Sea Level Rise
Resilience -- the capacity and capability to proactively counter hazardous events -- is highly dependent on clear information on the potential magnitude of those events. To develop the resilience of coastally situated cities information on local sea level rise is of utmost importance. The keyword here is “local”. Sea level rise around the planet is not homogenous, neither in space nor time, and detailed local information is hard to get, especially when other relevant processes like local vertical land movement need to be taken into account.
The inputs for SLR for the Adaptation Atlas expose information from global climate models combined with datasets on vertical land movement on a local level, and integrates this with local population information (which clearly shows the extent of coastally oriented cities), offering opportunities for data presentation previously unavailable to a wide audience.
Data and Methods
The GCM data used in the Adaptation Atlas is from CMIP5 (see http://cmip-pcmdi.llnl.gov/cmip5/guide_to_cmip5.html). It is important to note that not all models available in the CMIP5 database contain data on sea level rise.
The monthly changes in SLR of 28 GCMs between 1986-2005 and 2081-2100 under the RCP4.5 scenario were calculated using a pattern scaling method.
GCM data were retrieved from the Earth System Grid (ESG) data portal for CMIP5, including: the sea surface height (‘zos’), the global average thermosteric sea level change (‘zostoga’) and the global average sea level change (’zosga’) under the RCP4.5 scenario. For some GCMs, only ‘zosga’ was available and was therefore used instead of ‘zostoga’. The data availability is shown in Table 1. More information on the variables applied can be found at http://www.climatechange2013.org/images/report/WG1AR5_Ch13SM_FINAL.pdf and http://www-pcmdi.llnl.gov/ipcc/standard_output.pdf.
The change in sea surface height was scaled by the change in global average thermal expansion per month. The sea surface height from the GCMs includes the regional variability due to changes in water mass advection, thermohaline circulation, and wind-driven circulation, but does not include the tidal effects. The changes in global average thermal expansion were calculated by the changes in ‘zostoga’ if available, or by the changes in ‘zosga’.
The value X (cm/cm) in the normalized patterns is interpreted as “regional sea level rises X cm when the global average sea level rises 1 cm”. When the local SLR is faster than the global average, X>1, if it is slower, X<1.
An important element in the calculation of projected maximum sea level rise, is accounting for variations over the year which can change under climate change. This seasonal aspect has not been taken into account in the Adaptation Atlas data but can be found by exploring the CLIMsystems sea level rise for cities app http://slr-cities.climsystems.com/. It is computed by comparing the month with the highest current level with the month with the highest future level. The actual months could be different in each location.
The deviation from the yearly mean for the baseline runs between 12cm higher in February and 13cm lower in August. The changes in this variation by 2100 are small, then varying between +15cm and 16cm. The additional 3cm for the maximum month (February) have already been taken into account in the calculation of the 113cm SLR by 2100 for this location. In some location this approach adds more than 5% to the projected SLR.
The CLIMsystems sea level rise for cities web app depicts a global map of the combined processes of local (absolute) sea level rise and local vertical land movement. The sea level rise values are taken as the median value of an ensemble of 28 GCM’s, under the assumption of the largest greenhouse gas emissions as described by the RCP8.5 scenario in AR5. It also assumes a high climate sensitivity.
Processing of Vertical Land Movement Data
Vertical land movement (VLM) is a generic term for all processes that impact the elevation at a given location (tectonic movements, subsidence, groundwater extraction), causing land to move up or down. This is typically a slow process with magnitudes commonly between -10 (sinking) and +10 (rising) mm/year.
Local vertical land movement becomes relevant when looking at the local effects of sea level rise and hence has been included when generating the patterns for sea level rise for the Adaptation Atlas. The orders of magnitude are comparable, and VLM can thus either exacerbate or dampen the sea level rise experienced at a coastal location. In a place where VLM is upward (rising, like Norway), the local experienced SLR is smaller (local SLR can even be negative: sea level going down). When VLM is downward (sinking, like Manila), locally experienced SLR is stronger.
Because of its (potential) magnitude local VLM must be considered when sea level rise effects are determined on a local scale.
Note that local sea level rise is usually different from the global mean (regardless of VLM), because of variations in currents, the amount of heating of the sea water (responsible for thermal expansion component of overall sea level rise), as well as the volume (depth) of the sea water affected. This is expressed in the normalized change patterns extracted from GCM-results.
Vertical land movement can be observed directly, or inferred from related measurements.
Direct observations (SONEL)
Direct observations are available through the SONEL initiative (http://www.sonel.org/) whereby VLM is estimated from continuous GPS measurements at fixed locations, often coinciding with tidal observation stations. The latest set of “solutions” (http://www.sonel.org/-GPS-Solutions-.html?lang=en) contains location (lat/lon) and estimates of VLM (mm/year). As there are requirements for determining the trend (length of the period, completeness, quality, stability of the solution), not all stations have an associated value. With time more and more solutions will become available.
Indirect observations (PSMSL)
The Permanent Service for Mean Sea Level (http://www.psmsl.org/) maintains an archive of observed tides. An analysis of the data for these stations to estimate their trends (which are reflecting the rise in sea level), is also available (http://www.psmsl.org/products/trends/trends.txt). Note that not all stations meet the requirements of completeness, total number of observations and quality of measurements.
The local observed sea level rise and local vertical land movement have the following relation:
local observed SLR = local absolute SLR – local VLM
(with VLM>0 means that land is rising, VLM<0 land is sinking)
local absolute SLR = global SLR (over the observation period) * local normalized value (from an ensemble of GCMs)
To determine the global SLR over the period that the tidal observations were made, the following curve is used.
