Skip booklet index and go to page content

Technical Thematic Report No. 6. - Trends in large fires in Canada, 1959 to 2007

Ecozone Trends

 


Boreal Shield

Fires within the Boreal Shield Ecozone+ are characterized as intense crown fires (Amiro et al., 2004; Beverly and Martell, 2005) that are moderate in severity compared to the other ecozones+ (Burton et al., 2008). These types of fires are important to many boreal tree species, such as like Jack pine (Picea banksiana) and black spruce (Picea mariana (Mill.) B.S.P.) that require fire to initiate seedling germination. These and other boreal species require periodic, intense stand replacing fires, which have historically resulted in large areas burned (Bergeron et al., 2001; de Groot et al., 2003). Even within an area where most fires are actively suppressed, the Boreal Shield is consistently the largest contributor to the total area burned in Canada. The long-term annual area burned for this ecozone+ is 6,467 km², which is 37% of the area burned in Canada (see Table 1 on page 5 and Figure 3 on page 7). Similar to trends at the national scale, there is large variability in area burned from year to year (Figure 5). The most extreme year was in 1989 when almost 28,000 km² burned, compared to low years when area burned was less than 1,000 km² (for example, 1982).

Figure 5. Annual area burned by large fires for the Boreal Shield, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 5

This bar graph shows the annual area burned by large fires for the Boreal Shield from 1959 to 2007. There is large variability in area burned from year to year. The most extreme year was in 1989 when almost 28,000 km2 burned, compared to low years when area burned was less than 1,000 km2 (for example, 1982).

Top of Page

Fire frequency in the Boreal Shield Ecozone+, represented by the percent annual forested area burned, is not the largest in the country. At 0.49% of the forested area of the ecozone+, the Boreal Shield Ecozone+ burns less frequently than the Taiga Plains (0.71%) and Taiga Shield (0.77%) (See Table 1 on page 5). This lower percent area burned may be related to higher levels of fire protection in the Boreal Shield, especially in the southern part of the ecozone+ or alternatively it may be related to the regional variability in fire weather conditions (Girardin and Wotton, 2009). Approximately 64% of the Boreal Shield is protected through suppression efforts compared to 20% of the Taiga Plains and 2% of the Taiga Shield (Parisien et al., 2006).

The Boreal Shield, the largest of the ecozones+, has two distinct fire regimes. For this reason, most fire analyses divide it into east and west components (Stocks et al., 2003; Beverly and Martell, 2005; Flannigan et al., 2005; Kasischke and Turetsky, 2006; Parisien et al., 2006; Burton et al., 2008; Flannigan et al., 2009). The line of division is just east of Lake Nipigon in northwestern Ontario. Stocks et al. (2003) calculated the percent annual area burned for the Boreal Shield west and Boreal Shield east separately using the large fire database data and found that the Boreal Shield west is within the same range for percent annual forested area burned as the Taiga ecozones+ (0.60 to 0.80%), compared to the Boreal Shield east, which is significantly lower (0.00 to 0.20%). Stocks et al. (2003) analysis included the Newfoundland Boreal in Boreal Shield east.

The difference in fire regime across the Boreal Shield is attributed to more severe fire weather, such as longer dry spells and lack of precipitation, in the Boreal Shield west (Beverly and Martell, 2005; Girardin and Wotton, 2009), as well as a higher ratio of conifer to deciduous species in the Boreal Shield east (Parisien et al., 2006; Amiro et al., 2009). Another contributing factor is that a greater percent area of the Boreal Shield east is protected by suppression efforts compared to the Boreal Shield west (70 and 54% respectively) due to greater human population in eastern Ontario and Western Quebec. The end result is a lower average annual area burned in the Boreal Shield east (772 km²) compared to the Boreal Shield west (5,147 km²) (Parisien et al., 2006). This decreased fire frequency in the Boreal Shield east has resulted in longer fire cycles that have led to an increased proportion of uneven-aged old growth forests, characterized by more shade tolerant, less fire prone species, such as balsam fir (Abies balsamea (L.) Mill), compared with the Boreal Shield west (Bergeron et al., 2001; de Groot et al., 2003).

Using data from the large fire database, the average duration of the active fire season lasted 143 days from 1959 to 1999, with fires most commonly occurring between May and August (Figure 6). Fires can occur as early as March and into November but these occurrences are rare. The peak month of fire occurrence is June. There was little change in the seasonality of fire occurrence in this ecozone+ from the 1960s to 1990s (Figure 6).

Figure 6. Proportion of large fires that occur each month in the Boreal Shield and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 6

This bar graph shows the proportion of large fires that occur each month in the Boreal Shield and an overlain line shows the average duration of the active fire season from the 1960s to 1990s. The average duration of the active fire season lasted 143 days from 1959 to 1999, with fires most commonly occurring between May and August. Fires can occur as early as March and into November but these occurrences are rare. The peak month of fire occurrence is June. There was little change in the seasonality of fire occurrence in this ecozone+ from the 1960s to 1990s.

Top of Page

Long term trends in area burned in the Boreal Shield suggest significant increase from the 1960s to the 1980s (Figure 7). The reason for this increase is similar to that given for Canada overall: partly due to warmer temperatures (Podur et al., 2002; Gillett et al., 2004), and partly due to changes in detection methods and completeness of fire monitoring coverage, especially in areas outside the protection zones, such as northern Quebec and Ontario (Podur et al., 2002; Stocks et al., 2003). Increased human activity, which is one of the reasons attributed to the increase at the national level, is unlikely to be part of the cause here because the proportion of total area burned by human ignitions decreased between the 1960s and 1990s (Table 3).

Figure 7. Total area burned by large fires per decade for the Boreal Shield, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 7

This bar graph shows the total area burned by large fires per decade for the Boreal Shield from the 1960s to the 2000s. Areas are approximately 29,000 km2 in the 1960s, 48,000 km2 in the 1970s, 104,000 km2 in the 1980s, 83,000 km2 in the 1990's, and 66,000 km2 in the 2000's.

Top of Page

Table 3. Proportion of the number of large fires ignited by lightning compared to humans and the total area burned by each ignition source by decade for the Boreal Shield, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total4,01677Table 3, Footnote*23Table 3, Footnote*211,942.648,144.2
1960s673623819,231.99,701.9
1970s828712937,198.210,581.5
1980s1,356851587,392.216,559.5
1990s1,159891168,120.311,301.2

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 3, footnote * referrer

Top of Page

Following the peak in area burned in the 1980s, a decline in area burned is evident for the remainder of the record (Figure 7). This decline is not as dramatic as the increase from the 1960s to 1980s, which is similar at the national level (see Figure 4 on page 8). What is different from the national trend is the timing of the decline; it commences in the 1990s in the Boreal Shield compared to the 2000s at the national level. A similar earlier decline is shown in the Boreal Plains, but not in the Taiga Shield or Taiga Plains (see respective ecozone+ sections), which show similar trends to the national data. One of the primary differences between these ecozones+ is that larger proportions of the Boreal Shield and Boreal Plains are protected by suppression efforts compared to the Taiga Plains and Taiga Shield. Also, the Boreal Plains and Boreal Shield influence a larger proportion of the ecozone+ area as expressed by percent land use statistics in Parisien et al. (2006). The decline in area burned from the 1980s to the 1990s may be a result of successful forest suppression efforts (Cumming, 2005; Martell and Sun, 2008). The continued decline into the 2000s may be due to a combination of changing climatic oscillations that have been show to have an influence on fire weather in the Boreal Shield (Flannigan and Wotton, 2001; Skinner et al., 2002; Girardin et al., 2006a; Girardin et al., 2006b) and the increased management of human impacts.

