2012 Annual Report
The NRCS has been using various versions of the RUSLE2 for many years. RUSLE2 is an essential tool for NRCS and the programming is needed to integrate into the NRCS Conservation Delivery Streamlining Initiative (CDSI). RUSLE2 will provide before and after soil loss predictions for a conservation system and a Soil Conditioning Index estimate quantifying resource concerns for soil erosion and soil quality. The model requires programming to enable field offices to provide technical assistance to private landowners.
Tasks to be performed on a continual basis during the term of this agreement: a. Develop the RUSLE2D version of RUSLE2 that has the functionality of the latest version of RUSLE2 and the added spatial erosion output potentially useful for inclusion in the CDSI effort. b. Train the key NRCS personnel on database development. c. Provide technical support to NRCS throughout the term on this Agreement, including addressing any bugs or other problems with the RUSLE2 model due to the insertion of new code associated with this contract. d. Keep the NRCS Responsible Person informed of any changes to RUSLE2 model. e. Provide support to the DLL version in NRCS’s efforts to make it capable of utilizing the NRCS Land Operations Management Database (LMOD) and National Soils Database data. f. Evaluate the Soil Organic Matter subroutine across several areas nationwide to determine if it provides satisfactory results. g. Evaluate the accuracy of the newly inserted concentrated flow erosion subroutine into the RUSLE2 program. h. Evaluate the effect of degrading above ground and buried residue pools at different rates.
Ongoing activities include training of key NRCS personnel, technical support, communications. Work is commencing on the development of soil organic matter routines. Work is underway to increase the speed of the R2D application. It has been necessary to revise and update the RUSLE2 code base to correct “bugs”, to implement new capabilities, and to deliver recent scientific improvements. In addition, recently-developed methods now allow RUSLE2 to predict spatially-distributed erosion over an entire agricultural field using a Digital Elevation Model to represent terrain topography. The new computationally-demanding 2D computations required upgrades to RUSLE2’s simulation engine for improved efficiency.
A workshop gathering model developers was conducted at the NRCS Information Technology Center in Ft. Collins, CO, in May 2012. A thorough revision of RUSLE2’s source code was performed, and technical discussions were held to solve known issues. Model developers identified areas where improvements were needed, and determined modifications required for implementation of new erosion prediction science. New approaches to increase the model’s performance were proposed and tested. Technical discussions were conducted to define modifications to the RUSLE2’s computational engine so that it can be used more efficiently when predicting spatially-distributed erosion over large fields.
A more efficient procedure has been developed to speed up 2D grid-based computations. Implemented directly into the RUSLE2’s engine, it will reduce the number of required calculations and save computational time by saving and reusing certain computed parameters that do not vary in space.
The code was revised and tested to verify compatibility with the newest compilers, to improve software maintenance and to benefit from the most recent advances in computer hardware.
Automated terrain analysis methods were developed to determine surface drainage patterns, determine the presence and location of channels, and compute topographic properties. The new methods perform better than previously existing algorithms near field boundaries and flow barriers. The methods have been programmed as a software library, and have been used to describe surface drainage for 2D erosion predictions with RUSLE2, and to determine the location and extent of ephemeral gullies, whose size evolution and soil loss are estimated by the recently developed gully erosion model.