Location: Southwest Watershed Research Center
Project Number: 2022-13610-013-032-S
Project Type: Non-Assistance Cooperative Agreement
Start Date: Sep 15, 2025
End Date: Sep 14, 2026
Objective:
The Walnut Gulch Experimental Watershed (WGEW) Long-Term Agroecosystem Research (LTAR) site is an excellent research platform to study rangeland sustainability (especially related to soil and water), but not intensification of production. WGEW contains parts of 5 active ranches, each of which also contain grazed areas that are outside the watershed, and the ranchers do not share animal numbers, weight gains, or grazing regimes. In contrast, 80 kilometers to the west, the Santa Rita Experimental Range (SRER) has similar ecological conditions as WGEW, but is on land owned by the state of Arizona and managed for research purposes by the University. The primary rancher on the SRER shares the details of animal numbers and weight gains, with grazing timing and intensity mutually agreed upon. The SRER is a National Ecological Observatory Network (NEON) Core site, so it will benefit from annual, high resolution NEON Airborne Observation Platform (AOP) aerial photography with coincident hyperspectral and lidar imagery. The overall objective of this project is to develop a common set of field, drone, airplane, and satellite imagery processing protocols to better understand ecological and rangeland management processes at scale, especially related to vegetation production, lifeform, and erosion on both the SRER and WGEW sites.
Approach:
In Phase 1 of this project (2015-2020) we developed methods to better understand vegetation across research sites including: scale lidar data on rangeland vegetation, quantify forage utilization with drone collected imagery, collect and process georegistered drone imagery across large areas, map common ecological sites and states, and classify vegetation and bare soil. In Phase 2 of the project (2021-2025) we developed methods to scale drone imagery (5 cm) to Landsat (30 m) using machine learning models to estimate the connected bare soil area (largest patch index, or LPI) at risk of accelerated erosion from overland flow. LPI has been mapped across the area of Sandyloam Upland on the Santa Rita Experimental range. In Phase 3, we will extend the analysis using historical Landsat imagery starting in 1986 to present. We have assembled input datasets linking drought (as estimated by Standardized Precipitation Evapotranspiration Index, SPEI), grazing intensity, and fire to assess the interaction of those factors on bare soil LPI through time. This area is a mesquite savannah, so ground cover is grass (forage) growing in the interspace between mesquite trees. We will also develop tools to improve the application of the Rangeland Hydrology and Erosion Model (RHEM) in New Mexico, Arizona, Utah and Nevada. Specifically, we will develop a gridded database at 800 meter resolution of weather input files for RHEM and also calculate baseline erosion rates using cover values through time estimated using the Rangeland Analysis Platform (RAP).