Warning Decision Training Division Storm of the Month (April 2017)


Welcome everyone! Good morning, good afternoon, we have a very
special Storm of the Month for you today. It’s a multi-faceted approach: It’s Already
Flooding Again in Texas and this is presented to you by Trevor Boucher at the Austin-San
Antonio forecast office, Patrick Burke at the Weather Prediction Center and Jen McNatt
at Southern Region Headquarters, so we’ve got all three perspectives. I believe Patrick goes first so it’s all yours
team. Alright, thanks Jami, I hope everybody out
there is doing well and can hear me just fine. I know for as long as Jami’s been there she’s
been expressing this interest in having more case studies involving national centers and
I think we had a great event last year that kind of showcases how we can benefit one another,
both top down and from bottom up and I thank my, my friends there in Texas for helping
me present this case. There’s a story behind the title: �It’s
Already Flooding Again in Texas� Jen McNatt, one of the presenters, was here at WPC last
year for a meeting between the Regional Operations Center representatives in WPC and as she packed
up to go home I remember her saying �you know it’s probably already flooding again
in Texas.� And it seems like it has been that way the last few years with the Southern
Region keeping us here at WPC very, very busy with QPF and flooding issues. So here I was coming in to work on May 31st
last year and looked like we were heading into another multi-day heavy rain event and
this is, again, May and early June in Texas, you might expect this but it has been a very
wet couple of years and was already wet that month so this was on the heels of some flooding
that had already occurred just a week prior. And when I came in that day saw a very, well
what looked to be a very heavy QPF from WPC you can see in the lower left corner there,
that’s a five-day accumulation graphic but at the same time I think that rainfall in
the forecast was spread out over multiple days, there wasn’t any particular day that
really stood out at the time so we had kind of a series of slight risks for excessive
rainfall and there wasn’t maybe the sense of urgency there to have flash flood warnings
in place. But when I see that image of 5 inches of rain
falling over much of the state of Texas and Oklahoma I would expect there would be flash
flood impacts. So something didn’t quite add up there and,
and it was obvious that we needed to look into this. And I think there was this recognition at
all levels from the local offices, from the regional headquarters and at WPC that morning. That yeah this really doesn’t add up, if the
QPF is going to be that heavy and we’ve already had flooding in the last week, we’re going
to have issues here and things were about to get going later that day, we knew there
would be thunderstorms so something had to happen here and it was kind of out of sequence
from what you would normally expect in a typical forecast cycle. But fortunately over the past few years with
the emphasis on collaboration, the meetings that we’ve held with regional representatives,
the practice that we’ve gotten, especially in Southern Region, and in Central, Eastern
region as well, with some of these CWA heavy rain events. We had sort of the relationships in place
to make the phone calls, to bring things together in a large-scale way and by mid-day before
the event really began to unfold, we have this well coordinated flash flood watch that’s
covered multiple CWA s and some of the feedback that we got from the local level suggests
that �yeah the conditions on the ground warrants an upgrade to that risk category�
so we went up to moderate risk. So I’ll give you just a brief overview of
sort of what the weather pattern looked like heading into this multi-day event. I’ve taken a WPC surface chart and satellite
image there and I’ve just contoured a couple of the 500 millibar height contours. So you get the sense that there’s this split
flow regime around a ridge on the West Coast, fairly active northern stream with a wound
up system over the Dakotas and that�s dropping a front down into the Southern Plains early
on the morning on May thirty-first and there’s an outflow boundary even, ahead of the front,
that’s dropped their into southern Oklahoma North Texas. In the southern stream to the south of that,
that splitting ridge, we have a closed low, it’s not particularly deep but it is slow
moving and will approach the Southern Plains over the next few days. I’m showing the sounding that morning from
Fort Worth on the right-hand side, pretty typical for spring, there is instability there,
there’s plenty of moisture, you would expect there would be widespread thunderstorms over
the next few days but I think again there were questions on the mesoscale as to where
exactly is this going to be focused. It wasn’t as clear cut as maybe some of the
more synoptically driven events such as the, the one that happened in March in Louisiana
last year was ,was very blatant in some of the anomaly field, sort of moisture advection
and that kind of thing, this wasn’t quite so obvious. So again just kind of to recap what the inherited
message was in our products across the Weather Service that morning. Here you have this five-day outlook of particularly
heavy rain over a large area of the Southern Plains from WPC and yet the risk areas for
