Search

Understanding historical and projected coastal sea-level change is limited because the impact of large-scale ocean dynamics is not well constrained. Here, we use a global set of tide-gauge records over nine regions to analyse the relationship between coastal sea-level variability and open-ocean steric height, related to density fluctuations. Interannual-to-decadal sea-level variability follows open-ocean steric height variations along many coastlines. We extract their common modes of variability and reconstruct coastal sterodynamic sea level, which is due to ocean density and circulation changes, based on steric height observations. Our reconstruction, tested in Earth system models, explains up to 91% of coastal sea-level variability. Combined with barystatic components related to ocean mass change and vertical land motion, the reconstruction also permits closure of the coastal sea-level budget since 1960. We find ocean circulation has dominated coastal sea-level budgets over the past six decades, reinforcing its importance in near-term predictions and coastal planning. Coastal sea levels are impacted by local vertical land motion plus local and remote changes to ocean circulation, density and mass changes. Tide-gauge records are used to reconstruct the coastal sea-level budget over nine regions, showing its variability has been dominated by ocean circulation since 1960.

Global sea level reconstruction (RecSL) for 1900–2015 was used to estimate the variations in oceanic kinetic energy (OKE) and compare OKE with changes in wind patterns and wind kinetic energy (WKE); the comparison was done for each latitude and for 5 western boundary currents (WBCs). Two contributors to variability in sea level were analyzed: gravitational, rotational and deformational effects (GRD) related to changes in water masses (barystatic sea level change), and changes in the sterodynamic sea level (SDSL), associated with changes in wind, steric sea level and ocean circulation. GRD changes were responsible for latitudinal multidecadal variations with time scale of ~ 60 to 80 years, while SDSL changes were responsible for interannual and decadal variability, and together with the Greenland ice melt, to sea level acceleration since the 1960s. Regional changes near WBCs show a coherent upward trend in OKE (+ 24% ± 3 increase per century), while trends in WKE over the same regions changed widely from -11% (decrease) over the Gulf Stream region to + 28% (increase) over the Brazil Current region. Low frequency oscillations of wind and oceanic kinetic energy are correlated in some WBCs (e.g., R = 0.5 in the Kuroshio region) but not in others (e.g., R = -0.05 in the Gulf Stream region). The study suggests that several forcing mechanisms contribute to the increased OKE, they include an increased wind-stress curl over subtropical gyres, local changes in wind patterns that impact some WBCs, and large uneven warming near WBCs that increased sea level gradients and thus intensified OKE.

A new monthly global sea level reconstruction for 1900–2015 was analyzed and compared with various observations to examine regional variability and trends in the ocean dynamics of the western North Atlantic Ocean and the US East Coast. Proxies of the Gulf Stream (GS) strength in the Mid-Atlantic Bight (GS-MAB) and in the South Atlantic Bight (GS-SAB) were derived from sea level differences across the GS. While decadal oscillations dominate the 116-year record, the analysis showed an unprecedented long period of weakening in the GS flow since the late 1990s. The only other period of long weakening in the record was during the 1960s–1970s, and red noise experiments showed that is very unlikely that those just occurred by chance. Ensemble empirical mode decomposition (EEMD) was used to separate oscillations at different timescales, showing that the low-frequency variability of the GS is connected to the Atlantic Multi-decadal Oscillation (AMO) and the Atlantic Meridional Overturning Circulation (AMOC). The recent weakening of the reconstructed GS-MAB was mostly influenced by weakening of the upper mid-ocean transport component of AMOC as observed by the RAPID measurements for 2005–2015. Comparison between the reconstructed sea level near the coast and tide gauge data for 1927–2015 showed that the reconstruction underestimated observed coastal sea level variability for timescales less than ∼5 years, but lower-frequency variability of coastal sea level was captured very well in both amplitude and phase by the reconstruction. Comparison between the GS-SAB proxy and the observed Florida Current transport for 1982–2015 also showed significant correlations for oscillations with periods longer than ∼5 years. The study demonstrated that despite the coarse horizontal resolution of the global reconstruction (1∘ × 1∘), long-term variations in regional dynamics can be captured quite well, thus making the data useful for studies of long-term variability in other regions as well.

