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.

Identification of Key Factors of Fire Risk of Oil Depot Based on Fuzzy Clustering Algorithm

With the rapid development of the national economy, the demand for oil is increasing. In order to meet the increasing energy demand, China has established a number of oil depot in recent years, whose largest capacity reaching up to tens of millions of cubic meters. Due to the flammable and explosive nature of the stored medium, the risk of fire in the oil depot area has increased dramatically as the tank capacity of the storage tank area has increased. The intensification of the oil depot and the development of large-scale oil storage tanks have brought convenience to the national oil depot, but also brought many catastrophic consequences. In recent years, there have been many fires and explosions in the oil depot, causing major casualties and property losses, which seriously endangered the ecological environment and public safety. Based on the constructed oil depot fire risk index system, the fuzzy C-means algorithm (FCM) and fuzzy maximum support tree clustering algorithm is introduced. Through the two fuzzy clustering mathematical models, key factors in the established index system are identified. Firstly, the expert scoring method is used to evaluate the indicators in the oil depot fire risk index system, and the importance degree evaluation matrix of oil depot fire risk factors is constructed through the fuzzy analysis of expert comments. Then, the fuzzy C-means algorithm (FCM) and the fuzzy clustering tree algorithm are used to cluster the various risk indicators, and the key factors of the oil depot fire risk are identified. Through the comparative analysis and cross-validation of the results of the two fuzzy clustering methods, the accuracy of the recognition results is ensured. Finally, using an oil depot as a case study, it is found that passive fire prevention capability and emergency rescue capability are key factors that need to be paid attention to in the oil depot fire risk index. The fuzzy clustering algorithm used in this paper can digitize the subjective comments of experts, thus reducing the influence of human subjective factors. In addition, by using two fuzzy clustering algorithms to analyze and verify the key factors of the oil depot fire risk, the reliability of the clustering results is guaranteed. The identification of key factors can enable managers to predict high-risk factors in advance in the fire risk prevention and control process of the oil depot, so as to adopt corresponding preventive measures to minimize the fire risk in the oil depot, and ensure the safety of the operation of the oil depot.

A Fire-Risk-Breakdown System for Electrical Power Lines in the North of Brazil

Anthropogenic or spontaneous fires (hotspots) are the main causes of unexpected breakdowns of electrical power lines in the northern region of Brazil. This research has tested, adapted, and implemented a preoperational system aiming to prevent electrical breakdowns for 382 km of electrical transmission lines in the state of Maranhão. The breakdown electrical fire risk is based on a combination of three variables: 1) the fire risk index, 2) the remotely sensed hotspot presence in the vicinity of electrical power lines, and 3) the vegetation stage. These variables are converted into Boolean variables, and their combination will classify the electrical fire risk as extreme, high, medium, low, or null. In regard to the system input variables, the fire risk index carries the highest representativeness in composition value of the breakdown electrical fire risk. Therefore, the results of two fire risk indices, calculated on the basis of the (a) Monte Alegre and (b) Angstrom methods, are presented and discussed. The validation of the fire risk indices is based on six categorical statistics (with the obtained final values also indicated in parentheses for the Monte Alegre and Angstrom methods, respectively): accuracy (0.91, 0.92), bias (1.05, 1.06), probability of detection (0.98, 0.99), false-alarm ratio (0.07, 0.07), probability of false detection (0.90, 0.80), and threat score (0.91, 0.92). The system presented here may be used as a tool within the electrical sector to prevent and respond to sudden electrical power line breakdowns.