{ccc7c413-9908-4e20-83e1-acf01b05de38} eng; USA utf8 dataset EnviroAtlas Coordinator U.S. Environmental Protection Agency, Office of Research and Development-Sustainable and Healthy Communities Research Program, EnviroAtlas Geospatial Data Owner (919) 541-3832 109 T.W. Alexander Drive Research Triangle Park NC 27709 EnviroAtlas@epa.gov https://www.epa.gov/enviroatlas pointOfContact 2025-02-13 ISO 19115 Geographic Information - Metadata ISO 19115:2003/Cor.1:2006(E) EnviroAtlas -- Los Angeles County, CA -- Meter-Scale Urban Land Cover (MULC) Data (2016) 2019-09-06 publication U.S. EPA Office of Research and Development (ORD) - National Exposure Research Laboratory (NERL) Research Triangle Park NC publisher US Environmental Protection Agency (US EPA) (additional partners are listed in Supplemental) originator The Los Angeles County, CA Meter Urban Land Cover (MULC) dataset was generated from sub-meter image pixel resolution data created by the University of Vermont Spatial Analysis Laboratory (SAL) through the combined use of 2016 USDA National Agricultural Imagery Program (NAIP) four band (red, green, blue and near infrared) aerial imagery, 2016 LiDAR data and 2014 ortho-imagery. The sub-meter, thematic landcover data feature attributes were recoded and spatially resampled to a 1-meter spatial scale by EPA for MULC data product integration. The mapped area is confined to the boundaries of US Census Bureau's 2010 Urban Statistical Area for Los Angeles County, with an added 1km coastal water buffer area extension. The following eight land cover classes were mapped: Water, Impervious, Soil or Barren, Trees or Forest, Grass or Herbaceous, Woody Wetlands and Emergent Wetlands. Integrated wetland features were derived using ancillary National Wetlands Inventory (NWI) polygon data (version 2), downloaded from the Unites States Fish and Wildlife Service (USFWS) Wetland Mapper web mapping service (https://www.fws.gov/wetlands/data/mapper.html). Metadata for the NWI wetlands data layer can be found at http://www.fws.gov/wetlands/Data/Metadata.html. An accuracy assessment of the classified product, using 887 completely random and 40 stratified random photo-interpreted land cover reference sample points yielded an overall user's accuracy (MAX) of 61.1 percent and a fuzzy user's accuracy (RIGHT) of 89.2 percent. For data workflow processing details see Overview Description section. This dataset was produced by the University of Vermont Spatial Analysis Laboratory, the United States Forest Service Urban Tree Canopy (UTC) assessment program, and the US EPA to support research and online mapping activities related to the EnviroAtlas. EnviroAtlas (https://www.epa.gov/enviroatlas) allows the user to interact with a web-based, easy-to-use, mapping application to view and analyze multiple ecosystem services for the contiguous United States. The dataset is available as downloadable data (https://edg.epa.gov/data/Public/ORD/EnviroAtlas) or as an EnviroAtlas map service. Additional descriptive information about each attribute in this dataset can be found in its associated EnviroAtlas Fact Sheet (https://www.epa.gov/enviroatlas/enviroatlas-fact-sheets). The Meter-scale Urban Land Cover (MULC) map product serves as a spatially-detailed basemap from which multiple human and ecosystem health metrics and services are calculated using the EPA EnviroAtlas tool. This MULC data product was created collaboratively as part of an interagency agreement between the EPA and the University of Vermont Spatial Analysis Laboratory (SAL). The overall goal of EnviroAtlas is to employ and develop the best available science to map indicators of ecosystem services production, demand, and drivers for the nation. completed EnviroAtlas Coordinator U.S. Environmental Protection Agency, Office of Research and Development-Sustainable and Healthy Communities Research Program, EnviroAtlas Geospatial Data Owner (919) 541-3832 109 T.W. Alexander Drive Research Triangle Park NC 27709 EnviroAtlas@epa.gov https://www.epa.gov/enviroatlas pointOfContact asNeeded Ecology Ecosystem Environment Exposure Health Impact Land Natural Resources Recreation Water theme EPA GIS Keyword Thesaurus Ecosystem Services EnviroAtlas Environment Environmental Atlas theme EnviroAtlas California Los Angeles place None https://edg.epa.gov/EPA_Data_License.html otherRestrictions Use Constraints: None. Please check sources, scale, accuracy, currency and other available information. Please confirm that you are using the most recent copy of both data and metadata. Acknowledgement of the EPA would be appreciated. otherRestrictions Access Constraints: None PUBLIC https://project-open-data.github.io/schema/#accessLevel Standard Technical Controls eng; USA environment geoscientificInformation imageryBaseMapsEarthCover -118.983754 -117.630470 33.225470 34.836351 2014-01-01 2016-08-31 These data were created by University of Vermont