Innovations in machine learning and cloud‐based computing were merged with historical remote sensing and field data to provide the first moderate resolution, annual, percent cover maps of plant functional types across rangeland ecosystems to effectively and efficiently respond to pressing challenges facing conservation of biodiversity and ecosystem services. We utilized the historical Landsat satellite record, gridded meteorology, abiotic land surface data, and over 30,000 field plots within a Random Forests model to predict per‐pixel percent cover of annual forbs and grasses, perennial forbs and grasses, shrubs, and bare ground over the western United States from 1984 to 2017. Results were validated using three independent collections of plot‐level measurements, and resulting maps display land cover variation in response to changes in climate, disturbance, and management. The maps, which will be updated annually at the end of each year, provide exciting opportunities to expand and improve rangeland conservation, monitoring, and management. The data open new doors for scientific investigation at an unprecedented blend of temporal fidelity, spatial resolution, and geographic scale.

Operational satellite remote sensing products are transforming rangeland management and science. Advancements in computation, data storage and processing have removed barriers that previously blocked or hindered the development and use of remote sensing products. When combined with local data and knowledge, remote sensing products can inform decision‐making at multiple scales. We used temporal convolutional networks to produce a fractional cover product that spans western United States rangelands. We trained the model with 52,012 on‐the‐ground vegetation plots to simultaneously predict fractional cover for annual forbs and grasses, perennial forbs and grasses, shrubs, trees, litter and bare ground. To assist interpretation and to provide a measure of prediction confidence, we also produced spatiotemporal‐explicit, pixel‐level estimates of uncertainty. We evaluated the model with 5,780 on‐the‐ground vegetation plots removed from the training data. Model evaluation averaged 6.3% mean absolute error and 9.6% root mean squared error. Evaluation with additional datasets that were not part of the training dataset, and that varied in geographic range, method of collection, scope and size, revealed similar metrics. Model performance increased across all functional groups compared to the previously produced fractional product. The advancements achieved with the new rangeland fractional cover product expand the management toolbox with improved predictions of fractional cover and pixel‐level uncertainty. The new product is available on the Rangeland Analysis Platform ( https://rangelands.app/), an interactive web application that tracks rangeland vegetation through time. This product is intended to be used alongside local on‐the‐ground data, expert knowledge, land use history, scientific literature and other sources of information when making interpretations. When being used to inform decision‐making, remotely sensed products should be evaluated and utilized according to the context of the decision and not be used in isolation.

Abstract:Near surface (i.e., camera) and satellite remote sensing metrics have become widely used indicators of plant growing seasons. While robust linkages have been established between field metrics and ecosystem exchange in many land cover types, assessment of how well remotely-derived season start and end dates depict field conditions in arid ecosystems remain unknown. We evaluated the correspondence between field measures of start (SOS; leaves unfolded and canopy greenness > 0) and end of season (EOS) and canopy greenness for two widespread species in southwestern U.S. ecosystems with those metrics estimated from near-surface cameras and MODIS NDVI for five years (2012)(2013)(2014)(2015)(2016). Using Timesat software to estimate SOS and EOS from the phenocam green chromatic coordinate (GCC) greenness index resulted in good agreement with ground observations for honey mesquite but not black grama. Despite differences in the detectability of SOS and EOS for the two species, GCC was significantly correlated with field estimates of canopy greenness for both species throughout the growing season. MODIS NDVI for this arid grassland site was driven by the black grama signal although a mesquite signal was discernable in average rainfall years. Our findings suggest phenocams could help meet myriad needs in natural resource management.

No abstract

No abstract