Embedded Files
To assess the accuracy of the analysis (initial conditions), compare the analyzed fields to the rawinsonde data. Focus first on the low-level data, at 850 mb. Ask yourself the following questions:
- Does the analysis properly capture the intensity of cold fronts and their sharp temperature gradients?
- Are the coldest pockets of air properly represented by the model?
- Where do the 6-hour forecast and observations disagree?
- Does this suggest that the model forecast will be wrong for certain places in the future? If so, where?
Suppose the analysis, which is equivalent to the 0 hour forecast, shows a station behind a cold front with a strong wind from the northwest and a temperature of -6 C. There are no other observations nearby. Meanwhile, the 6-hour forecast shows the same station on the warm side of the -5 C isotherm, with the cooler air still a couple hundred kilometers upwind. If this observation is plausibly correct, the front may be a lot farther along than this forecast thinks it should be, and it will probably reach locations ahead of the front sooner than expected.
Report abuse