The importance of analyzing climate at high spatio-temporal resolution has been emphasized in the recent IPCC report. For instance, changes in trends at higher resolution are more likely to be attributed to the global warming than trends averaged over longer temporal and larger spatial scales. Several data sets of the precipitation have recently been created to meet this need. Data sets using independent methods and independent input data help verify that these data products are reliable for studying climate at the different spatial and temporal scales.

Here we present a new data set, CHOMPS – Cooperative Institute for Climate Studies (CICS) High Resolution Optimally interpolated Microwave Precipitation from Satellites, based on passive microwave satellite products, at 0.25o resolution. These data are developed using all available sensors with the latest and common retrieval scheme for the 1998 – 2007 (10 years in length) time period. The microwave estimates from the different sensors at hourly time scales are combined using Optimum Interpolation (OI) using estimates of the noise and spatial correlation scales, and using standardized analysis weights that ensure that the analysis is not damped when data are sparse. This technique reduces the random errors while still capturing the tails of the distribution of precipitation.

Evaluation of this data set with observational data such as the STAGE IV radar/gauge estimates and the TAO/TRITON buoy gauges and other high resolution precipitation products such as the GPCP-1dd, CMORPH and TMPA data have been performed and will be presented. For instance, correlations of the different high resolution products with STAGE IV radar over the US, indicates that CHOMPS has comparable results to CMORPH and TMPA with higher values over the south-east and lower values over the Rockies. Taylor diagrams of regional comparisons among the global high resolution products indicate that CHOMPS compares well with CMORPH and TMPA, with correlations > 0.9, and it has similar variability and has lower bias over the equatorial oceans. Regions of low rainfall corresponding to the downward branches of the Hadley and Walker Cells exhibit greater bias and lower correlations over the oceans. Comparison of monsoonal regions among the different blended precipitation products indicates that in general, CHOMPS, compares best with CMORPH. Both these products do not incorporate gauges in them. The biases with GPCP 1dd are the largest compared to all the blended products in all cases. This study shows that CHOMPS can serve as an independent high quality data product that can be used for examining precipitation at both local and global scales.