Regional and Mesoscale Meteorology Branch

00:00:00:00 – 00:00:18:11
Speaker 1
At the, WFO in Norman, Oklahoma, and he’ll talk about some, fires that they had recently and some, very interesting, goes our S.R. SLR imagery application. So without further ado, I will turn it over to Todd.

00:00:18:13 – 00:00:43:17
Speaker 2
Hi. Thanks, Dan. Yeah, I only have just, three slides here. Not a lot. I’ve been on, on travel and, having, put together, like, a formal presentation on this yet. But in the room with me, I have, Dave, Andrew, our Mr. Erin Anderson Itto, who’s largely, to credit for most of what I’m going to show you today and then, of of Agonal Holly, one of our meteorological interns.

00:00:43:17 – 00:01:05:14
Speaker 2
And, we’re just going to go through some of the work we did on February 18th, trying to take full advantage of some of the one minute, imagery we’re seeing. And also, we had, goes 15 also was in, rapid scan operations to support fire operations that day across the southern Great Plains.

00:01:05:14 – 00:01:29:21
Speaker 2
But the opening slide here is the Buffalo fire, which occurred in northwest Oklahoma. Just over 17,000, I believe, 17,200 acres consumed in about two and a half to three hours. So, very rapid moving grass fire out there in northwest Oklahoma. The smaller panoramic image there is just, pano I took on my workstation during the fire so you can see the different remote sensing.

00:01:29:29 – 00:01:52:11
Speaker 2
Here we have the the, 3.9 micron imagery on the left. And then, just a radar, from CBN’s advanced Air Force Base radar in northwest Oklahoma. And then the next one is actually a, a, image of GSV, where we were using the red flag threat Index, which happened to be peaked at about a seven on a scale of 0 to 10.

00:01:52:22 – 00:02:10:14
Speaker 2
Right there in a small area of northwest Oklahoma where the fire occurred. And then, of course, just Dr.. On the on the PC there looking at, the smoke plume from the, the buffalo fire.

00:02:10:16 – 00:02:37:23
Speaker 2
All right. So, the next slide, I have shows just, satellite blue visible and, I believe 3.9 micron, it goes 14 one minute data from the, from the buffalo fire. And what I mainly want to talk about is, you know, we knew we were going to have access to this. We, we were during one of the periods of the one minute data that we’ve been, bringing in from the hazardous weather testbed next door.

00:02:37:25 – 00:02:56:03
Speaker 2
And, we knew we were going to have, you know, extremely critical fire conditions on February 18th. So one of the first things, our Mr. Dave said to me when when I walked in the door that morning is today is the day we need to innovate. So I promptly went to talk to Eric and said, Dave wants us to innovate today.

00:02:56:05 – 00:03:42:17
Speaker 2
So the end result of his, innovation was, this tool within a website that simply uses the point tool to identify you manually drop the point onto an identified hotspot when you identify it. And, pops agree. Basically, you have a free text box where you can put in a message, but the point is that it will send a text message to some predetermined, phone numbers, and it will provide the, the lat and lon for the fire also provides the nearest, Oklahoma mesonet observations and then provides a link to the, activity planner from the industry.

00:03:42:19 – 00:04:03:13
Speaker 2
And we use the fire weather parameters for that. I believe we have temperature, RH, wind, wind gust and again, red flag threat index also. And we kind of had to go back and recreate some of this. We didn’t do the best of jobs archiving it all in real time. And, but, I think we set around eight of these text messages that day.

00:04:03:16 – 00:04:31:04
Speaker 2
We set it up in kind of partnership with Oklahoma Forestry. They were on the list and up to receive these notifications, as well as, the duty officer in the Oklahoma State EOC at the state capitol. But, what’d you see on the insert photo here? This is, our fire weather program leader, Scott Corral, in the foreground, actually issuing a fire warning at the request of the local emergency manager for some evacuations in Harper County.

