                 Changes in version 2.0.4 (2026-09-09)                  

The GitHub repository was renamed from nhdplusTools to hydrogeofetch,
which retired the old GitHub Pages site. The sample data files that
examples and tests download were still requested from the old address
and returned 404. All package URLs now point at
doi-usgs.github.io/hydrogeofetch and github.com/DOI-USGS/hydrogeofetch.

                 Changes in version 2.0.3 (2026-08-28)                  

get_drainage_area_estimates() now warns and returns NULL, rather than
stopping, when the NHDPlusV2 OGC API returns no flowline attributes for
the upstream network.

                 Changes in version 2.0.2 (2026-08-28)                  

Improve coordinate precision test file matching to be more robust.

                 Changes in version 2.0.1 (2026-08-21)                  

Bug fixes

  - Bounding box coordinates in OGC API Features requests are now
    rounded to six decimal places (about 0.1 m) instead of being sent at
    full floating point precision. Those coordinates come out of a
    projection round trip, and their last digits differ between PROJ
    builds, which made the request URL -- and the recorded test fixtures
    keyed on it -- platform dependent. Rounding is outward, so the
    queried area never shrinks.
  - Point coordinates that reach a request the same way -- through
    discover_nhdplus_id() and the processing-service calls behind
    get_raindrop_trace(), get_split_catchment(), get_xs_point(), and
    get_xs_points() -- are rounded to seven decimal places (about a
    centimeter) for the same reason.

                 Changes in version 2.0.0 (2026-08-20)                  

hydrogeofetch is the renamed successor to nhdplusTools. The package
scope has grown beyond NHDPlus, and the new name reflects its broader
role as a hydrologic geospatial data access and network tool chain.
nhdplusTools will remain on CRAN as a deprecation shim until
October 2028.

New features

  - New get_3dhp_index() addresses (snaps) point features to the 3DHP
    network using the HydroAdd3D web service, returning the snapped
    location and elevation, the matched mainstem (as a geoconnex uri),
    the measure along the mainstem, and associated GNIS attributes.
  - New add_mainstems() joins geoconnex mainstem identifiers onto a
    table containing NHDPlusV2 (comid/featureid) or NHDPlusHR
    (nhdplusid) identifiers, using a cached lookup table from the
    ref_rivers release. #469
  - New check_mainstems() checks whether mainstem identifiers have been
    superseded by a newer reference release; new update_mainstems()
    re-indexes points with a known mainstem id onto its replacement
    using hydroloom::index_points_to_lines(). #470

Removed functions (moved to hydroloom)

Network analysis and spatial indexing functions that duplicated
hydroloom have been removed. Users should call hydroloom directly. See
the hydroloom reference for full documentation, especially the indexing
and navigation sections.

  - get_flowline_index() → hydroloom::index_points_to_lines()
  - get_waterbody_index() → hydroloom::index_points_to_waterbodies()
  - disambiguate_flowline_indexes() → hydroloom::disambiguate_indexes()
  - get_UT(), get_DD() → hydroloom::navigate_hydro_network(x, start,
    "UT") / "DD"
  - get_UM(), get_DM() → hydroloom::navigate_hydro_network(x, start,
    "UM") / "DM" (note: the old wrappers also ran align_nhdplus_names,
    filtered, and supported sort/include arguments)
  - navigate_network() — workflow glue; use navigate_nldi() +
    subset_nhdplus() or hydroloom navigation directly
  - get_sorted() → hydroloom::sort_network()
  - get_pathlength() → hydroloom::add_pathlength()
  - get_streamorder() → hydroloom::add_streamorder()
  - get_streamlevel() → hydroloom::add_streamlevel()
  - get_levelpaths() → hydroloom::add_levelpaths()
  - get_pfaf() → hydroloom::add_pfafstetter()
  - get_terminal() → hydroloom::sort_network(x, split = TRUE)
  - calculate_total_drainage_area() →
    hydroloom::accumulate_downstream(x, "area")
  - calculate_arbolate_sum() → hydroloom::accumulate_downstream(x,
    "length")
  - get_path_members(), get_path_lengths() →
    hydroloom::navigate_connected_paths()
  - make_node_topology() → hydroloom::make_node_topology()
  - add_plus_network_attributes() → use hydroloom's add_levelpaths(),
    add_pathlength(), accumulate_downstream() etc. directly