excessive rainfall were all in the slight category at that time. You can see in the lower left-hand side image
that’s some of the antecedent rainfall that had occurred the past seven days, especially
impacting the New Braunfels or San Antonio/Austin CWA and the Houston CWA. So there was particularly concern down there
and at the same time, you know, there are always these other operational issues that
we’re juggling at one office or another. At WPC that morning we were issuing advisories
for post-tropical storm Bonnie, over the Carolinas, which does greatly increase our workload across
multiple desks. So that’s one of the things that we were juggling
and again there was this recognition there are no watches at the time in Texas and Oklahoma
and so we need to start making some calls. I think at this point I’ll hand it over to
Jen McNatt. Thanks Patrick, so that morning the ROC had
been facilitating the briefings for the Texas Department of Emergency Management and also
FEMA region six for several days, because like Patrick talked about, we had had rainfall
and flooding leading up to this particular event. And in the morning of the 31st, I was working
in the ROC and saw in the AWIPS collaboration chat room that San Antonio was looking to
coordinate a large-scale watch for that day because things were coming into place for
this to be a particularly big flooding event. And due to the large nature you saw the QPF
that Patrick showed there and how it covered two states, I had chatted all the offices
and asked if they wanted to do a collaboration call, a quick one in the morning, to talk
about the large-scale watch issuance. And then during that 0930 call to talk about
watches we also discussed agreeing to use WPC QPF to kind of make it easier on collaboration
later in the day and also agreed to reach out to WPC to talk about a QPF collaboration
call and that was scheduled for 1715Z later, well that was about midday. So I�m going to turn it over to Trevor to
kind of talk about, talk through the local operations down at the San Antonio office. Thanks Jen, and yeah so I’m going to not talk
a ton about the actual meteorology but more about our vulnerability here in the Austin/San
Antonio area and the graphic you’re seeing here is actually the seven days prior to this
event, as you see data there on May 31st at 10am. So the past seven days we had two rather major
events over the end of May, one including the over 16 inch rainfall that occurred in
Brenham, which is actually in Houston’s area, but we had significant impacts in our northeastern
counties, as you can see on the on the left there. But just a couple days later we also had a
10-plus inch report via CoCoRaHS in our hill country and a number of other locations that
had multiple inches of rain as well. And so we’ve, in my perspective, at the local
WFO, this was actually my eighth day in a row to work a day shift so I had worked all
of these events already. I had worked the long-term and the short-term
desk so I was issuing warnings and I was part of the grid process. So, as you can see, in the bottom left there,
the total warnings we issued over the past seven days were 69 total warnings, I had a
lot to do with most of them and then the flash flood warnings were 23, so we were busy all
the way up into this event. Just to talk about a little bit of the context
of our vulnerability, and if you’ll click forward here Jen, you’ll see that the, the
location for Medina Lake is right there where the star is located, it was a hundred percent
full as of 3Z on the 31st. And so eight days in a row working, we have
a full lake, high vulnerability, two major events and here comes our third significant
event that we were expecting. And as Pat had mentioned, there, there weren’t
any flash flood watches out yet, mostly because we were expecting the larger scale event not
to unfold until later that evening and then also you know the issue of the uncertainty
where it would actually develop and where was the higher end threat. So we had a level of uncertainty during the
day and then really thinking that the late, the major event would happen later that evening
and so we were expecting to have to issue the flash flood watch that day, typically
in the afternoon but you’ll see that, that changed quickly. So I put TGIT for you know Thank God It’s
Tuesday for me because that was my Friday and WWAMF which is What Was Actually My Friday. So I was very much looking forward to being
off for the next few days after these major events. So this is a look at what the Medina Lake
dam was doing and if you can see it was actually going over it’s spillway. In fact, we were getting a number of phone
calls from both Emergency Manager in Bandera County and Medina County, asking about any
rain not just, you know, multiple inches, any rain that might be moving into the watershed
here that could be impacting this dam. And so there was a particular level of concern
in the hill country, so the previous event that was on the 26th, we had our Emergency
Manager in Bastrop County that actually told us that the whole county is under water. So we had significant vulnerabilities to drive
that point home one more way I just want to take a real quick glance at the, kind of the
course flash flood guidance here, pick out a couple of numbers. You’ll see that at the top there, Bandera,
the one hour is 1.6 and of course this is the worst-case scenario and this is the county
