Skip Navigation Links 
NOAA logo - Click to go to the NOAA home page National Weather Service   NWS logo - Click to go to the NWS home page
Climate Prediction Center

About Us
   Our Mission
   Who We Are

Contact Us
   CPC Information
   CPC Web Team

HOME > Monitoring_and_Data > Oceanic and Atmospheric Data > Reanalysis: Atmospheric Data > CFSR plant hardiness zones

CFSR/GFS based growing degree days (GDD)


The Climate Forecast System Reanalysis (CFSR) and the Global Forecast System (GFS) creates analyses of many fields including the 2 meter temperature. The CFSR/GFS analyses were used to create the growing degree days. The growing degree days is a factor in the plant and pest development.

Celcius and Fahrenheit GDD

For corn, a commonly used GDD uses a minimum temperature (Tlower or Tbase) of 50F and a maximum temperature (Tupper or Tmax) of 86F. Now 50F and 86F corresponds to 10C and 30C. If you use the 10_30 model and multiply the results by 9/5, you'll get the GDD from 50F-86F model in Fahrenheit.


  1. For the CFSR, find the average daily 2m temperature (6 hour forecast)
  2. For the GFS, find the average daily 2m temperature by averging the daily Tmax and Tmin
  3. Convert the data from Kelvin to Celsius.
  4. Calculate the growing degree (crop dependent): GD = min(Tmax,max(Tbase,T)) - Tbase
    where Tbase and Tmax depend on the crop model. The formula is the same as:
    • if (T < Tbase) GD = 0
    • if (Tbase < T and T < Tmax) GD = T - Tbase
    • if (Tmax < T) GD = Tmax - Tbase
  5. For the grib files, the GDD is accumulated from January 1. To get the GDD for June 15 starting from March 15, you would take the June 15th GDD and subtract the GDD for March 14.
  6. The CFSR climatology GDD is the average GDD from 1979-2008.


The models are in Celcius and the first number is Tbase and the second number is Tmax.
  1. 5.5_30: (5.5C/30C) wheat, barley, rye, oats, flaxseed, lettuce, asparagus
  2. 10_30: (10C/30C) same as the 50F-86F model, corn, sorghum, rice, soybeans, tomato, Black cutworm, European Corn Borer

Details and Caveats

The CFSR uses a system very similar to those run by the National Weather Service to produce the 1-6 day weather forecasts that you see on TV. (Other countries run similar systems for their 1-6 day forecasts.) As part of producing a weather forecast, an analysis of the current atmospheric state is created. This analysis is made using data from satellites, aircraft, weather ballons and other sources but surprising no 2-meter temperature observations. The 2-m temperature is model derived product which depends on the land-surface model, the model physics, the model parameterizations and the analyses of the winds, temperatures and humidity at the model levels. There are many reasons why the analyses will differ from observations. For example, you make your temperature observation on a patch of grass and the model is using a forest. (The land surface is suppose to be representative of the average conditions in a 35x35 km grid box.) Yes, grass vs. forest can be a big difference. Another problem is model uses the average height of the grid box. Your measurement will be at a different elevation. An other problem is the model doesn't know the local orography. You could be on an exposed peak or in a sheltered valley. These features are not resolved within a 35-km grid used by the CFSR model.

One way to view the CFSR temperature is they are a model estimate of average temperature of a 35km x 35km grid box. It will not give you the same number as an observation.


plots and data


Calculations with different Tbase/Tmax can be accomodated. However, you will need to provide references for selecting the particular Tbase/Tmax values.

Please send comments to:

NOAA/ National Weather Service
National Centers for Environmental Prediction
Climate Prediction Center
5830 University Research Court
College Park, Maryland 20740
Climate Prediction Center Web Team
Page last modified: Aug 13, 2010
Disclaimer Privacy Policy