Predicting surface soil moisture of northern peatlands from hyper- and multispectral satellite data

Northern peatlands are critical carbon stores that are highly sensitive to hydrological conditions. Changes in these conditions, driven by climate change and land-use modifications, can even shift peatlands from carbon sinks to sources. Consequently, effective management of both natural and restored peatlands requires monitoring of hydrological parameters such as soil moisture. Recently launched hyperspectral satellites like EnMAP provide new capabilities for this monitoring through enhanced spectral resolution. This study presents a novel framework for estimating peatland soil moisture content by applying Continuous Wavelet Transform (CWT) with real-valued Morlet wavelet to EnMAP satellite data. We evaluated the performance of CWT-processed data against original EnMAP bands, multispectral Sentinel-2 data, and spectral moisture indices across nine peatland sites spanning hemiboreal, boreal, sub-Arctic, and Arctic zones in Finland, Sweden, and Estonia. The CWT-processed EnMAP model achieved the highest predictive accuracy (R2 = 0.67, RMSE = 14.02%), followed by original EnMAP bands (R2 = 0.50, RMSE = 17.23%), while Sentinel-2 bands showed substantially lower performance (R2 = 0.32, RMSE = 20.09%) despite having finer spatial resolution. These results suggest that spectral resolution outweighs spatial resolution for peatland soil moisture estimation. Spectral moisture indices performed poorly with both satellite sensors (R2 = 0.00-0.17), demonstrating the limitations of single band combinations compared to full spectral approaches. While spatial mismatch between ground measurements and satellite pixel size, along with limited hyperspectral data availability, constrained this study, our results demonstrate the potential of hyperspectral satellite data and CWT for peatland soil moisture monitoring. Future validation across diverse peatland types and restoration conditions would further strengthen these findings.