Arguments
Eastern U.S. broils after heat wave kills over 1,300 in Europe
Posted on 7 July 2026 by Guest Author
This is a re-post from Yale Climate Connections by Bob Henson and Jeff Masters
Update (Thursday, July 2, 3 p.m. EDT):
Extreme heat warnings covered much of the Midwest, mid-Atlantic, and Northeast U.S. on Thursday – including the heavily populated Interstate 95 corridor from Virginia to southern Maine – as a well-predicted heat wave intensified with the approach of a holiday weekend, bringing widespread temperatures above 100 degrees Fahrenheit. Millions of people are set to experience the hottest Fourth of July ever recorded in the history of their town or city. Parades, concerts, and other events were being rearranged, rescheduled, or moved indoors to avoid the worst of the dangerous heat.
As the United States prepared to commemorate the nation’s 250th birthday, the U.S. capital had entered one of the worst heat waves in its history. As of midday Thursday, the official forecast for Washington’s Reagan National Airport was for highs of 102°F, 104°F, and 103°F for Thursday through Saturday, July 2-4. These readings would melt the D.C. daily records of 101°F, 101°F, and 100°F. In observations that go back to 1871, Washington has never recorded two consecutive days of 103°F, and the only three-day streak of highs reaching at least 102°F occurred way back on July 19-21, 1930.
Further north, the nation’s first capital city, Philadelphia, was on track for predicted highs of 103°F, 103°F and 101°F on Thursday through Saturday, which would tie the daily record on Thursday and approach the city’s monthly record of 104°F set on July 3, 1966. For what it’s worth, the temperature recorded in Philadelphia by president-to-be Thomas Jefferson at 1:00 p.m. on the day the Declaration of Independence was signed – July 4, 1776 – was a mere 72.5°F.
In New York, Central Park had already reached 100°F by 2 p.m., the first time it has reached the century mark since July 18, 2012. The forecast was for 101°F on Friday and 97°F on Saturday. In 157 years of recordkeeping, Central Park has only had 5 pairs of consecutive days that both reached at least 101°F. The only pair of 102°F readings at Central Park occurred during the height of the central U.S. Dust Bowl on July 9-10, 1936.
Original article, posted on Monday, June 29:
As our fossil-fuel-warmed world careens into what’s likely to be months of record global-scale heat goosed by El Niño, early season heat waves are already proving tragic this summer. Most of central and northern Europe was assaulted over the past week by the continent’s worst heat wave ever recorded before the core summer months of July and August, and at least 1,300 people have died as a result.
Meanwhile, extreme heat watches extended along the U.S. East Coast on Monday from New Jersey and eastern Pennsylvania to southern Maine, foretelling a miserable and potentially record-setting heat wave that will run from midweek into the Fourth of July holiday across much of the U.S. east of the Appalachians. Heat indexes – reflecting air temperature plus humidity – may exceed 110 degrees across much of the extreme-heat watch area.
As of midday Monday, the official forecast for Washington’s Reagan National Airport was for highs of 99°F, 103°F, 103°F, and 101°F for Wednesday through Saturday, July 1-4. The city has never recorded two consecutive days of 103°F, and only a handful of heat waves have produced four-day strings of 99°F. In New York, Central Park’s forecast highs of 91°, 96°, 95°F, and 91°F for July 1-4 wouldn’t be as historic, but they would still make for a hot, sweaty lead-up to the holiday, especially with lows possibly staying above 80°F for two nights.
The National Weather Service office warned folks in and around Philadelphia: “Very warm low temperatures in the mid 70s to low 80s at night will not offer any relief from the heat. This combined with multiple days of near record-breaking temperatures will exacerbate the impacts from the heat and humidity.”
The climate change connection
In a rapid-response analysis released on Friday, “Fossil fuel emissions have rapidly worsened European heatwaves in just a few decades,” the nonprofit research group World Weather Attribution said: “In 1976, when some of the previous European records were set, the 2026 temperatures would have been virtually impossible to occur in June, while also highly unlikely at any time of the year. In 2003, the first major heatwave of this century, daytime heat like this would still have been very rare, about 10 times less likely than today, while nighttime temperatures such as this June would have been more than a hundred times less likely in 2003.”
“Extreme heat is already reaching the limits of our societies’ ability to cope,” the group added.
