Current Research

The staff and students associated with the Fire Management Systems Laboratory study the impact of fire on boreal forest ecosystems and use operational research methods and information technology to develop decision support systems for fire and forest managers.

Forest Fire Occurrence Prediction

Forest fire managers need to know when and where fires will occur so they can deploy fire fighting resources where they will be needed before the fires actually occur. Statistical methods can be used to analyze historical data to develop models that can be used to produce probabilistic predictions about fire occurrence processes. The procedures that most Canadian forest fire management agencies use to predict people-caused forest fire occurrence are based on fire occurrence prediction research carried out at University of Toronto in collaboration with the Canadian Forest Service and the Ontario Ministry of Natural Resources.

We are currently exploiting advances in statistics to develop new fire occurrence prediction models that can be used for fire management purposes. Mike Wotton, recent graduate and currently a visiting professor has developed a new people-caused fire occurrence prediction model and he is also developing a new lightning-caused fire occurrence prediction model.

Back to top of page

Forest Fire Management and Climate Change

The current generation of General Circulation Models (GCM's) project a major warming trend over the next 50-100 years, particularly in northern circumpolar countries. Forest fire occurrence and fire behavior can be expected to be affected early and dramatically by global warming. Canadian boreal forests are a source of timber and other forest resources and they also act as a major global carbon pool. Forest growth, timber harvesting and forest fires can contribute to significant flows to and from that carbon pool. It is therefore essential that we determine the extent to which climate change can both influence and be influenced by fire and forest management.

The Canadian forest fire record is often used to identify what many observers have described as an increase in fire activity due to climate change. However, simple plots of fire activity across Canada beginning in the early 1900's are difficult to interpret due to the fact that the area in which intensive fire management has taken place and fire management technology as well as land use patterns and forest vegetation have changed over that period. The analysis of fire activity trends is further complicated by the enormous year-to-year variability that is characteristic of fire activity in Canada.

One of our research interests is the extent to which there has been a significant change in fire activity across Canada since the early 1900's. The variability of fire activity and the fact that many factors may contribute to that variability are such that simple time series analysis methods that have been developed to detect trends in data cannot be applied to forest fire activity data. Justin Podur adapted statistical methods that have been developed for quality control methods in manufacturing and was able to apply them to the historical Canadian forest fire record. He applied those methods to all of Canada, the province of Ontario, and a portion of Northwestern Ontario. He partitioned the historical record into two time periods: 1921-1960 and 1961-2000. He found that there was a significant increase in both reported fire occurrence and reported area burned in all three areas. The results of his study are presented in Podur, Martell and Knight (2002). Since fire suppression effectiveness has increased over that period, these results suggest that the reported increase in activity may be due to climate change but since land use patterns and forest vegetation have also changed over that time, those changes cannot be interpreted as definitive evidence that fire activity has increased due to climate change.

Mike Wotton has used his new people-caused fire occurrence prediction model together with weather projections produced using GCM models to predict how fire occurrence might be influenced by climate change. His results (described in Wotton, Martell and Logan (2003)) indicate that people-caused fire occurrence in Ontario might increase by 18% for the period 2020-2040 and as much as 50% by the end of the 21^st century.

Back to top of page

Predicting Annual Area Burned for Strategic Forest Management Planning

Annual area burned depends on forest vegetation, weather and climate, and on fire suppression activity and fire managers need to know the extent to which those factors contribute to area burned. Although historical data and existing statistical models can be used to predict the average or expected area burned each year assuming average weather and no significant change in the current fire management system, such predictions are not adequate for the development of strategic plans that may call for different levels of fire protection and changes in climate. Justin Podur has therefore been developing simple analytical and computer simulation models for predicting annual area burned based on vegetation or fuel, weather and level of fire protection parameters. To date he has been working with a simple elliptical fire spread model, Bernie Todd's "Wildfire" fire spread model, and the new Prometheus fire spread model.

Back to top of page

Level of Fire Protection in Ontario Planning

Forest fire managers develop level of protection strategies that prescribe how fire is to be managed across vast areas as large as a province. They must determine the extent to which they will attempt to prevent people-caused fires and fight both human and natural origin fires in order to achieve burned area targets that are compatible with land management objectives. Once they have established burned area targets they must determine the mix of fire suppression resources such as fire fighters, transport helicopters and airtankers that is likely to achieve those targets at minimal cost. Understanding of the extent to which variations in level of protection, vegetation and weather influence area burned is crucial. David Martell is continuing his investigation of the impact of level of protection or fire suppression effort, forest vegetation and weather, on burned area in Ontario. His objective is to develop a simple model that can be used to predict how changes in level of protection can lead to increases or decreases in burned area across the province of Ontario.

Back to top of page

Intensive Forest Management, Fire and Climate Change

The Ontario Forest Accord Advisory Boards (OFAAB) "Room to Grow" policy framework supports investment in intensive forest management to provide more fiber and more parks and protected areas. Fire management strategies should be compatible with those land management objectives and other objectives (e.g., need to maintain some fire on the landscape and public safety) and land management objectives should be realistic given possible fire regimes. However, the development of truly integrated fire and forest management planning is a complex emerging area that will be complicated by the fact that climate change will influence forest growth and development, land use, fire and fire management productivity.

Forest managers that invest in intensive forest management in Ontario are expected to establish those investments at sites where biological productivity, delivered wood costs and other factors will enable them to maximize their economic returns. Since climate change is expected to influence both forest growth and fire management at rates that will vary over both time and space, there is a need for temporally and spatially explicit linkages between fire and forest management. We have developed computer-based decision support systems that address the needs of fire and forest management planners that must decide how to develop and evaluate integrated spatially explicit timber production and level of fire protection strategies. We are also exploring how they might satisfy the need to protect intensive forest management investments located at specific points on the landscape both now and as climate changes influences forest growth and fire activity in the future.

