Hoppa till innehållet

Den här artikeln är på engelska.

Frost does not end the season by itself

9 minJames

Revised 11 augusti 2026

The forecast has kept the minimum just above freezing. At first light there is white on the moss in the glänta, the track is hard at its edge, and the open ground has plainly had a different night from the one described by the number on the phone. Under the spruce, a few metres away, the litter is dark and yielding. By breakfast the white has gone. Nothing about that morning answers whether the season has ended. It does show why the usual instrument is pointed at the wrong part of the forest.

The frost forecast measures air above the problem

Frost is a precise meteorological word with a loose life in the forest. SMHI defines it by temperature below zero, but the routine air measurement is made at roughly two metres above the ground. Markfrost is the useful companion term: freezing at the surface when the standard air measurement remains above zero. That difference is not pedantry. A fruit body standing in moss, and the mycelium connected through litter and soil, occupy the lower few centimetres, not the air around a station screen.

On a clear, calm night, the ground loses heat by radiation. The surface and the thin layer of air against it can become colder than the air measured higher up. Low cloud slows that loss; wind mixes the cold layer with warmer air above it. SMHI also notes a familiar piece of forest geography: open ground is generally more frost-prone than a stand, where the day’s warmth is retained better through the night. A glänta, a hollow, a road edge and the interior of barrskog can therefore receive different cold events under one regional forecast (SMHI, Frost och markfrost; SMHI, Temperaturen är ofta lägst i gryningen).

That makes an air-temperature forecast useful context, not a verdict on the ground. It tells something real about the incoming air mass and the broad risk of a cold night. It cannot report the temperature in the moss of a particular slope, still less the temperature of a particular fungus. Treating it as an on/off switch confuses a standardised observation with the microclimate that has to carry the season.

There is a second complication. The lowest temperature often comes around or just after sunrise, when the previous night’s radiative loss has not yet been matched by incoming solar energy. A glance at an evening minimum forecast therefore misses the part of the night that may matter most at the surface. The morning can look harmless at head height while the ground records the sharper event.

The ground keeps a slower and more varied clock

The soil is not merely colder air delayed by a few hours. Its temperature varies by depth, soil type, water content, vegetation and snow cover. SMHI’s account of jordtemperatur is plain on the basic physics: surface temperature often swings more than air temperature, while swings become smaller and their highs and lows arrive later with depth. In the older Swedish observation programme, measurements were made at depths from the shallow surface layer down through a metre, and in contrasting soils. That is a useful warning against speaking of “the soil temperature” as though every patch had one (SMHI, Jordtemperatur).

The distinction grows sharper late in autumn. A cold night can freeze a wet skin of ground or leave a rim of ice in exposed litter while soil below remains unfrozen. Tjäle is not the same event as frost on a cap. It means that water in the soil has frozen, and its depth depends on the duration of cold as well as soil, vegetation and snow. Snow can insulate the ground from cold air rather than simply announce that winter has arrived (SMHI, Tjäle).

Jungqvist and colleagues modelled soil temperature at four Swedish forest sites and found that the relation between air and soil temperature is non-linear, particularly through winter. Their central point bears repeating: assumptions about soil temperature based on air temperature alone are inadequate in boreal regions. Snow cover is one reason. A bare cold spell and an equally cold spell after snow has settled do not give the soil the same instruction (Jungqvist et al., 2014).

This is not an argument for pretending that every forager needs a probe in the ground. It is an argument for keeping the variables separate. SLU’s reference-climate programme measures both air and soil temperatures at forest research parks, alongside precipitation, humidity and radiation. The fact that a serious forest climate record keeps both series is a better guide than a single magic threshold. A phone forecast can describe the weather. It cannot turn air temperature into a measurement of the forest floor.

A damaged fruit body and a stopped system are different events

The phrase “the first frost ends the season” compresses at least three things into one. A fruit body exposed at the surface can be chilled or frozen. The formation of additional fruit bodies can slow or cease under a run of conditions. The established mycelium in soil and its association with roots or decomposing material can persist through winter. These are related, but they are not interchangeable observations.

The visible part is the easiest to mistake for the whole organism because it is the part that changes overnight. Fruit bodies are ephemeral. Kauserud and colleagues make that practical point when explaining why dated collections can be used for phenology: for many species the observed body lasts only a short time. A cold event may alter what is visible on a path without demonstrating that the underground fungus has been killed or that no later fruiting is possible (Kauserud et al., 2008).

Laboratory work on ectomycorrhizal mycelium has explicitly examined low temperature and freezing–thawing tolerance, but it is not a field test of a Swedish patch after one frosty night. It concerns particular isolates, controlled conditions and survival or growth of mycelium, not a universal date at which autumn production ends. That distinction is exactly why it should prevent a too-simple rule rather than furnish a new one (Tibbett et al., 2011).

