runBeetle: Wrapper function to analyse the phylogenetic spread of...
In gisma/robubu: the rolling burning bus - whatever, it runs fine for now

Description Usage Arguments Details Value Author(s) References Examples

The phylogenetic spreading of the tibetan carabidae seems to be correlated to valley linkage and available humidity. Long term humidity is bound to precipitation and morphometry. It is obvious that the estimation of spreeading speed and range is in no case a simple euclidian one. runBeetle provides a first better estimation using a cost or friction analysis assuming that the spread is following natural lines of wetness e.g. valleys, humidity gradients...whatever I'am not a beetle ask the beetle beetle guy.

1 2	runBeetle(rootDir,workingDir="cost",inputData=NULL,costType="tci", externalCostRaster=NULL,internalCost=TRUE)

`rootDir`	project directory
`workingDir`	working directory
`inputData`	location data containing obligatory the cols lon,lat and optional a code col. The format has to be a data frame see example.
`costType`	used if internalCost = TRUE. default is "tci" you can choose "tci" "dem.filled" or "accu" see details for more information
`internalCost`	default = TRUE switches to external provided GTiff file wich has to be named cost.tifx
`dump`	default = FALSE if TRUE export r.terraflow products to GTiff
`usedump`	default = FALSE instead of running r.terraflow again use products products to GTiff
`externalCostRaster`	default = NULL you may provide a GTiff file with the name cost.tif NOTE you have to set internalCost = FALSE

The core of the analysis is an isotropic/anisotropic least cost path calculation. By default the cost surface is assumed to be a local derivate of the morphometry with respect to the potential soil humidity. A perfect approch to derive such information is the use of a Digital Elevation Model DEM and some corresponding derivates as the Topographic Convergence Index.

If you choose "tci" (default) the cost surface provides an estimation of rainwater runoff availability to plants based on specific catchment area (A) and local slope (b) such that TCI = ln(A/tan b) (Beven & Kirkby, 1979). This seems be pretty straightforward and fairly suitable for the beetles "behaviour.

If you chosse "accu" a typical accumulation cost grid from the original DEM will be used

If you choose "dem.filled" a typical accumulation cost grid from the hydrologically corrected DEM will be used

For the r.walk algorithm the non accumulated data is used.

runBeetle returns:
(a) a dataframe with the (a) euclidian distances, (b) cost distance (isotropic cost surface) and (c) the walk distance (anisotropic cost surface)

Chris Reudenbach
Maintainer: Chris REudenbach reudenbach@uni-marburg.de

Schmidt, J., B<c3><b6>hner, J., Brandl, R. & Opgenoorth, L. (in review): Mass elevation and lee effect override latitudinal effects in determining the distribution ranges of species: Ground beetles from the Himalaya-Tibet Orogen. <e2><80><93> PLoS ONE.

GRASS70 SAGA GIS

## Not run: 
#### NOTE: You obligatory need GRASS70

library(doParallel)
library(foreach)
library(raster)
library(sp)
library(gdalUtils)
library(rgdal)

### NOTE: the area is specifified and automatically downloaded (SRTM) by the input data extent plus a "zone"
### read a csv file
fn<-system.file("bug.csv", package="biomix")
beetleLocs<-read.csv2(fn,header = TRUE,sep = ',',dec = '.',stringsAsFactors=FALSE)

### use some arbitrary locations from scratch
beetleLocs <- as.list(c("86.83", "28.20", "100","84.58", "28.67", "200" ,"83.87", "28.80", "300"))
beetleLocs <- data.frame(matrix(as.numeric(unlist(beetleLocs)), nrow=3, byrow=T),stringsAsFactors=FALSE)
colnames(beetleLocs)<-c("lon","lat","code")
beetleDist<-runBeetle(rootDir = "/home/creu/proj/beetle" ,inputData = beetleLocs, usedump=TRUE, walk=TRUE)

## End(Not run)