Medium Range Forecasts (MRF)
One of the efforts of the California Applications Program (CAP) is to archive Medium Range Forecast (MRF) data that is available every day from the National Centers for Environmental Prediciton (NCEP). CAP's goal is to archive data for medium-range forecast applications in the California region.

Every day NCEP makes 15-day forecasts. 17 forecasts are available each day. 12 forecasts are made from 00Z conditions and 5 forecasts are made from 12Z conditions. The forecasts are global and data is saved every 12 hours. The forecast data is saved on a regular grid of 2.5 degrees latitude by 2.5 degrees longitude. Further information on the production of the MRFs is available at

In October 1999, there was a fire inside a supercomputer at NCEP. Because of this fire, NCEP was unable to make MRFs. After a few weeks, the Climate Diagnostics Center (CDC) began making MRF forecasts. CAP started archiving the forecast data from CDC when it became available in October 1999. In November 1999 NCEP again started the ensemble MRFs. CAP is now archiving data from both NCEP and CDC, saving the NCEP data when there is duplication.

MRF Varibles being saved by CAP
Accumulated Precipitation prcp (NCEP) &  apcp (CDC)
Mean Sea Level Pressure prmsl
Relative Humidity at 700mb rh
Temperature at 2 meters t2m
Temperature at 850mb t850
Maximum Temperature at 2 meters tmax
Minimum Temperature at 2 meters tmin
Zonal (u) wind at 10 meters u10m
Zonal (u) wind at 250mb u250
Zonal (u) wind at 500mb u500
Zonal (u) wind at 850mb u850
Meridional (v) wind at 10 meters v10m
Meridional (v) wind at 250mb v250
Meridional (v) wind at 500mb v500
Meridional (v) wind at 850mb v850
Geopotential Height at 1000mb z1000
Geopotential Height at 250mb z250
Geopotential Height at 500mb z500
Geopotential Height at 700mb z700
Accumulated Precipitation Probabilities ppf
MRF at 40N/120W
This figure was created by Mike Dettinger and shows an example of how CAP is using the MRF data. The figure shows MRF data at 40 degrees N latitude and 120 degrees W longitude. This grid point is over the central Sierra Mountain range of California. The figure shows two variables: precipitation and 850mb temperature (a figure for 1999 is shown below and includes insolation). Along the y-axis is the verification date for the forecast. Along the x-axis is the forecast day in lead time. The data shown in this figure reflect daily ensemble averages of the MRF data.

If the forecast was perfect, the values would be the same along the x-axis for a particular date. For example, if  a precipitation event was forecast for January 20th 15 days in advance (January 5th) and the precipitation event remained in the forecast for the next 15 days, one would see a line of constant shading or color parallel to the x-axis.

Mike Dettinger notes what is interesting about this figure is that it shows we can often reliably know if we can expect some weather event (say precipitation or warming) without having to know the magnitude of the event. Often forecast users don't need to know specific values but information about potential warming or precipitation can be crucial for planning. An example would be a dam operator who sees a warming 15 days in advance. Without knowing exactly how much warming to expect, the dam operator may expect stronger streamflow and may choose to release water from the dam before the warming event occurs to make room in the reservoir for the expected influx.

MRF at 40N/120W