ensemble.analogue: Climate analogues from climatic distance raster layers.

In BiodiversityR: Package for Community Ecology and Suitability Analysis

Climate analogues from climatic distance raster layers.

Description

Function ensemble.analogue creates the map with climatic distance and provides the locations of the climate analogues (defined as locations with smallest climatic distance to a reference climate). Function ensemble.analogue.object provides the reference values used by the prediction function used by predict .

Usage

ensemble.analogue(x = NULL, analogue.object = NULL, analogues = 1,
    RASTER.object.name = analogue.object$name, RASTER.stack.name = x@title,
    RASTER.format = "GTiff", RASTER.datatype = "INT2S", RASTER.NAflag = -32767,
    limits = c(1, 5, 20, 50), limit.colours = c('red', 'orange', 'blue', 'grey'),
    CATCH.OFF = FALSE)

ensemble.analogue.object(ref.location, future.stack, current.stack, name = "reference1", 
    method = "mahal", an = 10000, probs = c(0.025, 0.975), weights = NULL, z = 2)

Arguments

x	RasterStack object (stack) containing all environmental layers (climatic variables) for which climatic distance should be calculated.
analogue.object	Object listing reference values for the environmental layers and additional parameters (covariance matrix for method = "mahal" or normalization parameters for method = "quantile") that are used by the prediction function that is used internally by predict. This object is created with ensemble.analogue.object.
analogues	Number of analogue locations to be provided
RASTER.object.name	First part of the names of the raster file that will be generated, expected to identify the area and time period for which ranges were calculated
RASTER.stack.name	Last part of the names of the raster file that will be generated, expected to identify the predictor stack used
RASTER.format	Format of the raster files that will be generated. See writeFormats and writeRaster.
RASTER.datatype	Format of the raster files that will be generated. See dataType and writeRaster.
RASTER.NAflag	Value that is used to store missing data. See writeRaster.
limits	Limits indicating the accumulated number of closest analogue sites. These limits will correspond to different colours in the KML map. In the default setting, the closest analogue will be coloured red and the second to fifth closest analogues will be coloured orange.
limit.colours	Colours for the different limits based on number of analogues.
CATCH.OFF	Disable calls to function tryCatch.
ref.location	Location of the reference location for which analogues are searched for and from which climatic distance will be calculated, typically available in 2-column (lon, lat) dataframe; see also extract.
future.stack	RasterStack object (stack) containing the environmental layers (climatic variables) to obtain the conditions of the reference location. For climate change research, this RasterStack object corresponds to the future climatic conditions of the reference location.
current.stack	RasterStack object (stack) containing all environmental layers (climatic variables) for which climatic distance should be calculated. For climate change research, this RasterStack object corresponds to the current climatic conditions and range where climate analogues are searched for.
name	Name of the object, expect to expected to identify the area and time period for which ranges were calculated and where no novel conditions will be detected
method	Method used to calculate climatic distance: method = "mahal" results in using the Mahalanobis distance (mahalanobis); method = "quantile" results in dividing the differences between reference climatic values and climatic values in the 'current' raster by a quantile range obtained from the 'current' raster; method = "sd" results in dividing the differences between reference climatic values and climatic values in the 'current' raster by standard deviations obtained from the 'current' raster; and method = "none" results in not dividing these differences.
an	Number of randomly selected locations points to calculate the covariance matrix (cov) to be used with mahalanobis, therefore only used for method = "mahal". See also randomPoints.
probs	Numeric vector of probabilities [0,1] as used by quantile). Only used for method = "quantile".
weights	Numeric vector of weights by which each variable (difference) should be multiplied by (can be used to give equal weight to 12 monthly rainfall values and 24 minimum and maximum monthly temperature values). Not used for method = "mahal".
z	Parameter used as exponent for differences calculated between reference climatic variables and variables in the 'current' raster and reciprocal exponent for the sum of all differences. Default value of 2 corresponds to the Euclidean distance. Not used for method = "mahal".

Details

Function ensemble.analogues maps the climatic distance from reference values determined by ensemble.analogues.object and provides the locations of the analogues closest analogues.

The method = "mahal" uses the Mahalanobis distance as environmental (climatic) distance: mahalanobis.

