Regional Assessment of Global Change Impacts pp 639-650 | Cite as
Climate-Related Forest Fire Risk
- 1 Citations
- 1.1k Downloads
Abstract
An increase in the risk of forest fires in Central Europe is seen as a likely consequence of global warming. Therefore, timely planning of measures to adapt is necessary and requires the evaluation of specific hazards in affected regions. Knowledge about potential regional effects of climate change on the risk of forest fires is required to protect the forested regions in the Upper Danube basin. The forest fire index of Baumgartner, which is based on the forest fire statistics from Bavaria, was implemented in a forest fire module within DANUBIA. The future risk of forest fire was assessed in this study under the conditions of the Remo regional climate trend and the Baseline climate variant. The potential future development of the hydroclimatic conditions during March to September was evaluated using temperature, precipitation, potential evaporation and the climatic water balance. The model results indicate a significant increase in the forest fire risks within the Upper Danube basin as a consequence of climate warming. Under the assumbtions of the Remo regional climate trend and the Baseline climate variant, the number of days with “high” and “exceptionally high” forest fire risk more than doubled in the long-term average for the period from 2011 to 2060. Very dry years with extreme fire hazards may become more common. Such hazards were never reached in the past. For the Upper Danube area, an increase in days with significant forest fire risk may be expected especially for the summer months.
Keywords
GLOWA-Danube Upper Danube Forest fire risk Forest fire Modelling Climate change
Climate-related forest fire risk (REMO regional climate trend and Baseline climate variant)
74.1 Introduction
An increase in the risk of forest fires even in areas with previously low risk is seen as a likely consequence of climate warming in Central Europe (IPCC 2007). Therefore, timely planning of measures to adapt is necessary and requires the evaluation of specific hazards in affected regions. In light of the huge potential damage arising from forest fires, their precautionary prevention has great social relevance. Thus, knowledge about the regional effects of climate change on the risk of forest fires is required to protect the forested regions in the Upper Danube basin (approximately 32 % coniferous forest and 8 % deciduous and mixed forest). Extensive forested areas are mostly found in the Alpine region and the low mountain ranges (see Chap. D.3).
Generally, the formation of wildfires requires both direct sources of ignition and dispositive factors. The ignition potential of plants depends on the supply of water in and to the plant matter and is thus controlled by meteorological processes. Ignition itself is primarily caused by anthropogenic factors (mostly from arson and acts of negligence). At our latitudes, only 3–10 % of all forest fires arise as a result of lightning strikes (Badeck et al. 2003). Temperature and precipitation over the course of the year are the most important dispositive factors. Long periods without precipitation and concurrent high rates of evaporation associated with sunny weather and high temperatures lead to situations in which the litter layer and, in cases of low groundwater levels, the vegetation are very dry and hence susceptible to fire (Badeck et al. 2003). Moreover, the season and the forest structure, especially the composition of tree species and the stand coverage, are important factors. Young species up to 40 years and the generally drier pine stands of all age groups represent the forests with the greatest risks of fire (Wiese 2001). 40–60 % of all forest fires occur in the drier spring months (March to May), when the dried-out ground cover and vegetation are still not protected by fresh grass and herb growth (König 2007). The flammability of the forest ground cover significantly declines with the full development of the vegetation over the subsequent months.
For the assessment of forest fire risk and their prediction, respectively, the aim is to develop meteorological methods to assess the regional fire hazard. To respond to imminent forest fires with a preventative approach, mostly multiple day forecasts are applied. Various methods to assess the fire hazard have been used in different countries for many years (Alexander et al. 1996; Langholz and Schmidtmayer 1993; Lawson and Armitage 2008). The forest fire index of Baumgartner et al. (1967) has been used for the Upper Danube basin in this study; this method has also been the method of choice by the German Weather Service since the 1970s. This hazard key, which is based on the forest fire statistics from Bavaria for the years 1950–1959, provides the basis for the forest fire module implemented in DANUBIA; hence, this model can calculate the weather-related risk of a forest fire for each proxel. The future risk of forest fire was determined for this chapter under the conditions of the REMO regional climate trend and the Baseline climate variant (see Chaps. 47, 48, 49, 50, and 51). The development of the risk of forest fires under further GLOWA-Danube scenarios is presented in Chap. 75.
