Nicaragua: exposure in Lago Nicaragua

[LaurensJN]

I have already addressed this in #23 but there is a problem in Nicaragua, that we don't know how to deal with. There is one district in Nicaragua that is 'Lago Nicaragua'. I think this 'district' originates from the GADM boundaries, where it is (correctly) identified as water body. There is in the GADM dataset an attribute ENGTYPE_1, which for any of the other districts in Nicaragua ia 'Department' and for this lake is 'Water body'.

If there would just be a few occupants and buildings in the region, we would just ignore it, but there are almost 26.000 occupants at night in this area. In my opinion they should be redistributed over the neighbouring regions, because having a population in a lake does not make sense.

See: https://github.com/gem/global_exposure_model/blob/main/Caribbean_Central_America/Nicaragua/Exposure_Res_Nicaragua_Adm1.csv#L308-L342