Analysis of new NASA AIRS study: 80% of U.S. Warming has been at Night

By Dr. Roy Spencer

I have previously addressed the NASA study that concluded the AIRS satellite temperatures “verified global warming trends“. The AIRS is an infrared temperature sounding instrument on the NASA Aqua satellite, providing data since late 2002 (over 16 years). All results in that study, and presented here, are based upon infrared measurements alone, with no microwave temperature sounder data being used in these products.

That reported study addressed only the surface “skin” temperature measurements, but the AIRS is also used to retrieve temperature profiles throughout the troposphere and stratosphere — that’s 99.9% of the total mass of the atmosphere.

Since AIRS data are also used to retrieve a 2 meter temperature (the traditional surface air temperature measurement height), I was curious why that wasn’t used instead of the surface skin temperature. Also, AIRS allows me to compare to our UAH tropospheric deep-layer temperature products.

So, I downloaded the entire archive of monthly average AIRS temperature retrievals on a 1 deg. lat/lon grid (85 GB of data). I’ve been analyzing those data over various regions (global, tropical, land, ocean). While there are a lot of interesting results I could show, today I’m going to focus just on the United States.

AIRS temperature trend profiles averaged over the contiguous United States, Sept. 2002 through March 2019. Gray represents an average of day and night. Trends are based upon monthly departures from the average seasonal cycle during 2003-2018. The UAH LT temperature trend (and it’s approximate vertical extent) is in violet, and NOAA surface air temperature trends (Tmax, Tmin, Tavg) are indicated by triangles. The open circles are the T2m retrievals, which appear to be less trustworthy than the Tskin retrievals.

Because the Aqua satellite observes at nominal local times of 1:30 a.m. and 1:30 p.m., this allows separation of data into “day” and “night”. It is well known that recent warming of surface air temperatures (both in the U.S. and globally) has been stronger at night than during the day, but the AIRS data shows just how dramatic the day-night difference is… keeping in mind this is only the most recent 16.6 years (since September 2002):

The AIRS surface skin temperature trend at night (1:30 a.m.) is a whopping +0.57 C/decade, while the daytime (1:30 p.m.) trend is only +0.15 C/decade. This is a bigger diurnal difference than indicated by the NOAA Tmax and Tmin trends (triangles in the above plot). Admittedly, 1:30 a.m. and 1:30 pm are not when the lowest and highest temperatures of the day occur, but I wouldn’t expect as large a difference in trends as is seen here, at least at night.

Furthermore, these day-night differences extend up through the lower troposphere, to higher than 850 mb (about 5,000 ft altitude), even showing up at 700 mb (about 12,000 ft. altitude).

This behavior also shows up in globally-averaged land areas, and reverses over the ocean (but with a much weaker day-night difference). I will report on this at some point in the future.

If real, these large day-night differences in temperature trends is fascinating behavior. My first suspicion is that it has something to do with a change in moist convection and cloud activity during warming. For instance more clouds would reduce daytime warming but increase nighttime warming. But I looked at the seasonal variations in these signatures and (unexpectedly) the day-night difference is greatest in winter (DJF) when there is the least convective activity and weakest in summer (JJA) when there is the most convective activity.

One possibility is that there is a problem with the AIRS temperature retrievals (now at Version 6). But it seems unlikely that this problem would extend through such a large depth of the lower troposphere. I can’t think of any reason why there would be such a large bias between day and night retrievals when it can be seen in the above figure that there is essentially no difference from the 500 mb level upward.

It should be kept in mind that the lower tropospheric and surface temperatures can only be measured by AIRS in the absence of clouds (or in between clouds). I have no idea how much of an effect this sampling bias would have on the results.

Finally, note how well the AIRS low- to mid-troposphere temperature trends match the bulk trend in our UAH LT product. I will be examining this further for larger areas as well.

0 0 votes
Article Rating
Newest Most Voted
Inline Feedbacks
View all comments
mario lento
April 30, 2019 2:00 pm

It’s UHI right???? Isn’t that obvious or am I being ignorant here?

Bruce Hall
Reply to  mario lento
April 30, 2019 3:08 pm

Exactly. This phenomenon does not occur at rural stations.

Reply to  Bruce Hall
April 30, 2019 4:41 pm

Disagree. It’s water vapor that’s keeping the nights warmer.

Reply to  rah
April 30, 2019 5:54 pm

The article claims that this phenomena reverses over the oceans. If it was just water vapor, you should see the strongest signal over the oceans, not a reversed one.

Reply to  MarkW
May 1, 2019 12:51 pm

Never mind. Over the oceans, the amount of water in the air would be pretty much constant so there would be no change. I withdraw my objection.

Reply to  rah
April 30, 2019 6:48 pm

It is the dew point in fact!

Robert B
Reply to  mario lento
April 30, 2019 3:24 pm

3% of land surface is urban. It is only a problem with thermometer readings in those areas.

mario lento
Reply to  Robert B
April 30, 2019 3:40 pm

That’s a good point Bruce. I wonder if the datapoints can be separated out to see where the delta between day and night is most evident. When all data is mashed together, could the effect of the urban heat islands be so strong that it affects the overall results?

Izaak Walton
Reply to  mario lento
April 30, 2019 11:34 pm

Do you realise that this post is about satellite measurements?

Old England
Reply to  Izaak Walton
May 1, 2019 12:07 am

So are you suggesting that satellite measurements are unable to differentiate between UHI in cities and the lower rural temperatures?

A breakdown between geographic urban and rural temperatures as measured by satellite would be very helpful. Intriguingly, as stated, the greatest night time warming is in winter, but without comment that is the season when houses and offices are heated massively with attendant heat loss to atmosphere.

On what had been presented, with no urban / rural differentiation I suspect that UHI is a very significant component ov the revorded “warmjng”.

Reply to  Izaak Walton
May 1, 2019 1:23 am

Do you realise it is data measured from the surface upwards?