74.2 Description of the Forest Fire Module
Cumulative curve of frequency of the Baumgartner index (March–September, III to IX) of 1706 forest fires in Bavaria 1950–1959 (Baumgartner et al. 1967)
Mean air temperature March–September for the periods 1960–2005 and 2011–2060 in the Upper Danube basin
Standard values (in mm) of the Baumgartner index as a measure for the forest fire risk of the following day in Bavaria (Baumgartner et al. 1967); risk category: 1 = very low, 2 = low, 3 = moderate, 4 = high, 5 = very high
| Risk category | 1 | 2 | 3 | 4 | 5 |
|---|---|---|---|---|---|
| March | <−5 | −5 to 2 | 3 to 8 | 9 to 14 | >15 |
| April | <−3 | −3 to 7 | 8 to 15 | 16 to 26 | >27 |
| May | <3 | 3 to 15 | 16 to 24 | 25 to 34 | >35 |
| June | <12 | 12 to 23 | 24 to 31 | 32 to 40 | >41 |
| July | <12 | 12 to 23 | 24 to 30 | 31 to 39 | >40 |
| August | <8 | 8 to 19 | 20 to 27 | 28 to 36 | >37 |
| September | <8 | 6 to 17 | 18 to 25 | 26 to 34 | >35 |
In addition to precipitation sums and potential evaporation, the forest fire module calculates the meteorological risk of forest fires for each day.
The model results were compared to the results from the German Weather Office (Deutschen Wetterdienstes = DWD) for 2002 and 2003 in order to verify the forest fire module. The year 2002 was chosen as a comparatively average climatic year. In contrast, 2003 represents an exceptionally hot and dry year with particularly high values for forest fire risks. The evaluation was performed both visually and using a contingency matrix and a weighted kappa coefficient 0 ≤ ĸ ≤ 1 (Grouven et al. 2007) as a measure of the agreement between the two models.
The same risk value was calculated by both models for 2002 in 76.5 % of cases for 11 climate stations (2003: 61.1 %). The DWD model calculated a risk level that was higher in 11.4 % of cases (2003: 23.3 %) and lower in 10.1 % of cases (2003: 8.6 %). The kappa coefficient for 2002 was ĸ = 0.84 and for 2003 was ĸ = 0.52. The models are thus highly consistent for 2002 and moderately so for 2003.
The deviations for 2003 show a tendency for higher risk assessment in the DWD model. Differing methods for calculating the potential evaporation underlie these deviations to some extent. In the forest fire module, the Penman method was used, whereas the DWD selected the method derived by Haude.
74.3 Results
The results of the forest fire module are presented as regional mean values and specialised as maps for the past and a future scenario. The assessment of the future forest fire risk is based on the REMO regional–Baseline climate scenario. Only the months from March to September were analysed, since according to Baumgartner et al. (1967), there is a real risk of forest fires only during the growing season. The conditions in spring (March to May) and summer (June to August) are evaluated separately. A second focus was set on the climatic water balance (the difference between precipitation and potential evaporation), which is considered as a measure of the potential water yield in a region over the long-term average.
The regional mean values were analysed for the periods 1960–2005 and 2011–2060 and the modelled data by area were analysed for three decades (1991–2000, 2021–2030 and 2051–2060). The years 1991–2000 represent the present conditions in the Upper Danube basin. Hence, this period served as the reference decade. The decade 2021–2030 was chosen to estimate the trend in the near future, and the decade 2051–2060 should reveal the conditions that might dominate in around 40–50 years. The effects of climate change are clearly apparent by this time. The hydroclimatic conditions and their changes within the basin are presented first, and then the forest fire risk within the Upper Danube is described and compared among the decades.
74.3.1 Hydroclimatic Change
A potential future development of the hydroclimatic situation (March to September) is illustrated using temperature, precipitation and potential evaporation as well as the climatic water balance.
Temperature
Figure 74.2 shows the trend in mean temperatures within the growing season for 1960–2060. The consistent increase by 4 °C over the course of these 100 years is clearly evident. The highest values to date were recorded in the year 2003 with a mean of 12.0 °C. Such extreme years are likely to occur at greater frequencies over the next 30 years. Although the mean temperature increased by 0.38 °C per decade in the past, the increase per decade under scenario conditions is 0.52 °C, and hence a mean daily air temperature of 14.5 °C is projected.
Precipitation
Compared to the period 1991–2000, a decline in precipitation during the growing season by approximately 10 % is simulated for the future up to 2060 (see Map 74.1 – 1a–c): the means decrease from 785 to 700 mm. The driest regions are in the Nördlinger Ries district, the Upper Palatinate, in the region south of Eichstätt and in the Dungau district south of the Bavarian Forest between Regensburg and Straubing. The highest values can be seen in the high-altitude regions of the Alps, where the decrease in precipitation is the most marked (see also Chap. 53).
The precipitation sums fluctuate widely from year to year. In 2002, there was 949 mm of precipitation, but in the drier 2003, only 593 mm, which is less by a third. In the climate scenario period (2011–2060), a mean of only 520 mm precipitation sum was calculated for the growing season.
