A computationally efficient approach for watershed scale spatial optimization

A multi-level spatial optimization (MLSOPT) approach is developed for solving complex watershed scale optimization problems. The method works at two levels: a watershed is divided into small sub-watersheds and optimum solutions for each sub-watershed are identified individually. Subsequently sub-watershed optimum solutions are used for watershed scale optimization. The approach is tested with complex spatial optimization case studies designed to maximize crop residue (corn stover) harvest with minimum environmental impacts in a 2000 km2 watershed. Results from case studies indicated that the MLSOPT approach is robust in convergence and computationally efficient compared to the traditional single-level optimization frameworks. The MLSOPT was 20 times computationally efficient in solving source area based optimization problem while it was 3 times computationally efficient for watershed outlet based optimization problem compared to a corresponding single-level optimizations. The MLSOPT optimization approach can be used in solving complex watershed scale spatial optimization problems effectively. [Display omitted] •A novel spatial optimization approach (MLSOPT) is developed for complex spatial optimization.•MLSOPT reduced optimization complexity with multiple optimization levels.•Performance of MLSOPT with single level optimization test cases was evaluated.•MLSOPT was robust in convergence very efficient in solving spatial optimization problems.