As mentioned above, the number of human caused large fires declined over the period of record, decreasing from 38% in the 1960s to 11% in the 1990s (R² = 0.95, p = 0.019) (Table 3). It is important to note that these data do not include small fires that may have been ignited by humans but were suppressed before they were large enough to be recorded. This reduction is consistent with findings by Kasischke and Turetsky (2006) who state that government policies implemented in the 1970s and 1980s nation wide resulted in a 50% reduction in all human ignitions. Human ignitions of both small and large fires are still the dominant cause of fires in Canada; nevertheless, they result in a much smaller area burned compared to lightning ignitions (Weber and Flannigan, 1997; Wotton et al., 2003). Lightning fires were the cause of 80% of the large forest fires and made up the majority of the area burned in the Boreal Shield Ecozone+ from the 1960s to 1990s (Table 3).

Top of Page

Newfoundland Boreal

Like other coastal ecozones+, fire is not a significant natural disturbance in the Newfoundland Boreal. Average area burned is higher than the Pacific Maritime and Atlantic Maritime ecozones+ at 123 km² per year, and the percent annual area burned is also higher at 0.13%, but the contribution to total area burned in Canada is still less than 1% (see Table 1 on page 5). Its contribution to total area burned in Canada has changed over time (see Figure 3 on page 7), making a greater contribution in the 1960s at 4.7%. This was due to an extreme fire year that occurred in 1961 during which 3,962 km² burned. Apart from this, there is little variability in annual area burned and more commonly there are many years throughout the period of record were there are no large fires in this ecozone+ (Figure 8).

Figure 8. Annual area burned by large fires for the Newfoundland Boreal, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 8

This bar graph shows the annual area burned by large fires for the Newfoundland Boreal from 1959 to 2007. There is little variability in annual area burned, with the exception of an extreme fire year in 1961 where almost 4,000 km2 burned. More commonly, there are many years throughout the period without large fires.

Top of Page

The average active fire season duration, 35 days, is approximately the same as it is for the Atlantic Maritime Ecozone+ (see Table 2 on page 9). Fire occurrence peaks in May but fires commonly occur between May and July. Also, similar to the Atlantic Maritime and Pacific Maritime, the dominant cause of fire ignitions is humans at 96%. Lightning ignitions are rare occurrences that have only been documented four times in the large fire database for the Newfoundland Boreal.

Area burned decreased dramatically since the 1960s (Figure 9). The decline was most likely due to successful government policies aimed at preventing and suppressing fires (Stocks et al., 2003) and to the one extreme year, 1961. The doubling in area burned from the 1970s to the 1980s may be related to warmer temperatures (Podur et al., 2002; Gillett et al., 2004) that resulted in more fire escapes from suppression efforts. Area burned declined again in the 1990s and has remained small into the 2000s, which only had 12 large fires between 2000 and 2007. Similar to the Atlantic Maritime and Pacific Maritime ecozones+, these trends should be assessed with caution because they are based on a small number of fires, especially in more recent decades.

Figure 9. Total area burned by large fires per decade for the Newfoundland Boreal, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 9

This bar graph shows the total area burned by large fires per decade for the Newfoundland Boreal from the 1960s to the 2000s. Areas are approximately, 4,100 km2 in the 1960s, 550 km2 in the 1970s, 1,200 km2 in the 1980s, and below 50 km2 in the 1990s and the 2000s.

Top of Page

Boreal Plains

Similar to other areas of the boreal forest in Canada, fire is an important natural disturbance in the Boreal Plains Ecozone+. On average 2,213 km² burns in this ecozone+ each year, but this varied from less than 200 km² in some years, to over 6,000 km² in others (Figure 10). The most extreme fire year on record was 1981 when 18,760 km² burned. At the national level, the area burned in the Boreal Plains is a significant contribution to the total area burned in Canada. At 11%, it makes the fourth largest contribution to area burned in the country (see Table 1 on page 5). There has been some variability in this statistic over time, from a high of 19% in the 1960s to a low of 5% in the 1970s (see Figure 3 on page 7). Of the available forest fuels, the proportion that burn annually (0.47%) is similar to the nearby Boreal Shield and Taiga Cordillera ecozones+ (0.49 and 0.47% respectively) (see Table 1 on page 5) despite higher levels of fire protection. Approximately 90% of the Boreal Plains is protected by fire suppression activities, compared to 0% for the Taiga Cordillera and 62% for the Boreal Shield (Parisien et al., 2006). Although the proportion of their forests that burn each year are similar, the Boreal Plains has a much higher total annual area burned than the Taiga Cordillera, and is much lower than the Boreal Shield (see Table 1 on page 5).

Figure 10. Annual area burned by large fires for the Boreal Plains, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 10

This bar graph shows the annual area burned by large fires for the Boreal Plains from 1959 to 2007. The area burned varied annually from less that 200 km2 in some years to over 6,000 km2 in others. The most extreme fire year was in 1981 when over 18,000 km2 burned.

Top of Page

The Boreal Plains has the highest fire protection levels of all the boreal forest ecozones+ (Parisien et al., 2006). Protection levels are elevated to compensate for extensive human interaction within the forest and high values- at-risk in surrounding areas (Stocks et al., 2003). As a result, the fire regime in the Boreal Plains is highly influenced by humans. In addition to suppression efforts, humans also have a significant impact on the fire regime through ignitions. Human caused fires represented 57% of the total number of large fires that occurred in this ecozone+ between the 1960s and 1990s (Table 4). Over this period, the proportion of human caused fires decreased significantly (R² = 0.98, p = 0.004) such that, by the 1990s, lightning was the dominant cause. Overall, lightning fires resulted in a larger area burned; 42% of fires were caused by lightning but they were responsible for 69% of the area burned (Table 4). The reason for the discrepancy is because most human caused fires are commonly closer to human settlement resulting in earlier detection and response. Therefore, the success rate of fire suppression techniques is often higher on these fires, reducing their growth to larger areas compared to lightning fires that can often occur in more remote locations (Stocks et al., 2003).

Table 4. Proportion of the number of large fires ignited by lightning compared to humans and the total area burned by each ignition source by decade for the Boreal Plains, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total1,61343Table 4, Footnote*57Table 4, Footnote*69,733.330,859.5
1960s44426744,334.312,560.9
1970s32140603,846.53,310.7
1980s468475335,177.311,773.4
1990s380584226,375.23,214.4

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 4, footnote * referrer

Top of Page

The Boreal Plains also has the longest active fire season of all the ecozones+ (see Table 2 on page 9). The average duration of the active fire season was 158 days (approximately 5 months) from 1959 to 1999. Fires most commonly occur April to August, but they can extend into the month of December (Figure 11). The peak month of fire occurrence is May (Figure 11). The dominance of spring fires translates into less severe fires in this area compared to others (Burton et al., 2008). Based on fire weather indices the Boreal Plains does experience quite severe fire weather, but this does not translate into severe fires (Amiro et al., 2004; Parisien et al., 2006; Burton et al., 2008). The occurrence of both human and lightning caused fires is the reason behind the long fire season. Human caused fires are more prominent earlier in the season, while lightning caused fires occur more commonly later in the season. There were only minor changes in the duration of the active fire season, or shifts in the seasonality of fire occurrence based on the 40 years of available data.

