Research on attaining methods of rainfall "truth" by assessment of radar measurement of rainfall

Abstract

Using the precipitation data given by high-density regional automated weather stations during main flood season in Hunan Province, in 2013, Cressman algorithm is introduced to calculate the regional precipitation value in 0.01°×0.01° resolution and is compared with the methods of a reciprocal of distance square, distance exponent, and distance square exponent. It is shown that the precipitation construction of Cressman algorithm is finer and closer to the features of gage data than other methods. Precipitation accumulation amount given by Cressman algorithm is in agreement with those obtained using methods of reciprocal of distance square and range exponent as well. Cross-validation tests show that the number with the absolute error less than 0.5 mm accounts for 98.71% of total gage number. The standard deviation of error is 0.24 mm, which is the least of all methods. As to 0.01°×0.01° resolution, the regional precipitation calculated by Cressman algorithm is adaptable to validate the radar estimation rainfall.