|Model Run:||Table Region:||Plot Region:||Output:||
How to navigate:
On the main table: Click a field (e.g., 'Z') to switch to a sub-table with data for each vertical level
Click a value (e.g., '2.8') to plot a multi-panel image for that time and field at all levels
On a sub-table: Click a value (e.g., '2.8') to plot an image for that hour, field and level
Click a level (e.g., '500') to loop images for that field and level at all forecast hours
For a different table: Select the desired Model Run, Table/Plot Region, and Output Type from the drop-down menus above, and click View Table
The NAEFS Ensemble: A 42-member ensemble consisting of 21 GEFS ensemble members and 21 Canadian (GEPS) members. Each set of 21 members includes a control run and 20 initial condition perturbations. Although the GEFS and GEPS are run at native resolutions of 55 and 66 km, respectively, the NAEFS is distributed on a 1x1-degree grid.
NAEFS Standardized Anomaly: How different is the model forecast from the climatological mean? Compares the NAEFS ensemble mean forecast to a 3-week running mean and standard deviation derived from the 1979-2009 Climate Forecast System Reanalysis. Standardized anomaly = (NAEFS_forecast - CFSR_climatology_mean) / (CFSR_climatology_standard_deviation)
NAEFS Percentile (Recommended): Where would the model forecast fall with respect to climatology? Example: MAX at 00Z indicates that values in the current NAEFS forecast are greater than all 00Z values in the CFSR climatology for a 3-week period centered on the valid day. Forecasters are encouraged to focus on "MAX" and "MIN" values, indicating that the ensemble is forecasting an event that would fall outside the 1979-2009 climatology for this time of year
NAEFS Return Interval: How often do these forecast values show up in the climatology? Specifically, how often were the CFSR values (in a 3-week period centered on the valid time) more extreme than values in the NAEFS forecast. Example: a return interval of 5 on Feb 15th means that roughly every 5 years, there is a day in mid-February when values in the current forecast were met or exceeded. Another example: "outside CFSR climate" for temperature means that none of the mid-Febrary reanalyses were this warm between 1979 and 2009.
NAEFS Probabilities: How many of the ensemble members produce "extreme" values? Indicates the fraction of NAEFS members with values either higher or lower than any CFSR reanalysis (in a 3-week period centered on the valid time). 60% probability of a min for MSLP on 00Z 15 Sept means that 60% of the NAEFS members have MSLP values lower than any 00Z, mid-September reanalysis. We use the word "extreme" loosely because these are rarely all-time highs or lows -- they're just outside the 1979-2009 climatology for this time of year
GEFS Model Climate: How does this forecast compare to past forecasts? Same calculations as for the NAEFS outputs, but in this case the GEFS ensemble mean is compared to the GEFSv12 reforecast climatology (2000-2019). The current forecast is placed in the context of reforecasts with the same lead time and similar valid dates (e.g., the current 36-h forecast valid at 00Z on 15 Oct 2020 is compared to all 36-h reforecasts valid between 5 and 25 Oct 2000-2019). For example, large M-Climate temperature anomalies mean that it's unusual for the ensemble mean to already be this warm at this lead time. Prior to 23 September 2020, GEFS forecasts were compared to the GEFSv10 reforecast climatology (1985-2012).
For more information:
First VLab Webinar (~30 min) describing the table and verification
Slides from this webinar
Second VLab Webinar (~40 min) describing the table and verification
Slides from this webinar
STID Forecast Confidence Toolkit
Bill Lamberson (Weather Prediction Center - DTB)
Trevor Alcott, Chad Kahler (Western Region STID)
Randy Graham, Nanette Hosenfeld (WFO - Salt Lake City, UT)
Rich Grumm (WFO - State College, PA)