A student on Virginia Tech’s meteorology program is working on a project to improve high-impact flash flood warnings in rural Virginia by studying the density of current National Oceanic and Atmospheric Administration (NOAA) rain gauges.
Over the past four years, the region has experienced multiple high-impact flooding events that caused widespread damage and loss of life. However, the defining steep terrain in Appalachia complicates how rainfall is measured and monitored as the NOAA rain gauges have strict placement requirements. They need open, flat land with no nearby trees, which is difficult in a region where there are mountains, valleys and forests.
“There are very few NOAA rain gauges in the region,” said student McKenzie Tate. “For example, we estimated that Grundy, Virginia, is about 60km [37 miles] from the nearest NOAA rain gauge. What’s happening 60km away could be very different from what’s happening in a specific town surrounded by complex terrain.”
The Virginia Tech team evaluated how well radar-based precipitation estimates perform in this complex landscape. During major storms, the National Weather Service relies heavily on radar-driven rainfall estimates. But in mountainous areas, radar beams can be partially blocked or distorted by terrain, increasing the likelihood of error, particularly in communities located far from radar sites.
By comparing national rain gauge records with NEXRAD radar data and multi-radar multi-sensor (MRMS) estimates, the researchers found that results are highly sensitive to the availability of ground-based observations. With fewer gauges available to correct radar-only estimates, errors can become more pronounced during short-duration, high-intensity storms, the type most likely to trigger flash flooding.
“In one case, we noticed a ‘blue bubble’ of lower precipitation around a station, surrounded by higher precipitation estimates,” said Tate. “That suggested the density of stations affects how MRMS integrates observational data. MRMS uses radar and assimilates ground observations together. With fewer stations, the system has less ground data to refine its estimates, leading to generalization in areas farther from gauges.”
The findings highlight a broader challenge: areas that are already vulnerable to flooding often have limited monitoring infrastructure. Expanding ground-based observation networks across Appalachia could improve rainfall estimates, strengthen warning systems, and help protect communities facing increasingly extreme precipitation events.
When doing background research for the study, Tate read in a 2011 study that the central southern Appalachian region experiences some of the highest six-hour precipitation rates in the world. “That makes it even more important to improve forecasting and communication,” Tate added.
Tate is working on the project with Craig Ramseyer, associate professor in the Department of Geography, who said, “McKenzie’s research quantifies the spatial data inequity problems in far Southwest Virginia, specifically analyzing high temporal resolution rain gauges.
“These rain gauges are critically important for flood detection and measurement and yet, are sparsely available to forecasters, making floods in the region hard to diagnose in real time.
“McKenzie’s research serves as a critical building block for future funding proposals to install rain gauges in the towns and cities in Appalachia that are most in need of data for flood detection.”
Related news, Chongqing expands AI-powered weather services to improve warning times.
