Fig 3 - uploaded by Alan H. Taylor
Content may be subject to copyright.
Variable importance plot showing rank orders of variable importance for all four RF models using the Reduced Dataset and the spectral data from 2003 to predict the measured values of the canopy fuels parameters. The most important variables are at the top of the y-axis in each plot, and variables decrease in importance as one moves down the y-axis. The x-axis gives the mean percentage decrease in MSE for each variable. Abbreviations are: TC Bright—Tasseled Cap Brightness; TC Green—Tasseled Cap Greenness; TC Wet—Tasseled Cap Wetness; TM Band 1–5, TM Band 7 from Landsat 5, August 11, 2009; Slope—Slope of each plot in degrees; Aspect—Aspect of each plot in degrees from north; Elevation—Elevation of each plot; TP 150—Topographic Position of each plot relative to points within 150 m; TP 450—Topographic Position of each plot relative to points within 450 m. See the text for more details on each variable.
Source publication
Fire managers often need data that is spatially explicit at a fine scale (30 m) but is also laborious and time consuming to collect. This study integrates Landsat 5 imagery and topographic information with plot and tree based data to model and map four key canopy fuels variables: Canopy Bulk Density (CBD), Canopy Cover (CC), Canopy Base Height (CBH...
Context in source publication
Context 1
... results of the RF algorithm were consistently strong for CBD, CC, and HT, but weak for CBH (Table 8). Overall, topographic variables were largely unimportant in modeling and mapping the 2009 or 2003 canopy fuels parameters but vegetation indices and several at-satellite Landsat spectral bands were important (Figs. 2 and 3). NDVI, an important vegetation index for many types of landscape-scale and larger studies, was the most important predictor for most canopy fuel variables in both the Full and Reduced Datasets case. ...
Similar publications
Optimizing nitrogen (N) fertilizer combination is a crucial measure to maximize yield and production efficiency in a maize-soybean strip intercropping system (MSSI). In this field experiment, six maize/soybean N input combinations (0 kg ha−1, F0; 255/30 kg ha−1, F1; 255/45 kg ha−1, F2; 255/60 kg ha−1, F3; 210/45 kg ha−1, F4; 300/45 kg ha−1, F5) wer...
To determine the effect of Nitrogen in different yield attributing character of maize. A single-factor randomized complete block design was used with seven treatments and three replication. Plant height, leaf length, leaf breadth, Number of leaves, days to 50% tasselling and silking, number of cobs, cob length, 1000 grain weight, productivity, and...
Stemflow is vital for supplying water, fertilizer, and other crop essentials during sprinkler irrigation. Exploring the spatial and temporal variations of crop stemflow and its influencing factors will be essential to preventing soil water and nutrient ion’s migration to deeper layers, developing, and optimizing effective sprinkler irrigation sched...
In the second half of 20th century the awareness of importance of landraces for the future, led to organized collecting missions for numerous plant species. A total of 2217 maize (Zea mays L.) landraces, collected in the former Yugoslavia, are stored at Maize Research Institute (MRIZP) gene bank. During 2014, new collecting missions were organized...
This paper presents a method of mapping and monitoring ecological quality and environmental change using an ecological evaluation model (EEM), which is based on remote sensing data of the Pearl River Delta region in Guangdong, China. Five geographical indices were selected: Impervious Surface, Normalized Difference Vegetation Index, Land Surface Te...
Citations
... A review of the ML literature found that Random Forests are the most used algorithm in wildfire science [54]. For example, many papers have used Random Forests to perform fuel classification [55][56][57] and lightning prediction [58]. Random Forest models have also been shown to have superior performance in burned area assessment [59]. ...
As climate warming exacerbates wildfire risks, prompt wildfire detection is an essential step in designing an efficient suppression strategy, monitoring wildfire behavior and, when necessary, issuing evacuation orders. In this context, there is increasing demand for estimates of returns on wildfire investments and their potential for cost savings. Using fire-level data from Western Canada during 2015–2020, the paper associates variation in wildfire reporting delays with variation in suppression costs. We use machine learning and orthogonalization methods to isolate the impact of reporting delays from nonlinear impacts of the fire environment. We find that reporting delays account for only three percent of total suppression costs. Efforts to improve detection and reduce wildfire reporting delays by one hour lead to a modest 0.25% reduction in suppression costs. These results suggest that investments in detection systems that reduce wildfire reporting delays are not justified on suppression costs savings alone.
