The dataset provides an estimate of 2020 herbaceous mostly annual fractional cover predicted on May 1st with an emphasis on annual exotic grasses Historically, similar maps were produced at a spatial resolution of 250m (Boyte et al. 2019 https://doi.org/10.5066/P9ZEK5M1., Boyte et al. 2018 https://doi.org/10.5066/P9KSR9Z4.), but we are now mapping at a 30m resolution (Pastick et al. 2020 doi:10.3390/rs12040725). This dataset was generated using in situ observations from Bureau of Land Management’s (BLM) Assessment, Inventory, and Monitoring data (AIM) plots; weekly composites of harmonized Landsat and Sentinel-2 (HLS) data (https://hls.gsfc.nasa.gov/); relevant environmental, vegetation, remotely sensed, and geophysical drivers. These data were integrating into regression tree (RT) models for prediction of weekly cloud free Normalized Difference Vegetation Index (NDVI). A total 11,002 AIM plots from years 2016 - 2019 were used to train an ensemble of five-fold RT models using a cross-validation approach (each observation was used as test data once). Cheatgrass (Bromus tectrorum) is the most common species, however, number of other species were included in this study: Bromus arvensis L., Bromus briziformis, Bromus catharticus Vahl, Bromus commutatus, Bromus diandrus, Bromus hordeaceus L., Bromus japonicus, Bromus madritensis L., Bromus racemosus, Bromus rubens L., Bromus secalinus L., Bromus texensis (Shear) Hitchc., and Taeniatherum caput-medusae. The geographic coverage includes rangelands in the Great Basin, the Snake River Plain, the state of Wyoming, and contiguous areas. We applied a mask to areas above 2700-m elevation to target areas of unlikely substantial annual grass cover. To target likely sagebrush ecosystems, the mask also removed pixels classified as something other than shrub or grassland/herbaceous by the 2016 National Land Cover Dataset (NLCD).

The dataset provides a near real time estimation of 2020 herbaceous mostly annual fractional cover predicted on July 1st with an emphasis on annual exotic grasses Historically, similar maps were produced at a spatial resolution of 250m (Boyte et al. 2019 https://doi.org/10.5066/P96PVZIF., Boyte et al. 2018 https://doi.org/10.5066/P9RIV03D.), but starting this year we are mapping at a 30m resolution (Pastick et al. 2020 doi:10.3390/rs12040725). This dataset was generated using in situ observations from Bureau of Land Management’s (BLM) Assessment, Inventory, and Monitoring data (AIM) plots; weekly composites of harmonized Landsat and Sentinel-2 (HLS) data (https://hls.gsfc.nasa.gov/); relevant environmental, vegetation, remotely sensed, and geophysical drivers. These data were integrating into regression tree (RT) models for prediction of weekly cloud free Normalized Difference Vegetation Index (NDVI). A total 11,065 AIM plots from years 2016 - 2019 were used to train an ensemble of five-fold RT models using a cross-validation approach (each observation was used as test data once). Cheatgrass (Bromus tectrorum) is the most common species, however, a number of other species were included in this study: Bromus arvensis L., Bromus briziformis, Bromus catharticus Vahl, Bromus commutatus, Bromus diandrus, Bromus hordeaceus L., Bromus japonicus, Bromus madritensis L., Bromus racemosus, Bromus rubens L., Bromus secalinus L., Bromus texensis (Shear) Hitchc., and Taeniatherum caput-medusae. The geographic coverage includes rangelands in the Great Basin, the Snake River Plain, the state of Wyoming, and contiguous areas. We applied a mask to areas above 2700-m elevation to target areas of unlikely substantial annual grass cover. To target likely sagebrush ecosystems, the mask also removed pixels classified as something other than shrub or grassland/herbaceous by the 2016 National Land Cover Dataset (NLCD).

