Progress 07/01/09 to 06/30/14
Outputs
Target Audience: Government agencies; forest landowners; the general public Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest? Major study findings have been presented in a professional meeting and published on Forest Science, a leading refereed journal in forestry. What do you plan to do during the next reporting period to accomplish the goals? This is the termination report for this five-year project.
Impacts
What was accomplished under these goals?
Fires are an inherently multivariate phenomenon and can be characterized by several correlated features or processes: land area burned, and fire duration, which can be measured as the time interval from fire detection to containment. Area burned by fires has been widely reported in fire statistics, but fire duration has usually been ignored. The objective of this study was to analyze distribution patterns and behavior of large wildland fires through bivariate extreme value modeling of area burned and fire duration. A dataset of fire occurrences in Mississippi between January 1991 and December 2007 was used in the analyses. The methods employed were descriptive statistics, univariate extreme value statistics (i.e., generalized Pareto distribution), and bivariate extreme value statistics (i.e., asymmetric negative logistic distribution). Descriptive statistics and graphs revealed the overall patterns and guided the model selection for extreme value analyses. The extreme value models characterized the patterns of area burned, fire duration, and their association more completely. The patterns were assessed for all the fires combined, and then individually by period, ecoregion, cause, and extremal degree. After selecting appropriate distributions and fitting the corresponding models, the return levels of wildland fires by area burned and fire duration were predicted for all the wildland fires combined in Mississippi. This study reports the first application of bivariate extreme value statistics in analyzing wildland fires. Several similar studies have been published in the past, but all of them used univariate extreme value statistics only. The present study extends the above analyses by considering the joint distribution pattern of fire area and duration within the framework of bivariate extreme value models. Furthermore, the unique dataset used in this study allows for a thorough application of extreme value theory. The dataset contained all wildland fires in Mississippi as a typical southern state. The dataset for Mississippi allowed us to thoroughly explore the range of techniques related to descriptive and extreme value statistics in the analyses. As wildland fires are a universal phenomenon, these techniques employed in the present study can be applied to analogous datasets for other regions or countries. Finally, the results from these analyses of fire patterns and extremal behavior can provide assistances to forest planning and management. Wildland fire management is subject to numerous uncertainties. Forecasting large fires through return level analyses can provide valuable inputs for fire suppression and budget planning.
Publications
- Type:
Conference Papers and Presentations
Status:
Published
Year Published:
2013
Citation:
Sun, C. 2013. Bivariate extreme value modeling of wildland fire area and duration. Forest Science 59(6): 649-660.
|
Progress 01/01/13 to 09/30/13
Outputs
Target Audience: Government agencies; forest landowners; the general public Changes/Problems:
Nothing Reported
What opportunities for training and professional development has the project provided?
Nothing Reported
How have the results been disseminated to communities of interest? Major study findings have been presented in a professional meeting and published on a refereed journal. Presentations: Sun, C. 2012. Dependence between area and duration of wildland fires. The 42th Annual Southern Forest Economics Workshop. March 20, 2012. Charlotte, NC. What do you plan to do during the next reporting period to accomplish the goals? Work on the project further and finish this five-year project next year.
Impacts
What was accomplished under these goals?
While wildland fires are essential and beneficial to many ecosystems, catastrophic fires can generate severe economic and environmental consequences. Fires are a multivariate phenomenon that can be characterized not only by area burned as often analyzed in the literature, but also by other less scrutinized features such as fire duration. In analyzing the statistical properties of large wildland fires, there has been a critical need to consider the dependence between various features of wildland fires. In this study, two aspects of large wildland fires (i.e., area burned and fire duration) were assessed by bivariate extreme value statistics. A dataset of 64,474 fire records from Mississippi between January 1991 and December 2007 was used in the assessment. The descriptive statistics revealed that the average fire size was 5.8 hectares and the average fire duration was 100.3 minutes, while the largest fires burned 2,107 hectares or lasted 1,510 minutes. Large fires occurred mainly between February and May, in the southeastern region, or started as debris burning and incendiary fires in Mississippi. In comparison, area burned showed more extremal properties than fire duration, as supported by the scatterplot, histogram, and spatial density map. Overall, a positive but moderate association existed between area and duration of wildland fires in Mississippi over the study period. The descriptive statistics also indicated that an application of extreme value models would be appropriate. The univariate extreme value analyses concluded that the generalized Pareto distribution with a heavy tail characterized both the area burned and fire duration well. The threshold chosen was 40 hectares and 240 minutes for all the fires combined in Mississippi, while the threshold values varied slightly by type. The bivariate extreme value analyses found that a number of distributions fitted the data well, and the asymmetric negative logistic distribution provided the best fit. The Pickands dependence measure was 0.391 for all the fires combined and had similar values across the various types. The positive and moderate association between area burned and fire duration was accurately measured through the bivariate extreme models. The estimated parameter values were used to predict return levels associated with different return periods and risk. Incorporating the dependence between area and duration into the forecast generated more conservative estimates of return levels than these from the univariate analyses. Thus, fire budgets based on univariate analyses of wildland fires can become inadequate in case of catastrophic fires. The study outcomes have a number of implications for forest planning and management activities. In spite of various benefits, wildland fires pose threats to properties and the safety of human life, depending on natural resources available and fire policies adopted in a region or country. Information about the degree of extremal behavior and probability of catastrophic fires, as measured by multivariate extreme value statistics, can aid private and public landowners in their forest management activities. Risk assessments that take the probability of catastrophic fires into account may help landowners decide whether to engage in an investment opportunity. In addition, fire suppression costs have been rapidly rising over the past several decades as large wildfires have occurred more frequently. A more accurate prediction of area burned and fire duration of catastrophic wildland fires will help policy makers, landowners, and communities in making sound decisions related to fire budgets and facility investments.
