Model Output Statistics (MOS) are produced by taking raw numerical forecast model output and comparing it with observations to produce hopefully bias corrected results. Some of the models were predicting very cold temperatures yesterday and actual reported lows ended up being a bit warmer than predicted. The featured chart looks at the range of predicted morning low and afternoon high temperatures by the NAM MOS data along with the actual observation from the Des Moines Airport sensor. This range is produced by looking at the various MOS output from the model runs for each forecast time.

Voting:
Good = 19
Bad = 9

Tags:   mos  

Each day, computerized forecast models attempt to predict the weather and produce a lot of numeric output. One application of this output is to apply statistics to these values and attempt to produce more accurate forecasts (called Model Output Statistics [MOS]). The featured chart presents the range of MOS forecasts of daytime high and overnight low temperature for Des Moines from the GFS model along with the actual observations. In general, you can see that the MOS is by no means perfect, so there is currently room for humans to hopefully improve the process.

Voting:
Good = 29
Bad = 5

Tags:   mos