Summer is here: Let’s talk heat forecasting

In brief: Today’s post talks about the arrival of summer-like weather in Houston, and breaks down the four phases of summer. We also discuss a new offering from the National Weather Service called “HeatRisk.” Finally, we look ahead to a hot and sunny Memorial Day Weekend.

We’re starting to see 90-degree days on the regular which, in my mind, signals the start of summer in Houston. Long-time readers will know that I like to break summer down into four phases. Why? Because “summer” season lasts so long, nearly five months. This is contrast to meteorological summer, which runs from June through August, and “solstice” summer, which runs from June 20 through September 22 this year. For me, Houston’s summer typically runs from about mid-May through mid-October. Here are the four phases:

  • Early summer: When we first start to see 90-degree temperatures with regularity, but some nights in the 60s are still possible, and there’s still the thinnest hope of a weak front
  • Mid summer: When highs run from 90 to 95 degrees, and nights are sultry, but you know it could still get worse
  • High summer: Somewhere between late July and early September there’s a period where temperatures reach the upper 90s to mid-100s and you realize, “Ok, this really is the worst.”
  • Late summer: This is the period in September and early October when days grow shorter and we usually see the first front or two of the season. But most of the time it’s still hot.

Early summer began this week in Houston, and we have a chance to bump into Mid summer early next week before we drop back into early summer. You may think it’s crazy to have gradations of heat during summer, as Houston is invariably hot and humid during the summer months. But as we found out last year, there is heat and there is heat. Last summer we used the “wet bulb globe temperature”to measure how hot it really felt outside, and we’ll continue to use this tool. However there’s now another way to assess the heat.

HeatRisk map for Monday, Memorial Day. (National Weather Service)

This year the National Weather Service is introducing a “HeatRisk” color-based scale that takes into account the following factors to provide guidance for outdoor activities:

  • How unusual the heat is for the time of the year
  • The duration of the heat, including both daytime and nighttime temperatures
  • If those temperatures pose an elevated risk of heat-related impacts

The scale ranges from green (little or no risk) to magenta (extreme). If we look at the risk for Memorial Day we can already see a few areas of Houston reaching an ‘extreme’ level of risk. This is due to temperatures likely in the mid- to upper-90s and the time of year, late May, as we’ve not re-accustomed ourselves to the heat yet.

In any case, feel free to use this planning tool as we go through the summer.


We’ll see a few more clouds today than we saw on Monday, and this should help hold temperatures to about 90 degrees. Winds will be from the south, at 10 mph or so, with gusts up to 20 mph. Aside from that it’s going to be another humid day, with a warm night in the mid- to upper-70s.


Skies should be partly to mostly cloudy on Wednesday, and this should allow some parts of Houston to stay in the upper 80s. However, if we get a bit of sunshine during the afternoon hours, highs in the low 90s are possible. Still warm. Still humid.

Thursday and Friday

These days will probably bring mostly sunny skies, so temperatures from 90 to the low 90s are likely. Nights remain warm.

Wet Bulb Globe Temperatures will be high this week, but not extreme. (Weather Bell)

Saturday, Sunday, and Monday

Hot and mostly sunny. Highs this weekend are likely to reach the mid-90s, with upper-90s possible on Memorial Day. By Monday it looks like a weak front will approach the area, bringing a chance of rain with it. However I don’t believe this front will arrive in time to modify conditions during the daytime on Monday.

Next week

The front should arrive on Monday night or Tuesday, and at this point it does appear as though the boundary will indeed push all the way to the coast. Don’t expect miracles in late May, but we might see some slightly drier air and highs in the upper 80s, with modestly cooler nights. It isn’t much, but it’s the best we can reasonably hope for at least the next three months, and probably longer.

Answering a few questions about last week’s superstorm, and assessing how hot it will get this week

In brief: Houston will remain warm and humid through next weekend, with partly to mostly sunny days and highs generally in the low 90s. Rain chances are low throughout the period. Today’s post also addresses some lingering questions from last week’s storms, including why the event was not particularly well predicted.