The past week’s European heat wave was fierce, widespread, and prolonged – a perfect storm of torrid misery and danger, particularly in France. Early last week, temperatures soared as high as 44.6 degress Celsius (112 degrees Fahrenheit) in Bordeaux.
The brutal heat ignited a long-simmering debate on the role of air conditioning in France, where policy, custom, and legacy building styles have inhibited its use but where dense urban living without AC as heat waves worsen has led to horrific death tolls over the past 25 years.
Europe’s heat wave responsible for over 1,300 deaths
The World Health Organization said on Sunday that more than 1,300 excess deaths – those beyond the usual mortality rate for this time of year – have been recorded because of Europe’s record-breaking heat wave since June 21. Health officials in France independently reported 1,000 more deaths than expected in the country since Wednesday. The heat-related death toll is undoubtedly much higher, and we should expect the final toll from heat in Europe in 2026 will measure in the tens of thousands, as has occurred in each of the last four years.
Figure 1. Heat waves over the period 1900-2025 in which analyses have found that at least 100 heat-related deaths occurred.
The preliminary data for 2026 already makes the current heat wave the 15th-deadliest in world history. Most of these deadly heat waves have occurred in Europe. However, a 2026 paper found that one-day extreme heat waves in India in excess of the 97th percentile for temperature kill about 3,400 people, and five-day extreme heat waves kill about 30,000 people. There have been several heat waves of this magnitude in India in recent years, so India should have many more entries in Fig. 1 than shown.
The high heat death tolls in Europe are from a combination of factors:
- Building design: A lack of air conditioning, combined with architecture designed to retain heat during the long winters.
- Methodology: More rigorous techniques of reporting of heat deaths over time and as compared to other areas.
- Age: A relatively old population.
- Accelerated warming: According to the World Health Organization, Europe is the fastest-warming continent on Earth, heating at roughly twice the global average rate.
With heat, as with other phenomena such as hurricanes, scientists are now better able to comprehensively assess how extreme weather and climate events lead to premature death. National agencies and researchers once focused on heat-related deaths that were direct, such as those caused by heat stroke. Today, indirect heat-related deaths – for example, those caused by cardiovascular and respiratory conditions that are worsened by the heat and pollution trapped during heat waves – are being analyzed more comprehensively. This change is occurring at the same time extreme heat itself is getting measurably worse – and as heat-related adaptations actually improve in some places, in a high-pressure race with time and a warming planet.
More Euro heat to come?
Record and near-record heat may return over western Europe by late this week into the following week as the ridge of high pressure rebuilds. After highs in the 80s Fahrenheit this week, Paris is predicted to again approach 100°F by early next week. Likewise, London may boomerang from the 70s this week back toward or above 95°F next week.
The oceanic link to this summer’s heat in Europe
Because of the influence of the nearby Atlantic and Mediterranean, which take time to warm and cool with the seasons, the oceanic climates of Europe tend to see their most intense heat toward the latter months of summer. However, the waters adjoining Europe are enmeshed in a marine heat wave right now. Just south of France and Italy, the northwest Mediterranean sea surface temperature is running as high as eight degrees Celsius (14°F) above average.
West of Spain and Portugal, unusually high sea surface temperatures of 68°F (20°C) extend hundreds of miles to the west. This warm water is adding heat and moisture to the overlying atmosphere and stoking the humid heat plaguing France and the British Isles. As noted by meteorologist James Peacock (MetSwift) on X: “Usually if 20°C or higher [temperatures at 850 millibars] are to reach the UK, it has to come from the European landmass. Not so this summer, when it could just as easily come from the southwest …& be far more humid as a result. In the decades ahead, this may cease to be an unusual situation.”
Three firefighters die in Colorado
On Saturday afternoon, June 27, wildland firefighters with the Rifle Helitack crew engaged with the Snyder Fire in Mesa County, on the Colorado border with Utah. They experienced extreme fire weather conditions: high heat combined with winds gusting up to 57 mph (91 km/hr). The fast-moving flames overran their position, forcing five crew members to deploy their emergency fire shelters. Three firefighters were killed and two others were severely burned. The survivors were flown to a burn center, where they are being treated. Both U.S. Wildland Fire Service and Forest Service personnel were involved. The fire is estimated at over 28,000 acres.
The previous most recent line-of-duty fatality of a U.S. wildland firefighter occurred on September 26, 2025, when Isabella “Bella” Oscarson, a 26-year-old crew module leader with the Idaho Department of Lands, was killed while assisting with a prescribed burn in the Nez Perce-Clearwater National Forests in Idaho. She was fatally struck by a falling tree while working on the fire line.