Evaporation
The increasing sums for potential evaporation are closely associated with the increasing temperatures (see Figs. 74.2 and 74.3 and Map 74.1 – 2a–c). During the period 1960–2005, there was an increase in the evaporation sums of 11 mm per decade. The trend calculated for the future (2011–2060) is more than twice as high, at 23 mm per decade. The highest measured value was in 2003 (598 mm). Future annual evaporation sums of up to 656 mm are even possible (see also Chaps. 53 and 58).
Map 74.1 – 2a–c indicates that the potential evaporation is highly correlated with elevation: the lowest values are from the high altitudes in the Alps, whereas high values are found in the Hallertau, Dungau and Inn valley regions. Values above 600 mm, which are seen today only at few locations, are expected across wide areas of the lowlands. Evaporation will increase even at higher elevations. As a result, rates of evaporation will no longer differ in central uplands compared to the northern borders of the Alps.
Climatic Water Balance
A significant decrease in the potential water available during the growing season is evident for the Upper Danube basin (see Fig. 74.3 and Map 74.1 – 3a–c). This result is the consequence of decreasing precipitation and increasing rates of evaporation. The measurement data from the past (1960–2005) indicate a decrease of 7.3 mm per decade. This trend is projected to strengthen in the period from 2011 to 2060: the average decrease reaches 45 mm per decade. The conditions in 2003 provide a glimpse of what future conditions could be like: that year, a negative water balance was reached for the first time (−4.2 mm). Figure 74.3 indicates that a water deficit with maximum of up to −116 mm may arise more frequently in the future, especially after the second third of the twenty-first century.
Precipitation, potential evapotranspiration and climatic water balance in 1960–2005 and 2011–2060 in the Upper Danube basin
Map 74.1 – 3a–c shows the distribution of the potential water yield by forested area (see Chap. 9). The water deficits that prevail in the northern parts of the Upper Danube basin decrease towards the south. The forested regions of the low mountain ranges have values up to 600 mm in the high-altitude regions of the Bavarian Forest; these values represent a smaller water excess compared to the Alpine region, where maximum values are found around the northern Alpine fringe (up to 1,827 mm). Along the central alpine dry valleys, some regions have negative values.
In the decade from 2021 to 2030, there may be significantly less water available for the forests compared to today. The potential water yield is reduced by 34 % compared to 1991–2000. Regions with the greatest water deficits include Lower Bavaria and especially the districts of Dungau and Upper and Middle Franconia. In contrast, there is still a water excess (>100 mm) in the high-altitude regions of the low mountain ranges (the Bavarian Forest, the Swabian Jura) and in the south of Upper Bavaria, Swabia and in the Alps, although there are lower maxima in the high alpine regions.
On average during the period 2051–2060, only 45 mm of water is projected to be potentially available during the growing season. This is equivalent to a decrease of 83 % compared to 1991–2000. A marked water deficit dominates across all of Lower Bavaria, the Upper Palatinate and in Upper and Middle Franconia. In addition, lower values arise across the Alpine foothills, except in the highest altitudes of the low mountain ranges. Regions with a potential water deficit also include the river valleys of the Alps in the south-western part of the basin. The water surpluses of the Alps decline by more than 500 mm compared to 1991–2000. Even the regional maxima in the high alpine regions decrease by 36 % compared to 1991–2000.
Precipitation, potential evapotranspiration and climatic water balance in spring (March–May) 1960–2005 and 2011–2060 in the Upper Danube basin
Precipitation, potential evapotranspiration and climatic water balance in summer (May to August) 1960–2005 and 2011–2060 in the Upper Danube basin
A sharp decline in precipitation is calculated for the future under the assumed scenario. Along with a further increase in evaporation, the result will be a marked decline in the potential water that is available. After 2030, a positive climatic water balance is calculated only in exceptional cases. Already within 20 years under the scenario conditions, warm and dry summers like that experienced in 2003 will be the norm.
74.3.2 Forest Fire Risk
Number of days with forest fire risk category 4 or 5, after Baumgartner et al. (1967), 1960–2005 and 2011–2060 in the Upper Danube basin
Areal mean number of days in spring (Mar–Apr) and summer (Jun–Aug) with a forest fire risk category 4 or 5, after Baumgartner et al. (1967), 1960–2005 and 2011–2060 in the Upper Danube basin
For the decade 1991–2000 (see Map 74.1 – 4a), there are a negligible number of days (5–8) with severe forest fire risk over the entire basin. For regions with especially high rates of evaporation, the fire hazard is classed slightly higher (9–11 days). In the low mountain ranges and the Alpine region, there is a lower forest fire risk as a result of lower evaporation rates and higher total precipitations. One exception is the central alpine dry valleys, where there is a higher forest fire risk, primarily in the Inn valley, with a maximum of 14.5 days per year.