Spatial Analysis Laboratory and the United States Forest Service Urban Tree Canopy (UTC) assessment program as part of an interagency agreement between the US Environmental Protection Agency (EPA) and the US Forest Service. Corrections and modifications were done by the EPA. true Entity and Attribute Information Entity and Attribute Overview: The Los Angeles County, CA EnviroAtlas Meter-scale Urban Land Cover (MULC) dataset comprises approximately 11,336.10 km2 of mapped area, including the city of Los Angeles and its surrounding vicinity. The dataset boundary is defined by both the Los Angeles county boundary and the US Census Bureau's 2010 Urban Statistical Area for Los Angeles County, with an added 1-km coastal water buffer extension surrounding the mainland area and an approximate 6.5-km buffer extension surrounding Santa Catalina Island. The San Clemente Island has not been included as a mapped feature in the MULC dataset. Of this total mapped area, land and inland water features comprise nearly 92.02% of the mapped area, covering a 10,431.12 km2 area. The remaining coastal waters cover the remainder 7.98% of the total mapped area, with an estimated expanse of 904.98 km2. This EPA-modified dataset was derived using a sub-meter scale, base land cover dataset for Los Angeles County, CA that was developed by the University of Vermont Spatial Analysis Laboratory (SAL) under the direction of Jarlath O'Neil-Dunne as part of the United States Forest Service Urban Tree Canopy (UTC) assessment program. The sub-meter map product mapped a total of eight landcover classes: Tree Canopy, Grass/Shrub, Bare Earth, Water, Buildings, Roads, Other Paved Surfaces and Tall Shrubs. Updates to the data were completed by EPA in 2017 to correct for errors or data voids in the original dataset and modified, as necessary, to meet MULC mapping product standards. The eight landcover classes mapped by SAL were condensed by EPA into six baseline landcover class categories as follows: Water, Impervious, Soil/Barren, Tree/Forest, Shrub, Grass/Herbaceous. Corrections and modifications to the data were implemented using these six landcover classes prior to any additional landcover class category modifications. Agriculture, generally included as a MULC landcover feature class category, was excluded for the Los Angeles County area dataset because it did not meet the minimum area mapping threshold (5% of total mapped land area), as defined by EPA for feature class category inclusion. The total areal extent for cropland agriculture was estimated using combined data sources, including California 2014 Parcels Tax Roll data (available for download at https://egis3.lacounty.gov/dataportal/) and the U.S. Department of Agriculture National Agriculture Statistics Service (NASS) Cropland Data Layer (CDL) for years 2012 through 2017 (available for download at https://www.nass.usda.gov/Research_and_Science/Cropland/Release/index.php) . Total estimated agricultural land area corresponded to less than 1% of the total MULC mapped boundary area and less than 2% of the total mapped non-water land areas. Wetland landcover class categories (Woody Wetland and Emergent Wetland) were added to the map, subsequent all initial base map feature class modifications and corrections. Section 1. University of Vermont SAL Methods and Accuracy Assessment SAL Landcover Processing Steps: The landcover classification created by the University of Vermont Spatial Analysis Laboratory (SAL) was developed as part of the Urban Tree Canopy (UTC) Assessment for Los Angeles County, CA. Consequently, it represents a 'top down' mapping perspective in which tree canopy over hanging other features is assigned to the tree canopy class. SAL implemented the use of object-based image analysis (OBIA) methods (O'Neil-Dunne, MacFaden, and Royar, 2014) to derive its mapped land cover classes. OBIA systems work by grouping pixels into meaningful objects based on their spectral and spatial properties, while considering boundaries imposed by existing vector datasets. A minimum mapping unit of 9 square feet was defined for use in the determination of delineated boundaries for mapped features. Seven land cover classes were mapped: (1) Tree Canopy, (2) Grass/Shrub, (3) Bare Earth, (4) Water, (5) Buildings, (6) Roads, (7) Other Paved Surfaces, and (8) Tall Shrubs. A LiDAR-derived normalized Digital Surface Model (nDSM) height threshold of 2m was used to differentiate tall feature objects (>2m), including trees and tall shrubs, from short feature objects (< 2m). Visible RGB and NIR bands were used to further differentiate trees from tall shrubs. Feature object height, geometric morphology and LiDAR texture were used to differentiate tall impervious objects. Visible RGB and NIR bands, along with other ancillary data sources, were used to classify feature objects in the < 2m height threshold criteria into Grass/Shrub, Bare