Changelog

Updated the getting started article to not include a degree symbol in the Daymet graph, which caused compilation errors in Windows. Fixes Issue #106.
Updated get_nlcd() to include the 2021 NLCD as the default, in response to Issue #105.
Updated outdated package description
Bumped GDAL version req to >= 3.1.0 to accommodate storing spatial vectors as FlatGeoBufs
Added arcgislayers dependency and retired self-written esri functions. Closes Issue #109.

Updated the README and moved examples to an article
Added the PAD-US dataset in response to Issue #100.
Updated get_nlcd() to fix bug Issue #101 in tree canopy data, whose URL changed.
Updated get_ghcn_daily() to respond to slightly different file formatting on https server Issue #99.
Removed dependencies on sp and raster. All functions now return either sf or terra objects.

Updated get_ghcn_daily() and related functions to use https addresses instead of the legacy ftp server Issue #99.
Updated get_nhd() to gracefully handle situations where no point data (or any NHD data) are available within an area of interest, using solution offered in Issue #98.
Updated get_nhd() to handle some strange geometries present in the NHD Issue #98.

Updated get_nhd() to correctly access the extraction.dir if it already exists Issue #95. get_nhd() and get_ned() now both use file.path() in constructing file outputs.
Updated get_daymet() to use terra for writing raster data, in response to Issue #96. All functions in FedData now use terra for writing.
Cleaned up documentation.

Removed rgdal dependency throughout and instead require raster >= 3.6.3.
Switched back to using the Web Coverave Service for the NLCD, as it is now providing data in the native CRS.

Updated get_nlcd() to read CRS data correctly Issue #91 by requiring rgdal be installed.
Changed Daymet tempo codes Issue #92

All get_*() functions now return sf or raster objects.
Changed SDA_query() to soils_query() to avoid namespace masking with SoilDB.
Added get_nass_cdl() to retrieve the NASS Cropland Data Layer
Updated get_daymet() to pull from ORNL WCS, and fixed bug in Issue #49
Fixed issue where soils_query() was only returning first SSURGO study area (Issue #51)
Fixed issue where date parsing was USA-specific by leveraging lubridate::parse_date_time (Issue #61)
get_ssurgo() now saves in the GeoPackage file format
get_nhd() now access ESRI web services. Users can optionally get data from NHDPlus.
get_nlcd() now provides data in native CRS (CONUS Albers), rather than web-mercator (Issue #77) by pulling from self-hosted Cloud-Optimized GeoTiffs.
Issue #88 identified an issue with GDAL versions prior GDAL 3 reading COGs using the vsicurl functionality in GDAL. Accordingly, FedData3 requires GDAL >= 3.0.0.
Added Mesa Verde National Park as exemplar region, and removed use of paleocar package.
get_ned() now pulls from USGS NED Cloud-Optimized GeoTiffs available at https://prd-tnm.s3.amazonaws.com/index.html?prefix=StagedProducts/Elevation/.

Removing many internet resource tests from CRAN, to satisfy: ‘Packages which use Internet resources should fail gracefully with an informative message if the resource is not available (and not give a check warning nor error).’

Built-in access to the Soils Data Analysis query service to remove dependency on soilDB package.

Fixed issue (#41) that occurs when mosaicking NLCD tiles that are not cropped. When they aren’t cropped, the NLCD data is never read into memory, and the temporary file that the raster was created from gets destroyed. Solution: Force NLCD data into memory prior to mosaicking.
Added (non-CRAN) test for issue #41

SSURGO fixed test where supplying an unavailable survey area now returns NULL instead of an error.
SSURGO zip directory encoding changes as of late October 2017 forced changes in the FedData:::get_ssurgo_study_area function.
Fixed issue where NHD template wouldn’t load because they added a jpeg preview to the directory.

DAYMET functions now do not operate in parallel. This was breaking the download functions.
Final update for version 2 of FedData.
Accepted to ROpenSci! Migrating to the ROpenSci organization on GitHub.

writeOGR for SSURGO and NHD were failing on Windows when the extraction.dir included a trailing slash. Paths are now normalized to remove the trailing slash.

Updated the get_ned function to provide more useful errors and warnings when downloads are unsuccessful.

Added pkgdown site.
SSURGO functions (e.g., get_ssurgo) now doesn’t bomb on large (> 1 billion sq meter) requests. Now, the area of interest is broken into smaller chunks to build the download list.

FedData (development version)