Does leaving the car window down an inch actually cool it off?

[MsWhatsit]

I see people doing this a lot in the summertime around here. On a hot day, they'll crack their car windows by 1-2", I guess on the assumption that it will allow air to circulate so that their car will only be as hot as the surface of the Sahara Desert when they get back, instead of being as hot as the surface of the sun.

Does this actually work?

[scr4]

If you mean "is the logic valid," yes. Usually a surface exposed to direct sunlight will get hotter than the surrounding air. If that surface is the roof of a sealed car, the interior gets hotter than the outside air as well. So opening the window will lower the interior temperature.

If you mean "does a 1-inch gap really make a noticeable difference" - that depends on wind speed, outside temperature, color of the car, etc.

[Little Nemo]

My unproven assumption is that the air is circulating inside the car. As the sun heats the car, the hot air inside is rising to the top of the car. So with a gap in the car windows, some of the circulating air inside is going to move out of the car. And being as the gap is near the top of the car, this air is the warmest in the car and the air from outside that replaces it will be cooler.

[scr4]

I suspect outside wind is more important than any natural convection current inside the car, even on a seemingly calm day. Especially if you crack open windows on both sides of the car, so wind can push air into the car from one side and out the other.

[x-ray vision]

According to researchers at Stanford University School of Medicine:

Their study, published in the July 2005 Pediatrics, documented that on sunny days, even if the temperature is a relatively cool 72 degrees, the inside temperature of the car will reach 117 degrees within 60 minutes of turning the car off (and will top 108 degrees within 30 minutes!). Leaving the windows open 1.5 inches made no difference in the final temperature or the rate of rise.

http://www.drgreene.com/21_1957.html

This part is a little ambiguous:

On hot days, the rate of rise is the same, but dangerous temperatures can be reached within 5 minutes of closing the car doors, even if the windows are cracked.

The rate of rise is the same, but it's unclear whether or not the same final temperature will be reached on hot days.

[engineer_comp_geek]

My own personal experience strongly disagrees with your Stanfrod study. When I am at work, leaving the windows cracked just 1 inch makes a huge difference in the temperature of my car at the end of the day.

[x-ray vision]

Originally Posted by engineer_comp_geek

My own personal experience strongly disagrees with your Stanfrod study. When I am at work, leaving the windows cracked just 1 inch makes a huge difference in the temperature of my car at the end of the day.

What kind of day? A "relatively cool 72 degrees" day or a "hot" day?

[Quartz]

I think the problem here is that the air coming in to the cabin has to go via the car's ventilation system, so (a) there's a small volume of air coming in and (b) the air gets heated along the way before it enters the cabin.

I wonder what the effect of leaving the windows wide open would be - other than your car getting stolen, of course.

[Sapo]

Originally Posted by Quartz

I wonder what the effect of leaving the windows wide open would be - other than your car getting stolen, of course.

Then I guess it depends on whether the thief turns on the A/C or not.

[Zsofia]

I don't crack my windows when it's in the 70's. I do when it's over 100. It definitely makes a difference.

[zut]

Originally Posted by x-ray vision

According to researchers at Stanford University School of Medicine:

http://www.drgreene.com/21_1957.html

This part is a little ambiguous:

The rate of rise is the same, but it's unclear whether or not the same final temperature will be reached on hot days.

This is kind of an interesting drift of information. Your cite quotes the following paper:

"Heat Stress From Enclosed Vehicles: Moderate Ambient Temperatures Cause Significant Temperature Rise in Enclosed Vehicles," Catherine McLaren, MD, Jan Null, CCM and James Quinn, MD, PEDIATRICS Vol. 116 No. 1 July 2005, pp. e109-e112. (This might be restricted content, so I don't know if the paper is generally available.)

The paper says this:

On 2 of these days, recordings were made first with the windows closed. Then the doors were opened to return the vehicle to ambient temperature, and a second hour of measurements was made with the windows cracked 1.5 inches. It was ensured that no significant change in ambient temperature occurred during these 2 hours.

[...]

