Correcting wind-induced bias in solid precipitation measurements in case of limited and uncertain data

Abstract

Automatic precipitation gauges tend to underestimate solid precipitation in the presence of wind. Loss as a function of wind speed is typically evaluated by comparing the gauge with a more accurate measurement made using a double-fence intercomparison reference gauge (DFIR). For small precipitation events, small errors in the observations can induce large errors in the 'catch' ratio, i.e. the ratio of the automatic gauge measurement to the DFIR observation. For this reason, precipitation events of less than 3 mm are typically discarded before performing the regression analysis. This can mean discarding more than 90% of the observations. This paper shows how the method of weighted least squares can be used to perform a regression analysis that can take into account the whole sample to provide a more accurate estimation of the relationship between the catch ratio and the wind speed. This methodology is then used to obtain an adjustment curve for a shielded Geonor T-200B precipitation gauge in Northern Québec. Copyright © 2008 John Wiley & Sons, Ltd and Her Majesty the Queen in right of Canada.