Cutting Edge Research

Coastal Impact Study:
Nation Under Siege

Appendix

METHODOLOGY USED TO DETERMINE COASTAL FLOODING
DUE TO SEA LEVEL RISE

S. Guerin, J. Thorp, N. Thompson
Redfish Group, 624 Agua Fria, Santa Fe, NM 87501
info@redfish.com

Summary
The goal of this project was to provide an accurate estimate of coastal flooding due to sea level rise. Elevation data from the USGS National Elevation Dataset were used and augmented with higher resolution LIDAR data where available. Baseline high tide levels were established by incorporating measurements from the closest NOAA tide gauges. A flood-fill algorithm was developed to determine the extent of coastal inundation. The resulting data were imported into Google Earth to provide detailed images of coastal flooding for selected U.S. cities and towns.

Methodology For Predicting Coastal Flooding
The technical methodology developed to predict coastal flooding is comprised of five steps. These steps are: 1) compile the best available elevation data, 2) augment flood data with additional Light Detection and Ranging (LIDAR) and ground based sources 3) establish mean sea level (MSL) with respect to available elevation data, 4) execute the flood-fill algorithm, and 5) integrate flood maps with Google Earth [1].

Compile Available Elevation Data
The National Elevation Dataset (NED) was obtained from the United States Geological Survey (USGS) [2] for selected areas of interest. The NED is a seamless raster dataset of U.S. elevations. Within the NED, the U.S. is divided into 10-meter-by-10-meter squares whose elevations correspond to the average elevation within a square. The NED is a compilation of elevation data from many sources, including LIDAR and USGS digital elevation models. The NED is maintained by the USGS to contain the most accurate, up-to-date elevation of U.S. landforms.

Augment National Datasets With Local LIDAR Studies
Any attempt to assess the potential damage to coastline interests from sea level rise will require knowledge of how high the coastline is in relation to the water. Despite enormous strides that have been made in technology of elevation determination in recent decades, the quality of data maintained in the national databases remains uncertain. In 1998, the National Geodetic Survey provided Congress with a report detailing the woeful lack of coordination between various sources of information concerning elevation [3]. The national datasets used in this study, therefore, were supplemented with local LIDAR data where available.

The NED provides estimates of land elevation, which in most cases are clearly validated by aerial photography of the marks left on coastal zones by the ebb and flow of tides. It also, however, provides estimates that are clearly wrong. An example of the uncertainty in the NED is illustrated using the city of Long Beach, NY, a barrier island off the coast of Long Island. The NED elevation of an intersection near the center of town differs by approximately 2.25 m from the elevation listed on the U.S. Army Corps of Engineers flood plain map [4]. Thus, a flood-fill map generated using the NED data would show the island routinely flooded at high tide, which does not occur.

The reason that this situation has not already produced chaos is that city, county and state engineering departments have their own databases for the purposes of designing roads, sewage lines, drainage systems and flood control structures and for granting building permits. Some of these data are publicly available [5], but much is available for a cost.

LIDAR data at approximately 3-m resolution typically improves the horizontal resolution 9-fold over 1/3 arcsecond (~10 m) data. Figure 1 illustrates differences in resolution.

Figure 1. Comparison of 30-meter USGS, 10-meter USGS, and 3-meter LIDAR data for Seattle, WA
(Puget Sound LIDAR Consortium).

To validate and augment elevation data for this study, an effort was made to acquire additional LIDAR datasets in areas where visual inspection led to lower confidence with USGS NED data. Sources of LIDAR data were obtained using phone interviews with local governments’ Geographic Information System (GIS) users and Internet surveys. A summary of the locations investigated, along with the data sources used for this study, is shown in Table 1.

In addition to supplementing with LIDAR data, a survey team was sent to take actual spot elevations in lowerconfidence areas. These measurements consisted of determining the elevations of sea walls, berms, roads and piers relative to actual or apparent high tide. The local surveying allowed us to conservatively validate the amount of sea level rise required to breach a given area and result in significant flooding. Spot elevation data was gathered for the following locations: San Francisco, Boston, East Boston, Point Shirley, New York City, Brigantine, Point Pleasant, Atlantic City, Miami Beach and Miami, shown in Fig. 2. Example maps of survey points are in Fig. 3.

