Chantal floods North Carolina further underscoring the need to monitor forecast adjustments ahead of events

In brief: Tropical Storm Chantal’s remnants brought record flooding to parts of North Carolina overnight. It will exit off to the north and east later today. More heavy rain is causing flooding issues in Texas again. And Saharan dust is choking the Atlantic for now. For those that rely on email subscriptions to the site, please see the bottom for changes taking shape today.

Note: Most of the data in these posts originates from NOAA and NWS. Many of the taxpayer-funded forecasting tools described below come from NOAA-led research from research institutes that will have their funding eliminated in the current proposed 2026 budget. Access to these tools to inform and protect lives and property would not be possible without NOAA’s work and continuous research efforts.

North Carolina Chantal flooding

Overnight, rains pounded portions of North Carolina and southern Virginia.

Rain totals ending this morning from the last 48 hours show as much as 7 to 10 inches parts of interior North Carolina between the Triangle and Triad. (NOAA NSSL)

The Eno and Haw Rivers in North Carolina are in record flooding territory as a result of the rains. The Haw River’s records go back to the 1970s and the record level of about 22.5 feet was set earlier this morning southwest of Durham near Bynum, NC.

A preliminary record flood on the Haw River southwest of Durham. (NOAA)

The Eno River just northwest of Durham absolutely obliterated the previous record by nearly 10 feet. Records appear to only go back to the 1980s, but there’s no questioning that this was a very impressive new record.

The Eno River broke its previous flood of record by nearly 10 feet outside Durham, NC. (NOAA)

This was another instance where people were left scrambling in the middle of the night to get to higher ground. And it’s another instance where warnings may have been meteorologically adequate but receiving them was tougher. This topic is going to justifiably take up a lot of oxygen in the coming days.

Much like the flooding in Texas, this flooding in North Carolina while somewhat predictable escalated very late in the game forecast-wise. Compare the 48 hour rainfall forecasts using the probability matched mean product from NOAA’s HREF model below. The 00z run, what I was looking at yesterday while writing my post showed minimal risk of significant flooding rains. The 12z run, which was available a couple hours later? Much more adequate.

NOAA’s HREF model caught onto the rains late in the game but once it did, it did a very good job. (NOAA SPC)

The HRRR model also caught on late in the game too, though its performance actually worsened somewhat as the day progressed.

I think that the takeaway here is we had another event that showed us how model guidance in flash flood situations needs to be monitored religiously leading up to the event. If you are an end user of weather forecast information in Texas or North Carolina and you assumed the forecast was settled by 9 AM on the day of the nighttime event that caused historic flooding, you missed out on a ton of critically valuable information. Weather forecasting is so much better in 2025 than it ever has been. Our tools rarely miss something major at the end of the day. But they don’t always catch on 2 or 3 days ahead of time. But if you happen to use weather information to conduct your work or business or for public safety, it is absolutely critical that you monitor things closely leading up to any weather event. We can’t just assume anymore that it will be all’s well that ends well. And we can’t just check out on any given day.

Texas flooding continues

Another round of heavy rain impacted portions of Texas overnight, with upwards of 4 to 8 inches of rain falling in some small areas west of Killeen, farther north than the weekend’s rainfall.

Rainfall estimates for the last 24 hours ending this morning. (NOAA NSSL)

No major incidents have been reported yet, but flooding did occur in the Killeen area.

Rains are expected to continue through the day in spots. The overall weather pattern responsible for a lot of the rain and flooding is beginning to slowly unwind though, so as the week progresses, storminess should begin to ease up some.

Moderate risk (level 3/4) of flash flooding exists north and west of Kerrville in much of Hill Country today. (NOAA WPC)

Still the moderate risk for excessive rainfall is in place, as we still have some ingredients available for hefty rainfall. It is possible that heavy rains will disrupt recovery efforts, and all people in the region responding to Thursday night’s disaster should remain on guard for heavy rainfall and quickly changing conditions today.

Tropical Atlantic

Quiet in the Atlantic, with Chantal’s remnants continuing to lift northward up the Eastern Seaboard. Those will exit off the coast of New Jersey today and out to sea. Locally heavy downpours could occur in spots from the Chesapeake Bay into the Philadelphia and Trenton areas. A slight risk (2/4) for heavy rainfall exists there.