Publications
- Type:
Journal Articles
Status:
Awaiting Publication
Year Published:
2013
Citation:
Sun, C. 2013. Bivariate extreme value modeling of wildland fire area and duration. Forest Science. Forthcoming in December 2013.
|
Progress 01/01/12 to 12/31/12
Outputs
OUTPUTS: Wildland fires can cause enormous damages to environment and economy, so there is an urgent and ongoing need to fully comprehend the behavior of extremal wildfires. In this study, the patterns of wildland fires in the southern United States were assessed using the fire records in Mississippi. Descriptive statistics and graphs were used to reveal the patterns of these fires. Their extremal behavior was further evaluated with the generalized Pareto distribution and the peaks-over-threshold model. The analyses were conducted for all the fires combined, and also individually by year, period, ecoregion, and fire cause. The descriptive statistics revealed that wildland fires in Mississippi have been stable for the past 17 years. The presence of extremal behavior was clear with large fires contributing heavily to the area burned in Mississippi. The conclusions from the extreme value analyses were that the generalized Pareto distribution characterized the extremal behavior of forest fires in Mississippi well, and the peaks-over-threshold models could be used to predict future wildland fire sizes. These outcomes have vital implications for forest planning and management activities. Even for regions like Mississippi where wildland fires may appear less intense, fires still pose a huge threat to natural resources, economy, and human beings. The degree of extremal behavior and probability of catastrophic fires, as measured by extreme value statistics, should aid thousands of private and public landowners regarding their forest investments. Risk assessments that take into account the probability of catastrophic fires can help landowners decide whether to engage in an investment opportunity. In addition, fire suppression costs have been rising fast over the past several decades as the intensity of large wildfires has increased. A more accurate prediction of the size of catastrophic wildland fires will help landowners, communities, and policy makers in making sound decisions related to fire budgets and facility investments. PARTICIPANTS: Dr. Changyou Sun, Associate Professor, Department of Forestry, Mississippi State University, Mississippi State, MS 39762; Phone: (662) 325-7271; E-mail: csun@cfr.msstate.edu. Mr. Branden Tolver, Forester, Sam Houston National Forest, TX 77358. TARGET AUDIENCES: Government agencies; forest landowners; the general public PROJECT MODIFICATIONS: Not relevant to this project.
Impacts
This is the first application of extreme value statistics in analyzing wildland fires in the southern United States. The unique dataset allows for a more thorough application of extreme value models. The findings can provide great benefits to forest planning and management. Forecasts of large fires also can be a valuable input for fire suppression and budget planning. As wildland fires are a worldwide phenomenon, these methods can be applied to similar datasets for other regions or countries.
Publications
- Sun, C., and B. Tolver. 2012. Assessing administrative laws for forestry prescribed burning in the southern United States: a management-based regulation approach. International Forestry Review 14(3):337-348.
- Sun, C., and B. Tolver. 2012. Assessing the distribution patterns of wildfire sizes in Mississippi, USA. International Journal of Wildland Fire 21(5):510-520.
|
Progress 01/01/11 to 12/31/11
Outputs
OUTPUTS: Wildland fires can produce dramatic ecological and economic impacts. The objective of this study was to analyze the temporal and spatial distribution patterns of wildland fires using 64 474 fire records in Mississippi, collected between 1991 and 2007. The methodology employed was descriptive statistics and extreme value statistics. The analyses were conducted for all the fires combined, and also by year, period, ecoregion, and cause separately. Wildland fires occurred most frequently between February and May, with more than half of all the fires occurring in that period. The ecoregion of outer coastal plain mixed forest province had more fire occurrences, while the ecoregion of southeastern mixed forestry province had more catastrophic fires. By fire cause, debris burning and incendiary combined were responsible for 89.6% of the area burned. The top 10% of the largest fires burned 58.8% of the total area. The extreme value statistics revealed that wildfires in Mississippi displayed a generalized Pareto distribution. Based on predictions from the peaks-over-threshold models, the largest wildland fire in Mississippi within the next 10 years could burn 2171 hectares. These outcomes can help landowners and government agencies make better decisions related to forest investments, fire suppression, and budget planning. PARTICIPANTS: Dr. Changyou Sun, Associate Professor, Department of Forestry, Mississippi State University, Mississippi State, MS 39762; Phone: (662) 325-7271; E-mail: csun@cfr.msstate.edu. Mr. Branden Tolver, graduate research assistant, Department of Forestry, Mississippi State University, Mississippi State, MS 39762. TARGET AUDIENCES: Government agencies; forest landowners; the general public PROJECT MODIFICATIONS: Not relevant to this project.
