Page Banner

United States Department of Agriculture

Agricultural Research Service

Research Project: Predicting Impacts of Climate Change on Agricultural Systems and Developing Potentials for Adaptation

Location: Plant Physiology and Genetics Research

Title: Evaluation of satellite-based, modeled-derived daily solar radiation data for the continental U.S.

Authors
item White, Jeffrey
item Hoogenboom, Gerrit -
item Stackhouse, Paul -
item Hoell, James -

Submitted to: Agronomy Journal
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: May 4, 2011
Publication Date: June 6, 2011
Citation: White, J.W., Hoogenboom, G., Stackhouse, P.W., Hoell, J.M., 2011. Evaluation of satellite-based, modeled-derived daily solar radiation data for the continental U.S. Agronomy Journal. 103(4):1242-1251.

Interpretive Summary: Weather conditions have a major influence on agricultural production. Availability and quality of weather data, however, often constrain research and extension efforts. Daily values of incoming sunshine (solar radiation, abbreviated as SRAD) are especially problematic because the instruments used to measure SRAD need electronic integrators, the best sensors are expensive, and calibration to known values is difficult. The Prediction Of Worldwide Energy Resources (NASA/POWER; power.larc.nasa.gov) project at the NASA Langley Research Center estimates daily solar radiation based on data that are derived from satellite imagery, ground observations, windsondes, modeling and data assimilation. The solar data are available for each 1° cell of latitude and longitude over the globe. SRAD values can also be “generated” based on how radiation reaching the outside of the Earth’s atmosphere is reduced as it filters down to ground level. This study compares daily solar radiation data from NASA/POWER with instrument readings from 295 stations, as well as with values that were estimated with software to “generate” SRAD values using daily temperature and precipitation data from two sources. The NASA/POWER data consistently showed stronger relations with the observed data than did either set of generated values. However, while mean values of observed data and the two sets of generated data were comparable, the NASA/POWER data averaged 5 to 10% lower than observed SRAD. The differences were greater at lower latitudes and during summer months. Overall, the NASA/POWER solar radiation data are a promising resource for research and extension efforts where realistic accounting of historic variation in SRAD is required, but further research is needed to understand the reason for the consistent differences in average values.

Technical Abstract: Many applications of simulation models and related decision support tools for agriculture and natural resource management require daily meteorological data as inputs. Availability and quality of such data, however, often constrain research and decision support activities that require use of these tools. Daily solar radiation (SRAD) data are especially problematic because the instruments need electronic integrators, the best sensors are expensive, and calibration standards are seldom available. The Prediction Of Worldwide Energy Resources (NASA/POWER; power.larc.nasa.gov) project at the NASA Langley Research Center estimates daily solar radiation based on data that are derived from satellite imagery, ground observations, windsondes, modeling and data assimilation. The solar data are available for a global 1° coordinate grid. SRAD can also be estimated based on attenuation of extraterrestrial radiation (Q0) using daily temperature and rainfall data to estimate the optical thickness of the atmosphere. This study compares daily solar radiation data from NASA/POWER (SRADNP) with instrument readings from 295 stations (SRADOB), as well as with values that were estimated with the WGENR solar generator. WGENR was used both with daily temperature and precipitation records from the stations reporting solar data and records from the NOAA Cooperative Observers Program (COOP), thus providing two additional sources of solar data, SRADWG and SRADCO. Values of SRADNP for different grid cells consistently showed higher correlations (typically 0.85 to 0.95) with SRADOB data than did SRADWG or SRADCO for sites within the corresponding cells. Mean values of SRADOB, SRADWG and SRADNP for sites within a grid cell usually were within 1 MJm-2d-1 of each other, but NASA/POWER values averaged 1.1 MJm-2d-1 lower than SRADOB. The magnitude of this bias was greater at lower latitudes and during summer months. Overall, the NASA/POWER solar radiation data are a promising resource for regional studies where realistic accounting of historic variation is required.

Last Modified: 7/25/2014
Footer Content Back to Top of Page