“Burnover” incidents such as the one on Sunday, where firefighters are trapped by a fast-moving blaze, can be especially deadly. A total of 96 firefighters were killed and 78 injured in a total of 41 burnover incidents in the 28 years from 1990 through 2017, according to a USDA Forest Service report. Five incidents alone led to 44 of the fatalities.
Figure 2. A plume of dense smoke partially obscured the sun in Boulder, Colorado, on Sunday evening, June 28, 2026. The plume was generated by the Willow Fire, a few miles northwest of Leadville, Colorado, which exploded from five to about 1,000 acres on Sunday afternoon. As of midday Monday, the Willow Fire was 0% contained. (Image credit: Bob Henson)
An early and significant wildfire season in the Western U.S.
Significant wildland fire activity is occurring across multiple geographic areas in the U.S., with 3.1 million acres having burned so far in 2026. Over the past 10 years, only 2022 (with 3.6 million acres burned) was this active so early in the year. Today, the U.S. National Interagency Fire Center raised its national preparedness level to Preparedness Level 4 on a scale of one to five. The national preparedness level was last raised to the highest level, PL 5, on July 18, 2024. Reaching PL 5 indicates that the nation’s firefighting resources are severely stretched. It is triggered by a combination of high fire weather indices, widespread large fire activity, and an extreme demand for personnel and equipment. It is likely that PL 5 will be declared in July this year.
With an unusually strong trough of low pressure over the Southwest U.S. expected to bring strong winds and very little precipitation this week, firefighting resources will be strained, and there will be a continued high potential for new large fires to emerge across multiple geographic areas. The center wrote:
“The southern Intermountain West will continue to be plagued by dry and unusually windy conditions today. Extremely critical fire weather, with southwesterly winds gusting from 30-60 mph, locally stronger in favored downslope areas, along with relative humidity from 5-15% will be most likely across northern Arizona into eastern Utah, western Colorado and far northwestern New Mexico. These conditions are extraordinarily rare for late June, and impacts will likely be severe.”
Utah has seen the worst wildfire conditions in the U.S. this year, and the governor said that the 2026 wildfires have been the most destructive in Utah history. A State of Emergency has been declared for the state as five fires have run tens of thousands of acres, and a statewide ban on fireworks displays has been ordered for the July 4th weekend. Over 1,000 people statewide have been evacuated because of the fires.
The largest wildfire currently burning in the U.S. is the Cottonwood Fire in Utah, which is at 94,000 acres burned. Nearly 1,000 firefighters are deployed on the fire, which is 0% contained.
Smoke from the wildfires has thus far mostly avoided heavily populated areas, but you can follow the smoke forecasts for the fires by referencing our post from last year, 15 sources of wildfire smoke forecasts for North America.
Extreme heat records broken across Europe
At least 10 European nations or territories recently recorded all-time national highs, meaning the hottest single temperature ever reliably recorded at any town or city nationwide. Below are a few of these preliminary national records, as compiled by weather historian Christopher Burt (author of “Extreme Weather”), primarily from data posted by independent weather-records analyst Maximiliano Herrera. Note that all of the new records broke ones that occurred in late July or August, weeks later in summer than we are now:
Jersey (UK territory)
39.3°C (102.6°F) at Maison St. Louis on June 25
Old: 37.9°C (100.2°F) at Maison St. Louis on Jul. 18, 2022
Guernsey (UK territory)
36.4°C (97.5°F) at Rocquaine Bay on June 25
Old: 35.0°C (95.0°F) at Rocquaine Bay on Aug. 5, 2003
Luxembourg
40.9°C (105.6°F) at Reckange on June 26
Old: 40.8°C (105.4°F) at Steinsel on Jul. 25, 2019
Denmark
37.0°C (98.6°F) at Odum on June 27
Old: 36.4°C (97.5°F) at Holtstebro on Aug. 10, 1975
Germany
41.7°C (107.1°F) at Coschen on June 28
Old: 41.2°C (106.2°F) at two sites on Jul. 25, 2019
Czech Republic
41.9°C (107.4°F) at Doksany on June 28
Old: 40.4°C (104.7°F) at Dobrichovice on Aug. 20, 2012
Poland
40.5°C (104.9°F) at Slubice on June 28
Old: 40.2°C (104.4°F) at Proszkow on Jul. 29, 1921
Belarus
40.4°C (104.7°F) at Pinsk on June 29
Old: 38.9°C (102.0°F) at Gomel on Aug. 8, 2010
Slovakia
41.3°C (106.3°F) at Kamenice nad Hronom, June 30
Old: 41.0°C (105.8°F)at Tur?a nad Bodvou, June 29, 2026
