NEWS.md

In Rdistance: Distance-Sampling Analyses for Density and Abundance Estimation

title: "Rdistance NEWS"

Changes in version 3.0.0 (2023-06-12)

Version 3.0.0 is a substantial change and upgrade.

• BIGGEST CHANGE: Measurement units are now required on all distances (e.g., off-transect distances, strip widths, study area size, etc.), and it is possible to specify output units. Unit conversion makes use of the units package, is internal, and automatic.
• One-sided Transects: Added the ability to analyze single-sided transects by setting singleSided = TRUE in call to abundEstim.
• User Requested: Added computation of bootstrap confidence intervals for effective sampling distances (ESW and EDR).
• NA Length Transects: NA length transects are now allowed. Observations on NA length transects are used to estimate distance functions, but are dropped when estimating density. This allows, for example, off-transect distance observations from one study area or year to estimate a distance function which is then applied to the observations from another study area or year.
• Added calculation of target density (number per unit area) to abundEstim, and it is included in bootstrapping.
• Allowed expansion terms and covariates in the same distance function.
• Renamed uniform likelihood to logistic. Uniform is now deprecated.
• Added covariates to Gamma likelihood estimation.
• Change in default parameter: Changed default value of area in abundEstim from 1 to NULL. NULL now translates to 1 square output unit.
• Added color to print methods. Results are in green, if the R terminal allows it (i.e., in Rstudio, but not R gui). Errors are in red.
• Reformatted default output to include target density and abundance when printing an abund object.
• Fixed bug in F.gx.estim that occasionally popped up when sighting function was monotonically decreasing.
• Updated starting values for faster estimation.
• Decreased lower limit of negexp likelihood parameter to achieve more valid fits.
• Warning messages now report which parameters are at their boundary, and which boundary (lower or upper).
• Bug fixes:
• Fixed issue with contrasts in model.matrix
• Fixed issue of no observations during bootstrapping
• Fixed issues with changing scale locations (i.e., x.scl and g.x.sxl)
• Fixed ESW for w.lo > 0
• Restricted likelihood values to be positive
• Substantial documentation updates
• Substantial testing facilities added.

Changes in version 2.1.5 (2020-06-17)

• Fixed bug when bootstrap iteration selects zero targets
• Many updates to vignettes
• Updates to testthat files
• Added descriptive error for case when bySite==T and detection level covariates are present (Issue #72)

Changes in version 2.1.4 (2019-06-11)

• Fixed bug caused by empty factor levels
• Included automated testing (testthat)

Changes in version 2.1.3 (2019-01-02)

• Fixed scaling error in the Gamma likelihood causing mode to be less than 1.0
• Fixed inverted plots that occurred when w.lo > 0 (left-truncatation)
• Fixed bug in 'autoDistSamp' when 'bySite=TRUE'
• Fixed bug causing bootstrap failure of point transect surveys when 'siteData' contained only one column

Changes in version 2.1.2 (2018-08-23)

• Included methods for point transects. Previous version (v1.3.2) allowed only line transects.
• Implemented covariates in the detection functions.
• Implemented non-parametric (smoothed) distance functions
• Enhanced plotting abilities
• Enhanced vignettes
• Many bug fixes

Rdistance documentation built on July 9, 2023, 6:46 p.m.