Detection of Peat Fire Risk Area Based on Impedance Model and DInSAR Approaches using ALOS-2 PALSAR-2 Data

Joko Widodo, Yuta Izumi, Ayaka Takahashi, Husnul Kausarian, Daniele Perissin, Josaphat Tetuko Sri Sumantyo
2019 IEEE Access  
Forest fire in Indonesia occurs mostly in peatland area. Dry peatland areas with groundwater table (GWT) more than 40 cm from the soil surface have become degradation areas with high potentials to fire. This paper presents a new novel to detect a peat fire risk area by incorporating two methods: the impedance model and the differential interferometric SAR (DInSAR) technique which is based on the knowledge of annual subsidence rate associated with the GWT. The previous impedance model is
more » ... in this paper by integrating the surface roughness information in the model as a part of novelty. The proposed method was then validated with ground truth data of GWT. By using an impedance model, this paper successfully detected peat fire risk area based on the backscattering coefficient simulation of dry peatland. Based on the simulation model, the average, minimum, and maximum of backscattering coefficient of dry peat are −13.97, −11.5, and −17.29 dB, respectively. The correlation coefficient between the simulated backscattering coefficient and backscattering from ALOS-2/PALSAR-2 data is 0.8 with root mean square error of 1.4. By using the DInSAR method, detection of dry peatland area was successful. The significant relationships confirmed between GWT measurement and model are 0.71 for Pair A and 0.85 for Pair B. Both methods showed that peat fire risk areas were identified successfully. The dielectric constant of the peat soil also revealed that the soil condition of the area of interest is very dry indicating the potential to peat fire risk. Employing two models, respectively, were recommended to get precision of detection analysis. INDEX TERMS Detection of peat fire risk area, impedance model, DInSAR, ALOS-2 PALSAR-2 data.
doi:10.1109/access.2019.2899080 fatcat:5g52sibnazalrfnxh64g6hjcvm