Local random errors in drop-counting rain gauges

This is a continuation of the previous study by Chan and Li (2009) on the performance of drop-counting rain gauges at the Hong Kong International Airport. Three identical drop-counting rain gauges are working at the same time and data over a period of 15 months are considered. Two major issues have been studied in the present work:

(a) An improvement of the quality control of the rainfall data has been adopted, and the resulting data are used to find out if an analytical model could be used to represent the non-parametric estimate of the local standard error of the rainfall data from the three gauges at different accumulation time interval. It turns out that, contrary to the previous study by Chan and Li (2009), it may not be possible to use a single analytical model to fit the local standard errors at all the accumulation time intervals;

(b) A single value of the scaling factor (0.2) is employed in Chan and Li (2009) to calculate the non-parametric estimate of the local standard error of the rainfall data, as adopted from a previous study for tipping bucket rain gauges (Ciach, 2003). An optimum value of the scaling factor is determined in the present paper by minimizing the difference between the local rain gauge error and the non-parametric estimate of the local standard error for all the three gauges. It turns out that the optimum scaling factor is about 0.1.