Investigation of satellite-related precipitation products for modeling of rainfed wheat production systems

Precipitation is a very important weather variable for growth and yield of rainfed crops. In many agricultural regions of the world, high-quality precipitation records are not available, and thus, gridded precipitation products (GPPs) have to be applied as an alternative. The main objective of this study was to identify the most accurate GPP for simulating crop yield over a major rainfed wheat production zone in Iran. For this purpose, fifteen global GPPs were evaluated versus the observed precipitation records for the simulation of rainfed wheat growth and development and yield estimation using the Cropping System Model (CSM) CERES-Wheat model embedded in the Decision Support System for Agrotechnology Transfer (DSSAT). The findings showed that multisource GPPs had generally higher skill for the yield estimation. Considering all statistical and simulation results obtained from three sites during 2000-2010, MSWEP (Multi-Source Weighted-Ensemble Precipitation) was found as the best alternative GPP to the observed precipitation data for rainfed wheat grain yield simulation with normalized root mean square error (NRMSE) of 4.6 and Nash-Sutcliffe efficiency (NSE) of 0.79, while CMORPH (the Climate Prediction Center morphing method) was the weakest with NRMSE of 13.3 and NSE as - 0.81. The results point to differences among GPP, but there is a need to evaluate in other regions if multi-purpose GPPs are in general more reliable than GPPs based on specific sources.

Automated summary

This study compared different global GPPs and observed precipitation data to simulate crop yield in a major rainfed wheat production zone in Iran, with MSWEP identified as the best alternative GPP. The results suggested that multisource GPPs generally had higher skill for yield estimation, but further evaluation in other regions is needed to determine if they are more reliable than GPPs based on specific sources.