Coastal Flooding

Identification

1.	Indicator Description

This indicator describes how the number of coastal floods exceeding local flood threshold levels has
changed over time. Sea level rise related to climate change is a key driver of the increasing frequency of
coastal flooding. The 2017 Climate Science Special Report determined with "very high confidence" that
the number of tidal floods each year that cause minor impacts (also called "nuisance floods") has
increased in several U.S. coastal cities in response to rising sea levels (USGCRP, 2017).

Components of this indicator include:

•	A map that shows the change in flood days per year along U.S. coasts, comparing the first
decade of record with the most recent decade (Figure 1).

•	A more detailed graph that shows average flood days per year along U.S. coasts from 1950
to 2020 (Figure 2).

2.	Revision History

August 2016:	Indicator published.

April 2021:	Updated indicator with data through 2020 and a new set of flood thresholds.

Data Sources

3.	Data Sources

Coastal flooding trends are based on measurements from permanent tide gauge stations. The original
tide gauge data come from the National Water Level Observation Network (NWLON), operated by the
Center for Operational Oceanographic Products and Services (CO-OPS) within the National Oceanic and
Atmospheric Administration's (NOAA's) National Ocean Service (NOS). Daily maximum water levels are
derived from the hourly data set maintained by CO-OPS. Mike Kolian of EPA developed this indicator in
collaboration with William Sweet of NOAA. The analysis is adapted from NOAA (2018), which is an
update to an analysis published by Sweet and Marra (2015), NOAA (2014), and Sweet and Park (2014).

4.	Data Availability

Individual tide gauge measurements can be accessed through NOAA's "Tides and Currents" website at:
https://tidesandcurrents.noaa.gov/stations.html?type=Water+Levels. This website also presents an
interactive map that illustrates sea level trends over different timeframes.

EPA obtained the 1950-2020 analysis from NOAA's derived product data interface at:
https://api.tidesandcurrents.noaa.gov/dpapi/prod. Station-specific flood thresholds for locations
presented in this indicator, as well as other locations, are listed in Appendix 1 of NOAA (2018). Updates

Technical Documentation: Coastal Flooding

1


-------
to this indicator are likely to coincide with NOAA's annual report, The State of High Tide Flooding and
Annual Outlook, which is released each spring at:

https://tidesandcurrents.noaa.gov/HighTideFlooding AnnualOutlook.html.

Methodology

5. Data Collection

This indicator presents the change in flood days, which are defined as days in which a tide gauge records
water levels that exceed a threshold that NOAA has derived from the gauge data at each location. This
indicator captures all flooding events that exceed each location's threshold, which means it captures
floods that may range from small "nuisance" floods to much larger but less frequent major floods.

Coastal water levels have traditionally been measured using tide gauges, which are mechanical
measuring devices placed along the shore. These devices measure the change in sea level relative to the
land surface, which means the resulting long-term analysis reflect both changes in flood frequency
occurring from changing absolute sea surface height and local land levels.

Tide gauge data for this indicator come from NWLON, which is composed of 210 long-term, continuously
operating tide gauge stations along the U.S. coast, including the Great Lakes and islands in the Atlantic
and Pacific Oceans. This indicator shows trends for a subset of stations along the ocean coasts that met
the following criteria for sufficient data:

•	A total of at least 10 "flood days" during the entire period of analysis (1950-2020). This criterion
eliminates sites that have too few observed floods to support an analysis of change over time.

•	At least 60 years of data during the period of analysis. This criterion eliminates sites that are
missing too many years of data to allow for a reasonable analysis of change over time.

•	At least six years of data within each decade (1950-1959, 1960-1969, etc.). This criterion
eliminates sites with large gaps that might bias the results, including sites that might have come
online in the late 1950s but technically met the criterion for 60 years of data. The criterion is
prorated for the most recent decade if it is a partial decade.

Although many stations collected data before 1950, NOAA and EPA selected 1950 as a reasonable
starting point for this indicator to ensure adequate data for analysis. Choosing a much earlier starting
point would have greatly reduced the number of stations with sufficient data for a scientifically
defensible analysis, which in turn would have led to an indicator with less complete geographic
coverage.

NOAA (2014) describes tide gauge data and how they were collected. Data collection methods are
documented in a series of manuals and standards that can be accessed at:

https://tidesandcurrents.noaa.gov/pub.html. This indicator uses hourly averages based on each tide
gauge's continuous measurements.

Technical Documentation: Coastal Flooding

2


-------
6. Indicator Derivation

This indicator was derived by calculating each day's maximum water level based on hourly water level
data, then comparing these daily maxima with threshold levels for flooding at each tide gauge.