Figure 11. Proportion of large fires that occur each month in the Boreal Plains and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 11

This bar graph shows the proportion of large fires that occur each month in the Boreal Plains and an overlain line shows the average duration of the active fire season from the 1960s to 1990s. The average duration of the active fire season was 158 days (approximately 5 months) from 1959 to 1999. Fires most commonly occur April to August, but they can extend into the month of December. The peak month of fire occurrence is May.

Top of Page

There are trends in the area burned over the last 50 years, many of which are again tied to human interactions within this ecozone+. Over the past five decades the area burned peaked in the 1980s (Figure 12). Prior to this there was a decrease from the 1960s into the 1970s. A similar decline was also evident in the Boreal and Montane Cordilleras, Newfoundland Boreal, and Atlantic Maritime ecozones+ (see respective ecozone+ sections), all of which have significant human impacts on forest fire, either through suppression and/or ignitions. Therefore, the cause of the decline may be attributed to a change in human behaviour, either through advanced fire fighting techniques or increased prevention efforts. Monitoring and detection methods were also questionable during this early period, which could be another factor in the decline (Podur et al., 2002; Stocks et al., 2003). For the remainder of the record the pattern is similar to the nearby Boreal Shield Ecozone+. The peak in the 1980s has been attributed to a combination of warmer climate (Podur et al., 2002; Gillett et al., 2004), increased forest use by humans, and better detection methods (Stocks et al., 2003). Since the 1980s, area burned has declined in the Boreal Plains, similar to the pattern shown in the Boreal Shield. As is stated in the Boreal Shield section, the reason for the decline may be in part due to increased management of human impacts and changes in fire weather (Flannigan and Wotton, 2001; Skinner et al., 2002; Cumming, 2005; Girardin et al., 2006b).

Figure 12. Total area burned by large fires per decade for the Boreal Plains, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 12

This bar graph shows the total area burned by large fires per decade for the Boreal Plains from the 1960s to the 2000s. Areas are approximately 17,000 km2 in the 1960s, 7,000 km2 in the 1970s, 47,000 km2 in the 1980s, 24,000 km2 in the 1990s, and 16,000 km2 in the 2000s.

Top of Page

Taiga Plains

Some of the largest fires in the country occur in the Taiga Plains Ecozone+ (Parisien et al., 2006; Burton et al., 2008). The reason behind this is a cumulation of factors including: a dry continental climate (Stocks et al., 2003), remote location with little suppression efforts (approximately 20% of the area undergoes suppression) (Parisien et al., 2006), and a dominance of boreal fuel types at relatively high, average fuel loads that lead to higher consumption rates (Amiro et al., 2001; Amiro et al., 2009). These factors result in relatively severe fires that burn over large areas. Burton et al. (2008) found the Taiga Plains to have some of the most severe fires in the country.

On average 2,858 km² burn each year in this ecozone+ with large variability from year to year (Figure 13). There are many years where annual area burned is less than 100 km², while others have been as high as 17,354 km² (1995). Some low years early in the record may be due to limited monitoring in this northern ecozone+, however this trend continues in more recent decades validating the occurrence of very low fire years (for example, 1991, 1997, and 2002). Despite the occurrence of low fire years, the percent annual area burned is high at 0.71% of the forested ecozone+ area compared to other ecozones+ (see Table 1 on page 5). This is the second highest – only the Taiga Shield is higher at 0.77%. On a national level, the Taiga Plains constitutes 14% of the area burned in Canada, which is the third largest contribution following the Boreal Shield and Taiga Shield. Over time this contribution has varied from a low of 7.5% in 2000s to a high of 21.5% in the 1970s (see Figure 3 on page 7). There is no trend to this variation. The low number in the 2000s should be looked at with caution as it does not include data for the full decade.

Figure 13. Annual area burned by large fires for the Taiga Plains, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 13

This bar graph shows the annual area burned by large fires for the Taiga Plains from 1959 to 2007. There is large variability from year to year, with many years where the annual area burned is less than 100 km2 while others have had more than 17,000 km2.

Top of Page

The long term trend in area burned is similar to trends at the national level (Figure 14). Area burned increased from the 1960s until the 1990s and then fell sharply in the 2000s. Since the Taiga Plains Ecozone+ is located in northern Canada, the low numbers at the beginning of the record may be attributed to data collection techniques that improved starting in the 1970s (Stocks et al., 2003). Although the numbers presented for 2000s should be considered with caution because they do not include the full decade, the reason behind the recent decline requires further research to fully elucidate, but may be related to changes in large atmospheric oscillations as discussed in the National Trends section on page 5.

Figure 14. Total area burned by large fires per decade for the Taiga Plains, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 14

This bar graph shows the total area burned by large fires per decade for the Taiga Plains from the 1960s to the 2000s. Areas are approximately 10,000 km2 in the 1960s, 30,000 km2 in the 1970s, 40,000 km2 in the 1980s, 48,000 km2 in the 1990s, and 13,000 km2 in the 2000s.

Top of Page

The average duration of active large fire season from the 1960s through 1990s, was 81 days (approximately 4 months) and did not change much over this time period. Fires most commonly occur June through to August, but can occur as early as April and as late as September (Figure 15). Although the average duration of the period of fire occurrence did not change, the distribution of fires within the fire season made some subtle changes over the 40 year period. The proportion of fires that occurred in April shifted from zero in the 1960s to 1.2% in the 1990s and the proportion of fires that occurred in May steadily increased (R² = 0.93, p = 0.035) (Figure 15). All fires that occurred in April were human caused; those in May were equally distributed between humans or lightning. More data are needed to determine if these small changes are the start of a real phenomenon, such as lengthening of the fire season, or are artefacts of the large fire database limitations.

Figure 15. Proportion of large fires that occur each month in the Taiga Plains and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 15

This bar graph shows the proportion of large fires that occur each month in the Taiga Plains and an overlain line shows the average duration of the active fire season from the 1960s to 1990s. The average duration of active large fire season from the 1960s through 1990s, was 81 days (approximately 4 months) and did not change much over this time period. Fires most commonly occur June through to August, but can occur as early as April and as late as September.

Top of Page

Fires that occur later in the season do not appear to be changing in their distribution (Figure 15). These fires are predominately caused by lightning – an average of 83% of the total number of fires that occurred in the Taiga Plains from 1960s and 1990s were ignited by lightning (Table 5). Table 5 shows that the proportion of the number of fires caused by lightning increased compared to ignitions by humans. There is no clear trend in the lightning ignitions which differs from findings by Kasischke and Turetsky (2006) for the North American forest and Stocks et al.(2003) for Canada who found increases in the recent years. The total area burned as a result of lightning ignitions increased over the 40 years (Table 5). This increase in area burned by lightning is most likely due to warmer temperatures during the fire season in the 1990s (Podur et al., 2002; Gillett et al., 2004).

