# North Texas Summer Nights Are Heating Up Faster Than the Days > Thirty-five summers of NOAA data from DFW Airport: 100° days come and go, but nights that stay above 80°F have more than doubled. - Author: Ross Edman - Published: 2026-10-10 - Tags: research, texas, weather - URL: https://rossedman.io/blog/research/north-texas-summer-nights/ **TL;DR:** Since 1992, DFW's 100° days have bounced around too much to call a clean trend. Its nights haven't: the average summer low is up about 1°F a decade, and nights that never drop below 80°F went from about 10 a year to about 21. I moved to Texas in 1992. I was seven years old, and I remember that year well. So when I wanted to know whether North Texas summers have really gotten worse, or whether it only feels that way because the bad ones are the ones we remember, 1992 was the obvious place to start counting. It turns out Texas went easy on me. That first summer, DFW hit 100°F exactly once, on August 10. Only three other years since 1953 had one 100° day or none, and 1992's average summer high is tied for the third coolest in that whole stretch. The summers since have been a different story. I pulled every daily reading from the weather station at Dallas/Fort Worth International Airport since 1992 and counted, then went back to 1953 for context. The answer surprised me: the hottest afternoons haven't changed nearly as much as the nights. - **21 nights**: a year at 80°F or warmer, 2009–2026 (Up from 9.6 a year in 1992–2008) - **+1.7°F**: warmer average summer low (74.5°F → 76.2°F, same two periods) - **71**: 100° days in 2011, the most since 1953, when the station's daily highs begin (2025 had 7) ## The Data NOAA's [Global Historical Climatology Network](https://www.ncei.noaa.gov/products/land-based-station/global-historical-climatology-network-daily) publishes the daily high, low and rainfall for thousands of weather stations. I used station USW00003927, the official observing site at DFW Airport. Its daily highs start in 1953 (the record before that has only lows), and every summer day since 1953 is there, with nothing missing. Most of this post looks at 1992 on. The last section uses the full record from 1953. A few choices worth knowing about: - **Summer means June, July and August.** 100° days are counted for the whole year, since some land in April, May and September. Summer rainfall is the June-through-August total. - **Temperatures are whole degrees Fahrenheit**, as the weather service records them. NOAA stores them in tenths of a degree Celsius, and rounding converts them back exactly. - **Readings NOAA flagged as failing a quality check are left out.** Since 1992, none of the summer days were. - **2026 runs through October 7**, the last day in the file when I pulled it. The summer is complete. The [analysis script](analysis.py) turns NOAA's file into the three CSVs behind every chart here. Each chart has a download link and a table of its numbers. ## What I Found ### The hot days come in bursts Start with the number everyone talks about: days that hit 100°F. *Chart: 100° days come and go (interactive on the web page, with a data table).* There's no smooth climb here. 2011 towers over everything, and some of the worst years (1998, 2000, 2006) came early in the record. 2022, 2023 and 2026 were all brutal, but 2025 had only seven. Averaged out, the later half of the record has more 100° days: 26.6 a year from 2009 to 2026, against 18.8 from 1992 to 2008. But the swings from one year to the next are so big that a straight-line trend through these bars isn't statistically convincing. The afternoons are noisy. ### Every summer, day by day Here is every summer day since 1992, one row per year, so you can see when the heat arrives and how long it stays. *Chart: Every summer day since 1992 (interactive on the web page, with a data table).* The darkest runs are 100°+ stretches. In 2011 it hit 100° or more 40 days in a row, from July 2 to August 10. In most other years the heat comes in spells of a week or two, broken up by merely hot weather in the 90s. ### The nights tell a cleaner story Now look at the other end of the day. This chart puts each summer's average high and average low on the same scale, with a fitted line through each. *Chart: The nights warmed faster, and more steadily, than the days (interactive on the web page, with a data table).* Both lines rise. But the average summer low has climbed about 0.9°F a decade, and it does so steadily, year after year. The highs rose about 0.6°F a decade, with far more scatter around the line. Statistically, the trend in the lows is solid; the trend in the highs is suggestive at best. That gap matters, because the night is when bodies, houses and the power grid recover. A night that bottoms out at 80°F doesn't give you much of a break. *Chart: Nights that never drop below 80°F have more than doubled (interactive on the web page, with a data table).