Over the weekend, the greater Houston region continued to recover from damaging winds last Thursday. On Sunday evening, CenterPoint said that it had restored electricity to more than 700,000 people who lost power, with about 240,000 customers still without. The transmission company said it remains on track to be “substantially complete” with power restoration by Wednesday evening. Before jumping into the forecast here are some additional thoughts about the very strong storms that arrived last week.

Outside of north Texas, it won’t rain much this week. This will help recovery activities from last week’s storms proceed. (Weather Bell)

Why wasn’t this well predicted?

This is a great question. The answer is that we have seen these kinds of setups before in which there is an atmosphere primed with moisture and instability, and with a potential trigger for supercell storms to form many times before. However, even hours before their development just west of Houston on Thursday, there was no ready data to indicate the true severity of the blow-the-doors-off storm that was coming.

It’s the kind of thing where there maybe is a 1-in-50 or a 1-in-100 chance that something so severe, a supercell event with very strong straight-line winds directly over the city’s urban core, could develop. If we had messaged that an “extremely dangerous and destructive event” was coming the previous 10 times there was such an atmospheric setup, and nothing of any real significance happened, who would believe us on the 11th time? There is a boy-who-cried-wolf problem here. The only real solution here is that we try harder to find that bit of data that gives us more confidence in a rare event like this one. I’m confident that Thursday’s storms will be studied in depth to identify such clues.

I was lying in bed on Sunday morning thinking about all of this—yes, I lose sleep worrying about this kind of stuff—and of all things a 2002 Houston Texans football game popped into mind. It was December 8, the team’s first season, and being an expansion franchise the Texans were not great. That week the Texans had to go to Pittsburgh and play the Steelers in near-freezing temperatures for a meaningless game. The Steelers were a decent football team, nothing fantastic, but were expected to blow the Texans out. The line was Steelers -14. The Steelers outgained the Texans 442 yards to 47. They won the time of possession with 40 minutes to 20 minutes. Texans quarterback David Carr was 3-of-10 passing, for 33 yards. That was in line with expectations.

But the Texans won 24 to 6. They returned three turnovers for touchdowns, and kicked a field goal after a short drive. The Steelers had many more penalties. Their two long drives ended in field goals rather than touchdowns. It was just an odd day. While the final score was within the bound of possibilities, on any given Sunday anything can happen when two teams meet, it was the lowest probability event. That’s kind of like what happened Thursday. A super-storm was within the realms of possibility, but was like an expansion team playing terribly, but still blowing out the Pittsburgh Steelers on the road in very cold conditions.

Why weren’t there more real-time warnings?

This is something Matt and I are taking away from the storms. I feel like we do a good job of forecasting the weather here, but where we struggle is in real-time coverage. Some readers have asked why we did not send out warnings for tornadoes on Thursday. That is because this is the express function of the National Weather Service, which has a large staff and the technology to send out real time warnings for tornadoes and other life-threatening weather events. Here’s more information about the agency’s wireless emergency alerts program. They are the experts at that, and we defer to them.

Matt and I were both tracking Thursday’s storm in real time. We were on top of things. And there was perhaps a 30 to 60-minute window between 5:30 and 6:30 p.m. when we could have provided some actionable information. However, in the time it would take to write and disseminate a post about that in real-time, the dynamic event might have already passed for most people. One possible solution is that we are likely to start sending (albeit very rarely) “urgent notifications” through our app. It is available for free here for Apple ioS, and here for Android. Regardless, we recognize this as a weakness and will address it here at Space City Weather. For one thing, we are changing the frequency of updates on days when severe weather is possible.


Our overall warm, but not extremely hot pattern will continue as we inch toward the end of May. With high pressure largely in control we’ll see this pattern largely persist through the holiday weekend, with the only really noticeable change in the strength of southerly winds, or a few more clouds on some days. For today, we’ll see highs generally in the low 90s, with only coastal areas unlikely to reach 90 degrees. Skies will be mostly sunny. Winds will blow from the southeast at 5 to 10 mph. Overnight lows will be warmer than we’ve seen, dropping only into the upper 70s.

