Forest Fire Danger Rating System Research Articles

The Global Forest Fire Emissions Prediction System version 1.0

Abstract. The Global Forest Fire Emissions Prediction System (GFFEPS) is a model that estimates biomass burning in near-real time for global air quality forecasting. The model uses a bottom-up approach, based on remotely sensed hotspot locations, and global databases linking burned area per hotspot to ecosystem-type classification at a 1 km resolution. Unlike other global fire emissions models, GFFEPS provides dynamic estimates of fuel consumption, fire behaviour and fire growth based on the Canadian Forest Fire Danger Rating System, plant phenology as calculated from daily global weather and burned-area estimates using near-real-time Visible Infrared Imaging Radiometer Suite (VIIRS) satellite-detected hotspots and historical burned-area statistics. Combining forecasts of daily fire weather and hourly meteorological conditions with a global land classification, GFFEPS produces fuel consumption and emission predictions in 3 h time steps (in contrast to non-dynamic models that use fixed consumption rates and require a collection of burned area to make post-burn estimates of emissions). GFFEPS has been designed for use in operational forecasting applications as well as historical simulations for which data are available. A study was conducted showing GFFEPS predictions through a 6-year period (2015–2020). Regional annual total smoke emissions, burned area and total fuel consumption per unit area as predicted by GFFEPS were generated to assess model performance over multiple years and regions. The model's fuel consumption per unit area results clearly distinguished regions dominated by grassland (Africa) from those dominated by forests (boreal regions) and showed high variability in regions affected by El Niño and deforestation. GFFEPS carbon emissions and burned area were then compared to other global wildfire emissions models, including the Global Fire Assimilation System (GFAS), the Global Fire Emissions Database (GFED4.1s) and the Fire INventory from NCAR (FINN 1.5 and 2.5). GFFEPS estimated values lower than GFAS and GFED (80 % and 74 %) and had values similar to FINN 1.5 (97 %). This was largely due to the impact of fuel moisture on consumption rates as captured by the dynamic weather modelling. Model evaluation efforts to date are described – an ongoing effort is underway to further validate the model, with further developments and improvements expected in the future.

An adaptable dead fuel moisture model for various fuel types and temporal scales tailored for wildfire danger assessment

Estimating the Dead Fuel Moisture Content (DFMC) is crucial in wildfire risk management, representing a key component in forest fire danger rating systems and wildfire simulation models. DFMC fluctuates sub-daily and spatially, influenced by local weather and fuel characteristics. This necessitates models that provide sub-daily fuel moisture conditions for improving wildfire risk management. Many forest fire danger rating systems typically rely on daily fuel moisture models that overlook local fuel characteristics, with consequent impact on wildfire management. The semi-empirical parametric DFMC model proposed addresses these issues by providing hourly dead fuel moisture dynamics, with specific parameters to consider local fuel characteristics. A calibration framework is proposed by adopting Particle Swarm Optimization-type algorithm. In the present study, the calibration framework has been tested by using hourly 10-hours fuel sticks measurements. Implementing this model in forest fire danger rating systems would enhance detail in forest fire danger conditions, advancing wildfire risk management.

Development of national post-fire restoration system to assess net GHG impacts and salvage biomass availability

In light of the recent unprecedented wildfires in Canada and the potential for increasing burned areas in the future, there is a need to explore post-fire salvage harvest and restoration and the implications for greenhouse gas (GHG) emissions. Salvage logging and replanting initiatives offer a potential solution by regrowing forests more quickly while meeting societal demands for wood and bioenergy.This study presents a comprehensive modeling framework to estimate post-fire salvage biomass and net GHG emissions relative to a ‘do-nothing’ baseline for all of Canada's harvest-eligible forests. Forest ecosystem carbon emissions and removals were modeled at 1-ha spatial resolution for Canadian forests using the Generic Carbon Budget Model (GCBM) from 1990 to 2070 using several forest inventory data sources with future harvest and wildfires.Building upon previous research, our work integrated the Canadian Forest Fire Danger Rating System fire intensity to estimate fire severity of future wildfires. For 2024 to 2070, we assessed the changes in ecosystem carbon, emissions from harvested wood products, and substitution benefits from avoided emissions-intensive materials, relative to a forward-looking baseline. Our prototype system provides a comprehensive framework, configuration files, links to datasets to quantify the net GHG of post-fire restoration, and sample results for validation.•Developed spatially explicit forest carbon modeling system for all of Canada's forests.•Assessed the net GHG reduction from post-fire restoration.•Used system approach to consider forests, wood products and substitution benefits.