Before that: 40.3°C (104.5°F) at Hurbanovo on Jul. 20, 2007
Hungary: 42.0°C (107.6°F) at Szecseny on June 30
Old: 41.9° (107.4°F) at Kiskunhalas on Jul. 20, 2007
Perhaps more extraordinary was the extreme humidity that accompanied the heat wave, which helped allow new all-time national records for highest minimum temperature to be set in every country in Europe except Italy and Greece, according to Herrera. His nomination for craziest record set in Europe was the 40.4°C (104.7°F) at Pinsk on June 29, which set a new all-time heat record for Belarus. Herrera commented:
Just think this:
All time record at Pinsk was 36.3°C in August 1905
??It took 121 years but it got smashed by 4.1°C
INSANE
"As of midday Thursday, the official forecast for Washington’s Reagan National Airport was for highs of 102°F, 104°F, and 103°"
Highs were 102 102 and 103. But those were all about 2F too high when looking at nearby stations. I saw the same effect when DC hit 100 on June 12th. It was about 2-3F higher than it should have been. Definitely not comparable to past records.
As for the 1930 heat mentioned in the article, that was a dry heat from the NW and won't be matched until there is a similar pattern. But that might be a while. The lows in 1930 were in the 70s whereas the low last week at National airport was 84, a tie for the all time warmest low temperature set IIRC in 2011 and 2012.
Bottom line the very humid conditions we saw last week will never allow 102 or 103 unless there is a stream of exhaust heat from the terminals with a west wind at National airport like there was last week.
Seriously, Eric? "...stream of exhaust heat"???
What the heck do you mean by "exhaust heat"?
Looking on Google Earth, I can see a weather station at Reagan Airport, located in the large grassy area to the east of the runways. The closest runway is about 250m to the west of the weather station, and the main terminal buildings are about 1km away.
At a standard height of 1.5-2.0m above the ground surface, the temperature at that location will be well-adjusted to the surface conditions approximately 200m upwind. There is no wind direction that crosses anything but grass field for 200m before reaching that weather station. The idea that there are plumes of hot air from "exhaust" washing over the temperature sensor seems highly unlikely, if not imaginative.
That weather station is probably only mildly affected by surfaces further away - be it the runway and terminal buildings to the west, or the body of water to the east. With a mix of built-up areas and water bodies in the few km around it, it is probably pretty representative of all the areas close to the major parts of central Washington and the rivers. You are going to have to be a lot more convincing to claim that the airport weather station is anomalously warm for central Washington.
You claim to have looked at "nearby stations". What ones, exactly, and exact locations, please.
Bob the comparison stations can be found here www.weather.gov/wrh/hazards?obs=true&wfo=lwx&hazard=true&hazard_type=all&hazard_opacity=80&tab=legend but some are poorly sited CWOP which can run too warm or too cool (e.g. in shade on a front porch). The five minute readings from DCA, IAD and BWI can be found here www.weather.gov/wrh/timeseries?site=KDCA&hours=72 Just replace the KDCA with KIAD and KBWI.
The problem at National (DCA) may be some parked jets but mainly exhaust from the AC units at the newly expanded terminals. The evidence is fairly obvious in the five minute readings when a fresh west wind (e.g. from convective outflow) will spike temperatures 2-3F with a west wind 10 or more mph after lighter winds most of the day. That happened June 12 and July 4th around 6pm after less hot temperatures earlier at "peak heat".
The problem at BWI is package delivery jet parking at the relatively new Amazon terminal S and SW of the sensor there. Please examine Google earth view for both BWI and IAD as you did for DCA. You will see local problems at DCA and BWI but not IAD. At IAD any possible heat production is north of the sensor.
Eric @ 3:
Thanks for providing additional details. A good many points to consider.
First, the data source you link to is useful in providing coordinates for the actual weather stations. Often, sources referencing airport weather stations will provide the coordinates of the Airport Reference Point (ARP), which rarely matches the actual weather measurement coordinates. That web page does make it easy to find the weather instrumentation at BWI (Baltimore/Washington) and IAD (Dulles). Unfortunately, the coordinates it provides for DCA (Reagan/Washington National) appear to be about 150m north of the actual weather station, based on Google Earth's mapping.
