Den här artikeln är på engelska.

Trattkantarell follows the wet line, not the calendar

11 augusti 20268 minJames

Revised 11 augusti 2026

A spruce slope can be dry enough to crack under the heel at its crown and still hold a dark, mossy strip below. By October the two pieces of ground are no longer visually alike. The upper ground has shed water and litter; the lower line has collected both. That difference is worth noticing. It is not, however, a promise about what will be fruiting there.

Trattkantarell (Craterellus tubaeformis) is often given the broadest of habitat descriptions: mossy barrskog, autumn, acid ground. SLU Artdatabanken is more useful than that shorthand. Its account puts the species in barrskog and bokskog, seldom other deciduous forest, among moss, with spruce, pine and beech, preferably on acid and nutrient-poor ground. Naturhistoriska riksmuseet similarly calls barrskog its principal setting. Those are real constraints. At the scale of one hillside, though, each is a bundle of separate conditions: tree partners, humus, drainage, shade, accumulated litter and the history of the stand.

The late season is equally real, but is too often converted into a calendar rule. A mild November, an early frost, a wet week and a thin stand do not mean the same thing to the fungus. The fruit body is only the visible and short-lived part of a much older association in the soil. There is evidence for habitat and for broad phenology. There is much less for the particular switch that makes a Swedish patch produce fruit bodies in one autumn and remain quiet in another.

The mossy line below the ridge is a habitat distinction

On a hillside, water does not arrive evenly. It runs off convex ground, is slowed by small shelves and hollows, and is held or lost according to soil, organic layer and vegetation. Moss is therefore not merely scenery around trattkantarell. It is a useful visible sign that a particular piece of forest floor has been cooler, shadier or more continuously moist than the ground a few metres away. It is not a measurement of today’s water content, and it is not a substitute for the tree community.

This is where the common phrase “damp conifer forest” becomes too blunt. Permanently wet ground, fresh-fuktig ground and ground that merely remains damp after rain are different ståndorter. SLU’s Markfuktighetskarta makes the same distinction in mapped form. It is a model of long-term terrain, climate and soil controls, trained against Riksskogstaxeringen field data; SLU explicitly says that actual wetness at a particular time also depends on the weather. A blue or green patch is therefore evidence about the ground’s tendency, not a sensor reading from beneath the moss.

Terrain can help separate these strips of ground. Lantmäteriet’s national Markhöjdmodell is built on a one-metre grid, while SLU’s soil-moisture map uses terrain alongside other inputs. But one-metre input is not one-metre knowledge of forest-floor conditions, still less of fruiting. Roots, humus depth, an old drainage ditch under vegetation and a small break in canopy all escape a simple slope calculation. The ecologically honest grain for such a reading is tens of metres: a damp footslope, not a claim about an unwalked square metre.

Habitat suitability names ground where the association’s conditions appear favourable. It never becomes a probability that a person will find fruit bodies there. Where terrain or moisture data are missing, the honest description is unmapped, not empty.

Host, humus and water meet in the same place

The word “with” in an ecological catalogue matters. Artfakta’s spruce, pine and beech are not decorative neighbours. C. tubaeformis is a mycorrhizal forest fungus, dependent on a relationship with tree roots. A hillside can have the right shade and moss yet lack the host context or the established fungal network that makes the rest of the description meaningful.

The direct species study most often cited here is also a warning against easy Swedish conclusions. Trappe (2004) surveyed C. tubaeformis in north-western Oregon and confirmed its association with western hemlock, with further associations to Douglas-fir and Sitka spruce. In that study, stand age and well-decayed coarse woody debris were related to occurrence; slope, elevation and aspect were not. A striking share of the recorded fruit bodies was on very decayed woody debris despite that debris covering little of the ground.

That is valuable evidence that the species can be tied to the slow architecture of an old forest floor, not just last week’s rain. It is not evidence that a Swedish granbacke behaves like an Oregon hemlock forest. The host tree, climate, soil history and even the genetic limits of the taxon differ across that distance; Trappe also noted sequence differences among western North American, eastern North American and European populations. The study is useful precisely because it blocks two bad shortcuts: treating aspect as a universal rule, and treating every moist moss carpet as equivalent habitat.

For Swedish ground, the firmer statement remains modest. Acid, nutrient-poor mossy forest floor with the listed tree associates is the documented setting. The exact contribution of decayed wood, stand age, canopy density and humus depth to Swedish fruiting has not been isolated in a comparable field study. SGU’s jordartsdata, Skogsstyrelsen’s forest maps and a walked patch can describe parts of that setting, but none independently demonstrates the below-ground association.

A late season is not a date

The practical name höstkantarell has a sound observation behind it: the species’ fruit bodies are commonly recorded after the main summer flush of many familiar forest fungi. But “late” is relational. It does not mean that a particular week in September, October or November is biologically interchangeable from Skåne to Västerbotten, or from one year to the next.