... In Lassen Volcanic National Park (LVNP), Pierce et al. Pierce, Farris and Taylor (2012) aimed to analyse forest fuel patterns, evaluate the efficiency of RF as a method for modelling plot-level canopy fuel loads, and pre-map the loads. Their goals were to determine the differences in surface and canopy fuel loads across LVNP based on topography and vegetation type, map surface and canopy fuels using RF regression and topographic characteristics with LandSat data, compare our canopy fuel maps with standard datasets, and predict fire behaviour using the surface and canopy fuel maps. ...
Forest fires are a persistent global problem, causing devastating consequences such as loss of human lives, harm to the environment, and substantial economic losses. To mitigate these impacts, the accurate prediction and early detection of forest fires is critical. In response to this challenge and living in the digital era of Artificial Intelligence (AI) and smart economies, there has been a growing interest in utilising AI mechanisms for forest fire management. This study provides an in-depth examination of the use of AI algorithms in the fight against forest fires. In particular, our paper starts with an overview of the forest fire problem, followed by a comprehensive review of various systems and approaches. This review includes a thorough analysis of the various works that have evaluated the factors that influence fire occurrence and severity, as well as those that focus on fire prediction and detection systems. The paper also explores the use of AI in adapting and restoring after the occurrence of forest fires. The paper concludes with an evaluation of the potential impact of AI on forest fire management and suggestions for future research directions, taking full advantage of novel technologies, such as 5G communications, Software Defined Networking (SDN), digital twins, federated learning and blockchain. Finally, the paper draws lessons and insights on the potential and limitations of AI in forest fire management, highlighting the need for further research and development in this field to maximise its impact and benefits.
... Numerous studies have been carried out to estimate fuel load and fuel moisture using remote sensing techniques [49,[51][52][53][54][55]. Remote sensing has proved useful in live-canopy fuel characterization [56][57][58][59]; however, challenges lie in characterizing dead below-canopy surface fuels, which generally are available up to 2 m high from the forest floor. Efforts have previously been made to estimate surface fuel loads using satellite remote sensing-based inputs [60,61]. ...
In the present research, the open-source WRF-SFIRE model has been used to carry out surface forest fire spread forecasting in the North Sikkim region of the Indian Himalayas. Global forecast system (GFS)-based hourly forecasted weather model data obtained through the National Centers for Environmental Prediction (NCEP) at 0.25 degree resolution were used to provide the initial conditions for running WRF-SFIRE. A landuse–landcover map at 1:10,000 scale was used to define fuel parameters for different vegetation types. The fuel parameters, i.e., fuel depth and fuel load, were collected from 23 sample plots (0.1 ha each) laid down in the study area. Samples of different categories of forest fuels were measured for their wet and dry weights to obtain the fuel load. The vegetation specific surface area-to-volume ratio was referenced from the literature. The atmospheric data were downscaled using nested domains in the WRF model to capture fire–atmosphere interactions at a finer resolution (40 m). VIIRS satellite sensor-based fire alert (375 m spatial resolution) was used as ignition initiation point for the fire spread forecasting, whereas the forecasted hourly weather data (time synchronized with the fire alert) were used for dynamic forest-fire spread forecasting. The forecasted burnt area (1.72 km²) was validated against the satellite-based burnt area (1.07 km²) obtained through Sentinel 2 satellite data. The shapes of the original and forecasted burnt areas matched well. Based on the various simulation studies conducted, an operational fire spread forecasting system, i.e., Sikkim Wildfire Forecasting and Monitoring System (SWFMS), has been developed to facilitate firefighting agencies to issue early warnings and carry out strategic firefighting.
... They tested the random forest algorithm to estimate the above-ground biomass in southern Africa and found coefficient of determination (R 2 ) values ranging from 0.43 to 0.65. Pierce et al. [84] used the RF algorithm and information from Landsat 5 TM, field data, and topographic factors for modeling and mapping canopy fuels in California (USA) and attained pseudo-R 2 values ranging from 0.55 to 0.68. Frazier et al. [12] characterized the above-ground biomass in a boreal forest using Landsat temporal segmentation metrics and found R 2 values of 0.62 estimated using random forest models. ...