level so we would be doing a little bit more analysis in this but just to pick out some
of the, the numbers here that stand out, less than two inches in an hour in the end of May
we’re absolutely convectively going to get that down here in the PWAT environment that
we were looking at. We had Baylor and Travis County, our two major
metro areas, were only a little bit higher than that. You’ll see also the, what stands out to me,
is the three-hour levels, that they’re not much higher at all. So from our standpoint, significant vulnerability
so walking into a shift was pretty, pretty stressful for me. And moving into the next slide here, I want
to walk you through, sort of how things unfolded. So the previous day, SPC had a marginal for
our western areas and that’s the star there is Medina Lake, as well. Over the course of the normal, normal process
the 6 and 13Z issuances, the marginal threat, the more convective threat, it was farther
east now, including most of our CWA. And then by mid-morning of the day of the
31st, we actually had an upgrade to a slight, so the convective element was increasing too,
and we were actually seeing this in the high-resolution model output as well. And so we were actually looking at, while
initially thinking the major event would occur later that evening, it was, it was thinking
there was, a, becoming more apparent that this would actually begin, maybe as early
as late that morning so stepping into a shift was particularly stressful for me. Hence the title of this next slide, so I’ll
kind of walk you through my day and this was just the morning. 7 a.m. I come in and it’s the shift briefing and
there’s no flash flood watch issued because of the issues that we had mentioned before,
the uncertainty. So we jumped in and looking now at, at seven
to nine a.m. I was going through the typical process you
would go through, look through the models and see what the, you know, the mid-shift
was looking at, look at some of the flash flood guidance, how the rivers were doing,
after such a high vulnerability event we had lots of high streams and stuff and just a
little bit of context about me, I had only been a forecaster for less than a year, you
know, so from my experience made it handling major events like this and you know the, the
fatigue of eight days in a row I, I was getting pretty overwhelmed quickly. And that’s why I put this next point in red,
is that from nine to ten a.m. we are starting to see development, if you look at the radar
here at ten a.m., we’re starting to see convective development beginning already at ten o’clock
in the morning without a flash flood watch, without really communicating that there’s
really any threat during the day. So we had to, I was really starting to jump
in and get really stressed as far as being able to update the, the grids for pop chances,
weather, noticing we don’t have a flash flood watch, high levels of vulnerability and so
it became very important for me to coordinate something with our neighbors and I attempted
to do that in a collaboration chat. And that was, you know, that there’s some
limitations with chat, mainly did people see the chat? You know, what are they going to answer? Are they going to agree? A number of things like that, so what Jen
did, which was particularly important here for somebody with as little experience as
I had at this time, was really helping me out she noticed that I was trying to chat
with everybody and we were trying to do something locally through collaboration chat and that’s
when she jumped in and suggested that we do a collaboration for a watch on a large scale
as many different CWAs were going to have some impacts here and convection was also
occurring in Fort Worth�s area. So and then moving forward here, at 10:22
a.m. we had made our chat to our partners, was really the first time we had introduced
the prospect of a flash flood watch, and it was going to be coming very soon, an immediate
flash flood watch because of the changing atmospheric effects that are going on and
the additional conductive element. And then just 20 minutes later we issued our
first convective product, it was a significant weather advisory that was our very first product
that went out just 20 minutes after the chat saying we’re going to issue a flash flood
watch, it�s not out yet and then at 10:53 we issued the flash flood watch. So I come in at seven and less than four hours
later we’re issuing a flash flood watch for the entire CWA and remember this is in the
morning so we’re not getting the typical afternoon model suit that we would typically be looking
at, this is entirely based off of the high-resolution models that we were getting, namely the Texas
Tech WRF and the HRRR. And then moving forward again, only maybe
30 minutes later, we have an SPC upgrade just like the day one slight risk so the convective
development confidence increases even further, made me very happy that we did the flash flood
watch because our flash flood guidance was most likely going to be exceeded. And then if you look at the time there, 12:05
p.m. for the flash flood warning for Llano County, our first time it came out. And so we, very quickly, from a standpoint
of an inexperienced forecaster jumping in from 7am to the first four hours of my shift
were absolutely, you know, just kind of a nightmare when it comes to what’s going on. You’ll see here at the last point, 12:09 p.m.