Sea level rise (SLR) is causing acceleration in the frequency and duration of minor tidal flooding (often called “sunny-day” or “nuisance” flooding) along the U.S. East Coast. Those floods have a seasonal pattern that often follows the monthly mean sea level anomaly which peaks in September–October for stations between New York and south Florida. However, there are large differences between coasts: for example, over 75% of the minor floods occur during the fall in Florida, but during the spring and winter in Boston. Various data and forcing, such as tide gauge records, surface temperatures, winds, long-term tidal cycles, and the Gulf Stream flow, were analyzed to examine potential drivers and mechanisms that can contribute to the seasonal pattern of floods. The seasonal water temperature cycle, with maximum temperatures in August, could not by itself explain the seasonal sea level pattern, but two mechanisms that significantly correlate with the seasonal sea level and flooding patterns are the annual and semi-annual tidal cycles (correlation of ~ 0.97) and changes in the Gulf Stream (GS) flow (correlation of − 0.6; the GS shows a maximum decline in September–October during the period of peak flooding). The combination of the seasonal pattern of tropical storms and high coastal sea level can explain the high frequency of fall flooding along the Southeastern U.S. coasts, while winter storms have more influence on the northeastern coasts. In recent decades however, the seasonal pattern seemed to have shifted so that increased flooding is seen on the northeastern coasts during spring and summer, while in the Mid-Atlantic and southeastern coasts, a dramatic increase in flooding is seen almost exclusively during the fall. A long-term change in the mean zonal wind pattern along the coast can contribute to the recent shift in the seasonal flooding pattern. The study can help regional adaptation and resilience planning for flood-prone coastal cities and communities.

Global sea level rise (SLR) associated with a warming climate exerts significant stress on coastal societies and low-lying island regions. The rates of coastal SLR observed in the past few decades, however, have large spatial and temporal differences from the global mean, which to a large part have been attributed to basin-scale climate modes. In this paper, we review our current state of knowledge about climate modes' impacts on coastal sea level variability from interannual-to-multidecadal timescales. Relevant climate modes, their impacts and associated driving mechanisms through both remote and local processes are elaborated separately for the Pacific, Indian and Atlantic Oceans. This paper also identifies major issues and challenges for future research on climate modes' impacts on coastal sea level. Understanding the effects of climate modes is essential for skillful near-term predictions and reliable uncertainty quantifications for future projections of coastal SLR.

The study is focused on the disruption that a storm can cause to ocean stratification and ocean currents in a region dominated by a western boundary current and meso-scale variability. Sensitivity experiments with a regional numerical ocean model of the US East Coast are used to simulate different hurricane tracks to study the impact on the Gulf Stream (GS) flow, surrounding waters and coastal sea level. Realistic simulations of Hurricane Matthew (October 2016) using surface wind and heat flux from NOAA’s operational coupled forecast system were compared with idealized artificial hurricanes with tracks located at different distances from the coast (~ 200–600 km). Despite the limitation of representing realistic wind patterns by an analytical formula, coastal storm surge near the hurricane was simulated quite well. The height of the coastal storm surge was found to be very sensitive to the location of the hurricane track relative to the coast, but the impact of the hurricane on the GS flow was found to be less sensitive to the exact hurricane track, though the maximum influence was when the hurricane track passed ~ 100 km east of the GS with winds over the GS opposing the current direction. Hurricanes that passed within hundreds of kilometers from the GS caused disruption in the GS dynamics and weakening in the downstream flow of the GS that can last for many days after the storm disappeared. This indirect impact of hurricanes on the GS can elevate sea level along long stretches of the coast. The impact of a hurricane on a region dominated by meso-scale variability is complex, creating unpredictable spatial changes in temperatures and currents. After the hurricane disappeared and without additional surface heating, it may take the stratification as much as 2 months to recover to pre-hurricane conditions by advection alone. This lasting impact of a storm on ocean dynamics is consistent with observations that show minor tidal coastal flooding that lasts for days after hurricanes passed offshore.

Tropical storms and hurricanes in the western North Atlantic Ocean can impact the US East Coast in several ways. Direct effects include storm surges, winds, waves, and precipitation and indirect effects include changes in ocean dynamics that consequently impact the coast. Hurricane Matthew [October, 2016] was chosen as a case study to demonstrate the interaction between an offshore storm, the Gulf Stream (GS) and coastal sea level. A regional numerical ocean model was used, to conduct sensitivity experiments with different surface forcing, using wind and heat flux data from an operational hurricane-ocean coupled forecast system. An additional experiment used the observed Florida Current (FC) transport during the hurricane as an inflow boundary condition. The experiments show that the hurricane caused a disruption in the GS flow that resulted in large spatial variations in temperatures with cooling of up to ~ 4 °C by surface heat loss, but the interaction of the winds with the GS flow also caused some local warming near fronts and eddies (relative to simulations without a hurricane). A considerable weakening of the FC transport (~ 30%) has been observed during the hurricane (a reduction of ~ 10 Sv in 3 days; 1Sv = 106 m3 s−1), so the impact of the FC was explored by the model. Unlike the abrupt and large wind-driven storm surge (up to 2 m water level change within 12 h in the South Atlantic Bight), the impact of the weakening GS on sea level is smaller but lasted for several days after the hurricane dissipated, as seen in both the model and altimeter data. These results can explain observations that show minor tidal flooding along long stretches of coasts for several days following passages of hurricanes. Further analysis showed the short-term impact of the hurricane winds on kinetic energy versus the long-term impact of the hurricane-induced mixing on potential energy, whereas several days are needed to reestablish the stratification and rebuild the strength of the GS to its pre-hurricane conditions. Understanding the interaction between storms, the Gulf Stream and coastal sea level can help to improve prediction of sea level rise and coastal flooding.

