Sub-workflow parallel implementation of aerosol optical depth retrieval from MODIS data case on a Grid platform

摘要

Aerosol Optical Depth (AOD) is an significant parameter of aerosol optical properties. Operational production of AOD datasets over long time series, large-scale coverage puts on a severe challenge to computing technologies due to both the complexity of retrieval algorithm and the huge data amounts. The Grid computing solution-Remote Sensing Service Node (RSSN) was constructed as a high-throughput platform for remote sensing applications. Taking the sub-workflow level characteristics of some remote sensing retrieval applications into consideration, a sub-workflow parallel implementation for the Synergetic Retrieval of Aerosol Properties (SRAP) algorithm from the Moderate Resolution Imaging Spectroradiometer (MODIS) sensor data was taken on the RSSN, and an initial experiment result proved that the subworkflow parallel could further reduce the runtime of data parallel solutions commonly used.