Impacts of weather and growing season on the occurrence of fires in Finland

Weather is a strong driving factor behind wildland fire occurrence, especially in terms of fuel moisture control. Dead fuel moisture dynamics are mainly regulated by relative humidity, air temperature, wind, solar radiation, and precipitation, and these weather variables have been incorporated into most fuel moisture prediction systems. Live fuels usually have higher and less fluctuating moisture contents than dead fuels. Live fuel moisture is often considered in models as a retardant for ignition. Boreal forests in Finland receive precipitation at fairly regular intervals throughout the year. Extreme fire conditions resulting of prolonged drought occur only as isolated events in both space and time. In this mild fire weather environment, dead-live fuel dynamics may play a very important role in the determination of fire occurrence. There are several seasonal fuel quality processes that potentially influence burning susceptibility and, under some circumstances, override the impact of short-term weather. At the beginning of fire season, right after the snowmelt, fuel material consists in large part of dead or dormant vegetation that has relatively low moisture content and loses that moisture fast in favorable weather conditions. As the growing season proceeds, the accumulation of new vegetation is likely to start suppressing fire activity due to the higher moisture content of live herbs and shrubs. The slowing down of growth processes and the wilting of vegetation may later on again ease the occurrence of fires. So far, there has not been clear understanding of the effects of seasonal vegetation development on fire dynamics in Finland. In this study, we examine fire occurrence trends during fire season and analyze the combined effect of season advancement and fire weather. The results will aid the further development of Finnish Fire Risk Index and fire management practices.

To explore the seasonal trends of fire occurrence, we used national fire statistics 1996-2002 combined with the effective temperature sum (cumulative sum of mean air temperature consisting degrees above +5ºC) during growing season. To study the combined effect of weather and seasonal vegetation development, we analyzed the co-variation of fires and Finnish Fire Risk Index (FFI) in two locations - Kauhava (63º23' N, 24º02' E) and Sodankylä (67º22' N, 26º38' E) – during fire seasons 1996-2003. The FFI is designed to estimate the volumetric moisture content of organic layer (consisting of litter, moss, and humus) in clear-cut areas as a function of precipitation and evaporation. Weather data needed for the FFI calculation were interpolated for Kauhava and Sodankylä from the meteorological station network of the Finnish Meteorological Institute.

In fire statistics provided by the Rescue Service, information reported on each fire consisted of time, location, type of event, and fire size. Seasonal time span of the records ranged from the beginning of April to the end of October. Reported fires occurred within 59º90'-69º10' N and 21°20'-31°00'E. The events had been classified as forest fires, clearing and other open area fires, and peatland fires. Reported fire sizes ranged from 0 to 200 hectares. The reports did not include information on fire weather, fire behavior, or the intensity of suppression actions. In the analysis, we assumed that all fires get the same effort of fire suppression (i.e. they will be put out as soon as possible following the current fire policy in Finland) and as a result, the reported burned areas reflect the “goodness” of burning conditions. There is naturally some case-to-case variation in accessibility to fire events but generally topography and infrastructure do not impose remarkable obstacles to fire-fighting in Finnish landscape. Average daily temperature sum, number of fires, and daily burned area during were calculated to all forest fires reported in Finland during fire seasons 1996-2002. The co-variation of fires, seasonal vegetation development, and Finnish Fire Risk Index (FFI) was examined in Kauhava and Sodankylä.

The number of fires presented two seasonal peaks. The first one occurred during early fire season, from May 12 through June 8, at the average effective temperature sum of below 230 and the second in September 16-26, at the temperature sum of 1200. When average fire frequency was 4.4 fires per day for the total fire season, during the peaks, the number of reported ignitions was at highest 12.9 per day.

Daily burned area, calculated as a product of average size of fire event and number of daily fire events, reached its highest value at the temperature sum of 150-250 and began to decline from there on. The seasonal average of daily burned area was 1.7 ha for the total period of April-October. Daily burned area was highest (36.6 ha) from May 8 through June 15 and peaked again slightly (16.6 ha) later in the fire season, from August 27 through September 22. The comparison of burned area and the FFI showed that the index was prone to overestimate fire risk during mid- and late fire season.

The results reflected the expected impacts of vegetation development on fire activity, the most active fire season taking place from early May through mid-June, the occurrence of fires then declining due to the accumulation of new-growth, and slightly picking up during early autumn the new vegetation has started to wither. The highest values in the number of fire events and burned area in May and in September were slightly surprising since these months in many studies have been considered marginal parts of fire season. According to earlier statistics (mainly years 1900-1950), the three most active fire months in Finland have been July, June, and August. It is, however, difficult to draw conclusions about climate-induced seasonal shifts in fire activity because yearly burned area is nowadays only a fraction of that at the beginning of the 20th century.

The results of this study conclude that the accuracy of the FFI especially at the beginning of the season would benefit from the incorporation of the seasonal vegetation development factor.