The dataset provides a near real time estimation of 2020 herbaceous mostly annual fractional cover predicted on July 1st with an emphasis on annual exotic grasses Historically, similar maps were produced at a spatial resolution of 250m (Boyte et al. 2019 https://doi.org/10.5066/P96PVZIF., Boyte et al. 2018 https://doi.org/10.5066/P9RIV03D.), but starting this year we are mapping at a 30m resolution (Pastick et al. 2020 doi:10.3390/rs12040725). This dataset was generated using in situ observations from Bureau of Land Management’s (BLM) Assessment, Inventory, and Monitoring data (AIM) plots; weekly composites of harmonized Landsat and Sentinel-2 (HLS) data (https://hls.gsfc.nasa.gov/); relevant environmental, vegetation, remotely sensed, and geophysical drivers. These data were integrating into regression tree (RT) models for prediction of weekly cloud free Normalized Difference Vegetation Index (NDVI). A total 11,065 AIM plots from years 2016 - 2019 were used to train an ensemble of five-fold RT models using a cross-validation approach (each observation was used as test data once). Cheatgrass (Bromus tectrorum) is the most common species, however, a number of other species were included in this study: Bromus arvensis L., Bromus briziformis, Bromus catharticus Vahl, Bromus commutatus, Bromus diandrus, Bromus hordeaceus L., Bromus japonicus, Bromus madritensis L., Bromus racemosus, Bromus rubens L., Bromus secalinus L., Bromus texensis (Shear) Hitchc., and Taeniatherum caput-medusae. The geographic coverage includes rangelands in the Great Basin, the Snake River Plain, the state of Wyoming, and contiguous areas. We applied a mask to areas above 2700-m elevation to target areas of unlikely substantial annual grass cover. To target likely sagebrush ecosystems, the mask also removed pixels classified as something other than shrub or grassland/herbaceous by the 2016 National Land Cover Dataset (NLCD).

This dataset release provides historical (2016 - 2024) estimates of fractional cover for Exotic Annual Grass (EAG) species and a native perennial bunch grass in the western United States. The dataset includes five fractional cover maps per year, accompanied by five corresponding confidence maps, for a group of 15 species of EAGs. The following 15 species are included in the overall EAG cover estimate (species followed by * indicate specific maps for that species); field brome* (Bromus arvensis), medusahead* (Taeniatherum caput-medusae), cheatgrass* (Bromus tectorum), rattlesnake brome (Bromus briziformis), rescuegrass (Bromus catharticus), Bald brome (Bromus commutatus and Bromus racemosus), ripgut brome (Bromus diandrus), soft brome (Bromus hordeaceus and Bromus hordeaceus spp. hordeaceus), Japanese brome (Bromus japonicus), compact brome (Bromus madritensis and Bromus madritensis ssp. Rubens), red brome (Bromus rubens), rye brome (Bromus secalinus). Sandberg blue grass (Poa secunda) is not considered an EAG by this project or included in the EAG layer. We map Poa secunda separately as it can have similar phenology to many invasive grasses such as cheatgrass. The data were generated using a combination of field observations from Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) plots; remotely sensed data from the Harmonized Landsat and Sentinel-2 (HLS) product (specifically Normalized Difference Vegetation Index (NDVI) , and various environmental, vegetation, remotely sensed, and geophysical drivers. Additionally, artificial intelligence and machine learning techniques were employed in the data generation process. For 2024 maps we used a 40,154 AIM plots from 2016-2024. Pixels over 2350-m in elevation or classified as other than grassland/herbaceous or shrub (likely rangelands) were masked based on the 2023 National Land Cover Database (NLCD).

This dataset release provides historical (2016 - 2024) estimates of fractional cover for Exotic Annual Grass (EAG) species and a native perennial bunch grass in the western United States. The dataset includes five fractional cover maps per year, accompanied by five corresponding confidence maps, for a group of 15 species of EAGs. The following 15 species are included in the overall EAG cover estimate (species followed by * indicate specific maps for that species); field brome* (Bromus arvensis), medusahead* (Taeniatherum caput-medusae), cheatgrass* (Bromus tectorum), rattlesnake brome (Bromus briziformis), rescuegrass (Bromus catharticus), Bald brome (Bromus commutatus and Bromus racemosus), ripgut brome (Bromus diandrus), soft brome (Bromus hordeaceus and Bromus hordeaceus spp. hordeaceus), Japanese brome (Bromus japonicus), compact brome (Bromus madritensis and Bromus madritensis ssp. Rubens), red brome (Bromus rubens), rye brome (Bromus secalinus). Sandberg blue grass (Poa secunda) is not considered an EAG by this project or included in the EAG layer. We map Poa secunda separately as it can have similar phenology to many invasive grasses such as cheatgrass. The data were generated using a combination of field observations from Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) plots; remotely sensed data from the Harmonized Landsat and Sentinel-2 (HLS) product (specifically Normalized Difference Vegetation Index (NDVI) , and various environmental, vegetation, remotely sensed, and geophysical drivers. Additionally, artificial intelligence and machine learning techniques were employed in the data generation process. For 2024 maps we used a 40,154 AIM plots from 2016-2024. Pixels over 2350-m in elevation or classified as other than grassland/herbaceous or shrub (likely rangelands) were masked based on the 2023 National Land Cover Database (NLCD).