Removed re-exports

These hydroloom functions are no longer re-exported; call
hydroloom::fn() directly:

  - st_compatibalize(), rename_geometry(), get_node(), fix_flowdir(),
    rescale_measures(), get_hydro_location(), get_partial_length()

Caching changes

  - The memoise response cache now defaults to memory rather than
    filesystem. Cached web service responses live only as long as the R
    session unless you opt back in with
    hydrogeofetch_cache_settings(mode = "filesystem") or
    HYDROGEOFETCH_MEMOISE_CACHE=filesystem. The filesystem cache had no
    expiry and grew without bound in the user data directory.
  - New hydrogeofetch_cache_clear() removes data hydrogeofetch has
    cached in hydrogeofetch_data_dir(), either all of it or by category
    (responses, tiles, vaa, mainstems).

Bug fixes

  - get_3dhp() queries by universalreferenceid are now chunked, avoiding
    the service gateway timeouts (HTTP 504) that previously caused large
    reachcode lookups to return NULL. #475
  - get_catchment_characteristics() no longer surfaces the benign arrow
    "discarded from R metadata" warnings from the parquet metadata
    roundtrip. #475
  - Five stop() calls passed .call = FALSE instead of call. = FALSE,
    which appended FALSE to the error message the user saw. In
    get_nhdplus() and the query parameter checks.
  - Web requests now carry a 30 second connect timeout and a 300 second
    request timeout, so an unresponsive service can no longer hang a
    call indefinitely. Bulk downloads keep the connect timeout but are
    not capped on total duration.
  - check7z() no longer appends to PATH with Sys.setenv(); it returns
    the path to the 7z executable instead, leaving the user's
    environment alone.

Other changes

  - prepare_nhdplus() is no longer exported; it remains available
    internally. #475
  - Documentation and example fixes throughout, following package review
    (#475): corrected get_nldi_basin() and get_split_catchment() example
    plotting, clearer messaging when the experimental get_xs_point()
    cross-section service is unavailable, broadened README scope
    language, and vignette setup and typo cleanups.
  - Examples that cache data now redirect hydrogeofetch_data_dir() to a
    temporary directory and restore it, so running them does not write
    to user space.
  - Added missing @return documentation to
    get_catchment_characteristics(), get_characteristics_metadata(), and
    get_nldi_index().
  - Trimmed the get_3dhp() example, which downloaded a full work unit
    and several hundred reachcodes and took over three minutes.

Dependency changes

  - Removed pbapply from Imports (no longer needed)
  - Removed future, future.apply from Suggests (no longer needed)

Function Updates

  - get_drainage_area_estimates() now returns drainage_basin, the
    upstream NHDPlusV2 catchments dissolved into a single boundary with
    the portion below the submitted point trimmed off the outlet
    catchment, along with its area as basin_da_sqkm. No HUC boundaries
    are involved. Both are available when catchments = TRUE or
    catchment_data is supplied, and always for headwater starts. For a
    headwater the basin is the whole answer, so the drainage area
    estimates are taken from its geometry rather than from an NHDPlus
    areasqkm attribute with a geometric difference subtracted from it.
  - get_drainage_area_estimates() no longer reports a drainage_basin
    carrying area below the submitted point when a split of the outlet
    catchment was due but the processing service returned none. A failed
    split was previously indistinguishable from a start that needed no
    split, and the full outlet catchment was kept in both cases.
    drainage_basin is now NULL and basin_da_sqkm NA when the split is
    missing.
  - get_drainage_area_estimates() no longer drops catchments from
    drainage_basin when a start resolves to several outlet COMIDs and a
    split catchment is available. A split covers one outlet catchment,
    so substituting it while removing several lost the catchments of the
    outlets that were not split.

                        Changes in version 1.5.0                        

nhdplusTools 1.5.0 will be the last minor release of nhdplusTools before
the package is renamed. Path releases may occur to fix minor issues or
change things necessary for the name change, but the API is effectively
frozen and all new development will go toward the replacement package
hydrogeofetch. For more, see the README.md.

This release adds drainage area estimation tooling that combines HUC12
areas with NHDPlusV2 catchments to produce drainage area estimates at
arbitrary points on the network.

  - New get_drainage_area_estimates() returns total and
    contributing-only drainage area estimates at HUC12, HUC10, and HUC08
    resolutions, alongside a network-derived comparison value and an
    optional NHDPlusHR-catchment estimate. Non-contributing areas (from
    the HUC12 ncontrb_a attribute) are accounted for.