Establish Local Mean Sea Level
The elevations in the NED and LIDAR datasets are based on the North American Vertical Datum, 1988 (NAVD88). NAVD88 uses one base monument located at Father Point, Quebec in Canada as its zero elevation level. Elevations within the NED are calculated relative to that zero point (adjusted for the curvature of the earth). For general purposes, the NAVD88 represents an approximate standard for establishing elevations above mean sea level. Along a coastline, however, the level of the sea does not everywhere correspond to zero on the NAVD88. This discrepancy can range from a few centimeters in Florida to a few meters in the northwest United States. Thus, local tidal conditions must be established using local tide gauges. United States National Oceanic and Atmospheric Administration (NOAA) [6] datum tables were obtained from tide gauges. These datum tables included zero-level NAVD88 marks and the measured MSL and Mean Higher High Water (MHHW) levels. The MSL and the MHHW on these datum tables are averaged over the 19-year period from 1983 to 2001 (National Tidal Data Epoch). In cases where multiple gauges were available within a specified radius, the adjustment was calculated based on a simple average. The following equation was used to establish MSL with respect to the NAVD88-based elevation models:

EMSL = MSLGauge - NAVD88Gauge   (1)

where EMSL is the local MSL, MSLGauge is the MSL from the datum table, and NAVD88Gauge is the height of the zero-level NAVD88 mark from the datum table. To assess the impact of high tide, the elevation of the local MHHW can be established with the following equation:

EMHHW = MHHWGauge - NAVD88Gauge   (2)

where EMHHW is the elevation of local MHHW and MHHWGauge is the MHHW from the datum table.

An example NOAA datum table is shown in Fig. 4. This table was obtained from the New York tide gauge station shown in Fig. 5. In this table, the zero-level NAVD88 benchmark is 1.849 m high on this gauge. The MSL and the MHHW are 1.785 and 2.543 m, respectively. Using Eq. 1, the local MSL is -0.064 m with respect to NAVD88. The local MHHW is 0.694 m with respect to NAVD88 (Eq. 2).

 

To return to the map interface click here
or select a location from the menu below:


Foreword
Introduction
Sea Level Rise
Visual Imaging
One Meter of Sea Level Rise... and Rising
A Lesson Learned?
Current Trends
Timeline
Fossil Fuels and Climate Change
The Power of Coal
Silver Bullet: Moratorium on Coal
Replacing Coal
The 2030 Challenge
Been There, Done That
Revisiting Katrina
Conclusions
Appendix

Download the study:



Table 1. Locations Investigated and Data Sources Used

Location Sea Level Rise Population Data Source
Point Pleasant, NJ 1 meter 19,306 USGS 10M NED
Hollywood, FL 1 meter 139,357 LIDAR and USGS 10M NED
New Orleans, LA 1 meter unknown USGS 10M NED
Miami Beach, FL 1 meter 87,933 LIDAR IHRCS+
Lavalette/Dover Beaches, NJ 1 meter 6,044 USGS 10M NED
Silverton Area, NJ 1 meter 9,175 USGS 10M NED
Hampton, VA 1 meter 146,137 USGS 10M NED and NOAA NGDC*
Point Shirley, Winthrop, MA 1 meter 6,373 MassGIS LIDAR
Miami, FL 1.25 meter 362,470 LIDAR IHRCS
Cape Coral, FL 1.25 meter 102,286 USGS 10M NED
Cypress Lake, FL 1.25 meter 12,072 USGS 10M NED
Fort Lauderdale, FL 1.25 and 2 meter 152,397 LIDAR and USGS 10M
Foster City, CA 1.25 meter 23,803 LIDAR 2M BCDC (USGS 10M verified)
Oakland Airport, CA 1.25 meter NALIDAR 2M BCDC (USGS 10M verified)
East Boston, MA 1.5 meter 38,413 MassGIS# LIDAR
Brigantine, NJ 1.5 meter 12,594 LIDAR and USGS 10M NED
Naples/East Naples, FL 1.5 meter 66,878 USGS 10M NED
Galveston, TX 1.5 meter 57,247 NOAA NGDC
Atlantic City, NJ 1.5 meter 40,517 LIDAR and USGS 10M NED
Newport Beach, CA 1.75 meter 70,032 USGS 10M NED
Honolulu, HI 1.75 meter 371,657 LIDAR and USGS 10M NED
Freeport, TX 1.75 meter 57,247 USGS 10M NED
Port Aransas, TX 1.75 meter 3,370 LIDAR
San Francisco, CA 2.25 meter 776,733 USGS 10M NED
Boston, MA 3-5 meter 589,141 LIDAR and USGS 10M NED
New York City, NY 3-5 meter 1,537,195 USGS 10M NED
Savannah, GA 3-5 meter 131,510 NOAA NGDC
Coronado, CA 3-5 meter 24,100 USGS 10M NED
San Diego, CA 3-5 meter 1,223,400 USGS 10M NED
Marina Del Rey/Santa Monica, CA 3-5 meter 84,084 USGS 10M NED
Seattle, WA 3-5 meter 563,374 USGS 10M NED (LIDAR verified)
+International Hurricane Research Center, Florida International University
*National Geophysical Data Center
#Massachusetts Geographic Information System