* From 1992 to 2008, DFW averaged about 10 of these nights a year, and in a typical year (the median) just four. From 2009 to 2026 the average was 21 and the median 14.5. Every year with 40 or more, 2011, 2022, 2023 and 2026, came in the last fifteen years. The closest before that was 1998, with 39. ### The whole range of nights moved It isn't only the extremes. Lay every summer night from the two halves of the record side by side, and the whole spread has slid warmer. *Chart: A typical summer night is now warmer (interactive on the web page, with a data table).* The median summer night at DFW bottomed out at 75°F from 1992 to 2008 and at 77°F from 2009 to 2026. Nights that stayed at 78°F or warmer went from about a quarter of all summer nights to more than 40 percent. ### Why some summers are brutal: rain So if the afternoons don't trend cleanly, what makes one summer 1992 and another 2011? Mostly rain. When the ground is wet, a lot of the sun's energy goes into evaporating water instead of heating the air. When it's dry, there's nothing to slow the heat down. *Chart: Dry summers are the brutal ones (interactive on the web page, with a data table).* The summers drier than the median (7.35 inches) averaged 32.8 days at 100°F or hotter. The wetter ones averaged 13.3. 2023 was the driest summer since 1992, with 1.25 inches, and had 55 such days. 2011 got 3.9 inches. My first summer, 1992, got 9.8 inches and one 100° day. Rain explains a lot of the swings from year to year. It doesn't explain the warming nights, though. Summer rainfall hasn't changed: it averaged 7.85 inches from 1992 to 2008 and 7.82 inches from 2009 to 2026. And the nights warmed even in wet years. Comparing only the wetter-than-median summers, the average low went from 73.3°F in the first half of the record to 75.7°F in the second. ### The longer view, back to 1953 Starting in 1992 was personal, not scientific, so it's fair to ask what the decades before look like. The station's daily highs go back to 1953. *Chart: Seventy-four summers: the highs wobble, the lows climb (interactive on the web page, with a data table).* The longer record makes the same point, more firmly. From 1953 to 1991 the average summer high was 94.4°F; since 1992 it's been 95.3°F. The average low went from 72.9°F to 75.3°F, more than twice the change. Over all 74 summers the lows have risen about half a degree a decade, the most consistent trend in the data. The highs moved about a fifth of a degree a decade, which the year-to-year swings make hard to tell apart from no change at all. It also puts the hot afternoons in perspective. The drought summers of the 1950s were brutal: 1954 had 52 days at 100°F or hotter, as many as 2026, and 1980 had 69, nearly as many as 2011. Hot afternoons aren't new in North Texas. Hot nights are. *Chart: Nights that stay at 80°F were rare before the 1990s (interactive on the web page, with a data table).* From 1953 to 1991, DFW averaged about three of these nights a year. Since 1992 it has averaged 15.5. The six years with the most, 2011, 2023, 2026, 2022, 1998 and 2008, all came after I moved here. ## What It Doesn't Tell Us - **This is one weather station.** DFW Airport is one spot in a metro area of eight million people. Other stations would give somewhat different numbers. - **The land around it changed.** In 1992 there was a lot more open ground around DFW Airport. Concrete and buildings soak up heat during the day and release it at night, so some of the warmer nights here are likely local growth, not only a warmer climate. This data can't separate the two. - **Rain and heat feed each other.** Dry ground makes hotter days, but a stubborn high-pressure ridge also blocks rain and brings the heat. The rainfall chart shows they go together, not which one starts it. - **Thirty-five summers is a short record for climate.** Wet years and dry years, El Niño and La Niña, make the afternoons swing hard. That's exactly why the 100° day count doesn't produce a clean trend, and why I'd be careful reading too much into any single year, 2026 included. - **Averages hide the worst hours.** Daily highs and lows don't capture humidity, heat index or how long the hottest part of the day lasts. ## Get the Data - The original daily record: NOAA GHCN-Daily, [station USW00003927](https://www.ncei.noaa.gov/data/global-historical-climatology-network-daily/access/USW00003927.csv) (a 10 MB CSV covering 1948 to today). - The [analysis script](analysis.py). Run it on that file to rebuild everything here, or change `FIRST_YEAR` to look further back. - The cleaned data: [one row per year since 1992](dfw-summers.csv), [one row per summer day since 1992](dfw-summer-days.csv) and [one row per year since 1953](dfw-summers-history.csv).