  - Between the start point and the nearest upstream HUC12 outlets,
    NHDPlusV2 catchment areas are used; upstream of those outlets, HUC12
    areas are used. The outlet catchment is split at the start point
    when it is more than outlet_split_threshold_m upstream of the
    catchment outlet.

  - Supports fully offline operation via local_navigation = TRUE (uses
    get_vaa() for network navigation), plus huc12_data and huc12_outlets
    arguments for user-supplied HUC12 polygons and pour points.

  - HU_inclusion_override allows keeping HUC12s in landscape-connected
    regions (e.g. Prairie Potholes) where parent HUC outlets are not
    on-network.

  - New vignette("drainage_area_estimation") walks through the workflow,
    including comparisons against NWIS drainage areas.

  - get_catchment_characteristics() now batches variables by shared S3
    source, greatly improving performance when requesting multiple
    variables. Fixes #449.

  - get_catchment_characteristics() and get_characteristics_metadata()
    gain a source parameter supporting "usgs" (default) and "streamcat"
    for EPA StreamCat data via the StreamCatTools package.

  - rescale_catchment_characteristics() gains a source parameter passed
    through to get_catchment_characteristics().

                        Changes in version 1.4.2                        

This minor release fixes scalability issues in web service clients.

                        Changes in version 1.4.1                        

This minor release fixes a bug in get_huc() and adds several additional
layers.

                        Changes in version 1.4.0                        

This release migrates to a new web service provider for the key function
subset_nhdplus(). The change is mostly backward compatible but some
minor differences in performance and response data may be noticed.

  - subset_nhdplus() and get_huc() migrated to pygeoapi-based web
    services.

                        Changes in version 1.3.2                        

  - adjusted nldi client to work with python version of the NLDI.

                        Changes in version 1.3.1                        

  - added a web service client for the NHDPlusHR arcgis server
    get_nhdphr(). #415
  - removed deprecated discover_nldi_characteristics() and
    get_nldi_characteristics()

                        Changes in version 1.3.0                        

  - improve performance of discover_nhdplus_id(). #417
  - update urls for NLDI and Geoserver to use api.water.usgs.gov
  - no longer check ring direction in get_nhdplushr() to avoid very slow
    run time on some polygon layers. #411
  - fix splinters returned from get_split_catchment. #406
  - add support for more NHDPlusHR data and improve data loads. #404,
    #360
  - Deprecated discover_nldi_characteristics() and
    get_nldi_characteristics()
  - Bug fix for data cache directory #389
  - Improved warnings for web service requests. #400, #398
  - Improved documentation for get_split_catchment(). #399

                        Changes in version 1.2.0                        

This release introduces a new data source:
https://reference.geoconnex.us and a vignette providing a broad overview
of data access with nhdplusTools.

  - Fixed bug with use of memoise cache #365
  - Added vignette("get_data_overview") for WaterSciCon24
  - Added get_geoconnex_reference() and discover_geoconnex_reference()

                        Changes in version 1.1.0                        

This release has a significant consolidation of code to work with web
services. It incorporates an ArcGIS REST service for the first time. It
has very few backward compatibility issues with previous versions and
introduces new functions and behavior.

A key change in behavior is the introduction of caching with the memoise
package. See #366, pull request #364, and documentation in
nhdplusTools_cache_settings() for more.

  - fixed nhd and nhdplushr urls #368
  - fix issue with nhdplusTools_data_dir() #365
  - fixed bug with large nhdplus downloads with empty tiles. #361
  - Added 3DHP_all service client. #363
  - Removed deprecated function get_huc12 and get_huc8
  - Updated documentation of get_huc() and other web service functions.
  - added nhdplusTools_cache_settings() to control use of a memoise
    cache. #366
  - fix minor bug related to binding nhdplushr data together. #380

                        Changes in version 1.0.0                        

nhdplusTools has been split into two packages. A new package hydroloom
is now available. It contains all nhdplusTools functionality not related
to specific dataset schemas. See this issue for details.

hydroloom

v1.0.0 should be backward compatible with v0.7 but deprecation warnings
have been added and one previously deprecated function has been removed.

This release evolves nhdplusTools to use hydroloom as a core dependency.
nhdplusTools now depends on hydroloom for all network navigation and
indexing functionality that is not specific to the nhdplus data model or
specific to a US-context.

