Skip to main content
ARS Home » Southeast Area » Oxford, Mississippi » National Sedimentation Laboratory » Watershed Physical Processes Research » Research » Publications at this Location » Publication #320481
Title: Automated watershed subdivision for simulations using multi-objective optimization

Author
item MOMM, HENRIQUE - Middle Tennessee State University
item Bingner, Ronald - Ron
item EMILAIRE, R - Middle Tennessee State University
item Garbrecht, Jurgen
item Wells, Robert - Rob
item Kuhnle, Roger

Submitted to: Hydrological Sciences Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 5/2/2017
Publication Date: 8/2/2017
Citation: Momm, H.G., Bingner, R.L., Emilaire, R., Garbrecht, J.D., Wells, R.R., Kuhnle, R.A. 2017. Automated watershed subdivision for simulations using multi-objective optimization. Hydrological Sciences Journal. 62:10, 1564-1582 DOI: 10.1080/02626667.2017.1346794.

Interpretive Summary: Optimizing the level of watershed subdivision into homogeneous areas for watershed simulations improves the runoff and sediment yield results needed in watershed management plan development. An optimization approach was integrated with a watershed pollutant loading management model using landscape characterizations describing topographic, soil, and management information to minimize the number of watershed subdivisions. Thus, the approach improved the simulation results while reducing the computational effort required for the simulations. This study provides an approach that can be used as a guideline describing watershed landscapes when applying watershed technology in developing management plans for conservation practice implementation by action agencies.

Technical Abstract: The development of watershed management plans to evaluate placement of conservation practices typically involves application of watershed models. Incorporating spatially variable watershed characteristics into a model often requires subdividing the watershed into small areas to accurately account for topographic, landuse or soil features. A qualitative trial-and-error approach is commonly used to define these watershed subdivisions. In the majority of cases, a single constant topographic threshold value is considered for the entire watershed. A methodology has been developed to quantitatively determine the set of spatially variable threshold values using two landscape attributes: topography and another user-defined reference layer. An optimization algorithm was integrated with topographic parameterization to create solutions that meet two objectives: minimize the number of watershed subdivision and maximize the agreement between the discretized watershed and the reference layer. The optimized framework was illustrated on a North Central Mississippi watershed where some reference layers were observed to be more significant than others when evaluating estimated sediment loads using the USDA watershed model, AnnAGNPS. An agricultural management layer utilized as the reference layer in subdividing the watershed yielded results closer to observed sediment loads than scenarios subdivided with approximately 10 times more sub-catchments. This can be attributed to the significant effect of land use/land cover to the model’s ability to estimate daily variations in water balance properties and curve number. Watershed modeling and simulation technology has been recognized as a potential tool to support sustainable watershed management through timely and cost-effective allocation of resources and proposed adoption of conservation practices. The quality and applicability of this technology depends on the capability of representing and describing physical processes accurately. Adoption of washed modeling technology in management planning starts with the appropriate subdivision of the watershed into modeling basic units and their suitable discretization. The optimization techniques produced in this study can be used as guidelines in applying watershed technology in developing management plans for conservation practice implementation.