Concrete and mortar have specific densities. The land is very flat, and the elevation is known. Average angle of repose for concrete rubble is known as well. Go out and sample 5,000 pieces of rubble, measure length in each dimension an categorize by size. Classify entire sampled rubble piles reasonably. Each category will have a percentage air to ruble ratio, starting around 30% and moving up from there. Survey the land with satellite imagery and count rubble piles. Multiply by actual specific density, and add the average vehicles and appliances. Voila.
Yes those are outliers, although I would bet money that there is good literature from post-war europe on the mathematics of estimating these things. Not to mention, lots of experience of workers who can walk up to a building and say "Yeup, that's 14-15 truckloads, 300 ton"
But this assumes piles of rubble not half standing buildings in big giant jagged pieces