GRASS logo

NAME

r.exdet - Quantification of novel uni- and multi-variate environments

KEYWORDS

similarity, multivariate, raster, modelling

SYNOPSIS

r.exdet
r.exdet --help
r.exdet [-pde] reference=raster[,raster,...] [projection=raster[,raster,...]] [region=region] output=raster [--overwrite] [--help] [--verbose] [--quiet] [--ui]

Flags:

-p
Most influential covariates (MIC)
-d
Mahalanobis distance in projection domain?
Keep layer Mahalanobis distance in projection domain?
-e
Mahalanobis distance in reference domain
Keep layer Mahalanobis distance in reference domain?
--overwrite
Allow output files to overwrite existing files
--help
Print usage summary
--verbose
Verbose module output
--quiet
Quiet module output
--ui
Force launching GUI dialog

Parameters:

reference=raster[,raster,...] [required]
Reference conditions
Reference environmental conditions
projection=raster[,raster,...]
Projected conditions
Projected conditions to be compared to reference conditions
region=region
Projection region
Region defining the area to be compared to the reference area
output=raster [required]
Suffix name output layers
Root name of the output layers

Table of contents

DESCRIPTION

Correlative species distribution models (SDMs) often involve some degree of projection into conditions non-analogous to those under which it has been calibrated. This projection into areas with novel environmental conditions is risky as it may be ecologically and statistically invalid. However, depending on the research question it may be difficult to avoid or indeed the objective of research to do so. An example is the prediction of potential distribution of species under future climates. The latter may include conditions hitherto not encountered in the area of interest. It is important to identify such areas and to interpret model results with care.

The r.exdet function allows you to identify areas with novel conditions, following methods developed by Mesgaran et al. (2014) [1][2]. This includes areas where conditions fall outside the range of values observed in the reference / calibration data set ( NT1: Type 1 novelty) or areas with novel combinations between the environmental variables (NT2: Type 2 novelty), which Mesgaran et al. call the multivariate combination novelty index.

The type 1 (NT1) similarity is similar to how the multi-environmental similarity measure (MESS) computes novel climates [3]. In both cases if a point is outside the range of a given covariate, it gets a negative value based on its distance to the minimum/maximum of that covariate. The difference is that the MESS is based on the most negative value amongst these covariates. The NT1, on the other hand, is the sum of all these distances. The NT1 thus accounts for all variables [2]. The NT1 can have infinite negative values to zero where zero indicates no extrapolation beyond the univariate coverage of reference data.

The type 2 (NT2) similarity is based on the Mahalanobis distance and is used to identify areas where conditions are within the range of univariate variation but which exhibits novel combinations between covariates. NT2 can range from zero up to infinite positive values. Values ranging from 0 to 1 indicate similarity (in terms of both univariate range and multivariate combination), with values closer to zero being more similar. Values larger than one are indicative of novel combinations.[1]

r.exdet can also compute the most influential covariate ( MIC ). For areas with novel conditions, this is the variable that has the lowest NT1 value. For areas with multivariate combination novelty, this is the variable that yields the largest percentage reduction in the Mahalanobis distance if dropped.

The function can be used to compare (1) conditions at two different times (e.g., current climate conditions and climate conditions in 2085). As input, the user needs to provide two different sets of environmental variables, each representing conditions at different times. The function can also be used to compare the conditions in two different areas. This can be done in three different ways:

Some of the options can be combined. For example, the use can set a mask, a set of layers describing current conditions (reference) and a set of layers providing future conditions (projection). In this case, the future conditions in the whole region are compared to the current conditions within the area defined by the MASK. (todo: provide some examples)

EXAMPLES

You can download a sample data set from http://www.climond.org/ExDet. The sample data contains 4 clipped Bioclim variable layers for Australia and South Africa sourced from the CliMond dataset. In this example we will use the Australia data as reference and the South Africa data as projected or test. In the example below I will assume you have downloaded the data and imported it in the currently open location/mapset (the coordinate system is latlon, EPSG 4326).
g.region raster=AusBio13
r.exdet -p reference=AusBio13@example,AusBio14,AusBio5,AusBio6 projection=SaBio13,SaBio14,SaBio5,SaBio6 output=AusSa

CITATION

When using this tool, please cite the paper describing the method in your publications or other derived products.

If you want, in addition, to cite this tool, you can use:

REFERENCES

[1] Mesgaran, M.B., Cousens, R.D. and Webber, B.L. (2014) Here be dragons: a tool for quantifying novelty due to covariate range and correlation change when projecting species distribution models. Diversity & Distributions, 20: 1147-1159, DOI: 10.1111/ddi.12209.

[2] ExDet: An stand alone extrapolation detection tool for the modelling of species distributions. URL: http://www.climond.org/ExDet.aspx

[3] Elith, J., Kearney, M. and Phillips, S. 2010. The art of modelling range-shifting species. Methods in Ecology and Evolution 1:330-342.

SEE ALSO

r.mess

AUTHOR

Paulo van Breugel, paulo at ecodiv.earth

SOURCE CODE

Available at: r.exdet source code (history)

Latest change: Thursday Feb 03 09:32:35 2022 in commit: f17c792f5de56c64ecfbe63ec315307872cf9d5c


Main index | Raster index | Topics index | Keywords index | Graphical index | Full index

© 2003-2024 GRASS Development Team, GRASS GIS 8.4.1dev Reference Manual