Table 5. Proportion of the number of large fires ignited by lightning compared to humans and the total area burned by each ignition source by decade for the Taiga Plains, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total99183*17*125,531.410,687.6
1960s14958423,455.66,987.9
1970s262873527,463.52,199.3
1980s326961339,920.4217.8
1990s254912254,691.91,282.6

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 5, footnote * referrer

Top of Page

Taiga Shield

The fire regime in the Taiga Shield Ecozone+ is characterized by large, severe fires (Stocks et al., 2003; Parisien et al., 2006; Burton et al., 2008), similar to the fire regime in the Taiga Plains. The annual area burned in the Taiga Shield is 3,789 km², which is higher than the nearby Taiga Plains (2,858 km²), but second to the Boreal Shield (6,467 km²). The large annual area burned results in a contribution of 18% to the Canadian total (see Table 1 on page 5). The low contribution to annual area burned in the 1960s is most likely due to limited detection methods during this time. At the ecozone+ scale, 0.77% of the available forested area burns on average each year. This is the largest percent annual area burned in Canada (see Table 1 on page 5).

Like the Boreal Shield, the Taiga Shield is commonly broken into two separate areas for analysis in the fire literature (Amiro et al., 2001; Stocks et al., 2003; Parisien et al., 2006; Burton et al., 2008; Amiro et al., 2009). The divide is on either side of the Hudson Plains Ecozone+ separating the area into eastern and western portions based on different fire regimes. The reason for the difference in regimes is significantly different climates. This occurs despite other similarities between the two sides including: fuel types, both sides are predominately conifer fuels (Parisien et al., 2006; Amiro et al., 2009); ignitions, both are dominated by lightning; and human influence, both have little to no suppression (Parisien et al., 2006). The Taiga Shield west has more severe fire weather due to its continental location, compared to the Taiga Shield east, which is cooler and wetter because of its coastal location (Parisien et al., 2006; Burton et al., 2008). As a result, the fire regime in the Taiga Shield west is represented by larger, more severe and frequent fires than in the Taiga Shield east (Stocks et al., 2003; Amiro et al., 2004; Parisien et al., 2006; Burton et al., 2008). Burton et al. (2008) calculated the percent annual area burned to be 0.83 in the Taiga Shield west and 0.25 in the Taiga Shield east (based on data from 1959 to 1999), using an average annual area burned of 2,632 and 1,126 km² respectively.

The area burned varies greatly from year to year in the Taiga Shield, with some years early in the record having no fires greater than 2 km² compared to extremes in 1989 (28,120 km²) and 1994 (31,279 km²) (Figure 16). The extremely low numbers in the 1960s do not occur again later in the record, thus one may assume these are due to detection problems (Podur et al., 2002; Stocks et al., 2003). Over the long-term the area burned increased from the 1960s until the 1990s (Figure 17). The same pattern is shown in the Taiga Plains. The increase has been attributed to a combination of changes to detection methods and warmer temperatures during this time frame (Podur et al., 2002; Gillett et al., 2004). The decline since the 1990s is also similar to the pattern shown for the Taiga Plains. The magnitude of the decline is of the same magnitude as changes between previous decades; therefore it may signify a real response despite being based on only seven years of data pro-rated over the decade. The reason for the decline may be related to changes in large atmospheric oscillations as discussed in the National Trends section on page 5.

Figure 16. Annual area burned by large fires for the Taiga Shield, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 16

This bar graph shows the annual area burned by large fires for the Taiga Shield from 1959 to 2007. The area burned varies greatly from year to year in the Taiga Shield, with some years early in the record having no fires greater than 2 km2 compared to extremes in 1989 (28,120 km2) and 1994 (31,279 km2).

Top of Page

Figure 17. Total area burned by large fires per decade for the Taiga Shield, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 17

This bar graph shows total area burned by large fires per decade for the Taiga Shield from the 1960s to the 2000s. Areas are approximately 1,300 km2 in the 1960s, 1,700 km2 in the 1970s, 6,500 km2 in the 1980s, 10,500 km2 in the 1990s, and 8,400 km2 in the 2000s.

Top of Page

The average duration of the active fire season is relatively short compared to the other boreal ecozones+ (see Table 2 on page 9). The long-term average between 1959 and 1999 was 75 days, which did not change much since the 1970s. Fires most commonly occur between June and August, peaking in July (Figure 18). There was a significant increase in the occurrence of fires in the month of May (R² = 0.99, p = 0.002), from no fires in the 1960s to 2.4% of the fires in 1990s (Figure 18). Fires in the Taiga Shield east generally occur earlier than fires in the Taiga Shield west (Kasischke and Turetsky, 2006), therefore the increase in the number of May fires may be from this side of the ecozone+. There was also an increase in fires late in the fire season (August).

Figure 18. Proportion of large fires that occur each month in the Taiga Shield and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 18

This bar graph shows the proportion of large fires that occur each month in the Taiga Shield and an overlain line shows the average duration of the active fire season from the 1960s to 1990s. The average duration of the active fire season has been 75 days, and relatively stable, for the past three decades. Fires most commonly occur between June and August, peaking in July. There was a significant increase in the occurrence of fires in the month of May.

Top of Page

Similar to the other northern boreal ecozones+, fires are predominately caused by lightning ignitions. On average 92% of the large fires that occur in the Taiga Shield are caused by lightning (Table 6). The ratio of human to lightning caused fires is consistent from one side of the ecozone+ to the other (Parisien et al., 2006). The proportion of fires caused by lightning (due to a decrease in human caused fires) and the number of fires caused by lightning increased from the 1960s to 1990s but this increase is not statistically significant. Kasischke and Turetsky (2006) also found an increase in lightning ignitions over time. The average area burned by lightning ignitions increased between the 1960s and 1990s, but again the increase was not statistically significant.

Table 6. Proportion of the number of large fires ignited by lightning compared to humans and the total area burned by each ignition source by decade for the Taiga Shield, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total1,71192*8*150,553.66,842.7
1960s15188124,739.21,263.6
1970s468871332,711.13,627.2
1980s50995544,744.61,830.1
1990s58397368,358.6121.8

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 6, footnote * referrer

Top of Page

Hudson Plains

The Hudson Plains Ecozone+ has one of the world's largest network of wetlands that extend across the flat landscape (Amiro et al., 2001; Parisien et al., 2006). As a result of this extensive network, fuels are discontinuous, which leads to smaller fires in the Hudson Plains than the surrounding Boreal Shield and Taiga Shield. The average annual area burned of 547 km² in the Hudson Plains contributes only 3% to the total national area burned, compared to 18% from the Taiga Shield and 37% from the Boreal Shield (see Table 1 on page 5). Over time, the percent contribution to the total area burned in Canada increased form 1.5% in the 1960s to 4.8% in the 1990s (see Figure 3 on page 7). Comparatively, the area burned in this ecozone+ is more significant at the ecozone+ scale. The percent annual area burned for the Hudson Plains is 0.17%, which is much smaller than the Boreal Plains which is 0.47%. The reason for the difference my be related to the extensive network of wetlands which have been shown to be more resistant to fires (Burton et al., 2008). Burton et al. (2008) also showed that fire severity in the Hudson Plains was moderate compared to other ecozones, and that individual fires were more homogenous in severity, which is probably related to the extensive and relatively homogenous wetlands common to this region.