I agree that the instrument location at BWI is pretty poor - near the button of runway 10, between the runway and the apron to the south. Dulles is definitely better, being at least 500m away from most large pavement surfaces.
...but I disagree strongly with your interpretation of the Reagan data for July 4 around 6pm. (As of today, the 7-day data from the page you link to only goes back to July 2.)
Here is my graph of the July 4 temperature data - both air temperature and dew point (humidity).
Temperatures were pretty stable at 100F from noon onward, and then there is a slight rise to 102F before 6pm, a drop to 97F immediately after 6pm, and then a temporary rise to 99F before dropping to 94F (and lower) after 6:30pm.
Note that dew point also fluctuates during this period. Having varied between 68F and 72F in the latter part of the afternoon, it shows a drop to 65F at 5:40pm, then rises to 74F by 6:10pm, then drops more or less steadily to the low 60s in the evening.
There is more variation in atmospheric moisture around your 6pm time range than there is in atmospheric heat.
Let's look at wind direction. Your source lists wind direction in 22.5 degree increments (S, SSW, SW, WSW, W, etc.) I have transformed that into standard degree coordinates in this graph. (The wild variations in values between 0 and 337.5, early in the day, are just small variations around north).
You claim that the temperature spike around 6pm is due to west winds bring exhaust air from the airport facilities. But the timing of the wind shift does not match this. There is one 5-minute value of west winds at 5:40pm, but from 5:45pm to 7:30pm, the winds are almost entirely from the S and SW. There is an increase in wind speed around 6pm, but the stronger winds do not happen until 6:30-ish and later, when a thunderstorm moves through.
Using the correct location of the DCA weather station, the land surface to the SW is predominantly grass, runways, and (to the WSW) aprons (pavement) a few buildings, and parking areas. Anything that looks like "newly expanded terminals" on Google Earth is to the west (or north of west) of the weather station. It is not upwind for S or SW winds.
And frankly, the amount of energy coming off the AC units you speak of is going to be dwarfed by energy coming of large areas of pavement. Any hot plume from AC is going to be mixed in with the normal air flow - spreading both vertically and horizontally in the turbulent air.
So, your "hot AC plume" is not from the correct direction, it does not explain the variations in humidity (dew point), and it is physically unlikely. What I see is normal variations in temperature, humidity, and wind speed/direction on a hot summer day as some thunderstorms move in.
I think you are over-interpreting the data, trying to seek justification for a viewpoint that the data do not support.
[BL] When making later comments, I realized that the second graph in this comment suffered an incomplete copy-paste-revise process. The vertical axis units are wind direction in degrees, not temperature.
Thank you Bob for looking more closely at the data. The reading I remembered (incorrectly) as 6pm was actually at 5:40 Jul 4, 5:40 pm 102 65 30 106 W 9 The 9 or more mph west is the likely time when the 103 occurred although that five minute interval (and prior and subsequent) had 102 . The one minute data from Iowa Mesonet would likely reveal more but I'm having trouble accessing that.
The 103 likely coincided with the drop back to dew point of 65 where it had been earlier. In general lower dewpoints allow hotter temperatures. There is certainly a valid synoptic explanation for that: mixing, drier air, temporary heating. Such events can happen in small or large scales any time of day. Thus it need not be a plume to cause a 1F bump (the minimum possible delta).
While 103 was an outlier for northern Virginia: mesonet.agron.iastate.edu/sites/hist.phtml?station=DCA&network=VA_ASOS&mode=daily&year=2026&month=07&day=04 it was however not an outlier for MD. Andrews AFB to the SE was also 103 that day: mesonet.agron.iastate.edu/sites/hist.phtml?station=ADW&network=MD_ASOS
Eric:
This is a subject where precision is important, so please try to be very specific in timing of events, etc. In this case, there is a noticeable difference between 5:40pm and 6pm.
...but first, are you backing off your claim of AC exhaust plumes? Or are you still considering it as a viable explanation?
Returning to the Reagan July 4 data, the 5:40pm time you call to attention is the one 5-minute period where winds are recorded from the west. There are several factors to consider when interpreting that value:
A couple of relevant points on the temperature data. The 5-minute readings are (from what I can find) actual 5-minute averages. The reported maximum temperatures are based on 1-minute averages, though, so the maximum can exceed the highest 5-minute average. Also, there is a sequence of conversions between Celcius and Fahrenheit as data are collected, processed, and displayed. So rounding can occur.
