using distance path with uniform cost - is the result simply path length?

Discussion created by kkober on Jul 12, 2011
Hi all,
I am trying to find a tool which basically helps me to identify  which locations can be reached by an animal from a single source location if there are barriers in the way which force the animal to walk around it. (in this case I am talking about seabirds in a colony and the question where they can go to feed at sea, within a given maximum foraging range (e.g. 580km), and given the fact that they will not fly over land).

I think distance path should be able to give me the solution, if I give the colony location as the source (file attached "rathlin_point") and as cost a raster file which holds the area within the foraging range (580km in this case) with a uniform  cost of 1 for all at sea areas and no data provided for terrestrial areas (file attached "r-rathlin"). If I understand the description of Distance Path correctly, the tool will then calculate the distance to each of the raster cells based on costs and distance to that cell, which means in this case (with a uniform cost) it should simply give me the distance to the cell. It should also "walk around" areas where no data is given (the terrestrial areas in this case).

The setting for the analysis:
(1) input raster or feature source data: "rathlin_point"
(2) output distance raster: I simply accepted the suggested name
(3) input cost raster: "r_rathlin"

I assumed that locations which were within 580km and which can be reached by a linear (Euclidian) distance from the source location (without any barriers in the way) should all be included, but that is not the case. It appears that at locations identified by Distance path as 580km away from the source are usually 540km or less away from the source, even if no barrier was in the way which might have forced a detour.

My question is: what exactly is the output distance given by Distance path if calculated in this way (without costs)? Is it a "Manhatten block distance" rather than an Euclidian distance?
And is there any other way I should have run the analysis to get the result I am looking for?

Thank you very much for your help!
I'll attach my input files, in case you want to try this out yourself.