream. Journal of Korean Society of Agricultural Engineers. 57(3):21-39. doi.org/10.5389/KSAE.2015.57.3.021.

Interpretive Summary: Watershed-scale simulation models provide useful tools for evaluating loading of sediment and nutrients from agricultural watersheds. The ability of the SWAT model to simulate agricultural runoff and loading was evaluated using data collected in subwatershed K of the Little River Experimental watershed located in South Atlantic Coastal Plain of the USA. The sediment and nutrient simulation results agreed with understandings of hydrologic characteristics of the watershed. The calibrated model provided good performance in reproducing monthly streamflow of the watershed. However, the sediment and TP simulations provided unsatisfactory performance in the calibration and validation periods, implying that the multi-scale calibration processes applied here may not help obtain good model performance in predicting water quality at a watershed outlet.

Technical Abstract: The capability of SWAT for simulating long-term hydrology and water quality was evaluated using data collected in subwatershed K of the Little River Experimental watershed located in South Atlantic Coastal Plain of the USA. The SWAT model was calibrated to measurements made at various spatial scales from 1979 to 1986. The calibrated model was then validated by comparing the simulated and observed variables of interest for two validation periods, 1987 to 1995 and 1996 to 2004. Typical agronomic management schedules for conventional and conservation practices were defined for each major crop. Watershed and field-scale sediment and nutrient transport along known loading pathways, including upland, riparian buffer, and in-stream processes, were considered. Riparian forest buffers were identified as an important hydrologic entity to be represented in the model. The sediment and nutrient simulation results agreed with understandings of hydrologic characteristics of the watershed. The multi-scale model evaluation processes described in this paper demonstrated a way to consider upland, riparian buffer, and in-stream processes and incorporate measurements made at different spatial scales into model calibration. The calibration practices considering in-stream nutrient transformation processes represented by the QUAL2E module of SWAT provided better performance in predicting total N and total P than did the default simulation of SWAT. The calibrated model provided good performance in reproducing monthly streamflow of the watershed with percent errors of -19.9 to 1.2% and Nash-Sutcliffe Efficiency indexes of 0.89 to 0.93. However, the sediment and TP simulations provided unsatisfactory performance in the calibration and validation periods, implying that the multi-scale calibration processes may not help obtain good model performance in predicting water quality at a watershed outlet.