Study on Forest Fire Risk Analysis Based on Drought Scenarios

The number of forest fires is increasing worldwide owing to climate change. Forest fires linked to drought are becoming more frequent and severe than ever before. This causes massive damage. However, instead of quantification, studies have been conducted on the relationship between drought and forest fires. Therefore, this study quantitatively analyzed the increase in forest fire risk based on a drought scenario. We defined a 30-year frequency drought scenario for the Gyeongsangbuk-do Province, where the most forest fire damage occurred in 2022, and analyzed the risk of forest fires by calculating the Fire Weather Index (FWI) used in the Canadian Forest Fire Danger Rating System (CFFDRS). The precipitation in the drought scenario was 581 mm/year that was equivalent to 52% of the average annual precipitation. We noted that the drought Fire Weather Index (DFWI) was approximately 1.6 times greater than the average Fire Weather Index (AFWI). We quantitatively confirmed the extent to which the risk of forest fire increased under the influence of drought. Therefore, measures should be established not only for drought but also for forest fires when a drought occurs.

Environmental Forest Fire Danger Rating Systems and Indices around the Globe: A Review

The objective of the present review is to analyze and evaluate the most used and well-performing environmental forest fire danger rating systems and indices globally, aiming to the creation of an integrated forest fire danger system for Greece. The analysis emphasizes the core input parameters that have been associated with forest fire danger (i.e., weather, vegetation, topography, and hydrology) and the computational procedure of each system index as well as the categorization of the output values. Online search engines such as Scopus, Google Scholar, WorldWideScience, ScienceDirect, and ResearchGate were used in the search for relevant literature published in scientific journals, manuals, and reports. The retrieved studies were classified and reviewed. Studies were selected for analytically describing the calculation process related to forest fire danger ignition and not being strictly geographically bound. A total of 210 studies were included in the current review, describing 63 forest fire danger systems and indices. These were analyzed and evaluated based on a scoring system. Overall, the top-rated indices were the: Nesterov’s index, Sharples’ index, Keetch and Byram’s drought index, Telicyn logarithmic, and vapor pressure deficit, and the 3rd and the 4th also proved to be the most accurate for fire-prone regions. Remote sensing indices also proved to be promising in forest fire danger estimation.

Evaluation of new methods for drought estimation in the Canadian Forest Fire Danger Rating System

Background Canadian fire management agencies track drought conditions using the Drought Code (DC) in the Canadian Forest Fire Danger Rating System. The DC represents deep organic layer moisture.Aims To determine if electronic soil moisture probes and land surface model estimates of soil moisture content can be used to supplement and/or improve our understanding of drought in fire danger rating.Methods We carried out field studies in the provinces of Alberta and Ontario. We installed in situ soil moisture probes at two different depths in seven forest plots, from the surface through the organic layers, and in some cases into the mineral soil.Results Our results indicated that the simple DC model predicted the moisture content of the deeper organic layers (10–18 cm depths) well, even compared with the more sophisticated land surface model.Conclusions Electronic moisture probes can be used to supplement the DC. Land surface model estimates of moisture content consistently underpredicted organic layer moisture content.Implications Calibration and validation of the land surface model to organic soils in addition to mineral soils is necessary for future use in fire danger prediction.

Predicting Forest Fire Danger Using Fuel Characteristics of Forest

The fuel characteristics of a forest are important factors in designing a forest fire danger rating system and must be analyzed to understand forest fire behavior. Therefore, in this study, forest fire intensity was predicted depending on the amount of fuel in the forests in the Yeongdong and Yeongseo regions of Gangwon-do from November 2020 to December 2021. The possibility of forest fire occurrence, from surface fire to crown fire, was analyzed depending on the fuel conditions. Regional differences were found in the amount of fuel and the fuel structure of the surface layer in the forest. The intensity and type of forest fire changed according to these differences. The Yeongdong area is more susceptible to crown fires and more intense of surface fire than the Yeongseo area. These results show the need to prevent early forest fires by learning the characteristics of forests in different regions and analyzing the intensity and type of forest fire predicted.

