Jeremy Maestas Matthew Jones Neal Pastick Matthew Rigge Bruce Wylie Lindy Garner Michele Crist Collin Homer Stephan Boyte Bill Witacre 20200531 Raster Digital Data Set (https://chohnz.users.earthengine.app/) https://doi.org/10.5066/P9VL3LD5 Cheatgrass (Bromus tectorum) and other invasive annual grasses represent one of the single largest threats to the health and resilience of western rangelands. To address this challenge, the Western Governors Association (WGA)-appointed Western Invasive Species Council convened a cheatgrass working group to develop a new regional vision for invasive annual grass management across the West. Foundational to implementing this new vision is the creation of a common spatial map to guide strategic actions. The WGA cheatgrass working group sought to develop a 30-m base map of annual herbaceous cover to support a common spatial strategy for tackling invasive annual grasses across the western U.S. Here, we leverage three large-scale datasets to provide land managers with a product estimating the recent extent (2016-2018) of annuals across western rangelands. Input annual herbaceous datasets include Rangeland Analysis Platform (Jones et al. 2018), US Geological Survey (USGS) Harmonized Landsat and Sentinel (Pastick et al. 2020, Pastick et al. in prep) and USGS National Land Cover Database (NLCD) (Rigge et al. 2020). These three datasets are combined using a weighted mean approach to generate the final annual herbaceous mean cover product across the sagebrush biome (Jeffries and Finn 2019). References: Jeffries, M.I., and Finn, S.P. 2019. The Sagebrush Biome Range Extent, as Derived from Classified Landsat Imagery: U.S. Geological Survey data release, https://doi.org/10.5066/P950H8HS. Jones, M.O., Allred, B.W., Naugle, D.E., Maestas, J.D., Donnelly, P., Metz, L.J., Karl, J., Smith, R., Bestelmeyer, B., Boyd, C., Kerby, J.D., McIver, J.D. 2018. Innovation in rangeland monitoring: annual, 30m, plant functional type percent cover maps for U.S. rangelands, 1984-2017. Ecosphere 9, e02430. https://doi.org/10.1002/ecs2.2430. Pastick, N.J., Dahal, D., Wylie, B.K., Parajuli, S., Boyte, S.P., Wu, Z. 2020. Characterizing Land Surface Phenology and Exotic Annual Grasses in Dryland Ecosystems Using Landsat and Sentinel-2 Data in Harmony. Remote Sens. 12, 725. Pastick, N.J., Dahal, D., Wylie, B.K., Rigge, M.B., Jones, M.O, Allred, B.W., Boyte, S.P., Parajuli, S., and Wu, Z. In prep. Rapid monitoring of the occurrence and spread of exotic annual grasses in the western United States using remote sensing and machine learning. Global Change Biology. Reeves, M., and Mitchell, J. 2011. Extent of Coterminous US Rangelands: Quantifying Implications of Differing Agency Perspectives. Rangeland Ecology and Management 64: 585-597. Rigge, M., Shi, H., Homer, C., Danielson, P., Granneman, B. 2019. Long-term trajectories of fractional component change in the Northern Great Basin, USA. Ecosphere: e02762. Rigge, M., Homer, C., Cleeves, L., Meyer, D., Bunde, B., Shi, H., Xian, G., Bobo, M. 2020. Quantifying Western U.S. Rangelands as Fractional Components with Landsat. Remote Sensing. 12: 412. The goal of the annual herbaceous base map is to support a common spatial strategy for tackling invasive annual grasses across the western U.S. As with all remote sensing-based products, the map presented here is best used alongside local knowledge and data. The map is intended to facilitate cross-boundary regional planning, and it is anticipated that state and local partners will further refine priority areas for management using additional information. Although this Federal Geographic Data Committee-compliant metadata file is intended to document the data set in nonproprietary form, as well as in Esri format, this metadata file may include some Esri-specific terminology. 20160101 20181231 ground condition Complete None planned -122.116081408 -102.260259357 49.0016614443 34.2918384983 USGS Thesaurus terrestrial ecosystems vegetation invasive species grassland ecosystems remote sensing Alexandria Digital Library Feature Type Thesaurus grasslands None cheatgrass Great Basin ISO 19115 Topic Category biota geoscientificInformation USGS Metadata Identifier USGS:5ec5159482ce476925eac3b7 Common geographic areas United States Washington Oregon Montana North Dakota South Dakota Wyoming Idaho Nebraska Nevada Utah California Colorado Arizona New Mexico Great Basin Any downloading and use of these data signifies a user's agreement to comprehension and compliance of the USGS Standard Disclaimer. Insure all portions of metadata are read and clearly understood before using these data in order to protect both user and USGS interests. None. Users are advised to read the data set's metadata thoroughly to understand appropriate use and data limitations. USDA-NRCS Jeremy Maestas Ecologist Mailing and Physical