Comparison of Closed Versus Cracked Open Windows
Figure 4 illustrates a nonsignificant trend to faster heating in the first 20 minutes with the windows closed 6.25°F per 5 minutes (95% confidence interval [CI]: 1.0–11.5) versus 5.5°F per 5 minutes (95% CI: 1.2–9.8) with the windows cracked open. However, overall, the rate of temperature rise for windows that were cracked open was 3.4°F per 5 minutes (95% CI: 2.4–4.4°F) compared with the rate for closed windows of 3.1°F per 5 minutes (95% CI: 1.4–4.8°F), with the final temperature for both circumstances being identical. (my bold)

Note that this confirms that there is no difference in the final temperature, but the rate of rise is different.

Now, this paper, in turn, cites a previous paper:

King et al found that in an ambient temperature of 36.8°C (98.2°F), 75% of the maximum temperature rise occurred within 5 minutes of closing the doors and maximized within 15 minutes to 51 to 67°C (124–153°F). Opening the windows 20 cm (8 inches) had minimal effect on the temperature rise and maximum temperature attained. (my bold)

So let's look at that paper:
"Heat Stress in Motor Vehicles: A Problem in Infancy," K. King, K. Negus and J. C. Vance, Pediatrics1981;68;579-582. (a pdf; again I don't know if it's generally available.)

The effects of ventilation were assessed using three large white sedans with brown interior trim. One car had closed windows; the second and third had all four windows opened 50 mm and 200 mm (halfway), respectively.

[...]

The vehicle with windows open 200 mm was significantly cooler than the vehicle with windows open only 50 mm. Likewise, the latter was cooler than the closed vehicle. The stabilized interior ternperatures for the three vehicles were as follows: closed vehicle, 66.4 C; windows open 50 mm, 50.5 C; windows open 200 mm, 40.9 C (ambient temperatare 36.8 C).

[...]

Substantial reductions in interior temperature were only achieved by opening the vehicle windows by at least 200 mm (halfway) for the test sedans. This action gives internal temperatures approaching ambient, approximately 40 C compared with 36.8 C. This degree of window opening increases the risk of theft,2 and restraining the child within the car may be more difficult. A “compromise” opening of 50 mm, although resulting in a temperature drop of more than 15 C over the closed vehicle temperatures, still produced a temperature 13 C above ambient. Thus, “cracking” the windows contributes little to the creation of an acceptable interior environment. (my bold)

Note that the actual data from the King paper directly contradicts what the later McLaren paper says--not just contradictory data between the two experiments, but McLaren actually mischaracterized the data obtained by King.

In addition, the McLaren paper shows temperature rise for "cracked" windows that has an unexpected, unusual shape over time. I suspect their experimental procedure ("[R]ecordings were made first with the windows closed. Then the doors were opened to return the vehicle to ambient temperature, and a second hour of measurements was made with the windows cracked 1.5 inches.") is flawed, although admittedly that's hard to say with certainty. However, the mischaracterization of the earlier results does not give me confidence in the McLaren paper, and thus the Dr Greene site--although admirably cautionary--does not strike me as conclusive evidence.

[Squink]

On 2 of these days, recordings were made first with the windows closed. Then the doors were opened to return the vehicle to ambient temperature, and a second hour of measurements was made with the windows cracked 1.5 inches. It was ensured that no significant change in ambient temperature occurred during these 2 hours.

I'm curious as to whether they also controlled for the 30° movement of the sun over the course of these two hours, and if so, how.

[Cheesesteak]

Originally Posted by zut

I suspect their experimental procedure ("[R]ecordings were made first with the windows closed. Then the doors were opened to return the vehicle to ambient temperature, and a second hour of measurements was made with the windows cracked 1.5 inches.") is flawed, although admittedly that's hard to say with certainty.

It all depends on how long the doors were left open when returning the vehicle to ambient. If they opened the doors, flapped them a few times and waited a couple of minutes, there's no way this procedure is sound. The seats, floor, roof and dash would all still be significantly hotter than ambient, and will foul up the results. Use 2 identical cars, and alternate cracked vs. closed instead of always doing cracked second.

