Traps were used to quantify charcoal production and transport during three experimental forest fires in Boreal Scandinavia. The traps were spatially arranged to collect charcoal particles inside burn areas, and outside burn areas at different distances (0.1–100 m) from the fire edge. The number of inside and outside traps was 280 and 424, respectively. Trap area was 48 cm2. After the burn, trap content was sorted and sieved in two size-classes of charcoal particles, namely small (0.5–2.0 mm) and large (. 2.0 mm), and number and mass of particles were determined. The production and distribution of charcoal were highly variable at fine spatial scales inside burn areas. On average, inside traps contained 12.1 small and 10.1 large particles, and the average charcoal mass was 0.112 g per trap (corresponding to 235 kg ha-1). The largest size-class made up 94% of the mass. Outside traps contained 0.3 small and 0.1 large particles per trap, and 45% of the outside particles were distributed, 1 m from the fire edge. It is concluded that the occurrence of macroscopic charcoal ($ 0.5 mm) in forest soils provides a solid evidence for local fire influence, and that the presence of large charcoal particles can be used to distinguish between fire-prone and fire-free areas with high spatial precision. Absence of large particles must, however, be more carefully interpreted as 14% of the inside traps lacked macroscopic charcoal. We argue that the charcoal in Boreal forest soils should be less persistent than previously suggested because documented fire-return intervals result in an unrealistic charcoal accumulation presupposing high persistence.

No abstract

I have used occurrence of macroscopic charcoal particles, pollen analyses and radiocarbon datings to examine local forest fire abundance in southern and central Norway. Peat cores, covering the last 1000 to 6000 yr, were sampled from 20 bog margin and swamp forest sites, and the charcoal records documented local fire occurrence in 10 of the sites. Forest fires have not occurred in the sites located in central Norway, whereas the fire occurrence in southern Norway showed large variation among the sites. However, forest fires ceased prior to the establishment of Norway spruce Picea abies in seven of the sites, whereas the establishment of spruce preceded the fire decline in three of the sites. Odds ratio calculations indicated that it is several hundred times more likely that fires occurred prior to, than after, the spruce establishment. Although time spans between fire decline and spruce establishment showed some variation, they did not increase along a gradient from east to west in Norway, suggesting that the establishment of spruce might have initiated a change from fire‐prone to fire‐free ecosystems.

No abstract