The impact of sea level rise on increased tidal flooding and storm surges in the Hampton Roads region is demonstrated, using ~90 years of water level measurements in Norfolk, Virginia. Impacts from offshore storms and variations in the Gulf Stream (GS) are discussed as well, in view of recent studies that show that weakening in the flow of the GS (daily, interannually, or decadal) is often related to elevated water levels along the U.S. East Coast. Two types of impacts from hurricanes on flooding in Hampton Roads are demonstrated here. One type is when a hurricane like Isabel (2003) makes a landfall and passes near the Chesapeake Bay, causing a large but short-term (hours to a day) storm surge. The second type is when Atlantic hurricanes like Joaquin (2015) or Matthew (2016) stay offshore for a relatively long time, disrupting the flow of the GS and leading to a longer period (several days or more) of higher water levels and tidal flooding. Analysis of the statistics of tropical storms and hurricanes since the 1970s shows that, since the 1990s, there is an increase in the number of days when intense hurricanes (Categories 3‐5) are found in the subtropical western North Atlantic. The observed Florida Current transport since the 1980s often shows less transport and elevated water levels when tropical storms and hurricanes pass near the GS. Better understanding of the remote influence of the GS and offshore storms will improve future prediction of flooding and help mitigation and adaptation efforts.

In October 7–9, 2016, Hurricane Matthew moved along the southeastern coast of the U.S., causing major flooding and significant damage, even to locations farther north well away from the storm’s winds. Various observations, such as tide gauge data, cable measurements of the Florida Current (FC) transport, satellite altimeter data and high-frequency radar data, were analyzed to evaluate the impact of the storm. The data show a dramatic decline in the FC flow and increased coastal sea level along the U.S. coast. Weakening of the Gulf Stream (GS) downstream from the storm’s area contributed to high coastal sea levels farther north. Analyses of simulations of an operational hurricane-ocean coupled model reveal the disruption that the hurricane caused to the GS flow, including a decline in transport of ∼20 Sv (1 Sv = 106 m3 s−1). In comparison, the observed FC reached a maximum transport of ∼40 Sv before the storm on September 10 and a minimum of ∼20 Sv after the storm on October 12. The hurricane impacts both the geostrophic part of the GS and the wind-driven currents, generating inertial oscillations with velocities of up to ±1 m s−1. Analysis of the observed FC transport since 1982 indicated that the magnitude of the current weakening in October 2016 was quite rare (outside 3 standard deviations from the mean). Such a large FC weakening in the past occurred more often in October and November, but is extremely rare in June-August. Similar impacts on the FC from past tropical storms and hurricanes suggest that storms may contribute to seasonal and interannual variations in the FC. The results also demonstrated the extended range of coastal impacts that remote storms can cause through their influence on ocean currents.

Power draw is a complex physical response to the workload of a given application on the hardware, which is difficult to model, in part, due to its variability. The empirical mode decomposition and Hilbert–Huang transform (EMD/HHT) is a method commonly applied to physical systems varying with time to analyze their complex behavior. In authors’ work, the EMD/HHT is considered for the first time to study power usage of high-performance applications. Here, this method is applied to the power measurement sequences (called here power traces) collected on three different computing platforms featuring two generations of Intel Xeon Phi, which are an attractive solution under the power budget constraints. The high-performance applications explored in this work are codesign molecular synamics and general atomic and molecular electronic structure system—which exhibit different power draw characteristics—to showcase strengths and limitations of the EMD/HHT analysis. Specifically, EMD/HHT measures intensity of an execution, which shows the concentration of power draw with respect to execution time and provides insights into performance bottlenecks. This article compares intensity among executions, noting on a relationship between intensity and execution characteristics, such as computation amount and data movement. In general, this article concludes that the EMD/HHT method is a viable tool to compare application power usage and performance over the entire execution and that it has much potential in selecting most appropriate execution configurations.