Check bias with weather data (cloud bias)

[linaskerath]

The last weighty bias in the data set concerns cloud cover. The missing data in the optical data set that we hope to fill stems from obstructions by clouds. The existing optical data, however, only covers days without or minimal cloud cover. Even though it seems reasonable that direct sunlight affects ice melt differently than cloud cover, we have no data to test that assumption thus far. We must be aware, that our data is inherently biased against cloudy days.

[linaskerath]

Might need to do postprocessing.

Need to check if there is a performance difference between cloudy and not- cloudy days (once we have weather station data).

Discussed in meeting #6