1201 NE Lloyd Blvd

Suite 801

Portland OR 97232 U.S. 503-273-2425 jeremy.maestas@usda.gov USGS, University of Montana, USDA-NRCS,USFWS, WGA Environment as of Metadata Creation: Microsoft [Unknown] Version 6.2 (Build 9200) ; Esri ArcGIS 10.6.1 (Build 9270) Service Pack N/A (Build N/A) The Rangeland Analysis Platform (RAP continuous land cover data were externally validated with three independent collections of field data from the Sagebrush Steppe Treatment Evaluation Project (SageSTEP; a collaborative Great Basin effort to evaluate sagebrush restoration), the Restore New Mexico Collaborative Monitoring Program initiative (RNMCMP; BLM and USDA-ARS Jornada Experimental Range collaborative effort to evaluate restoration treatments), and a collaborative project from the USDA-Agricultural Research Service and The Nature Conservancy collocated at the Eastern Oregon Agricultural Research Center (EOARC). Data were aggregated to percent cover per plot for the annual forbs and grasses, perennial forbs and grasses shrubs, and bare ground (bare ground measurements were only available for EOARC). Differences were calculated between plot measurements and the average predicted land cover values for all pixels that intersected the plot boundaries. Annual herbaceous cover was found to have a MAE and RMSE of 8.2 and 11.5 with respect to Sage STEP data, 7.5 and 14.8 with respect to RNMCMP data, and 7.3 and 10.0 with respect to EOARC data. The USGS-Harmonized Landsat and Sentinel-2 (USGS-HLS) driven product was developed using regression tree models calibrated and validated using Bureau of Land Management Assessment, Inventory, and Management field observations (2016-2019). For model development and accuracy reporting, the model database was split into 10 training and validation (80/20 split) sets to develop 5 regression tree models for the estimation of exotic annual grass (%) cover. Stratified random sampling was used during data splitting to ensure that the distribution of the response variable in each fold approximates the distribution of the entire dataset. Optimal model parameters (i.e. number of committee members and rules) were identified using the grid search method and cross validation procedures. Calibrated models were then applied to spatial inputs which resulted in five maps which were used to calculate median annual (2016 to 2019) values of exotic annual grass (%) cover estimates at each 30-m pixel. Comparisons with independent test data across each model iteration suggest fair agreement (average Pearson’s r = 0.67; Mean Absolute Error =11%) with observed exotic annual grass (%) cover conditions. The USGS-NLCD data have been validated with two robust approaches. First, an approach capitalizing on field data collected concurrently with high-resolution satellite (HRS) images over multiple locations (n = 42) and years (Rigge et al. 2020). HRS sites spanned a broad range of vegetation, biophysical, climatic, and disturbance regimes in Wyoming, Nevada, and Montana. Field observations were used to train regression tree models, predicting the component cover across each HRS image. We evaluated the spatial and temporal relationships between HRS and USGS-NLCD component cover and compare spatio-temporal climate responses. For each HRS site-year (n = 77) we averaged both the HRS and USGS-NLCD predictions within each site separately and regressed the averages to quantify the temporal accuracy. Next, we regressed individual pixel values of corresponding HRS and USGS-NLCD predictions to quantify the spatio-temporal accuracy. Results showed strong temporal correlations with an average R2 of 0.63 and Root Mean Square Error (RMSE) of 5.47% as well as strong spatio-temporal correlations with an average R2 of 0.52 and RMSE of 7.89% across components. The second major validation approach for USGS-NLCD data through comparison of results to data from two long term monitoring sites in southwest Wyoming (Shi et al. In Review). The field data included ten years of consistent observation at 126 plots over the period of 2006-2008. Field observations and Landsat times-series predictions generally responded similarly to interannual variation in weather, chiefly driven by precipitation. The spatial-temporal correlation across all 126 plots showed robust correlations for all components, with an R2 of 0.69 for bare ground and 0.40 for shrub cover and averaging 0.46 across components. Changes observed in both the field and USGS-NLCD data were typically gradual, within-state, changes which are most difficult to resolve, which were successfully captured. To develop a common map for the WGA cheatgrass strategy, we leveraged three published datasets that provide estimates of continuous percent cover of annual herbaceous plants across all, or large portions of, the study area: 1) Rangeland Analysis Platform ([RAP]; Jones et al. 2018), 2) USGS-Harmonized Landsat and Sentinel ([HLS]; Pastick et al. 2020, Pastick et al. in prep), and USGS-National Land Cover Database ([NLCD]; Rigge et al. 2020) (Fig. 1). Products were chosen because of their spatial and temporal coverage at 30-m resolution which facilitated development of a regional map using multiple predictions. Specific methodologies used to create each product vary but all combine remotely sensed data with field-based observations to predict fractional, per-pixel vegetation cover on western rangelands. Emphasis was placed on mapping rangelands in the region where invasive annual grasses, such as, cheatgrass, medusahead (Taeniatherum caput-medusae), and ventenata (Ventenata dubia) are most problematic and resulting in vegetation state shifts from shrublands to annual grasslands. While invasive annuals do exist on other land cover types (e.g., forests), the vegetation datasets we used were optimized primarily for arid rangelands. Estimates of annual herbaceous cover on higher productivity lands may be less useful as a surrogate for invasive annual grasses where the amount of native annuals is naturally higher. This fractional estimation of annual herbaceous cover is a weighted average across three data inputs. Data set is considered complete for the information presented, as described in the abstract. Users are advised to read the rest of the metadata record carefully for additional details. A formal accuracy assessment of the horizontal positional information in the data set has not been conducted. A formal accuracy assessment of the vertical positional information in the data set has either not been conducted, or is not applicable. Matt Jones et al. 20180101 Vector Digital Data Set University of Montana University of Montana https://rangelands.app/ Digital and/or Hardcopy Resources 20160101 20181231 ground condition