Forecast rain totals in excess of 4 inches are possible in Delmarva today through Wednesday from Chantal’s remnants and a humid, warm weather pattern. (Pivotal Weather)

Beyond Chantal I think the Atlantic looks quiet this week. We have a lot of Saharan dust in place, and that should help suppress any development. After this week, the next area to watch may be in the northeastern Gulf of Mexico next week. A handful of ensemble members and AI ensemble members from various model suites indicate some potential there in about 10 days or so. Nothing to worry about at this time.

Saharan dust is choking most of the Atlantic basin right now, which should help limit any tropical concerns over the next 7 to 10 days there. (University of Wisconsin Cooperative Institute for Meteorological Satellite Studies)

Last note, our housekeeping update:

We will be migrating our email newsletter over to the Substack platform today. If you currently subscribe to The Eyewall via email, your email should roll over to our Substack without you having to do anything. We understand a few people have various concerns about the Substack platform, and some of the content posted there by other writers. If you decide to unsubscribe from emails, we understand. But the benefits of this route outweigh the potential costs. The primary issue is that WordPress caps our email subscribers, and the process to manage an account is not user-friendly. For a site with a limited budget such as ours, this is the best way to ensure the continued delivery of an email newsletter with our content. We will still continue to post here at The Eyewall.com, but updates will not be sent via email from here.

Also: Nothing at all is changing with Space City Weather. So for our Houston readers, our site, email, and annual fundraiser is not changing! This only applies to The Eyewall.

We hope that this will offer us the flexibility to do more things, offer more options, more content, and a better overall user experience as time goes on. Thanks for your support!

Chantal makes landfall near Myrtle Beach and now moving inland as a rainmaker

In brief: After yesterday’s post on Texas, we wanted to raise a couple additional points after some reactions. Plus, TS Chantal made landfall near Myrtle Beach and is now a flooding threat for North Carolina.

Texas comments

There remains a lot to unpack still after yesterday. The reaction to our post covered a rather broad spectrum, and I just want to emphasize a couple points.

1.) We have been on record as saying the proposed budget cuts to NOAA would be disastrous. We’ve had people throwing that back at us and saying “Wait a minute…” The reality is that there are a handful of key ways the budget cuts will impact offices, alerts, and warnings. First, the primary current cuts have led to understaffing in many offices. Relative to other offices, the NWS office in San Antonio/Austin is fairly well off. For example, they currently have seven more full-time staff members than the neighboring Houston office.

They are missing the warning coordination meteorologist, who primarily focuses on building understanding of warnings on what we call “blue sky” days. That work gets done before tragedies occur. That position has only been vacant for a couple months, so it likely did not play a role here. But it will if it is not filled in the coming weeks.

But the office has most of their meteorologist positions filled, as well as most of their hydrologists. From the warning perspective in real time, this plus the timeliness of issued warnings is why I don’t believe staffing was an issue here.

2.) Weather balloon launch cuts probably had little to no impact on this specific forecast. Again, I want to emphasize that the launch of the evening balloon at Del Rio was crucial to understanding the potential severity of the event, and this emphasizes that fully funding and staffing NWS offices to be able to launch balloons is critically important. The takeaway here is that funding NOAA leads to better understanding of unfolding weather, forecast weather, and real-time decision-making.

3.) Weather model forecasts were imperfect ahead of time, but as the event closed in, they performed as designed and did better. Why is this? Because of NOAA-led research. The HRRR and HREF models have been instrumental in improving our understanding of localized (or mesoscale) forecasts like this. We have a long way to go to perfect those forecasts, but they more often than not do their job in guiding the horses (meteorologists) to water. How will we improve those guidance tools more? By fully funding NOAA research. How will we reduce false alarms during potential high impact weather? By fully funding NOAA research. It comes back to that in the end.