Earth, Water, Roads and Other Paved Surfaces land cover categories. The primary sources used to derive the mapped landcover classes were 2016 National Agricultural Imagery Program (NAIP) color infrared imagery, 2016 LiDAR data and 2014 ortho photography, in combination with ancillary data sources provided by Los Angeles County and those created by SAL. Morphological procedures were implemented as a final processing step to ensure data accuracy and to create a cartographically-pleasing map product. Upon completion, no accuracy assessment was conducted; however, a detailed manual review of the dataset was carried out at a scale of 1:2,000 and all observable errors were corrected. The SAL sub-meter landcover data is considered current as of summer 2016. Section 2. EPA Methods and Accuracy Assessment EPA Landcover Corrections, Additions and Modifications: Updates to the SAL data product were completed by EPA in 2017 to correct for classification errors or data voids in the original dataset and modified, as necessary, to meet MULC mapping product standards. All data were processed and updated in their original, distributed map projection and mapping units (EPSG:2229) NAD_1983_StatePlane_California_V_FIPS_0405_Feet and subsequently reprojected to (EPSG:26911) NAD_1983_UTM_Zone_11N in meters using a nearest neighbor resampling option for online publication and distribution. To meet MULC mapping product standards, the SAL sub-meter spatial resolution landcover data was resampled to 1-meter spatial resolution using a nearest-neighbor resampling option and SAL landcover classes were recoded from eight classes to six. The eight landcover classes defined within the SAL landcover were condensed by EPA into the following six baseline landcover class categories: Water, Impervious, Soil/Barren, Tree/Forest, Shrub, Grass/Herbaceous. The resampled data were visually assessed for data voids (data pixels that lack, but require, a landcover classification code) and classification errors due to data artifacts (e.g., misclassified data pixels due to imagery seamline errors or inherent data noise resulting from data acquisition equipment). Identified data voids or artifacts that did not exceed 1 row or 1 column in width, such as misclassified pixels resulting from imagery seamline feature artifacts, were removed using a 3x3 mode filter, followed by a 3x3 max filter to fill in scene edge pixels. Those data voids or artifacts that exceeded the aforementioned width criteria were corrected using the iterative, unsupervised classification technique described below. EPA corrected, modified and quality-assessed the Los Angeles County, CA land cover dataset using United States Department of Agriculture (USDA) National Agriculture Imagery Program (NAIP) Digital Orthophoto Quarter Quadrangle (DOQQ) ("quarter quads"). A total of 334 spatially-correspondent DOQQ imagery tiles pertaining to the year 2016 were used within the Google Earth Engine (GEE) cloud-computing code-editor platform to assist in data additions, corrections and modifications. Missing coastal features (e.g., marina infrastructure, roadway bridges or overpasses, wave breakers, and piers) and solar farm infrastructure were mapped using iterative, unsupervised cluster Learning Vector Quantization (weka-LVQ) algorithms in GEE and map feature coverage was extended to include areas within a 1km delimiting buffer boundary for coastal waters. Further information about Google's code editor may be found at: https://developers.google.com/earth-engine/. Further information pertaining to Google's GEE implementation of the LVQ algorithm can be found at the following referenced source: T. Kohonen, "Learning Vector Quantization", The Handbook of Brain Theory and Neural Networks, 2nd Edition, MIT Press, 2003, pp. 631-634. Results from the weka-LVQ classifications were analyzed iteratively, extracting non-water classified features of interest by unsupervised cluster group grid value and recoding the selected features to binary data value 1. The binary data values were then subjected to a pixel connectivity analysis in order to identify connected pixel groups by their assigned unique group-id value. Only group ids pertaining to features that were directly connected, by one or more pixels, to the inland land area and that met a minimum pixel count threshold of 10 were selected and recoded to correspond with a mapped land cover class category and added to the finalized land cover map. This post-unsupervised classification data filtering step resulted in the removal of most disconnected, undesired floating features (e.g., undocked boats, sun glint, floating vegetation, etc.) from inclusion in the finalized data. Additional manual digitization was used to remove residual undesired feature objects from water bodies (e.g., boats, ships, overhanging cranes, sunglint, etc.) where pixel connectivity analyses removal was not deemed feasible