The writer is a professor of atmospheric science in the department of geosciences at Texas Tech University and an affiliate of the university’s National Wind Institute.
Tornado Alley covers some 367,000 square miles of the central United States, from Texas to North Dakota and from Iowa to Wyoming. During the spring seasons of 2019 and 2020, this is where you’ll find me, collecting data on storms that can spell danger, destruction and even death.
I’m an atmospheric scientist at Texas Tech University, and I took our research group to the Great Plains in mid-May to join with about 50 other scientists from across the country on a pioneering study called TORUS — Targeted Observation by Radars and Unmanned Aircraft Systems of Supercells.
Our goal is simple: to save lives. Between 2004 and 2018, tornadoes caused 1,237 deaths in the United States. Our means: to better understand the phenomenon and, in turn, improve the accuracy of forecasting.
Recent research has shown that about 75% of tornado warnings turn out to be false alarms, due in part to the inability to operationally observe the lowest portions of these supercells, which are the parent storms for most large tornadoes. TORUS brings together a unique ensemble of field equipment to give us an unprecedented look at the boundaries produced by supercell thunderstorms and how they may influence the development of tornadoes. TORUS enables us to target specific areas within a storm that numerical models and theory tell us are critical to this process.
A list of the big-ticket technology we’re deploying includes the Texas Tech Ka-band Mobile Doppler Radars; unmanned aircraft from the University of Colorado-Boulder; surface thermodynamic instruments from the University of Nebraska-Lincoln and the University of Oklahoma; radar instrumentation from the National Severe Storms Laboratory; and the WP-3D Orion “Hurricane Hunter” aircraft from the National Oceanic and Atmospheric Administration. All of these assets work together to obtain key measurements of the changes in temperature, water vapor and wind that are believed to be important for tornado genesis.
Decades of research have contributed a base of knowledge that informs the placement of our measurements. One of the more recent methods is that of ensemble sensitivity analysis, an approach that has precedent in forecasting on the broader scale of weather systems but is just starting to be used at the scale of individual storms.
An ensemble sensitivity analysis starts with a numerical simulation of a storm where we’ve disturbed the initial state just a bit, representing the uncertainty of that state. For example, we might choose to increase temperature up or down, or adjust the wind a little faster or slower.
The computer then takes the differences in that initial state and lets them go forward in time to see how much growth occurs. This process allows us to associate the uncertainty in a specific outcome at the end of a timeline (e.g., a tornado) with what was changed initially. It helps us know what specific types of features to look for, and where. Based on this information, we know the preferred areas to send an aircraft, radar or weather balloon to make measurements that will yield a more accurate forecast ultimately.
Our Ka-band Mobile Doppler Radars show fine scales of motion, measured in meters, at a range of 3 or 4 miles. They are primarily tasked with identifying horizontal and vertical spin associated with outflow boundaries, the leading edge of cold air that spreads out from the thunderstorm. This data, combined with radar observations from the ground-based NOAA X-Pol (NOXP) platform and the WP-3D Hurricane Hunter at 4,000 feet above ground level, allow us a holistic picture of the storm’s structure.
Given the sometimes elusive nature of these tornado-producing storms, it’s not always easy to deploy our instruments at the optimum position with time to collect the data we need. Sometimes storms don’t evolve the way we expect. Sometimes storms come at night when it’s too dangerous to work. Sometimes we miss the forecast and end up in the wrong state.
But when our forecasting is accurate, our road network has many options, and we have an unobstructed view of the horizon with our radars, we’re ready to put our game plan and our equipment into action.
A year from now, when the TORUS fieldwork is complete, we’ll have collected two months’ worth of measurements: mid-May through mid-June 2019 and mid-May through mid-June 2020, followed by a period of thorough analysis.