Figure 2. Map of Survey Locations

  
Figure 3. Example Maps of Survey Points (Left: Boston; Right: Manhattan)



Execute Flood-Fill Algorithm
Once local MHHW levels were established, the flood-fill algorithm was used. This algorithm determines contiguous inland access from the coastline for increased sea levels. The algorithm was executed for the following values of sea level rise: 0, 0.25, 0.5, 0.75, 1.0, 1.25, 1.5, 1.75, 2.0, 3.0, 4.0, 5.0 and 6.0 m. (A sea level rise of 0 m was evaluated as an algorithm check.) For each area studied, the land-water edge, based on local MHHW, is determined. The algorithm uses this edge as the starting point of the flood-fill and moves inland. From each flooded point, the algorithm selects neighboring pixels that are at, or below, the local MHHW plus the amount of simulated sea level rise. The algorithm continues recursively from these neighboring points until no new points are selected.

Integrate Flood Maps with Google Earth
The flood maps generated by the flood-fill algorithm were saved as portable network graphics (PNG) images. The PNG is a bitmap image format that uses lossless data compression. From these PNG images, KML files were generated for direct import into Google Earth. The KML format is used to display geographic data in Google Earth and is based on the XML (extensible markup language) standard file format. At its best, Google Earth imagery is sufficiently detailed to show a bird's eye view of buildings, streets, parks, waterways, and even individual automobiles. The resulting, superimposed flood maps show in detail how localities will be flooded on a calm, rain-free day. An example of a flood map for Miami, FL is shown in Fig. 6 for a 1.25-m sea level rise above MHHW.


  
Figure 4. Example NOAA tide gauge station datum table.


Figure 5. Tide gauge station in New York Harbor
("The Battery", ID-8518750).

Figure 6. Miami flood map for a 1.25-meter sea level rise above MHHW.


REFERENCES

1. Google, Inc., Mountain View, CA, (http://earth.google.com).
2. National Elevation Dataset available from the U.S. Geological Survey, EROS Data Center, Sioux Falls, SD.
3. National Digital Elevation Project, 2004. Guidelines for Digital Elevation Data, V1.0. Published May 10, 2004, and available at http://www.ndep.gov/NDEP_Elevation_Guidelines_Ver1_10May2004.pdf.
4. The intersection is Magnolia and West Park. See US Army Corps of Engineers, National Flood Insurance Rate Map for the City of Long Beach, Long Island, New York. See http://msc.fema.gov/ to generate a FEMA flood map for this area.
5. GIS Technical Committee (Paul McCombs, Chair). Production Operations and Maintenance Plan for 2007, King County, Washington State.
6. National Oceanic and Atmospheric Association’s Center for Operational Oceanographic Products and Services, Silver Spring, MD.