Coupling of sea level and tidal range changes, with implications for future water levels

www.nature.com/scientificreports OPEN Received: 9 June 2017 Accepted: 10 November 2017 Published: xx xx xxxx Coupling of sea level and tidal range changes, with implications for future water levels Adam T. Devlin1,2, David A. Jay2, Stefan A. Talke2, Edward D. Zaron2, Jiayi Pan1,3,4 & Hui Lin1 Are perturbations to ocean tides correlated with changing sea-level and climate, and how will this affect high water levels? Here, we survey 152 tide gauges in the Pacific Ocean and South China Sea and statistically evaluate how the sum of the four largest tidal constituents, a proxy for the highest astronomical tide (HAT), changes over seasonal and interannual time scales. We find that the variability in HAT is significantly correlated with sea-level variability; approximately 35% of stations exhibit a greater than ±50 mm tidal change per meter sea-level fluctuation. Focusing on a subset of three stations with long records, probability density function (PDF) analyses of the 95% percentile exceedance of total sea level (TSL) show long-term changes of this high-water metric. At Hong Kong, the increase in tides significantly amplifies the risk caused by sea-level rise. Regions of tidal decrease and/or amplification highlight the non-linear response to sea-level variations, with the potential to amplify or mitigate against the increased flood risk caused by sea-level rise. Overall, our analysis suggests that in many regions, local flood level determinations should consider the joint effects of nonstationary tides and mean sea level (MSL) at multiple time scales. Worldwide, ocean tides have exhibited trends over the past century that cannot be explained by the orbital mechanics underlying predictable tidal variability1,2. Regional studies have shown changes in major diurnal and semidiurnal tides in the Eastern Pacific3, in the Gulf of Maine4, in the North Atlantic5,6, in China7,8, in Japan9, and at islands throughout the Pacific10. Alongside these changes, mean sea level (MSL) has been increasing at a global rate of +1.7 ± 0.2 mm yr−1 as estimated from coastal and island tide gauge measurements from 1900–2009, and +3.4 ± 0.4 mm yr−1 for 1993–2016 as estimated from satellite altimetry11,12. A recent calculation shows the satellite era rate to be closer to +3.0 mmyr−1, though the last decade showed a faster rise, presumably from increased ice melt in Greenland13. However, this increase is not spatially uniform; Western Pacific MSL rise rates exceed +10 mmyr−1 in some locations, whereas Eastern Pacific rates are often zero or negative14,15. Climate models predict that MSL rates will accelerate in upcoming decades through global climate change mechanisms16 such as ice sheet melt and thermosteric MSL rise due to upper-ocean warming17–24. Both MSL and tides exhibit short-term (seasonal to decadal) variability, in addition to long-term trends25. Tide properties can be altered by depth changes26 including those caused by MSL rise27, or by mechanisms driving MSL rise, such as increased surface water temperatures28. In either case, societal consequences are often due to total sea level, which is the sum of MSL and the time-variable water level (including tides). Therefore, both tides and MSL should be considered together in assessing the significance of future sea-level changes29. Figure 1 shows a simple cartoon of the various mechanisms that can affect MSL and tides. Changes in water depth (ΔH) can be due to changes in density (Δρ), which is a function of water temperature (Tw). An increase in Tw will decrease the density of the upper layer (ρu), increasing ΔH17. Altered frictional effects (Δr) can manifest from changes in ΔH27. The “coupled oscillator” effect between the shelf regions and the open ocean can lead to modifications of tidal amplitudes via changing Δr and frequency-dependent tidal responses to astronomical forcing (ΔΨω) for multiple tides29. Internal tides can change via modulations of ΔH and Δρ, which can change the wave phase relative to the barotropic tide, which can modulate the surface expression of the tidal amplitude30. Resonance is a function of ΔH and ΔΨω; small changes in ΔH may change the resonance of enclosed bays 1 The Chinese University of Hong Kong, Shatin, New Territories, Hong Kong, SAR, China. 2Portland State University, Portland, OR, United States of America. 3College of Marine Science, Nanjing University of Information Science and Technology, Nanjing, 210044, China. 4Shenzhen Research Institute, The Chinese University of Hong Kong, Shenzhen, Guangdong, 518057, China. Correspondence and requests for materials should be addressed to A.T.D. (email: ) SCIeNTIfIC REPorTS | 7: 17021 | DOI:10.1038/s41598-017-17056-z 1 www.nature.com/scientificreports/ Figure 1. Schematic cartoon showing some of the mechanisms that can affect MSL and tides. See text above for details. Relevant reference citation numbers are indicated in figure by superscript. and harbors through modification of ΔΨω that may either increase or decrease tidal amplitudes31. Water depth changes are also dependent on ΔQr in estuarine locations. An increase in ΔQr will certainly increase MSL locally, but the increased friction of the incoming tide interacting with the outgoing river discharge will change Δr, decreasing tidal amplitude32. Finally, on a larger scale, changes in ΔH can also affect the structure of basin-wide amphidromic systems via movement of the co-tidal lines33. Another mechanism not shown on this figure is that of resonant triad interactions, such as between the M2, K1, and O1 tides, which can be amplified through changes in ΔH and Δρ that change local geometry and modulate all three tidal amplitudes in the relationship25. The correlations of multiple forcing mechanisms make it challenging to determine the causes behind the linked variability of MSL and tides that is observed at many locations34,35. However, the presence of correlations between sea-level and tides indicates that one or more of these processes is at work. In summary, previous studies suggest that changes in the amplitude and phase of a tidal constituent may be a function of multiple variables: ΔAmptidal = f (ΔH , ΔQr , Δρ , Δmx , Δr, ΔΨω , ...) (1) where H is the water depth, ρ is water density, r is friction, mx is mixing, Qr is river discharge, and Ψω represents frequency-dependent tidal response to astronomical tidal forcing. The “…” indicates other variables not listed here, such as wind. Since many of the variables which affect sea-level (river flow, density, wind) can also affect tidal amplitudes, a correlated response is frequently observed. Identifying connections and correlations between tidal range and MSL is critical for making reliable predictions of coastal water levels and inundation risk. When combined with storm surge, larger tides and higher MSL may amplify flood risk, coastal inundation, damage to infrastructure, and population displacement. Even without the consideration of storm surge, change (...truncated)