Long records show why a calendar cannot carry that weight. In a Swiss forest plot surveyed over 21 years, Straatsma, Ayer and Egli (2001) found seasonal fungal production related to rainfall and the timing of appearance related to summer temperature. That result concerns a whole fungal community in Swiss beech and pine woodland, not trattkantarell on Swedish podzol. It supports the idea of several weather windows rather than one rain trigger. It cannot supply a Swedish threshold or a number of waiting days.

Kauserud and colleagues (2008), using Norwegian herbarium records, reported a shift of the autumn mushroom-fruiting season towards later autumn. Krah and colleagues (2023) likewise found that temperature shapes timing and duration of fungal fruiting across major terrestrial biomes. Both studies make late fruiting biologically plausible in a changing climate. Neither tells us which combination of moisture, host carbon, soil temperature and developmental state initiates a trattkantarell fruit body on a given slope.

Capinha and colleagues (2024) came closer to the species by modelling the timing of winter chanterelle fruit-body observations in Denmark from citizen-science records. Their method explicitly contrasts the weather conditions associated with dated observations against conditions available at the same locations and accounts for spatial and temporal observation bias. This is good work on phenological timing. It remains a model of reported emergence, not a controlled test of the mechanism in the mycelium, and Denmark is not Sweden. It should not be turned into a Swedish calendar or an outcome forecast.

Frost marks a change but does not explain the change

“Trattkantarell handles frost” is an observation with more versions than tellers. It may mean that an already visible fruit body survives a light frost, that it remains recognisable after it has frozen, or that new fruit bodies continue to appear during a run of cold nights. Those are three different claims. They concern persistence, appearance and growth, and they need different evidence.

SMHI defines a frostdygn by an air-temperature minimum of minus 0.1°C or below. That is a precise meteorological definition, not a temperature measurement in the humus. SMHI also notes that soil-surface temperature varies with the underlying surface and that temperature variation diminishes with depth. Under moss and litter, beside a trunk or at the foot of a slope, the relevant temperatures can part company with the nearest station’s overnight minimum.

So a frost reading can mark a meaningful shift in the forest, especially when repeated cold is paired with drying or frozen ground. It cannot, on its own, establish a physiological tolerance limit for this species. I could find no Swedish experiment that exposes C. tubaeformis mycelium and fruit bodies to defined cold treatments while controlling moisture and host condition. Nor could I find a field study that separates a single light frost from a sustained cold period and measures subsequent Swedish fruiting.

The careful conclusion is narrower than the folk rule but not dismissive of it. The species has a conspicuously late season and is regularly associated with conditions in which frost is possible. Whether cold is a trigger, a tolerated consequence of late fruiting, or mainly a companion to the moisture and seasonal changes that matter more is not settled.

Observation records carry a date but not an absence

A dated record is evidence that a fruit body was observed at a place on a day. It is not evidence that nobody looked there the day before, that the patch was unproductive the following week, or that comparable habitat nearby lacked the species. This is the difficult fact underneath every apparent season chart.

Artportalen is valuable precisely because it gathers observations from private people, professionals, researchers and organisations, and SLU describes several levels of quality assurance, including later expert verification of selected records. Position, date and a photograph can substantially improve a record. Yet reporting effort follows access, weekends, known patches, conspicuous flushes and the attention of people already interested in fungi. A chart of reports is consequently a mixture of biology and observers.

Capinha and colleagues’ Danish analysis is important because it confronts rather than hides this problem. Their framework treats observations as a sample of environmental conditions in which fruit bodies were reported and builds spatial and temporal bias into the comparison. That can support a cautious read of when conditions resemble the observed phenological window. It cannot transform a blank cell into evidence of absence, or a favourable cell into a certainty of fruiting.

The same limit applies to personal memory, though memory has a different strength. Repeated, dated notes from ground already walked can reveal whether one patch tends to retain moisture, whether the host stand has changed, and whether fruiting has occurred after particular autumn conditions. That is fine-grained knowledge because it is a record of real experience. It is not a fine-grained prediction for ground that has not been walked.

The useful uncertainty is still large

Settled: trattkantarell is a mycorrhizal forest fungus with a documented Swedish association with mossy barrskog and, more broadly, acid and nutrient-poor ground with spruce, pine or beech. It is a late-fruiting species in the ordinary sense, and terrain, soil and canopy can make sharply different moisture settings within one slope. Weather and temperature matter to fungal phenology at broader scales.

Also settled: none of those facts reduces the hillside to a rain gauge. The relevant substrate is a living forest-floor system. Moisture maps model tendencies rather than current conditions at a particular patch. A frost statistic describes air temperature by a standard definition, not the temperature of the fungus. A reported observation is presence evidence, never a confident zero around it.

Not settled: the Swedish trigger for fruit-body initiation in C. tubaeformis. There is no defensible national rainfall total, temperature threshold, frost rule or fixed lag that does that job. The Oregon work is a different forest; the Swiss work is a different community; the Danish work models records in a neighbouring but different country. These are useful pieces of the mechanism, not permission to pretend it has been solved.

That uncertainty is not a gap to paper over with a better-looking map. It is the boundary between describing favourable ground and claiming an outcome. The first can be careful, local and useful. The second remains something the forest has not promised.