The use of subjective probability assessments to predict forest fire occurrence

1976 ◽  
Vol 6 (3) ◽  
pp. 348-356 ◽  
Author(s):  
A. A. Cunningham ◽  
D. L. Martell

This paper addresses the problem of predicting forest fire occurrence. A simple methodology is developed to elicit information, from experienced fire managers, for deriving subjective probability assessments concerning the number of fires that will be reported in their districts each day. The approach is based upon the assumption that such individuals can identify and classify similar fire environments with satisfactory consistency. Using the concept of subjective probability, a methodology is developed for combining classified experience with a decision maker's initial assessment concerning assessor behavior. A scoring rule is used to measure the accuracy of the assessments. The results of an experiment which was conducted in northern Ontario during the 1973 fire season are presented.

1987 ◽  
Vol 17 (5) ◽  
pp. 394-401 ◽  
Author(s):  
D. L. Martell ◽  
S. Otukol ◽  
B. J. Stocks

The authors describe the development of a procedure that can be used to predict daily people-caused forest fire occurrence in the Northern Region of the province of Ontario. The procedure is based on the use of logistic regression analysis techniques to predict the probability of a fire day and the assumption that a Poisson probability distribution can be used to model daily people-caused forest fire occurrence. The results of a field test that was conducted during the summer portion of the 1984 fire season indicate the procedure works well during relatively wet periods.


2017 ◽  
Author(s):  
Mohamed Elhag ◽  
Slivena Boteva

Abstract. The Fire Weather Index (FWI) module was tested under the Mediterranean- type conditions of Crete (Greece) for the two fire seasons 2008–2009. High correlations were found between the Fine Fuel Moisture Code (FFMC) and the Duff Moisture Code (DMC. The Drought Code (DC) was insignificantly correlated with the soil moisture content. No significant correlation was found between the area burned by wildfires and any component of the FWI system during the studied period, unlike fire occurrence with which most of the components were highly correlated. Meanwhile, the Keetch-Byram Drought Index (KBDI) of the American Forest Fire Danger Rating System (NFFDRS) was also examined under the same conditions. It provided a useful means of monitoring general wetting and drying cycles, but is inadequate for indicating daily fire danger throughout the fire season in our region. Weak correlations between the KBDI- the fire occurrence and the area burned were found for the two fire seasons studied-2008–2009. Correlations between the KBDI and litter, duff and soil did not give statistically sound results. On the contrary, the KBDI seemed to predict with high accuracy the moisture content of three annual plants (Piplatherum miliaceum, Parietaria diffusa, Avena sterillis) with a shallow rooting system of Pinus halepensis forest understory in the region. This indicated that the index was adequate, to a certain extent, to represent the upper soil layers' water status, while it is unsuitable to predict needles moisture content of Pinus halepensis, which has a deep rooting system.


Forests ◽  
2020 ◽  
Vol 12 (1) ◽  
pp. 5
Author(s):  
Slobodan Milanović ◽  
Nenad Marković ◽  
Dragan Pamučar ◽  
Ljubomir Gigović ◽  
Pavle Kostić ◽  
...  

Forest fire risk has increased globally during the previous decades. The Mediterranean region is traditionally the most at risk in Europe, but continental countries like Serbia have experienced significant economic and ecological losses due to forest fires. To prevent damage to forests and infrastructure, alongside other societal losses, it is necessary to create an effective protection system against fire, which minimizes the harmful effects. Forest fire probability mapping, as one of the basic tools in risk management, allows the allocation of resources for fire suppression, within a fire season, from zones with a lower risk to those under higher threat. Logistic regression (LR) has been used as a standard procedure in forest fire probability mapping, but in the last decade, machine learning methods such as fandom forest (RF) have become more frequent. The main goals in this study were to (i) determine the main explanatory variables for forest fire occurrence for both models, LR and RF, and (ii) map the probability of forest fire occurrence in Eastern Serbia based on LR and RF. The most important variable was drought code, followed by different anthropogenic features depending on the type of the model. The RF models demonstrated better overall predictive ability than LR models. The map produced may increase firefighting efficiency due to the early detection of forest fire and enable resources to be allocated in the eastern part of Serbia, which covers more than one-third of the country’s area.


2014 ◽  
Vol 14 (6) ◽  
pp. 1477-1490 ◽  
Author(s):  
A. Venäläinen ◽  
N. Korhonen ◽  
O. Hyvärinen ◽  
N. Koutsias ◽  
F. Xystrakis ◽  
...  

Abstract. Understanding how fire weather danger indices changed in the past and how such changes affected forest fire activity is important in a changing climate. We used the Canadian Fire Weather Index (FWI), calculated from two reanalysis data sets, ERA-40 and ERA Interim, to examine the temporal variation of forest fire danger in Europe in 1960–2012. Additionally, we used national forest fire statistics from Greece, Spain and Finland to examine the relationship between fire danger and fires. There is no obvious trend in fire danger for the time period covered by ERA-40 (1960–1999), whereas for the period 1980–2012 covered by ERA Interim, the mean FWI shows an increasing trend for southern and eastern Europe which is significant at the 99% confidence level. The cross correlations calculated at the national level in Greece, Spain and Finland between total area burned and mean FWI of the current season is of the order of 0.6, demonstrating the extent to which the current fire-season weather can explain forest fires. To summarize, fire risk is multifaceted, and while climate is a major determinant, other factors can contribute to it, either positively or negatively.


FLORESTA ◽  
2002 ◽  
Vol 32 (2) ◽  
Author(s):  
Ronaldo Viana Soares ◽  
Juliana Ferreira Santos

O conhecimento do perfil dos incêndios florestais é muito importante para o planejamento do controle dos mesmos. O objetivo deste trabalho foi estabelecer o perfil dos incêndios florestais no país através de dados coletados, em áreas protegidas, no período de 1994 a 1997, através de formulários preenchidos por empresas e instituições florestais. Foram registrados e informados 1.957 incêndios e apesar deste número não representar a totalidade dos incêndios ocorridos no período estudado, constituiu-se numa base confiável para se conhecer as principais características dos incêndios. Os resultados mostraram que a área média atingida por incêndio no período analisado foi de aproximadamente 135 ha, sendo Minas Gerais o estado líder, tanto em número de incêndios informados (62,7% do total) como em área queimada (25,2%). O grupo Incendiários foi a principal causa dos incêndios, com 56,6% das ocorrências, vindo a seguir as Queimas para limpeza com 22,1%. Com relação à área queimada o grupo Queimas para limpeza , com 74,1% da superfície atingida, foi a principal causa, ficando o grupo Incendiários em segundo lugar com 19,8%. A principal estação de incêndios no país se estende de julho a novembro, quando ocorreram 79,2% dos incêndios, correspondendo a 98,6% da área atingida. O maior número de incêndios (39,7% das ocorrências) foi registrado em Outro tipo de vegetação, que inclui cerrado, capoeira e campo. Com relação à área atingida, entretanto, 92,5% foi registrada em Florestas Nativas. Quanto à distribuição dos incêndios através das classes de tamanho, 23,9% foi enquadrado na classe I ( 0,1 ha). É importante ressaltar que quanto maior a eficiência no combate aos incêndios, maior é a concentração dos mesmos na classe I. Apesar de corresponder a apenas 2,4% das ocorrências, os incêndios da classe V ( 200,0 ha) foram responsáveis por 94,5% da área queimada. FOREST FIRE STATISTICS IN BRAZIL FROM 1994 TO 1997 Abstract Forest fire statistics knowledge is an important tool for fire control planning. The objective of this research was to collect information on forest fire occurrence in Brazilian protected areas in the period of 1994 to 1997. The analyzed variables were the number of fires and burned areas per state of the federation, monthly distribution, probable causes, affected vegetation, size class distribution, and average burned area per fire. Results showed that the average burned area per fire was approximately 135 ha and Minas Gerais ranked first, both in number of registered fires (62.7%) and burned surface (25.2%). Incendiary, with 56.6% of the occurrences was the leading cause, followed by debris burning with 22.1%. However, as for the affected area, Debris burning was the leading cause (74.1%), followed by Incendiary (19.8%). The fire season extends from July to November, when 79.2% of the fires occurred, corresponding to 98.6% of the burned surface. Miscellaneous, that includes savanna, secondary growth forest, and grassland were the most affected vegetation type (39.7% of the occurrences). In relation to the burned surface, Native Forest (92.5%) ranked first. The distribution of the registered fires through the size classes presented 23.9% of the occurrences in Class I ( 0.1 ha), whereas 94.5% of the burned area were result of Class V ( 200 ha) fires. Size Class II (0.1 to 4.0 ha), with 49.1% of the occurrences, ranked first in number of registered fires during the analyzed period.


2008 ◽  
Vol 77 (1) ◽  
Author(s):  
Álvaro Corral ◽  
Luciano Telesca ◽  
Rosa Lasaponara
Keyword(s):  

Author(s):  
David MJS Bowman ◽  
Grant J Williamson

Fire risk can be defined as the probability that a fire will spread. Reliable monitoring of fire risk is essential for effective landscape management. Compilation of fire risk records enable identification of seasonal and inter-annual patterns and provide a baseline to evaluate the trajectories in response to climate change. Typically, fire risk is estimated from meteorological data. In regions with sparse meteorological station coverage environmental proxies provide important additional data stream for estimating past and current fire risk. Here we use a 60-year record of daily flows from two rivers (Franklin and Davey) in the remote Tasmanian Wilderness World Heritage Area (TWWHA) to characterize seasonal patterns in fire risk in temperate Eucalyptus and rainforests. We show that river flows are strongly related to landscape soil moisture estimates derived from down-scaled re-analysis of meteorological data available since 1990. To identify river flow thresholds where forests are likely to burn, we relate river flows to known forest fires that have occurred in the previously defined ecohydrological domains that surround the Franklin and Davey catchments. Our analysis shows that the fire season in the TWWHA is centered on February (70% of all years below the median threshold), with shoulders on December-January and March. Since 1954 forest fire can occur in at least one month for all but four summers in the ecohydrological domain that includes the Franklin catchment, and since 1964 fire fires could occur in at least one month in every summer in the ecohydrological domain that includes the Davey catchment. Our analysis shows that mangers can use river flows as a simple index that provide a landscape-scale forest fire risk in the TWWHA.


2020 ◽  
Vol 2020 (1) ◽  
Author(s):  
M. Dennekamp ◽  
K. Martin ◽  
S. Coutts ◽  
J. Choi ◽  
A. Hinwood ◽  
...  

Sign in / Sign up

Export Citation Format

Share Document