Linear Disturbances at the 2004 Freeman River Fire
Introduction
FERIC is currently investigating the influence linear disturbances have on potential fire behaviour. To date, four sets of experimental fires have been carried out (see FERIC website at http://fire.feric.ca) comparing fire behaviour on mowed versus unmowed plots of grass. FERIC saw the Freeman River Fire as an opportunity to collect some data on a fire containing an abundance of linear disturbances.
The Freeman Fire began at approximately 4 pm on May 18th (the cause is still under investigation) under a powerline in close proximity to a gravel road. The fire grew rapidly to 100 hectares in the first four hours and then to just under 1000 hectares by 10 AM the next morning. A quick look at the fire map shows there were many linear disturbances inside the fire boundary and quick and rough estimates show that there were roughly:
· 25 km of pipelines
· 10 km of powerlines
· at least 13 wellsites, and
· many kilometres of roads
Easy access gave FERIC the opportunity to view the fire area in detail and to collect samples for fuel load and degree of curing in the grass along these right-of-ways (see map below).
Map of Freeman River Fire and the abundant linear disturbances.
Data Collection
FERIC travelled to the Freeman River Fire on May 25th in order to collect fuel load and degree of curing samples from linear disturbances directly adjacent to the fire boundary. This was 6 days after the fire started. Most grass within the fire was burned although one site was found where the fire left some grass untouched within the fire boundary. The objectives of the data collection were to:
· Determine grass fuel loads and degree of curing along the linear disturbances.
· Compare areas that were mowed to those of natural, standing grass in terms of fuel load and degree of curing.
· Estimate fire behaviour based on the collected data and fire weather observations.
Fuel load data was collected at five locations around and within the fire perimeter (this includes the point of origin of the fire). Degree of curing samples were also collected at the same sites and at one location where it appears the amount of green grass stopped fire spread. Fuel load data was collected by cutting all the grass under a disc with a diameter of 46 cm (1662 cm2) and then oven drying the sample for 24 hours at 80oC to determine an oven dry weight (this was then converted to tons/hectare). The grass samples collected were separated into live and dead (i.e., green and cured) components. These separate samples were then oven dried and a degree of curing was calculated based on the dry weight of each sample.
Results
Fuel Load Data
Sample Number |
Location |
t/ha |
Comments |
1 |
Pipeline |
10.04 |
Standing grass |
2 |
Powerline |
4.66 |
Standing grass |
3 |
Powerline |
1.56 |
Mowed |
4 |
Powerline |
5.37 |
Point-of-origin, standing grass |
5 |
Powerline |
1.43 |
Standing grass, retardent |
Degree of Curing
Sample Number |
Location |
Degree of Curing % |
Comments |
1 |
Pipeline |
89 |
Standing grass |
2 |
Powerline |
82 |
mowed |
3 |
Powerline |
88 |
Point-of-origin, standing grass |
4 |
Edge of burn |
74 |
Standing grass |
5 |
Powerline |
92 |
retardent |
Photographs of Fuel Loads
The following photos are presented as a visual ‘database’ of what specific fuel loads (in t/ha) of the grass fuel complex ‘look’ like. This information may assist fire behaviour specialists when estimating fire behaviour potential in grass fuels by gaining an understanding of the amount of fuel available to burn.
Plot 1. 10 tons/ha.
Plot 2. 4.66 tons/ha.
Plot 3. 1.56 tons/ha (mowed)
Plot 4. 5.37tons/ha (point of origin)
Plot 5. 1.43 tons/ha.
Degree of Curing
The degree of curing at this location was 74%. This amount of green grass appeared adequate to slow fire spread. There were no visible signs of any initial attack at the location to show that firefighting efforts created this break suggesting this amount of ‘green’ grass was responsible for limiting the spread. On the other hand, there is experimental evidence indicating fire spread in grasslands down to 50% curing or less (Cheney and Sullivan 1997).
Discussion
According to Phil Robert, an Air-attack Officer and Wildfire Behaviour Specialist course graduate, who was working the fire, he felt that the linear disturbances did not contribute appreciably in any way to the difficulty of controlling the fire. Both forest and grass fuels carried the fire, making control of this fire difficult in terms of its rapid spread and growth. Although linear disturbances did not play a key role in the behaviour of this fire, the fire did start on a powerline and did consume most of the grass along rights-of-way inside the fire boundary. Because full green-up still had not occurred along the lines, the fire provided us with an opportunity to collect data on a fire site that is typical of springtime fire conditions that occur in Alberta.