4.) We need to understand the limits of predictability in these scenarios. Models often tell us about the potential for a high-end event. But in Texas in particular the specific placement of a high-end event is extremely tough to pin down. A few miles makes a world of difference, and I think it’s evident that we need to build some pad into rainfall forecasts beyond what we’re currently doing. These are works that are accomplished by having NWS meteorologists working together with local emergency managers to help improve knowledge and refine trigger points for warnings. Again, this points back to fully funding and staffing NOAA and NWS offices to accomplish this critical work. The private sector cannot fill this gap at all.

For the families dealing with unimaginable tragedy right now, none of this matters. They deserve time, privacy, and space to grieve. But for those not directly impacted, it’s important for us to push this dialogue to try to play a small part in our broader understanding of the problems and realities. And it’s important to do it immediately because those on the outside have short memories. But I again want to share our condolences to all those affected in some way.

Tropical Storm Chantal

Chantal made landfall earlier this morning near Myrtle Beach as a fairly spicy little tropical storm with winds of about 50 mph. We had a handful of gusts to 48, 54, and 56 mph between Bald Head Island in North Carolina and Murrells Inlet just south of Myrtle Beach, SC.

Chantal is rolling inland across North Carolina this morning. (Tropical Tidbits)

The main threat going forward from Chantal will be heavy rainfall. Flood Watches are now posted for the Piedmont in North Carolina and include Raleigh, Durham, and the Triad, as well as Fayetteville and Goldsboro to the south and east. Model forecasts suggest an average of 1 to 4 inches of rain, but depending on exactly how quickly Chantal pushes northward and weakens, there could be isolated pockets of 4 to 6 inches in spots, leading to some more serious flash flooding.

A slight risk (2/4) for excessive rainfall and flash flooding is posted for much of central North Carolina. (NOAA WPC)

Given what has happened lately, I think it would be important for folks in central North Carolina, particularly in flood-prone areas to be acutely aware of conditions through the day.

Chantal will dissipate over North Carolina later today, but some locally heavy rainfall could extend north into parts of Virginia, Delmarva, or even southern New Jersey tonight and tomorrow.

We don’t see anything else on the horizon right now in the Atlantic, but as always we’ll keep watching.

Housekeeping reminder

A lot of you have reached out to request ways to support our work beyond just reading, and we have (to this point) resisted that, other than various small sponsorships last hurricane season. And we are grateful for the interest many potential sponsors have shown this year too! However, we are going to try something else this year: subscriptions.

To be abundantly clear, our daily content will remain free, never paywalled, and that is not going to change as long as we operate this site. So, forever. This year, however, we will give readers who wish to support the site the option to do so at a modest cost, as well as a premium option for people who want a little bit more access. We will have more information on this soon.

The first step in this process is to migrate our email newsletter over to the Substack platform. If you currently subscribe to The Eyewall via email, your email should roll over to our Substack without you having to do anything. We understand a few people have various concerns about the Substack platform, and some of the content posted there by other writers. If you decide to unsubscribe from emails, we understand. But the benefits of this route outweigh the potential costs. The primary issue is that WordPress caps our email subscribers, and the process to manage an account is not user-friendly. For a site with a limited budget such as ours, this is the best way to ensure the continued delivery of an email newsletter with our content. We will still continue to post here at The Eyewall.com, but updates will not be sent via email from here.

Also: Nothing at all is changing with Space City Weather. So for our Houston readers, our site, email, and annual fundraiser is not changing! This only applies to The Eyewall.

Our target date for this migration of The Eyewall newsletter to Substack is tomorrow, Monday, July 7th. We hope that this will offer us the flexibility to do more things, offer more options, more content, and a better overall user experience as time goes on. Thanks for your support!

Making sense of the weather that led to a horrible Texas flooding tragedy, plus Tropical Storm Chantal

In brief: Invest 92L became Tropical Depression 3 yesterday and is now Tropical Storm Chantal. It’s en route to the Carolinas. We take a detailed look at what occurred in Texas Thursday night and Friday morning.

Note: Most of the data in these posts originates from NOAA and NWS. Many of the taxpayer-funded forecasting tools described below come from NOAA-led research from research institutes that will have their funding eliminated in the current proposed 2026 budget. Access to these tools to inform and protect lives and property would not be possible without NOAA’s work and continuous research efforts.

