Methodological Guidelines for Circadian Modeling of Daylight Saving Time

“Methodological Guidelines 
for Circadian Modeling 
of Daylight Saving Time:
Application to the United States”

Most mornings I get an email from arXiv, an archive of preprints in the sciences, especially physics. (Here at the start of this post I should warn you: Preprints on arXiv are not peer reviewed. They are preliminary, and are meant to get results out quickly, avoiding the delays associated with publishing in a scientific journal. But peer review is not just slow, it is also valuable. Almost all preprints contain some mistakes and some are garbage. So, reader beware!)

In my daily email from arXiv, I get a list of new preprint titles and abstracts submitted under the category of biological physics. Most days, I scan the titles and don’t find anything that interests me. But last week (Thursday, June 18), I saw this: 

arXiv:2606.19541 (*cross-listing*) 

Date: Wed, 17 Jun 2026 19:33:48 GMT   (587kb)

Title: Methodological guidelines for circadian modeling of Daylight Saving Time: application to the United States 

Authors: Jose Maria Martin-Olalla, Jorge Mira 

Categories: physics.soc-ph physics.bio-ph physics.data-an physics.pop-ph 

Comments: 2037 words, 7 pages, 4 figures 

\\   

Modeling the circadian impact of seasonal clock changing requires precise synchronization between solar and social time. This report critiques a recent study that associated disease prevalence in the United States with seasonal clock exposure. We identify a fundamental computational error in which a sign reversal of the longitudinal offset effectively inverted the US East-West axis, cross-correlating local health data with the circadian burden of hypothetical locations on the opposite side of a time zone. We outline the methodology for a correct modelization [note: I hate the word “modelization”] of the circadian process in the context of US geography. 

\\ 

(https://www.google.com/url?q=https://arxiv.org/abs/2606.19541&source=gmail-imap&ust=1782438904000000&usg=AOvVaw2-3WiUhnifCPipklvdW2B2, 587kb)

This preprint, which I downloaded and read as a pdf, illustrates science in action. It critiques a previously published paper and claims to show there is a fundamental flaw in its analysis. (Another disclaimer: I am not an expert in circadian rhythms, and I cannot say from my own research who is right: the original authors Lara Weed and Jamie Zeitzer from Stanford University, or the Spanish physicists Jose Martin-Olalla and Jorge Mirab.) 

Intermediate Physics for Medicine and Biology features similar scientific debates. For instance, in Chapter 16 on the Medical Uses of X-Rays Russ Hobbie and I outline a debate between Bernard Cohen and Jay Lubin about the danger of the radioactive gas radon for causing lung cancer, and in Chapter 9 on Electricity and Magnetism at the Cellular Level we analyze another exchange between Joseph Kirschvink and Robert Adair about the biological effects of weak magnetic fields. We include these tales to illustrate how science is done. Students using IPMB need to realize that science is more than a collection of facts. It’s a process that can sometimes be messy but is vital for the advance of science. 

Weed and Zeitzer’s original paper looked at the health hazards caused by daylight saving time. It’s fairly well known that changing the clocks twice a year at the beginning and end of daylight saving time can be associated with health problems. But how significant are those problems? And, more specifically, if we eliminate the time change, should we replace it with permanent standard time or permanent daylight saving time? Weed and Zeitzer used mathematical modeling to analyze the degree of “circadian burden” on health. Circadian burden is the physiological strain on your body’s biological clock when your daily schedule in not in sync with the sun’s natural rhythms. Weed and Zeitzer concluded that

As compared to current time policy [switching back and forth between standard time and daylight saving time], permanent SDT [standard time] and [permanent] DST [daylight saving time] reduce circadian burden and are anticipated to reduce the prevalence of obesity and stroke with SDT having a more positive impact than DST.

The preprint by Martin-Olalla and Mirab claims to have found a bug in Weed and Zeitzer’s computer program.

We identify a fundamental computational error in which a sign reversal of the longitudinal offset effectively inverted the US East-West axis, cross-correlating local health data with the circadian burden of hypothetical locations on the opposite side of a time zone.