Techniques and tools meant to aid fire management activities in the Cerrado, such as accurately determining the fuel load and composition spatially and temporally, are pretty scarce. The need to obtain fuel information for more efficient management in a considerably heterogeneous, biodiverse, and fire-dependent environment requires a constant search for improved remote sensing techniques for determining fuel characteristics. This study presents the following objectives: (1) to assess the use of data from Landsat 8 OLI images to estimate the fine surface fuel load of the Cerrado during the dry season by adjusting multiple linear regression equations, (2) to estimate the fuel load through random forest and k-nearest neighbor (k-NN) algorithms in comparison to regression analyses, and (3) to evaluate the importance of predictor variables from satellite images. Therefore, 64 sampling units were collected, and the pixel values associated with the field plots were extracted in a 3 × 3-pixel window surrounding the reference pixel. For multiple linear regression analyses, the R2 values ranged from 0.63 to 0.78, while the R2 values of the models fitted using the random forest algorithm ranged from 0.52 to 0.83 and the R2 values of those fitted using the k-NN algorithm ranged from 0.30 to 0.68. The estimates made through multiple linear regression analyses showed better results for the equations adjusted for the beginning of the dry season (May and June). Adopting the random forest algorithm resulted in improvements in the statistical metrics of evaluation of the fuel load estimates for the Cerrado grassland relative to multiple linear regression analyses. The variable fraction-soil (FS) exerted the most significant effect on surface fuel load estimates, followed by the vegetation indices NDII, GVMI, DER56, NBR, and MSI, all of which use near-infrared and short-wave infrared channels in their calculations.
... They tested the Random Forest algorithm to estimate aboveground biomass in southern Africa and found coefficient of determination (R²) values ranging from 0.43 to 0.65. Pierce et al. [77] used the RF algorithm and information from Landsat 5 TM, field data, and topographic factors for modeling and mapping canopy fuels in California (USA) and attained pseudo-R² values ranging from 0.55 to 0.68. Frazier et al. [11] characterized the aboveground biomass in a boreal forest using Landsat temporal segmentation metrics and found R² values of 0.62 estimated by Random Forest models. ...
Techniques and tools meant to aid fire management activities in the Cerrado, such as accurately determining fuel load and composition spatially and temporally, are pretty scarce. The need to have fuel information for more efficient management in a considerably heterogeneous, biodiverse, and fire-dependent environment makes a constant search for the improvement of the use of remote sensing techniques in determining fuel characteristics. This study presents the following objectives: (1) to assess the use of data from Landsat 8 OLI images to estimate the fine surface fuel load of the Cerrado during the dry season by adjusting multiple linear regression equations; (2) to estimate fuel load through Random Forest and k-Nearest Neighbor (k-NN) algorithms in comparison with re-gression analyses; and (3) to evaluate the importance of predictor variables from satellite images. Therefore, 64 sampling units were collected, and the pixel values associated with the field plots were extracted in a 3 x 3 pixel window surrounding the reference pixel. For the multiple linear regression analyses, the R² values ranged from 0.63 to 0.78, and the models fitted by the Random Forest al-gorithm ranged from 0.52 to 0.83, while those fitted by k-NN ranged from 0.30 to 0.68. Adopting the Random Forest algorithm resulted in improvements in the statistical metrics of evaluation of the fuel load estimates for Cerrado grassland relative to the multiple linear regression analyses.
... Existing fire behaviour models can capture the impacts of weather on fire spread rate and intensity but require spatially explicit information about a range of fuel attributes as input (Tolhurst et al., 2008). Although current fuel hazard can be mapped using a combination of groundbased and remotely sensed observations (Pierce et al., 2012), quantification of future changes in these patterns in response to climate change requires predictive models. ...
... space-for-time substitution; Pickett, 1989), these ground-based fuel surveys contain possibly the best information about the potential for changes in fuel hazard across Victoria in response to projected climate conditions. Random forest models have been used to synthesise field-based fire observations and environmental drivers with demonstrated success (Jenkins et al., 2020;McColl-Gausden et al., 2020;Pierce et al., 2012). However, previous machine learning approaches have generally ignored plant responses to climate change (e.g. ...
