Skip to main content
ARS Home » Plains Area » Temple, Texas » Grassland Soil and Water Research Laboratory » Research » Publications at this Location » Publication #319158

Title: Assessment of model predictions and parameter transferability by alternative land use data on watershed modeling

item YEN, HAW - Texas Agrilife Research
item SHARIFI, AMIRREZA - University Of Maryland
item KALIN, LATIF - Auburn University
item MIRHOSSEINI, GOLBAHAR - Auburn University
item Arnold, Jeffrey

Submitted to: Journal of Hydrology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/29/2015
Publication Date: 5/5/2015
Publication URL:
Citation: Yen, H., Sharifi, A., Kalin, L., Mirhosseini, G., Arnold, J.G. 2015. Assessment of model predictions and parameter transferability by alternative land use data on watershed modeling. Journal of Hydrology. 527(2015):458-470.

Interpretive Summary: Large scale watershed models are increasingly used to determine the impacts of land management and climate change on water supply and water quality in lakes and rivers. Large amounts of input data are required to drive the models including data on land use. In this study three different land use maps, derived from satellite data, were obtained and input to the SWAT (Soil and Water Assessment Tool) watershed model. It was determined that SWAT performed best using a regional land use data set. This study will help decision makers recognize the importance of utilizing appropriate input data and subsequently improve engineering applications and policy development.

Technical Abstract: In recent years, complex large-scale watershed models have been developed to perform simulations of hydrologic and nutrient processes. The potential impact caused by human activities such as agricultural implementations against the environment can be evaluated under future scenarios. Meanwhile, large amount of input data are required to enhance the performance of simulated results. For some natural or urban regions, it is possible to have multiple sources of geophysical data available but the associated effects of using alternating data sources on modeling results is not yet evaluated. In this study, three sources of land use data (Mid-Atlantic Regional Earth Science Applications Center (RESAC 2000), National Land Use Cover Dataset (NLCD 2001), and State Land Use/Cover Maps (STATE) were implemented on the Greensboro watershed, Maryland, USA. The Alternative Dataset Scheme (ADS) and the Parameter Transferability Scheme (PTS) were applied to investigate model predictive uncertainty and the potential impact of cross transferring optimal calibration parameters between models. It was demonstrated that model predictions simulated by SWAT model had better performance when RESAC land use map was used, followed by STATE, and NLCD land use maps. In addition, calibrated best parameter set from RESAC has presented relatively more transferable compared to NLCD and STATE. The use of varying data source may not only alter model predictions and the associated predictive uncertainty but also have direct impact on the transferability of model parameters. The major findings in this study may help future modelers and decision makers to recognize the importance of alternative data source selection. Therefore, the quality of subsequent research work, engineering applications or policies can be further improved.