They conclude that

Consequently, the original study’s conclusions currently lack empirical support.

Who’s right? I can’t say for sure, and I would like to read Weed and Zeitzer’s response to this preprint before I draw any definite conclusion. But the physicists make a convincing case for a fundamental flaw. They quote Weed and Zeitzer’s computer code line by line, highlighting where the bug occurs. It seems to be a sign error in how the burden shifts with position in the time zone, inverting east and west. 

Martin-Olalla and Mirab include a “manuscript timeline” in their preprint that may not survive peer review and journal policy, but is interesting nonetheless.

The curiosity bug bit us. We decided to pull down and run the original code ourselves to compute and observe the yearly burden… all of a sudden, we uncovered the inversion error. It... was sitting right there in plain sight, but we had initially trusted the comments embedded within the script rather than checking the raw mechanics. We checked that the light diets were inverted: Western locations had brighter mornings and darker evenings in their clock time analysis.

Before I go on, I should confess to you some of my biases. First, I’m a physicist, so I naturally suspect that the physicists will be more skilled at mathematical and computer modeling. Second, I live in Michigan, just north of Detroit, which is pretty far north and is near the western edge of the Eastern Time Zone. In the winter, it doesn’t get light until about 8 am. If we were on permanent daylight time, it wouldn’t get light until 9. I’m a morning person, and this would annoy me. My circadian clock would suffer from a whole lotta burden. So I’m passionately against the idea of permanent daylight time, and would rather keep switching twice a year if we can’t get permanent standard time.

And now back to the story. What lessons can be learned from this scientific debate?

1. Science is difficult. Sometimes you’re wrong. Wrong doesn’t mean evil, and wrong doesn’t mean stupid. Although I don’t know who is wrong in this case, I’m pretty sure one pair of authors is right and one is wrong (I don’t see how both pairs could each be a little right and a little wrong). That’s why science has peer review. In this case, peer review by the journal doesn’t seem to have spotted the problem. But peer review also occurs when other scientists get bit by the curiosity bug. It may take a long time, but the truth is usually uncovered by the peer review process. It can be painful on a personal level, but is vital for science as a whole. 
2. Mathematical modeling is difficult. I should know, because I did mathematical modeling and computer simulation for a living for nearly 40 years. It’s easy for a bug to creep into a computer program. That’s why you have to test, test, test each part of the code. Toy models with analytical solutions of special cases can help verify your program. They can’t prove the code is correct, but they can uncover flaws. In general, I always tried to analyze special cases and examine intermediate results “by hand” to make sure they make sense. It’s slow, painstaking work, but minimizes the chances of being wrong. 
3. Sharing computer code is good. I rarely shared my code early in my career; it just wasn’t a thing people did back then (and was difficult to do before the internet). But the trend now is to share code openly. Some journals insist on it. This can be embarrassing if your program is wrong, and even if it’s right but written in an ugly, complicated way. But sharing is essential. Share you code, and write your code in a way so you aren’t embarrassed to share it.
4. Sign up for arXiv. It’s one of the best ways I know of to keep up with the current literature. Nowadays the scientific literature sprays out like from a fire hose, so we need all the help we can get.
5. This is a story about science, not antiscience. Over the last couple years, the forces of antiscience are everywhere (consider vaccine hesitancy and climate change denial for starters). Antiscience advocates demonize scientists and ignore evidence. The authors I’m talking about today—Weed, Martin-Olalla, Mirab, and Zeitzer—have nothing to do with antiscience. They are merely engaged in a scientific debate. They are showing us how science should be done. Compared to advocates of antiscience, both pairs of authors in this debate are heroes. I love them all, no matter who turns out to be right in the end. 

If the Week and Zeitzer article is flawed, then is permanent standard time really better than permanent daylight time? Frustratingly, Martin-Olalla and Mirab don’t say. They don’t reproduce Week and Zeitzer’s analysis using a corrected program. So, after all this debate, we still don’t know what is best. I’ll hope for permanent standard time, but policy should depend on scientifically-established, country-wide analysis and not on Brad Roth’s personal preferences. I’m not convinced we have the answer to this issue right now. Stay tuned.