Reply to  mario lento
April 30, 2019 3:49 pm

I live in the rural mountains of Northern California, 100 miles from the Pacific. The night time temps have also been warmer for years, although that has changed in recent years with night temps returning to average temps at times. Also, I spent many hundreds of hours looking at evening temps at other locations around the globe using weather sites who have historical data, and have noted a similar change where the above average night temps have returned closer to average, or even below average at times in some regions. There has been a definite change underway over the last 4 years.

Steven Mosher
Reply to  mario lento
May 1, 2019 11:48 pm

Not uhi.

Urban area is less than 5%.

Robert of Texas
April 30, 2019 2:03 pm

I am uncertain why the greater warming at night is at all a surprise. It is a prediction of both global warming caused by CO2, and the Urban Heat Island Effect.

The warming at an altitude at night would have to be due to either CO2 or additional water in the atmosphere (possibly as clouds) – can’t blame a nighttime rise of temperature on the Sun. People sometimes forget it isn’t one or the other – its both. Some temperature rise do to CO2 and some to UHIE which would explain nicely the sudden increase near the surface.

To separate UHIE and CO2, just look at a large rural area and see if the curve changes.

Or look at northern Canada…if its CO2 then nighttime temperatures in northern Canada have had to go up even more dramatically.

If its clouds, then coasts where cloud formation is likely would show the greatest changes. southern Texas would be a good location to look for that (Gulf warm air).

Dr Deanster
Reply to  Robert of Texas
April 30, 2019 7:35 pm

Try looking at the Sahara desert …. ain’t no water vapor, and probably not a whole lotta clouds either. ….. if it’s a CO2 signal, that would be where I would expect to see it.

James Clarke
Reply to  Dr Deanster
May 1, 2019 4:57 am

Maybe it is mostly the CO2 signal, demonstrating the benign nature of CO2 induced warming. The main impact of such warming would be longer growing seasons, wider growing regions and less chance of killing frosts and freezes during growing seasons.

If we weren’t adding CO2 to the air accidentally, we would have to do it on purpose! The benefits are extraordinary!

M Courtney
April 30, 2019 2:12 pm

My guess is some form of UHI related to more luxury – more 24hr heating.

Both day and night are warmed but night would be affected by UHI to a greater degree.

Reply to  M Courtney
May 1, 2019 4:06 am

the massive shift to reverse cycle aircons pumping out hot air? instead of more confined older radiant or indoor fan heaters?

Steven Mosher
Reply to  M Courtney
May 1, 2019 11:49 pm

It’s a satellites view. Less than 5 percent of what is viewed is urban

April 30, 2019 2:40 pm

cloud cover at night holds the heat in. Water Vapor is the culprit…if natural or artificial (chemtrails)… Greenhouse effect as cause of global warming is a nighttime phenomenon…the heat of the day is held in and is not going out into space. I am very concerned about the geoengineering that is going on, the weaponization of weather, chemtrails, and the “Global Dimming Project” being funded by Bill Gates.

Robert B
Reply to  Michael Phillip Miller
April 30, 2019 3:46 pm

Looks like the new approach is to take a lead from sceptics and plant a conspirator in the oppositions ranks to make them look dumb like we did with AOC.

Zig Zag Wanderer
Reply to  Robert B
April 30, 2019 7:05 pm

That’s funny, right there 🙂

Michael Jankowski
Reply to  Michael Phillip Miller
April 30, 2019 5:48 pm

“… I am very concerned about…the weaponization of weather, chemtrails…”

Stick to InfoWars and tinfoil hat functions.

Reply to  Michael Phillip Miller
May 1, 2019 4:08 am

even if it isnt chemtrails but just the massive increase in airflights heating the upper areas from engines and the vapourtrails as well.

April 30, 2019 2:41 pm

Dr Spencer, is there any indication cloudiness has increased? That clouds are forming at higher temperatures? If so that might explain the night time bias at least.

Reply to  Eric Worrall
April 30, 2019 4:47 pm

My understanding is that all the temperature data was collected when there was no cloud. Cloudless nights got warmer faster than cloudless days and the effect was more pronounced on the coldest nights.

mario lento
Reply to  Thomas
April 30, 2019 4:52 pm

Could it be said that the data intentionally then filters out much of the effect of clouds?

Old England
Reply to  Thomas
May 1, 2019 12:33 am

Seem to recall that the sensors can’t measure temperature through cloud.

Pop Piasa
Reply to  Eric Worrall
April 30, 2019 7:23 pm

Eric, where I live the dew point is what sets the nighttime low temps. When RH is 85% of saturation, there won’t be much cooling, clouds or clear (or shall we say hazy summer nights). The colder the air mass is, the more any moisture in it retards further cooling. That’s why the Arctic can’t get as cold in the winter, because at -60 F, a few molecules of water per cubic foot raise the temperature considerably.

Ian MacCulloch
April 30, 2019 2:45 pm

In China it is the change from the 40w light bulb to the more common 100w bulb as the country gentrifies. At night time this must translate into greater temperatures Dr Spencer has indicated. The sheer volume of ‘new’ power being added in China and India is huge. It is no surprise that it is going up at night.

Joel O’Bryan
April 30, 2019 2:52 pm

Increasing UHI effect across much of the US will also slow night time cooling to a greater extent than pushing up daytime warming. UHI can be thought of as more thermal mass (ever higher amounts of concrete, steel, asphalt) with more automobiles spewing heated exhaust and AC moving warmth to the “outside” air.

The vast open ocean areas and vast deserts with no urban growth is where AGW would be seen as a signature of GHE warming, separate from increasing UHIE.

Clyde Spencer
Reply to  Joel O’Bryan
May 1, 2019 9:54 am

Along with CO2, cars also emit considerable water vapor. While it has a short resident time, it is being replenished continuously.

Steven Mosher
Reply to  Joel O’Bryan
May 1, 2019 11:51 pm


Clyde Spencer
Reply to  Steven Mosher
May 2, 2019 12:29 pm


Thank you for your informative, authoritative response that so succinctly demolishes Joel’s opinion.

