Skip to content

Rain works on two clocks

6 minJ

Revised 10 August 2026

You know the disappointment. A dry August, then a proper Thursday downpour — twenty millimetres, the kind that runs off the track in rivulets. Saturday you walk the ground you have walked for thirty years and find nothing. Not thin pickings. Nothing.

The usual explanation is that it needed a few more days. Usually it did not. It needed a different month.

Moisture reaches a fruiting body through two separate processes with two separate lags, and they are not interchangeable. One decides whether a flush initiates at all. The other decides whether the fruiting bodies are up and findable right now. Nearly every conditions score on the market collapses both into a single figure, which makes it structurally unable to tell you the one thing worth knowing.

The trigger clock runs in weeks

The most direct evidence for a long window comes from a five-year commercial yield study in the Boreal Plain of Saskatchewan. Ivanochko et al. (2021) regressed chanterelle (Cantharellus cibarius) yield against hourly weather and reported that growing degree days at a base of 5 °C, combined with soil temperature and with either soil moisture or cumulative precipitation of 50–100 mm, gave the strongest relationship 6 to 13 weeks prior to first appearance. Not six days. Six to thirteen weeks.

The longest continuous record in Europe points the same way while splitting the job between two variables. Straatsma, Ayer and Egli (2001) counted fruit bodies weekly on a plot at La Chaneaz in western Switzerland across 21 years and found that productivity correlated with precipitation from June until October, whereas the time of fruit body appearance correlated with temperatures in July and August. Moisture set how much; temperature set when. Those are two questions, and the data answered them with different predictors.

Finland gives a third angle. Tahvanainen, Miina and Kurttila (2019) analysed 39 years of national purchase statistics and found chanterelle volumes correlated most strongly with July precipitation (r = 0.53), with August second (r = 0.38); their fitted supply model used combined June-and-July rainfall. Ceps keyed to August rainfall instead. That is market volume, not plot yield, so picker effort is baked in — but the offset between the two species is what a trigger window predicts.

The expansion clock runs in days

Now the short window. Rochon et al. (2011), working in two jack pine stands in eastern Canada, found positive correlations between sporocarp productivity and total rainfall one week prior to fructification, and air temperature two weeks prior. One week, two weeks — the same species, a completely different timescale from the Saskatchewan trigger.

Karavani et al. (2018) put both windows inside one model. Across 28 monitored plots in Mediterranean pine, mushroom yield depended primarily on weather and soil moisture during the same month, with one exception: the effect of precipitation showed a one-month delay. Same dataset, two lags, depending on which variable you look at.

The physiology explains why. A primordium is largely finished before you ever see it. Money (2002) describes how stipe, cap and gills are demarcated within the embryonic fruit body long before it expands, and that the expansion itself is driven by water uptake and cell wall loosening rather than by a burst of new cell division. Mushrooms appear overnight because inflation is fast. But you cannot inflate something that was never built. A dry August means no primordia; Thursday’s twenty millimetres has nothing to push on.

Why the rain gauge is the wrong instrument

Rainfall totals are a poor proxy for the water a fungus actually experiences. Lilleskov et al. (2009) used stable isotopes to trace sporocarp water during summer drought and concluded that in drier soils fruiting bodies likely derived 25–80% of their water from deep (>30 cm) or hydraulically lifted sources. Twenty millimetres onto baked ground wets the litter and evaporates. The same twenty onto soil that has been damp since mid-August recharges the profile.

This is the case for reading soil water rather than precipitation. ERA5-Land (Muñoz-Sabater et al., 2021) resolves soil moisture in four layers — 0–7 cm, 7–28 cm, 28–100 cm, 100–289 cm — on a 9 km grid. The 7–28 cm band integrates weeks of weather rather than yesterday’s shower, and it sits where ectomycorrhizal mycelium mostly lives. That is defensible. It is not validated: I could find no published study establishing which depth best predicts fruiting in Nordic forest soils, so anyone naming a depth is reasoning from first principles, as am I.

The temperature gate, and the fact that it moves

Warmth gates both clocks. The Saskatchewan work put the threshold near 500 ± 70 growing degree days above 5 °C — the same base SMHI uses for the Swedish vegetation period, and the same threshold Tahvanainen and colleagues used for their effective temperature sum. Andrew et al. (2018), analysing fruiting across central and northern Europe, found temperature the dominant driver for autumn-fruiting ectomycorrhizal fungi, with mean fruiting varying about 25 days with latitude and up to 30 days with altitude.

And the gate drifts. Kauserud et al. (2008) found Norwegian autumn fruiting delayed by an average of 12.9 days since 1980, and the follow-up (2012) traced the shift across Europe. A calendar rule inherited from your grandfather describes a gate that has since moved by nearly two weeks.

Why one number cannot answer this

Average a weeks-scale accumulation with a days-scale one and you get a figure that reads high in two opposite situations. Loaded soil plus a dry warm week: the flush is coming, but not yet — wait. Dry soil plus three days of drizzle: the litter is damp and nothing is on its way — do not bother. Those demand opposite decisions, and a single score cannot separate them, because the information was destroyed in the averaging.

Split them and the picture becomes actionable. “The ground has been loaded since mid-August; you are waiting on warmth” is a different sentence from “it rained hard on Thursday, but the profile has been dry since July, and this buys you nothing.” Both are honest. Neither is a number.

What is settled, and what is extrapolation

Settled: two windows exist and separate cleanly; moisture governs quantity while temperature governs timing (Straatsma et al., 2001; Andrew et al., 2018); expansion is water-driven inflation of a pre-formed primordium (Money, 2002); soil moisture works as a contemporaneous predictor while precipitation acts with a lag (Karavani et al., 2018).

Not settled: every specific number above. Six to thirteen weeks is Saskatchewan jack pine. One and two weeks is Québec. The 500 GDD figure is one regional regression, not a constant. The Spanish work is Mediterranean pine under a water regime nothing like Småland, and De la Varga et al. (2013) found Boletus edulis soil mycelium there peaking in February and bottoming in October — belowground biomass and fruiting are not in step, and transplanting a lag across forest types is a guess.

I could find no published, validated lag model for chanterelle fruiting in Swedish forest. The Finnish evidence is market volumes, not plot yields. So: the two-clock structure is real and well supported. The Swedish constants are not known, by anyone. Anybody quoting you a precise number of days for your parish is extrapolating, and the honest ones say so.

Sources