Case 3: Sweden 1 - Wildfires

The following is a preliminary assessment of how Sweden will be affected by climate change - the main hazard in focus, in this case, is wildfires. The assessment will be developed further as the project progresses

Sweden is, short of Russia, the 2nd most forested country in Europe with about 75 % forest coverage of its landmass outside its mountainous regions. Sweden also produces 8 % of the world’s sawn wood export. A few large recent incidents, such as the intensive mega-fire in Västmanland in 2014 and the dry summer of 2018 which resulted in over 25 000 ha burnt in Sweden, have highlighted the potential effect of wildfires on people and communities, especially in the rural inland of the north.

Fire regimes in a century-long retrospect

Before 1870, fire return intervals (i.e. time between fires in a small defined area) averaged at 80-200 and 30-50 years in the northern and southern parts of Fennoscandia, respectively (Niklasson & Granström, 2000; Niklasson & Drakenberg, 2001; Wallenius et al., 2010; Rolstad et al., 2017). These intervals are on par with modern-day intervals in some other parts of the boreal region, such as Canada (Whitman et al., 2019; Ter-Mikaelian et al., 2009; Rogeau, 2016).

After relatively stable high fire return intervals for centuries, dramatic changes occurred during the late 1800s. Fire frequencies dropped dramatically with the onset of organized industrial forestry. There was a state-supported incentive for both more organized suppression efforts and a decline in the cultural use of fire (i.e. release of nutrients, removal of unwanted crops, and clearing of land) (Granström & Niklasson, 2008). Fire suppression became highly effective, originally based on mass mobilization of the rural population and in later decades it became mechanized, with the support of smaller suppression crews. As a result, the burnt area per year was reduced by more than 99 % (Figure 2) and the fire return interval today is 15 000 years in southern Sweden and 7 000 years in northern Sweden (recalculated from Sjöström & Granström, 2020).

Figure 1 Remains of a destroyed recreational house in a clear-felled and completely scorched landscape after the fires in Ljusdal, 2018. Photo: Johan Sjöström

Figure 2 Cumulative number of fires from 1300 to 2000 in a 600 km2 study area in Västerbotten, north Sweden. The figure shows observed (solid line) and estimated (dashed line) fire. Adapted from Niklasson and Granström (2000).

Projected fire danger for future climate

Studies of three projected IPCC emission scenarios have shown that increased frequencies of fire-prone weather and a prolonged fire season are expected (3-20 days by 2050 and >20 days by 2099) resulting in severities increasing by a factor of 2-3 during this century (Flannigan et al., 2013). Other studies show that the north is likely to be less impacted by heavy droughts but that the effect will be stronger in the southeast (Yang et al., 2015; Sjökvist et al., 2013).

However, ongoing studies of the past 160 years of fire danger show that the combination of higher temperatures and increased precipitation have no clear effect on the resulting fire danger. So far, no clear trend can be noticed other than that large variations exist between years and that the last two decades (with 2018 as an exception) constitute a period of relatively low fire danger (Figure 3), especially in terms of long periods of very high danger (Sjöström and Granström, in progress). Thus, there is so far no clear trend towards increased fire danger due to changing climate, although this could be the case for the coming 30-50 years.

Figure 3 Cumulative yearly Duff Moisture Code (CDMC) – a scalar estimate of a season's fire danger for Storuman (inland north), Umeå (north coast), and Västervik (south-east) from 1860-2020 (Sjöström and Granström, in progress).

Trends in land-use

Three native species dominate the Fennoscandian forests, Scots pine (Pinus sylvestris), Norway spruce (Picea abies), and Wart birch (Betula spp.). Since the end of the 1960s, Lodgepole pine (Pinus contorta – Figure 4) has been imported from North America and covers part of the inland north, currently over 600 000 ha. P contorta is more fire-prone compared to the native species due to its larger surface fuel bed mass (Lundmark et al., 1982), more porous structure due to its longer needles (Granström, 1998), and increased fuel ladder towards the crown by less shredding of dead branches (Bäcklund et al., 2018).

The pre-commercial thinning (Figure 4) of birch, without which many stands on fertile soil would become dominated by deciduous trees, has been increasingly aggressive during the 20th century. The impact of broad-leaves in Fennoscandia is not clearly known but generally, crown fires do not occur if stands are broad-leaved dominated (van Wagner, 1977) and even minor inclusion will impact the surface fuel by smothering the moss and lichen component (Schimmel & Granström, 1997).

