kerneloverlaphr and BRB

Dear all,

I apologize in advance for any stupid question. But I am completeIy new 
in this kind of analyses and confused with some of the necessary input 
parameters...
I want to compare my BRB outputs of different individuals (gulls) from 
DIFFERENT COLONIES (i.e. single breeding colonies are about 5 to 160 km 
away (beeline) from each other) by calculating their volume of 
intersection with kerneloverlaphr.
Seems as if I  don't really understand the grid and extent option within 
the BRB command. Increasing the grid means increasing resolution, 
doesn't it? But what does increasing of the extent mean at the same time?

I need to use the same grid for all individuals to calculate the 
overlap, right? The problem is to find the right grid for all 
colonies/individuals.
Additionally, the output of kernel.area always differs when I use other 
grids and extents and I often get the following error message:

"The grid is too small to allow the estimation of home-range for the 
following value of percent: 
20,25,30,35,40,45,50,55,60,65,70,75,80,85,90,95. You should rerun 
kernelUD with a larger extent parameter."

Why are the areas not the same irrespective of the grid size?

Is it just not possible to calculate the overlap of individuals from 
different colonies?

Hope, anyone can help me out of this confusion.

Thanks a lot in advance,
Anna