Do you really think that anyone here so highly values your opinion that a single word is an adequate response to what wasn’t a solicitation of your opinion?

mario lento
April 30, 2019 2:58 pm

UHI mainly affects night time temperatures by buffering cooling as heat is released out of the heated surroundings…

April 30, 2019 2:59 pm

This has been told many, many times. Global warming is not global. It is mainly over land and not over sea. It is mainly in the Northern Hemisphere and not in the Southern Hemisphere. It is mainly during winter and not during summer. And it affects mainly minimal (night) temperature and not maximal (day) temperature.

It does not fit the theory on how global warming should work, and since it affects mainly winter-night temperature it is not dangerous in the least. Quite the contrary.

Reply to  Javier
May 1, 2019 5:56 am

Excellent summary. Why are the facts ignored by the MSM and Democrat politicians who constantly spew of gross exaggerations? Also even in the Northern Hemisphere the warming is greater in the ARCTIC and northern climes.

HD Hoese
April 30, 2019 3:00 pm

That makes sense if you go buy killing freezes in the Gulf of Mexico as we have not had one since 1989. This is when there is ice on the bays, shallow water especially in lower salinity. Fish die in significant numbers, shallow water oysters the same. Effect, often blamed on other things, is obvious in populations for a few years. They vary some in intensity and location falling a little more to the east or west. Last bunch 1940, 1947 1951, 1962, 1983, 1989. 1977 more towards Florida. 1962 and 1899 were real bad. Fed the Taylor army in 1845-46 with fish and turtles. Minor freezes since 1989 have had relatively small effects. Interesting to study however, as sea levels fall from very high pressure. Before modern regulations, people could gather them for food. Nice and fresh, well refrigerated.

Sweet Old Bob
April 30, 2019 3:02 pm

Read Willis’ post . Yesterday .
Increased clouds/water vapor .
Looks like he’s ahead of the curve again ….. 😉

Peter Fraser
April 30, 2019 3:03 pm

If I remember correctly the late John L Daly suggested in the 1980’s that one way of assessing CO2 impact on global temperature was to set up a large number of maximum/minimum thermometers worldwide at carefully selected sites away from UHI effects. A narrowing of the gap between Max. and Min. would indicate the effect of increasing CO2

Reply to  Peter Fraser
April 30, 2019 5:50 pm

Yes indeed Peter, a great citizen scientist . . .
For those who don’t know his work, check out:

Steven Mosher
Reply to  Peter Fraser
May 1, 2019 11:53 pm

Yup. We see that

April 30, 2019 3:15 pm

I know. Let’s spend $5 trillion on more roads, parking lots, and dense urban housing project incentives.

R Shearer
April 30, 2019 3:21 pm

CO2 is a lady of the evening don’t you know?

April 30, 2019 3:22 pm

Fifty years of riding a motorcycle have taught me that UHI effect is growing here in the US of A and much more noticeable at night.

mario lento
Reply to  JimG1
April 30, 2019 3:42 pm

And 30 years of me driving a convertible have shown me the same. It’s abundantly obvious that there are a good several up to and over 10 degrees F difference in night time temperatures near man made thermal masses. Highway overpasses come to mind.

Randle Dewees
Reply to  JimG1
April 30, 2019 3:51 pm

48 years of riding motorcycles has taught me to get and use an electrically heated vest

Steven Mosher
Reply to  JimG1
May 1, 2019 11:53 pm

Not uhi.
This is a satellite view of the entire surface

Solomon Green
April 30, 2019 3:32 pm

When I look at a satellite picture of the world at night (not that I can understand how they manage to film so that it is night at all points) I am struck by how much of the Northern Hemisphere and other populated areas is glowing. Is this light produced without hear?

Reply to  Solomon Green
April 30, 2019 4:06 pm

That brightness is mostly produced by overexposure. From those satellite photos, one can calculate approximately what percentage of the Earth is the same temperature as the surface of the Sun. Incorrect of course.

R Shearer
Reply to  Solomon Green
April 30, 2019 4:10 pm

Yes, light is silent.

Reply to  R Shearer
April 30, 2019 5:03 pm

…. but deadly

Sorry, I couldn’t resist.

Clyde Spencer
Reply to  philincalifornia
May 1, 2019 9:58 am

Deadly if you are a vampire.

Hank McCard
April 30, 2019 3:34 pm

I share Dr. Spencer’s suapicion:

“If real, these large day-night differences in temperature trends is fascinating behavior. My first suspicion is that it has something to do with a change in moist convection and cloud activity during warming. For instance more clouds would reduce daytime warming but increase nighttime warming. But I looked at the seasonal variations in these signatures and (unexpectedly) the day-night difference is greatest in winter (DJF) when there is the least convective activity and weakest in summer (JJA) when there is the most convective activity.”

Robert B
April 30, 2019 3:38 pm

Not UHI which only effects thermometer readings in urban areas.

My guess would be drought followed by heavy rains and snow at the end of the trend period. An independent researcher found a good correlation of maximum temperatures with rainfall in Australia. The more humid air keeps maximum temps down. Contra to that is that cold fronts bring in rain but most of the rain comes from huge downpours. There was no correlation with minimum temperatures even though you would expect more humidity to keep temperatures up. I’m guessing that is unique to Australia where half the country gets most of its rain in monsoonal downpours in warmer months when it doesn’t get close to the dew point at night. The US might be significantly different but if its possible, it would be good to look at different zones and rainfall.

April 30, 2019 3:41 pm

What I have always thought about this phenomenon is that the oceans are the reason for the night time warming. Surface winds carry the warmed air over land in the form of water vapor. In the daytime that leads to a slight cooling, while at night it means warmer temps. I think the current offshore surface winds are an example of how this works.

