SBIR-STTR Award

Central Iowa-based small business, Agren, in collaboration with the ARS Sedimentation Laboratory and the University of Tennessee, has developed two GIS-based soil loss modeling tools. Each tool combines remotely-sensed, high-resolution digital elevation models (LiDAR) and newly developed hydrologic algorithms with long standing soil-loss modeling principles. The results to date are two-dimensional, distributed (grid-based) soil loss models for sheet and rill erosion and ephemeral gully erosion. However, with flow networks that may approach one-million cells on one to three-meter raster spacing, runtimes for a single rasterized field take several hours. The models are not practical to deploy at a large scale for localized conservation planning. The proposed Phase I work plan will determine the technical feasibility of optimizing the SoilLoss Calculator code set to meet user requirements in a web-based system. In order to be commercially viable, the model needs to process an entire field and return a result to the user in under one-minute, requiring the execution time to be reduced by about 400 times. Research questions that will be explored include: Where within the model infrastructure are opportunities for significant speed increases? What alterations to the code set can be proposed as possible options for further optimization? What approximate decrease in execution time may be realized by each proposed change? What are the estimated time and resource requirements for each proposed code modification? The proposed project expands the research and development of models for distributed erosion modeling, which are substantial scientific improvements over existing soil loss modeling tools. Moreover, if successful, these efforts will make the concepts of precision conservation a practical reality for land managers. Conservation practitioners will be able to quickly and easily target appropriate management practices to landscape positions that contribute the most significant sediment loads and evaluate which practice alternatives provide the most environmental benefit. If optimization efforts prove successful, Agren intends to incorporate both the SoilLoss Calculator and ephemeral gully calculator models into their commercially-available suite of conservation planning tools. Agren provides access to the conservation planning tools through a web-based, Software as a Service model. Subscribers pick and choose from a suite of conservation practice planning tools to create a package that best suits their local needs. Agren’s target markets include both public sector conservation planning agencies and private sector audiences such as engineering firms, land improvement contractors, and agricultural retailers.

Central Iowa-based small business, Agren, Inc., has collaborated with the USDA-ARS Sedimentation Laboratory and the University of Tennessee since 2007 to develop two soil loss modeling tools currently referred to as the Revised Universal Soil Loss Equation- Raster (RUSLER) and the Ephemeral Gully Erosion Estimator (EphGEE). Each tool combines remotely-sensed, high-resolution digital elevation models and newly developed hydrologic algorithms with long standing soil-loss modeling principles. The results are two-dimensional, distributed soil loss predictions for sheet, rill, and ephemeral gully erosion. Agren has incorporated the RUSLER methodology and RUSLE2 modeling engine as the core technologies underlying their SoilCalculator tool. With an easy to use interface and intuitive, captivating map-based outputs, SoilCalculator has the potential to revolutionize soil conservation planning. However, long runtimes, excessive use of server resources, and computational issues with the legacy RUSLE2 modeling make it impossible to offer SoilCalculator as a commercially viable product. Progress made during Phase I SBIR indicates scaling SoilCalculator is technically feasible and attainable in the near-term. Code changes internal to the RUSLE2 model have reduced runtimes by 100. This, along with a prototype parallel computing architecture, shows promise to further increase computational efficiency and result in reliable, cost-effective simulations. The objective of Agren's Phase II work is to continue to overcome the technical and scientific limitations of offering a full-suite of erosion prediction services using the core science and modeling framework of RUSLE2. Research questions include:How are proposed modifications to the distributed SoilCalculator framework and RUSLER DLL related to scalability and compute time? Can these modifications result in a tool that meets end-user requirements for delivery via a cloud-based precision ag platform? Can UAV-collected photogrammetry yield elevation models compare to proven ground-collection methods for measuring small changes in ephemeral gullies? What is the relationship of EphGEE model results to measured values of sediment loss in ephemeral gullies in different geographic areas and under different management systems? How can EphGEE be configured to improve correlation with measured values? How can the RUSLER science and modeling framework be leveraged to create added saleable value for SoilCalculator? The proposed project expands the research, development and validation of models for distributed erosion modeling, which are substantial scientific improvements over existing modeling tools. These efforts will lead to transformational methods of conservation planning by delivering erosion prediction services commercially, through the private sector. If Agren is successful in overcoming the technical limitations to scaling SoilCalculator, they have commitment from a distribution partner to deliver SoilCalculator on over 40 million acres of U.S. cropland.