I don't think you can average the averages. For instance
10M @ 100% + 5 M @ 50% gives you 7.5 M @ 75% average.
If you then take 10M @ 50% + 5M @ 100%, you get the same average but the amount of potash would be significantly different
I think a better calculation would be calculate the equivilant width of pure potash for each sample and then average that number out.
A quick calculation gives me an average of 1.75 M of 100% KCL over the area you describe. The final number (13.26 MT of KCL) is less but is, I think, mathmatically more accurate.