An analysis of hourly precipitation is constructed on a 0.125 deg lat/lon grid over Guangdong province in southern China for 2005 by interpolating gauge observations at over 400 stations. First, analyzed fields of ratio between hourly observations and daily precipitation climatology of Xie et al. (2006) are created by interpolating the corresponding station values. Gridded values of hourly precipitation are then computed by multiplying the hourly ratio with the climatology. This two-step approach is adopted to improve the quantitative accuracy of the analysis and to take into account of the orographic effects in precipitation included in the definition of the daily climatology. The target domain of ~150,000 km^2 is covered by a dense network of over 400 gauges, ensuring reasonable quantitative accuracy of analyzed hourly precipitation.

The gauge-based precipitation analysis is applied to examine the performance of five selected high-resolution satellite precipitation estimates including the CMORPH of Joyce et al. (2004), the TRMM 3B42RT of Huffman et al. (2004), TRMM 3B42 of Huffman et al. (2006), the NRL blended product of Turk et al. (2004), and the PERSIANN of Hsu et al. (1997). Satellite estimates of the five products are integrated into fields of 0.25 deg lat/lon and 3-hourly resolution and compared with the gauge-based analysis for a period from April to August, 2005. Precipitation over the region in southern China is characterized by the evolution of East Asia monsoon from April to June, and by scattered convections in July to August. While our comparison covers a wide range of time and space scales, the focus of this work is the performance of the satellite products in depicting precipitation variations of sub-daily time scales. Our preliminary results showed the following:

1) All of the five satellite products are capable of capturing the overall spatial distribution and temporal variations patterns of precipitation reasonably well; 2) Performance of the satellite products varies for different regions and different precipitation regimes. In particular, bias changes with the monsoon evolution; 3) Diurnal cycle of precipitation derived from satellite estimates shows slightly different phase with that from the gauge observations; and 4) Satellite products tend to under-estimate the local maxima of precipitation.

Further work is underway to perform case studies and to extend the statistical analysis. Detailed results will be reported at the workshop.