Back to top of page

Integrated Fire and Forest Management in the Boreal Forest

We are using decision analysis methods to develop a formal methodology for describing and evaluating integrated fire and forest management policies; and fuel management, road construction and cut block layout strategies that will shape flammable boreal forest landscapes in such a way that they have the potential to both provide an economic supply of industrial wood fiber and support healthy wildlife populations.

We are studying the effectiveness of what can perhaps best be described as "FireSmart" forest management strategies that are designed to reduce area burned and reduce the detrimental impact of fire on timber supply. Traditional fuel management relies on the construction of fuel breaks and other fuel modification measures designed to reduce the flammability of forest landscapes. We also consider the fact that logging access roads, harvest cut block layout, and forest regeneration activities also influence the flammability of the landscape. We are developing landscape models that can be used to evaluate integrated strategies that include both traditional fuel management measures and the impact of harvesting operations on forest landscape flammability. We are also developing spatially explicit harvest planning and road construction mathematical programming planning models that can be used to help evaluate such strategies.

Our approach will help managers decide how to strategically fragment forest landscapes but if such strategies were designed to satisfy timber production objectives alone, they could easily lead to the development of strategies that would fragment landscapes to the detriment of other forest values. We are therefore also investigating the impact of these strategies on wildlife habitat

This is a collaborative research project involving Kelvin Hirsch of the Canadian Forest Service, Rob McAlpine of the Ontario Ministry of Natural Resources, Mauricio Acuna, Cristian Palma, Dario Robak and Andres Weintraub of the University of Chile, and Wenbin Cui, Ana Espinoza, Jay Malcolm and David Martell of the University of Toronto. We are working with two forest companies: Millar Western Forest Industries in Alberta and Tembec in Ontario.

Back to top of page

Fire Management in Wildland Urban Interface (WUI) Areas

Forest fires often threaten public safety and property in what fire managers refer to as wildland urban interface areas, the relatively sparsely inhabited zone that separates urban areas from the forest in which such communities are enclosed. Canberra (2003),California (2003) and the southern interior of British Columbia (2003) serve as stark examples of what can happen when fuel, weather and fire combine to wreak havoc upon such areas.

Many fire management specialists have advocated the development of what is commonly refereed to as FireSmart forest management, the establishment of strategically placed fuel breaks and the conversion of some forest stands to less flammable fuel types to minimize the spread of large escaped fires and thereby reduce fire losses in WUI areas.

Mariam Sanchez-Guisandez worked with Wenbin Cui to develop a computer-based decision support system that can be used to help develop and evaluate FireSmart fuel treatment strategies for WUI areas. The DSS can be used to display a map that indicates the probability that each point on the landscape will be burned in the near future. A planner or fire manager can use that map along with a display of homes and other values at risk to identify 'hot spot' that are potential candidates for fuel treatment activities. He or she can then delineate selected forest stands that should receive fuel treatment or delineate potential fuel breaks on the landscape, The "modified" landscape can be used to produce a revised map of the predicted burn probability that would result if the proposed fuel treatment activities were implemented

Back to top of page

Fire Management in Parks and Protected Areas

Fire often poses threats to public safety, property and forest resources but fire is a natural ecosystem process that cannot and should not be excluded from all forested areas. The need for a balanced approach to fire management is particularly important in parks and protected areas where fire managers must work with other land managers to ensure that fire activity on the landscape is compatible with forest conservation and other objectives.

Jennifer Beverly investigated several aspects of fire regimes for protected areas. She carried out a statistical analysis of extreme fire events (the largest fire each year) and extreme dry-spell events (the longest run of consecutive days with little or no rain each year) and found that regional differences in the characteristics of extreme fire event processes across Ontario are

related to the occurrence of dry-spell events, the ecoclimatic characteristics of each region and the level of fire protection.

She also studied the impact of a prescribed fire in Quetico Park in Northwestern Ontario and developed a statistical model that can be used to relate white pine (Pinus strobus L.) mortality to tree size and fire intensity. She incorporated that model in simulation and goal programming models of white pine stand dynamics that can be used to determine the optimal or preferred prescribed fire interval to produce and maintain a sustainable population of large dead and live white pine trees in a healthy stand of white pine to enhance the stand's wildlife habitat and visual quality attributes.

Back to top of page

Sustainable Forest Management Under Uncertainty

Forest management planners develop plans that describe how the land for which they are responsible, typically a forest management unit several hundred thousand hectares in size, is to be managed during the near and far distant future. Such plans may include access road construction plans, descriptions of when and where timber harvesting and other resource management activities are to take place, the designation of protected areas such as nesting sites and other important wildlife habitat areas that are not to be disturbed by human activities, and fire management strategies. In their efforts to achieve sustainability they must decide what should be done while considering what might happen over very long time periods, typically 150 years or more into the future. Fire, insects and disease, changes in technology, the prices of forest products and how society views its forests and how they should be managed can be expected to change over time as the future unfolds and plans are implemented. Forest managers must therefore deal with enormous uncertainty and develop robust flexible plans that can be modified as the future as it unfolds.

We are continuing our investigation of sustainable forest management under uncertainty with emphasis on dealing with fire as a source of uncertainty. David Martell has been investigating the long term structure and sustainability of flammable forests that are managed according to plans prescribed by mathematical planning models. One of our objectives is to determine the extent to which such models produce "good" sustainable forest management plans. He has developed a simulated planning environment that includes a flammable forest that is managed by an embedded planning model. He can modify the embedded planning model to investigate how its structure including the re-planning interval, the length of the planning horizon, and the accuracy of the fire loss estimates influence the sustainability of the simulated forest.

Back to top of page