Most weather forecasts are made by computer, and some of these forecasts are further enhanced by humans. Computers can keep track of myriad complex nonlinear interactions among winds, temperature, and moisture at thousands of locations and altitudes around the world — an impossible task for humans. Also, data observation, collection, analysis, display and dissemination are mostly automated.
Fig. 20.1 shows an automated forecast. Produced by computer, this meteogram (graph of weather vs. time for one location) is easier for non-meteorologists to interpret than weather maps. But to produce such forecasts, the equations describing the atmosphere must first be solved.
- 20.7: Forecast Quality and Verfication
- NWP forecasts can have both systematic error and random error. By making ensemble forecasts you can reduce random errors caused by the chaotic nature of the atmosphere. By post-processing each ensemble member using Model Output Statistics, you can reduce systematic errors (biases) before computing the ensemble average. After making these corrections, the forecast is still not perfect. But how good is the forecast?