Statistical Prediction of the Temperature in Norman
Start with:
As downloaded:
plot_t.py makes this plot (from the included data file):
The NCEP reanalysis for the boundary layer is actually a six hour forecast. The 18Z forecast is stored in files timestamped as 12Z. (It is my understanding that surface obs are not actually used in this forecast.) Thus 0Z refers to the 6Z forecast, or midnight local time (and so is not really "morning" as indicated in some of the files). Likewise 12 Z is the forecast for 18 Z or noon local time.
With the included data, lsq.py predicts the temperature in Norman at the same time as a bunch of other observations. This is as if you lost your thermometer in 1998, but you had your thermometer (and other obs) in 1997 to train the model. Here is the verification when applied to 1998:
Your Task:
In lsq.py You will find the following words: Your task is to modify both the data files to forecast the temperature 24 hours after the observations. ...and more.
But apparently some more clarification is needed. One students found these words helpful:
You could do this with a spreadsheet. Cut the temperature column, except for the first two elements. Paste the values as a new column, shifted up two rows. You now have tomorrow's temperatures in today's record. Predict tomorrow's temperature using the other values in the record. Beware of December 31.