00:04:31:06 – 00:04:59:24
Speaker 2
And then you can see Aaron in the background is working on this tool, but this is an actual screen capture from my phone that shows the text message that went out, minutes after the, the hot spot was detected in Harper County for this particular fire. So some, thoughts on the utility of this tool. Like I said, we used it to, notify those officials of eight, new hotspots during the course of the day.

00:05:00:25 – 00:05:24:18
Speaker 2
Oklahoma forestry, you know, that there’s they’ve still been busy. We still have, fairly critical fire conditions, across the state of Oklahoma. So we haven’t had a, debrief or anything like that with them, other than very informal on the phone at this point. But they’d have indicated that they were receiving some of these notifications and they were able to offer them on to the local jurisdictions prior to nine, 11 calls being received.

00:05:24:20 – 00:05:56:00
Speaker 2
And that’s that’s actually not new. That’s something we’ve been able to do even using just RSA. So as far back as 2011, with some of the fires, that we’ve seen out here on the, on the Southern Plains, but the definitely a service for those folks are very appreciative of have been contacted since by the Oklahoma EOC and they would like to integrate this into some sort of consolidated situation awareness display, along with real time, ground reports coming in from the local jurisdictions.

00:05:56:00 – 00:06:20:06
Speaker 2
So, you know, it’s definitely got their attention. It’s something they want to build upon. So, you know, we don’t really know what the next step is. But, you know, as we get further down that goes our path. I think, the possibilities are endless. And speaking of that, you know, this is something that could not only be used for initial notification of new hotspots developing, but also for tactical support.

00:06:21:08 – 00:06:46:15
Speaker 2
Maybe we could sample trends in the hotspots, you know, certainly more as we get as it goes. Our air. And also not just for satellite derived information, but maybe radar dual pole attributes of fires and smoke points to, just submitted a paper from the Bastrop fire September 2011, looking at at, horizontal convective rolls, interacting with that fire and how it changed its behavior on the ground.

00:06:46:15 – 00:07:07:19
Speaker 2
So there are some real tactical information that we can, you know, just like in the in the near storm environment that we can provide, decision makers on the near fire environment to and this might be a, you know, or the next iteration of this or this tool in general might be a very efficient, way to communicate that information to decision makers very quickly.

00:07:07:22 – 00:07:11:06
Speaker 2
So that’s really all I have.

00:07:11:08 – 00:07:17:05
Speaker 1
Okay. I’ll open the floor for, discussion or questions.

00:07:17:07 – 00:07:49:16
Speaker 2
Yeah. This is, Tim Schmidt. No, I notice, first of all, I’ve never been so glad we got the right sector. Looking at goes 14 one minute date. And, of course, that was made easier by space’s, fire outlook. I guess my comment is, you know, with goes are there is a fire hotspot product, you know, to give you a quantitative answer, although it is not currently envisioned to run on the massive scale, i.e. the one minute, small sectors.

00:07:49:24 – 00:08:00:19
Speaker 2
So a when you find that useful and if that is if the case, does anybody have a plan to maybe to someday make that happen?

00:08:00:21 – 00:08:18:20
Speaker 2
Well, this is Todd. From the operational perspective, I think it would be very useful. You know, the one minute was almost the resolution was almost so high that, you know, you would initially see flickering and then, you know, you kind of waited a few minutes to see if it would be sustained or whether it would take off.

00:08:18:20 – 00:08:45:05
Speaker 2
So some of the early stuff we we relied on, you know, were almost false alarms where there were some hot spot signatures that flickered and that did not really persist or take off. But in general, I think that’s just incredibly valuable information when, you know, because you get these these very fast moving grass fires out here on the plains and, you know, the quicker that they can get on that and prevent it from becoming a very large fire.

00:08:45:05 – 00:09:13:08
Speaker 2
And, the less damage that we’re going to incur. So, you know, if we could get one minute, resolution data permanently with Goes-r, I just can’t imagine why we wouldn’t want to do that. As far as fire detection. Okay. Thanks. So this is Brad Pierce. Known as this as well? I think, you know, the the grass fires is a really interesting issue, particularly the rapid spread.

