Location: Southwest Watershed Research Center
Project Number: 2022-13610-013-011-S
Project Type: Non-Assistance Cooperative Agreement
Start Date: Aug 1, 2020
End Date: Jun 30, 2022
Objective:
Sustainability and ecological function of over 160 million hectares of rangelands in the central and western United States depends upon effective management of woody species. Brush management is one of the most cost-shared and implemented NRCS conservation practices on grazing lands. Historically, the USDA-Natural Resources Conservation Service (NRCS) spends an average of $34 million/year on brush management nationwide (2005-2016 data). There is little to no documentation to support or refute the need for brush management in the conservation planning process. Thus, there is a need for an efficient, repeatable way to determine woody canopy cover baseline and change data for documentation in the planning process, and to optimize conservation efforts and targeting at larger scales. The Rangeland Brush Estimation Toolbox (RaBET) is an operational ArcGIS toolbox developed to assess spatial and temporal changes in woody vegetation over large heterogeneous landscapes at the 30m brush management as a conservation practice on rangelands at the field/pasture scale and larger. In addition, RaBET maps of woody canopy cover can be used as planning tools for practice implementation or program initiatives. The objective of this project is to complete development the of RaBET for Major Land Resource Areas (MLRAs) in the western rangelands within five years, and then continue with a “maintenance and update plan” to continue bringing new classified imagery into the toolbox.
Approach:
We anticipate that collaboration between parties will greatly advance the production and quality of an operational RaBET. Completion of RaBET will require four main tasks. Task 1 is the production of MLRA-specific algorithms which create the woody cover maps. This task utilizes a workflow of classifying NAIP or other high resolution imagery to identify woody vegetation, aggregating it up to the Landsat scale, and pairing it with Landsat imagery processed into vegetation indices to develop multiple regression algorithms to produce woody cover maps. Task 2 is the validation effort to discern how closely the maps produced resemble conditions on the ground. This task involves field data collection campaigns to obtain ground measurements to compare against the woody cover maps. Task 3 is the creation of a species-specific woody vegetation product. This task will utilize hyperspectral remotely sensed data which uses electromagnetic spectra unique to individual vegetation types, rather than just vegetation greenness to produce species-specific woody cover maps. It will also explore using LiDAR data for structure and data fusion techniques to produces these products. Task 4 is the development of an online platform to deliver RaBET to users. This task will work with ARS personnel to create an ArcGIS Online version of RaBET that will allow NRCS client data to remain protected and private, create fundamental RaBET mobility, ease-of-use, and the ability to refresh/improve the RaBET tools without interrupting individual users.