Tropical Storm Chantal

Hot off the presses this morning, Tropical Depression 3 is now Tropical Storm Chantal, giving us a 3 for 3 conversion rate of depressions to storms so far this season.

(NOAA/NHC)

Chantal’s forecast has not changed a whole heck of a lot since yesterday, as impacts will remain primarily heavy rain and localized flooding, as well as rip currents on the beaches. Please exercise caution, particularly on the Carolina coasts and a little farther north. Chantal will remain over fairly warm water for the next 18 to 24 hours, so it may try to strengthen a bit further. Tropical storm warnings include the upper South Carolina coast, including Myrtle Beach, as well as extreme southeast North Carolina.

Chantal is slowly organizing off the coast of the Carolinas. (Tropical Tidbits)

Chantal will continue north and east after tomorrow dissipating before it gets to Delmarva.

Unraveling what happened in Texas

First, as someone based in Texas and with multiple members of our large but still close community still missing, I want to extend whatever sort of condolences are possible to friends and family that are dealing with what is truly an unspeakably horrific tragedy. On a holiday weekend that should be celebratory for kids and families no less, in the middle of the night when they’re at their most vulnerable. There truly are no words.

What we can do is take stock of what happened, explain why, explain what was known and what wasn’t know, and try to piece together how something like this happens in 2025.

The cause

As is often the case in interior Texas, one of the culprits involved last night was the remnant of a tropical system. Remember Tropical Storm Barry? It lasted all of 12 hours before coming ashore in Mexico. We can track Barry’s remnants from landfall last weekend through Thursday evening by looking at its “fingerprint” about 10,000 feet above the surface.

The path of Barry’s remnant circulation. (Tropical Tidbits)

So you had the remnants of a tropical storm. Because of this, you had abundant moisture coming from that storm’s source region in the Gulf. You had strong moisture transport coming northward as well with a strong low-level jet stream (a common feature in Texas located about 5,000 feet above the surface). But the jet was oriented to allow for maximum upsloping, aimed right at Hill Country. You had plentiful instability in the atmosphere as well.

So, tallying all that together: A remnant tropical system, moisture levels in the 99th percentile or higher, forced upward motion due to geography and wind direction, and plentiful instability. That’s a recipe for flash flooding. So how do you go from flash flooding to catastrophic flash flooding, because the difference is clearly enormous. When you put those parameters in concert with a weather pattern that allows for maximum efficiency of rainfall, a monsoonal pattern, and slow movement, as well as geography that allows for rapid build up of water on dry ground and riverbeds that “funnel” that through an area, that’s when you flip from ordinary to potentially tragic.

The forecast

Beginning last Sunday morning, forecast discussions from the NWS office in San Antonio and Austin noted the potential for heavy rain through the week. By Monday morning, they also noted the potential for nighttime warm rain processes in the western part of their coverage area, which would probably include Kerrville. By Tuesday afternoon they had specifically mentioned the potential for flooding on Thursday. Nothing really changed messaging wise on Wednesday or Thursday morning. By the afternoon, flood watches had been hoisted as the potential for significant rain became more evident.

Modeling? Well, it was so-so. When it comes to heavy rain, the lower resolution global models can give you a sense of what may happen, but they’re generally unable to resolve where the heaviest rains will fall. I went back and looked at some of the model rainfall forecasts for the event this past week. While some of the global models did indicate heavy rainfall potential, none were really flagging a high-risk type event. Even on Thursday morning, the WPC Excessive Rain Outlook showed a slight risk (2/4) in the area. We could probably say that the catastrophic rains were isolated enough that this was still reasonable. But truth be told, this probably deserved a moderate.

The Weather Prediction Center’s morning outlook had the Hill Country included in the slight risk (2/4) on Thursday morning. (NOAA WPC)

Within their discussion, they did emphasize risks of 3 inches or more, which is a reasonable note to make.

The higher resolution models did do better and did lead everyone down the correct path. On Wednesday night, the 00z HRRR model had about 7-9″ in a few bullseyes between Mexico and Texas by Friday morning. By Thursday morning, the model showed as much as 10 to 13 inches in parts of Texas. By Thursday evening, that was as much as 20 inches. So the HRRR model upped the ante all day.