From one year to the next there is large variability in the area burned in the Hudson Plains (Figure 19). Some years, particularly early in the fire record do not show any fires that burned over 2 km². Low years are also present in more recent times however, for example 1993 and 2004, indicating these lows are common and not just a problem with limited monitoring early in the record. Years of extreme fire activity include 1989 (4,572 km²) and 2003 (3,455 km²). Trends in decadal data indicate a significant increase in area burned after the 1970s (Figure 20). Based on the limited area burned in the 1960s and 1970s and the remote, northern location it is likely that the frequency of low area burned values may be to limitations in detection methods. Fires in the Hudson Plains go largely unsuppressed (Ontario Ministry of Natural Resources, 2004; Parisien et al., 2006; Burton et al., 2008) and other authors have documented that detection was non-existent or limited in these non-suppressed areas at the beginning of the large fire database records especially in Quebec and Ontario (Stocks et al., 2003). The increase in area burned from the 1980s to the 1990s, may be climate induced (Podur et al., 2002; Gillett et al., 2004; Girardin, 2007) similar to trends at the national level. Finally, the data show a slight decline in the last decade, although not as marked as the changes between other decades. Therefore, it is difficult to determine if this is real or a result of pro-rating the data for the 2000s based on the first seven years only.

Figure 19. Annual area burned by large fires for the Hudson Plains, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 19

This bar graph shows the annual area burned by large fires for the Hudson Plains from 1959 to 2007. From one year to the next there is large variability in the area burned in the Hudson Plains. Prior to 1983, there were no recorded years where the area burned exceeded 2000 km2. Although some years have little fire activity, such as 1993 (8 km2) and 2004 (73 km2), there have also been years of extreme fire activity including 1989 (4,572 km2) and 2003 (3,455 km2).

Top of Page

Figure 20. Total area burned by large fires per decade for the Hudson Plains, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 20

This bar graph shows the total area burned by large fires per decade for the Hudson Plains from the 1960s to the 2000s. Areas are approximately 1,300 km2 in the 1960s, 1,700 km2 in the 1970s, 6,500 km2 in the 1980s, 10,500 km2 in the 1990s, and 8,400 km2 in the 2000s.

Top of Page

Similar to other remote northern ecozones+, the Hudson Plains is dominated by lightning caused fires. Lightning ignitions on average account for 89% of the large fires that occur (Table 7). This proportion varies slightly depending on the decade but lightning is consistently the dominant cause of fires. The proportion of fires caused by lightning does not appear to be changing, but the number of large fires caused by lightning has increased. In the 1980s, only 80 fires where caused by lightning, this almost doubled to 143 in the 1990s. More records are required to determine if this is a significant upward trend in lightning occurrence.

Table 7. Proportion of the number of large fires ignited by lightning compared to humans by decade for the Hudson Plains, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused ignitions (%)Proportion of human caused ignitions (%)
Total29889*11*
1960s12928
1970s468020
1980s908911
1990s150955

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 7, footnote * referrer

Top of Page

Corresponding to the cause of fires, the occurrence of fires is predominately later in the season. Most fires in the Hudson Plains occur between May and August, peaking in July (Figure 21). The duration of the active fire season is on average 40 days, which is shorter than the nearby Taiga Shield and Boreal Shield (see Table 2 on page 9). The duration of the active fire season increased from 9 days in the 1960s to 59 days in 1990s. This increase is significant (R² = 0.97, p = 0.026), but the results should be looked at with caution due to the high probability that there are missing data in the 1960s and 1970s (Stocks et al., 2003). Lastly, the seasonality of fire occurrence appears to be shifting to later in the season (Figure 21) with an increased proportion of fires in August, although this pattern was not statically significant. These changes are consistent with changes noted by Kasischke and Turetsky (2006) and predictions made by Wotton and Flannigan (1993) that warmer temperatures, due to climate change, would result in a lengthening of the fire season later into the year in eastern Canada.

Figure 21. Proportion of large fires that occur each month in the Hudson Plains and the average duration of the activefire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 21

This bar graph shows the annual area burned by large fires for the Taiga Cordillera from 1959 to 2007. The area burned varies annually; there are years without large fires (1970 and 1997) as well as extreme fire years (2004).

Top of Page

Taiga Cordillera

Despite the similarities in fuel type to the other taiga ecozones+, Parisien et al. (2006) showed that fires in the mountainous regions of the Taiga Cordillera are more similar to the Montane and Boreal Cordillera than the taiga or boreal ecozones+. The continental location of the cordillera areas results in drier burning conditions (Burton et al., 2008) and the mountainous terrain creates fuel discontinuity limiting fire spread (Parisien et al., 2006). The result is smaller, more severe fires that occur less often (Amiro et al., 2001; Stocks et al., 2003; Burton et al., 2008).

On average, 857 km² burn in the Taiga Cordillera Ecozone+ each year, which translates into a 4.5% contribution to the total area burned in Canada (see Table 1 on page 5). This proportion has fluctuated over time, from a low of 2.2% in 1980s to a high of 7.4% in the 2000s (see Figure 3 on page 7). On an annual basis, as with all the other ecozones+, there is much variability in area burned from year to year (Figure 22). During extreme fires years, for example 2004, 7,000 km² burned compared to years where no large fires occurred, such as 1970 and 1997. When area burned is calculated as a proportion of the total forested area, the Taiga Cordillera is larger than the national average at 0.47% (see Table 1 on page 5). It is also much higher than the Montane Cordillera (0.10%) and Boreal Cordillera (0.38%). The higher number may be related to differences in suppression efforts. Fires that burn in the Taiga Cordillera reflect the natural fire regime because there are no areas of intense fire suppression; compared to 100% suppression in the Montane Cordillera and 41% in the Boreal Cordillera (Parisien et al., 2006).

Figure 22. Annual area burned by large fires for the Taiga Cordillera, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 22

This bar graph shows the annual area burned by large fires for the Taiga Cordillera from 1959 to 2007. On an annual basis there is much variability in area burned from year to year. There are years where no large fires occurred (1970 and 1997) as well as extreme fire years, for example 2004 7,000 km2 burned.

Top of Page

At the decadal scale there is no consistent pattern in area burned for the Taiga Cordillera Ecozone+ (Figure 23). The increase from the 1960s to 1970s is most likely due to limited monitoring in this northern ecozone+, which is a problem throughout the country (Podur et al., 2002; Stocks et al., 2003). The small decline in the 1980s is not seen in any of the other limited or non-suppressed ecozones+ (see Hudson Plains or Taiga Shield sections), which all showed increases in area burned attributed to warmer climate conditions (Podur et al., 2002; Gillett et al., 2004). Area burned doubles between the 1980s and 1990s and remains high, but drops slightly into the 2000s. The increase into the 1990s is consistent with the Montane and Boreal cordilleras, but the decline in the 2000s is not. A decline in area burned was shown, however, in the taiga and boreal ecozones+. Hence, the Taiga Cordillera may be responding to the same climate patterns as the other boreal areas, or the decline may be an artefact of using only eight years of data to represent 2000s.

Figure 23. Total area burned by large fires per decade for the Taiga Cordillera, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 23

This bar graph shows the total area burned by large fires per decade for the Taiga Cordillera from the 1960s to the 2000s. Areas are approximately 2,500 km2 in the 1960s, 6,900 km2 in the 1970s, 6,000 km2 in the 1980s, 15,000 km2 in the 1990s, and 13,000 km2 in the 2000s.

