Identify cut-off values in the probability density functions of different samples

I have created 30 probabaility density function plots for a material sample, with each pdf representing a different grid size (from 300 to 10) used to analyse porosity across the sample. The curves get flatter and wider with decreasing grid size. I have calculated an experimental value for porosity of ~35% for the sample but I need to select the ideal grid size in order to analyse how porosity varies across the material. The mean values are all around 35%, although the standard deviation/variance does increase significantly with decreasing grid size. Is there any way of quantifying a cutoff point which will help determine which grid size to use? I am looking for a grid size which is small enough to be able to find high resolution detail in the material but not too small as to start causing scatter (i.e. the grid measures either 100% porosity if it is in a pore or 0% porosity if it is on a particle, as is happening with the smaller grid sizes and results in large standard deviations around the mean). I have tried a few statistical approaches but have not had too much success.

Any help would be greatly appreciated.