Assessment of lake-level fluctuation as an indicator of fire activity in boreal Canada

Fuel moisture content is a highly variable property that dictates the ignition and combustion potentials of wildfire fuels. The Canadian Fire Weather Index (FWI) System—a subset of the established Canadian Forest Fire Danger Rating System—relies solely on weather observations to track fuel moisture and compute indices of fire potential. Ancillary environmental measures, such as lake levels in non-regulated lake systems, may integrate climatic and landscape factors that better reflect environmental moisture deficits and thereby complement the FWI System. Accordingly, we investigated whether inter-annual variation in lake levels has potential to serve as an indicator of regional fire activity using a quasi-binomial regression approach. This study combines annual measurements of fire activity, lake level, and other climatic variables from the Boreal Shield and Boreal Plain ecozones of Canada from 1980 to 2019 to determine whether variation in regional area burned can be explained by lake-level fluctuation, and whether this explanatory power differs by ecozone. Results indicated that lake-level fluctuation is a significant predictor of regional area burned for many lakes (28–48%). It often performs as well as or better than traditional fire-weather metrics (such as temperature and precipitation), and this relationship is stronger on the Boreal Shield compared to the Boreal Plain. Lake-level fluctuation provides information on fuel availability unaccounted for by traditional fire-weather metrics, suggesting that this variable could be added to the suite of environmental indicators of regional fire activity in Canada.

Relationship between Forest Fuel Moisture Contents and Weather Factors During the Forest Fires Danger Season in Chuncheon, Gangwon

The amount of moisture contained in fallen leaves in the forest is closely related to the possibility of the ignition and spread of a forest fire. This moisture content is an important fundamental component of the forest fire danger rating system, as well as an essential parameter in forest fire management. This study examined the relationship between the moisture content of fuel in the forests and weather information obtained from HOBO data loggers installed in forests during the forest fire seasons of 2018 and 2020. The moisture content of the forest fuel was divided into two slopes (south and north), and the fallen leaves used were broad-leaved (Prunus serrulate and Quercus dentata) and coniferous (Pinus densiflora and Pinus koraiensis). The temperature (mean, minimum, and maximum) and relative humidity (mean and minimum) data obtained from the weather measuring devices installed in the forests were used to correlate weather factors with fuel moisture content. During the spring forest fire period, the temperature (mean, minimum, and maximum) and relative humidity (mean and minimum) of each slope and tree species were significant, while the average temperature and relative humidity were significant during the autumn. According to the correlations between the fuel moisture content and weather, factors with high significance were selected, and multiple regression analyses of the slope and species were found to be significant at 1 percent. The results of this study are expected to be used as primary data for calculating the fuel index of the forest fire forecasting system by accounting for the effects of weather factors on the moisture content of the fuel.

Interaction of elevation and climate change on fire weather risk

Most wildfire studies are regional to global in scale; however, many of the values of interest and the weather are local phenomena that may give rise to large spatial variability in risk. We assessed the interaction of elevation and climate on fire weather for the Penticton Creek watershed in southwestern Canada for historic weather, and five climate change scenarios. Daily temperature and precipitation records spanning 100 years were generated using the LARS-WG5 weather generator and were used to calculate the fire weather indices of the Canadian Forest Fire Danger Rating System. Fire season length, restricted activity season, and fire season severity are all projected to increase by the 2050s and in some scenarios to increase further by the 2080s. Low and mid-elevations had substantially worsening risks, whereas at the highest elevations, risks were mitigated by the continuation of the snowpack. Increasing temperatures lengthened the fire season while decreasing (increasing) precipitation exacerbated (ameliorated) the intensity of the fire risk. These results indicate more variable climate change effects than in the literature. Over 24 million km2globally of forestland have a similar or higher degree of slope, highlighting the need for detailed topographic integration in fire studies to meet the needs of local planners.

A case-study of wildland fire management knowledge exchange: the barriers and facilitators in the development and integration of the Canadian Forest Fire Danger Rating System in Ontario, Canada

Background Among the most successful examples of Knowledge Exchange (KE) between researchers and practitioners in Canadian wildland fire management is the development and integration of the Canadian Forest Fire Danger Rating System (CFFDRS) into operational use. Aims Our aim was to identify key factors for this success. Methods Through a case study, we investigated historical KE of two CFFDRS components in Ontario, Canada. We held semi-structured interviews with principal Canadian Forest Service researchers and Ontario fire management practitioners active in development and implementation of CFFDRS from the late 1960s to 2010s. Key results The importance of both formal and informal facilitators to support KE was emphasised. Conclusion Participants were most likely to associate successful implementation with informal facilitators such as personal relationships, shared field-based experiences, and opportunities for dialogue between researchers and practitioners. Critical to success were the credibility and soft skills of the knowledge brokers, early engagement, and consideration of training needs for end users in the design of products. Implications This identification of factors that facilitated or hindered the development and implementation of CFFDRS can enhance the impact of research that will help wildland fire management deal with its present and future challenges.