Before looking at dew point data, I'll cover a bit of basic micrometeorology. In comment 2, I mentioned that temperature measurements at 2m height are responding to what the surface is doing over a distance of 200m upwind. This is important in considering the effects of the surface a short distance upwind.
The measured air temperature will still be dependent on larger-scale regional effects, though. The rates of adding heat and water vapour to the air from the local surface will have stabilized 200m downwind of the surface change, but ultimately the observed temperature and humidity at 2m will also depend on the ability of the atmosphere to move that heat and humidity up and away into the synoptic-scale air motions above that location.
So, to get back to the Reagan July 4 data, my interpretation is as follows:
To me, I see nothing in this data that suggests any effects that can be attributed to AC exhaust plumes or any other artifacts that are not part of a normal environmental response in a moderately complex area (combination of grass fields, water bodies, and built-up areas).
Also Eric @ 5:
In your first comment, you mentioned "nearby stations", but ended up bringing in Dulles and Baltimore-Washington, neither of which I would consider "nearby" (being 40-50km away). Now you mention Andrews AFB, which is closer but still 15km away. Each of these also has quite different meso-scale surroundings compared to Reagan airport.
Trying to make comparisons of the sort you are doing for stations that far apart and subject to different local and meso-scale environments is much more complex than you seem to be making it.
Bob, my reason for bringing up only IAD and BWI along with DCA is I look at them daily for precipitation as part of early morning commentary at a local weather site. While the June 12th 100 degree event happened I had the CWOP and other station site up looking for other 100 degree readings. I came up with the plume explanation on that day based on wind switching to the west and nearby comparisons.
I saw the Andrews max at 103 for the first time before the post with that link. I was otherwise occupied on the 4th, and not able to look at the local stations. A general explanation for July 4th is synoptic scale warming from thunderstorms blowing themselves out to the west, heating both DCA and a number of MD stations with hotter and a bit drier air.
A more specific explanation for DCA is that the 9mph west wind at 5:40 droppd the dew point from 20C to 18C due to a drier air source and the potential of drier air mixing down. You noted those possibilities and the complexity and I agree. An AC exhaust plume is not a good explanation for that.
Eric:
Another factor for short-term variations in local temperature is sunshine. The data source you linked to for DCA in your earlier comment does not include radiation measurements (it's not critical for aviation weather), but it does include cloud cover information. (I don't know if this is from human observers or an automatic system.) Most of the day on July 4 had clear skies. Cloud cover began to increase late afternoon, and some clouds remain after the brief rain/thunder spell around 6:30pm.
As for comparing temperatures regionally - still somewhat of a mug's game when it comes to absolute values. There is a reason why global temperature changes are evaluated using anomalies, not absolute values. There is an excellent 4-part series on this question starting here.
Regional precipitation patterns are even harder, as so many precipitation events (especially summer thunderstorms) are so localized. It's easy for significant events to slip between observing stations (or park over one and make it look like the whole region is getting dumped on). In Canada, a Canadian Precipitation Analysis product (CaPA) combines surface measurements, radar, and forecasting model results to provide precipitation maps and such. Documentation here, page with links to current maps here. I do not know if the US produces anything similar for public consumption.
As for the "exhaust plume" issue (to exhaust that discussion), there is much literature on the subject, from the perspective of pollution sources and how they spread locally or regionally. A basic model is the "Gaussian Plume". This link seems to give a decent overview. From a point source, the turbulent wind dynamics will spread a pollutant (or any property, such as temperature) both vertically and horizontally. As the spread increases, the concentration decreases. The further downwind, the harder it is to detect a weak source.
If the plume is hot (as you would expect for a source of heat, as you originally supposed) the plume will also tend to rise through the cooler air. At greater heights, turbulent mixing is increased as the mixing eddies increase in size.
What is seen at a point sink (e.g. a weather station) is the opposite: the further upwind you look, the wider the area the weather station sees. It will be getting air from a long line of (possibly differing) point sources.
Pollution spread science also looks at linear sources (think "highway vehicle exhaust in an urban area") and area sources (think "city surrounded by rural areas").
It is for these reasons that I rapidly discounted your AC heat plume hypothesis - before looking at data and hypothesizing alternate explanations for the temperature pattern you identified.