Note how there are no clouds offshore until way down by Los Angeles area. The surface winds then turn eastward and clouds immediately form, and then move east across the US. I have been watching this for the last 5 days. Prior to that the surface winds flowed south to around Monterey approximately, before turning to the west. Clouds then formed and were driven back up the middle of the Pacific towards Alaska and the Aleutian islands. Thus the warmer Alaska which alarmists like to make a big deal over. That has also meant warmer air flowing in to the west side of the Arctic which has led to reduced sea ice extent mainly in the Bering Sea. Thus the alarmists crowing over sea ice loss. I would like to hear their explanation how CO2 can drive surface winds. …,37.03,1107/loc=-122.748,34.888

mario lento
Reply to  goldminor
April 30, 2019 3:50 pm

We are talking about a warming trend. Given that the data shows a warming (nighttime) trend, that implies (not proves) something is changing that is causing that. One thing that we know is changing (increasing) is urban development. This factor cannot be ignored… and one would think that 3% of the landmass where there is a potential multiple degree warming trend would contribute to the overall trend.

Reply to  mario lento
April 30, 2019 5:12 pm

And so my comment that warmer oceans is what leads to warmer temps at night for rural areas such as where I live. There is no UHI up here. I agree that UHI around urban areas is a strong effect. An example of what warmer night means for the mountain area where I live is that back in the 1960s/70s it was hard to grow tomatoes in Trinity County because the temps at night were much cooler than in recent decades. Night time temps have been running at least 6 degrees F above average from prior to the 1990s.

Over the last 9 years where I have lived up here it has been easy to grow nice tomatoes with the exception of last year where there was a definite change, and almost no spring time warmth. Many of the locals struggled with their tomato plants last year. This spring was a bit warmer earlier on, but cool spells keep intruding. Temps dropped 6 F to 38F in the last 2 nights.

Tonight is forecast to drop to 36 F and 35F tomorrow night. Then evening temps are slated to rise 10 F back into the mid 40s on Thursday. Typically, a gardener never set their plants out before mid to late May in this county. I will have to cover my tomatoes for the next several nights.

mario lento
Reply to  goldminor
April 30, 2019 5:20 pm

Goldminor: I did miss that you have no U affect in your area. Your information is terrific! Wow… there seems to be a real trend that you have felt because you are tuned to it. Farmers really need to know their conditions especially with organic sensors such as tomatoes. Thank you!

Dennis Sandberg
Reply to  goldminor
April 30, 2019 7:53 pm

Thanks for explaining the lack of ice extent in the Bering Sea. I recently noticed those high billowing clouds between Paso Robles and Monterey. Beautiful but unusual here on the Central Coast.

April 30, 2019 3:50 pm

It would be very interesting to see the difference between Arid and Tropical sites as this would help identify if the culprit is actually water ?

April 30, 2019 3:55 pm

Anecdotally, I would whole heartedly agree with that assessment for my area in central Indiana. I can’t remember so may consecutive warm nights as we had during last summer. Usually on a Friday or Saturday night on a weekend you’ll find me and my neighbor and some family or friends out around one of the fire pits solving the problems of the world and on clear nights , checking out the night skies with my telescope. Last summer it was just too darn warm to enjoy a fire almost every evening.

April 30, 2019 4:00 pm

The nighttime warming effect is much more noticeable in built-up areas and is likely due to the Urban Heat Island Effect. Increasing clouds and water vapour would add to the effect. All this leaves a very little role for the dreaded CO2.

Loren Wilson
April 30, 2019 4:01 pm

We have balloon data that give humidity over this same period. What does it say about humidity in the atmosphere over the U. S. of A.?

Rud Istvan
April 30, 2019 4:25 pm

Would be UHI except there isn’t nearly enough U.
Would be WVP except humidity lowers at night as air cools.
So, the science IS NOT settled.

mario lento
Reply to  Rud Istvan
April 30, 2019 4:37 pm

Isn’t 3% of a big number, significant when we’re talking about fractions of a degree change? I do not think it is necessarily true that there is not enough U.

Gary Pearse
Reply to  mario lento
April 30, 2019 6:17 pm

Mario. Especially when there is a clustering of thermometers in Urban areas. Also suburban is less heated, but more than rural. IIRC, Roy did a study that showed the effect present in communities of less than 100,000. I’d give the UHI effect maybe double 3% or more.

Randy Bork
Reply to  Gary Pearse
April 30, 2019 8:29 pm

The data in this study isn’t from thermometers, in urban areas or anywhere else. This is looking at the IR spectrum from the AIRS satellite.

mario lento
Reply to  Randy Bork
April 30, 2019 10:15 pm

Yes, that is correct. Is their any position granularity that can be gleaned from the AIRS data? That may help dispel the UHI issue.

Gary Pearse
Reply to  Randy Bork
May 1, 2019 5:20 am

Randy, no, but the its the source of the heat! detected by AIRS.

Reply to  Rud Istvan
April 30, 2019 4:44 pm

“except humidity lowers at night as air cools.”

This is not true. For example, if the air has 30% relative humidity at 50F, when the temperature drops to 40F, the relative humidity increases. This continues until the air temperature drops low enough, reaching the dew point, then, and only then does the water vapor condense out of the air. When the dew point is reached the relative humidity is at 100%.

mario lento
Reply to  Mike Borgelt
April 30, 2019 4:50 pm

All true about RH vs AH (relative humidity vs absolute humidity)… Another thing that most non science people miss is that while AH decreases normally reduces as it cools… RH may increase when it cools. During this period, heat of condensation is released, thus resisting said cooling. Conclusion, AH does in fact store heat and resist cooling.

Clyde Spencer
Reply to  Rud Istvan
May 1, 2019 10:06 am

How does RH or AH lower as the air cools without precipitation? I would expect the RH to increase and the AH to stay constant. What am I missing?

Richard M
April 30, 2019 4:41 pm

One negative feedback to any kind of surface warming is convective cooling. This occurs predominately at the warmest times of the day. The atmosphere is always trying to cool things off using buoyancy and latent heat.

At night it really isn’t warm enough to have much convective cooling going on. Since, winters are cooler to begin with and have more night time hours you should see less convective cooling during those periods as well. So, I believe Roy’s first thoughts were mostly correct but didn’t take in account the longer nights that allow them to warm more.