NOAA derived a consistent set of location-specific flood thresholds using a statistical regression model.
This model sets each location's flooding threshold at a specific height above the average local tide range.
NOAA determined the regression coefficients by analyzing flood impact levels (minor, moderate, major)
that have been established for selected coastal cities by local National Weather Service (NWS) weather
forecasting offices, based on many years of impact monitoring. NOAA's derived thresholds for this
indicator are based on NWS minor flood impact levels and the observed relationship between these
NWS thresholds (where they exist) and the corresponding long-term local tide range. NOAA (2018)
describes the derivation of thresholds for this indicator in detail.

Of the 210 NWLON tide gauges, NOAA determined that 98 had adequate data for an initial analysis.
Accordingly, NOAA derived thresholds for these 98 locations. Applying the more restrictive criteria
described in Section 5 above led to the selection of 33 sites with adequate 1950-2020 data for inclusion
in this indicator.

The total number of days exceeding the derived flooding threshold was calculated for each tide gauge
and for every calendar year. Annual totals were averaged together over multi-year periods for Figures 1
and 2. Figure 1 provides a simple comparison between the first and last decades of record: the 1950s
(1950-1959) and the most recent decade (2011-2020). Figure 2 covers the entire period of record by
sorting the data into bins, most of which are 20 years in length.

Indicator Development

A previous version of this indicator used a different set of flood thresholds: location-specific "minor
flooding" thresholds established by local NWS weather forecasting offices. Direct use of NWS thresholds
offered the advantage of leveraging local knowledge tied to observed impacts, but this approach also
had disadvantages:

•	Because thresholds were derived differently for each location, results did not lend themselves
to comparison across multiple locations.

•	Only 75 of the 210 NWLON tide gauges have corresponding NWS thresholds, which meant that
trends in coastal flooding could not be analyzed for other locations that otherwise had ample
data.

Source publications before 2018 all describe the original, NWS-threshold-based method. In 2018, NOAA
published a new methodological approach that used the derived thresholds described at the beginning
of this section (NOAA, 2018). This approach results in broader applicability (more sites) and more
comparability between sites. EPA adopted this method for the 2021 indicator update to be consistent
with NOAA's new recommended approach. This change expanded the indicator to cover 33 sites,
compared to 27 using the previous method.

Technical Documentation: Coastal Flooding

3


-------
7. Quality Assurance and Quality Control

Quality assurance and quality control procedures for U.S. tide gauge data are described in various
publications available at: https://tidesandcurrents.noaa.gov/pub.html.

Analysis	

8.	Comparability Over Time and Space

All of the tide gauges included in this indicator have used the same methods for determining hourly
water levels. These methods remained constant over time and across gauges, except as documented in
NOAA (2014). Tide gauge measurements at specific locations are not indicative of broader changes over
space, however, and the network is not designed to achieve uniform spatial coverage. Rather, the
gauges tend to be located at major port areas along the coast, and measurements tend to be more
clustered in heavily populated areas like the Mid-Atlantic coast. Nevertheless, in many areas it is
possible to see consistent patterns across numerous gauging locations—for example, increases in the
frequency of flooding along the Atlantic and Gulf Coasts.

Flooding thresholds have been established using a statistical model that applies the same formula across
all locations. Each location's flood threshold has been applied consistently throughout the period of
record, which supports this analysis of trends over time.

9.	Data Limitations

Factors that may impact the confidence, application, or conclusions drawn from this indicator are as
follows:

1.	Coastal flooding relates to relative sea level change, which is the height of the sea relative to the
height of the land. Changes in coastal flooding frequency cannot be solely attributed to absolute
sea level change, but instead may reflect some degree of local changes in land elevation (e.g.,
subsidence). Tide gauge measurements generally cannot distinguish between these two
influences without an accurate measurement of vertical land motion nearby.

2.	Some changes in coastal flooding may be due to multiyear cycles such as El Nino/La Nina and
the Pacific Decadal Oscillation, which affect coastal ocean temperatures, water density (due to
salt content), winds, atmospheric pressure, and currents.

3.	The flood thresholds used for this indicator are derived from a statistical model, not directly
based on local conditions, so they do not necessarily correspond to a consistent level of
observed disruption or damage across different locations. Every location has its own unique
characteristics in terms of land cover, topography, elevation of critical infrastructure, and the
presence or absence of flood defenses such as seawalls. Thus, flooding that reaches the derived
threshold in one city might correspond to much more damage and disruption than a flood that
reaches the threshold in a different location. For this reason, when considering the impacts of
flooding, it is more useful to compare change over time at a single location than it is to compare
patterns across different locations.

