Assessing the Effectiveness of Using Prescribed Fire to Create Bison Pasture in Heart Lake, Alberta
Sebastian Hetsch and Greg Baxter
Forest Engineering Research Institute of Canada (FERIC)
2601 East Mall, Vancouver, B.C.
March 2005
Introduction
A prescribed burn was conducted at the Heart Lake Indian Reserve north of Lac La Biche on May 15, 2003. The burn was conducted by the Lac La Biche Fire Management Area (FMA) for the Heart Lake Indian Band with the goal of improving bison pasture.
Following the harvest of the spruce-aspen forest cover in 1995, only aspen regenerated. At the time of the burn, the area was fully stocked with aspen suckers 3–6 m in height. The intention of this prescribed burn was to remove the regenerating aspen and harvesting debris piles to allow the area to re-colonize with grass for improved bison pasture (Figure 1).
FERIC staff monitored the burn with the following objectives:
· Correlate pre-burn fuel loading and fire weather data with actual fire behaviour.
· Assess the effectiveness of the burn in killing the aspen re-growth on the site.
Figure 1: Bison grazing on Heart Lake reserve.
Methods
Prior to the burn, sampling was done to estimate potential fire behaviour and fire effects. Fuel loading was measured using a standard fuel triangle to determine surface fuel loads, and the percentage of cured grass was estimated with the collection of samples. During the burn, fire rates of spread were observed and documented, and fire weather data were collected on site and from the nearest fire tower. Still photos and video footage of the burn were also taken.
The site was revisited two months later on July 17 and re-measurements were taken. The fuel triangle was used to re-sample the site and the mortality of the aspen due to the fire was determined. Aspen mortality was sampled using 18 line transects (30 m long and 1 m wide). Burned trees, living trees, and new shoots that had sprouted from roots were counted.
Results and discussion
Fire weatherThe fire weather observations taken on the site (Table 1) and those calculated from the Heart Lake tower (Table 2) both indicated a high fire danger rating. These conditions suggested that an intense surface fire was possible.
Table 1: On-site weather observations recorded on day of burn
| Relative |
|||
| Time |
Temp |
Wind speed |
humidity |
| (deg. C) |
(km/h) |
(%) |
|
| 1520 |
21 |
18 |
20 |
| 1540 |
22.7 |
9 |
18.5 |
| 1600 |
21.2 |
16 |
19 |
| 1620 |
22.2 |
17 |
14.3 |
| 1640 |
23.3 |
7.6 |
17 |
| 1700 |
22.7 |
12.6 |
17 |
| 1710 |
n.a. |
9.6 |
n.a. |
| 1730 |
23 |
11.3 |
13 |
| 1745 |
23.2 |
6.1 |
15 |
| 1812 |
24 |
2.2 |
16 |
| 1900 |
21.7 |
12.4 |
17 |
| 1912 |
24.8 |
1.5 |
19 |
Table 2: Fire Weather Index (FWI) values taken from Heart Lake tower on day of burn
Fine Fuel Moisture Code (FFMC) 90
Duff Moisture Code (DMC) 20
Drought Code (DC) 177
Initial Spread Index (ISI) 11.3
Fire Weather Index (FWI) 19
Results from the post-fire fuel load measurements show that little of the woody material on the ground had burned; however, the cured grass (estimated at 5 t/ha) and the debris piles had burned completely. The litter and duff layers were for the most part unburned. This was probably due to the wet spring soils at the time of the burn. These fire effects indicate that the intensity of the fire was low, which was a contradiction from predictions by the Canadian Fire Behaviour Prediction (FBP) model in the grass fuels. Although some flare-ups were observed, flames were about 1 m in height (Figure 2), which indicates intensity values <500 kW/m. The intensity values predicted using the FBP system were 2000–3000 kW/m, which are rated as very high.
Figure 2: Fire behaviour in grass fuels at the Heart Lake burn.
Aspen mortality
The literature indicates that aspen mortality in a low intensity fire is often low, and that aspen regeneration following low intensity fires can be vigorous, especially in the first two years (Brown and DeByle 1987; Brown and DeByle 1989; Ryan 1982). This was the outcome observed after the Heart Lake burn. Even though the tops of most (99.6%) of the aspen stems died in the fire (Figure 3), the stocking density had increased from 15 833 to 18 574 stems/ha (117%) two months after the burn (Table 3). New shoots developed from either the aspen roots (Figure 4), or from the stem just below the height at which the cambium was killed by the flames (Figure 5).
Figure 3: Post-burn site conditions -- aspen stems with dead tops and dense sucking.
Figures 4 and 5: Shoots developing from the roots and from the stem.
Table 3: Aspen stocking data from transect sampling two months following the burn
| Transect line reference |
Stems (no.) |
Stems (no.) |
New shoots on dead (no.) |
Post-burn stocking change (%) |
| F1 |
0 |
70 |
76 |
109 |
| F2 |
0 |
63 |
68 |
108 |
| F3 |
0 |
53 |
38 |
72 |
| R1 |
0 |
96 |
102 |
106 |
| R2 |
0 |
42 |
51 |
121 |
| R3 |
0 |
73 |
106 |
145 |
| S1 |
2 |
32 |
40 |
125 |
| S2 |
2 |
26 |
65 |
250 |
| S3 |
1 |
53 |
59 |
111 |
| B1 |
0 |
23 |
27 |
117 |
| B2 |
0 |
5 |
18 |
360 |
| B3 |
0 |
15 |
10 |
67 |
| P1 |
0 |
63 |
55 |
87 |
| P2 |
0 |
89 |
71 |
80 |
| P3 |
0 |
47 |
40 |
85 |
| P4 |
0 |
41 |
56 |
137 |
| P5 |
0 |
45 |
67 |
149 |
| P6 |
0 |
19 |
54 |
284 |
| sum |
5 |
855 |
1 003 |
|
| average |
48 |
56 |
117 |
|
| stems/ha |
93 |
15 833 |
18 574 |
Conclusions
The initial post-burn stocking survey at the Heart Lake prescribed burn found the treatment to be ineffective in creating bison pasture by killing aspen. Aspen stocking density increased by 117% as a result of the treatment.
It is believed the spring conditions had an impact on actual fire behaviour as compared to what was predicted by the Canadian FBP model. The soil was likely more moist than that predicted by the FBP model and thus fire behaviour was less intense than predicted.
Implementation
This prescribed burn showed that a more intense fire is required to remove aspen from a site. The intensity of the required burn is unknown at this time, but it should be greater than 500 kW/m.
Moisture contents of the deeper layers should be sampled to help improve model predictions. Samples of the duff layer (down to 5 cm in depth) could be collected to initialize the Fire Weather Index (FWI) model, as these values are then used in the FBP model.
The FBP model is appropriate for this work, but accurate season starting values are required.
The site will be revisited during 2005 and re-sampled. Debris pile sites will also be sampled to see how these sites have re-vegetated, if at all. A FERIC Advantage report will then be produced.
References
Brown, James K.; DeByle, Norbert V. 1987. Fire damage, mortality and suckering in aspen. Canadian Journal of Forest Research 17:1100–1109.
Brown, James K.; DeByle, Norbert V. 1989. Effects of prescribed fires on biomass and plant succession in western aspen. USDA Forest Service Research Paper INT-412. 16 pp.
Ryan, K.C. 1982. Evaluating potential tree mortality from prescribed burning. Pages 167–179 in D. M. Baumgartner, editor. Site preparation and fuel management on steep terrain. Washington State University, Cooperative Extension, Pullman, Wash.