Quality Control of 11-Year Hourly Rain Gauge Data Over CONUS Based on Radar and Atmospheric Environmental Data

Abstract:

Rain gauge networks have provided primary in situ observations of precipitation for over centuries and are widely used in meteorological and hydrological operations. In the last several decades, many new automated rain gauge networks were deployed to increase the spatial coverage and resolution of the precipitation observations. However, automated gauges are subject to a range of error sources, including clogging by clutter or frozen precipitation, undercatch in strong wind, wind turbulence, evaporative losses, double tipping of the sensor, or signal failure. Quality control of the gauge data has been a challenge, especially at hourly or sub-hourly scales. The current study takes advantage of a high-resolution, radar-based quantitative precipitation estimation (QPE) and develops an automated gauge data quality control (QC) using the radar QPE and atmospheric environmental data. The gauge QC technique compares hourly gauge observations with the radar QPE and identifies five types of potentially erroneous gauge data: 1) stuck gauges; 2) false precipitation; 3) frozen gauges; 4) spurious low outlier; and 5) spurious high outlier. An 11-year dataset (2001-2011) of hourly gauge observations, including 1) Hydrometeorological Automated Data System (HADS); 2) US Climate Reference Network (USCRN) and 3) Automated Surface Observing System (ASOS) networks obtained from the National Climatic Data Center (NCDC), were analyzed through the new gauge QC process. The analysis indicates that during the winter months (Dec-Feb), ~6% of the gauges per hour was in a below freezing environment and would not be able to provide reliable precipitation observations. On average, ~1.5% of gauges per hour were found stuck and ~1% reported false precipitation. Detailed results will be presented at the conference.