Comparing water level estimation in coastal and shelf seas from satellite altimetry and numerical models

Accurately resolving coastal Total Water Levels (TWL) is crucial for socio-economic and environmental reasons. Recent efforts in satellite altimetry and numerical modelling have improved accuracy over near-shore areas. In this study we used data from tide gauges (TGs), SAR-mode altimetry from two satellites (Sentinel-3A (S3) and CryoSat-2 (C2)), and a state-of-the-art high-resolution regional coupled environmental prediction model (Amm15) to undertake an inter-comparison between the observations and the model. The aim is to quantify our capability to measure TWL around the UK coast, and to quantify the capacity of the model to represent coastal TWL. Results show good agreement between the satellite and TG data (the mean correlation (R) over seventeen TGs between June 2016 and September 2017 is 0.85 for S3 and 0.80 for C2). The satellite-model comparison shows that the variability is well captured (R=0.98 for both satellite), however there is an offset (-0.23m for S3, -0.15m for C2) between the satellite and model data, that is near-constant across the domain. This offset is partly attributed to the difference in the reference level used by the satellites and the model, and residual differences linked to the temporal resolution of the model. The best agreement between model and satellite is seen away from the coast, further than 3-4km offshore. However, even within the coastal band, R remains high, ~0.95 (S3) and ~0.96 (C2). In conclusion, models are still essential to represent TWL in coastal regions where there is no cover from in-situ observations, but satellite altimeters can now provide valuable observations that are reliable much closer to the coast than before.