is when the MPD comes out, you know, explaining why is it a flash flood watch and the convective
element going on with the rain here. So we were, for from my standpoint, without,
you know, I had a whole lot of support from my lead forecaster, Clad at the time and our
management to jump in and really help us out because we need to get into warning ops right
away and, and also for Jen for jumping and recognizing, you know, what was happening
here and how how maybe region can come in and help facilitate this role here. So I think there’s one more slide which shows
you later on that afternoon what our CWA ended up looking like, with rain right over Medina
Lake of course, and, and then moving into the eastern areas that were high vulnerability. So we had a very busy day as far as warnings
go after that point. So from this, once again, it was just very,
you know, very busy for me on a local level. So I’ll pitch it back up to Pat to tell you
how he went through his Monday. Yeah, thanks Trevor, nice to hear some of
the human factors in the operation area in your talk and we have that going on here at
WPC too. This was my Monday, so where Trevor’s coming
from having been on seven plus days and seeing everything I’ve been off for four plus days
and you kind of, you know, you lose touch in that first day, you’re trying to get your
feet under you and figure out well what are the big issues that I need to deal with. And so one of the ways we do that, some of
the national centers and regional centers will participate in a 745 a.m. stand up briefing
through the headquarters and with the National Operations Center and during that briefing
there’s a slideshow, and I remember after seeing that slideshow, that the WPC Deputy
Director Kathryn Gilbert came by my desk and just sort of asked me �I noticed there was
really heavy QPF but we didn’t have any flash flood watches out,� so she kind of planted
that bug in my ear and that got me thinking, of course at that point I’m everybody’s filing
into the office, we have eight desks, forecasters are coming in, I’m briefing them we’re getting
going and before I get a chance to get on the phone with Southern Region, I end up on
a half hour long, particularly lengthy, interview with Reuters over the Texas flooding that
was already ongoing from the previous rainfall. So there’s a little delay there as soon as
I get off of that call, the Hurricane Center calls and wants to talk about our strategy
for dealing with the post-tropical storm in the Carolinas and so finally about an hour
plus after getting this initial awareness about the flooding events in Texas, I�m
finally able to call and talk to, I can’t remember if it was Jen or somebody else at
the region. Fortunately, this is east coast time, so it’s,
it’s now earlier there in central time and that’s one of the fortunate
advantages of having WPC on the east coast is kind of getting out ahead of some of these
collaborative issues as we arrive early in the day. But you can see this next slide here is a,
an entry from the WPC log that we use in operations at 1210 that afternoon, a lot of what happened
throughout the morning really was at the local and regional level, and WPC after making that
initial contact with Southern Region went back about the business of preparing our products
and trying to get ahead of the game as much as possible with the next model runs. But this log entry you see their Southern
Region call will take place at 1715Z. They’re worried about getting into warning
mode. We may not have our preliminary full QPF package
at that time but we can discuss any trends that we’ve noted in the models and what we
plan to do with the excessive rainfall outlooks. And this idealized timeline that you see along
the bottom of the screen, I think this is one of probably the most important things
that we have to figure out as a Weather Service if we really want collaboration on QPF to
work regularly, to work in a consistent manner, because we really do our best this, this collaboration
timeline there that you see that gives about a two-hour window between the time we issue
our preliminary products and our final products, it’s worked very well for our winter weather
desk over the years. We’ve tried to do that now with QPF the past
couple of years and it’s with variable success, you have offices across four different time
zones, everyone has their own individual needs and workflows and the way your shifts are
structured at a local level and what your partners want. So it can be very difficult to, to serve everyone
including the River Forecast Centers with this one timeline. So that would be a good topic for discussion
actually after today�s talk. The next WPC log entry there, the collaboration
call with Southern Region was went pretty quick and I think that’s because we didn’t
dwell on the details of the QPF, I mean some of these bigger events there is an advantage
there to just making a choice we’re going to load WPC. The collaboration call gives offices an opportunity
to chime in if they have any particular input as far as the models or what they see from
a WPC preliminary graphic but by, by making this choice we can then move on and talk about
messaging and there really is valuable input, I think, probably more so with, with excessive