Our weather will be nothing if not consistent this week. (Weather Bell)

Tuesday and Wednesday

In response to low pressure over the central United States, we’ll see an uptick in southerly winds with gusts up to 25 mph. Both days will see partly to sunny skies, with highs around 90 degrees. Nights remain quite warm.

Thursday and Friday

These will be partly to mostly sunny days, with winds perhaps slackening a bit. Highs in the low 90s. There is perhaps a 10 percent chance of rain showers each afternoon.

Saturday, Sunday, and Monday

Memorial Day weekend looks warm and sunny. Expect highs in the low 90s, with warm nights. Rain chances remain near nil. Plan your outdoor activities with confidence.

Next week

There are some hints of a pattern change by Tuesday or Wednesday of next week, with possibly a weak front sneaking into the area at the end of the month, and bringing with it a smattering of rain chances. This is far enough into the forecast, however, that my overall confidence is quite low. We shall see.

Power remains out for a quarter of Houston as we begin to assess carnage from Thursday’s storms

In brief: Ironically, Thursday was “National Love a Tree Day.” Instead of loving trees, however, a line of storms ripped through the heart of Houston and tore many thousands of them down, killing a handful of people, damaging homes and vehicles, and bringing down power lines. In this post we’ll provide some meteorological data about what happened Thursday, and a forecast looking ahead to calmer weather.


The National Weather Service has completed a survey of damage from two confirmed tornadoes that struck Houston on Thursday evening, one at 5:44 pm near Pine Island in Waller County, and the second at 6:08 pm near Jersey Village in Harris County. Here is the summary of those investigations:

Waller County tornado

  • Rated EF1
  • Peak winds: 100 mph
  • Path length: 0.71 miles
  • Path width: 100 yards
  • No fatalities or injuries
  • Duration: 5:44 to 5:45 pm CT
  • Significant tree damage, large metal barn destroyed, metal debris thrown up to 1,000 yards away.
Radar signature of the Cypress tornado. (National Weather Service)

Cypress tornado

  • Rated EF1
  • Peak winds: 110 mph
  • Path length: 0.77 miles
  • Path width: 100 yards
  • No fatalities or injuries
  • Duration: 6:08 to 6:09 pm CT
  • Numerous single family homes had roof damage and broken windows. Damage path well defined.

This is a preliminary report that may be adjusted in the future.

Straight-line wind damage

As Matt noted yesterday, most of the damage in the region was caused by straight-line winds, and what appeared to be something of a mini-derecho. The National Weather Service has also collected maximum wind reports from around the area on Thursday afternoon and evening. Some of the highest values were recorded at Texas A&M University, 71 mph, and the Houston Ship Channel, 74 mph.

Very strong winds just above the surface were able to break through, and get down to the ground. These winds a few hundred feet above the ground were even stronger than what was observed at ground level. This is likely the reason why some skyscrapers in downtown Houston observed significant damage at their upper levels.

Power outages

About one-quarter of customers in Harris County are still without electricity this morning. In an update to the media on Friday evening, CenterPoint Energy said restoration could take “several days or longer” for the hardest hit areas.

“As crews continue to uncover damage and encounter new challenges while making repairs, restoration may take more time than customers typically experience following a routine storm event,” the company said. “CenterPoint Energy appreciates customers’ patience, and the company will work around-the-clock until the last customer is restored.”

We genuinely appreciate the hard work by all of the lineman working to restore power. It is a hard and hot business to rapidly repair storm damage, I know first-hand how hard the men and women are working out there.


The heaviest rainfall on Thursday occurred north of the area, in locations such as College Station and Livingston, which have already been inundated over the last 30 days. Significant river flooding is likely along the East Fork of the San Jacinto River later this weekend, whereas the Trinity River is already cresting today before a slow fall next week.

