Estimation and validation of InSAR-derived surface displacements at temperate raised peatlands

Peatland surface motion derived from satellite-based Interferometry of Synthetic Aperture Radar (InSAR) is potentially a proxy for groundwater level variations and greenhouse gas emissions from peat soils. Ground validation of these motions at equivalent temporal resolution has proven problematic, either because of limitations of traditional surveying methods or because of limitations with past InSAR time-series approaches. Novel camera-based instrumentation has enabled in-situ measurement of peat surface from mid-2019 to mid-2022 at two large temperate raised bogs undergoing restoration – Cors Fochno and Cors Caron, in mid-Wales, United Kingdom. The cameras provided continuous measurements at sub-millimetre precision and sub-daily temporal resolution. From these data and Sentinel-1 acquisitions spanning mid-2015 to early-2023, we demonstrate that accurate peat surface motion can be derived by InSAR when a combination of interferometric networks with long and short temporal baselines is used. The InSAR time series data closely match the in-situ data at both bogs, and in particular recover well the annual peat surface oscillations of 10-40 mm. Pearson's values for the point-wise correlation of in-situ and InSAR displacements are 0.8–0.9, while 76% of differences are < ±5 mm and 93% are < ±10 mm. RMSE values between multi-annual in-situ and InSAR peat surface displacement rates are ~7 mm·yr−1 and decrease to ∼3.5 mm for individual peat surface motion measurements. Larger differences mainly occur during drought periods. Multi-annual displacement velocities rates based on InSAR indicate long-term subsidence at Cors Caron (maximum −7 mm·yr−1), while Cors Fochno exhibits subsidence at the centre and uplift at the margins (−9 mm·yr−1 to +5 mm·yr−1). Because of the annual peat surface oscillations, however, more robust ground validation of the long-term peat surface motion rates derived from InSAR requires longer time-series of in-situ measurements than are presently available. Nonetheless, the InSAR-derived surface motion rates correlate well spatially with both peat dome elevation and peat thickness. In addition, the annual oscillations in surface motion are synchronous with or lag slightly behind groundwater level changes. A coarse ratio of 10:1 is observed between annual changes in groundwater level and peat surface displacement. Satellite-based InSAR derived from a fusion of short- and long-term temporal baseline networks can thus enable accurate monitoring of hydrologically driven surface motions of moderately degraded to intact temperate raised peatlands.