Review for memory leaks/performance

[peterdesmet]

[bartaelterman]

Possible idea proposed by Adriaan:

Calculate the inverse distance weighting only once. Then, when a new grid needs to be calculated, you can quickly interpolate.

There is however a downside: this will not work when one of the radars has no data, because then you only use the other radars in the inverse distance weighting.

[vlasvlasvlas]

im running a localhost version and it crashes/leaks memory all the time.

any tips for performance? i'm running same numbers (stations, birds, just different longlats for testing)