or effective, or where specific connected features (e.g., large docked commercial ships) were to be excluded from the final land cover. The delineation of boundaries for added coastal features were aided using 1-meter resolution digital elevation model (DEM) data available from the U.S. Geological Survey web viewer at https://viewer.nationalmap.gov/basic/. A total of 310 DEM data tiles for the Los Angeles County area were mosaicked and data grid values that exceeded a 0.5-meter elevation minimum threshold were recoded to a binary value of 1 to delineate inland area boundaries and to facilitate in the identification and the delineation of areas where coastal feature land cover classifications were incomplete or incorrect. A total approximated area of 902 Sq. Kilometers of unclassified map extent areas were classified to coastal or inland water landcover features and an additional approximated area of 1.78 Sq. Kilometers were added to other feature class categories. A total approximated area of 6.79 Sq. Kilometers of existing landcover features were modified or recategorized from their original baseline landcover class category to a new land cover class category. These added or modified features are not meant to be comprehensive in spatial coverage. Impervious features pertaining to solar farm infrastructure were mapped primarily for areas that were readily discernible due to their overall spatial extent. Coastal infrastructure features were mapped according to their current spatial location and extent in 2016 NAIP imagery. A known caveat is that marina features are dynamically-managed systems with movable and configurable parts and hence their spatial location may change at any given time. Wetland class categories, Woody Wetland and Emergent Wetland, were added to the landcover as a final step, subsequent all previous corrections, modifications and QA assessments to the base classified landcover class categories. Integrated wetland features were derived using ancillary National Wetlands Inventory (NWI) polygon data (version 2), downloaded from the Unites States Fish and Wildlife Service (USFWS) Wetland Mapper web mapping service (https://www.fws.gov/wetlands/data/mapper.html). Metadata for the NWI wetlands data layer can be found at http://www.fws.gov/wetlands/Data/Metadata.html. NWI vector polygons were dissolved by their feature attribute "ATTR" and converted to raster features, mapped at a 1-meter spatial resolution. The MULC base landcover class categories were cross-referenced with the NWI rasterized feature attributes at the pixel-to-pixel level to derive a unique cross-referencing code used to determine each pixel's appropriate pixel-level landcover class re-assignment. Tree or Forest and Shrub landcover class feature pixels located within NWI wetland areas were re-assigned to Woody Wetland landcover class. Grass or Herbaceous landcover class feature pixels located within NWI wetland areas were re-assigned to Emergent Wetland landcover class. Additionally, Soil or Barren landcover class feature pixels located within NWI wetland areas were re-assigned to Emergent Wetland landcover class, under the assumption that the soil substrate could theoretically support emergent vegetation with mutable hydrological conditions. All delineated wetland feature area pixels that intersected the Impervious landcover class category were re-assigned to the Impervious class, as these areas are generally considered to be highly compacted and impermeable surfaces. NWI wetland delineated boundaries were not modified by the EPA and therefore no guarantee regarding the accuracy of wetland feature spatial location or extent is expressly given. Please refer to the NWI metadata to determine the potential limitations that any temporal and/or spatial inaccuracies in the NWI dataset may pose for your particular application (http://www.fws.gov/wetlands/Data/Metadata.html). QA Accuracy Assessment: A shapefile of 2400 stratified random point samples (spaced at a minimum 50 meters) was created for the Los Angeles County area, distributed equally between urban and non-urban areas (1200 Urban + 1200 non-urban = 2400 total # sampling points). Urban versus non-urban area boundaries were derived from the City Boundaries data for Los Angeles County available for download at (https://egis3.lacounty.gov/dataportal/). A random number generator (https://www.random.org/sequences/) was used to create a sequence of 2400 random numbers and these numbers were added as a new attribute column within the sample point shapefile attribute table. Sample points were then ordered by their random number in ascending order and the first 1000 points in the table were selected for use in a QA assessment of the six-class landcover map. Wetland features are incorporated into the dataset once all base landcover types have been completed and assessed and are not included the accuracy assessment. A