This is more evidence of negative water vapor feedback to whatever is causing the warming. Note in this case a lot of the warming “trend” was created by the recent super El Nino. The planet’s hydrology didn’t care, it still tried to cool us off.

April 30, 2019 4:44 pm

How much of this is due to irrigation? I know a lot of the farming areas are more humid places than they were in years past, thanks to irrigation. Throw in lawn watering, and that’s a lot of extra moisture being put into the air.

Steven Mosher
Reply to  ToddF
April 30, 2019 6:04 pm


Well there are are GIS datasets showing the part of the planet that are irrigated
If you take the temperature stations that are in irrigated areas and compare them to non irrigated
there is no statistical difference.

OR, you can search around and find data that comes exlcusively from Ag nets
( temerature stations NOT used by NOAA etc)
and you can compare these to the stations used by NOAA.


No difference

Been at this since 2009, 10 years of looking at these kinds of things.


Reply to  Steven Mosher
May 1, 2019 1:33 am

Oh ten whole years! That’s like 0.0000000000000001% of the time we need to understand climate. Keep going, you’ll get there one day.

And yes, it may well be getting warmer. But relative to what? When it was cooler? The 1930s? The MWP? The Roman Warm Period?

Reply to  Phoenix44
May 1, 2019 5:45 am

I think “Mosher’s” point is don’t point to irrigation as a cause of warming if you want to contribute substantively to the skeptical position.

Clyde Spencer
Reply to  sycomputing
May 1, 2019 10:21 am


There is little doubt that places like Phoenix are more humid than they were 50 years ago. Besides the abundant well-watered golf courses, WV from automobile tailpipes, and the proliferation of swimming pools, there are the ubiquitous misters at gas stations, bus stops, and people’s patios.

While the actual temperatures may not be increasing, the heat index surely is. More importantly, the evaporation of the abundant water transfers the heat upwards and downwind and, while in the air as the vapor phase, WV helps retain heat radiated from the surface.

Reply to  sycomputing
May 1, 2019 11:05 am

While the actual temperatures may not be increasing . . .

I suspect this is “Mosher’s” point. Unless you would argue that a “heat index” is a more accurate measurement of warming than actual temperatures.

Clyde Spencer
Reply to  sycomputing
May 1, 2019 11:24 am

The latent heat of evaporation may be stabilizing the temperature at the source areas of WV. However, that doesn’t mean the increased water vapor isn’t impeding cooling that would otherwise take place. The radiated heat is absorbed by the WV and transported elsewhere. So, while the source areas may not be warming, that doesn’t mean that the globe isn’t heating from the increased WV.

Mosher was typically cryptic. What he didn’t say may be more important than what he did say. That is, he denied that there is a statistical difference in temperatures between irrigated and non-irrigated areas. The implication is that is the annual, average temperature. What is happening to the TMin temperatures? How is it possible to find microclimates that are comparable to irrigated areas that aren’t under cultivation unless there is sufficient water at the surface to not need irrigation? Thus, there is plentiful water to supply WV through evapo-transpiration. In my example (elsewhere) of Phoenix, the immediate metropolitan area doesn’t grow as much as it did 50 or 60 years ago, but the agricultural water has been supplanted by numerous other sources. Is Phoenix one of the “non-irrigated” areas used to compare to irrigated areas? Mosher’s statement is virtually meaningless without all the details.

Joel Snider
Reply to  sycomputing
May 1, 2019 12:34 pm

‘Mosher was typically cryptic. What he didn’t say may be more important than what he did say.’


Steven Mosher
Reply to  sycomputing
May 2, 2019 1:22 am

Yup. There are a few places where you can see the difference. Not a big issue

Clyde Spencer
Reply to  sycomputing
May 2, 2019 12:47 pm


Dale Quatrocchi has demonstrated that urban air masses affect weather for tens of miles downwind. If you are comparing the temperatures of urban areas as defined by corporate boundaries, or even impervious surface percentages, then you are not getting the right results because the rural areas are being contaminated with the urban air masses.

Also, one has to define just what it is that is being warmed. The ground level air is warmed substantially by dark surfaces such as asphalt, whereas irrigated areas are cooled at the surface by evapo-transpiration. However, the water vapor-enriched air rising especially from the irrigated areas offers more opportunity for absorption of IR at night, thus warming the upper air downwind.

Anecdotally, years ago, I used to commute regularly between San Jose (CA) and Sacramento. My car didn’t have air conditioning, so I usually drove with the window down. Somewhere around Davis I would encounter a noticeable drop in temperature, probably around 10 deg F. I would look over and there were several acres of well-watered alfalfa. Clearly, irrigated fields have a significant impact on the local temperatures, although it might not show up in a 25 x 25 Km square.

The coarseness of sampling may be why the impact isn’t more noticeable to you.

Reply to  Steven Mosher
May 1, 2019 3:19 am

And before that (1940-1980)

Anthony Banton
Reply to  Graemethecat
May 1, 2019 6:08 am


Because there were -ve forcings in place that overrode the +ve of CO2 – which was only at 310ppm in 1940 (0.5 W/m^2), and 340 ppm in 1980 (1.2 W/m^2).
It’s now ~410ppm or 2.1 W/m^2.
comment image
comment image

And increasing SO2 aerosols ….
comment image

Then we have the influence of the PDO ….
comment image

Reply to  Steven Mosher
May 1, 2019 4:55 am

Um, they have looked at places like rural Iowa and the central valleys of California. They are moister. Their nights are warmer. There is a lot of crop irrigation.

Steven Mosher
Reply to  ToddF
May 2, 2019 1:24 am

Yes. Isolated cases.
Not global.

It’s getting warmer.
There was an LIA

Clyde Spencer
Reply to  Steven Mosher
May 1, 2019 10:12 am


An inquiring mind might want to understand just why increasing humidity doesn’t cause an increase in temperature. Might it be the latent heat of evaporation is cooling the surface while transporting the heat up and downwind?

Joel Snider
Reply to  Steven Mosher
May 1, 2019 12:10 pm


Or are you just back here arguing your angels on the head of a pin again?