Specific functions imported from hydroloom include:

  - hy
  - accumulate_downstream
  - add_levelpaths
  - add_pathlength
  - add_pfafstetter
  - add_streamlevel
  - add_streamorder
  - add_toids
  - disambiguate_indexes
  - fix_flowdir
  - get_hydro_location
  - get_node
  - get_partial_length
  - index_points_to_lines
  - index_points_to_waterbodies
  - make_fromids
  - make_index_ids
  - navigate_connected_paths
  - navigate_hydro_network
  - rename_geometry
  - rescale_measures
  - sort_network
  - st_compatibalize

See the hydroloom reference list for details about these functions.

With these functions migrated to hydroloom, a number of nhdplusTools
functions are now deprecated in favor of the hyroloom implementation.

  - make_node_topology()
  - get_pfaf()
  - get_path_members()
  - get_path_lengths()
  - get_levelpaths()
  - get_terminal()
  - make_node_topology()

In v1.0.0, the download_nhd() function was added.

In v1.0.0, the stage_national_data (previously deprecated) function was
removed.

                        Changes in version 0.7.1                        

This release fixes a minor example issue for CRAN checks

                        Changes in version 0.7.0                        

This release adds three new functions.

  - get_characteristics_metadata()
  - get_catchment_characteristics()
  - rescale_catchment_characteristics()

These functions are designed to provide access to:

Wieczorek, M.E., Jackson, S.E., and Schwarz, G.E., 2018, Select
Attributes for NHDPlus Version 2.1 Reach Catchments and Modified Network
Routed Upstream Watersheds for the Conterminous United States (ver. 3.0,
January 2021): U.S. Geological Survey data release,
\doi{10.5066/F7765D7V}.

See https://github.com/DOI-USGS/hydrogeofetch/issues/304 and
https://github.com/DOI-USGS/hydrogeofetch/issues/303 for details.

                        Changes in version 0.6.2                        

Update a test for a change in sf

                        Changes in version 0.6.1                        

No functional changes. Package migrated to new repository:
https://github.com/DOI-USGS/hydrogeofetch

                        Changes in version 0.6.0                        

  - minor change to behavior of navigate_nldi the "origin" will always
    be named as such rather than the named 1-element list being dropped
    to a data.frame.
    https://github.com/DOI-USGS/dataRetrieval/issues/623
  - Address scalability issue with web service downloads. #308
  - Add ability to add node topology back to a edge list topology. #302

                        Changes in version 0.5.7                        

  - Update reference for updated flowline attributes.
  - Fix failing example for cran checks.

                        Changes in version 0.5.6                        

  - Update URL for NHDPlusV2 download
  - make geometry validity checks more robust

                        Changes in version 0.5.5                        

Note v0.5.4 was burned in a CRAN release mistake

fixes a cran test failure in an example and an issue with precision of
get_flowline_index()

                        Changes in version 0.5.3                        

v0.5.3 is primarily bug fixes and enhancements.

One notable addition is the addition of the rpu_boundaries and
vpu_boundaries datasets which provide simplified polygons around the RPU
and VPU subset domains of the NHDPlusV2 dataset.

                        Changes in version 0.5.2                        

Mike Johnson is now recognized as an author of nhdplusTools.

New citation:

> citation("nhdplusTools")

To cite nhdplusTools in publications, please use:

  Blodgett, D., Johnson, J.M., 2022, nhdplusTools: Tools for Accessing and
  Working with the NHDPlus, https://doi.org/10.5066/P97AS8JD

New Functions

  - map_nhdplus() works much the same as plot_nhdplus() but creates a
    leaflet map rather than a static plot.
  - get_nldi_index() is used to call the NLDI web service to retrieve a
    hydrologic location.

Improvements

  - get_flowline_index() and get_waterbody_index() units and projection
    handling was altered. Defaults for the search_radius input were
    changed. Coordinate reference system handling in geoprocessing now
    relies on the CRS of the provided points rather than the flowlines.
  - get_nldi_basin() now supports precise basin delineation via a
    logical "split" parameter as well as the option to retrieve
    un-simplified basin geometry.
  - All network-dependent functions in nhdplusTools now return a warning
    and NULL if the web service is unavailable or fails for other
    reasons. This is according to CRAN policy.