rainfall outlook,s compared to some of the other National Center products with that value
coming from field up to us. You know, we try to make the best of the gridded
flash flood guidance that the River Forecast Centers provide us but we know that that has
a lot of shortcomings to it and we are plugged into using some of the new output from the
National Water Model, we do look at things like surface runoff for experimental things
coming out of high-resolution models but really there’s no substitute for hearing from Trevor
in San Antonio that �hey the lake is at one hundred percent capacity, the flash flood
guidance is pretty low on a three hour time scale, you know, we think that we fully support
a moderate risk in this kind of situation even despite all the mesoscale uncertainty. And so speaking of the mesoscale, at that
point we have the QPF settled, we’ve decided to go up to moderate risk for excessive but
we don’t stop there, you have this forecast funnel approach that includes the mesoscale
precipitation discussions at WPC and so I think, you know, still trying to get the message
out to all the field. That when you think of WPC, traditionally
and you think of that model diagnostic discussion, that’s very popular, it talks about primarily
global models and the NAM, that kind of thing, there can be this impression that that’s what
WPC is about but we really are plugged into every imaginable high-resolution model run
that there is and we have a desk that is dedicated to this flash flood problem. And so this kind of allows the National Centers
to pay equal attention, deservedly so, to the flash flood problem at times when there
is also severe weather concern. And so during this event, we, we issued 28
of these mesoscale precipitation discussions to kind of keep that information flowing as
you guys are dealing with all the phone calls and all the distractions and warning mode
that happen at the local level. A few other tidbits to wrap up from our perspective
here, we also supported southern Region and Florida offices later in that week for Tropical
Storm Colin, so very busy week. We had a high-risk upgrade that involves just
a couple of offices so we were able to do that over the chat software and just to show
the impacts, this did include the tragic drowning deaths that occurred at Fort Hood, Texas,
I think the day after, all of this collaboration occurred on May 31st. And Mike Eckert from the FAA Command Center
I recall gave us a ring to talk about some of the air traffic around Houston being affected
by the convection. So it’s great to see how far-reaching our
mission is and then as we work together that we can address all of these different things. Thanks Patrick and so to build on that with
some other outcomes. There the day after the particular day we
just discussed there were 31 counties in Texas that declared disasters due to the flooding
and just some of the DSS that was being done the ROC and the RFC and the WFO were coordinating
with the state Division of Emergency Management on a daily basis to help provide them with
information on where to place resources such as search and rescue personnel and equipment
such as swift water rescue boats and, and then all of the localized DSS that was being
provided at all the WSOs and the RFCs within this area too to county and city Emergency
Managers were taking place. And so from this event some of the things
that the three of us kind of thought were some key takeaways was how important it is
to really recognize quickly when to collaborate and when to ramp up the message. You heard from Trevor at the local level and
then from the ROC perspective, kind of trying to help facilitate that communication and
coordination and then from the WPC perspective and I think we need to be nimble on when we
collaborate and how we all help together to ramp up that message and as a result of this
particular event we had a seamless watch issued for that during that warning before the event
started. The moderate risk upgrade, I think, also helped
with the messaging to partners and the key point in that too. that Patrick talked about was that the local
input really informed that upgrade of the excessive rainfall outlook because they were
talking to Emergency Managers like Trevor outlined and know what kind of local effects
are going to happen, so that was really key. And this collaboration really is more successful
now because the relationship between the ROC and WPC has strengthened in order to reach
out rather quickly to each other when we see a need for collaboration or when offices request
collaboration. And really the last one is that you know a
lot of coordination, collaboration on a daily basis happens via chat via AWIPS collaboration
and that’s totally effective but when you get to large-scale events such as this where
you have watches that cover multiple states, then a quick phone call, a five-minute phone
call, can get a lot of discussion and a lot of coordination done a lot faster than 10
offices trying to chat each other and figure out where to put the watch and time lines
of the watch and. and excessive rainfall outlook things like that. So those were our key takeaways and we will
stand by for any questions or discussion.

Leave a Reply

Your email address will not be published. Required fields are marked *