This dataset release provides historical (2016 - 2022) estimates of fractional cover for exotic annual grass (EAG) species and a native perennial bunch grass in the arid and semi-arid rangelands of the western United States. The dataset includes four fractional cover maps per year, accompanied by corresponding confidence maps, for a group of 16 species of EAGs, cheatgrass (Bromus tectorum); medusahead (Taeniatherum caput-medusae); and Sandberg bluegrass (Poa secunda). The data were generated using a combination of field observations from Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) plots; remotely sensed data from the Harmonized Landsat and Sentinel-2 (HLS) product (specifically Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI)), and various environmental, vegetation, remotely sensed, and geophysical drivers. Additionally, artificial intelligence and machine learning techniques were employed in the data generation process. It should be noted that the training of regression-tree models and the development of historical maps (2016-2020) utilized a total of 17,536 AIM plots from years 2016 – 2019. For the creation of 2021 maps, 19,415 AIM plots from years 2016 - 2021 were employed, and 2022 maps, 28,901 AIM plots from 2016-2022 were used. In the case of 2016 – 2020 maps, areas above 2250-m elevation and pixels classified other than grassland/herbaceous or shrub (likely rangelands) were masked based on the 2016 National Land Cover Dataset (NLCD). For the 2021 and 2022 maps, areas above 2350-m elevation and pixels classified as other than grassland/herbaceous by the 2019 NLCD were masked. The seed source variable from the Rangeland Analysis Platform (RAP) [Jones et al., 2018]) was used as one of the drivers for modeling of 2016 – 2020 maps but was not utilized for modeling of 2021 and 2022 maps. Additionally, HLS NDWI were not used for 2021 and 2022 maps. All other predictor variables are identical for all sets of maps. For details, please check data quality information section.

This dataset release provides historical (2016 - 2023) estimates of fractional cover for Exotic Annual Grass (EAG) species and a native perennial bunch grass in the arid and semi-arid rangelands of the western United States. The dataset includes four (five for 2023) fractional cover maps per year, accompanied by corresponding confidence maps, for a group of 16 species of EAGs, Cheatgrass (Bromus tectorum); Medusahead (Taeniatherum caput-medusae); and Sandberg Bluegrass (Poa secunda). Field Brome (Bromus arvensis) is added as individual map species in 2023. The data were generated using a combination of field observations from Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) plots; remotely sensed data from the Harmonized Landsat and Sentinel-2 (HLS) product (specifically Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI)), and various environmental, vegetation, remotely sensed, and geophysical drivers. Additionally, artificial intelligence and machine learning techniques were employed in the data generation process. It should be noted that the training of regression-tree models and the development of historical maps (2016-2020) utilized a total of 17,536 AIM plots from years 2016 – 2019. For the creation of 2021 maps, 19,415 AIM plots from years 2016 - 2021 were employed, and 2022 maps, 28,901 AIM plots from 2016-2022 were used. In the case of 2016 – 2020 maps, areas above 2250-m elevation and pixels classified other than grassland/herbaceous or shrub (likely rangelands) were masked based on the 2016 National Land Cover Database (NLCD). 2021 onward maps, areas above 2350-m elevation and pixels classified as other than grassland/herbaceous by the 2019 NLCD for 2021 and 2022 and 2021 NLCD for 2023 were masked. The seed source variable from the Rangeland Analysis Platform (RAP) [Jones et al., 2018]) was used as one of the drivers for modeling of 2016 – 2020 maps but was not utilized for modeling of 2021 and later maps. Additionally, HLS NDWI were not used after 2021 maps. All other predictor variables are identical for all sets of maps. For details, please check data quality information section.

This dataset release provides historical (2016 - 2023) estimates of fractional cover for Exotic Annual Grass (EAG) species and a native perennial bunch grass in the arid and semi-arid rangelands of the western United States. The dataset includes four (five for 2023) fractional cover maps per year, accompanied by corresponding confidence maps, for a group of 16 species of EAGs, Cheatgrass (Bromus tectorum); Medusahead (Taeniatherum caput-medusae); and Sandberg Bluegrass (Poa secunda). Field Brome (Bromus arvensis) is added as individual map species in 2023. The data were generated using a combination of field observations from Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) plots; remotely sensed data from the Harmonized Landsat and Sentinel-2 (HLS) product (specifically Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI)), and various environmental, vegetation, remotely sensed, and geophysical drivers. Additionally, artificial intelligence and machine learning techniques were employed in the data generation process. It should be noted that the training of regression-tree models and the development of historical maps (2016-2020) utilized a total of 17,536 AIM plots from years 2016 – 2019. For the creation of 2021 maps, 19,415 AIM plots from years 2016 - 2021 were employed, and 2022 maps, 28,901 AIM plots from 2016-2022 were used. In the case of 2016 – 2020 maps, areas above 2250-m elevation and pixels classified other than grassland/herbaceous or shrub (likely rangelands) were masked based on the 2016 National Land Cover Database (NLCD). 2021 onward maps, areas above 2350-m elevation and pixels classified as other than grassland/herbaceous by the 2019 NLCD for 2021 and 2022 and 2021 NLCD for 2023 were masked. The seed source variable from the Rangeland Analysis Platform (RAP) [Jones et al., 2018]) was used as one of the drivers for modeling of 2016 – 2020 maps but was not utilized for modeling of 2021 and later maps. Additionally, HLS NDWI were not used after 2021 maps. All other predictor variables are identical for all sets of maps. For details, please check data quality information section.

