Home > Global Futures for Agriculture, Simulation modeling > Crop model calibration with yield trial data: Dealing with missing weather data

Crop model calibration with yield trial data: Dealing with missing weather data

One of our tasks in the Global Futures project was to calibrate potato cultivars in the DSSAT-SUBSTOR potato crop model with field trial data requested from CIP breeders. A common problem with the use of that kind of data was that weather data was missing. In most cases, only maximum and minimum temperature, as well as rainfall measured during the cropping season are available. The crop model, however, in addition requires solar radiation data. Furthermore, in order to carry out simulations with different planting or harvest dates, data which goes beyond the original cropping period is needed.

The approach we took to obtain a complete set of weather data that can be used with the crop model rests upon data provided by the NASA Langley Research Center POWER Project funded through the NASA Earth Science Directorate Applied Science Program. It consists of the following steps:

  •  Retrieve NASA weather data for the trial site you have data for from the NASA Agroclimatology database.
  •  Create a combined data set from the site data you have and the NASA data you need.
  • Create DSSAT weather files from the combined data set.

This typically leads to a data set which contains temperature and rainfall data measured at the site during the cropping period, NASA solar radiation data during the whole year and NASA temperature and rainfall data for the periods before and after planting and harvest dates, respectively.

Thereby, some aspects are worth considering:

  1. Since the NASA data is derived from satellite data, it may not be accurate for purposes of crop model calibration. In general, solar radiation data has been considered to be reliable, in parts even more reliable than data measured at weather station.* Temperature and rainfall data is less accurate and should not be used for calibration purposes.
  2. The spatial resolution of the NASA data is 1° latitude by 1° longitude. This may be too rough if point data is needed and may lead to errors, in particular in hilly/mountainous regions. For example, we tried to retrieve data for CIP’s test sites in La Molina. Entering the respective coordinates, we received data for a site at 1920 masl, which implied significantly higher values for solar radiation than those known for La Molina. We corrected this by moving further West, thus obtaining NASA data from a pixel with lower altitude.
  3. Of course it’s always better to have measured data from the site.

*Evaluation of Satellite-Based, Modeled-Derived Daily Solar Radiation Data for the Continental United States (2011), Jeffrey W. White, Gerrit Hoogenboom, Paul W. Wilkens, Paul W. Stackhouse Jr., and James M. Hoel, Published in Agron. J. 103:1242-1251.


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: