Wildfire Susceptibility Mapping using NBR Index and Frequency Ratio Model
Quantifying fire hazards in natural areas and their spatial patterns are essential for developing appropriate fire management strategies, especially in countries with limited historical data on past fires. In this study, a fire hazard map for the Andika region of Iran was constructed by examining th...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
OICC Press
2022-06-01
|
| Series: | Geoconservation Research |
| Subjects: | |
| Online Access: | https://gcr.isfahan.iau.ir/article_693957_51d612d5c845a288fe43e6c437809179.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1832568689095868416 |
|---|---|
| author | Farzad Bolouk Heidari Ramin Arfania |
| author_facet | Farzad Bolouk Heidari Ramin Arfania |
| author_sort | Farzad Bolouk Heidari |
| collection | DOAJ |
| description | Quantifying fire hazards in natural areas and their spatial patterns are essential for developing appropriate fire management strategies, especially in countries with limited historical data on past fires. In this study, a fire hazard map for the Andika region of Iran was constructed by examining the correlation of past fires with the criteria of topography, meteorology, land cover, and human factors. The locations of eight-year fire points from 2013 to 2020 of Nova satellite sensor VIIRS were received and the fire map of each was constructed using the NBR (Normalized Burn Ratio). The wildfire events distribution maps were randomly divided into 70 and 30 percent ratios for training (modeling) and testing (validation) data, respectively. Using the frequency ratio, a fire hazard map of the region was created. Four fire hazard areas ranging from very high to low were identified. The results of past fires and the frequency ratio model showed that in the study area, land cover (2.982), elevation (2.778), and annual precipitation (2.419) have the greatest prediction rate and influence on fire occurrence. The results also showed that a large proportion of past fires (71.37%) were located in high and very high-risk areas. The evaluation results of the area under the curve method showed an accuracy of 71.1% using evaluation data and 74.4% using training data, which can be considered desirable. The small differences between the validation results using test and training data indicate an unbiased fire hazard map. |
| format | Article |
| id | doaj-art-183440c243804387a0caab42a3cabf04 |
| institution | Kabale University |
| issn | 2645-4661 2588-7343 |
| language | English |
| publishDate | 2022-06-01 |
| publisher | OICC Press |
| record_format | Article |
| series | Geoconservation Research |
| spelling | doaj-art-183440c243804387a0caab42a3cabf042025-02-03T00:39:24ZengOICC PressGeoconservation Research2645-46612588-73432022-06-015124026010.30486/gcr.2022.1961153.1107693957Wildfire Susceptibility Mapping using NBR Index and Frequency Ratio ModelFarzad Bolouk Heidari0Ramin Arfania1Department of Geographic Information System and Remote Sensing, Lenjan Branch, Islamic Azad University, Lenjan, IranDepartment of Geology, Isfahan (Khorasgan) Branch, Islamic Azad University, Isfahan, IranQuantifying fire hazards in natural areas and their spatial patterns are essential for developing appropriate fire management strategies, especially in countries with limited historical data on past fires. In this study, a fire hazard map for the Andika region of Iran was constructed by examining the correlation of past fires with the criteria of topography, meteorology, land cover, and human factors. The locations of eight-year fire points from 2013 to 2020 of Nova satellite sensor VIIRS were received and the fire map of each was constructed using the NBR (Normalized Burn Ratio). The wildfire events distribution maps were randomly divided into 70 and 30 percent ratios for training (modeling) and testing (validation) data, respectively. Using the frequency ratio, a fire hazard map of the region was created. Four fire hazard areas ranging from very high to low were identified. The results of past fires and the frequency ratio model showed that in the study area, land cover (2.982), elevation (2.778), and annual precipitation (2.419) have the greatest prediction rate and influence on fire occurrence. The results also showed that a large proportion of past fires (71.37%) were located in high and very high-risk areas. The evaluation results of the area under the curve method showed an accuracy of 71.1% using evaluation data and 74.4% using training data, which can be considered desirable. The small differences between the validation results using test and training data indicate an unbiased fire hazard map.https://gcr.isfahan.iau.ir/article_693957_51d612d5c845a288fe43e6c437809179.pdfremote sensinggisnbr indexfrequency ratiowildfirefire hazard |
| spellingShingle | Farzad Bolouk Heidari Ramin Arfania Wildfire Susceptibility Mapping using NBR Index and Frequency Ratio Model Geoconservation Research remote sensing gis nbr index frequency ratio wildfire fire hazard |
| title | Wildfire Susceptibility Mapping using NBR Index and Frequency Ratio Model |
| title_full | Wildfire Susceptibility Mapping using NBR Index and Frequency Ratio Model |
| title_fullStr | Wildfire Susceptibility Mapping using NBR Index and Frequency Ratio Model |
| title_full_unstemmed | Wildfire Susceptibility Mapping using NBR Index and Frequency Ratio Model |
| title_short | Wildfire Susceptibility Mapping using NBR Index and Frequency Ratio Model |
| title_sort | wildfire susceptibility mapping using nbr index and frequency ratio model |
| topic | remote sensing gis nbr index frequency ratio wildfire fire hazard |
| url | https://gcr.isfahan.iau.ir/article_693957_51d612d5c845a288fe43e6c437809179.pdf |
| work_keys_str_mv | AT farzadboloukheidari wildfiresusceptibilitymappingusingnbrindexandfrequencyratiomodel AT raminarfania wildfiresusceptibilitymappingusingnbrindexandfrequencyratiomodel |