[zut]

Originally Posted by Cheesesteak

It all depends on how long the doors were left open when returning the vehicle to ambient. If they opened the doors, flapped them a few times and waited a couple of minutes, there's no way this procedure is sound. The seats, floor, roof and dash would all still be significantly hotter than ambient, and will foul up the results. Use 2 identical cars, and alternate cracked vs. closed instead of always doing cracked second.

Exactly my point. The irregular shape of the data curve and the non-intuitive result (as well as the mischaracterization of earlier data) makes me think something is awry. If nothing else, consistently running the "cracked" trial after the "closed" trial, rather than alternating or randomizing the trials, is poor experimental design and the results should be called into question just because of this.

[cjepson]

Originally Posted by Quartz

I wonder what the effect of leaving the windows wide open would be - other than your car getting stolen, of course.

That definitely makes a huge difference. If I'm at home and know I'm going to be using the car in, say, half an hour, sometimes in summer I'll go out and open all the windows. Over the course of that half hour the temperature in the car will drop from oven-like to pretty much whatever the outside temperature is.

[x-ray vision]

Nice work, zut.

[Turek]

And after the nice work?

[x-ray vision]

Pardon?

[silenus]

Bad, wicked zut.

X, it's a Monty Python reference.

[Sapo]

Originally Posted by zut

Exactly my point. The irregular shape of the data curve and the non-intuitive result (as well as the mischaracterization of earlier data) makes me think something is awry. If nothing else, consistently running the "cracked" trial after the "closed" trial, rather than alternating or randomizing the trials, is poor experimental design and the results should be called into question just because of this.

Good catch.

Do we have any reason to suspect an intention to eschew the results for a particular purpose? Is any of these studies funded by a company selling cooling fans, a group against animal cruelty, child abuse, etc? Or is it just sloppy science?

[Boyo Jim]

FYI, there is a product designed for colling a car heating in the sun.

Solar powered fan set in cracked open car window.

It get really bad reviews.

[MsWhatsit]

Wow, this thread became a lot more informative than I had anticipated. OK, so maybe all those people who crack their windows are really on to something.

[Squink]

Originally Posted by Sapo

Good catch.

Do we have any reason to suspect an intention to eschew the results for a particular purpose?

Getting a paper published with your name on it is sufficient purpose for some.

[Sapo]

Very true, but what makes you think that one result is more publishable than the other?

[lazybratsche]

To be fair, the overall conclusion for both papers is still pretty sound: don't leave babies in cars on sunny day, even if it's not that hot out. The authors (both sets) look like they just wanted a quick bit of evidence, though they seemed to get sloppy in the details, they always saw dangerous increases in temperatures.

As to how this got published? Both were "letter" sized publications -- typically short papers sharing a single experiment. These are published to get a piece of data out fast, or to put out something that you're not interested in following up with a longer paper. The peer review process is usually faster and not quite as rigorous for this. Adding to that, both papers were published in Pediatrics, where the reviewers will be other pediatricians who will be focused on the medical accuracy and validity of the papers, rather than the details of physics underlying each experiment.

[Early Out]

Originally Posted by lazybratsche

To be fair, the overall conclusion for both papers is still pretty sound: don't leave babies in cars on sunny day, even if it's not that hot out.

And if you must leave them in the car, pack them in ice so they don't spoil.

What?

[zut]

Originally Posted by Sapo

Do we have any reason to suspect an intention to eschew the results for a particular purpose? Is any of these studies funded by a company selling cooling fans, a group against animal cruelty, child abuse, etc? Or is it just sloppy science?

Never attribute to malice, etc. etc.

1. The main thrust of the McLaren paper is to determine whether the interiors of closed vehicle reached temperatures that would be dangerous to children, even on days with relatively mild ambient temperatures. The characterization of the diffrences between open and closed windows was (I suspect) just a side project.

2. The authors have a background in medicine, not heat transfer. Although they should certainly have known enough to run randomized trials, they probably didn't know enough to question their results.