Michael Jankowski
Reply to  Steven Mosher
May 1, 2019 5:24 pm

Gee Mosh, it took me ten seconds to find opposing claims.

The first is from 2008, before you had been “at this.”

The second is from 2015.

The third is from 2017.

No, irrigation does not create warming…but it certainly affects temperature and apparently even climate.

Steven Mosher
Reply to  Michael Jankowski
May 2, 2019 1:26 am

Does not effect the global mean

And yes you can find isolated areas where it makes a small difference.

We are about 1.2 c warmer than 1850.


70 percent of the globe is ocean


Joel Snider
Reply to  Steven Mosher
May 2, 2019 8:50 am

And isn’t the majority of C02 coming from the ocean – something a bit less than 60%?

Michael Jankowski
Reply to  Steven Mosher
May 2, 2019 5:00 pm

FFS you claimed, “If you take the temperature stations that are in irrigated areas and compare them to non irrigated
there is no statistical difference.”

FFS the links I provided suggested otherwise.

Now FFS you’re claiming these are just “isolated areas where it makes a small difference.”

FFS here is another that finds a small global impact

FFS it also says irrigation can match or exceed the effect of GHGs at the regional level.

FFS I never said “it is irrigation” that is causing warming or a difference between the trends in day vs night temperatures, just said FFS that your baseless and unsubstantiated claim was false.

April 30, 2019 4:54 pm
Reply to  Chaamjamal
May 1, 2019 1:28 am

I can confirm that for SE Australia, Tmax shows a distinct Pause in the 21st century, whereas Tmin has warmed in a bit in that period.

Richard M
April 30, 2019 4:58 pm

I think it would be informative to look at the data from 2003 up to 2014. This was before the super El Nino. Over this period I doubt there is much of a trend in the overall data. So, is there still a difference in the day-night values? Are the days cooling and the nights warming?

If the answer to these questions is yes then it could provide some answers. It would tell us how much of the trends of both day and night were not due to the El Nino and how much was. It may be this would provide good data that we need to understand climate sensitivity. It would also tell is how much of the warming was due to the El Nino.

David S
April 30, 2019 5:00 pm

I don’t get it. From wood for trees the linear trend for the UAH data over that time period is .0155C per year or .155C per decade. But the graph looks like it’s about double that. What am I missing?

Gary Pearse
April 30, 2019 5:04 pm

Joel, certainly they should use AIRS to check out the responses from the middle of the Sahara, the middle of the Amazon/Congo, boreal forests, Antarctica, mountainous areas – all uninhabited regions, and then in urban, suburban, and rural areas of a variety of climate conditions. Who said you couldn’t do experiments in climate?

I think it’s high time we made real scientific use of all this superb instrumentation to advance the science. Even this simple example proves to be surprising to a climate scientist of Roy Spencer’s caliber. Willis uses satellite data in interesting ways and manages to surprise himself every time with the results. When I read the literature, I see mainly that the science of Tyndall and Arhenius hasn’t been improved upon vis a vis CO2 GHG effect. Com’on, let’s finally get a handke on this science. The “Team” actually has no intention or motivation to get at the truth it seems.

April 30, 2019 5:19 pm

I would further want to investigate solar cycle/cosmic ray effects on cloud formation. I have long suspected the effects of cloud cover drive our climate much more than previously understood.

Steven Mosher
Reply to  SteveC
April 30, 2019 5:59 pm

One reason I got the AIRS data was to study cosmic rays

NO effect. no matter how you cut the data

Clyde Spencer
Reply to  Steven Mosher
May 1, 2019 10:26 am

You said, “NO effect. no matter how you cut the data” At least none that you could find. “Lack of evidence in not evidence against something.”

Steven Mosher
Reply to  Clyde Spencer
May 2, 2019 2:29 am

Clyde people predicted that a change in GCRs would change the clouds.

I looked found nothing.

They made a prediction. I looked.


at some point you actually need some evidence for your theory,
other than your desire for it to be true.

Now, YOU never looked. you never practiced self skepticism
You just believe your theory because you read something.
maybe a paper.


I’m skeptical. There was cloud data … I looked.




from some of the best data there is.

global… nothing
NH? nothing
SH? nothing
Different latitude bands? Nothing
low clouds? nothing
high clouds? nothing
over land? nothing
over ocean? nothing
individial grid cells? nothing


Clyde Spencer
Reply to  Steven Mosher
May 2, 2019 12:57 pm


In other words, there were no obvious, first-order effects that jumped out at you after looking at several data sets, and you feel confident that you have demonstrated there are no effects. Once again, “A lack of evidence is not evidence against something.”

You are making assumptions about my past history that are unwarranted as you know nothing about my background or work history. Do you do that with everything?

It might take a more sophisticated approach to teasing out the relationships between variable such as temperature, humidity, clouds, and GCRs than what you have applied.

Steve O
April 30, 2019 5:33 pm

What is the temperature impact of complete cloud cover?
Is it different in the winter compared to the summer?

April 30, 2019 5:46 pm

The plot indicates that UAH LT temperature trend is about 3.4 °C/decade. Is it correct? This UAH plot shows lower global trend IMO. Does it mean that the US is heating up faster?

April 30, 2019 5:47 pm

From the suburbs of Northwest NJ to New York City there is about a 3C temperature increase. I fear the alarms about the end of humanity from such warming are already in evidence – there is something drastically wrong with many of the inhabitants of NYC, starting with their mayor and recent Congressperson.

Reply to  BobM
April 30, 2019 9:08 pm

Catastrophic Anthropogenic Political Climate (PC) Change

As for circumstantial warming, there are open questions. Is it normal? Is it beneficial? Is it anthropogenic, and, if so, how do we sustain it? Will disassembling high-density population center heat islands help people?

Steven Mosher
April 30, 2019 5:56 pm


I did the Airs work a few years back comparing both Skin and SAT products


I was going to investigate Min Max and compare to UAH, but at that time I couldnt figure out
How to get Absoluate T from your files. I figured that out now.

