Development of a Tool to Process Gridded Climate Scenario Data from Coordinated Regional Climate Downscaling Experiment.

Future climate data at high resolution would help regional impact assessment of climate change using agricultural models. Those climate scenario data are usually provided as gridded data. For example, the Coordinated Regional Climate Downscaling Experiment (CORDEX) provides multiple sets of gridded data by regions, e.g., East Asia. To make the best use of CORDEX data, it would be preferable to have a data tool for processing of different spatial or temporal properties. A data tool for the CORDEX data, CORDEX Data Support Library (CDSL), was developed to support preparation of weather input data for agricultural models. The CDSL was implemented based on the Grid Analysis and Display System (GrADS). The CDSL has functionalities to parse metadata of spatio-temporal properties, to make ensemble datasets and, to support high-performance computing. In a case study, reference evapotranspiration (ET₀) was calculated using the CDSL and R with four different sets of CORDEX East Asia datasets. ET₀ was estimated during March to April for 25 years (1981 – 2005), which would require water resources to rice planting. Four sets of climate data were combined to create an ensemble set, which was used as inputs for calculation of ET₀. Another set of ensemble data was also prepared by averaging four sets of ET₀ outputs. All of the ET₀ output was compared to ET₀ calculated from AgMERRA data. The degree of agreement statistics for the ensemble of ET₀ estimates was greater than that of other ET₀ estimates. The running time of ET₀ estimation was reduced by 25% when 16 processors were used in the CDSL. Those results suggested that the use of CDSL would help timely and reliable estimation of environmental variables in agricultural ecosystems, which merits further studies on the application of the CDSL under different data products, e.g., EOBS data.