3. As far as I can tell, the results from the McLaren paper for closed vehicles fall in line with the results from earlier papers. Temperatures in vehicles with cracked windows, although lower than the temperatures in vehicles with closed windows, were still dangerous--and so the results were "consistent" among the papers in that respect.

4. Finally, the fact is that this: "Opening the windows 20 cm (8 inches) had minimal effect on the temperature rise and maximum temperature attained." is a complete mischaracterization. It seems unbelievable that that would be done intentionally on something that's so easy to check.

[Really Not All That Bright]

zut, I love you. If anybody needs me, I'll be in the parking lot with a meat thermometer.

ETA: I have an AutoCool somewhere which I bought solely for the purpose of testing it- I was pretty sure it would be useless- and it was. It does, however, make a minimal but significant difference to the smell of an older car (subjectively, of course).

[engineer_comp_geek]

Originally Posted by x-ray vision

What kind of day? A "relatively cool 72 degrees" day or a "hot" day?

A "hot" Baltimore (technically, Owings Mills, MD) day, which is going to be in the 80 to 90 deg range. I don't bother cracking the windows on cooler days. The car sits for about 9 hours in the parking lot, with no shade.

[zut]

Originally Posted by Really Not All That Bright

zut, I love you. If anybody needs me, I'll be in the parking lot with a meat thermometer.

For reference, I believe temperatures obtained rectally are more reliable than those obtained orally. Please make sure you have a comfortably shaped experimental device.

[Really Not All That Bright]

Originally Posted by zut

For reference, I believe temperatures obtained rectally are more reliable than those obtained orally. Please make sure you have a comfortably shaped experimental device.

"Sarah? How comfortable is the shape of your rectum?"

[Johnny L.A.]

Originally Posted by zut

For reference, I believe temperatures obtained rectally are more reliable than those obtained orally. Please make sure you have a comfortably shaped experimental device.

But air has less mass than rectums. Especially if someone has a massive rectum. So temperature change would be difficult to measure.

[BMalion]

Originally Posted by Boyo Jim

FYI, there is a product designed for colling a car heating in the sun.

Solar powered fan set in cracked open car window.

It get really bad reviews.

Yeah but it says "as seen on TV", that's got to count for something.







Rectum? Damn near killed 'im.

[Sapo]

Originally Posted by BMalion

Yeah but it says "as seen on TV", that's got to count for something.

Yeah, it counts for TV cars. If you use it, no cars on TV will get hot.

[Richard Pearse]

Unfortunately, at least some of the people who are leaving their windows cracked open do it so the windows don't explode from the pressure of the expanding hot air inside .

[Really Not All That Bright]

Originally Posted by 1920s Style "Death Ray"

Unfortunately, at least some of the people who are leaving their windows cracked open do it so the windows don't explode from the pressure of the expanding hot air inside .

Well, that's just silly. I, of course, do it because hot air rises, and I don't want my car floating off while I'm doing my shopping.

[Nava]

Originally Posted by zut

4. Finally, the fact is that this: "Opening the windows 20 cm (8 inches) had minimal effect on the temperature rise and maximum temperature attained." is a complete mischaracterization. It seems unbelievable that that would be done intentionally on something that's so easy to check.

One of the most-quoted papers in Protein Chemistry is misquoted (people quote a paper from the right author but which talks about sugars and doesn't even mention proteins; what should be quoted is a presentation, which is less prestigious). I got a lowered grade on a presentation I did on the subject for NOT quoting the wrong paper... even after I explained the misquotation. People are very, very bad at verifying quotes and for some Tradition beats Accuracy every time.

I don't know what's required of MDs in the US, but most people I know who studied Statistics in college (in several countries including the US) never saw anything about Design of Experiments, or saw everything in a completely theoretical fashion, or never learned any continuous probability distribution other than Gauss' "Normal" Distribution (which they then proceed to apply to any data distribution, statistical or not, because, you see, it's "normal") - that's when they went beyond combinations. It's pretty scary given how many of them work in research.