Quantifying Precipitation Undercatch in a Semi-arid Watershed in Southeastern Arizona

The observed difference in precipitation measured at above ground level (AGL) and ground-surface (PIT) rain gauges is referred to as wind-induced undercatch (U). Quantification of U is important to accurately assess the water balance and eco-hydrologic response of watersheds and for the modeling of precipitation driven processes. U is a well-known phenomenon having been documented for over one hundred years. Neff (1977), among many others, provides historical perspective on the "Jevons" effect, the increase in U with increasing height of the rain gauge above the earth's surface. U is primarily an effect of wind on precipitation whereby wind and precipitation particles interact such that U increases with increasing wind velocity and increases with smaller and lighter particles, liquid and solid. In recent decades much research on U has been undertaken in field, laboratory, and numeric modeling studies in the U.S. and Europe (e.g. Sieck et al. 2007). Much variability of U is exhibited by years, seasons and storm events. The Walnut Gulch Experimental Watershed and Long Term Agro-ecosystem Research (LTAR) site located in southeastern Arizona has been measuring precipitation at a AGL and PIT rain gauge, wind profiles, and drop size distribution for the period 2010-2015. Our results show that the cumulative precipitation difference between AGL and PIT average 6% for the six year period, but vary from 1% to 12% annually and more so seasonally. Although winter (Nov 1 - Mar 31) has greater U expressed as percentage, more than 2/3 of the total U amount occurs in summer (Jun 15-Oct 15), in the same proportion as seasonal precipitation. Regression estimated event U is greater than daily and cumulative, but may be much greater for individual storm events. The undercatch amount is linearly related to storm event intensity, increasing with increasing intensity, but the U percentage is non-linearly related and increases with decreasing intensity. In agreement with previous studies, U percentage is greater for shorter intervals, greater in winter during non-convective events than summer convective events at low intensities, and greater for faster wind speeds. Similar results are found for U amount for winter events and for wind speeds, but U amount is greater for longer intervals.