G44A-01 INVITED
Satellite and In Situ Observations of Regional Sea Level Change: What can they tell us about future changes?
Satellite altimetry has provided precise globally-distributed sea level measurements since the early 1990s. With 15 years of measurements now in hand, we can start to examine the regional patterns of the trend in sea level change over this interval. In situ ocean termperature measurements (XBTs, ARGO) show that the geographic variations in sea level change are dominated by variations in thermosteric sea level change. However, the melting of mountain glaciers and the polar ice sheets are also expected to create distinct patterns of regional sea level change. While these have yet to be detected in the altimeter data, satellite gravity missions such as GRACE provide a means of monitoring the ice melt contributions directly. With these satellite tools, knowledge of how the sea level responds to each contribution (thermal expansion, mountain glaciers, Greenland, etc.), and output from coupled climate models, we can begin to put together a framework for predicting what regional sea level change will look like in the future. We will discuss the current limitations of such predictions, and what other information is needed to enhance the reliability of these estimates. A critical missing piece required for these predictions is knowledge of how land motion contributes to the relative sea level change seen along the coasts. http://sealevel.colorado.edu
G44A-02 INVITED
Regional variability of sea level trends
Satellite altimetry has allowed precise mapping of the geographical variability of the rates of sea level change and showed that sea level is far from rising uniformely. In some regions, rates are up several times the global mean rise, while in other regions sea level is falling. Like observed sea level trends, observed and model-based ocean thermal expansion trends are not spatially uniform and at least for the recent years, closely resemble the sea level trends patterns. Such a result suggests that non uniform thermal expansion (i.e. non uniform ocean heat content) is the main cause of observed sea level regional variability. However, salinity changes are not negligible at regional scale. In many oceanic regions, temperature and salinity contributions compensate each other and it is indeed the net effect that is responsible of the patterns seen in altimetry-derived sea level trend maps. Self- gravitational effects of present-day meltwater over the oceans also contribute to the spatial trend patterns. The spatial patterns observed in the rates of thermal expansion seem to be subject to high interannual/decadal variability related to ENSO, PDO and NAO variability. Is this also true for sea level, considering that temperature and salinity partially compensate at regional scale? In this presentation, we discuss the stationarity of the spatial trend patterns over the past few decades in sea level, thermal expansion and halosteric effects using observations, global ocean circulation models outputs and sea level reconstructions. The question we intend to address is : are the altimetry-derived spatial patterns transient structures or do they reflect long-term regional trends? This is an important topic for identifying the most vulnerable regions affected by sea level rise.
G44A-03 INVITED
The Impact of the Annual Continental Water-Storage Cycle on Coastal Sea-Level Variations
Geographic variations in coastal sea-level change, as observed by tide gauges, are driven not only by ocean dynamics and freshwater flux, but also crustal motion and equipotential height variations caused by varying mass loads on the continents. These patterns of sea-level change have been used in the past to infer the mass balance of the large ice sheets. However, GRACE observations suggest that even larger amplitude geographic variations may be produced on shorter time scales by mass changes associated with the hydrological cycle. In this talk, we examine the impact of the hydrological cycle on tide-gauge observations, focusing only on the static variations in sea level. Previous studies have shown that the non-steric, globally-averaged, annual sea-level change is primarily due to mass exchange between the continents and the oceans. This does not imply, though, that the sea level varies uniformly. Indeed, large regional variations in this signal exist along the coasts, depending upon the phase difference between the local water storage cycle and the mean global ocean signal. During late summer, when the annual ocean cycle is at its maximum and the water stored in most of the Northern Hemisphere is at a minimum, the local crustal uplift and equipotential subsidence due to the decrease of mass in the northern latitudes cancels the impact of the increase water volume in the oceans. However, when the ocean signal and hydrological signal are in phase, the loading effects and increased mean sea level contribute to a sea-level annual cycle amplitude of up to 20~mm. In particular, we focus on regions where this signal is the largest, such as the Bay of Bengal and the South China Sea. The results also demonstrate the importance of not assimilating the entire signal present in tide-gauge records into ocean models.
