When analyzing the shoreline data we focused on two key questions:
“How much has the shoreline changed?” (How far has it moved?) and
“How fast has the shoreline changed?” (At what rate is it moving?)
For each question we also considered:
the “long” term (i.e., using shoreline data from 1880s — 2006)
the “short” term change (i.e., using just shoreline data from 1983 — 2006)
To answer these, shoreline data was processed using freely available USGS software (Digital Shoreline Analysis System (DSAS) using the following metrics:
How much has the shoreline changed?
Net Shoreline Movement (NSM): When looking at all the shorelines at a location, it is the distance between oldest & newest shorelines.
How fast has the shoreline changed?
End Point Rate (EPR): The End Point Rate is simply the Net Shoreline Movement divided by the time interval.
Linear Regression Rate (LPR): uses shoreline locations to “fit” a line that approximates the trend of the data. The trend line is computed so that the offset between the line and the data points is minimized.
The slope of the dashed line is the linear regression rate of change, and the offsets between the line and the data points are used to define a confidence value for the rate at a given location. Note that ALL shorelines are used, not simply the oldest and newest. With respect to these calculations, some pros and cons are noted below:
End Point Rate Pros:
- A simple calculation that’s easily understandable
- Can be used essentially anywhere there are data (only need 2 shorelines.)
- Easily applied to both Long Term and Short Term analyses
End Point Rate Cons:
- Ignores other shorelines so the rate can be idealized
- Assumes a linear fit; not always the case
- Can be highly influenced by the quality of either (or both) of the shorelines
- Provides no measure of confidence in the rate
Linear Regression Rate Pros:
Relatively easy to implement
Uses all shoreline data
Provides a rate and an estimate of confidence in it
Allows user to specify level of confidence (in this case, 86.5% or 1.5 Standard Deviations)
Linear Regression Rate Cons:
Assumes a linear fit; not always the case
Requires at least 3 data points (ideally more)
Can return “inconclusive” results (e.g., where the measure of uncertainty is greater than the rate) — requires user to interpret results
There may be areas where no output can be used.
Across all the metrics, negative values characterize loss of land and positive values characterize land gain. These were derived using a common baseline created to standardize results and provide a consistent framework for the analysis.
Continue to Shoreline Sections