研究目的
To assess forest fire risk using entirely remote sensing derived variables and construct the relationship between fire influencing variables and the Fire Risk Index (FRI) using a logistic regression model.
研究成果
The logistic regression model effectively assesses forest fire risk with AUC values of 0.8 for grassland and 0.81 for woodland, indicating good performance. FMC is verified as a critical variable. Combining with meteorological data could enhance accuracy for future studies.
研究不足
The study relies solely on remote sensing derived variables and does not incorporate meteorological forecast data, which could improve accuracy. The complexity of FMC data acquisition and inversion may limit application in some contexts.
1：Experimental Design and Method Selection:
The study uses a logistic regression model to integrate remote sensing derived variables for forest fire risk assessment. Variables include FMC, NDVI, LAI, Elevation, Slope, and derived Difference and Anomaly of FMC. The model is developed and validated using spatial sampling and data analysis techniques.
2：Sample Selection and Data Sources:
Burned and unburned points are selected based on MODIS MCD64A1 burned area data product from 2007 to 2016 for Yunnan Province, China. Data is split into subsets for model development (2007-2015) and validation (2016).
3：6).
List of Experimental Equipment and Materials:
3. List of Experimental Equipment and Materials: Remote sensing data from MODIS images, specifically MCD64A1 product, is used. No specific equipment or materials are detailed beyond data sources.
4：Experimental Procedures and Operational Workflow:
Variables are derived from remote sensing data; Pearson correlation and Wald Chi-square tests are used for data analysis. The logistic regression model is built by iteratively excluding insignificant variables.
5：Data Analysis Methods:
Pearson correlation coefficient to identify variable correlations, Wald Chi-square test for significance, and AUC from ROC curves for model performance assessment.
独家科研数据包，助您复现前沿成果，加速创新突破
获取完整内容
