A Study on the Susceptibility of Wildfire Disasters in Wuzhou City Based on Interpretability Model
Frequent wildfires in Wuzhou City, Guangxi, pose serious threats to the ecological environment and the safety of people’s lives and property. Accurately assessing wildfire susceptibility is crucial for effective disaster prevention and management decisions. This study aims to create an ac...
Saved in:
| Main Authors: | , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IEEE
2025-01-01
|
| Series: | IEEE Access |
| Subjects: | |
| Online Access: | https://ieeexplore.ieee.org/document/11050435/ |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1839633695107973120 |
|---|---|
| author | Lei Zou Hongjuan Shao Yintao Liu Chao Ren Qinyi Chen Haoming Bai Zhengzhong Huang Yao Gu |
| author_facet | Lei Zou Hongjuan Shao Yintao Liu Chao Ren Qinyi Chen Haoming Bai Zhengzhong Huang Yao Gu |
| author_sort | Lei Zou |
| collection | DOAJ |
| description | Frequent wildfires in Wuzhou City, Guangxi, pose serious threats to the ecological environment and the safety of people’s lives and property. Accurately assessing wildfire susceptibility is crucial for effective disaster prevention and management decisions. This study aims to create an accurate wildfire susceptibility model for Wuzhou City and examine the influence of various features on the prediction results through an explanatory approach. Based on historical wildfire data from 2017 to 2023, the study comprehensively considered the spatial distribution characteristics of the samples, selecting 15 evaluation factors such as elevation, NDVI, average annual precipitation, and average temperature. The Extreme Gradient Boosting (XGBoost) algorithm was adopted to construct the prediction model, aiming to provide precise wildfire susceptibility assessment. The SHapley Additive exPlanations (SHAP) method was then used to explain the global and local models, analyzing the importance and dependence of each feature. XGBoost is widely used in wildfire prediction due to its efficiency and accuracy. It not only provides strong predictive performance but also helps identify key factors influencing wildfire occurrences. The SHAP method further enhances the XGBoost model by improving the interpretability of the results. It quantifies each feature’s contribution to the predictions, enabling decision-makers to better understand the model’s decision-making process. The model’s performance was further validated using typical wildfire data from 2024. The results indicate that: 1) the wildfire susceptibility model for Wuzhou City exhibits excellent performance, with an Area Under the Curve (AUC) value of 0.981, accurately identifying high-susceptibility areas and providing essential decision-making support for relevant departments; 2) nine key factors affecting wildfire susceptibility in Wuzhou City were identified, including average annual precipitation, average temperature, and average wind speed, offering scientific evidence for wildfire risk management; 3) a comparison between typical historical wildfire samples from 2017 to 2023 and the latest wildfire data from January 2024 to November 2024, further validated the model’s accuracy. The findings of this study hold significant theoretical and practical value for wildfire prevention and control in Wuzhou City. |
| format | Article |
| id | doaj-art-e3c8e3d6e3af40ee9ed0682ee654d12c |
| institution | Matheson Library |
| issn | 2169-3536 |
| language | English |
| publishDate | 2025-01-01 |
| publisher | IEEE |
| record_format | Article |
| series | IEEE Access |
| spelling | doaj-art-e3c8e3d6e3af40ee9ed0682ee654d12c2025-07-10T23:01:11ZengIEEEIEEE Access2169-35362025-01-011311586011588010.1109/ACCESS.2025.358308911050435A Study on the Susceptibility of Wildfire Disasters in Wuzhou City Based on Interpretability ModelLei Zou0https://orcid.org/0009-0004-1758-9357Hongjuan Shao1https://orcid.org/0009-0003-2408-7702Yintao Liu2https://orcid.org/0009-0004-8096-3647Chao Ren3https://orcid.org/0000-0002-2591-6619Qinyi Chen4Haoming Bai5Zhengzhong Huang6Yao Gu7College of Geomatics and Geoinformation, Guilin University of Technology, Guilin, ChinaCollege of Physics and Electronic Information Engineering, Guilin University of Technology, Guilin, ChinaCollege of Geomatics and Geoinformation, Guilin University of Technology, Guilin, ChinaCollege of Geomatics and Geoinformation, Guilin University of Technology, Guilin, ChinaCollege of Geomatics and Geoinformation, Guilin University of Technology, Guilin, ChinaCollege of Geomatics and Geoinformation, Guilin University of Technology, Guilin, ChinaCollege of Geomatics and Geoinformation, Guilin University of Technology, Guilin, ChinaCollege of Geomatics and Geoinformation, Guilin University of Technology, Guilin, ChinaFrequent wildfires in Wuzhou City, Guangxi, pose