Top of Page

The fire season in the Taiga Cordillera is relatively short at 44 days on average (see Table 2 on page 9). From the 1960s until the 1980s, the duration the active fire season was consistently around 41 days, until the 1990s which saw an increase of 14 days (Figure 24). More data is required to determine if this increase is significant. Fires most commonly occur in June and July in this ecozone+. There has been a steady increase in the occurrence of fires in the month of August (R² = 0.91, p = 0.04). The number of fires that occurred during this month increased three fold from the 1960s to the 1990s (Figure 24).

Figure 24. Proportion of large fires that occur each month in the Taiga Cordillera and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 24

This bar graph shows the proportion of large fires that occur each month in the Taiga Cordillera and an overlain line shows the average duration of the active fire season from the 1960s to the 1990s. The duration of the active fire season was stable, around 41 days, until the 1990s which saw an increase to 65 days. More data is required to determine if this increase is significant. Fires most commonly occur in June and July in this ecozone+.

Top of Page

The short duration during which fires occur and their occurrence late in the fire season are due to the prevalence of lightning ignitions in the Taiga Cordillera. Ninety-five percent of fires are caused by lightning in this ecozone+ (Table 8). Human caused large fires are rare and appear to be becoming more rare overtime (Table 8), although this decline was not shown to be significant. The ratio of human to lightning caused fires in the Taiga Cordillera is the lowest in the country at 0.05 (see Table 2 on page 9).

Table 8. Proportion of the number of large fires ignited by lightning compared to humans by decade for the Taiga Cordillera, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total42895*5*31,654.71,206.9
1960s559372,480.870.4
1970s969556,636.5286.1
1980s1579555,697.5311.3
1990s13999116,839.9539.0

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 8, footnote * referrer

Top of Page

Boreal Cordillera

The fire regimes in the cordillera ecozones+ have been shown to be unique when compared to the boreal forest (Parisien et al., 2006). The continental location of the cordillera results in drier burning conditions (Burton et al., 2008) while the mountainous terrain creates fuel discontinuity limiting fire spread (Parisien et al., 2006). Of the three cordillera ecozones+, the Boreal Cordillera has the highest area burned and most severe fires (Stocks et al., 2003; Parisien et al., 2006; Burton et al., 2008). Average annual area burned is 1,206 km² (see Table 1 on page 5) which varies annually from less than 10 km² (for example, 1973 and 1987) to 11,014 km² (2004) during extreme years (Figure 25). The average relative contribution to total area burned in Canada is 8%, which is essentially double the other two cordillera ecozones+ (see Table 1 on page 5). At a local scale, the average percent of forested area that burns annually in the Boreal Cordillera is 0.38%, which is slightly higher than the Canadian average.

Figure 25. Annual area burned by large fires for the Boreal Cordillera, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 25

This bar graph shows the annual area burned by large fires for the Boreal Cordillera from 1959 to 2007. The area varies annually from less than 10 km2 (1973 and 1987) to over 11,000 km2 (2004).

Top of Page

From a decadal perspective, there has been little variation in area burned over the last five decades, with the exception of the 1970s (Figure 26). The decline in area burned in the 1970s can be attributed to changes in human behaviour, either through advanced fire fighting techniques or increased prevention efforts. This trend is also shown in other ecozones+ that have some degree of suppression efforts (such as, Montane Cordillera and Boreal Plains). According to Parisien et al. (2006), 41% of the Boreal Cordillera is protected by intensive suppression activities. The increased success of suppression efforts has been attributed to the introduction of water bombers in the 1970s (Bergeron et al., 2001). An alternative explanation for the decline in burns in the 1970s is poor detection methods that lead to questionable values in the 1960s and 1970s, especially in northern regions (Podur et al., 2002; Stocks et al., 2003). Total area burned rose by about 9,000 km² between the 1970s and 1980s, after which it remained stable into the 1990s and rose slightly in the 2000s (Figure 26). The large increase from the 1970s to 1980s is not seen in either the Montane or Taiga cordillera ecozones+. This trend was evident in some of the boreal ecozones+ and was attributed to a combination of warmer climate (Podur et al., 2002; Gillett et al., 2004), increased forest use by humans, and better detection methods (Stocks et al., 2003). The increase in area burned in the 2000s is evident in all of the western ecozones+. There are a number of factors that may be causing these increases (see Montane Cordillera section on page 34). Changes in climate oscillations are most likely the cause of the increase in the Boreal Cordillera (Volney and Hirsch, 2005).

Figure 26. Total area burned by large fires per decade for the Boreal Cordillera, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 26

This bar graph shows the total area burned by large fires per decade for the Boreal Cordillera from 1960s-2000s. Areas are approximately 12,000 km2 in the 1960s, 5,000 km2 in the 1970s, 14,000 km2 in the 1980s and the 1990s, and 17,000 km2 in the 2000s.

Top of Page

The duration of the active fire season is relatively short (37 days) in this ecozone+, compared to the nearby Taiga Cordillera (44 days), Taiga Plains (81 days) and Montane Cordillera (101 days) (see Table 2 on page 9). Fires most commonly occur between May and August, peaking in June or July (Figure 27). Fires have been recorded as early as April and as late as September, but these are rare occurrences. No significant changes occurred between the 1960s and 1990s in the duration or seasonality of fires in the Boreal Cordillera (Figure 27).

Figure 27. Proportion of large fires that occur each month in the Boreal Cordillera and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 27

This bar graph shows the proportion of large fires that occur each month in the Boreal Cordillera and an overlain line shows the average duration of the active fire season from the 1960s to 1990s. At 37 days, the duration of the active fire season is relatively short in this ecozone+. Fires most commonly occur between May and August, peaking in June or July. On rare instances fires were recorded as early as April and as late as September.

Top of Page

The majority, 78% on average, of large fires that occur in the Boreal Cordillera are caused by lightning ignitions (Table 9) and human influence is relatively low. Parisien et al. (2006) identified 19% of the area as affected by human land use. This is similar to the Boreal Shield west and Taiga Plains. There was no significant change in the proportion of lightning to human ignitions up to between 1959 and 1999 (Table 9). The area burned by human ignitions increased over time, but this trend was not statistically significant.

Table 9. Proportion of the number of large fires ignited by lightning compared to humans by decade for the Boreal Cordillera, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total73878*22*40,736.35864.5
1960s161762410,978.5611.6
1970s10084165,242.584.0
1980s213693110,581.13,630.6
1990s264841613,934.31,538.4

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 9, footnote * referrer

Top of Page

Montane Cordillera

The Montane Cordillera Ecozone+ experiences the most severe fire weather of all the ecozones+. Despite this strong positive climatic influence, the fire regime is characterized by many small, low intensity fires (Stocks et al., 2003; Parisien et al., 2006). Topography and humans have a greater influence on fires in the Montane Cordillera limiting fire spread. Area burned for this ecozone+ ranged from zero (1993) to about 2,500 km² (1961 and 2003) during extreme years Figure 28), with a long-term average of 316 km² per year. The average contribution to the total area burned in Canada is almost 3.0% (see Table 1 on page 5), which fluctuated from a high of 5% in the 1960s to a low of 0.6% in the 1980s and 1990s (see Figure 3 on page 7). The percent of forested area that burns is also small at 0.10% per year on average (see Table 1 on page 5).