April 30, 2019 5:59 pm

Here in Raleigh during the winter every house and building is heated. Exhaust gases from NG furnaces constantly pour into the atmosphere heating the air. The daytime requires much less heating. The heat generated at night is not inconsequential.

Donald Kasper
April 30, 2019 8:40 pm

First, there is no instrument at each elevation so the sensor in the satellite is getting convolved sample set of all values integrated together. Then some principle are applied to a deconvolution operator to say how much of the signal comes from each elevation of the atmosphere. So from a sensor standpoint at the top of the graph looking down, the convolution operator error increases with distance from the sensor, essentially exponentially worse. Nothing else is going on. Day vs night? Infrared specular reflectance effect which increases closer to the surface. Essentially, IR scatter.

April 30, 2019 11:43 pm

“the day-night difference is greatest in winter”

Warming always was predicted to be greatest at the cold times and places. It seems this has be confirmed.

It also means that warming means less cold, not more warmth. So in fact global warming is a reduction in extremes of temperature. It means plants an animals are not exposed to such cold temperatures, which is surely beneficial to them.

Reply to  MattS
May 1, 2019 1:41 am

Agreed, that is the conclusion I arrived at when I first heard of global warming (about 1990).

Warming reduces the problems of extreme cold a lot more that it increases the problems of extreme heat; a net benefit.

Also weather is driven by temperature differences, so raising minimum temperatures more that maximum ones reduces extreme weather.

donald penman
April 30, 2019 11:43 pm

I think that because we are using a large land mass here that there will be a difference between night and day in the south and north both in winter and summer but I don’t know how that would affect the data.

Mike Maguire
May 1, 2019 12:27 am

It’s not the UHI effect.

The exact same thing is going on at the highest latitudes but is stretched out into a seasonal variation vs diurnal because of the great disparity in the length of day between Winter and Summer.

Look at the Artic for instance. It’s the Winters that have become much warmer because the Winters are an extended version of night. The Arctic Summers, however, have not warmed that much. Not coincidentally, they are also an extended version of daytime with solar radiation around the clock(though from a low angle).
There is no UHI effect in that part of the world.
I believe it’s from increased water vapor/higher dew points. Put more moisture into the air and it reduces the spread/disparity between daytime temperatures…….which warm slower and nightime temperatures ……that cool slower.
It takes more solar energy to heat humid air and that humid air takes longer to cool it at night/radiate heat out because it’s holding more energy.

Mike Maguire
Reply to  Mike Maguire
May 1, 2019 6:09 am

We can all agree that H2O is the main greenhouse gas. This diurnal temperature signature is exactly what you see when you increase water vapor.

It’s more amplified at the higher latitudes because the air is driest there and the contribution of increased water vapor is more significant.

As an operational meteorolist for 37 years, I’ve noted big positive temperature anomalies over the highest latitudes for very long stretches in the Winter months at both the Arctic and the Antactic.
In the Summer months, it’s just the opposite, with negative temperature anomalies often prevelant and almost never the positive anomalies that we commonly have witnessed during the Winters for the past 2 decades

Mike Maguire
Reply to  Mike Maguire
May 1, 2019 6:24 am

You will also note the same difference between regions with dry air masses compared to more humid regions. The diurnal temperature spread decreases with increasing H2O in the air.

What would happen to our planet if we took all the water vapor out of the atmosphere?
Nights would be brutally cold.

Just the opposite is happening because of a bit of an increase in water vapor across the globe, with the effects greatest over the driest areas……..high latitudes……land.

May 1, 2019 12:48 am

One thing that has changed a lot during this time is aerosols.
If aerosols cools when present it will give less cooling when lacking.
Less cooling is warming.
But why not warming at daytime?

May 1, 2019 1:50 am

Interesting stuff. I am slightly confused by using a decadal trend for 16 years worth of data though? That seems really inappropiate for a data set less than two “units” long. Its also intriguing and rather counter-intuitive that higher nighttime temperatures don’t feed through into higher daytime temperatures. It suggests a high degree of independence that I would not have suspected.

Tim Gorman
May 1, 2019 6:32 am

I have collected 5 minute temperature data at my rural KS site on a 24/7/365 basis since 2002. My data shows very similar results as Dr. Spencer shows. My data shows that daytime temperatures are actually trending downward and not upward as Dr. Spencer shows. My data also shows that nighttime temperatures, however, are trending up. A linear regression on all of the data shows a slight upward trend, mainly fueled by the higher nighttime temperatures.

I don’t have the knowledge to explain why this is but I believe that our rural agriculture transition away from low evapotranspiration crops such as wheat, milo, and pasture grasses to higher evapotranspiration crops such as soybeans, rice, and corn has something to do with it. This evapotranspiration provides water vapor that cools during the day but insulates during the night. One result of this is longer growing seasons. That is not a bad thing.

I also suspect that this transition to high evapotranspiration crops is happening on a global basis as these crops represent high energy crops that can feed a growing population. Is this enough to cause what Dr. Spencer has identified? I simply don’t know. But it is possible that it is a contributing factor.

Mike Maguire
Reply to  Tim Gorman
May 1, 2019 9:01 am

Great observations. This also effects the entire Cornbelt/Midwest in similar fashion………..during the growing season.

Tightly packed rows of corn(twice as dense as 40 years ago) are adding tremendous amounts of moisture to the air from evapotranspiration……….especially during the months of June/July when evapotranspiration peaks.

From September to the following April(into much of May for seasons like this one), this can’t be a factor because the crop is no longer actively growing.

Tim Gorman
Reply to  Mike Maguire
May 1, 2019 10:02 am

Mike, the graph appears to show the average deviation from some seasonal average cycle which I’m not sure its origin. If the deviation of the nighttime temperatures go up during the summer and fall seasons then the overall average will go up as well. It’s like have a data set of 1,2,3,2 for the four seasons (winter, spring, summer, fall) and then having a dataset of 1,2,4,3 equivalent to having higher deviations in the summer and fall due to evapotranspiration. The average deviation goes up, it doesn’t stay the same or go down. So just because it happens during part of the year doesn’t prevent it from impacting the overall average deviation.

