Sensitivity of the Precipitation Gauge Correction for the Estimation of Global and Continental Water Balance

GSWP-I was an intercomparison project of eleven LSMs to evaluate their results and to discuss problems among them. Runoff observations at 250 stations had been collected for 1987 and 1988, and were compared with the LSM results. It was found that land surface models (LSMs) could estimate annual runoff fairly well in general, but, all of the models tend to underestimate the annual runoff in the high latitudes. It was suggested that the underestimation is caused by the observational bias of forcing data, especially for solid precipitation. The aim of this study to investigate the sensitivity of the precipitation gauge correction for the estimation of the global water balance by LSMs, based on a compensated global precipitation dataset. According to the correction, the precipitation increases by 100mm and the runoff increases by 50mm on average in the high latitude of the northern hemisphere. It is found that: 1. The runoff underestimation in the high latitude, which was found in the previous GSWP experiment, was reduced by the gauge correction. 2. After the compensation, global precipitation increased approximately 100 mm/yr and global runoff increase was 40 to 50 mm/yr. 3. Global precipitation and runoff and those on continental basin scale were reasonable from the atmospheric water balance. The bias of runoff underestimation in a high-latitude became smaller than that of the previous GSWP experiment, however, the variance of error didn't become smaller yet. This fact shows that the direction of compensation is proper but it needs still more improvement in estimating more realistic runoff amount.