Figure 28. Annual area burned by large fires for the Montane Cordillera, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 28

This bar graph shows the annual area burned by large fires for the Montane Cordillera from 1959 to 2007. Area burned for this ecozone+ ranged from zero (1993) to about 2,500 km2 (1961 and 2003) during extreme years, with a long-term average of 316 km2 per year.

Top of Page

The Montane Cordillera is the only ecozone+ that is completely encompassed by fire suppression activities (Parisien et al., 2006). These forest management efforts are evident in the long-term pattern of area burned (Figure 29), which is strongly influenced by humans. There is a sharp decline in area burned from the 1960s to the 1970s, most likely due to increased fire management and suppression efforts (Stocks et al., 2003). Area burned remained stable over the next three decades unlike any of the other ecozones+. The reason again is likely tied to successful suppression efforts due to high values at risk in this area (Stocks et al., 2003; Parisien et al., 2006). During the 2000s, there was a dramatic increase in area burned (Figure 29). Extreme fires years in 2003, 2004, and 2006 contributed to this increase (Figure 28). A number of factors may be responsible for this increase including: climate oscillations such as an El Niño event in 2002/2003 and a warm phase PDO from 2003 to 2005 (Volney and Hirsch, 2005), increased fuel loads in the forest due to long-term suppression of fires (Allen, 2001; Parker et al., 2006), and a positive interaction between forest fires and the mountain pine beetle epidemic that struck this ecozone+ in the last decade (Jenkins et al., 2008).

Figure 29. Total area burned by large fires per decade for the Montane Cordillera, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 29

This bar graph shows the total area burned by large fires per decade for the Montane Cordillera from the 1960s to the 2000s. Areas are approximately 4,000 km2 in the 1960s, 1,600 km2 in the 1970s, 1,700 km2 in the 1980s, 1,600 km2 in the 1990s, and 7,200 km2 in the 2000s.

Top of Page

The period during which fires occur in the Montane Cordillera spans from April to November, lasting 101 days on average (Figure 30). Fires are most common in June through August, peaking at the end of the summer, but can continue into the month of November. From the 1960s to 1990s, the duration of the active fire season did not changed, but the distribution of fires within the fire season did. Since the 1960s, there was an increase in the proportion of fires in both the spring and fall (Figure 30). The proportion of fires in the month of May increased from 2.7 to 17.9% (R² = 0.94, p = 0.03) and from 3.2 to 17.9% for the month of September (R² = 0.93, p = 0.03). These changes are consistent with predictions made by Wotton and Flannigan (1993) that the fire season would lengthen in both the spring and fall in British Columbia as global temperatures increase.

Figure 30. Proportion of large fires that occur each month in the Montane Cordillera and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 30

This bar graph shows the proportion of large fires that occur each month in the Montane Cordillera and an overlain line shows the average duration of the active fire season from the 1960s to 1990s. The period during which fires occur in the Montane Cordillera spans from April to November, lasting 101 days on average. Fires are most common in June through August, peaking at the end of the summer, but can continue into the month of November.

Top of Page

On average, fire ignitions over the long-term have been equally caused by humans and lightning in this ecozone+ (Table 10), thus the reason for the increase in both seasons. As stated in the National Trends section, spring fires are predominately human caused and summer fires are primarily caused by lightning (Stocks et al., 2003). There were slightly more lightning caused large fires than human caused fires on average between the 1960s and 1990s, at 52.8 and 47.2% respectively (Table 10). Over time these proportions changed significantly. At the beginning of the period of record lightning ignitions were responsible for 61% of the large fires; this declined to only 27% by the 1990s (R² = 0.95, p = 0.03) (Table 10). Area burned by lightning ignitions also declined significantly over the 40 year period by over 2,000 km² (Table 10). Despite the increase in the proportion of human ignitions there was also a decline in area burned by humans, although this decrease was not significant. Both sources of ignitions lead to similar areas burned.

Table 10. Proportion of the number of large fires ignited by lightning compared to humans by decade for the Montane Cordillera, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total58047*53*4,417.43,888.9
1960s22261392,345.52,095.6
1970s15955451,073.2578.0
1980s1374555817.0901.0
1990s622773181.6314.2

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 10, footnote * referrer

Top of Page

Western Interior Basin

Large forest fires do not play a significant role in the disturbance regime of the Western Interior Basin. Previous ecological classifications have included this area as part of the Montane Cordillera Ecozone+ (see Preface on page i), therefore any supplementary information available in the fire literature for this area is the same as that presented for the Montane Cordillera. One can assume that similar conditions apply. Fire weather is severe for this region (Parisien et al., 2006), but even fewer large fires occur than in the Montane Cordillera. Fires in the Western Interior Basin range from years with no large fires to extreme years where up to 800 km² burned (Figure 31). On average, only 54 km² burned each year. This area contributes less than 1% to the total area burned in Canada (see Table 1 on page 5) and has varied little over the last 40 years (see Figure 3 on page 7). Within the ecozone+ the percent forested area that burns annually is slightly higher than the Montane Cordillera at 0.11% per year.

Figure 31. Annual area burned by large fires for the Western Interior Basin, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 31

This bar graph shows the annual area burned by large fires for the Western Interior Basin from 1959 to 2007. Fires in the Western Interior Basin range from years with no large fires to extreme years where up to 800 km2 burned (2003).

Top of Page

Until the 2000s, area burned in the Western Interior Basin was relatively stable (Figure 32). There was a slight decline in the 1980s, a trend that also occurred in the nearby Pacific Maritime Ecozone+, and was attributed to a reduction in fires as a result of prevention education and aggressive fire attack (Stocks et al., 2003). This level was not sustained into the 1990s, which returned to levels of area burned similar to the 1960s and 1970s. The modest increase in the 1990s may be attributed to warmer temperatures (Stocks et al., 2003; Gillett et al., 2004). The most significant change over the past five decades occurred from the 1990s to the 2000s. Area burned increased by over 1,000 km². This trend also occurred in the Montane Cordillera. The same potential drivers apply to this ecozone+ including: positive climate oscillations such as an El Niño and a warm phase PDO (Volney and Hirsch, 2005), increased fuel loads in the forest due to long-term suppression (Allen, 2001; Parker et al., 2006), and a positive interaction between forest fires and the mountain pine beetle epidemic (Jenkins et al., 2008).

Figure 32. Total area burned by large fires per decade for the Western Interior Basin, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 32

This bar graph shows the total area burned by large fires per decade for the Western Interior Basin from the 1960s to the 2000s. Areas are approximately 300 km2 in the 1960s, 400 km2 in the 1970s, 200 km2 in the 1980s, 500 km2 in the 1990s, and over 1,500 km2 in the 2000s.

Top of Page

Fire suppression and human land use are also high in the Western Interior Basin. Areas such as the Okanagan Valley contain high values at risk for protection and increased concentration of land use due to extensive population in the area. All of the area is protected by fire suppression activities based on statistics for the Montane Cordillera (Stocks et al., 2003; Parisien et al., 2006). Compared to the Montane Cordillera, the average duration of the active fire season is much shorter – 45 days compared to 101 days (see Table 2 on page 9). Most fires occur from June to August, with the peak in July, but fires have been recorded as early as April and as late as October (Figure 33). There was no significant change in the duration of the fire season or in the distribution of fires during the season from 1959 to 1999. The duration of the active fire seasons decreased by 27 days but this decline was not statistically significant and may be linked to a significant decrease in the number of fires that have occurred over the last two decades (Figure 33). More data are required to make any definitive insights into these changes.

