Canopy active fuel loads estimation with C-band PolSAR by combining polarization decomposition and vegetation scattering model: a case study in Southwestern China

ABSTRACT

The forest canopy active fuel load (CAFL) is easily flammable that is usually consumed in the flaming front, and is a critical factor in crown fire early prediction, suppression and response. Previous studies universally relied on optical remote sensing data, whose effectiveness is extremely limited in regions with cloud cover and haze, such as the southwest of Sichuan, China. Synthetic aperture radar (SAR) can work 24 h and is not affected by weather conditions. This study aims to evaluate the potential of volume backscattering coefficients extracted through decomposition of polarimetric SAR (PolSAR, Radarsat-2 C band) for CAFL estimation. The polarimetric decomposition method used in this study was based on Freeman-Durden decomposition. The separated volume backscattering coefficients were used to estimate CAFL based on the variant of the extended water cloud model (VEWCM). For comparison, the SAR raw observations and extended water cloud model (EWCM) were also used to estimate CAFL. The performance of these two models was validated through quantitative assessment, utilizing CAFL measurements obtained from southwest Sichuan and the leave-one-out cross-validation method. Results show that the VEWCM (R² = 0.69, RMSE = 0.87 tons/ha) performed better than EWCM (R² = 0.60, RMSE = 1.48 tons/ha) especially in HV polarization. We concluded that PolSAR has the potential to improve CAFL estimation and further assist with pre-fire risk warnings, fire suppression responses, and post-fire effects assessment.

KEYWORDS:

• Fire risk
• Forest canopy active fuel load
• Model-based polarimetric decomposition
• Extended water cloud model

Acknowledgements

This work was supported by the National Key R&D Program of China (Contract No. 2022YFC3003001) and the Sichuan Science and Technology Program (Contract No. 2023YFS0432). The authors would like to appreciate Tengfei Xiao, Lin Chen, Chunquan Fan and Jianpeng Yin in Quantitative remote sensing team at University of Electronic Science and Technology of China for their help of the data collection.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was supported by the National Key R&D Program of China [Contract No. 2022YFC3003001] and the Sichuan Science and Technology Program [Contract No. 2023YFS0432].