Monitoring water level dynamics in neotropical peatlands with earth observation data

Intact tropical peatlands are globally important carbon stores, yet their hydrology remains poorly understood due to limited accessibility and sparse field measurements. In this study, we evaluate the potential of L-band Synthetic Aperture Radar (SAR) backscatter to monitor above-ground water level variation across diverse lowland peatland ecosystems in Colombia and Peru. Using vegetation structure metrics from GEDI with ancillary remote sensing data, we assess the sensitivity of L-band HH (L-HH) backscatter to water level changes. We observed significant linear correlations between water level and L-HH backscatter in white-sand ecosystems, palm swamp peatlands (open and forested) and seasonally flooded forests. Pole forest peatland water levels showed no correlation with L-HH backscatter. To predict these regressions, we developed ecosystem-specific multiple linear regression models using L-band HV backscatter, NDVI, and GEDI metrics, achieving strong predictive performance (R 2 = 0.8–0.94). We further tested the temporal robustness of these relationships by predicting water levels across different years. Our results demonstrate the potential of combining L-band SAR with vegetation metrics derived from spaceborne data for regional monitoring of peatland hydrology. This provides a methodological pathway for integrating tropical peatland dynamics into carbon cycle models. • L-band SAR HH backscatter correlates with water level in tropical peatland sites. • Vegetation structure controls L-band SAR sensitivity to hydrological changes. • Ecosystems with low vegetation density show strong monitoring potential. • Palm swamps and flooded forests exhibit good monitoring potential with L-band SAR. • Our approach supports the integration of peatland hydrology into carbon cycle models.