Change Details - Precip Mesonets Statistics

5-Sept-2019 (beta)
Messing with hard-wiring error bars to duplicate what I've been doing with Tom Hamill's 'permutation' method of estimating errors is contingency-table-based statistical model differences. error bars are not to be believed!. Also, only a few stats (like CSI) actually work.

22-July-2019 (beta)
Corrected an error with the TSS calculation. It now is correct and matches MADIS.

11-July-2019 (beta)
Corrected an error with formatting times. We can now see times to the nearest second in the "show text" window. This needed for looking at, among other things, the number of stations in the CUM_PCP "model", which varies by minute. Routine(s) changed: gov/noaa/pmel/util/

13-Feb-2019 (beta)
Added the ability to select individual networks, but only for 'models' CUM_PCP and OBS_ONLY, and only for stats: 'Observed', 'Ntot', and 'Nsta'.

8-Feb-2019 (beta)
Added the HRRR_OPS and RAP_OPS models.

7-Feb-2019 (beta)
Added the 'models' CUM_PCP and OBS_ONLY. These look only at observations--not model forecasts. CUM_PCP looks at the precip_mesonets2.cum_pcp2 table. OBS_ONLY looks at the precip_mesonets2.hourly_pcp2 table.

Prepared by Bill Moninger,
Last modified: Thu Sep 5 20:22:50 GMT 2019