serious threats to the ecological environment and the safety of people’s lives and property. Accurately assessing wildfire susceptibility is crucial for effective disaster prevention and management decisions. This study aims to create an accurate wildfire susceptibility model for Wuzhou City and examine the influence of various features on the prediction results through an explanatory approach. Based on historical wildfire data from 2017 to 2023, the study comprehensively considered the spatial distribution characteristics of the samples, selecting 15 evaluation factors such as elevation, NDVI, average annual precipitation, and average temperature. The Extreme Gradient Boosting (XGBoost) algorithm was adopted to construct the prediction model, aiming to provide precise wildfire susceptibility assessment. The SHapley Additive exPlanations (SHAP) method was then used to explain the global and local models, analyzing the importance and dependence of each feature. XGBoost is widely used in wildfire prediction due to its efficiency and accuracy. It not only provides strong predictive performance but also helps identify key factors influencing wildfire occurrences. The SHAP method further enhances the XGBoost model by improving the interpretability of the results. It quantifies each feature’s contribution to the predictions, enabling decision-makers to better understand the model’s decision-making process. The model’s performance was further validated using typical wildfire data from 2024. The results indicate that: 1) the wildfire susceptibility model for Wuzhou City exhibits excellent performance, with an Area Under the Curve (AUC) value of 0.981, accurately identifying high-susceptibility areas and providing essential decision-making support for relevant departments; 2) nine key factors affecting wildfire susceptibility in Wuzhou City were identified, including average annual precipitation, average temperature, and average wind speed, offering scientific evidence for wildfire risk management; 3) a comparison between typical historical wildfire samples from 2017 to 2023 and the latest wildfire data from January 2024 to November 2024, further validated the model’s accuracy. The findings of this study hold significant theoretical and practical value for wildfire prevention and control in Wuzhou City.https://ieeexplore.ieee.org/document/11050435/Wildfire disasterwildfire susceptibility assessmentmachine learning modelsSHAPmodel interpretation |
| spellingShingle | Lei Zou Hongjuan Shao Yintao Liu Chao Ren Qinyi Chen Haoming Bai Zhengzhong Huang Yao Gu A Study on the Susceptibility of Wildfire Disasters in Wuzhou City Based on Interpretability Model IEEE Access Wildfire disaster wildfire susceptibility assessment machine learning models SHAP model interpretation |
| title | A Study on the Susceptibility of Wildfire Disasters in Wuzhou City Based on Interpretability Model |
| title_full | A Study on the Susceptibility of Wildfire Disasters in Wuzhou City Based on Interpretability Model |
| title_fullStr | A Study on the Susceptibility of Wildfire Disasters in Wuzhou City Based on Interpretability Model |
| title_full_unstemmed | A Study on the Susceptibility of Wildfire Disasters in Wuzhou City Based on Interpretability Model |
| title_short | A Study on the Susceptibility of Wildfire Disasters in Wuzhou City Based on Interpretability Model |
| title_sort | study on the susceptibility of wildfire disasters in wuzhou city based on interpretability model |
| topic | Wildfire disaster wildfire susceptibility assessment machine learning models SHAP model interpretation |
| url | https://ieeexplore.ieee.org/document/11050435/ |
| work_keys_str_mv | AT leizou astudyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT hongjuanshao astudyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT yintaoliu astudyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT chaoren astudyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT qinyichen astudyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT haomingbai astudyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT zhengzhonghuang astudyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT yaogu astudyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT leizou studyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT hongjuanshao studyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT yintaoliu studyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT chaoren studyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT qinyichen studyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT haomingbai studyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT zhengzhonghuang studyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel AT yaogu studyonthesusceptibilityofwildfiredisastersinwuzhoucitybasedoninterpretabilitymodel |