Fuel loads varied considerably along these disturbances from 1.43 to 10.04 t/ha, with a mean value of 4.61 t/ha. This average fuel load is slightly greater than that used as a standard value in the FBP System Fuel Type O-1 models of 3.5 t/ha, and this heavier fuel load leads to a difference of 1382 kW/m in fire intensity using the inputs observed at the fire, potentially leading to a fire that has an intensity that is far more difficult to control.
Mowed versus Standing grass
One site was sampled that had been mowed. Although this is only one sample, it is compared to the other sites using the FBP system model. The degree of curing and fuel loading varied between the mowed and unmowed plots, as well as the FBP fuel type (O-1a versus O-1b). These values were entered into the model to illustrate the potential differences in fire behaviour. The results are as follows:
Standing Grass Mowed Grass
Tons/ha 5.4 1.5
Degree of Curing 86% 82%
Head Fire Intensity 6724 kW/m 1441 kW/m
Rate of spread 41.5 m/min 32.0 m/min
Estimated Flame length* 4.7 m 2.2 m
* The flamelength was estimated using the equation that estimates flamelength from intensity. L = √(I/300) (Newman, 1974). Keep in mind we are estimating flamelengths using estimated fire behaviour outputs!
The above comparison readily shows the difference in fire behaviour between the mowed and standing grass. Firstly, the FBP model predicts a difference of 5300 kW/m for fire intensity. This is a very significant output, and would lead to completely different strategies when controlling a fire. The standing grass had a calculated intensity of 6724 kW/m, which is above the EXTREME threshold of 4000 kW/m and suggests that an indirect attack strategy would have to be used (See Alexander and Degroot 1988 poster). If you combine this intensity with a fire estimated to be travelling at 41.5 m/min you have a fire that requires a fuel change or a change in the weather to control it. The simulated fire behaviour characteristics for the mowed grass suggest a more readily controllable situation. Rates of spread are 10 m/min slower and intensity values are just over 1000 kW/m (flamelengths are also half the length). A fire of this intensity gives firefighters a greater chance of stopping the fire using more direct attack techniques. Research (Alexander 1994) suggests that fire has a 44% probability of breaching a 4 m firebreak in a grassfire with an intensity of 6500 kW/m (with no trees present) and a 44% probability of breaching a firebreak of 2 m with a fire with an intensity of 1000 kW/m. If trees are present firebreaks must be considerably wider.
Summary
The trip to the Freeman River Fire site contributed to FERIC’s growing knowledge base on the effects linear disturbances have on fire behaviour potential. Data was collected on fuel loads and degree of curing along linear disturbances that had standing grass and one site that had been mowed. The data collected will hopefully help fire behaviour specialists estimate fuel loads and this information can be used in the FBP System rate of spread models to predict potential fire behaviour. The degree of curing information ascertained for this fire is also important as potential fire behaviour is strongly influenced by the amount of live versus dead grass.
Another benefit of this work, along with the use of the photographs in helping to estimate fuel loads, includes the collection of basic fuel load data using a ‘disc meter’ that is designed to provide an estimate of fuel load based on the resting height of the disc on the grass fuel bed. This technique has not been calibrated for grass fuel complexes in Canada as of yet and the data will be used to eventually provide fuel load-height relationships for standing and mowed grass. At this stage only a very rough relationship exists – there is a need for the collection of more samples to develop a stronger relationship.
References
Alexander, M.E. 1994. Proposed revision of fire danger class criteria for forest and rural areas in New Zealand. National Rural Fire Authority, Wellington, NZ. Circ. 1994/2. 73 p.
Alexander, M.E.; De Groot, W.J. 1988. Fire behaviour in jack pine stands as related to the Canadian Forest Fire Weather Index (FWI) System. Can. For. Serv., North. For. Cent., Edmonton, AB. Poster (with text).
Cheney, P. and Sullivan, A. 1997. Grassfires: fuel, weather and fire behaviour. CSIRO Publishing, Collingwood, Australia. 102 p.
Newman, M. 1974. Toward a common language for aerial delivery mechanics. Fire Management 35(1): 18-19.