Predicting Forest Fire Using Remote Sensing Data And Machine Learning

Over the last few decades, deforestation and climate change have caused increasing number of forest fires. In Southeast Asia, Indonesia has been the most affected country by tropical peatland forest fires. These fires have a significant impact on the climate resulting in extensive health, social and economic issues. Existing forest fire prediction systems, such as the Canadian Forest Fire Danger Rating System, are based on handcrafted features and require installation and maintenance of expensive instruments on the ground, which can be a challenge for developing countries such as Indonesia. We propose a novel, cost-effective, machine-learning based approach that uses remote sensing data to predict forest fires in Indonesia. Our prediction model achieves more than 0.81 area under the receiver operator characteristic (ROC) curve, performing significantly better than the baseline approach which never exceeds 0.70 area under ROC curve on the same tasks. Our model's performance remained above 0.81 area under ROC curve even when evaluated with reduced data. The results support our claim that machine-learning based approaches can lead to reliable and cost-effective forest fire prediction systems.

Characterisation of initial fire weather conditions for large spring wildfires in Alberta, Canada

We evaluated surface and 500-hPa synoptic weather patterns, and fire weather indices from the Canadian Forest Fire Danger Rating System for 80 large wildfires during 1990–2019 in Alberta that started in May and grew to over 1000 ha. Spread days were identified during the first 4 days of wildfire activity. We observed two distinct synoptic weather patterns on these days. Pre-frontal and frontal passage activity was the predominant feature associated with 48% of the calendar spread days. Strong south–south-east winds from a surface high centred east of Alberta (west of Hudson Bay) and supported by an upper ridge, and a surface low located south-west of the ridge occurred on 26% of the calendar spread days. Surface analysis indicates the spring wildfire season in Alberta is driven by very high to extreme Initial Spread Index, a rating of the expected wildfire rate of spread based on Fine Fuel Moisture Code and wind. Very high to extreme values of Buildup Index, a rating of the amount of fuel available for consumption, are not a prerequisite for large wildfires in May. For Alberta, this means large wildfires in May can occur after only a few days of dry, windy weather.

Developing a New Hourly Forest Fire Risk Index Based on Catboost in South Korea

Forest fires can cause enormous damage, such as deforestation and environmental pollution, even with a single occurrence. It takes a lot of effort and long time to restore areas damaged by wildfires. Therefore, it is crucial to know the forest fire risk of a region to appropriately prepare and respond to such disastrous events. The purpose of this study is to develop an hourly forest fire risk index (HFRI) with 1 km spatial resolution using accessibility, fuel, time, and weather factors based on Catboost machine learning over South Korea. HFRI was calculated through an ensemble model that combined an integrated model using all factors and a meteorological model using weather factors only. To confirm the generalized performance of the proposed model, all forest fires that occurred from 2014 to 2019 were validated using the receiver operating characteristic (ROC) curves and the area under the ROC curve (AUC) values through one-year-out cross-validation. The AUC value of HFRI ensemble model was 0.8434, higher than the meteorological model. HFRI was compared with the modified version of Fine Fuel Moisture Code (FFMC) used in the Canadian Forest Fire Danger Rating Systems and Daily Weather Index (DWI), South Korea’s current forest fire risk index. When compared to DWI and the revised FFMC, HFRI enabled a more spatially detailed and seasonally stable forest fire risk simulation. In addition, the feature contribution to the forest fire risk prediction was analyzed through the Shapley Additive exPlanations (SHAP) value of Catboost. The contributing variables were in the order of relative humidity, elevation, road density, and population density. It was confirmed that the accessibility factors played very important roles in forest fire risk modeling where most forest fires were caused by anthropogenic factors. The interaction between the variables was also examined.

