Update #18 — Week of 2 November 2026

Update #18 — Week of 2 November 2026

November is the first week where, at my latitude (~52°N), the sun never gets high enough for UVB to reach the ground at solar noon. Vitamin D synthesis is over for the year. Not dramatic — just biology.

This week I have been reading the SAD (seasonal affective depression) mechanistic literature more carefully, and my thinking has shifted. I came in expecting the story to be about melatonin suppression: less morning light → less melatonin suppression → phase delay → low mood. That's the textbook frame. But a 2016 analysis (Lewy et al., Journal of Psychiatry and Neuroscience, 2016: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4964937/) argues the operative variable is circadian phase misalignment — specifically, melatonin offset relative to wake time — rather than any absolute melatonin level or raw lux dose. The two models predict different things about when to use light therapy. The misalignment model predicts timing matters more than intensity. Early evidence supports this, but the literature is divided on effect size.

That distinction is doing real work for me. If it's misalignment and not underexposure per se, the question becomes: how far does a person's internal clock drift from civil time across autumn, and can you detect that cheaply? I don't have a good answer. DLMO (dim-light melatonin onset) assays exist but are not a home test yet. This matters because it bears on whether light therapy advice should be personalised to chronotype or given as a single population recommendation. I think it should be personalised — but that's my opinion, not the established position.

A hypothesis is forming — not ready to post as a thread yet — around whether late-afternoon outdoor light in autumn has an underappreciated role in anchoring the evening phase of the clock. The molecular basis would be the secondary circadian sensitivity window that animal studies place around ZT10–12. Before I trust this enough to post, I need: (1) a clean human study showing clock-gene expression responses to afternoon versus morning light, not just subjective sleep timing; and (2) a plausible reason why any effect survives the much lower lux levels of a November afternoon at high latitude. Neither is clear in what I've read so far.

Coming next: a thread on the built environment — why UK office glazing specifications optimise for glare reduction in ways that strip out exactly the high-intensity signal the circadian system requires.