Another factor of land-use is the drainage of wetlands that have been common practice throughout the regions since the 19th-century onset of commercial forestry. The wetlands do not always act as fire barriers, but their presence locally decreases fire intensity, aiding suppression efforts (Bohlin et al., 2017). Drained wetlands also exhibit deep dry peat layers which favor deep smoldering, which in turn complicates the mop-up and final extinguishment. The drainage rate of Scandinavian forests increased with the aid of more accessible machinery up to the 1980s but is now banned on previously undrained land for biodiversity reasons (Päivänen & Hånell, 2012).

Finally, the road network in Sweden and Finland is very dense from a boreal perspective. Continuous development of forestry roads is widespread and they now constitute about 40 % of all roads outside residential areas (Karlson & Mörtberg, 2015). The road network aids suppression efforts by allowing access of firefighting crews closer to the fires and by the easier establishment of fire breaks. Even without suppression efforts, the presence of roads can constitute spontaneous fire stops. To be even more effective, especially as a means for evacuation of residents, there is a need for connecting the dead-end roads that now are prevalent across the rural areas of the region (Vermina Plathner & Sjöström, 2021).

Although the inclusion of P contorta, the drainage of wetlands as well as the densification of the forestry road network comes from century-long traditions, their legacy remains and has a large impact on today’s fire occurrence, burn probability, and fire behavior. Thus, decisions taken today concerning these issues will have long-standing effects on the situation by the end of this century. Pre-commercial thinning of broad-leaved species and prescribed burning of heather has, on the other hand, a shorter response time for the fire situation and is thereby effective for the short-term future (i.e. decades).

In short, to date, there is no clear indication that the periods of fire-prone weather have increased since the 1800s but projections suggest this can be the case for the coming decades. The dramatic change in fire activity from the mid-1800s is a direct consequence of effective and organized suppression efforts. Usually, it takes decades to see the consequences of changing land-use policies, and recent trends such as highly aggressive thinning of birch, drainage of wetlands, and introduction of P contorta all suggest an increased risk of large forest regardless of the weather situation.

Figure 4 Right: Mixed conifer forest in Sweden including Norway spruce and Scots pine. Upper left: Pine forest subject to effective pre-commercial thinning. Lower left: Needles and cone from the exotic Lodgepole pine, introduced in Sweden for logging. Photos: Johan Sjöström

Vulnerability

Trends in organization for warning and response

As a tradition in many parts of the world, fire has long been a tool for forest management in Fennoscandia. In pre-industrial times the whole community had a strong incentive for using fire as a land management tool and it balanced its usefulness against the threat it constituted for crops, cattle, and dwellings. Should wildfires occur, even neighboring villages were obliged by law to assist in Sweden and Finland since the 18th century (Granström & Sjöström, in progress). The use of municipal fire brigades also in wildfire suppression increased with the denser road network and motorized fire engines and is now completely dominating total suppression strategies.

Today, the Civil Protection Act (2003:778) dictates that municipal fire brigades are responsible for suppression. Landowners are then obliged to monitor scorches after the immediate danger has passed. Thus, landowners should operate mop-up and monitoring after the spread of a fire is controlled. This work is often performed by low-wage seasonal employees whose main job description is site preparation, tree planting, and pre-commercial thinning and thereby lack experience in managing wildfires. Poor mop-up is one of the major key components in the devastating summer of 2018 (Granström, 2020).

The ongoing and several decade-long trends of depopulation of rural areas (as well as throughout the world) is particularly strong in some areas in Sweden – up to 30 % in some areas in the last decade (Lindblad et al., 2015). This also has strong implications for the vulnerability of the remaining inhabitants. Since tax revenues decrease with the population, so does the firefighting resources in already sparsely populated areas. In 2014, the ten municipalities in Sweden with the highest cost per capita for fire rescue services were all in the sparsely populated north, whereas the ten with the lowest costs were all in or near the larger population centers in the south (MSB, 2015). The average cost between these two groups varied by a factor of 4.4.

With fewer firefighters and fewer manned stations, the time from alarm to arrival at the ignition site increases. The chance of an ignition growing out of control increases exponentially with increasing time to arrival (Sjöström & Granström, 2020). During the last two decades, a trend of increasing time to arrival for the municipal forces is noticed throughout the whole country (Figure 5) but is most significant in regions with a low population density. Thus, the fraction of large fires is larger in rural Sweden compared to more densely populated areas and this difference is increasing.