visual interpretation of landcover was performed for the first 1000 selected validation points using 2016 NAIP ortho imagery. Visual interpretations also incorporated the use of image viewers such as Google Satellite and Street Views and Bing Aerial and Birdseye views, as needed, to substantiate the landcover class interpretation. A scaled fuzzy validation score, ranging from whole number values 1 to 5 was assigned to each point to describe its user-interpreted degree of correspondence to each of the six baseline land cover classes. The fuzzy classification scheme used is defined by Gopal and Woodcock (1994). The scaled fuzzy QA scores are interpreted as follows: 1: Absolutely wrong: classification value was unacceptable (Very Wrong); 2: Understandable but Wrong: classification value was not good. There was something about the site that made the answer understandable, but there was clearly a better answer. Classification would pose a problem for users of the map. (Not Right); 3: Reasonable or Acceptable: Maybe not the best possible classification but it was acceptable; the classification did not pose a problem to users of the map. (Right); 4: Good Answer: Would be happy to find this classification given on the map (Very Right); 5: Absolutely Right: No doubt about the match. (Perfect) After the initial assessment, in situations where a landcover class was not represented by at least 50 QA sample points, additional stratified random points were added to reach the minimum 50-sample point representation minimum. These additional points were interpreted and included in the final accuracy assessment. For the Los Angeles County study area, an additional 50 stratified random points were added to the 1000 original points to meet the 50-point minimum requirement for the Grass/Herbaceous landcover. QA points located in shadow or feature edge areas were discarded from interpretation due to the lack of a readily discernible and prominent landcover class feature. Of the total 1050 QA sample points, a total of 123 were discarded due to shadow and edge locations, leaving a total of 927 valid, interpreted QA sample points. The accuracy assessment produced two confusion (error) matrices. The "MAX" view is more conservative, and the "RIGHT" view is more liberal. The MAX interpretation is considered correct if the classified land cover coincides with the user's interpreted highest assigned score, illustrating the interpreter finds this class to be the most appropriate for that location. The RIGHT interpretation is considered correct if the classified land cover matches any landcover class to which the interpreter has assigned a score of 3 or greater, illustrating that the interpreter finds the classification to be acceptable, although another class may be more appropriate. While other EnviroAtlas MULC mapped product QA accuracy assessment confusion matrices have been calculated using a minimum user-interpreted score of 3 or greater for correct/acceptable interpreted landcover class features, this dataset has been assessed using a user-modified minimum interpreted valid score value of 2 or greater for the reasons explained below. Generally, the availability of LiDAR height-above-ground data to assist with the visual interpretation of features in reference imagery is a particularly useful aid for both the visual discernment and the feature class determination of vegetation classes and renders the default QA scoring system to be adequate. The base landcover map for Los Angeles County created by SAL incorporated the use of 2016 LiDAR-derived height-above-ground data to assist in the classification of vegetation. However, the visual interpretation for this QA was performed without the aid of LiDAR height-above-ground data to assist with the visual interpretation of features and to aid in the discernment of vegetation classes. Consequently, a modification was introduced to the original confusion matrix computing code to allow for leniency/ flexibility in the QA for the RIGHT interpretation of vegetation class designations. The default minimum user-interpreted score of 3 or greater for QA confusion matrix calculations was modified and reduced to include user-interpreted scores of 2 or greater. User-interpreted assigned scores of 2 o greater could qualify as "correct" or "acceptable" alternative landcover classes to account for vegetation feature interpretations that could reasonably belong to any of 3 landcover classes where height-above-ground determination was not available to make a confident vegetation class designation. The modification was introduced to also accommodate for the fact that the unmodified SAL dataset under-represents the Grass/Herbaceous landcover class for the Los Angeles County area, attributing a significant portion of natural dry grassland landscape to the Barren/ Soil landcover class. As these extensive areas are subject to seasonal