The HREF from NOAA’s Storm Prediction Center indicated risk of 10″ or more as early as Thursday morning. (NOAA SPC)

The HREF model on Thursday morning, a tool developed from NOAA research also indicated the risk of 10″ or more in spots, using the “probability matched mean” product which can identify higher risk areas for heavy rainfall. I’ve used this in Houston many times, often with considerable forecast success ahead of flooding events. It has a knack for cutting through some of these higher end events and highlighting those risks.

So the signals were there and got worse as Thursday progressed. Messaging and flood watches responded to this appropriately and expanded into Thursday evening.

The warnings

Flash flood warnings were issued for areas before midnight as radar rain totals began to inflate up and over 3 to 4 inches. A flash flood emergency was issued at 4 AM for the Kerrville storms and 4:15 AM for storms near San Angelo. Rain totals were estimated to be encroaching on 10 inches at that point. So there was warning. This NWS office is acutely aware of the threats to the area from flooding, and the history is there. So I am assuming they were timely warnings unless I hear otherwise.

EWX continues Flash Flood Emergency [flash flood: radar and gauge indicated, flash flood damage threat: catastrophic] for Bandera, Kerr [TX] till Jul 4, 7:00 AM CDT Link

— IEMBOT Emergency (@emergency.weather.im) 2025-07-04T09:03:48.353981+00:00

Issuing the warning is half the process. Were the warnings received and acted on? That’s another story. And that will also come out in the days ahead. More on that below.

Did budget cuts play a role?

No. In this particular case, we have seen absolutely nothing to suggest that current staffing or budget issues within NOAA and the NWS played any role at all in this event. Anyone using this event to claim that is being dishonest. There are many places you can go with expressing thoughts on the current and proposed cuts. We’ve been very vocal about them here. But this is not the right event for those takes.

In fact, weather balloon launches played a vital role in forecast messaging on Thursday night as the event was beginning to unfold. If you want to go that route, use this event as a symbol of the value NOAA and NWS bring to society, understanding that as horrific as this is, yes, it could always have been even worse.

What should we be asking about then?

Beyond the fact that this was truly a tragedy that is extremely difficult to disseminate warnings on, I think we need to focus our attention on how people in these types of locations receive warnings. This seems to be where the breakdown occurred.

It’s not as if catastrophic flash flooding is new in interior Texas. There are literal books written about the history. The region is actually referred to as “Flash Flood Alley.” But how we manage that risk is crucially important context here. Are there sirens in place? Do there need to be sirens in place? Would people even hear sirens in the middle of the night in cabins or RVs or wherever they were? Tornado sirens have traditionally been used in parts of the country for people outdoors to get warnings. Is that an appropriate method in this region for the middle of the night and indoors?

Do we need to start thinking of every risk of flooding in Texas as a potential high-end event we should pre-evacuate the highest risk people (like children and elderly in floodways) for? Is that even practical? We can critique the answer given by the Kerr County judge here all day, but he’s correct in that the reality is they deal with flooding a lot. What is actually practical? I don’t know the answers to these questions. But it’s been a little over 10 years since Wimberley, which was a wake up call in some ways too. It’s time for another, and we need to think much bigger than just the areas impacted this time and more about Flash Flood Alley as a whole. Flooding risk is high in Texas. People learn to live with it in some ways. But something like this absolutely cannot happen again. The Texas legislature meets for a special session beginning on July 21st. This may be an important topic to add to the agenda.

Invest 92L a minor threat to the Southeast and some Eyewall housekeeping

In brief: Today we tackle Invest 92L, a minimal threat for the Southeast. We also have some important housekeeping to address, and we also have a quick update on some horrific flooding in parts of Central Texas overnight.

First, some housekeeping

We are going to begin this morning with some housekeeping regarding The Eyewall. Eric and myself, along with Dwight Silverman, have been discussing the future of the site. A lot of you have reached out to request ways to support our work beyond just reading, and we have (to this point) resisted that, other than various small sponsorships last hurricane season. And we are grateful for the interest many potential sponsors have shown this year too! However, we are going to try something else this year: subscriptions.

