This section describes the methods used to compile shoreline data and calculate rates of change for the Kauai, Oahu, and Maui coasts. 

Shoreline Data

Historical shoreline data were acquired from multiple sources and are documented in the shoreline metadata files available in this report.  Shoreline data were digitized from National Oceanographic and Atmospheric Administration (NOAA) topographic sheets (T-sheets) and aerial photographs. The dates of historical shorelines range from 1899 to the 2000s. 

The historical shorelines are based on an interpreted low-water mark (LWM) that was digitized from topographic maps and aerial photographs following the methods of Fletcher and others (2003) and Romine and others (2009).  Previously published U.S. Geological Survey (USGS) reports in the series used the high water line (HWL).  In Hawaii, the high reflectivity of the white carbonate beaches reduces the visibility of the HWL on historical aerial photographs.  Excellent water clarity relative to most continental beaches allows the delineation of the LWM on historical aerial photomosaics, which is distinguished by a black and white or color tonal change at the base of the foreshore.  The shorelines extracted from T-sheets are referenced to a HWL.  To include T-sheet shorelines in the time series of historical shorelines, the HWL is migrated to a LWM in our study using an offset calculated from measurements in beach profile surveys at the study beach or a similar nearby location. Uncertainty resulting from positional and mapping errors is calculated for each shoreline segment and is included in an "Accuracy" attribute field in the shoreline data. 

Calculation of Shoreline-Change Rates

Rates of shoreline change were generated within ArcMap version 9.3 using the Digital Shoreline Analysis System (DSAS) version 4.2, an ArcGIS tool developed by the USGS (Thieler and others, 2009). The tool is a freely available application designed to work within the ESRI ArcGIS software.  DSAS is used to generate orthogonal transects starting from a reference baseline and intersecting the shoreline positions at 20-meter intervals.  The distance measurements between the transect/shoreline intersections and the baseline are then used to calculate the rate-of-change statistics.  

Shoreline-change rates were calculated using weighted least squares regression (WLS), which accounts for uncertainty in each shoreline position when calculating a trend line. The weight for each shoreline position is the inverse of the positional uncertainty squared, so that shorelines with higher uncertainty have less influence on the trend line than shorelines with smaller uncertainty. Rates of change were calculated in units of meters per year along with the uncertainty value (at the 95 percent confidence interval) over the long-term (all available years) and short-term (post-World War II) to capture potential changes in trends or rates. The linear regression method of determining shoreline-change rates was based on an assumed linear trend of change between the earliest and latest shoreline dates.  Data for shoreline areas where such a linear trend did not exist and shoreline-change rates have not remained constant through time would produce a poor linear fit to the data with a higher reported uncertainty.  The long- and short-term transect metadata files in this report provide descriptions of the two fields associated with the linear regression rate calculation, and additional information can be found in Section 7 of the DSAS user guide.