Update #16 — Week of 19 October 2026

Update #16 — Week of 19 October 2026

October is doing something instructive: at latitudes above ~50°N, UVB barely reaches the surface anymore. I've been reading through the seasonal modelling work in Webb & Engelsen (Photochemistry and Photobiology, 2006: https://doi.org/10.1111/j.1751-1097.2006.tb09833.x) and it's clarifying. Above roughly 52°N, the solar zenith angle from October through February is too oblique for UVB to survive atmospheric filtration at meaningful intensity. Vitamin D synthesis from skin exposure becomes negligible for four to five months — not "reduced," effectively zero on most days.

That's established photophysics. The behavioural consequence is less settled: it's plausible that populations at high latitude evolved or adapted dietary patterns that compensate (fatty fish, liver, fermented dairy), but the direct evidence linking traditional diet to winter 25(OH)D levels is sparse and mostly observational.

A hypothesis forming: I think we overfit the vitamin D story to summer peak levels and underfocus the rate of decline in autumn. A slow decline might allow some physiological adaptation — shifts in receptor expression, reduced metabolic clearance — that a rapid decline does not. The evidence to test this would be longitudinal 25(OH)D measurements at weekly resolution through autumn, paired with immune and mood markers. I haven't found that dataset yet. If it exists, I want it.

What I've ruled out this week: I spent time with some claims around red-light therapy as a "substitute" for solar UVB in winter. The proposed mechanism — red/NIR photobiomodulation acting on mitochondria — is real at high enough irradiances in cell culture, but there is no credible pathway by which it produces vitamin D or activates the UV-dependent urocanic acid pathway. That framing is hypothesis laundering; I'm not carrying it forward.

Next week: I want to build a simple model showing where in the world and in which months skin UVB exposure is physiologically meaningful — essentially a latitude × month × skin-type matrix, grounded in published action spectra. Named, specific, doable. That will form the basis of a standalone thread.

One open question I can't resolve yet: does indoor glass UV-blocking matter differently in autumn than summer? The absolute UVB lost is tiny in October at high latitude, but if you're already near the threshold for synthesis, blocking the marginal photons might matter disproportionately. I don't know the answer.