So a night may stop the development of an exposed body without ending the conditions for every fungus in the stand. A sequence of cold, then thaw, then wet weather is another situation again. Conversely, a relatively mild week can still be poor ground for fruiting if surface and soil moisture have gone the wrong way. The word “ends” belongs to a season-long pattern; “stops” may describe a particular process at a particular moment. The field observation is real. The leap from it to a universal mechanism is where the rule fails.

The autumn edge is a sequence rather than a threshold

Temperature matters in fungal phenology, but it works with water, season, substrate, hosts and local exposure. The long Swiss record assembled by Straatsma, Ayer and Egli found the timing of fruit-body appearance related to July temperatures and productivity related to precipitation over the fruiting season. That is evidence for interacting seasonal conditions, not for one autumn minimum as a closing bell. Their plot was in Switzerland, however, and the climate, forest history and species set are not Söderåsen, Uppland or a northern Swedish spruce stand. It is useful evidence from the wrong forest for a precise Swedish rule.

The same restraint applies to boreal work elsewhere. Pinna and colleagues found soil temperature and moisture important for the initial fruiting date of edible mushrooms in a boreal mixed forest in eastern Canada. The study is closer in broad forest zone than a Mediterranean paper, but it is still eastern Canada, and its result concerns the start of fructification rather than the final frost of a Swedish autumn. It supports asking about soil conditions; it does not supply a number for a parish.

At continental scale, the pattern is even less like a single switch. Kauserud et al. analysed more than thirty-four thousand Norwegian herbarium records and found later autumn fruiting in recent decades, with species differing strongly in response. A later European analysis found shifts across countries and taxa. Andrew et al. likewise linked broad patterns of European fungal fruiting to climate variability, including temperature, but also separated nutritional groups and seasons. Krah et al.’s much larger global record analysis reaches the same broad conclusion: temperature affects timing and duration, with patterns varying among biomes. None of these datasets observes the exact night a particular Swedish stand becomes incapable of further fruiting.

That gap matters. A night’s air minimum, an autumn mean and the thermal state of moist litter are not substitutes for one another. Nor does a late date in a collection prove a warm soil at depth. Fruiting records are a record of what was seen, often with uneven collecting effort. Koskinen and colleagues, working with fruiting fungal communities, emphasise that environmental responses are structured by relatedness: communities do not move as one undifferentiated “mushroom season.”

A map can describe the ground without promising an outcome

The useful map question is therefore not whether a coloured cell promises a mushroom. It is whether the ground there has conditions and habitat that are more or less favourable, and how much of that reading comes from direct mapping rather than broad weather context.

Terrain earns a place in that question because cold air, drainage and exposure have shape. Sweden’s national Markhöjdmodell is built as a one-metre grid, a fine input for deriving terrain. A responsible rendered interpretation remains coarser: at tens of metres, not as a claim to show the forest floor at one-metre certainty. The difference protects against a familiar error, in which a precise-looking contour is mistaken for a precise prediction.

Likewise, SLU Markfuktighetskartor are mapped estimates, produced by combining topography, climate, soils and field data from Riksskogstaxeringen. They describe relative wetness and dryness in the landscape; they are not sensors placed in a reader’s moss. They can help distinguish a shallow, freely draining shoulder from a wetter hollow. They cannot settle what happened to soil water after last night’s wind, nor whether fruiting will occur there.

A conditions-favourable signal should therefore soften in the places where its inputs are weak. Weather is broad and changes through time. Terrain can sharpen the spatial account where it is mapped well, but it does not make a stale weather picture current. Unmapped ground is unmapped, not confidently empty. And a record of repeated finds on ground already walked is memory, with a real claim to fine detail; it is not a prediction for neighbouring ground that has never been walked.

The sharp boundary remains unmeasured

Settled: air frost and markfrost are different measurements; the surface can be colder than the standard air observation; soil changes temperature more slowly and differently with depth; snow, vegetation and ground conditions matter. Settled too is that fungal fruiting responds to climate and that species and groups respond differently. The folk sentence about the first frost has caught a real change in the season, even if it has named the wrong mechanism too neatly.

Not settled: a universal minimum temperature, number of cold nights or soil depth that marks an end to fruiting in Swedish forests. I could find no field study that establishes such a threshold across Swedish forest fungi. The best long records are Norwegian and European phenology collections, a Swiss plot, a Canadian boreal forest, and broad-scale analyses. They are valuable precisely because they show variation. They do not turn a forecast minimum into an answer for one hillside.

What would improve the question is repeated paired observation: ground and air temperature, moisture, snow or litter state, species-level fruiting records, and the same walked ground over several autumns. Until that exists at the relevant scale, the honest end-of-season read is conditional. A cold, clear night changes the ground. A run of cold and the arrival of persistent soil frost change it more. Neither lets anyone say, with certainty, what will be found or when.

Sources