Joe G
May 1, 2019 6:54 am

OK, wait. The Sun heats the Earth during the day. When the Sun goes down, the Earth becomes the heat source and emits its blackbody radiation. That’s when the GHG’s are at work.

That means we would expect what the NASA graph depicts

Mark Pawelek
Reply to  Joe G
May 1, 2019 8:36 am

No. Earth emits blackbody radiation all day and night. It emits most OLR during the day because it is warmest during the day! According to theory GHGs work 24/7/365. They have no holidays.

50% of weather stations are located in urban or urbanized areas like airports. But globally only 5% of earth’s population live in 95% of land area. So 95% of the population live in only 5% of the area! In a country such as UK, 3% of our land is built on. It seems likely that some 50% of weather stations are located in or near the 5% of land where 95% of people like; and that the other 50% of weather stations are in the 95% of land where only 5% of people live.

I suggest we move all weather stations away from built up areas. Stop measuring daily average temperature by taking the average of the highest and lowest readings for the day. Outfit all weather stations with hourly, or half-hourly recording equipment. Get them to phone in, or use satellite relays. That way we avoid human error too when measuring. Ensure all stations use 3 separate thermometers which work on 3 separate temperature measuring principles. In that manner, each acts as a control for the other two. This kind of thing should’ve been done decades ago. Stop wasting money on useless desk-bound computer gamers. Start spending money gathering excellent quality data.

The Urban heat island effect best explains this warming at night.

Tim Gorman
Reply to  Mark Pawelek
May 1, 2019 10:08 am

+1. I would only add that an average daily temperature, even if determined by hourly readings, is useless for determining what is actually happening. You can calculate an average speed for a train that leaves Point A and derails at Point B. But it won’t tell you that the train was going to fast for the curve at Point B!

Clyde Spencer
Reply to  Mark Pawelek
May 1, 2019 10:36 am

Mark Pawelek
But, the nighttime temperatures coincide with an emission peak that is nearly identical to the overlapping absorption features of H2O and CO2, thus closing the door. However, the daytime summer highs move the emission peak to a region that is more transparent. That is, a hot daytime temperature can radiate to space more effectively than a cool nighttime temperature.

Mark Pawelek
Reply to  Clyde Spencer
May 2, 2019 4:51 am

My post above is in error. Dr Roy is almost certainly right: higher night surface temperatures are due to more cloud cover. Of course – this is something one needs to show. No evidence suggests that, with a clear sky, an atmosphere with 400ppm CO2 will be noticeably warmer at night than one with 300ppm.

Steven Mosher
Reply to  Mark Pawelek
May 2, 2019 2:22 am

“The Urban heat island effect best explains this warming at night.”


The AIRS satillite views all of the USA at a 25km resolution.
The portion of US teritory that is urban is TINY,

For example

For every 25km grid that is urban, there are 20 that have zero people

Clyde Spencer
Reply to  Steven Mosher
May 2, 2019 1:14 pm


A bit of hyperbole! Except for some of the highest mountains and hottest deserts, there are probably few areas in the country that have “zero people” Even Death Valley has permanent residents and transient tourists year round.

Compared to even Landsat with a nominal 28m pixel resolution, 25KM is extremely coarse. You can’t see the trees, only the forest. When doing multispectral classification, resolution can affect the result significantly because with Landsat, shadows are a part of the spectral signature. At the resolution of SPOT or the Worldview satellites, shadows become a separate spectral class that are only separable by context, as with object-oriented classifiers.

The approaches that you have used suggest that you have little direct acquaintance with remote sensing.

Steven Mosher
Reply to  Mark Pawelek
May 2, 2019 2:23 am

“50% of weather stations are located in urban or urbanized areas like airports. ”


Joe G
Reply to  Mark Pawelek
May 2, 2019 7:00 am

The earth emits only when it is the heat source, that is when it is hotter than the atmosphere and incoming solar radiation.

Clyde Spencer
Reply to  Joe G
May 2, 2019 5:39 pm

Joe G,
You are wrong. Any body above absolute zero will radiate, with the peak IR emission governed by the temperature. I think you are confusing IR radiation with conduction, where heat moves from the higher temperature to the body with the lower temperature.

May 1, 2019 6:55 am

So Earth Day should be replaced with National Asphalt Parking Lot Day.

The Urban Planning Profession is complicit.

May 1, 2019 10:34 am

Daytime highs hardly changing and nighttime lows increasing agrees w/my own observations here in the rural central Appalachians. A reasonable person would think that would be a good thing….

mike macray
May 2, 2019 4:49 am

At the risk of stating the Obvious…
1. Temperature T is relevent only for dry air, for moist air T is “swamped” (sorry) by Enthalpy making T irrelevent as an indicator of energy content.
2. If warming is predominantly reflected in Tmin (night) then the temperature gradient driving weather should be weakening ergo reducing the probability of violent storms.
Good post with some enlightening comments!

Johann Wundersamer
May 3, 2019 2:00 am

“But I looked at the seasonal variations in these signatures and (unexpectedly) the day-night difference is greatest in winter (DJF) when there is the least convective activity and weakest in summer (JJA) when there is the most convective activity.

One possibility is that there is a problem with the AIRS temperature retrievals (now at Version 6).”

It is probably in winter

a greater difference between air surface temperature and temperature throughout the troposphere

than in the summer.

Johann Wundersamer
May 3, 2019 2:07 am

“But I looked at the seasonal variations in these signatures and (unexpectedly) the day-night difference is greatest in winter (DJF) when there is the least convective activity and weakest in summer (JJA) when there is the most convective activity.

One possibility is that there is a problem with the AIRS temperature retrievals (now at Version 6).”

It is probably in winter

a greater difference between air surface temperature and temperature throughout the troposphere

than in the summer.


%d bloggers like this:
Verified by MonsterInsights