alterInclProbs.cluster: Alters inclusion probabilities to accommodate a 2-stage...

In MBHdesign: Spatial Designs for Ecological and Environmental Surveys

Alters inclusion probabilities to accommodate a 2-stage cluster sampling process

Description

Alters inclusion probabilities so that the cluster sampling approach respects specified inclusion probabilities.

Usage

 alterInclProbs.cluster( nCluster, clusterSize, clusterRadius, inclusion.probs, 
                                    maxIter=50, tolerance=NULL, 
                                    mc.cores=parallel::detectCores()-1, doPlot=FALSE)

Arguments

nCluster	an integer specifying the number of clusters.
clusterSize	an integer specifying the number of sites within a cluster.
clusterRadius	a numeric scalar giving the geographical extent of each cluster (from cluster centre to most extreme). Must be specified in metres (see extract and buffer).
inclusion.probs	a RasterLayer specifying the inclusion probability for each of the potential sampling sites. The values contained are the probability that each site will be included in the final sample. The values of this rasterLayer will be scaled so that they sum to ncluster*clusterSize. This is due to the alteration not depending on the number of clusters (otherwise they would be scaled to clusterSize * nCLuster. Must be specified.
maxIter	an integer giving the maximum number of iterations that the algorithm is allowed to perform before termination.
tolerance	a scalar numeric specifying the desired tolerance between observed inclusion probabilities (from 2-stage cluster sampling process) and the specified inclusion probabilities (from user). This tolerance specifies the accuracy with which the working inclusion probabilities are calculated.
mc.cores	an integer defining the number of processes to for some of the calculations on (in particular the calculation of distances to legacy sites). See parLapply(qv) in package parallel. Defaut is quite greedy, all but one of the cores available.
doPlot	a boolean indicating if a set of diagnostic plots should be produced within each iteration. Can be useful to help identify problems with algorithm convergence.

Details

Using the spatially-balanced 2-stage clustering process of Foster et al (in prep) will not, without careful consideration and adjustments, respect the inclusion probabilities specified by the user. This function provides the means to create _working_ inclusion probabilities, which when used in the 2-stage process _will_ produce the specified inclusion probabilies.

Value

The output of this function is designed for direct input into quasiSamp.cluster. It contains a SpatRaster with four layers: the user-specified inclusion probabilities (IP.s), the observed inclusion probabilities (IP.o, those inclusion probabilities obtained when using 2-stage sampling with the working inclusion probabilities), the working inclusion probabilities and the area-summed working inclusion probabilities (IP.bar). The IP.bar are based on the working inclusion probabilities and are used to sample the centre of clusters.

Author(s)

Scott D. Foster

References

Foster, S.D., Lawrence, E., and Hoskins, A. (in prep). Spatially-Balanced Cluster Sampling.

See Also

quasiSamp.cluster

MBHdesign documentation built on Sept. 25, 2023, 5:08 p.m.