evaluation.strip.data {BiodiversityR}

Evaluation strips for ensemble suitability mapping

Description

These functions provide a dataframe which can subsequently be used to evaluate the relationship between environmental variables and the fitted probability of occurrence of individual or ensemble suitability modelling algorithms. The biomod2 package provides an alternative implementation of this approach (response.plot2).

Usage

evaluation.strip.data(xn = NULL, ext = NULL,
    models.list = NULL, 
    input.weights = models.list$output.weights,
    steps=200, CATCH.OFF = FALSE
)

evaluation.strip.plot(data, TrainData=NULL,
    variable.focal = NULL, model.focal = NULL,
    ylim=c(0, 1.25),  
    dev.new.width = 7, dev.new.height = 7, ...
)

Arguments

xn

RasterStack object (stack) containing all layers that correspond to explanatory variables of an ensemble calibrated earlier with ensemble.calibrate.models. See also predict.

ext

an Extent object to limit the prediction to a sub-region of xn and the selection of background points to a sub-region of x, typically provided as c(lonmin, lonmax, latmin, latmax); see also predict, randomPoints and extent

models.list

list with 'old' model objects such as MAXENT or RF.

input.weights

array with numeric values for the different modelling algorithms; if NULL then values provided by parameters such as MAXENT and GBM will be used. As an alternative, the output from ensemble.calibrate.weights can be used.

steps

number of steps within the range of a continuous explanatory variable

CATCH.OFF

Disable calls to function tryCatch.

data

data set with ranges of environmental variables and fitted suitability models, typically returned by evaluation.strip.data

TrainData

Data set representing the calibration data set. If provided, then a boxplot will be added for presence locations via boxplot

variable.focal

focal explanatory variable for plots with evaluation strips

model.focal

focal model for plots with evaluation strips

ylim

range of Y-axis

dev.new.width

Width for new graphics device (dev.new). If < 0, then no new graphics device is opened.

dev.new.height

Heigth for new graphics device (dev.new). If < 0, then no new graphics device is opened.

...

Other arguments passed to plot

Details

These functions are mainly intended to be used internally by the ensemble.raster function.

evaluation.strip.data creates a data frame with variables (columns) corresponding to the environmental variables encountered in the RasterStack object (x) and the suitability modelling approaches that were defined. The variable of focal.var is an index of the variable for which values are ranged. The variable of categorical is an index for categorical (factor) variables.

A continuous (numeric) variable is ranged between its minimum and maximum values in the number of steps defined by argument steps. When a continuous variable is not the focal variable, then the average (mean) is used.

A categorical (factor) variable is ranged for all the encountered levels (levels) for this variable. When a categorical variable is not the focal variable, then the most frequent level is used.

Value

function evaluation.strip.data creates a data frame, function evaluation.strip.data allows for plotting.

Author(s)

Roeland Kindt (World Agroforestry Centre)

References

Kindt R. 2018. Ensemble species distribution modelling with transformed suitability values. Environmental Modelling & Software 100: 136-145. doi:10.1016/j.envsoft.2017.11.009

Elith J, Ferrier S, Huettmann F & Leathwick J. 2005. The evaluation strip: A new and robust method for plotting predicted responses from species distribution models. Ecological Modelling 186: 280-289

See Also

ensemble.calibrate.models and ensemble.raster

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)
# subset based on Variance Inflation Factors
predictors <- subset(predictors, subset=c("bio5", "bio6", 
    "bio16", "bio17"))
predictors <- stack(predictors)
predictors
predictors@title <- "base"

# presence points
presence_file <- paste(system.file(package="dismo"), '/ex/bradypus.csv', sep='')
pres <- read.table(presence_file, header=TRUE, sep=',')[,-1]

# the kfold function randomly assigns data to groups; 
# groups are used as calibration (1/5) and training (4/5) data
groupp <- kfold(pres, 5)
pres_train <- pres[groupp !=  1, ]
pres_test <- pres[groupp ==  1, ]

# choose background points
background <- randomPoints(predictors, n=1000, extf=1.00)
colnames(background)=c('lon', 'lat')
groupa <- kfold(background, 5)
backg_train <- background[groupa != 1, ]
backg_test <- background[groupa == 1, ]

# calibrate the models
# MAXLIKE not included as does not allow predictions for data.frames
# ENSEMBLE.min and ENSEMBLE.weight.min set very low to explore all
# algorithms.
# If focus is on actual ensemble, then set ENSEMBLE.min and 
# ENSEMBLE.weight.min to more usual values
ensemble.calibrate <- ensemble.calibrate.models(x=predictors, 
    p=pres_train, a=backg_train, 
    pt=pres_test, at=backg_test,
    ENSEMBLE.min=0.5, ENSEMBLE.weight.min = 0.001,
    MAXENT=0, MAXNET=1, MAXLIKE=1, GBM=1, GBMSTEP=0, RF=1, CF=1,
    GLM=1, GLMSTEP=1, GAM=1, GAMSTEP=1, MGCV=1, MGCVFIX=1, 
    EARTH=1, RPART=1, NNET=1, FDA=1, SVM=1, SVME=1, 
    BIOCLIM.O=1, BIOCLIM=1, DOMAIN=1, MAHAL=0, MAHAL01=1,
    Yweights="BIOMOD", 
    PLOTS=FALSE, models.keep=TRUE)

# obtain data for plotting the evaluation strip
strip.data <- evaluation.strip.data(xn=predictors, steps=500,
    models.list=ensemble.calibrate$models)

# in case predictions for DOMAIN failed
# however, ENSEMBLE should also be recalculated
DOMAIN.model <- ensemble.calibrate$models$DOMAIN
strip.data$plot.data[, "DOMAIN"] <- dismo::predict(object=DOMAIN.model, 
    x=strip.data$plot.data)

# in case predictions for MAHAL01 failed
predict.MAHAL01 <- function(model, newdata, MAHAL.shape) {
    p <- dismo::predict(object=model, x=newdata)
    p <- p - 1 - MAHAL.shape
    p <- abs(p)
    p <- MAHAL.shape / p
    return(as.numeric(p))
}

MAHAL01.model <- ensemble.calibrate$models$MAHAL01
MAHAL.shape1 <- ensemble.calibrate$models$formulae$MAHAL.shape
strip.data$plot.data[, "MAHAL01"] <- predict.MAHAL01(model=MAHAL01.model, 
    newdata=strip.data$plot.data, MAHAL.shape=MAHAL.shape1)

# create graphs
evaluation.strip.plot(data=strip.data$plot.data, variable.focal="bio6",
    TrainData=strip.data$TrainData,
    type="o", col="red")
evaluation.strip.plot(data=strip.data$plot.data, model.focal="ENSEMBLE",
    TrainData=strip.data$TrainData,
    type="o", col="red")


## End(Not run)