Observations from Grass Burns in Fort McMurray April 25, 2005 Mowed and Standing Grass Plots Fall Mow

Transcription

1 Observations from Grass Burns in Fort McMurray April 25, 2005 Mowed and Standing Grass Plots Fall Mow Greg Baxter FPInnovations Wildfire Operations Research In attendance: Greg Baxter (FERIC) Gary Dakin (FERIC Consultant) Casper de Groot (FERIC summer student) Fort McMurray SRD (led by Darrell Beam) Background Two sets of burns were completed in Fort McMurray on April 25. Each set contained two plots (one mowed and one standing ), each 50 x 50 m in size (Fig.1). The plots were treated the previous fall by mowing. This is in contrast to the burns carried out in 2004 where mowing was completed during the same spring prior to the burns. The two plots were ignited at the same time and fire behaviour was documented. Four sets of burns were completed in 2004 as part of the linear disturbance project. The objective of this research is to determine the influence that mowing has on potential fire behaviour. A number of industries use mowing to keep lines clear and reduce the fuel load but little research has taken place to document its influence. Figure 1. Plot 1a (mowed on right) and 1b (standing ).

2 Fire Weather * April 25, 2005 FFMC 94 DMC 53 DC 323 ISI 10 BUI 75 FWI 26 *calculated by Darrell Beam (SRD) On Site: T RH W SE E E E (observed using a hand held Kestral 3000) Ignition Times Plot Plot Fire Behaviour In 2005, there was a distinct difference in fire behaviour between the mowed plots and the plots containing standing. In 2004, fire in the mowed plots travelled at roughly 85% the ROS as fire in the standing. These plots were mowed during the spring a month prior to the burns. In 2005, the plots in Fort McMurray were mowed during the previous fall and allowed to settle over the winter. This resulted in reduced fuel loads, a quicker green-up and thus a reduction in rates-of-spread. Initial indications are that fall mowings can have an important influence on potential fire behaviour (although this thought is based on only two trials). There were some unburned patches in Plot 1a (mowed). More clover (sweet clover) existed in Plot 2a than in 1a. Flamelength (FL) Plot 1a (mowed) had observed FL s of about 50 cm at the highest these occurred when the fire hit jackpots of fuel. Mean flamelengths were about 25 cm. Plot 1b (standing) had flamelengths of 2-4 m in that was about 50 cm in height (average). Flamelengths were slightly longer in Plot 2a (mowed), where they averaged about 40 cm with flashes up to 1.0 m. Plot 2b had similar flamelengths to those observed in Plot 1b. Rates of Spread (ROS) 25 m 50 m Plot 1a (mowed) 9:38 (2.4 m/min) did not reach

3 Plot 1b (standing) 1:01 (24.6 m/min) 1:41 (30 m/min) Plot 2a (mowed) 2:50 (9 m/min) 3:32 or 6:06 (13.8 or 8.4 m/min) Plot 2b (standing) 1:27 (16.8 m/min) 1:54 (26.4 m/min) Figure 2. Difference in fire behaviour between Plots 1 a and b. ROS and flamelength are far more pronounced in the standing. Fire spread in plot x faster in the standing than in it s associated mowed plot. In plot 2 the fire spread 2 3 x faster in the standing. We recorded two times for plot 2b as the fire spread along one side (in heavier fuels) to reach the end of the plot whereas it spread more slowly through the middle of the plot where fuel loads were less. (It appeared to get in the standing beside the plot on the south side to increase its ROS). The official time will be the slowest time to 50 m (6:06 minutes). Temperature Data FERIC tested newly acquired Hobo dataloggers during the burns. Two thermocouples were placed in each plot at the 25 m mark. One was placed at the surface and the other 1.0 m above the surface. Three of the four dataloggers provided data. Data lines 1 and 2 are from the mowed plots and data line 3 is from the standing plots. All data is contained in Figure 3. Dataloggers were moved from one fire to the next while running so the data shows two peaks for each data logger. Lines 1 and 2 are from the mowed plots; line 1 is on the surface of the ground while line 2 collected data 1 m above the surface. Line 3 is from the standing plots and was located on the ground. All dataloggers were started at the same time and Figure 3 shows the fire reaching Logger 3 at 7597 seconds and then at Logger 3 at 8111 seconds a difference of 9.4 minutes. The temperature difference between the mowed and standing on the surface was 393 o C, showing the fire moved much faster and was hotter in the standing. There was a 2-