00:09:13:08 – 00:09:44:27
Speaker 2
How as you look towards some kind of consolidated display for situational analysis, how, for example, would you fold in, you know, ten meter winds or something like that so that you could use this also combined with, smoke imagery to understand where things are likely to go? I think there’s a lot of opportunities within that. I think kind of a blended product we’re doing has similar kind of activity under the JFS, Fire and Smoke initiative.

00:09:44:27 – 00:10:21:01
Speaker 2
And, and Goes-r certainly will be able to bring a lot, particularly for these rapidly evolving fires. Right. I don’t really know if, you know, I don’t really have a good grasp yet on how this could evolve, but yeah. So, I mean, I think you’re right. If you start compositing a lot of different parameters, maybe, you know, ten meter wind, RH, fuel temperature or any of these sort of things that, you know, the sky could really be the limit, where this goes, maybe spread, spread components or, you know, some some, you know, there are a number of parameters that you could use to composite with this type of imagery that could really provide

00:10:21:01 – 00:10:37:21
Speaker 2
a very short term, you know, tactical 1 to 3 hour, piece of guidance for, for decision makers, emergency managers and just the firefighters on the ground.

00:10:37:28 – 00:11:06:12
Speaker 2
This is Chris Schmidt. Do you have access to the current ghost fire detection through your. Buildings really get to know the bill’s 14. I believe we only get, four different bands from those 14. That one minute data. Okay. All right. And that was something we we actually kind of scrambled on the afternoon of the 17th, the afternoon prior to the event, to make sure we could get the 3.9 micron.

00:11:06:12 – 00:11:38:21
Speaker 2
And it was just in time for this. Yeah. This this is Tim Schmidt. I think Chris event was asking if you routinely get the W off of, this operational fire detection product from the operational goes into your ellipse. No we don’t. We usually just operate with 3.9 micron. That would be able to provide you with, the detection and, fire rated of power and size and temperature estimates as well.

00:11:38:24 – 00:12:00:09
Speaker 2
But the power would be no need to be most interested in probably. Right. Give you an idea of intensity. But I know that that’s been an issue with it not necessarily being available to everybody through a website. And it’s been a problem we’ve been chasing for many years now. Chris, is that generated when you when the satellite’s an hour or so as well?

00:12:00:09 – 00:12:20:14
Speaker 2
Now at ops, I don’t know. Okay. And I don’t know that it’s coming. It’s not going to eight weeks from now. I know that it’s like having a day. Control frequency is very important.

00:12:20:16 – 00:12:23:29
Speaker 1
Okay. Any other questions or comments?

00:12:24:01 – 00:12:44:06
Speaker 2
Well, I guess if there are any kind of, or there can be any kind of written summary that this is something that you guys want to see because I, it was a bit interesting to see it because I actually proposed something just like this about five, six years ago. And, using the satellite data and other data and kind of got shot down.

00:12:44:08 – 00:13:04:23
Speaker 2
So it’s good to see that the officers are looking for it. But to get it supported in the long run, there needs to be some sort of documentation trail. From the staff, I guess. Is there going to be one or, you know, we’ll be able point to that and say, hey, this is what they want to use our fires for on it before.

00:13:04:25 – 00:13:24:02
Speaker 2
Yeah, I was very much, thinking ahead, to writing this up in some form or fashion. And I don’t know what avenue to pursue that with yet. It’s so recent after the event and with other business we’ve had since then, I haven’t, I don’t have a plan. Yeah, but yeah, I think we’ll get we’re definitely going to write something up on this and.

00:13:24:04 – 00:13:28:07
Speaker 2
And leave a paper trail. Yeah.

00:13:28:09 – 00:13:53:18
Speaker 2
And Tim Schmidt, if I can comment on that, whatever you have, even if it’s a, a website site or maybe even ultimately there’s visit training. I’m happy. Dude, link to it. As you know, we have, a source. So our page at we try to, you know, we have all the imagery from previous years, but then also interesting cases we link to however, and people are using the data and this would be a great case.