RAP Source information used in support of the development of the data set. Matthew Rigge et al. 20191021 Raster Digital Data Set MRLC USGS/MRLC https://www.mrlc.gov/ Digital and/or Hardcopy Resources 20160101 20181231 ground condition NLCD-USGS Source information used in support of the development of the data set. Neal Pastick et al. 20200601 Vector Digital Data Set USGS Science Base USGS Science Base https://doi.org/10.5066/P9ZZSX5Q Digital and/or Hardcopy Resources 20160101 20181231 ground condition HLS Source information used in support of the development of the data set. Three datasets were used to calculate the annual herbaceous weighted mean. 1. The RAP dataset provides an estimation of percent cover of annual forbs and grasses at 30-m resolution across the western U.S. each year from 1984 to the present (Jones et al. 2018). Percent cover values are predicted using a Random Forests machine learning model that incorporates Landsat satellite data and a suite of geo-spatial land surface variables which is trained using over 27,000 field plots from the BLM-AIM and NRCS NRI field inventory databases. Complete access and further information on the data are available on the Rangeland Analysis Platform (https://rangelands.app) and a detailed description of the methods can be found in Jones et al. (2018). 2, The HLS dataset represents exotic annual grass percent cover (i.e. exotic annual Bromus spp. and medusahead) estimates made using Bureau of Land Management (BLM) Assessment, Inventory, and Monitoring (AIM) field data (i.e. TerrAdat, LMF; 2016-2019), weekly harmonized Landsat and Sentinel-2 composites and derived phenological metrics, other geospatial inputs (e.g. National Elevation Data [NED]), and Cubist regression tree models and machine learning techniques. Refer to Pastick et al. (2020) for a more detailed description of this mapping approach and inputs. 3. The NLCD dataset quantified the percent cover of annual herbaceous species across the western U.S. using 30-m Landsat imagery from 1985-2018. The dataset was trained using a circa 2016 fractional cover product developed with extensive ground measurements (Rigge et al. 2020). Time-series NLCD predictions were completed by applying spatio-temporal training data to regression tree modeling, change detection among years, and post-processing to ensure accurate post-burn trajectories, eliminate noise, and illogical change in the predictions (Rigge et al. 2019). The NLCD annual herbaceous component included all grasses and forbs whose life history is complete in one growing season. Refer to Rigge et al. (2019) and Rigge et al. (2020) for a more detailed description of this mapping approach and inputs. For the inter-comparison among the RAP, NLCD, and HLS annual herbaceous products, a consistent masking of products was applied to exclude non-rangeland areas. Pixels identified as non-rangeland by the USGS shrub mapping project (Rigge et al. 2020) were combined with those included in the rangeland extent product of Reeves and Mitchell (2011). The USGS non-rangeland mask included forested areas with greater than 40% tree canopy cover in the NLCD 2016 fractional tree canopy cover product and additional forested pixels identified using Normalized Difference Vegetation Index (NDVI) or Modified Soil-Adjusted Vegetation Index (MSAVI) thresholds (Rigge et al. 2020). Next, urban areas, major roads, snow, and ice were masked according to NLCD 2011 land cover classes. Third, cultivated crop and pasture/hay fields were masked using a combination of the 2013 Cropland Data Layer from the National Agricultural Statistics Service and the NLCD 2011 agricultural classes. Fourth, open water areas were identified using Normalized Difference Water Index (NDWI) thresholds. Supplemental hand edits were then applied where issues needed correction. The Reeves and Mitchell (2011) mask included data from LANDFIRE, Natural Resources Inventory (NRI), and Forest Inventory and Analysis. The NRI model used data on the identity, height, and cover of the dominant life form in each pixel to define pixels as rangeland/non rangeland. In species that can be considered both a shrub and tree, pixels with vegetation less than 5 m in height were considered shrub. Pixels with up to 25% tree canopy cover are included as rangeland. Pixels identified as not rangeland, agriculture, water, ice, snow, and urban were combined into a non-rangeland mask, following Jones et al. (2018). After combining the USGS and Reeves and Mitchell (2011) non-rangeland masks, playas and salt flats were captured by intersecting pixels with NLCD 2011 barren land cover and 0% topographic slope within the Northern, Central, and Mojave Basin and Range Level III ecoregions. These additional pixels were added to the combined non-rangeland mask. A per-pixel weighted averaging approach was used to combine 30-m mean (2016-2018) estimates of exotic annual grass and annual herbaceous (%) cover across available products (Jones et al. 2018, Pastick et al. 2020, Rigge et al. 2020, Pastick et al. in prep). An internal weighting scheme was used to determine per-pixel weight coefficients (wt) for each product (x), such that the sum of weight factors equals 1. The approach penalizes (give less weight) to an input data has an annual herbaceous cover estimate that is a statistical outlier relative to the other two datasets in a pixel. Determination of outlier status is based on absolute difference of each product from the unweighted mean of all layers raised to the fourth power to penalize for outliers. If only two datasets are available a simple average is taken of the inputs. We opted to use an internal weighting scheme, as opposed to determining weights externally (e.g. comparing annual estimates to field observations), for the sake of simplicity and time constraints. Finally, if only one dataset is available, that dataset is used directly. 20200601 Raster Grid Cell 54583 73678 1 0.000269494585236 0.000269494585236 Decimal seconds D_WGS_1984 WGS_1984 6378137.0 298.257223563 Attribute Table Table containing attribute information associated with the data set. Producer defined Value Unknown Producer defined 0 81 Percent cover The entity and attribute information provided here describes the tabular data associated with the data set. Please review the detailed descriptions that are provided (the individual attribute descriptions) for information on the values that appear as fields/table entries of the data set. The entity and attribute information was generated by the individual and/or agency identified as the originator of the data set. Please review the rest of the metadata record for additional details and information. USGS Science Base Mailing and Physical

