Improved Quality Control of Mesonet Observations in the Real-Time Mesoscale Analysis System

The Real Time Mesoscale Analysis (RTMA)/UnRestricted Mesoscale Analysis (URMA) system generates gridded surface analyses for situational awareness and serves as an ‘analysis of record’ for use in bias correction, verification and calibration within the National Weather Service. Observations used in the RTMA/URMA system come from a variety of sources, including mesonet observations across the country provided through MADIS. While mesonet observations often provide important observations where traditional observing systems cannot, lack of consistent siting, equipment, maintenance and documentation/metadata standards often pose a significant quality control challenge. This challenge is especially great when dealing with crowd-sourced mesonet observations.

In RTMA/URMA, land surface observations are generally divided into two categories for processing and assimilation purposes: METAR and non-METAR. This fails to account for the variation of observation quality or representativeness between and within various mesonets. Previous attempts to derive station quality based on observation vs. background statistics have increased the number of observations being used, but have often resulted in larger observation vs. analysis differences at highly-trusted sites. While METARs and observations at airports are of great importance, especially for aviation, the mesonets located in other locales are also representative of their unique regimes (e.g. urban, forested, etc.) or may be located at a site of particular interest. The challenge is especially great when analyzing wind, as wind observations are more easily modified by highly localized settings and sensor setup than pressure, moisture or temperature observations. The presentation will cover recent and ongoing efforts on improving the use of this valuable, yet challenging data set.