Other methods use a normalization method to handle scale differences between environmental (climatic) variables:

ClimaticDistance = (\sum_i(weight_{i}*(|T_{i}-C_{i}| / norm_{i})^z))^(1/z)

where T_{i} are the target values for environmental (climatic) variable i, C_{i} are the values in the current environmental layers where analogues are searched for, weight_{i} are the weights for environmental variable i, and norm_{i} are the normalization parameters for environmental variable i

Value

Function ensemble.analogue.object returns a list with following objects:

name	name for the reference location
ref.location	coordinates of the reference location
stack.name	name for time period for which values are extracted from the future.stack
method	method used for calculating climatic distance
target.values	target environmental values to select analogues for through minimum climatic distance
cov.mahal	covariance matrix
norm.values	parameters by which each difference between target and 'current' value will be divided
weight.values	weights by which each difference between target and 'current' value will be multiplied
z	parameter to be used as exponent for differences between target and 'current' values

Author(s)

Roeland Kindt (World Agroforestry Centre) and Eike Luedeling (World Agroforestry Centre)

References

Bos, Swen PM, et al. "Climate analogs for agricultural impact projection and adaptation-a reliability test." Frontiers in Environmental Science 3 (2015): 65. Luedeling, Eike, and Henry Neufeldt. "Carbon sequestration potential of parkland agroforestry in the Sahel." Climatic Change 115.3-4 (2012): 443-461.

See Also

ensemble.novel

Examples

## Not run: 
# get predictor variables
library(dismo)
predictor.files <- list.files(path=paste(system.file(package="dismo"), '/ex', sep=''),
    pattern='grd', full.names=TRUE)
predictors <- stack(predictor.files)
predictors <- subset(predictors, subset=c("bio1", "bio5", "bio6", "bio7", "bio8", 
    "bio12", "bio16", "bio17"))
predictors
predictors@title <- "base"

# instead of searching for current analogue of future climate conditions,
# search for analogue in southern hemisphere
future.stack <- stack(crop(predictors, y=extent(-125, -32, 0, 40)))
future.stack@title <- "north"
current.stack <- stack(crop(predictors, y=extent(-125, -32, -56, 0)))
current.stack@title <- "south"

# reference location in Florida
# in this case future.stack and current.stack are both current
ref.loc <- data.frame(t(c(-80.19, 25.76)))
names(ref.loc) <- c("lon", "lat")

# climate analogue analysis based on the Mahalanobis distance
Florida.object.mahal <- ensemble.analogue.object(ref.location=ref.loc, 
    future.stack=future.stack, current.stack=current.stack, 
    name="FloridaMahal", method="mahal", an=10000)
Florida.object.mahal

Florida.analogue.mahal <- ensemble.analogue(x=current.stack, 
    analogue.object=Florida.object.mahal, analogues=50)
Florida.analogue.mahal

# climate analogue analysis based on the Euclidean distance and dividing each variable by the sd
Florida.object.sd <- ensemble.analogue.object(ref.location=ref.loc, 
    future.stack=future.stack, current.stack=current.stack, 
    name="FloridaSD", method="sd", z=2)
Florida.object.sd

Florida.analogue.sd <- ensemble.analogue(x=current.stack, 
    analogue.object=Florida.object.sd, analogues=50)
Florida.analogue.sd

# plot analogues on climatic distance maps
par(mfrow=c(1,2))
analogue.file <- paste(getwd(), "//ensembles//analogue//FloridaMahal_south_analogue.tif", sep="")
plot(raster(analogue.file), main="Mahalanobis climatic distance")
points(Florida.analogue.sd[3:50, "lat"] ~ Florida.analogue.sd[3:50, "lon"], 
    pch=1, col="red", cex=1)
points(Florida.analogue.mahal[3:50, "lat"] ~ Florida.analogue.mahal[3:50, "lon"], 
    pch=3, col="black", cex=1)
points(Florida.analogue.mahal[2, "lat"] ~ Florida.analogue.mahal[2, "lon"], 
    pch=22, col="blue", cex=2)
legend(x="topright", legend=c("closest", "Mahalanobis", "SD"), pch=c(22, 3 , 1), 
    col=c("blue" , "black", "red"))

analogue.file <- paste(getwd(), "//ensembles//analogue//FloridaSD_south_analogue.tif", sep="")
plot(raster(analogue.file), main="Climatic distance normalized by standard deviation")
points(Florida.analogue.mahal[3:50, "lat"] ~ Florida.analogue.mahal[3:50, "lon"], 
    pch=3, col="black", cex=1)
points(Florida.analogue.sd[3:50, "lat"] ~ Florida.analogue.sd[3:50, "lon"], 
    pch=1, col="red", cex=1)
points(Florida.analogue.sd[2, "lat"] ~ Florida.analogue.sd[2, "lon"], 
    pch=22, col="blue", cex=2)
legend(x="topright", legend=c("closest", "Mahalanobis", "SD"), pch=c(22, 3 , 1), 
    col=c("blue" , "black", "red"))
par(mfrow=c(1,1))

## End(Not run)

BiodiversityR documentation built on June 22, 2024, 11:57 a.m.