Denver Federal Center, Building 810

Mail Stop 302

Denver CO 80225 U.S. 1-888-275-8747 sciencebase@usgs.gov Unless otherwise stated, all data, metadata and related materials are considered to satisfy the quality standards relative to the purpose for which the data were collected. Although these data and associated metadata have been reviewed for accuracy and completeness and approved for release by the U.S. Geological Survey (USGS), no warranty expressed or implied is made regarding the display or utility of the data for other purposes, nor on all computer systems, nor shall the act of distribution constitute any such warranty. Although several data inputs have been processed successfully on a computer system at the USGS, no warranty expressed or implied is made by the USGS regarding the use of the data on any other system, nor does the act of distribution constitute any such warranty. Data may have been compiled from various outside sources. Spatial information may not meet National Map Accuracy Standards. This information may be updated without notification. The USGS shall not be liable for any activity involving these data, installation, fitness of the data for a particular purpose, its use, or analyses results Raster Digital Data Set https://rangelands.app/cheatgrass/ None. No fees are applicable for obtaining the data set. 20200818 Matthew B Rigge (CTR) U.S. Geological Survey, LAND RESOURCES mailing address

47914 252nd Street

Sioux Falls SD 57198 US 605-594-2894 mrigge@contractor.usgs.gov FGDC Content Standard for Digital Geospatial Metadata FGDC-STD-001-1998