G44A-04 INVITED
On the use of Local Sea Level Scenarios for Managing and Mitigating the Impact of Coastal Inundation
Coastal inundation is increasingly recognized at national and international levels as an issue with potentially extreme societal impact. Consequently, there is an urgent need for decision-support tools that would help to manage and mitigate the impacts of coastal inundation, storm surges, and human activities on coastal communities and ecosystems. Decision making with respect to mitigation in the coastal zone is an extremely complicated issue for various reasons, including but not limited to: (i) The time scales involved are long from a human perspective, with coastal engineering typically dealing with infrastructure with a life time of 50 to 200 years. (ii) The economic scale of the problem is extreme: For example, the costs for increasing the height of the coastal dikes in Germany by 1 m are estimated to be of the order of 300 billion Euro; the flood gates being built in Venice are an estimated 5 billion Euro. The scale of the required investments is often seen as prohibitive for precautionary action without solid scientific basis, and failing to invest where needed may lead to large economic losses as demonstrated in New Orleans. (iii) Coastal zones are a magnet for human activities (one could say that society tends to put its "jewelry" in the coastal zone): the main increase in vulnerability in the coastal zone is not expected to come from increased hazards due to climate change but rather from increased risks due to continuing migration of population into the coastal zone and an associated increase in key infrastructure. Decisions on mitigation and adaptation in the coastal zone are likely to affect the life and prosperity of people in the future. Reliable and precise predictions of coastal inundation risks, for example through local sea level rise, would be invaluable for decision support. However, considering the aleatory and epistemic uncertainties in the processes that contribute to the hazards and risks in coastal zones over the 50 to 100 year time scale, accurate predictions cannot be made. What can be provided at best are reasonable scenarios, which describe a set of plausible trajectories based on the best information available about the present trends and specific assumptions about future evolution of the system. Scenarios thus give a better indication of the range of plausible futures than analyses based solely on aleatory uncertainty of present trends. Unfortunately, applying this approach to the coastal zone often emphasizes the large uncertainties and wide range of plausible futures (particularly if a realistic variety of assumptions is considered). We have developed an observation-based approach to scenarios for future local sea levels which allows us to consider a wide range of assumptions concerning the main contributions (vertical land motion, steric changes in the ocean volume, atmospheric circulation changes, and ocean-ice mass exchange) and thus to assess the full range of plausible futures a given location might be facing, including the associated uncertainties. We will demonstrate the approach for three example locations (Venice, New Orleans, Boston) and discuss the relative weight of the uncertainties in the forcing factors at these locations. Communicating the range of plausible futures and the uncertainties to decision makers in a proper way is a key problem that we as scientists too often tend to ignore.
G44A-05
High Subsidence and Low Elevation in New Orleans
On-going problems in New Orleans related to recovery from Hurricane Katrina highlight important issues related to the role of Earth science in society. Earth science has not played a significant role in New Orleans' current recovery or past planning, but probably should. Low elevations and subsidence are clearly key factors in past and future flood hazard, but have not yet had a major influence on zoning or rebuilding guidelines for the city. One problem is that the Earth science community has not spoken with a clear voice on this topic. A recent AGU report on New Orleans by a group of experts concluded that "Presently, there is considerable discussion and debate among the scientific community regarding mechanisms and rates of subsidence in the Mississippi delta area ("Hurricanes and the U.S. Gulf Coast…", http://www.agu.org/report/hurricanes/). Here I argue that this assessment is flawed: there is clear evidence of rapid subsidence in the city, and this subsidence explains the current pattern of low elevation. Our community's failure to issue clear statements on this relatively simple problem has confused the public. In particular, we have mixed legitimate scientific discussion of the larger issue of relative sea level change over a range of time scales and over a broad section of the Gulf Coast, a complex problem, with the narrower and more easily understood issue of subsidence in New Orleans over the last 100- 150 years. In the latter case, geodetic techniques (leveling, GPS, INSAR) indicate high subsidence rates, up to 15-25 mm/yr in some areas. These high rates, presumably applicable for the period after major drainage projects began in the mid 1800's, explain the current low elevations for most parts of the city. Assuming elevations close to sea level prior to major drainage, 20 mm/yr subsidence for 100 years yields 2.0 meters elevation below sea level, typical of regions that experienced flooding from Hurricane Katrina. While several processes contribute to high subsidence rate and low elevation, oxidation of organic-rich soils in former marshland, drained for agriculture or urbanization, is probably the major contributor in the lowest-lying areas of the city. These areas have high flood hazard, and should either be converted to other uses, or re-built with elevated structures.