The East Fork of the San Jacinto River will make a steep rise on Saturday. (National Weather Service)

Fortunately the entire southeast Texas region will now get a chance to dry out. Our weather is about to get boring.


After the fog lifts this morning we’ll see sunny and hot conditions today, with highs likely in the low 90s. And that’s pretty much your forecast for the next week, at least. The only change will be gradually warming nights. Lows this morning got to below 70 degrees for most locations. It was, almost, nice? Well, except for the mosquitoes. By Wednesday or Thursday morning, our lows are likely to only dip into the upper 70s. Rain chances are nil for the foreseeable future.

We wish everyone the best this weekend as you recover from Thursday’s storms, and thank everyone who is out there helping to pick up the pieces. This community is at its best in moments like these, when we’re all working together.

Houston, what the heck happened on Thursday?

In brief: There will be some showers this morning south and east of US-59 to the coast. Some thunder is possible. No severe weather is expected. The rest of the forecast through Monday and Tuesday is quiet and turning hotter.

We have a few showers southwest of Houston that will push in this morning. The steadiest rain will be south and southeast of the city, areas that saw a bit less action yesterday. We then clear out and dry out for later and tomorrow, Sunday, and Monday. Highs will nudge into the 90s with lows slowly increasing through the 70s into next week.

Trying to make sense of Thursday

This will be a different post than typical. I want to walk through what happened yesterday from a forecaster’s perspective. About 750,000 customers remain without power this morning, and because of the extensive, widespread damage, this number will very slowly decrease today and tomorrow. Some may be without power until next week.

So how did we get here? As a refresh, here is Eric’s post from yesterday morning. We were all really focused on the threat for heavy rain, and with the high risk in place yesterday to our north, that shouldn’t be a surprise. In fact, 4 to 5 inches of rain did fall as expected, basically north of highway 105 through Conroe.

Rain totals on Thursday lined up really well with the high risk bullseye that was in place to our north. Flash flooding was a common problem in many places yesterday. (NSSL MRMS)

Eric did note the severe weather and correctly underscored the chances of wind and an isolated tornado. The Storm Prediction Center had the right idea on severe weather yesterday too. But again, I think most of us the significant messaging was heavily focused on the rain.

Through most of yesterday morning, not a whole lot seemed to change. Eric, Dwight, and I met up for lunch at a Pappas BBQ but not the one we originally planned on, which, thanks Apple Maps. Or Google Maps. Who’s to say? Anyway, we touched briefly on the day’s weather but were not particularly concerned about anything else happening. We checked radar while leaving and all looked good.

I got back to my desk and noticed a few people pinging me about a Reed Timmer tweet discussing rain-wrapped tornado potential in southeast Texas. Reed’s a good guy and a very smart meteorologist, but he also has tons of enthusiasm. His brand is to dominate and never stop chasing. My first reaction admittedly was to roll my eyes, but then I checked out the HRRR model, one of our hourly updating high-resolution weather models. Indeed, it lit up with supercells by 2 PM. But it was 2 PM. And there were no supercells.

So clearly it was overdoing it, and it would be necessary to watch subsequent runs to see what changes, as well as radar to observe the evolution of the storms. I decided to look at some other data, and I was surprised to see that despite clouds, haze, and mist (or even a heavy drizzle at times), the atmosphere was extremely unstable over the area.

The atmosphere over Houston around 2 PM on Thursday could be classified as highly unstable. (NOAA SPC)

And when you would dig deeper into the models, everything pointed to severe potential. But we also just went through this a few days ago with a major hail threat that basically failed to materialize with any consequence in Houston. In that case, we had the same situation in theory: Impressive instability and an atmosphere primed to rock. As a meteorologist, you look at this two ways: We just came off a semi-bust and you need to make sure you are more confident in something than normal before pushing it, and secondly you also can’t deny that the ingredients were there. A tornado watch was issued, which was a little surprising, though given the parameters there was no good reason to argue much against it.