Technical Documentation: Coastal Flooding

4


-------
4.	This indicator only includes tide gauges with sufficient hourly data since 1950, which results in
sparse coverage on the Pacific and Gulf Coasts and no coverage of Hawaii.

5.	The statistical model used to derive flood thresholds for the contiguous 48 states (NOAA, 2018)
is not valid for Alaska. Therefore, NOAA has not derived flood thresholds for Alaska and this
indicator does not include Alaska.

6.	Impacts are localized and not necessarily readily observable. When water levels are expected to
exceed an official NWS flooding threshold, coastal flood advisories are typically issued. Minor
impacts typically, but not always, manifest.

7.	Local topography may affect the relative influences of various environmental processes on a
specific site's flooding. For example, offshore barriers such as coral reefs or barrier islands may
help to buffer certain contributing effects, such as wind. By contrast, other areas may have
topographical features that amplify the flooding caused by slight changes in the environment.
Although these differences do not negate the site-specific trends observed, they do contribute
to differences between stations.

10.	Sources of Uncertainty

Error measurements for each tide gauge station are described in NOAA (2009), but many of the
estimates in that publication pertain to longer-term time series (i.e., the entire period of record at each
station, not the exact period covered by this indicator). Uncertainties in the data do not impact the
overall conclusions. Tide gauges provide precise, reliable water level data for the locations where they
are installed.

11.	Sources of Variability

Changes in sea level and corresponding changes in coastal flooding can be influenced by multi-year
cycles such as El Nino/La Nina and the Pacific Decadal Oscillation, which affect coastal ocean
temperatures, salt content, winds, atmospheric pressure, and currents.

12.	Statistical/Trend Analysis

This indicator does not report on the slope of the apparent trends in flood frequency, nor does it
calculate the statistical significance of these trends. Separately, NOAA (2018) presented statistical trend
analyses for four example locations using quadratic and linear fits for data from 1950 to 2016. NOAA
reported that two East Coast locations (Atlantic City, New Jersey, and Norfolk, Virginia) fit a quadratic
regression, representing acceleration. San Diego, California, and Seattle, Washington, each fit a linear
increasing trend. All four regressions were significant to at least a 90 percent level.

In a previous analysis using NWS minor flood thresholds, NOAA (Sweet and Park, 2014) analyzed trends
in the annual number of flood days from 1950 to 2013 at 27 of the locations included in the current
indicator. Of the 27 stations, 19 had significant quadratic fits and four had significant linear fits, where
significance was defined at the 90 percent level.

Technical Documentation: Coastal Flooding

5


-------
References

NOAA (National Oceanic and Atmospheric Administration). 2009. Sea level variations of the United
States 1854-2006. NOAA Technical Report NOS CO-OPS 053.
www.tidesandcurrents.noaa.gov/publications/Tech rpt 53.pdf.

NOAA (National Oceanic and Atmospheric Administration). 2014. Sea level rise and nuisance flood
frequency changes around the United States. NOAA Technical Report NOS CO-OPS 073.
https://tidesandcurrents.noaa.gov/publications/NOAA Technical Report NOS COOPS 073.pdf.

NOAA (National Oceanic and Atmospheric Administration). 2018. Patterns and projections of high tide
flooding along the U.S. coastline using a common impact threshold. NOAA Technical Report NOS CO-OPS
086. https://tidesandcurrents.noaa.gov/publications/techrpt86 PaP of HTFIooding.pdf.

Sweet, W.V., and J.J. Marra. 2015. 2014 state of nuisance tidal flooding.

www.ncdc.noaa.gov/monitoring-content/sotc/national/2015/aug/sweet-marra-nuisance-flooding-
2015.pdf.

Sweet, W.V., and J. Park. 2014. From the extreme to the mean: Acceleration and tipping points of
coastal inundation from sea level rise. Earth's Future 2(12):579-600.
https://agupubs.onlinelibrarv.wilev.com/doi/full/10.1002/2014EF00Q272.

USGCRP (U.S. Global Change Research Program). 2017. Climate science special report: Fourth National
Climate Assessment, volume I. Wuebbles, D.J., D.W. Fahey, K.A. Hibbard, D.J. Dokken, B.C. Stewart, and
T.K. Maycock (eds). https://science2017.globalchange.gov. doi:10.7930/J0J964J6

Technical Documentation: Coastal Flooding

6


-------