isotopeAssignmentModel(), schoenersD(),
and surfaceSimilarityMatrix() now normalize correctly over
isoscapes with masked (NA) cells, passing
na.rm = TRUE in their internal sums; previously a single
NA cell turned the whole surface NA (and
errored in schoenersD()).schoenersD() returns a length-1 numeric value rather
than a 1x1 data.frame, and coerces legacy
Raster* inputs with the correct class check.isotopeAssignmentModel() now uses terra
instead of raster, and its in-function parallel-processing
capability is deprecated.schoenersD() now expects SpatRaster inputs
and automatically converts legacy Raster* objects.surfaceSimilarityMatrix() builds a Schoener’s D
similarity matrix across all layers of a SpatRaster. It is
vectorized (one value extraction plus base-R pairwise arithmetic),
roughly 100x faster than per-pair evaluation. The legacy
simmatrixMaker(), which takes a RasterStack,
is retained but deprecated.isotopeAssignmentModel() no longer silently discards a
supplied SD_indv vector whose first element is 0. The
default is now SD_indv = NULL (no individual-level error;
isoscape error only), replacing the ambiguous 0 that caused
the bug.makeMultiMonthIsoscape() now uses terra
and corrects two errors in the monthly combination: the
precipitation-weighted mean previously divided by the number of months
(returning half-magnitude values), and the equal-weight
(precip_stack = NULL) path errored. The combined-error
surface is unchanged (root-sum-of-square).meanAggregateClusterProbability() accepts any cluster
labels (integers, letters, factors), not just 1..K; the
previous version indexed clusters by position and silently failed for
other codings (#5). It also errors clearly when indivIDs do
not match the surface layer names, and names each output layer by its
cluster.makeMultiMonthIsoscape to combine multiple
monthly isoscapes and error maps into one of each, for making
month-specific isoscape assignments.schoenersD function now automatically normalizes
surfaces to sum to 1, rather than just advising it in the
documentation..assignmentMakercitation("isocat").isotopeAssignmentModeldplyr.meanAggregateClusterProbability function.makeCumSumSurface, function iterates
faster over RasterLayers.purrr::map and pipe operators
within function.purrr package removed from suggested packages.cumsumAtSamplingLocation and clusterSimmatrix
functions, in anticipation of change to matrix objects inheriting from
class “array” (which broke old class check code).isotopeAssignmentModel.makeQuantileSimulationSurface function.schoenersD
function.cumsumAtSamplingLocation
function.snow package with
calls to parallel.cumsumbelow, makeCumSumSurface,
cumsumAtSamplingLocation.meanAggregateClusterProbability function to
create mean probability-of-origin map based on cluster assignments (for
individuals of similar origin).projectSummarySurface function to show which
RasterLayer in a stack has highest value at any given location.makeQuantileSimulationSurface on object of class
RasterLayer.clusterSimmatrix function to match
pvclust version 2.0-1 with debug of r
parameter.projectCluster function with alternative
function (projectSummarySurface; see isocat
vignette).schoenersD function that compares the spatial
similarity of two surfaces.makeQuantileSimulationSurface function that
converts a surface to a quantile-simulation surface.NEWS.md file to track changes to the
package.IsoModAT (Isotope Models of Animal
Origin Analytical Toolset) to isocat (Isotope Clustering
and Assignment Tools).makeQuantileSurface function to converts
normalized probability surfaces to quantile surface. Matches
makeOddsSurface.oddsAtSamplingLocation function that finds odds
ratios of origin probability at a set of coordinates within a normalized
probability surface. Matches
quantileAtSamplingLocation.
Need a high-speed mirror for your open-source project?
Contact our mirror admin team at info@clientvps.com.
This archive is provided as a free public service to the community.
Proudly supported by infrastructure from VPSPulse , RxServers , BuyNumber , UnitVPS , OffshoreName and secure payment technology by ArionPay.