Rangeland Productivity Monitoring Service

Rangeland managers and livestock producers need more timely and consistent tools that produce information to inform grazing strategies, risk management, and allotment management plans. On the ground monitoring is extremely expensive and is therefore difficult to employ consistently limited resources, fewer trained staff and shifting priorities. In addition, National Forest systems are now in various stages of Forest Plan Revisions which require assessments of current rangeland conditions and past vegetation performance in a clear, unbiased manner. In response to this need, Matt Reeves, with the Rocky Mountain Research Station, in partnership with private industry, has developed several new data services freely available to all stakeholders and managers. Specifically, we have developed the Rangeland Production Monitoring Service (RPMS) which consists of 2 components.

First, it includes a retrospective dataset with the objective of mapping and quantifying annual production of all 662 million acres of coterminous US rangelands from 1984 to present. To create this dataset we leveraged the Thematic Mapper data suite warehoused on Google Earth Engine and generated Normalized Difference Vegetation Index (NDVI). These originally data are offered at 30 m spatial resolution and to reduce processing constraints and file size (to greatly enhance download capability) the final product is resampled to 250 m. These data were converted to annual production by processing them through the Rangeland Vegetation Simulator (RVS) (Link to the PDF). The RVS is a simulation system that enables quantification of 1, 10, 100, 1000 hour fuels, standing carbon in shrubs, annual production of herbs, stems per acre, and vegetation response to fire and herbivory. Importantly this new simulation program was used to calibrate the NDVI to annual production. This unprecedented time series enables users of this service to quantify trends in production through time, evaluate inter-annual variability, and quantify recovery from drought and wildfire. This type of information has long been the target of many producers and managers so our development of this innovative service is timely and needed to aid efforts aimed at increasing resiliency and creating better grazing management strategies which can improve economic and ecological resiliency alike.

The second component of the RPMS is a forage projection system that utilizes machine learning to process near real time climate and remote sensing data to estimate the magnitude and timing of annual production across all rangelands in the Northern Region of the USDA, US Forest Service (Region 1). This automated projection system operates between March and July of the growing season and is updated every 2 weeks allowing a new estimate of the total annual yield is made in concert with an estimate of the timing of the peak of the growing season. To ensure full disclosure and transparency we also offer a 95% prediction Interval about the estimate. These data are complementary to the GrassCast offered at: http://grasscast.agsci.colostate.edu/.