Fifty years of wildland fire science in Canada

We celebrate the 50th anniversary of the Canadian Journal of Forest Research by reflecting on the considerable progress accomplished in select areas of Canadian wildland fire science over the past half century. Specifically, we discuss key developments and contributions in the creation of the Canadian Forest Fire Danger Rating System; the relationships between wildland fire and weather, climate, and climate change; fire ecology; operational decision support; and wildland fire management. We also discuss the evolution of wildland fire management in Banff National Park as a case study. We conclude by discussing some possible directions in future Canadian wildland fire research including the further evaluation of fire severity measurements and effects; the efficacy of fuel management treatments; climate change effects and mitigation; further refinement of models pertaining to fire risk analysis, fire behaviour, and fire weather; and the integration of forest management and ecological restoration with wildland fire risk reduction. Throughout the paper, we reference many contributions published in the Canadian Journal of Forest Research, which has been at the forefront of international wildland fire science.

Downscaling fire weather extremes from historical and projected climate models

An important aspect of predicting future wildland fire risk is estimating fire weather—weather conducive to the ignition and propagation of fire—under realistic climate change scenarios. Because the majority of area burned occurs on a few days of extreme fire weather, this task should be able to resolve fire weather extremes. In this paper, we present a statistical downscaling procedure based on distribution based scaling (DBS) to bias correct the Fire Weather Index (FWI), part of the Canadian Forest Fire Danger Rating System, as calculated from modeled climate data. Our study area is western Canada (British Columbia and Alberta) and we consider both an historical control period (1990–2000) and three future time periods (2020–2030, 2050–2060, and 2080–2090). The historical data used to calibrate the DBS procedure comprises weather station data and weather from the North American Regional Reanalysis (NARR), whereas the future climate projections are the output of three regional climate models, corresponding to different model parameterizations and downscaled from the NCAR Community Earth System Model under the RCP 8.5 scenario. By fitting a truncated Weibull distribution to observed and modeled FWI values, our method is able to reproduce historical extremes in fire weather indices as determined by the distribution of annual potential spread days, which are defined as days with FWI values greater than 19. Moreover, by calibrating the DBS procedure with gridded reanalysis data as well as station observations, we are able to project future spread day distributions over the entire study area. The results of this study show the DBS procedure leads to a greater number of projected annual spread days at most locations compared with estimates using the uncorrected model output, and that all three RCM models show positive increases in potential annual spread days for the 2050–2060 and 2080–2090 time periods.

Evaluation of IMERG-E Precipitation Estimates for Fire Weather Applications in Alaska

AbstractLarge wildfire outbreaks in Alaska are common from June to August. The Canadian Forest Fire Danger Rating System (CFFDRS) is used operationally by Alaskan fire managers to produce statewide fire weather outlooks and forecast guidance near active wildfires. The CFFDRS estimates of fire potential and behavior rely heavily on meteorological observations (precipitation, temperature, wind speed, and relative humidity) from the relatively small number of in situ stations across Alaska with precipitation being the most critical parameter. To improve the spatial coverage of precipitation estimates across Alaska for fire weather applications, a multisatellite precipitation algorithm was evaluated during six fire seasons (1 June–31 August 2014–19). Near-real-time daily precipitation estimates from the Integrated Multisatellite Retrievals for the Global Precipitation Mission (IMERG) algorithm were verified using 322 in situ stations across four Alaskan regions. For each region, empirical cumulative distributions of daily precipitation were obtained from station observations during each summer, and compared to corresponding distributions of interpolated values from IMERG grid points (0.1° × 0.1° grid). The cumulative distributions obtained from IMERG exhibited wet biases relative to the observed distributions for all regions, precipitation amount ranges, and summers. A bias correction approach using regional quantile mapping was developed to mitigate for the IMERG wet bias. The bias-adjusted IMERG daily precipitation estimates were then evaluated and found to produce improved gridded IMERG precipitation estimates. This approach may help to improve situational awareness of wildfire potential across Alaska and be appropriate for other high-latitude regions where there are sufficient in situ precipitation observations to help correct the IMERG precipitation estimates.