00:13:53:21 – 00:14:01:08
Speaker 2
Okay. Good. This is Brad. I think this would make a really nice bands article.

00:14:01:11 – 00:14:10:08
Speaker 2
That, you know, a short piece on on this particular event. Okay.

00:14:10:10 – 00:14:50:03
Speaker 2
Hold on. This is, Greg Patrick from Southern Region headquarters. This is, this is fantastic stuff, and I appreciate you sharing it. Sharing it with us. Can you, I would just kind of interested in what was your latency in getting the one minute, super rapid scan observation images? And also, could you comment or speculate on fast, the relative the usefulness of one minute versus five minute imagery, you know, because we’ll be we’ll be able to get the one minute, sectors one minute updated sectors with, with goes-r, but the five minute, updated imagery will be more common.

00:14:50:03 – 00:15:11:28
Speaker 2
And, you know, pretty much in every day thing. So can you comment on, on, how you perceive the usefulness and utility of one minute versus five minute? Yeah, absolutely. Greg. The latency was not very bad. I’m thinking it was within 2 to 3 minutes or so, and I didn’t take detailed notes on there, but I think it was within 2 to 3 minutes.

00:15:11:28 – 00:15:35:12
Speaker 2
It was certainly under five minutes. There was a major difference. So between the five minutes, goes 14 and the one minute to goes, I’m sorry, goes 15, five minute and one minute goes 14. We were getting that the five minute data hour or so. It was very, very noisy. And there were a lot of noisy artifacts that were showing up that looked like hot spots.

00:15:35:12 – 00:16:05:12
Speaker 2
So we, you know, we were using the, the, the looping feature, which we expanded to 128 and eight was to and to to verify very quickly in the one minute data if the spots were, were noise that we were seeing in the five minute data, or if it was a real hot spot in the one minute data. But that was very much a very frustrating thing through the day, was the the amount of noisy artifacts that we were seeing in a five minute or so much later.

00:16:05:14 – 00:16:16:18
Speaker 2
Yeah, that one minute data was much, much cleaner, much easier to determine what was real and what wasn’t. Yeah. All right. Thank you.

00:16:16:21 – 00:16:28:04
Speaker 2
It goes 14 is a much, much better instrument as far as that goes. That has been showing a lot of noise.

00:16:28:06 – 00:17:01:16
Speaker 2
Yeah, I guess I have a question to follow up on that, because Tony came with my ear. You know, I, was that a difference between him both working one minute and, the 30 at their timeout for five minutes? I don’t understand the noise. Noise in the picture. Okay. I think there are two, two distinctions here. One, again, if you have multiple views of the same event, then and there’s a again, at the beginning it was talked about this flicker at the hotspot coming up to the detection level.

00:17:01:18 – 00:17:18:26
Speaker 2
If you have multiple looks that in a sense gives you a cleaner signal because you could say, oh, I get it. Maybe, maybe yes, yes yes yes. Now yes it’s good as opposed to getting sampled every five minutes. And you maybe just see that maybe case and then you’re not sure. You would have to then wait another five minutes.

00:17:19:11 – 00:17:44:03
Speaker 2
As far as the different instruments, yes, they all have their different instruments. Signal noise, of course, goes 14 has a much better look since it’s at 105 West instead of 135 West. But again, I do agree. And we’ve looked at this with some of the other fires, California rim, etc.. Well, we had one metadata and it does show unique things that happened between the five minutes.

00:17:44:05 – 00:18:14:19
Speaker 2
It was making a good point. I kind of skipped over that aspect of it. But the fire behavior, you’ll see it when you look at one minute imagery and then slice it every five minutes. You can see flickering because it has become too damn or something has gotten in the way. Briefly. And so that doesn’t make it harder to detect that the instrument noise I was talking about is something that algorithms are good at finding because they can, they can see it more readily than the human eye sometimes based on correlation to it to 11 microns.

