Effects of gauge under-catch on precipitation observation and long-term trend estimates in Beijing area
摘要
Among all of the uncertainties of precipitation records, the gauge under catch effect is one of the most important error sources. It not only affects the accuracy of the observational value and area-averaged precipitation calculation, but also leads to a false estimate of the long-term trends of the precipitation for any sites and regions. An accurate evaluation of the gauge under-catch effect is thus of great significance for understanding long-term change of precipitation and its possible causes and impacts. Based on observational data of daily precipitation and daily mean wind speed from 20 meteorological stations in Beijing area during the period 1976-2015, the effects of gauge under-catch on precipitation records and the estimates of the long-term precipitation trends were evaluated, and the urbanization effect on near-surface mean wind speed change was also discussed. The results showed that the mean values of precipitation catch rate ranged from 90% to 95% over last 40 years, but they had a significant upward trend and a heterogeneous spatial distribution with larger catch rates in urban area than in rural area. The urbanization process, which led to a lower wind speed in urban area, was the main reason for the observed difference of the under catch rates between the urban and rural stations especially after 2006. The effects of under catch differed among years and seasons significantly. In the past 40 years, the measured area-averaged annual precipitation was 552.2 mm, while the under catch adjusted area-averaged annual precipitation was 575.3 mm, with an average absolute error of 23.1 mm and an average relative error of 4.0%. The area-averaged precipitation intensity before and after the under catch adjustment were 7.9 mm/d and 8.3 mm/d respectively. The downward trends of the area-averaged annual precipitation for the past 40 years changed from -34.4 mm/10 a for the pre-adjusted data to -37.0 mm/10 a for the aft-adjusted data. Therefore, the original obser