                        Changes in version 0.5.1                        

This is a very minor release required by CRAN checks.

Some scalability issues in plot_nhdplus() were identified and fixed.

                        Changes in version 0.5.0                        

Documentation updates

  - A new "advanced network attributes" vignette was added.
  - All vignettes were revisited to ensure they are up to date.
  - The introduction vignette is now very brief with links to others.

Web Service Wrapper Functions

  - New functions get_raindrop_trace(), get_split_catchment() use a web
    service to access elevation data and retrieve a raindrop trace path
    to the nearest flowline in nhdplus V2 and a split catchment or basin
    to a precise point location respectively.
  - New functions get_xs_point(), get_xs_points(), and
    get_elev_along_path() access an elevation data service from the 3D
    Elevation Program and return cross section / path elevation data.

Utility Functions

  - New functions st_compatibalize() and rename_geometry() provide
    utility functionality to make two layers compatible, including
    projection and rename geometry columns respectively.

  - New function get_sorted() generates a sorted and optionally
    partitioned version of a dendritic tree.

Subsetting and data utilities

  - New function subset_vpu() wraps subset_rpu() and will subset a
    vector processing unit.
  - New function fix_flowdir() will re-order geometry nodes such that
    their order corresponds to the convention used in nhdplus data.
  - get_vaa() will now return an updated nhdplus network attribute set
    derived from multiple improved sources.

Network Navigation and Attributes

  - New function navigate_network() provides a wrapper around a number
    of network navigation capabilities and will support local or web
    service data.
  - New function get_tocomid() encapsulates a suite of functionality for
    converting edge node network representations to edge list format.
  - New function get_path_lengths() finds network distances between
    pairs of points.
  - New function get_partial_length() finds flowpath length up and
    downstream of a hydrologic location.

Plotting

  - plot_nhdplus() now includes on network and off network waterbodies.
  - New function get_wb_outlet() finds the outlet flowline of an
    on-network waterbody. This function works for both NHDPlusV2 and
    hires.

Package Internals

  - nhdplusTools testing has been updated to use testthat edition 3.
  - All terminal "toID" values are 0 rather than NA to avoid confusion
    with a mix of both.
  - igraph package is no longer a dependency. All graph algorithms have
    been implemented in base R for performance and simplicity.
  - Many package functions have been converted to use lowercase
    attributes. Attribute naming is settling into two modes -- ID, toID,
    and lowercase otherwise -- and comid, tocomid, and lowercase. The ID
    syntax is meant to be a temporary ID while comid is meant to be some
    external identifier that should be treated as such.
  - The master branch was renamed main.

                        Changes in version 0.4.3                        

  - New function added: get_streamlevel() calculates stream level, a
    bottom up level path order used by nhdplus to categorize complete
    river levels.

  - nhdplusTools internals have started to be refactored to use all
    lower case attribute names. No changes should be seen from outside
    the package, but please report issues with attribute naming.

  - Package testing now runs in parallel and uses revision 3 of the
    testthat package.

                        Changes in version 0.4.2                        

Version 0.4.2 further cleans up temporary and cache data functionality
for CRAN policy.

New functionality was added to indexing functionality.

  - When multiple flowlines are returned,
    disambiguate_flowline_indexes() supports for disambiguating them
    based on numeric or text attributes that should match. This
    functionality is not heavily tested, but forms a basis for expansion
    of this functionality going forward.
  - A function to find the point location of a flowline index,
    get_hydro_location() was added. It will return point geometry given
    a geometry identifier and measure along that identifier.

                        Changes in version 0.4.1                        

Version 0.4.1 is a minor release with bug fixes and updates for CRAN
policy.

A notable addition is handling for strictly dendritic topology encoded
using a "tocomid" attribute rather than "fromnode" and "tonode"
attributes. This functionality was added in subsetting functions and is
backward compatible with previous versions.

                        Changes in version 0.4.0                        

Version 0.4.0 adds new functionality for data access via web services
and updates the usability and flexibility of data discovery and
subsetting functions. This release includes significant rework of
functions related to network navigation and web-service data subsetting.