4 Temperature (C) minute and 261 o C difference between the fires in Plot 2. The extra peaks in the data are caused by electrical fluctuations not related to the fire Fire 1: standing Fire 1: mow ed Fire 2: standing Fire 2: mow ed 1 Ground 2 Aerial Ground Time (seconds) Figure 3. Surface and aerial temperatures from the mowed and standing plots. Both burns are included in the figure. In the standing surface temperatures peaked around 975 o C in Plot 1 and 700 o C in Plot 2 on the surface. The mowed plots had lower surface temperatures and very low temperatures 1 m above the surface. Arrows point at the temperatures recorded during the burn. FBP Predictions Using the predicted daily FWI values, the Red Book predicts a ROS for FBP fuel type O- 1a (matted ) of 27 m/min. Actual ROS was 2.4 m/min in Plot 1a. FBP fuel type O- 1b (standing ) was predicted to travel at 29 m/min whereas actual ROS was 30 m/min in Plot 2b. Rates of spread in plots 2 a and b were 8.4 and 26.4 m/min respectively. The degree of curing was estimated to be % for plots 1 and about 60-90% for plots 2. Actual degree of curing values is as follows: Plot 1a (mowed) 92%, 84% (estimated at 100 and 90%) Plot 1b (standing) 78%, 99% (estimated at 60 and 90%) Plot 2a (mowed) 91%, 84% (estimated at 90 and 60%) Plot 2b (standing) 96%, 95% (estimated at 80 and 80%)

5 Samples show there was a slightly higher degree of curing in the standing than in the mowed areas (92 vs 89.25%), although the green was much more visible in the mowed areas (Figure 4) resulting in lower percent cured estimations (visual assessments). Figure 4. Visible green-up in Plot 1a. Fuel Loads There was a dramatic difference in fuel loading between spring and fall mowings. Spring mowings (2004) had a mean load of 5 t/ha. With fall mowings (2005), these loads decreased to a mean of 2.5 t/ha thus the loads were reduced by half. This would have an influence on potential fire behaviour (fire intensity). Plots 1a 1b 2a 2b 2.85 and 2.7 t/ha 2.4 and 5.85 t/ha 1.83 and 2.64 t/ha 2.6 and 3.6 t/ha Moisture Contents 1a (mowed) 33.6% 1b (standing) 20.6% 2a (mowed) 4.4% 2b (standing) 5.3% Moisture contents were significantly higher in Plots 1 a and b compared to Plots 2.

6 FBP Fire Intensity Estimates Plot 1 Standing ROS Fuel load 30 m/min 3.61 t/ha I = 300 w (kg/m2) r (m/min) I = 3249 kw/m Mowed I = kw/m ROS 2.5 m/min Fuel load 2.5 t/ha Plot 2 Standing ROS Fuel load 26.4 m/min 3.1 t/ha I = 300 w (kg/m2) r (m/min) I = 2455 kw/m Mowed I = kw/m ROS 8.4 m/min Fuel load 2.23 t/ha The calculated intensities show the standing has values that would make containment of a similar fire difficult. Intensity values averaged 375 kw/m in the mowed plots. At this intensity hand crews should be able to control the fire with handtools. Containment would be very difficult in the standing as rates of spread were high and the fires were very hot which would limit the ability to use direct attack. Comments There was a significant difference in spread rates in the mowed plots between 2004 and 2005, and this difference may be attributed to the time when the mowing of the plots took

7 place. In 2004 mowing occurred just a few weeks before burning took place, and the mowed was still on site available as fuel as decomposition did not have time to occur. Because of this, surface fuel loads appeared to be heavier and fire was able to spread more quickly through the cuttings than in areas where this fuel was not present (as in the 2005 burns). The same plots were mown in the fall previous to when the burns took place in 2005, resulting in half the fuel on site. Fall precipitation would have acted to speed up the decomposition process and the winter snow load acted to compressed the fuel. This set of burns illustrated the potential changes in fire behaviour in fuels based on when a mowing treatment takes place. These were only two trials, but they did reveal a significant difference in fire behaviour between mowed and standing plots compared to the fire behaviour observed in 2004 at the same sites. This raises the issue of the timing of the treatment and or repetitive burns. These trials should be completed again next spring (using a fall mowing) to understand the fire behaviour observed this year. More trials are scheduled to take place in Slave Lake this spring. These will be sites that are treated this spring.