meteo program to view the tutorial forecasts.
By viewing forecasts for several consecutive days, verify that the persistence forecast is indeed the same as the observations from the previous day.
In the winter time, the raw forecast has a warm bias during the day (around noon). The Kalman Filter tries to fix this problem. On 2012/01/01 the Kalman Filter starts cold, without any correction. Roughly how many days of training does the Kalman Filter need to correct this bias?