G44A-06
Spatial Variations of Subsidence, Uplift, and Sea-level Rise in Western Canada and Northwestern U.S.: Implications for Coastal Communities
Sea-level rise in the Northeast Pacific Basin is a significant source of hazard to the coastal urban centers and infrastructures in western Canada and northwestern U.S. Crustal vertical motions, and hence relative sea-level (RSL), are expected to vary considerably across the 2000 km-long coast as a result of spatial variations in active tectonics and Holocene post-glacial rebound. Thus, tide gauge data from this region has so far been rejected from most global analyses to avoid the effects of uncorrected vertical motions. Recent studies suggest large spatial variations in the 20th Century rates of sea-level rise raising the issue of applicability of global mean eustatic estimates to the Northeast Pacific Basin. To address this question, we analyze a subset of about 35 collocated or nearly located continuous GPS and tide gauge sites. Data from Canadian and U.S. tide gauges are quality controlled, filtered, and corrected for common seasonal and inter-annual signals to derive robust RSL trends. In particular, we estimate linear trends and associated standard errors that account for the spatially and temporally correlated characteristics of the RSL data. Vertical velocities at the GPS sites are derived in the ITRF2000 reference frame. Alternative reference frames (e.g., IGb00, ITRF2005) and other GPS-specific scale issues are considered as part of the overall uncertainties in the absolute GPS vertical velocities. The combined tide gauge/GPS analyses indicate a 20th Century Northeast Pacific rate of sea-level rise of 1.5-2.0 mm/yr, in good agreement with global average eustatic estimates. Based on our assessments of coastal vertical motions from GPS and tide gauge data, we derive maps of predicted RSL rise over the next 50 and 100 years in western Canada and northwestern U.S. Large-scale sea-level rise predictions are based on IPCC scenarios, but the main factors controlling the spatial RSL patterns are the uplift and subsidence variations along the coast. In particular, we show that tectonic and sediment-loading subsidence in the Puget Lowland-Strait of Georgia area have the potential to significantly aggravate the predicted RSL rise in Greater Vancouver and Seattle-Tacoma. Conversely, tectonic uplift along parts of the west coast can compensate for about half of future sea-level rise.
G44A-07
Unraveling the Complexity of Sea-Level Trends in North America: Results From a New Generation of Glacial Isostatic Adjustment Models
Estimating the impact of global change on regional sea-level trends within North America requires that observations be corrected for a suite of geophysical processes, most notably ongoing glacial isostatic adjustment (GIA). Unfortunately, the `decontamination' of the observations using numerical models of the GIA process is problematic, both because of uncertainties in our understanding of the Late Pleistocene ice history and viscoelastic Earth structure and as a consequence of assumptions adopted in the underlying post-glacial sea-level theory. An archetypal example of this issue was the long-standing suggestion that GIA-corrected tide gauge rates along the U.S. east coast were characterized by anomalous geographic trends. These trends were interpreted as reflecting, for example, oceanographic and neotectonic signals, and they were sometimes cited to justify ignoring the regional rates in estimates of global sea-level rise. However, subsequent work demonstrated that moderate changes in the radial viscosity profile adopted in the GIA predictions could nearly eliminate the trends. In this talk, we present sea-level predictions along the eastern and Gulf coasts of the U.S. based on a new generation of high-resolution GIA models. These models are based on a finite element numerical framework and they incorporate an accurate treatment of shoreline migration and surface loading processes, as well as realistic 3-D Earth structure. The latter includes lateral heterogeneity in mantle viscosity as well as transition in lithospheric strength from continent to oceanic environments. These new aspects of the modeling are shown to have a large effect on peripheral bulge and continental levering dynamics associated with GIA, and hence conclusions based on earlier modeling results (e.g., the regional GIA-corrected sea-level rate) may require significant revision.
G44A-08
Nonlinear trends and multi-year cycles in regional and global sea level records
We analyze the Permanent Service for Mean Sea Level (PSMSL) database of sea level time series using a method based on Monte Carlo Singular Spectrum Analysis (MC-SSA). We remove 2-30 year quasi- periodic oscillations and determine the nonlinear long-term trends for 12 large ocean regions. Our global sea level trend estimate of 2.4 ± 1.0 mm/yr for the period from 1993 to 2000 is comparable with the 2.6 ± 0.7 mm/yr sea level rise calculated from TOPEX/Poseidon altimeter measurements. However, we show that over the last 100 years the rate of 2.5 ± 1.0 mm/yr occurred between 1920 and 1945, is likely to be as large as the 1990s, and resulted in a mean sea level rise of 48 mm. We evaluate errors in sea level using two independent approaches, the robust bi-weight mean and variance, and a novel "virtual station" approach that utilizes geographic locations of stations. Results suggest that a region cannot be adequately represented by a simple mean curve with standard error, assuming all stations are independent, as multi-year cycles within regions are very significant. Additionally, much of the between-region mismatch errors are due to multi-year cycles in the global sea level that limit the ability of simple means to capture sea level accurately. We demonstrate that variability in sea level records over periods 2-30 years has increased during the past 50 years in most ocean basins.