Spatial logistic regression models for predicting peatland fire in Bengkalis Regency Indonesia

Lailan Syaufina, . and Sandhi Imam Maulana, . and Lilik Budi Prasetyo, . and Muhammad Nur Aidi, . Spatial logistic regression models for predicting peatland fire in Bengkalis Regency Indonesia. pp. 55-66. ISSN 2672-7226

Peat land fires have received increasing attentions as a major recurrent environmental problem in Indonesia particularly across the eastern coast of Sumatera Island where Bengkalis Regency is located. Although peat land fire prediction analysis has become an essential aspect of fire management however it seems that studies are still very limited in Indonesia. Therefore this study objective is to develop a prediction models for peat land fire particularly in the Bengkalis Regency Riau Province-Indonesia. This study was conducted based on spatial logistic regression method in which that a prediction model that can provide preparation time up to 2 months before the beginning of peat land fire season was selected to produce a prediction map based on physiographic variables peat physical characteristics variables human activity variables and climate variables. Performance of the selected peat land fire prediction model has been verified and validated using an independent testing subset and the results showed are consistently reliable (Chisquare p 0.001; Nagelkerke R20.314; AUC0.8309; Overall accuracy of correct predicted value85.16). This finding is useful to improve peat land fire management in Bengkalis Regency and can also be used to help the authorities in the spatial domain to make more appropriate decisions related to fire prevention strategies.

Full text available from: Official URL: http://jssm.umt.edu.my/wp-content/uploads/sites/51...

Abstract

Peat land fires have received increasing attentions as a major recurrent environmental problem in Indonesia particularly across the eastern coast of Sumatera Island where Bengkalis Regency is located. Although peat land fire prediction analysis has become an essential aspect of fire management however it seems that studies are still very limited in Indonesia. Therefore this study objective is to develop a prediction models for peat land fire particularly in the Bengkalis Regency Riau Province-Indonesia. This study was conducted based on spatial logistic regression method in which that a prediction model that can provide preparation time up to 2 months before the beginning of peat land fire season was selected to produce a prediction map based on physiographic variables peat physical characteristics variables human activity variables and climate variables. Performance of the selected peat land fire prediction model has been verified and validated using an independent testing subset and the results showed are consistently reliable (Chisquare p 0.001; Nagelkerke R20.314; AUC0.8309; Overall accuracy of correct predicted value85.16). This finding is useful to improve peat land fire management in Bengkalis Regency and can also be used to help the authorities in the spatial domain to make more appropriate decisions related to fire prevention strategies.

[error in script]

Item Type:	Article
AGROVOC Term:	Fires
AGROVOC Term:	Peatlands
AGROVOC Term:	Forest fires
AGROVOC Term:	forest fire management
AGROVOC Term:	Prediction
AGROVOC Term:	Data collection
AGROVOC Term:	Regression analysis
AGROVOC Term:	Mathematical models
AGROVOC Term:	Physiographic features
AGROVOC Term:	Soil physical properties
Depositing User:	Mr. AFANDI ABDUL MALEK
Last Modified:	24 Apr 2025 00:54
URI:	http://webagris.upm.edu.my/id/eprint/8302

Actions (login required)

View Item