This dataset release provides historical (2016 - 2023) estimates of fractional cover for Exotic Annual Grass (EAG) species and a native perennial bunch grass in the arid and semi-arid rangelands of the western United States. The dataset includes four (five for 2023) fractional cover maps per year, accompanied by corresponding confidence maps, for a group of 16 species of EAGs, Cheatgrass (Bromus tectorum); Medusahead (Taeniatherum caput-medusae); and Sandberg Bluegrass (Poa secunda). Field Brome (Bromus arvensis) is added as individual map species in 2023. The data were generated using a combination of field observations from Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) plots; remotely sensed data from the Harmonized Landsat and Sentinel-2 (HLS) product (specifically Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI)), and various environmental, vegetation, remotely sensed, and geophysical drivers. Additionally, artificial intelligence and machine learning techniques were employed in the data generation process. It should be noted that the training of regression-tree models and the development of historical maps (2016-2020) utilized a total of 17,536 AIM plots from years 2016 – 2019. For the creation of 2021 maps, 19,415 AIM plots from years 2016 - 2021 were employed, and 2022 maps, 28,901 AIM plots from 2016-2022 were used. In the case of 2016 – 2020 maps, areas above 2250-m elevation and pixels classified other than grassland/herbaceous or shrub (likely rangelands) were masked based on the 2016 National Land Cover Database (NLCD). 2021 onward maps, areas above 2350-m elevation and pixels classified as other than grassland/herbaceous by the 2019 NLCD for 2021 and 2022 and 2021 NLCD for 2023 were masked. The seed source variable from the Rangeland Analysis Platform (RAP) [Jones et al., 2018]) was used as one of the drivers for modeling of 2016 – 2020 maps but was not utilized for modeling of 2021 and later maps. Additionally, HLS NDWI were not used after 2021 maps. All other predictor variables are identical for all sets of maps. For details, please check data quality information section.

This dataset release provides historical (2016 - 2023) estimates of fractional cover for Exotic Annual Grass (EAG) species and a native perennial bunch grass in the arid and semi-arid rangelands of the western United States. The dataset includes four (five for 2023) fractional cover maps per year, accompanied by corresponding confidence maps, for a group of 16 species of EAGs, Cheatgrass (Bromus tectorum); Medusahead (Taeniatherum caput-medusae); and Sandberg Bluegrass (Poa secunda). Field Brome (Bromus arvensis) is added as individual map species in 2023. The data were generated using a combination of field observations from Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) plots; remotely sensed data from the Harmonized Landsat and Sentinel-2 (HLS) product (specifically Normalized Difference Vegetation Index (NDVI) and Normalized Difference Water Index (NDWI)), and various environmental, vegetation, remotely sensed, and geophysical drivers. Additionally, artificial intelligence and machine learning techniques were employed in the data generation process. It should be noted that the training of regression-tree models and the development of historical maps (2016-2020) utilized a total of 17,536 AIM plots from years 2016 – 2019. For the creation of 2021 maps, 19,415 AIM plots from years 2016 - 2021 were employed, and 2022 maps, 28,901 AIM plots from 2016-2022 were used. In the case of 2016 – 2020 maps, areas above 2250-m elevation and pixels classified other than grassland/herbaceous or shrub (likely rangelands) were masked based on the 2016 National Land Cover Database (NLCD). 2021 onward maps, areas above 2350-m elevation and pixels classified as other than grassland/herbaceous by the 2019 NLCD for 2021 and 2022 and 2021 NLCD for 2023 were masked. The seed source variable from the Rangeland Analysis Platform (RAP) [Jones et al., 2018]) was used as one of the drivers for modeling of 2016 – 2020 maps but was not utilized for modeling of 2021 and later maps. Additionally, HLS NDWI were not used after 2021 maps. All other predictor variables are identical for all sets of maps. For details, please check data quality information section.