Weather radar is a vital tool for monitoring severe weather and extreme rainfall. My Master’s Thesis at Florida State University explored the use of early NEXRAD radar data for tracking landfalling hurricanes, and I teach a Radar/Mesoscale Meteorology course at the University of Georgia. As such, a Tweet by Cornell University meteorology student Jack Sillin caught my eye this week as a meteorologist, an African American scientist, and someone who grew up in a rural community. Sillin tweeted the graphic below and said, “Following severe storms through the Southeast this week, it’s been hard not to notice some large gaps in radar coverage of the area.” He went on to suggest that many of those gaps are in areas where significant Black populations live. This set off a debate on Twitter. Of course it did -It’s Twitter and race was mentioned. Here’s my perspective.

First off, Sillin’s observation is a fair one. However, people are very uncomfortable talking about race and often exhibit knee-jerk reactions when it is applied in such situations. From personal experiences, this is something very familiar to me, and I could write a book about them. I often watch as people squirm and contort themselves when a problem is mentioned that is race-specific. I saw it in Jack Sillin’s Tweet thread on this topic. When I saw Sillin’s Tweet, there was nothing in my interpretation that assumed he meant that other demographics or races in radar gaps were less important. I also did not take from his Tweet that this was deliberately done to oppress a group of people. Stephen Strader is a Professor at Villanova University that studies intersections of tornadoes and geography. His Tweet sums things up perfectly, “Turns out a lot of people don’t understand the historical relationships between minority populations, rural areas, and poverty across the Southeast…Two things can be true. Urban dwellers and rural dwellers can both be vulnerable.”

What Sillin revealed is a convergence of science, geography and history that is worth a discussion. The “Black Belt” is a region of the South that has its roots in the era of slavery. According to the Southern Legislative Conference website, the Black Belt is, “a string of counties that stretches from east Texas, through the deep South, and up into eastern Virginia…. It is the largest, poorest, most rural region in the country.” It is very apparent in Sillin’s map. Sillin told me in a message, “Many storms (in the South this week) were located far from radar sites and thus it was tougher to assess low-level rotation.” This can make nowcasting of storm risks a challenge. Sillin went on to say, “the gaps are still frustrating, impactful, and consistent with a long legacy of underinvestment in these communities.”

Sillin’s point about disparities in weather radar coverage are not unique. A Five Thirty Eight wrote an article noting contemporary health and death patterns in the Black Belt are disproportionate and likely have roots in slavery and a segregated society that only ended several decades ago. It still blows my mind that my parents lived in a segregated society, and they are not that old. These things are not “race cards.” They are facts that still shape society today as uncomfortable as it is to talk about.

The 2017 Five Thirty Eight article points out that, “Maps of the modern plagues of health disparities — rural hospital closings, medical provider shortages, poor education outcomes, poverty and mortality — all glow along this Southern corridor.” The article also mentioned that other parts of the U.S., including significant Native American regions had similar health disparities.

Sillin’s map above shows the relationship between population density and radar coverage. This suggests that less populated areas are likely to experience coverage gaps. In the South, this will bring a large Black population into play. In the West, it is likely to disproportionately impact Hispanic or Native American populations. Of course, these gaps will affect White populations in these regions too. It is important to shatter the narrative that highlighting one group’s challenge minimizes another group, but that is an article for a different day.

Radar coverage has long been discussed within the meteorological community. Radars work by transmitting pulses of microwave energy into the storm systems. The radars typically scan a volume over a series of incremental elevation angle changes. This means that the radar (and rotation-detecting capacity) begins to degrade at distances further away from the radar. As such, a dense network is required for adequate coverage. A good case study in gap-filling is the weather radar at the University of Louisiana – Monroe. Before this system went in around 2016, the closest radars were in Shreveport, Louisiana and Jackson, Mississippi. Officials now credit the new radar for increasing severe storm monitoring capacity in the region.

In 2020, the National Weather Service (NWS) reported to Congress that weather radar gaps are not hindering its warning accuracy, according to the Washington Post. This assessment was in response to a Congressional request in 2017. However, critics point to radar gaps such as those in North Carolina and the Pacific Northwest as very problematic. NBC Charlotte Chief Meteorologist Brad Panovich told Andrew Freedman of the Washington Post that the NWS report itself was “full of holes.”

For the Southeast, such issues are already compounded by a 2020 Carnegie Mellon University study that found people living in the Southeast are burdened by several misconceptions about tornado risks. According to a university press release, the study in the journal Weather, Climate, and Society revealed that, “most people are more familiar with the risk associated with Great Plains tornadoes, leaving residents in the southeastern United States vulnerable to increased storm risk.” Storms in the Southeast can have very different attributes (nocturnal, quasi-linear convective systems, secondary peak in late Fall) and socio-geographic considerations (more trees, dense population, less resilient structures). Many tornadoes in the Southeast are on the weaker end of the spectrum too. Even the aforementioned NOAA report (found here) acknowledged some warning accuracy degradation for EF-0 to EF-2 storms due radar coverage or beam issues.

Sillin told me, “To be honest the map could and should be improved and the analysis is far from complete but as you know the underlying concern stands.” Sillin’s analysis focused on where the beam is below 6,000 feet above the ground since this is where a storm’s low-level mesocyclone is located and thus important for detecting tornadoes. He overlayed coverage data from NOAA’s NCEI with demographic data from ESRI. While different methodologies might show differences in coverage, the bottom line is that coverage gaps exist irrespective of who lives there. We even have one in parts of the area near the University of Georgia. Let’s fix them.