Wildfire growth modelling in Sweden - a suitability assessment of available data

Abstract

Wildfires are worldwide problems that stress societies by damaging their economies and they cause serious, sometimes fatal, health implications among the populations. In recent years, the rate of wildfire occurrence has increased as the global average temperature has risen. Because of global climate change, this rate is expected to continue rising. In Sweden, in the aftermath of the 2014 Västmanland wildfire, the Swedish Civil Contingencies Agency called for an urgent need to develop the national emergency preparedness for disastrous events like wildfires. Maps of forecasted wildfire growth can facilitate the work of emergency services trying to limit the consequences of a wildfire. Hence, the introduction of a wildland fire growth simulation model has the potential to strengthen the Swedish societal resilience to wildfires. However, no Swedish guidelines outlining the necessary collection and preparation of spatial information, required by wildland fire growth simulation models, currently exist. Therefore, the aim of this thesis project was to assess the suitability of available spatial data in Sweden for wildfire growth modelling. This was achieved by first reviewing literature to gain the needed understanding of general wildfire behaviour to allow for the theorising of an expected wildfire propagation in the area burnt by the 2014 Västmanland wildfire. Then, the 2014 Västmanland wildfire was modelled using two wildland fire growth simulation models. Their modelled extents were reviewed and compared with real final perimeter of the modelled wildfire. Finally, based on observed deviations, weaknesses in the spatial data were identified. The study concludes that the, in Sweden, available spatial data is sufficient to allow for wildfire growth modelling. However, the temporal resolution of the ground cover information must increase to reflect changes in vegetation before the modelled forecasts can be suitable for any use.

Abstract (Swedish)

Vegetationsbränder är världsomspännande problem som skakar samhällen genom att skada deras ekonomier och orsaka allvarliga, ibland dödliga, hälsokomplikationer bland deras befolkningar. Takten av vegetationsbrandsförekomst har ökat de senaste åren till följd av en stigande global medeltemperatur. Denna takt förväntas att fortsätta öka på grund av den globala uppvärmningen. I efterdyningarna av branden i Västmanland, 2014, rapporterade Myndigheten för Samhällsskydd och Beredskap om ett akut behov av att öka den nationella krisberedskapen i Sverige för katastrofala händelser såsom vegetationsbränder. Kartor över prognostiserad vegetationsbrandstillväxt kan underlätta räddningstjänstens arbete med att begränsa konsekvenserna av löpeld. Om ett modellsystem för simulering av löpeldstillväxt implementeras i Sverige kan motståndskraften i samhället mot vegetationsbränder stärkas. Det finns dock inga riktlinjer som beskriver den nödvändiga insamlingen och beredningen av de rumsliga datamängder som krävs av ett sådant modellsystem. Därför var syftet med detta examensarbete att bedöma lämpligheten av tillgänglig rumslig data för modellering av vegetationsbrandstillväxt. Detta uppnåddes genom att först utföra litteraturstudier för att skapa en nödvändig teoretisk förståelse av förväntad brandtillväxt under en referensbrand. Därefter modellerades referensbranden med hjälp av två befintliga simuleringsmodeller. Den prognostiserade löpeldstillväxten jämfördes med referensbrandens kända perimeter och med hjälp av den teoretiska förväntningen kunde slutligen svagheter i den rumsliga datan identifieras. I detta arbete dras slutsatsen att upplösningen av den geografiska data som finns tillgänglig i Sverige är tillräcklig för modellering av brandtillväxt i vegetation. Däremot måste marktäckesinformationens temporala upplösning öka från dagens fem år till en upplösning som kan återspegla naturliga variationer i vegetationen. Fram till dess förblir modellering av vegetationsbränder olämpligt för all typ av användning i Sverige.

Popular Abstract

Wildfire growth modelling can facilitate the work of emergency services – but currently not in Sweden

Because of the wildfires causing damages amounting to hundreds of millions euro in recent years in Sweden, the Swedish Civil Contingencies Agency reported the need to increase the emergency preparedness for complex events such as wildfires. Wildfire growth modelling can strengthen the societal resilience to wildfires. However, the guidelines outlining the collection of the necessary spatial data currently do not exist in Sweden. Are data sets already available sufficient for wildfire growth modelling?

In order to understand how the growth of a raging wildfire can be modelled it is important to introduce two concepts describing a fire. Firstly, a fire is not as different to water as you might think. Figuratively speaking that is. Water is commonly referred to as a body of water. This body consists of a constellation of water molecules. Similarly, a fire is a constellation of flames. Secondly, a common misconception is that a fire spreads. A fire cannot spread. Instead, it grows when new flames are added to its body. Every flame constantly transfers energy in all directions, just like a radiator in your home, meaning the fire heats surrounding matter. Eventually, pyrolysis (i.e. the process in which matter decomposes into fire fuel) starts. Subsequently, the fire grows when ignition of the newly released fuel takes place. In contrast, a fire reduces in size when flames go extinct. Hence, the fire may appear to be spreading when in fact it is growing in one direction while decreasing in size in another. Continuous ignitions of new flames happen as long as a mixture composition of fuel, oxygen and heat remains balanced.

A wildfire is an uncontrolled fire that grows in vegetated areas. Since the amount of oxygen and pyrolysable matter in nature is almost inexhaustible, a wildfire has the potential to grow indefinitely. The rate at which the wildfire grows is controlled by environmental factors that describe the terrain, ground cover and weather of the area where the wildfire exists. Data sets containing this spatial information are currently available in Sweden. The spatial resolution (i.e. resolution of the area represented by a single value in a data set; denoted as the length of one side of a squared pixel) of these data sets is adequate to allow for wildfire growth modelling either as it is or after a resolution increasing process. Likewise, the temporal resolution (i.e. sampling rate; update frequency) of the terrain and weather information is sufficient. However, the ground cover information has a current planned update interval of five years meaning it cannot reflect variations (e.g. droughts; harvests; storms) in nature. In fact, its temporal resolution does not even reflect natural variations during the growing season. Therefore, the sampling rate of the ground cover data must increase before modelling of wildfire growth can be used in Sweden to strengthen the societal resilience to wildfires.

• author: Hagelin, Henrik
• supervisor: Veiko Lehsten
• organization: Dept of Physical Geography and Ecosystem Science
• course: GISM01 20222
• year: 2023
• type: H2 - Master's Degree (Two Years)
• subject: Earth and Environmental Sciences
• keywords: Geography, GIS, Geographic Information Science, Physical geography, Prometheus, WiSE-FASS, The 2014 Västmanland wildfire, Wildfire growth modelling, Wildfire suppression support, Wildfire behaviour, MESAN
• publication/series: Master Thesis in Geographical Information Science
• report number: 153
• language: English
• id: 9106431
• date added to LUP: 2023-01-11 16:05:10
• date last changed: 2023-01-11 16:05:10