Drones: The Future of Warnings and Damage Rating?

[Joey Prom]

Hey All,
I am interested to get your thoughts on the topic, whether it is a critique/suggestions to my idea described below, or other ways to best utilize recent advances in technology to better our understanding of storms/tornados, and to increase the NWS ability to get up-to-date information on a storm and its intensity.
I have been stuck on this idea of using a small, rapidly deployable fleet of drones to track severe weather. All you need is 1 or 2 trucks in the field to service and launch the drones, then they can be piloted back at some remote, air conditioned location, via satellite uplink from one of the trucks, or even something like the Starlink network. They would fly well ahead (or behind) of the possible tornado, then use their high angle to "see above the trees" and gather accurate windspeed data using small radar or possibly lidar. A low cost option might be to just use a high resolution high speed camera with good zoom, and just track debris velocity, then use momentum equations, estimating debris mass, and compute the "true" windspeed. This data could then be independently analyzed and compared to the damage surveys. Additionally, if the system can be proven to work and if it can be proven to be low cost and reliable, the data could be directly relayed to the closest NWS office, to assist with warning wording and give a blow by blow account of the tornados intensity. From an engineering standpoint (my specialty) I think the tech is there. It just needs to be utilized.
I think the best type of drone would be a fixed wing vehicle, with 2 propellers, for redundancy. Basically a scaled down version of the a10 Warthog. Fixed wing to increase the range/orbit time with the storm, and to provide enough velocity to stay out in front of the storm. I have not run any numbers, but I think a good design goal would be a UAV with a flight time of ~1 hour at a speed of 60mph, with the ability to fly at up to 80 mph if needed. The craft would have to be stable in crosswind and able to withstand hail up to lets say 1.5". I think its unrealistic to design it to withstand substantial downdraft winds (ie the RFD) or to withstand sig-severe hail, so the OP would be to try to avoid those areas of the storm. That should be much less of a problem than for a ground based chaser, as a UAV is not affected by road grid or geography. Anyways, I think this is a really interesting topic and I hope this sparks an insightful and creative discussion on what advances in technology can do for the weather community, specifically from a research and warning perspective. Excited to hear what yall have to say!

 

[Warren Faidley]

Good thought but there are some serious issues. You would still have to involve ATC with any drone large enough to withstand heavy winds or long distances. Even news station helicopters have some issues with extreme wind. What if winds or a malfunction crashes a drone into a home or car? How do you deploy the drones to risk areas? What do you do with multiple storms over a large area? Another problem would be line of sight for low level flying, required to see / confirm many tornadoes. The cost vs. benefit would likely not make the risk and cost justified, especially when radar has advanced to the point of allowing radar issued warnings. I do believe drones have been deployed for research.

 

[Joey Prom]

Warren, thanks for the response. Here is my current thinking on those points you raised.
I believe that it would be possible to program some kind of fail safe mode into the UAV's if there is a loss of communication with the operator. (The drones will be controlled remotely, but still manually). A code could analyze the location and velocity of "hazard areas" and if signal was lost or a mechanical problem detected the UAV could avoid those areas and go into a holding pattern until signal is reestablished. Or it could just land. It is mostly a software issue. These things could be made as complicated as you want, but that comes with its own problems and $$$. I am not sure if there is off the shelf UAV autopilot technology. As for the likelihood of crashes that is definitely a concern. The system could be verified in more rural areas first, where the risk of impact is less. Or possibly a greater distance from the storm would have to be maintained in more populated areas.

Drone deployment I see as being from some sort of carrier vehicle. The drones would be driven to a location, then launched. The service vehicle would then pace the storm from a safe distance, but still being close enough to recover the vehicle if there were an issue.

I see the system as initially being more of an academic/research focus, so dealing with multiple storms would be similar to how we chasers deal with it, pick a storm, follow it, then move on to another. The caveat I guess is that while UAV's don't care about road grid, they would probably not be very good at core punching. If the drones are small and easy to deploy then they could just return to the service truck and move on to the other storm.

I guess the trade descision would be: do we want a big drone with long range, that just stays flying the whole time, has all the bells and whistles, but costs a lot of $$$ and thus wouldn't want to risk it. Or do we have multiple smaller drones, where we just launch, see storm, launch another, and bring the first one back to swap out the batteries.


The low level flying would definitely be a problem when dealing with low lcl/high moisture environments. Especially if the sensors were just visual based. But if it was radar then it can see through rain.

Conventional doppler can detect storm rotation, but it cannot "see" tornados as easily because they are so low to the ground and doppler s so widly spaced. Likewise, while DOW trucks can see tornados, they also struggle to see the lowest parts, which are the parts directly affecting humanity. They are also bound to the road grid.

Thanks again for the comments. Not trying to be long winded, just sounding out the idea.

 

[Jeff Duda]

I largely agree with Warren's response. I think you would have a really hard time getting this to work reliably, although I do think a proof-of-concept trial would be worthwhile and I think this method would work in special situations. But the big and common problems I see are:

• storm coverage/competition: getting a crew to position itself and the drones in relatively clear air between storms on days where storms or precip are densely packed (a good example was in MS/AL this past Sunday)
• Getting a clear shot of the tornado itself. Lidar beams would attenuate too much to get any meaningful data on any tornado that is at all rain-wrapped visually, and possibly even some that are seemingly clear to sight, but have a small number of very large hydrometeors orbiting them.

 

[Joey Prom]

On a related note, is dropping sensor pods in the path of tornados considered useful anymore? Especially given what we know about large tornados concentrating the majority of their vorticity into erratic subvortexes. Would one or a small number of sensors at ground level glean any important information? Lets presume for this question that there is no human risk to placing pods (ie placed by a drone) and that the pods can also be placed essentially anywhere you would want them, not just along roads. Is there any point to having pods be directly impacted by tornados? What about gathering atmospheric data such as ground level winds and humidity in the immediate vicinity of an updraft?

 

[Jeff Duda]

On a related note, is dropping sensor pods in the path of tornados considered useful anymore? Especially given what we know about large tornados concentrating the majority of their vorticity into erratic subvortexes. Would one or a small number of sensors at ground level glean any important information? Lets presume for this question that there is no human risk to placing pods (ie placed by a drone) and that the pods can also be placed essentially anywhere you would want them, not just along roads. Is there any point to having pods be directly impacted by tornados? What about gathering atmospheric data such as ground level winds and humidity in the immediate vicinity of an updraft?

It certainly doesn't hurt to get more data points that come from pods. But putting a pod in a tornado only gives you one piece of information on tornado structure, whereas science has a desire to know more about the 3D structure of a tornado. There have been past measurements of surface pressure and winds, which is good. But that's about all you're gonna get.

 

[James K]

There is Project TORUS where they are using drones (and multiple other sources) to study tornadoes...

 

[rdale]

Not this year It was canceled.

 

[Chris Gullikson]

One would need a BVLOS Waiver from the FAA to fly a drone without visual reference. They are hard to come by.

https://www.faa.gov/uas/resources/events_calendar/archive/2019_uas_symposium/media/How_To_Get_Approval_to_Fly_BVLOS-Part_107.pdf