... Despite the high resolution, it has relatively low computational demand for regional projections (i.e.does not require high-performance computing). The comprehensiveness separated this approach from previous empirical models that only consider a subset of predictors (e.g.Jenkins et al., 2017;McColl-Gausden et al., 2020;Pierce et al., 2012) while the high resolution (90 m) is what current process-based modelling cannot provide(Rabin et al., 2017). The random forest model predicted increased P 4_5 in the future for areas currently under semiarid climates(Figure 4). ...
Bushfire fuel hazard is determined by the type, amount, density and three‐dimensional distribution of plant biomass and litter. The fuel hazard represents a biological control on fire danger and may change in the future with plant growth patterns. Rising atmospheric CO 2 concentration (C a ) stimulates plant productivity (‘fertilisation effect’) but also alters climate, leading to a ‘climatic effect’. Both effects have impacts on future vegetation and thus fuel hazard. Quantifying these effects is an important component of predicting future fire regimes and evaluating fire management options.
Here, we combined a machine learning algorithm that incorporates the power of large fine spatial resolution (i.e. 90 m) datasets with a novel optimality model that accounts for the climatic and fertilisation effects on vegetation cover. We demonstrated the usefulness and practicality of this framework by predicting fuel hazard across the state of Victoria in Australia. We fitted and evaluated the models with long‐term (i.e. 20 years), ground‐based fuel observations.
The models achieved strong agreement with observations across the fuel hazard range (accuracy >65%). We found fuel hazard increased more in dry environments due to future climate and C a . The contribution of the ‘fertilisation effect’ to future fuel hazard varied spatially by up to 12%.
The predictions of future fuel hazard are directly useful to inform fire mitigation policies and as a reference for climate model projections to account for fire impacts.
Synthesis and applications : Climate change and rising C a have profound impacts on vegetation and thus fuel load. Operational fire management and future fire risk forecasts will benefit from our realistic fuel load prediction framework that incorporates plant responses and fine soil and terrain attributes.
... Gale et al. (2021) divided fuel attributes based on their vertical position, and more specifically in overstory, understory, surface, and bark fuel, as well as their condition, i.e., dead or live fuel and fuel model. Overstory main attributes, like canopy cover and height, have been modeled with reasonable accuracies, using statistical and machine learning techniques alongside with remote sensing methods (Heiskanen et al., 2011;Pierce et al., 2012;Palaiologou et al., 2013;Korhonen et al., 2017). Nevertheless, the other three types of fuel attributes have merely been modeled using passive remote sensing, as the reflectance of the understory, surface and bark fuel is interrupted by the canopies of the 3] A general term used to express an assessment of both fixed and variable factors of the fire environment that determine the ease of ignition, rate of spread, difficulty of control, and fire impact; often expressed as an index GFMC-FAO, 1999 Fire model 1] A computer program which, with given information, will predict the rate of spread of a fire from a point of origin GFMC-FAO, 1999 2] A computer model which, with given information, will predict the spread of fire as influenced by meteorological conditions, fuel characteristics, and topography Fire risk ...
Wildfires have been systematically studied from the early 1950s, with significant progress in the applied computational methodologies during the 21st century. However, modern methods are barely adopted by administrative authorities, globally, especially those considering probabilistic models concerning human-caused fires. An exhaustive review on wildfire danger studies has not yet been performed. Therefore, the present review aims at collecting and analyzing integrated modeling approaches in estimating forest fire danger, examining the driving factors, and evaluating their influence on fire occurrence. The main objective is to propose the top performing methods and the most important risk factors for the development of an Integrated Wildfire Danger Risk System (IWDRS). Studies were classified based on the applied technique, i.e., geographic information systems, remote sensing, statistics, machine learning, simulation modeling and miscellaneous techniques. The conclusions of each study concerning the relative importance of model input variables are also reported. Online search engines such as 'Scopus', 'Google Scholar', 'WorldWideScience', 'ScienceDirect' and 'ResearchGate' were used in relevant literature searches published in scientific journals, manuals and technical documentation. A total of 230 studies were gathered with a selected subset being evaluated in a meta-analysis process. Machine learning techniques outperform average classic statistics, although their predictability relies heavily on the quantity and the quality of the input data. Geographic information systems and remote sensing are considered valuable yet supplementary tools. Modeling techniques apply best to fire behavior prediction, while other techniques referenced in the current review are potentially useful but further investigation is needed. In conclusion, wildfire danger is a function of seven thematic groups of variables: meteorology, vegetation, topography, hydrology, socio-economy, land use and climate. Ninety-five explanatory drivers are proposed.