To be abundantly clear, our daily content will remain free, never paywalled, and that is not going to change as long as we operate this site. So, forever. This year, however, we will give readers who wish to support the site the option to do so at a modest cost, as well as a premium option for people who want a little bit more access. We will have more information on this soon.

The first step in this process is to migrate our email newsletter over to the Substack platform. If you currently subscribe to The Eyewall via email, your email should roll over to our Substack without you having to do anything. We understand a few people may have various concerns about the Substack platform, and some of the content posted there by other writers. If you decide to unsubscribe from emails, we understand. But the benefits of this route outweigh the potential costs. The primary issue is that WordPress caps our email subscribers, and the process to manage an account is not user-friendly. For a site with a limited budget such as ours, this is the best way to ensure the continued delivery of an email newsletter with our content. We will still continue to post here at The Eyewall.com, but updates will not be sent via email from here.

Also: Nothing at all is changing with Space City Weather. So for our Houston readers, our site, email, and annual fundraiser is not changing! This only applies to The Eyewall.

Our target date for this migration of The Eyewall newsletter to Substack is Monday, July 7th. We hope that this will offer us the flexibility to do more things, offer more options, more content, and a better overall user experience as time goes on. Thanks for your support!

Invest 92L

First, apologies for a couple days of lapse. Your primary author has been juggling family priorities the last couple days. Let’s get into this. First off, we now have Invest 92L. It’s sitting off the Southeast coast.

Invest 92L looks rather robust this morning, but despite the serious thunderstorm coverage, there’s not a whole lot of apparent organization yet. (Tropical Tidbits)

While it looks fairly impressive on satellite, underneath all the thunderstorms there is not a whole lot of apparent organization to this. That said, we will find out more later as hurricane hunters fly into the system to investigate. I think we could briefly squeeze a depression or low-end tropical storm out of this, but it’ll be close. Thankfully, it’s one of those where despite the classification, the impacts should be the same. We expect that this will migrate to the north over the next day or so, and it will eventually come ashore in the Carolinas before lifting north and east and out to sea.

92L will come ashore in the Carolinas and then exit quickly out to sea. (Tropical Tidbits)

The main threat from 92L will be locally heavy rainfall. The Carolinas look particularly prone to this where 1 to 4 inches of rain could fall near the coast. Not to minimize any threat, but there are currently not any flood watches posted. Still, localized flooding is a possibility with this in those areas. Thankfully, threats up the coast look limited as well.

The five-day NOAA rainfall forecast. (Pivotal Weather)

One other item to watch: Rip currents. Always a hazard but especially with the beaches *packed* this holiday weekend. Please use caution while in the water and always swim with a lifeguard present if that’s possible.

Beyond 92L

Nothing yet. It looks quiet for the time being.

Catastrophic Texas flooding

I want to close with one other major weather story you may hear about. Here’s what I wrote this morning at our Houston companion site, Space City Weather.

As noted yesterday morning by Eric, the forecast changed rather abruptly. And yesterday’s storms, while fairly quick movers, hit with some punch. Thankfully they did, as we saw instantaneous rain rates as high as 4 to 6 inches per hour in spots! Rainfall varied overall, with as much as 2 inches or a bit more in northeast Harris County and about a quarter-inch in western Harris County.

On a more serious note, parts of Hill County have seen some severe rain overnight. South-Central Kerr County is estimated to have received as much as nearly a foot of rainfall. This has led to flash flood emergencies there and along the Guadalupe River. The river at Hunt, TX is approaching 30 feet, beyond the 1987 flood event at that gage.

Historic flooding on the Guadalupe River in Kerr County. (NOAA NWS)

Downstream in Kerrville, they aren’t expected to hit the 37 feet of 1987, but it still looks to be a top 3 flood. Meanwhile, San Angelo saw around 8 inches of rain, with over 12 inches just north of the city.

Some of the pictures and stories from along some of these rivers, particularly the Guadalupe are pretty awful. We are hopeful that the warnings were received in time and that the human toll from this event will be minimal.