The curve is fitted to data generated by Church & White, which can be downloaded from http://www.psmsl.org/products/reconstructions/ and http://www.cmar.csiro.au/sealevel/sl_data_cmar.html.
Note: The longest part of the global curve is based on tidal observations (up to 1992, after 1992 satellite observations are used). In order to do that, assumptions needed to be made about the local VLM at each tidal station. A global model (mostly for tectonic movements) was used to do this. This creates a “thinking loop” as we are trying to estimate local VLM from data that has been corrected with a modelled VLM. The assumption is that the averaging of the data around the globe minimizes this bias.
From points to raster (IDW, parameters)
To be able to use VLM in places where it has not been observed, the VLM values in the (SONEL or PSMSL) point locations needed to be interpolated spatially over a grid. ArcGIS has multiple models for spatial interpolation of point values which were tested on their performance noting:
- Ease of use (parameters and inputs needed)
- Ability to reproduce known values at point locations
- Overall “look-and-feel” of the output (homogeneity, reproduction of expected behaviour [like land rising close to the Arctic region, due to rebound from retracting land-ice]))
The conclusion is that the IDW (Inverse Distance Weighted) model is the most useable. The following parameter choices were made:
Note: The description of the IDW tool is unclear on how it deals with coordinate system issues. As the station locations are in LAT/LON, the distances between the stations are not simple equations. It is assumed ArcGIS deals with this issue. If this is not the case, there is a bias in the inverse distance weighing. As most stations are relatively close to each other, this will only result in a small error.
The IDW tool does not wrap around the globe (crossing the 360° to 0° meridian). This was managed by executing the IDW tool twice (only for the combined result, see below): once with the 180° meridian centered and once with the 0° meridian centered (the longitudes of the observation stations were either mapped on 0° to 360°, or on -180° to 180°). The resulting images were joined using the following scheme:
Combing the SONEL and PSMSL datasets results in the following image (contour lines in mm/year) (with wrap-around correction):
Note that as all tidal stations (green), and most continuous GPS stations (red) are located along coastlines, resulting in this area having the most reliable estimate of local VLM. This is also where local sea level rise, corrected for local vertical land movement, is the most relevant for assessing the risks to coastal cities.
Note: as there are many smaller islands which are not part of the very coarse coastline definition shape file that was used (from NOAA’s National Geophysical Data Centre, http://www.ngdc.noaa.gov/mgg/shorelines/gshhs.html), it is still possible to click and retrieve values in the obfuscated sea areas.
The final map, as displayed in the Adaptation Atlas was generated with the SimCLIM-for-ArcGIS/Marine toolbar (http://www.climsystems.com/simclimarcgis/marine/) in ArcGIS for Desktop 10.2.2 and exported as a tiled Map at 1 kilometer resolution.
The methodology described combines multiple techniques and datasets to get a best estimate of local vertical land movement around the globe. Its practical usage is limited to the coastlines, where most data is collected, making use in those regions more reliable. The methodology can be reapplied to update the resulting VLM-image when new information from SONEL and/or PSMSL becomes available.
Sea level rise is a very local affair. Multiple variables interact resulting in considerable variation around the world’s coastlines in how sea level rise is expressed locally. The sea level rise maps in the Adaptation Atlas represent values realized through the application of an ensemble of all the CMIP5 general circulation models that have sea level rise patterns. Across the 28 models there are some where the modeled sea level rise for a location will exceed the ensemble mean while others will be less. People who will use this type of sea level rise data need to be aware of this variability in model outcomes. Moreover, the models used in this App are only used to express potential change in sea level rise out to the year 2100 while sea level rise will continue for centuries given the lag in the global climate system set in train by current greenhouse gas concentrations. The Adaptation Atlas maps only express the potential sea level rise levels that could be realized given current trends in greenhouse gas emissions that are in line with the representative concentration pathway of 8.5 W/m2 in 2100. The future could be better or worse than this assessment. The information provided by the Adaptation Atlas must not be considered the only scenario for the future. It is one of many that could evolve and therefore this one should not be construed as a prediction of any future level for a given location.
For more information on sea level rise and computation of extreme sea level rise levels at a local scale for consideration in high resolution inundation mapping for disaster risk management and planning purposes contact the world leaders at CLIMsystems Ltd. We have the expertise, data and modeling capacity to provide the most reliable outputs in support of risk assessments.
Church, J. A., & White, N. J., 2011. Sea-level rise from the late 19th to the early 21st century Surveys in Geophysics, 32(4-5), 585–602. doi:10.1007/s10712-011-9119-1.Douglas, B. C. (1991) Global sea-level rise. Journal of Geophysical Research-Oceans, 96, 6981-6992.
Douglas, B. C. (1997) Global sea rise: A redetermination. Surveys in Geophysics, 18, 279-292.
Permanent Service for Mean Sea Level (PSMSL), 2014, "Tide Gauge Data", Retrieved 17 Mar 2014 from http://www.psmsl.org/data/obtaining/.
Simon J. Holgate, Andrew Matthews, Philip L. Woodworth, Lesley J. Rickards, Mark E. Tamisiea, Elizabeth Bradshaw, Peter R. Foden, Kathleen M. Gordon, Svetlana Jevrejeva, and Jeff Pugh (2013) New Data Systems and Products at the Permanent Service for Mean Sea Level. Journal of Coastal Research: Volume 29, Issue 3: pp. 493 – 504. doi:10.2112/JCOASTRES-D-12-00175.1.
Solomon, S. 2007. Climate change 2007: the physical science basis : contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge; New York: Cambridge University Press.
Zervas, C. E. 2001. Sea level variations of the United States, 1854-1999. Silver Spring, Md.: U.S. Dept. of Commerce, National Oceanic and Atmospheric Administration, National Ocean Service.