Comparing altimetry-derived coastal sea level anomaly with tide gauge observations along the global coastline by accounting for shelf-width 

Abstract

It is expected that sea level rise and resulting coastal flooding will cost us over 1 trillion dollars annually by 2050. Therefore, understanding and monitoring coastal sea level rise is vital. Tide Gauges are in-situ instruments that have been providing sea level measurements since the 1800s, but they are sparse, and data availability is limited. Therefore, ocean altimetry has been the preferred observational tool for monitoring global sea levels.   Satellite altimetry has been providing extensive and continuous global sea level data for more than three decades now. However, extracting reliable data close to the coast has been problematic due to signal contamination from land or calm water and lack of accurate geophysical corrections. Recently dedicated coastal altimetry products were proposed to provide better coastal sea level change product.    In this study, we compare coastal altimetry products XTRACK-1Hz, XTRACK/ALES-20Hz in observing Sea Level Anomalies (SLA) with Tide Gauges (TG) along the global coastline from 2002-2019. 458 stations were selected for the study after applying several selection criteria that address data gaps, data availability from TG, altimetry, and correction products. The SLA signals from TG were decomposed into non-linear trend, seasonal, and residual components using Seasonal-Trend Decomposition using Loess (STL) method. The correlation coefficient, Root Mean Square Error (RMSE), and Index of Agreement (IOA) were computed for interannual and residual signals from TG and coastal altimetry products. Linear sea level trends at each station were also estimated from altimetry and TG observations after correcting for GPS-derived vertical land motion (VLM).  When using altimetry for sea level signals near the coast, it is important to select point observations carefully instead of using a search radius that may take points from adjacent regions that could behave differently due to different coastal ocean processes. We developed a dynamically varying search radius for each TG, a function of the coastal shelf width near that station, to collate satellite observations as a representative of coastal sea level change. All the altimetry observations that fall within the search radius and are less than 25 km along the coast are used for comparison. In several cases, due to the sharp changes in the coastal morphology, the sea level signals seen by the adjacent TG stations are quite different, and thus, the reliability of altimetry suffered.  With our analysis approach, we found good agreement between all altimetry products (XTRACK, XTRACK/ALES), and TG at residual and non-linear trend scales. A few stations near the fault lines and other tectonic regions disagree with altimetry trend estimates due to strong VLM signals that are not completely resolved by the VLM product used for correction. Around 70% of stations had a good agreement (r > 0.7) with trend and 55% with residual components. High-resolution (20Hz) XTRACK/ALES provided more observations near the coast. Nevertheless, both XTRACK/ALES-20Hz and XTRACK-1Hz performed well. This novel approach to select representative observation points from altimetry for a coastal zone will provide improved coastal sea level products from satellites, which can be considered at par with TG observations.