When I looked at model data, it appeared that any supercell risk would track from the southwest to northeast. I even highlighted a cell around Spring at about 4:25 that I thought was showing signs of trending toward something more meaningful. I went to pick up my oldest from school and came back intending to help him do some homework. Even by about 5:10 or so, it seemed pretty straightforward: A line of severe storms with gusty, maybe localized damaging winds was moving through Brenham and toward Waller County, and we’d probably get it through by 7 or 8 PM and that would be that.

Radar at 5:10 PM showing a line of severe storms west of Waller County with heavy rain as predicted to the north of The Woodlands and Conroe. (RadarScope)

Things began to change quickly about 15 minutes later. It was evident that rotation had begun to develop on the leading edge of the bowing line near Bellville. And it seems likely that a tornado may have been put down just east of there shortly thereafter. That is not necessarily uncommon. It’s often how we get our tornadoes locally, but it’s usually brief and disappears after 5 to 10 minutes. And indeed, the rotation weakened some, but at 5:35, it flared back up again, just west of FM 359 to the east of Bellville. Thereafter, it absolutely exploded near Pine Island and just south of 290 in Prairie View. By 5:40 to 5:45, we clearly had a problem.

A two panel radar image at 5:43 PM showing tremendous rain on the left and vigorous, violent winds and rotation on the right, centered just south of Pine Island. (RadarScope)

I have been in Houston for about 12 years as of this week, and I cannot recall seeing this type of velocity signature (the right-hand panel) show up in this area. When you see this as a meteorologist, it either means a violent tornado is underway, or destructive winds are probable. At times, it looked like you could pick out a debris signature on radar that would essentially confirm a tornado, but it never took off, which led me to believe that this was becoming a major straight line wind event.

I had been texting a bit with Justin Ballard, the Houston Chronicle’s fine meteorologist earlier about the tornado watch. We both expressed some skepticism it would produce. He texted me at 6:07 in the middle of this saying, “Yeah, that doesn’t look like a bad decision after all.” Yes, many of us talk to one another. Yes, we occasionally have opinions on things.

Anyway, this continued to march east-southeast and slowly expand. At this point, it becomes straightforward: Monitor it and warn and clear. I had posted to Twitter in a tone I very, very rarely ever use. I don’t throw around language like “Treat this like a tornado” very often. We had gotten very few damage reports up to that point, but the radar was indicating 110 mph winds down to about 2,500 feet. By 6:23 PM, radar showed 120 mph winds down below 2,000 feet approaching Oak Forest. Does all that reach the ground? No. But a lot of it can. It also made me gravely concerned for the downtown high rises.

Insane winds down to about 2,000 feet moving into the northwest Loop at 6:23 PM. Whether or not it was a tornado, the outcome would be similar. (RadarScope)

We’ll find out more today and tomorrow about specifics on damage and what was a tornado or straight line winds. Whatever the case, this was one of the most ferocious storms I’ve ever seen. This was a smaller scale version of what occurred in Iowa a few years back, when they had 140 mph winds down to about 1,000 feet but over a wider area. Whether or not this gets classified as a derecho will remain to be seen. I think it probably falls just short of that metric because of some discontinuity in the damage report path, but honestly, does it matter? It will take time to pick up from this one, and we hope our readers are safe. A major kudos goes out to the many media meteorologists and NWS meteorologists that assisted in keeping as many people safe and informed as possible. Saving lives is rarely a literal thing for a meteorologist. I am thinking that it was for many last night.

I’ll close with a bit of a sobering note: Hurricane season begins in about 2 weeks. What many of you witnessed last night would be experienced not over a few minutes but over several hours over a large area if a truly potent hurricane found its way into the Houston area. By living in this region, you have to accept the risks associated with that. We know a lot about flooding. Most of us know about surge. Very few knew about wind and what it’s really like. Many do now. Use this experience to inform your preparation for hurricane season just in case. Houston has been through an absolute meat grinder of weather disasters in the last 10 years. Candidly, it sucks, but we should know enough now to prepare for the next one.