Figure 33. Proportion of large fires that occur each month in the Western Interior Basin and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 33

This bar graph shows the proportion of large fires that occur each month in the Western Interior Basin and an overlain line shows the average duration of the active fire season from the 1960s to 1990s. Most fires occur from June to August, with the peak in July, but fires have been recorded as early as April and as late as October. There was no significant change in the duration of the fire season or in the distribution of fires during the season from 1959 to 1999.The duration of the active fire season decreased by 27 days but this decline was not statistically significant.

Top of Page

Humans are the dominant cause of fires in the Western Interior Basin (Table 11). This trend is consistent from the 1960s to the 1990s. Lightning ignited fires are less common here than in the Montane Cordillera. The Western Interior Basin is more similar in this regard to the Pacific Maritime Ecozone+.

Table 11. Proportion of the number of large fires ignited by lightning compared to humans by decade for the WesternInterior Basin, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total11422*78*234.2989.1
1960s38217975.5268.6
1970s36287254.0346.8
1980s20257582.3108.5
1990s20158522.4265.2

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 11, footnote * referrer

Top of Page

Pacific Maritime

Fire is not a significant natural disturbance in the Pacific Maritime Ecozone+ (Stocks et al., 2003; Amiro et al., 2009), nor has it been in the distant past (> 6,000 years) (Lertzman et al., 2002). The contribution to annual area burned in Canada is the lowest of all ecozones+ at 0.3% (see Table 1 on page 5). It is also insignificant at the ecozone+ scale, where only 0.02% of the forested area of the ecozone+ is affected by fires on average per year. Despite the inclusion of some subalpine and montane species along the eastern boundary of the ecozone+, fire is not a prominent feature on the landscape. These species readily burn in the adjacent Montane Cordillera but the wet, maritime climate of the Pacific Maritime is not conducive to fire.

Fires that do occur are predominately caused by humans (72%) through recreational activities and logging practices (Pew and Larsen, 2001). The average annual area burned is only 20.8 km² and ranges from zero in some years to an extreme of 144 km² in 1961 (Figure 34). The majority of fires that do occur are very small, in the range of 0.1 ha or less (Pew and Larsen, 2001). These small fires are not included in the data from 1959 to 1994. Over the last five decades there has been some change in annual area burned (Figure 35). From the 1960s through to the 1980s, the area burned decreased. This is most likely due to a reduction in fires caused by humans as a result of prevention education and aggressive fire attack (Stocks et al., 2003). Since the 1980s, the area burned has increased by 84 km², which is not a significant increase over 20 years. The trends in long-term area burned are purely speculation at this point as it is difficult to elucidate trends when the fire activity is infrequent and area burned is so low.

Figure 34. Annual area burned by large fires for the Pacific Maritime, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 34

This bar graph shows annual area burned by large fires for the Pacific Maritime from 1959 to 2007. The average annual area burned is only 20 km2 and ranges from zero in some years to an extreme of 144 km2 in 1961.

Top of Page

Figure 35. Total area burned by large fires by decade for the Pacific Maritime, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 35

This bar graph shows the total area burned by large fires by decade for the Pacific Maritime from the 1960s to the 2000s. Areas are approximately 370 km2 in the 1960s, 210 km2 in the 1970s, 120 km2 in the 1980s, 150 km2 in the 1990s, and 180 km2 in the 2000s.

Top of Page

Atlantic Maritime

Forest fire is not a common or significant natural disturbance in the coastal climate of the Atlantic Maritime Ecozone+. On average 34 km² burns in this ecozone+ each year. This area is less than 1% of the total area burned in Canada and represents only 0.02% of the forested area of the ecozone+ (see Table 1 on page 5). There is some variability in area burned from year to year (Figure 36), but the magnitude is insignificant when compared to the range in the other ecozones+. Extreme fire years in 1986 and 1995 burned 374 and 296 km² respectively. It is common in the Atlantic Maritime to have years where there are no large fires.

Figure 36. Annual area burned by large fires for the Atlantic Maritime, 1959-2007.

Source: data from 1959-1994 are from the large fire database; data from 1995-2007 are derived from remote sensing.

Long Description for Figure 36

This bar graph shows the annual area burned by large fires for the Atlantic Maritime from 1959 to 2007. There is some variability in area burned from year to year, but the magnitude is insignificant when compared to the range in the other ecozones+. Extreme fire years in 1986 and 1995 burned 374 and 296 km2 respectively. It is common in the Atlantic Maritime to have years where there are no large fires.

Top of Page

The active fire season in the Atlantic Maritime Ecozone+ lasts an average of 32 days and fires most commonly occur in May to June, with the peak occurrence in May (Figure 37). The early peak in the fire season is related to the primary cause of fire ignition in this ecozone+, which is humans (Table 12). Areas dominated by human ignitions tend to have greater fire occurrence earlier in the year (Stocks et al., 2003). On average, humans ignited 86% of the fires in the Atlantic Maritime from the 1960s to 1990s and are the dominant contributor to area burned (Table 12). Lightning ignited fires are relatively rare in the Atlantic Maritime; there was an absence of lightning caused fires for almost 20 years in this ecozone+.

Figure 37. Proportion of large fires that occur each month in the Atlantic Maritime and the average duration of the active fire season (purple line), 1960s-1990s.

Monthly numbers are the percentage of the total number of fires that occurred during the month.
Source: data are from the large fire database.

Long Description for Figure 37

This bar graph shows the proportion of large fires that occur each month in the Atlantic Maritime and an overlain line shows the average duration of the active fire season from the 1960s to 1990s. The active fire season in the Atlantic Maritime Ecozone+ lasts an average of 32 days and fires most commonly occur in May to June, with the peak occurrence in May.

Top of Page

Table 12. Proportion of the number of large fires ignited by lightning compared to humans and the total area burned by each ignition source by decade for the Atlantic Maritime, 1960s-1990s.
DecadeNumber of firesProportion of lightning caused fire ignitions (%)Proportion of human caused fire ignitions (%)Total area burned as a result of lightning ignitions (km²)Total area burned as a result of human ignitions (km²)
Total8214*86*263.81,129.1
1960s30237748.4356.0
1970s40100023.1
1980s2601000514.9
1990s223268215.4235.2

Footnotes

Footnote *

average not total
Source: data are from the large fire database

Return to table 12, footnote * referrer

Top of Page

Similar to the Pacific Maritime, it is difficult to determine significant trends in long-term area burned because the frequency of fires is so small; for example, there were only four fires in the 1970s (Table 12). Figure 38 shows the change in area burned over the last five decades. There was a sharp decline in the 1970s, which could be attributed to prevention and suppression efforts, but this would not explain the following rise in the 1980s that was sustained into the 1990s. The decline in the 2000s is similar to what occurred at the national level, but again it is difficult to explain why these trends are occurring without additional information.

Figure 38. Total area burned by large fires per decade for the Atlantic Maritime, 1960s-2000s.

The value for the 2000s decade was pro-rated over 10 years based on the average from 2000-2007.

Long Description for Figure 38

This bar graph shows the total area burned by large fires per decade for the Atlantic Maritime from the 1960s to the 2000s. Areas are approximately 400 km2 in the 1960s, 20 km2 in the 1970s, 500 km2 in the 1980s and 1990s, and over 100 km2 in the 2000s.