The search code in find_potential_matches needs an estimate of the fraction of pixels in the matching area actually end up in M. It doesn't need to be accurate (although search should run a bit faster with an accurate number.)
It would also be better if the number were calculated for pixels with a matching ecoregion/country/luc0/5/10 which will increase the fraction.
When we have all 30 projects running, calculate this (we could easily instrument find_potential_matches to output it, including the exclusion of pixels outside the country/ecoregion/luc0/5/10).
The search code in find_potential_matches needs an estimate of the fraction of pixels in the matching area actually end up in M. It doesn't need to be accurate (although search should run a bit faster with an accurate number.)
It would also be better if the number were calculated for pixels with a matching ecoregion/country/luc0/5/10 which will increase the fraction.
When we have all 30 projects running, calculate this (we could easily instrument find_potential_matches to output it, including the exclusion of pixels outside the country/ecoregion/luc0/5/10).