Impacts
This is the first application of extreme value statistics in analyzing wildland fires in the southern United States. The unique dataset allows for a more thorough application of extreme value models. The findings can provide great benefits to forest planning and management. Forecasts of large fires also can be a valuable input for fire suppression and budget planning. As wildland fires are a worldwide phenomenon, these methods can be applied to similar datasets for other regions or countries.
Publications
- Sun, C. and B. Tolver. 2011. Assessing the extremal behavior of wildland fires in Mississippi, USA. International Journal of Wildland Fire (in press).
|
Progress 01/01/10 to 12/31/10
Outputs
OUTPUTS: Many reforms in administrative laws and regulations have been made for forestry prescribed burning in the southern United States. Florida has been leading other states across the region in these reforms. Various regulations have been adopted in the South to relieve the burden of liability for private landowners and ensure the public that every measure has been taken to reduce risk. A key feature of these regulations is that they focus on the production and management process of prescribed burning to identify and reduce risk. In this project, we examine administrative laws for forestry prescribed burning in the South within the theoretical framework of management-based regulation (MBR). The MBR approach has been proven effective in comprehending various environmental regulations. The origin, condition, advantage, and design choices of MBR are analyzed to elaborate this theoretical framework. Major production stages in conducting prescribed burning and associated risk locations are identified, and the MBR approach is applied to design a system of regulations on forestry prescribed burning. Current administrative laws in the South are examined through a review of the evolution of Florida's prescribed burning laws and a comparison of regulations in the 13 southern States (i.e., Alabama, Arkansas, Florida, Georgia, Kentucky, Louisiana, Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee, Texas, and Virginia). Furthermore, how the theoretical analysis fits with existing laws is assessed. PARTICIPANTS: Dr. Changyou Sun, Associate Professor, Department of Forestry, Mississippi State University, Mississippi State, MS 39762; Phone: (662) 325-7271; E-mail: csun@cfr.msstate.edu. Mr. Branden Tolver, graduate research assistant, Department of Forestry, Mississippi State University, Mississippi State, MS 39762. TARGET AUDIENCES: Government agencies; forest landowners; the general public PROJECT MODIFICATIONS: Not relevant to this project.
Impacts
This analytical assessment of prescribed burning regulations can greatly assist the public in comprehending the rationale and trend of administrative law reform for prescribed burning. This assessment will also benefit private landowners in designing or evaluating new practices for using fires on forest lands with legitimate purposes.
Publications
- Tolver, B.T. 2010. Fire in southern U.S.: administrative laws and regulations in the southeast and wildfire distributions in Mississippi. Master`s thesis, Mississippi State University, Mississippi State, MS. August 2010. 90 p. Advisor: C. Sun.
|
Progress 01/01/09 to 12/31/09
Outputs
OUTPUTS: The southeastern United States has experienced many reforms in terms of state administrative laws and regulations for prescribed burning. Management-based regulations have been implemented in recent years to help relieve the burden of liability for private landowners. There is now a need for a comprehensive examination of current administrative laws and regulations of prescribed burning throughout the southeastern region in order to better understand the prescribed fire policy of today. In this project, we first examine case study examples from both Florida and Mississippi in order to outline the evolution of administrative regulations in each state over the past 30 years. We also include a 13 state comparison of current administrative laws and regulations for the southeastern U.S within the theoretical framework of management-based regulation. This new approach has been proven effective in other environmental fields and is in the early stages of origination in prescribed burning. In addition, the linkage between administrative and statutory laws is discussed to clarify the origination of the various components of each administrative law and or regulation. Management-Based Regulation's origin, its processes, and examples are used to help better understand the trend with prescribed burning. Providing a review of current regulations and administrative laws aides the general public who genuinely needs to know what is being done to protect their interests. It also benefits private land managers by helping them to better understand the trend of administrative law reform and its future. We had two presentations. 1. Sun, C. 2009. Wildfire regimes and management-based regulation of prescribed fire. MSU Forest and Wildlife Research Application Conference. September 30, 2009. Jackson, MS. 2. Tolver, B.T. and C. Sun. 2009. A management-based regulation approach to prescribed fire: Review of administrative law and regulation in the southeastern United States. 39th Annual Southern Forest Economics Workshop 2009. March 14, 2009. Chapel Hill, NA. PARTICIPANTS: Dr. Changyou Sun, Associate Professor, Department of Forestry, Mississippi State University, Mississippi State, MS 39762; Phone: (662) 325-7271; E-mail: csun@cfr.msstate.edu. Mr. Branden Tolver, graduate research assistant, Department of Forestry, Mississippi State University, Mississippi State, MS 39762. TARGET AUDIENCES: Government agencies; forest landowners; the general public. PROJECT MODIFICATIONS: Not relevant to this project.
Impacts
No outcomes/impacts yet in the first year of this five-year project.
Publications
- No publications reported this period
|
|