Tags

maxent eval table row names correction

Improved the mean response curve plot of glm modelling.
Keep one y axis title, make title in the center.
2nd try

Improved the mean response curve plot of glm modelling.
Keep one y axis title, make title in the center.

- Alternative approach to creating constrained climate layers (plus other minor improvements)

Alternative approach to creating constrained climate layers (plus other minor improvements)

Entend "base layer" changes below to climate change experiments

## Failed test fixes
-   Ensure dashes (-) are substituted for underscores (\_) in predictor layers
    -   data.frame(x) converts '-' to '.' causing name mismatch errors

## Improvements
-   Option for users to select a "base layer" for experiments which determines:
    -   The resolution for resampling spatial rasters;
    -   The CRS for reprojecting the raster stack, occurence/absence points and training/prediction constraints
-   Additional handling allowing models to be trained under a different constrained region to which predictions are made.

## Failed test fixes
-Ensure all study area constraint longitudes are from -180 to 180 degrees
-Fixes failed test 00041 all algorithms
-Add implementation for Maxent bias layer
-Fixes for failed test 00042 maxent

## Improvements
-Alternative CRS for occurrences
-Allow non-WGS84 coordinates for occurrences via EPSG codes

## Failed test fixes

- Ensure/force GeoTiffs within ensemble experiments have matching extents
  - Fixes failed test 00028 ensemble job
- Removal of missing occurrence records
  - Fixes failed test 00031 maxent job
- Climate change experiments using terra (as per SDMs)
  - Fixes failed test 00035 cc (maxent) job
  - Fixes failed test 00040 cc (bioclim) job
- Added MaxEnt replicate capability to climate change experiments
  - Fixes failed test 00003 cc (maxent with replicates) job
  - Avoid unnecessary warnings

## Climate change experiment improvements

- Corrected datatype in categorical change GeoTiff
- Avoid empty trim when all cells less than threshold when calculating centre 
  gravity metric

# ecocommons 1.2.6

## Large constraint fix

- Re-implemented the filtering of occurrence points to within the area 
  constraint using terra in EC_SDM_geoconstrained.R
  - Deals with large global and complex constraint areas that were causing 
    errors
  - Better captures the occurrence points within the constraint area. A side 
    effect of this, is that some models sensitive to randomness (via seed) may 
    fail to rerun/calibrate successfully 

## Climate change experiment extent fix

- Ensure/force GeoTiffs within climate change experiments have matching 
  extents with those from the original experiments

More constraint fixes

- Fix constraint issue with SDM
- Fix extent issue with CC

1.2.4 ECR package as base image + constraint fixes

NOTE: pkg description is 1.2.3 but this matches the current build version for this ref.
- vignettes in progress
- add example data with inst/
- adding more user friendly functions for local use, exp setup and repeat platform exp
- adding parallel workflow for resampling etc in progress
- preparing to add test cases for climatch
- fixing bugs in CC
- fix bug in CC with SRE not working because of predictors in different order to original model
- enable clamping mask for biomod future projection
- fix bug in reading current + future predictions to stack
- adding some developer functions in zzz.R for debugging platform jobs quickly