DESCRIPTION

Analysis of landscape connectivity using individual-based dispersal model inside a moving window. This module provides information about the patch connectivity within a moving window.

NOTES

The suitability matrix impacts the step direction of individuals. If individuals are moving beyond the mapset borders the indivuals are set back to their original source patches.

EXAMPLE

An example for the North Carolina sample dataset: The patch relevance concerning connectivity are based on patches of the landclass96 raster class 5 amd the time (amount of steps) from emigration to immigration is computed. The step length is set to 5 pixel, the output statistics are set to average time and variance of searchtime. For each patch 1000 individuals were released and the model stopped when at least 80% of all individuals sucessfully immigrated. The command below is computing the time for the whole landscape and the result is prompted to the screen
r.pi.searchtime.mw input=landclass96 output=searchtime_mw1 keyval=5 step_length=5 stats=average,variance percent=80 n=1000
for a graphical output a moving window size needs to be defined (here 7x7 pixel) and the amount of steps until immigration per moving window is provided in the output raster: 
r.pi.searchtime.mw input=landclass96 output=searchtime_mw1 keyval=5 step_length=5 stats=average,variance percent=80 n=1000 size=7
setting the perception range to 10 pixel: 
r.pi.searchtime.mw input=landclass96 output=searchtime_mw1 keyval=5 step_length=5 stats=average,variance percent=80 n=1000 size=7 perception=10
increasing the attraction to move towards patches to 10: 
r.pi.searchtime.mw input=landclass96 output=searchtime_mw1 keyval=5 step_length=5 stats=average,variance percent=80 n=1000 size=7 multiplicator=10
the previous examples assumed a homogeneous matrix, a heterogenous matrix can be included using a raster file which values are taken as costs for movement (0-100): 
# it is assumed that our species is a forest species and cannot move 
# through water, hence a cost of 100, does not like urban areas (class: 
# 6, cost: 10) but can disperse through shrubland (class 4, cost=1) 
# better than through grassland (class 3, cost: 2):
r.mapcalc "suit_raster = if(landclass96==5,1,if(landclass96 == 1, 10, if (landclass96==3,2, if(landclass96==4,1,if(landclass96==6,100)))))"
r.pi.searchtime.mw input=landclass96 output=searchtime1 keyval=5 step_length=5 stats=average,variance percent=80 n=1000 size=7 suitability=suit_raster

SEE ALSO

r.pi.searchtime, r.pi.searchtime.pr, r.pi

AUTHORS

Programming: Elshad Shirinov
Scientific concept: Dr. Martin Wegmann
Department of Remote Sensing
Remote Sensing and Biodiversity Unit
University of Wuerzburg, Germany

Port to GRASS GIS 7: Markus Metz

Last changed: $Date$