Figure 5 Area intervals of fires against the median time to arrival within each interval. The red line shows an exponential fit to the data. The inset shows the median arrival time in the whole country for fires on open land, forested land, and forested land where the fire grows >0.5 ha. The solid lines are polynomial fits to the data.

The 19th-century north-southern gradient for fire return intervals has today changed to a sharp negative gradient with population density. Thus, fewer fires occur in sparsely populated regions. But the ones that do can spread and grow to much larger areas (Sjöström & Granström, 2020).

Due to the variability between years, it is impossible to maintain the capacity to meet demands in peak years. During 2018, the suppression forces were quickly exhausted since the high fire danger was prevalent in most of Sweden. During the summer, up to 30 % of the suppression efforts came from abroad. Cooperation within the rescEU program opens the possibility for resources to be dispatched towards the greatest current need. MSB has now signed an agreement for both scooping aircrafts and helicopters as part of rescEU which are stationed within Sweden. This has boosted the financial possibilities and willingness of municipalities to use aerial support earlier in the response, which was rather low back in 2014.

Impact on people and property

A few Swedish cities (e.g., Umeå, Sundsvall, Luleå, and Karlskrona) were completely destroyed by roaring fires during the late 19th-century. The rapid fire spread between buildings was attributed to the wooden facades, combustible roofing, and the closeness between homes that were typical of cities around that time (Palmgren, 2006). Only one timber building survived the 1888 Sundsvall fire out of all buildings in the city center. This building was located behind the city park and was thereby protected by broad-leaved vegetation (Boström & Dahlgren, 1988).

Along with the build-up of the destroyed towns, city planning evolved to take the risk of fire destruction into account. The inclusion of broad-leaved species within and in the surroundings of cities, broader streets, brick exteriors, and tile roofs on buildings have been the main precautionary measures (Palmgren, 2006). The effect of these preparedness actions is still discernible, as only a handful of low-intensity wildfires during the past decades have approached city interfaces, without really threatening the structures. Herbaceous species and broad-leaved vegetation are the most common surrounding the buildings in medium- and high-density populated areas. However, in rural areas (i.e. low-populated wildland areas and intermix areas), it is more common that fire-prone wildland such as pine stands surround the structures (Vermina Plathner & Sjöström, 2021). The main feature of passive protection to houses in rural areas is the existence of a managed lawn around the structure and the presence of deciduous trees around the garden (Figure 6). Also, avoiding concentrations of combustibles close to the facade is shown to contribute to the survival of a house (Vermina Plathner et al., in progress).

Figure 6 After the Sala fire, April 2015. The foreground shows what previously was a vegetated clear-felled area. The surrounding coniferous forest burned with crown fire. However, the deciduous forest strip containing birch, sallow, and aspen stopped the fire before it reached the building in the background. Photo: Kenneth Risberg.

Even without the direct economic losses, such as forested land or buildings, the changes in the landscape can have high consequences. Areas that have to be clear-felled are sometimes much larger than a typical clear-felled area and therefore the whole surrounding environment can change drastically. People no longer recognize places they grew up in and that they might have a strong attachment to. Studies from similar events show that some social bonds grow stronger while old conflicts can also continue. A stronger sense of distance between people affected by and not affected by the crisis is also a feature that is stronger in rural areas compared to regions with a denser population (Sellerberg, 2011). However, community bonds are generally shown to be strengthened within the community after, for example, the Västmanland fire 2014.

