Toward peatland fire vulnerability monitoring using surface water maps: A geospatial and temporal analysis of peatland water bodies using Synthetic Aperture RADAR

Abstract

Peatlands, essential ecosystems with unique hydrological characteristics, are increasingly vulnerable to fire events due to climate change. The primary objective of this thesis is to test fire vulnerability monitoring in peatlands through the mapping of peatland water bodies, aiming to assess the hydrologic conditions over time. Employing geospatial analysis techniques, the study investigates the backscatter attributes and spatial patterns of these water bodies, elucidating their role in peatland hydrology and their impact on fire dynamics. This information is applied to test the design of Random Forest surface water classification models which test model parameters and training data schemes to determine the best surface water classification model. The characterization of small peatland water bodies and the testing of these models offers a strong foundation for a surface water classification that can capture surface water in peatlands, provided more intentional training data which represents the various conditions in peatlands is added.