00:18:14:19 – 00:18:25:10
Speaker 2
So, but the flickering doesn’t always mean it’s full fire. It could mean it’s real, but, it’s just one know for sure for little. Well.

00:18:25:13 – 00:18:58:19
Speaker 3
And this is burning at ten, I’ll make another comment because sometimes during strong winds here, our responses panel gets a little noisy. So, if I’m understanding it correctly, your your goes 15 or 13, 13 whatever. Which one it was, is coming through the ESPN and the other one was coming through the the internet. So that could be a difference in a little bit of the noise you’re seeing between the two of them.

00:18:58:22 – 00:19:15:27
Speaker 2
Yeah. Our one minute data was not coming in via the internet. Okay. When we were getting it directly parked in from NFL and the hazardous weather test said, I believe we’re looking to our RPO here for, exactly. Direct path.

00:19:15:29 – 00:19:26:27
Speaker 2
I think that pass was via the dam via sera, I think I think either Colorado. Yes, that’s correct. Bill had to call right here though, right?

00:19:27:00 – 00:19:38:09
Speaker 3
Yeah. So not understanding exactly what the noise is in your goes, 15ft. Do you ever do see that sometimes during strong winds that you get noisy pixels.

00:19:38:12 – 00:20:04:00
Speaker 2
This was the worse I have seen it. And our winds that day were were strong. Of course, they were around 30 miles an hour or so with some gusts. But, so that is a possible explanation for why we were seeing the noise. It did seem unusually noisy. Yes. And to the point where it was impacting these operations, because we’re looking for hotspots, which, you know, could easily look like noise early on.

00:20:04:00 – 00:20:09:00
Speaker 3
So yeah. So it could it could have been a combination of all those things.

00:20:09:05 – 00:20:15:08
Speaker 2
Sure.

00:20:15:10 – 00:20:22:22
Speaker 1
Okay. Any other comments or questions.

00:20:23:11 – 00:20:27:26
Speaker 1
Scott Lindstrom, I don’t know if you want to show anything else or do you think we should conclude it here?

00:20:28:00 – 00:20:46:19
Speaker 2
I can show you two things. I have a graph of one minute versus five minutes. Yeah, that’s right. Imagery. And I can show you a burn scar and the .86. So I just want to, This is not this. I hope you’re seeing this on the screen. This is a trace of brightness temperature from Goes 14 one minute data.

00:20:46:22 – 00:21:13:24
Speaker 2
So the blue line is the one minute data. And the red points are what it would look like with the 50 if was just sampled every five minutes. So hopefully you’re seeing the red points come on and off. So when you have five minutes sampling you are missing, some detail. I don’t know if that you could want to call that the flickering, but the change in brightness temperature can be very rapid as a as a fire flares up, especially in strong winds.

00:21:13:24 – 00:21:33:19
Speaker 2
So, you do lose something when you subsample down to five minutes. And I just wanted also to show, this is from a blog post on the same satellite blog from that particular day. So there is a post on this particular fire event. And, Scott Buck Meyer found this image from the 0.86, which of course will be on goes.

00:21:34:01 – 00:22:04:12
Speaker 2
So, you know, post storm, there is a burn scar that you can see here in the one kilometer data in the point at six. It’s very sensitive to recent burn scars. So that’s something you can look at to see where the fire has been. If you if you’re familiar with what the 1.86 looked like the day before, you know, the lakes are showing up very nicely as well, because 0.86 discriminates nicely between land and water, but it also pulls out, recent burn scars.

00:22:04:12 – 00:22:08:13
Speaker 2
So that’s all I really wanted to show you.

00:22:08:15 – 00:22:16:05
Speaker 1
Okay. Any more questions or comments before we conclude?

00:22:16:07 – 00:22:24:14
Speaker 1
Okay. Thanks, everybody. And a special thanks to, Todd for presenting. That was very interesting information. Thanks, everybody, and have a great day.

00:22:24:17 – 00:22:26:01
Speaker 2
Thanks. Yeah. Thanks. So.