The Canadian versus the National Forest Fire Danger Rating Systems tested in Mediterranean forests fire Crete, Greece

The Canadian Forest Fire Danger Rating System (CFFDRS) and the National Forest Fire Danger Rating System (NFFDRS) were exercised under the Mediterranean forest type conditions of Crete Island, Greece in 2008–2009. The CFFDRS shows a high association between the assessed values of the duff moisture content on one hand and the Fine Fuel Moisture Code and the Duff Moisture Code on the other hand. During the designated fire seasons, there was an insignificant association between the Fire Weather Index and the wild forest fires which are contrary associated with wildfire occurrences. Meanwhile, the NFFDRS showed monitoring capabilities but with the insufficient prediction of the daily fire danger in the Mediterranean forests. During the tested forest fire seasons, there was an unstable association between the burnt area and the fire occurrences. Consequently, there was no significant correlation between the CFFDRS and the litter soil content. Although, CFFDRS has a higher prediction accuracy of moisture content of some shallow roots annual plants within the studied area. The findings of the current research emphasized that there is no single system that is adequate enough for both soil moisture content and the annuals moisture content simultaneously.

A Comparison of Fire Weather Indices with MODIS Fire Days for the Natural Regions of Alaska

Research Highlights: Flammability of wildland fuels is a key factor influencing risk-based decisions related to preparedness, response, and safety in Alaska. However, without effective measures of current and expected flammability, the expected likelihood of active and problematic wildfires in the future is difficult to assess and prepare for. This study evaluates the effectiveness of diverse indices to capture high-risk fires. Indicators of drought and atmospheric drivers are assessed along with the operational Canadian Forest Fire Danger Rating System (CFFDRS). Background and Objectives: In this study, 13 different indicators of atmospheric conditions, fuel moisture, and flammability are compared to determine how effective each is at identifying thresholds and trends for significant wildfire activity. Materials and Methods: Flammability indices are compared with remote sensing characterizations that identify where and when fire activity has occurred. Results: Among these flammability indicators, conventional tools calibrated to wildfire thresholds (Duff Moisture Code (DMC) and Buildup Index (BUI)), as well as measures of atmospheric forcing (Vapor Pressure Deficit (VPD)), performed best at representing the conditions favoring initiation and size of significant wildfire events. Conventional assessments of seasonal severity and overall landscape flammability using DMC and BUI can be continued with confidence. Fire models that incorporate BUI in overall fire potential and fire behavior assessments are likely to produce effective results throughout boreal landscapes in Alaska. One novel result is the effectiveness of VPD throughout the state, making it a potential alternative to FFMC among the short-lag/1-day indices. Conclusions: This study demonstrates the societal value of research that joins new academic research results with operational needs. Developing the framework to do this more effectively will bring science to action with a shorter lag time, which is critical as we face growing challenges from a changing climate.

Predicting forest fires burned area and rate of spread from pre-fire multispectral satellite measurements

Operational forest fire danger rating systems rely on the recent evolution of meteorological variables to estimate dead fuel condition. Further combining the latter with meteorological and environmental variables, they predict fire occurrence and spread. In this study we retrieved live fuel condition from MODIS multispectral measurements in the near infrared and shortwave infrared. Next, we combined these retrievals with an extensive dataset on actual forest fires in Campania (13,595 km2), Italy, to determine how live fuel condition affects the probability distribution functions of fire characteristics. Accordingly, the specific objective of this study was to develop and evaluate a new approach to estimate the probability distribution functions of fire burned area, duration and rate of spread as a function of the Perpendicular Moisture Index (PMI), whose value decreases with decreasing live fuel moisture content (LFMC). To this purpose, available fire data was intersected with MODIS 8-day composited reflectance data so to associate each fire event with the corresponding pre-fire PMI observation. Fires were then grouped in ten decile bins of PMI, and the conditional probability distribution functions of burned area, fire duration and rate of spread were determined in each bin. Distributions of burned area and rate of spread vary across PMI decile bins, while no significant difference was observed for fire duration. Further testing this result with a likelihood ratio test confirmed that PMI is a covariate of burned area and rate of spread, but not of fire duration. We defined an extreme event as a fire whose burned area (respectively rate of spread) exceeds the 95th percentile of the frequency distribution of all observed fire events. The probability distribution functions in the ten decile bins of PMI were combined to obtain a conditional probability distribution function, which was then used to predict the probability of extreme fires, as defined. It was found that the probability of extreme events steadily increases with decreasing PMI. Overall, at the end of the dry season the probability of extreme events is about the double than at the beginning. These results may be used to produce frequently (e.g. daily) updated maps of the probability of extreme events given a PMI map retrieved from e.g. MODIS reflectance data.