... Using a large dataset, FFNN achieved an accuracy of 70.5% compared to the 50.4% accuracy achieved by GDA. Pierce et al. [37] used a RF model to map the forest canopy fuel in Lassen Volcanic National Park, California, by integrating ML and satellite imagery data. Patil and Sivagami [38] used a stacking of ensembles, namely RF, Extra Trees (XT) classifier, and Adaptive Boosting (AdaB) of XTs, integrated with a FFNN to improve the accuracy of forest cover classification. ...
Predicting the behaviour of wildfires can help save lives and reduce health, socioeconomic, and environmental impacts. Because wildfire behaviour is highly dependent on fuel type and distribution, their accurate estimation is paramount for accurate prediction of the fire propagation dynamics. This paper studies the effect of combining automated hyperparameter tuning with Bayesian optimisation and recursive feature elimination on the accuracy of three boosting (AdaB, XGB, CatB), two bagging (Random Forest, Extremely Randomised Trees), and three stacking ensemble models with respect to their ability to estimate the vegetation cover type from cartographic data. The models are trained on the University of California Irvine (UCI) cover type dataset using five-fold cross-validation. Feature importance scores are calculated and used in recursive feature elimination analysis to study the sensitivity of model accuracy to the different feature combinations. Our results indicate that the implemented fine-tuning procedure significantly affects the accuracy of all models investigated, with XGB achieving an overall accuracy of 97.1% slightly outperforming the others.
... In addition, many other intelligent algorithms have also been investigated [65,66,125,126], including random forest, SVM, ANN, DNN, and MDP. ...
The intensity and frequency of bushfires have increased significantly, destroying property and living species in recent years. Presently, unmanned aerial vehicle (UAV) technology advancements are becoming increasingly popular in bushfire management systems because of their fundamental characteristics, such as manoeuvrability, autonomy, ease of deployment, and low cost. UAVs with remote-sensing capabilities are used with artificial intelligence, machine learning, and deep-learning algorithms to detect fire regions, make predictions, make decisions, and optimize fire-monitoring tasks. Moreover, UAVs equipped with various advanced sensors, including LIDAR, visual, infrared (IR), and monocular cameras, have been used to monitor bushfires due to their potential to provide new approaches and research opportunities. This review focuses on the use of UAVs in bushfire management for fire detection, fire prediction, autonomous navigation, obstacle avoidance, and search and rescue to improve the accuracy of fire prediction and minimize their impacts on people and nature. The objective of this paper is to provide valuable information on various UAV-based bushfire management systems and machine-learning approaches to predict and effectively respond to bushfires in inaccessible areas using intelligent autonomous UAVs. This paper aims to assemble information about the use of UAVs in bushfire management and to examine the benefits and limitations of existing techniques of UAVs related to bushfire handling. However, we conclude that, despite the potential benefits of UAVs for bushfire management, there are shortcomings in accuracy, and solutions need to be optimized for effective bushfire management.
... Riano et al. (2005) used artificial neural network (ANN) to predict and map the water content of leaves in Italy. Pierce et al. (2012) used ML to classify important canopy fuel variables using field measurements and topographic data to produce forest canopy fuel maps in California. Riley et al. (2014) performed fuel classification and mapping in eastern Oregon with 84% accuracy in identifying existing vegetation groups (i.e. ...
There is an urgent need to develop new data-driven methods for assessing wildfire-related risks in large areas susceptible to such risks. To assess these risks, one of the critical parameters to analyse is fuel. This research note presents a framework for classifying fuels through the analysis of roadside images to complement the current practice of assessing fuels through aerial images and visual inspections. Some of the most prevalent fuel types in North America were considered for automated classification, including grasses, shrubs and timbers. A framework was developed using convolutional neural networks (CNNs), which can automate the process of fuel classification. Various pre-trained neural networks were examined and the best network in terms of time efficiency and accuracy was identified, and had ~94% accuracy in identifying the chosen fuel types. This framework was initially applied to street view images collected from Google Earth. Indeed, the results showed that the framework has the potential for application for fuel classification using roadside images, and this makes it suitable for crowdsensing-based fuel mapping for wildfire risk assessment, which is the future goal of this research.