variation that may strongly influence vegetation detection in the NIR-band and cannot be readily differentiated, even with the additional aid of LiDAR, EPA has considered the Soil/Barren landcover classification as a reasonable alternative class to Grass/ Herbaceous for QA purposes wherever a QA sample point has been determined to intersect dry grassland features. The following confusion matrices summarize the MAX and RIGHT accuracy assessment results, respectively. The following confusion matrix summarizes the MAX accuracy assessment data for the EPA study area. These matrices can be copied and pasted from ArcCatalog or a text editor into Excel for readability: Grass_Herb Impervious Shrub SoilBarren TreeForest Water Row_Total User's_Accuracy Grass_Herb 57 16 4 2 5 0 84 0.678571 Impervious 3 229 0 10 3 0 245 0.934694 Shrub 19 0 57 2 2 0 80 0.7125 SoilBarren 212 2 27 56 2 0 299 0.187291 TreeForest 20 8 21 3 103 0 155 0.664516 Water 0 0 0 0 0 64 64 1 Column_Total 311 255 109 73 115 64 927 nan Producer's_Accuracy 0.18328 0.898039 0.522936 0.767123 0.895652 1 nan nan Overall_Accuracy 0.610572 K_Hat 0.534088 K_Variace 0.000354 The following confusion matrix summarizes the RIGHT accuracy assessment data for the EPA study area: Grass_Herb Impervious Shrub SoilBarren TreeForest Water Row_Total User's_Accuracy Grass_Herb 63 16 1 1 3 0 84 0.75 Impervious 3 239 0 1 2 0 245 0.97551 Shrub 10 0 68 2 0 0 80 0.85 SoilBarren 6 2 14 275 2 0 299 0.919732 TreeForest 20 8 6 3 118 0 155 0.76129 Water 0 0 0 0 0 64 64 1 Column_Total 102 265 89 282 125 64 927 nan Producer's_Accuracy 0.617647 0.901887 0.764045 0.975177 0.944 1 nan nan Overall_Accuracy 0.892125 K_Hat 0.861833 K_Variance 0.000166 References: Gopal, S. and Woodcock, C. (1994). Theory and Methods for Accuracy Assessment of Thematic Maps Using Fuzzy Sets. Photogrammetric Engineering and Remote Sensing 60(2), 181-188. O'Neil-Dunne J, MacFaden S, Royar A. (2014). A Versatile, Production-Oriented Approach to High-Resolution Tree-Canopy Mapping in Urban and Suburban Landscapes Using GEOBIA and Data Fusion. Remote Sens. 2014, 6, 12837-12865. http://www.mdpi.com/2072-4292/6/12/12837. U.S. Department of Agriculture. (2016). National Agriculture Imagery Program imagery. Farm Service Agency. Aerial Photography Field Office: U.S. Department of Agriculture Web page, https://www.fsa.usda.gov/programs-and-services/aerial-photography/imagery-programs/naip-imagery/index. U.S. Fish and Wildlife Service. (2006 (Imagery date)). California National Wetlands Inventory digital data. Accessed September 2018. http://wetlands.fws.gov/. Entity and Attribute Detail Citation: https://www.epa.gov/enviroatlas/enviroatlas-fact-sheets EnviroAtlas Coordinator U.S. Environmental Protection Agency, Office of Research and Development-Sustainable and Healthy Communities Research Program, EnviroAtlas Geospatial Data Owner (919) 541-3832 109 T.W. Alexander Drive Research Triangle Park NC 27709 EnviroAtlas@epa.gov https://www.epa.gov/enviroatlas distributor https://enviroatlas.epa.gov/download/LACA_MULC_tif.zip download Download The URL providing direct access to the downloadable dataset. download https://enviroatlas.epa.gov/arcgis/rest/services/Supplemental/Landcover_AllCommunities/ImageServer ESRI:ArcGIS API The endpoint of a web service to access the dataset (REST endpoint, WMS GetCapabilities URL, or a SOAP WSDL endpoint). webMapService https://www.epa.gov/enviroatlas/enviroatlas-data information Documentation URL for accessing related documents such as data dictionary, technical information about a dataset, developer documentation, etc. information https://www.epa.gov/enviroatlas information Home Page Alternative landing page used to redirect user to a contextual, Agency-hosted "homepage" for the Dataset or API information dataset Horizontal Positional Accuracy Report Data were collected using methods that have unknown accuracy (EPA National Geospatial Data Policy [NGDP] Accuracy Tier 10). For more information, please see EPA's NGDP at http://epa.gov/geospatial/policies.html Features represented have not been tested for completeness Features represented have not been tested for completeness Tests for integrity have not been performed. This lineage was generated by the FGDC CSDGM to 19139 transformation See Overview Description for detailed data processing workflow 2019-07-01T00:00:00 unknown 2029-02-13 This metadata was automatically generated from the FGDC Content Standard for Digital Geospatial Metadata standard (version FGDC-STD-001-1998) using the 2024-02-09T11:11:00 version of the FGDC CSDGM to ISO 19139 transform. EnviroAtlas Coordinator U.S. Environmental Protection Agency, Office of Research and Development-Sustainable and Healthy Communities Research Program, EnviroAtlas Geospatial Data Owner (919) 541-3832 109 T.W. Alexander Drive Research Triangle Park NC 27709 EnviroAtlas@epa.gov https://www.epa.gov/enviroatlas pointOfContact