References

Boström, S. & Dahlgren, D. (1988) Den tändande gnistan: Staden som reste sig ur askan, Sundsvall Tryckeribolaget Sundsvall AB (in Swedish)
Bohlin, I., Olsson, H., Bohlin, J. & Granström, A. (2017) Quantifying post-fire fallen trees using multi-temporal lidar. International Journal of Applied Earth Observation and Geoinformation, 63, 186-195.
Bäcklund, S., Jönsson, M., Strengbom, J. & Thor, G. (2018) Tree and stand structure of the non-native Pinus contorta in relation to native Pinus sylvestris and Picea abies in young managed forests in boreal Sweden. Scandinavian Journal of Forest Research, 33, 245-254.
Flannigan M. D., Cantin A. S., de Groot W. J., Wotton M., Newbery A. & Gowman L. M. (2013) Global wildland fire season severity in the 21st century, Forest Ecology and Management 294, 54–61.
Granström A. (1998) Framtidens skogsbränder : ändrad brandrisk genom förändrad skogsskötsel. Räddningsverket (in Swedish)
Granström A. (2020) Brandsommaren 2018: Vad hände, och varför?, MSB1496, Myndigheten för Samhällsskydd och Beredskap, Karlstad (in Swedish).
Granström, A. & Niklasson, M. (2008) Potentials and limitations for human control over historic fire regimes in the boreal forest. Philosophical Transactions of the Royal Society B-Biological Sciences, 363, 2353-2358.
Granström, A. & Sjöström J. (in progress) Is fire coming back to boreal Sweden? Countering negative impact of societal trends and land use legacies on future wildfire situation
Karlson, M. & Mörtberg, U. (2015) A spatial ecological assessment of fragmentation and disturbance effects of the Swedish road network. Landscape and Urban Planning, 134, 53-65.
Lindblad, S., Tynelius, U., T, D., W, P. & Anderstig, C. (2015) Demografins regionala utmaningar. Bilaga 7 till Långtidsutredningen 2015. Statens offentliga utredningar 2015:101 (in Swedish)
Lundmark, J., Berg, B. & Nilsson, Å. (1982) Contortatallens inflytande på mark och markvegetation i jämförelse med sylvestristallens. Sveriges Skogsvårdsförbunds Tidskrift, 43-48 (in Swedish)
MSB (2015) Räddningstjänst i siffror 2014. Karlstad, Sweden: Swedish Civil Contingency Agency (in Swedish)
Niklasson, M., & Granström, A. (2000). Numbers and sizes of fires: Long-term spatially explicit fire history in a Swedish boreal landscape, Ecology, 8, 1484–1499.
Niklasson, M., & Drakenberg, B. (2001) A 600-year tree-ring fire history from Norra Kvills National Park, southern Sweden: Implications for conservation strategies in the hemiboreal zone, Biological Conservation 101, 63-71
Palmgren, G. (2006) Stadsingenjörer i Luleå 1864-1984 Från den förste till den sista, Luleå. Luleå kommun repro (in Swedish)
Päivänen, J. & Hånell, B. (2012) Peatland ecology and forestry–a sound approach, University of Helsinki.
Rogeau M.-P (2016) Fire regimes of southern Alberta, Canada, PhD Thesis, University of Alberta, Edmonton.
Rolstad, J., Blanck, Y.l. & Storaunet, K.O. (2017) Fire history in a western Fennoscandian boreal forest as influenced by human land use and climate. Ecological Monographs, 87, 219-245.
Sellerberg, A. M. (2011) Efter stormen, En sociologisk undersökning av skogsägarfamiljer, Bokbox förlag. Malmö (in Swedish)
Schimmel, J. & Granström, A. (1997) Fuel succession and fire behavior in the Swedish boreal forest. Canadian Journal of Forest Research, 27, 1207-1216.
Sjökvist E., Axén Mårtensson J., Sahlberg J., Andréasson J. & Hallberg K. (2013) Framtida perioder med hög risk för skogsbrand - Analyser av klimatscenarier, MSB535, Myndigheten för Samhällsskydd och Beredskap, Karlstad (in Swedish).
Sjöström J., & Granström A. (2020) Wildfires in Sweden - trends and patterns during recent decades, MSB1536, Myndigheten för Samhällsskydd och Beredskap, Karlstad (in Swedish).
Ter-Mikaelian M. T., Colombo S. J. & Chen J. (2009) Estimating natural forest fire return interval in northeastern Ontario, Canada, Forest Ecology and Management 258, 2037-2045.
van Wagner, C.E. (1977) Conditions for the start and spread of crown fire. Canadian Journal of Forest Research, 7, 23-34.
Vermina Plathner F. & Sjöström J. (2021) The wildland-urban interface in Sweden, Technical Note TN 7.1 from the WUIVIEW project, www.vuiview.org.
Vermina Plathner, F., Sjöström, J. & Granström, A. (in progress) Factors of structure survivability in Swedish wildfires
Wallenius, T.H., Kauhanen, H., Herva, H. & Pennanen, J. (2010) Long fire cycle in northern boreal Pinus forests in Finnish Lapland. Canadian Journal of Forest Research, 40, 2027-2035.
Whitman E., Parisien M.-E., Thompson D. K. & Flannigan M. D. (2019) Short-interval wildfire and drought overwhelm boreal forest resilience, Scientific Reports 9, 18796.
Yang, W., Gardelin M., Olsson J., & Bosshard T. (2015) Multi-variable bias correction: application of forest fire risk in present and future climate in Sweden, Natural Hazards and Earth System Sciences 15, 2037-2057.

This project has received funding from the NordForsk Nordic Societal Security Programme under Grant Agreement No. 97229