  - Previously, subset_nhdplus() included internal functionality for
    downloading data via web service. This code for subsetting NHDPlus
    via web service was completely rewritten and a number of new
    web-service data access functions were added. The get_nhdplus()
    function has been added and is for web-service data access only.
    subset_nhdplus(), which will output a subset to a local .gpkg file,
    now uses get_nhdplus() for download functionality.
  - navigate_nldi() and most other NLDI functions are now based on an
    NLDI client in the dataRetrieval package. Functionality has been
    maintained backward compatible with nhdplusTools v0.3 as much as
    possible with some minor modifications to accommodate changes in
    dataRetrieval. The response format from navigate_nldi() has been
    changed slightly in v0.4, it now includes both the origin feature
    and navigation type(s) as a list. Two un-needed nldi discovery
    functions were removed.
  - Functions to download and cache all NHDPlus VAAs (in a 195mb .fst
    file) for use in network navigation and other applications have been
    added. These can be used with any nhdplusTools functions that do not
    require geometry.
  - get_levelpaths() now has an override_factor parameter that will
    cause stream leveling to follow the weight rather than the name
    input if the weight is override_factor times larger than the name
    indication.

Functions Added

  - get_nhdplus()
  - get_nhdarea()
  - get_waterbodies()
  - get_gagesII()
  - get_huc12()
  - get_huc8()
  - download_vaa()
  - get_vaa()
  - get_vaa_names()
  - get_vaa_path()

Functions Removed

  - discover_nldi_navigation()
  - discover_nldi_sources()

Functions Deprecated

  - stage_national_data()

Functions with Consolidated Internals

  - discover_nhdplus_id()
  - get_nldi_basin()
  - get_nldi_feature()
  - subset_nhdplus()
  - plot_nhdplus()

Finally, the handling of data and caching in the package has been
overhauled. These changes are non-functional, but aim to decrease data
downloads by caching rather than using temp directories.

                       Changes in version 0.3.16                        

  - subset_nhdplus() now validates geometry and ensures all outputs are
    in NAD83
  - subset_nhdplus() queries the NHDPlus database rather than loading
    then filtering
  - get_levelpaths() rewritten for performance. Package now uses
    data.table.

                       Changes in version 0.3.15                        

  - Added discover_nldi_characteristics() and get_nldi_characteristics()
  - Changed navigate_nldi() to use the new NLDI navigation end point.
    Distance is now required.
  - Fixed a bug in get_flowline_index() to handle multipart lines.
  - Added flowline_only input to plot_nhdplus() to improve scalability
  - Added streamorder filtering to subset_nhdplus() for download
    filtering.
  - Default behavior of plot_nhdplus() updated to improve scalability of
    large downloads.

                       Changes in version 0.3.14                        

  - Added get_node() to get top and bottom of a flowline.
  - Switched geospatial data dependency to use labs.waterdata.usgs.gov
  - Added subset_rpu() to enable subsetting the NHD by Raster Processing
    Unit
  - Added get_waterbody_index() to find associations between point
    locations and waterbodies.
  - Added ability to get multiple matches from get_flowline_index() with
    max_matches parameter.
  - Added ability to download flowlines in get_flowline_index() rather
    than requiring data to be loaded locally.

                       Changes in version 0.3.13                        

  - tested with dplyr 1.0 and sf 0.9.

                       Changes in version 0.3.12                        

  - Added plot_nhdplus() with both outlet-based and bbox based data
    subsetting as well as rudimentary style modification capabilities.
  - Implemented stream order and Pfafstetter code generators
    (get_streamorder() and get_pfaf()) for dendritic networks.
  - subset_nhdplus() can now return data without writing a file to disk.
  - NLDI feature specification is more flexible, doesn't require names
    anymore.
  - get_nhdplushr() has been updated to modify terminal path and other
    identifiers making nhdplusHR subsets "stand alone". A function
    make_standalone() is now exported.
  - Added get_terminal() and get_pathlength() functions to generate
    nhdplus network attributes.

                       Changes in version 0.3.11                        

  - Updated link to NHDPlusHR data
  - Empty tibble rather than NULL response from empty NLDI result
  - Slides from AWRA National 2019, link in README

                       Changes in version 0.3.10                        

  - Added name alignment function
  - Added ability to sort response of network navigation

                        Changes in version 0.3.9                        

  - Added downloader functions for more hydrography datasets
  - Change navigations to exclude requested catchment

                        Changes in version 0.3.8                        

  - Modified vignettes and examples to avoid web-service calls breaking
    CRAN build.
  - Added error handling to some web service requests.
  - Added CITATION and NEWS.

                        Changes in version 0.3.7                        

  - First CRAN release!