Compared to the reference decade, the decade 2021–2030 (see Map 74.1 – 4b) shows an increase in the number of forest fire days by 84 % over the whole basin, except in the high-altitude regions of the low mountain ranges. The main reason is the gradual increase in potential evaporation. Precipitation totals remain largely unchanged in the lowlands. Mostly, the regions in Franconia and the Upper Palatinate are affected by a slight decline in precipitation sums. A negative climatic water balance is calculated for the northern portion of the drainage basin and in the Tertiary Hills. The greatest decrease in precipitation takes place in the high alpine regions, but this area remains quite moist as a result of low evaporation rates. The entire Alpine region, with the exception of the central alpine dry valleys, is also with low risk of forest fires in the future. However, primarily in the Inn valley, the risks that are already quite high today may double. The most notable increase, at 150 %, was calculated for the number of days with greater and unusually higher risk of forest fires in the Alpine foothills, especially in the Fünfseenland region of Bavaria. The highest forest fire risks (13–17 days) were calculated for the Upper Palatinate, Franconia and large areas of Lower Bavaria, with increases of 100 %.
The scenario period 2051–2060 also shows an increase in the forest fire risk by 86 % compared to the reference decade. However, there is a rather different picture for the region north of the Alps. With only a negligible increase in forest fire risk, the western part of the drainage basin (Swabia, the Swabian Jura) has little risk (<8 days) – in some areas there is even a slight decrease. The forest fire risk may double in the Upper Palatinate, Central Franconia, the Tertiary Hills and in eastern Upper Bavaria and the Bavarian Fünfseenland region (10–13 days). In some regions of Central Franconia, Lower Bavaria and the Upper Palatinate, the risk even triples (16–23 days). In the high-altitude regions of the low mountain ranges and in the Alps, the forest fire risk is projected to remain low (0–5 days) despite an ongoing slight increase. Although rates of evaporation increase and precipitation totals decrease, the water balance in these regions continues to be distinctly positive. In the central alpine dry valleys, the forest fire risk is quite high (up to 23 days; an increase of up to 150 %); this result is quite likely the consequence of the highest evaporation totals in the basin.
74.4 Conclusion
The model results indicate significant changes in the forest fire risks within the Upper Danube basin as a consequence of climate warming. Under the assumptions of the REMO regional climate trend and the Baseline climate variant, the forest fire risk more than doubles in the long-term average for the period from 2011 to 2060. Very dry years with total days with extreme fire hazards may become more common; such totals were never reached in the past. An increase in forest fire risk is expected especially for the summer months. The comparison to the situation today indicates a tendency for elevated forest fire risks associated with climate both in regions that are already quite dry today and in moist areas.
References
- Alexander ME, Stocks BJ, Lawson BD (1996) The Canadian forest fire danger rating system. Initial Attack 1996(Spring):6–9Google Scholar
- Badeck F-W, Lasch P, Hauf Y, Rock J, Suckow F, Thornicke K (2003) Steigendes klimatisches Waldbrandrisiko. AFZ/Der Wald 59(2):2–4Google Scholar
- Baumgartner A, Klemmer L, Raschke E, Waldmann G (1967) Waldbrände in Bayern 1950–1959. Mitteilungen der Staatsverwaltung Bayern 36, MünchenGoogle Scholar
- Grouven U, Bender R, Ziegler A, Lange S (2007) Der Kappa-Koeffizient – Artikel Nr. 23 der Statistik-Serie in der DMW. In: Deutsche Medizinische Wochenschrift 2007/132, pp 65–68Google Scholar
- IPCC – Intergovernmental panel on climate change, WMO/UNEP (ed) (2007) Klimaänderung 2007: Zusammenfassungen für politische Entscheidungsträger. Vierter Sachstandsbericht des IPCC (AR4), Bern/Wien/BerlinGoogle Scholar
- König HC (2007): Waldbrandschutz. Kompendium für Forst und Feuerwehr. Edition Gefahrenabwehr; Supplement (Band 1), BerlinGoogle Scholar
- Langholz H, Schmidtmayer E (1993) Meteorologische Verfahren zur Abschätzung des Waldbrandrisikos. Allgemeine Forstzeitschrift 48(8):94–396Google Scholar
- Lawson BD, Armitage OB (2008) Weather Guide for the Canadian Forest Fire Danger Rating System. Natural Resources Canada, Canadian Forest Service, Northern Forestry Centre, Edmonton, AlbertaGoogle Scholar
- Penman HL (1948) Natural evaporation from open water, bare soil and grass. Proc Roy Soc Lond A 194:120–145CrossRefGoogle Scholar
- Wiese A (2001) Waldbrandversicherung als Ergänzung der Waldbrandprävention. In: AFZ/Der Wald, 11/2001, pp 568–569Google Scholar