fd_canopy_height()
: fix an error that was provoking the
cached rasters to be deleted in the current session.fd_allometry_tallo()
: new function to retrieve data
from the TALLO database (thanks to @i-c-grant in #1).
fd_pathogens_defid2()
:
quiet
argument, which defaults to
FALSE
.metadata_forestdata
:
List of chorological species updated
New list containing metadata of possible options for
fd_pathogens_defid2()
.
New list containing metadata of codes for discrete and forest
layers in fd_landcover_copernicus()
.
fd_landcover_esri()
: now includes land cover data
for 2023.
fd_canopy_height()
: gains a new argument
model
with two possible options:
model = "eth"
: default value. It returns the ETH
Global Canopy Model with 10 meters of spatial resolution.
model = "meta"
: it returns the Meta Global Canopy
Model with 1 meter of spatial resolution. Thanks to @Wycology in #2.
fd_forest_glad()
: new function that deprecates
fd_forest_extent_glad()
. It includes a new argument
model
to choose forest extent, forest height, or land cover
models from GLAD. It also fixes #7.
Properly name land cover instead of land use to
fd_landcover_copernicus()
and
fd_landcover_esri()
in their documentation.
fd_forest_eutrees4f()
: now gives an error with a
better message when
distrib == "nat" & period != 2005
.
Improve documentation, fix mistakes in urls to other functions,
add unit tests for all functions, correct wrong
fd_forest_extent_glad()
reference.
Remove dependency on RODBC
, and use DBI
and odbc
instead in fd_inventory_spain()
.
Improve error message when data is not available.
fd_canopy_height()
: now crops and then merges
multiples tiles, so it’s much faster when using
crop = TRUE
. It acquires argument mask
, and
losses ...
passed to terra::crop()
because it
fails to mask when used within crop()
.
Now every function have the argument quiet = FALSE
by default, and retrieves the citation of the dataset.
fd_forest_chorological()
: fixes #5. Species
Cedrus atlantica and Phoenix theophrasti are
eliminated from the metadata since they have no range data.
Chamaerops humilis is now well spelled. The function now
retrieve an error message when data is not available (fixes
#5).
fd_inventory_spain()
It was returning metadata with the same name always. Now it will be different depending on the dataset that we download, and it will be checked if it exists before downloading it (according to #3).
Some provinces in IFN4 were incorrectly georreferenced. Now they are correctly georreferenced based on the Documentador. IFN3 provinces are now also georreferenced
fd_landcover_copernicus()
and
fd_canopy_height()
now crop data regardless of the
CRS.
fd_forest_extent_glad()
: incorrectly crops to the
argument x
(fixes #7).