Cumulative Wave Impact Height (CWIH)
The two plots below show CWIH versus recession rates for 1966-1999. The first plot is for CWIH hindcast from two percent runup, and the second plot is for CWIH hindcast from mean runup.
Cumulative Wave Impact Height (CWIH) versus recession rates for 1966-1999. CWIH from Nielsen and Hanslow 2% runup equations.
Cumulative Wave Impact Height (CWIH) versus recession rates for 1966-1999. CWIH from Nielsen and Hanslow mean runup equations.
The mean runup has a lower R2 value for the CWIH versus recession rate correlation than the two percent runup. The reason why the mean is lower is because the two percent wave runup accounts for many storms that impact the bluff relatively few times a storm. The mean runup correlation give greater weight to storms that potentially do more damage, but occur less often. For mean runup to cause positive WIH the average wave hits the bluff toe or above and, for example, the two percent waves are impacting higher up and with greater impact. Though there is much scatter in the data, the mean runup CWIH correlation provides a stronger relationship with recession rates (i.e. steeper slope of trend line). Also, there are some trends that are different for the mean CWIH and the two percent CWIH. For example, look at the five points to the lower right of the data. These points yield a positive trend for the mean CWIH correlation but only a scattering of points for the two percent wave runup CWIH correlation. The mean runup CWIH also performed the best for the adjustments discuss below and will be the only correlation discussed from this point on.
As seen in the above plots there is a considerable amount of data with low CWIH and variable recession rates (i.e. many points along the y-axis). Mentioned previously was the fact that WIH does not account for all of the variables contributing to erosion. Thus, some bluffs may be receding due to subaerial processes, and are less dependent on wave erosion. While wave energy is still required to carry failed material away, it is not the driving erosive force. One way of dealing with these sites is to account for beach width. Minimum beach widths were calculated based on highest water levels from 1966-1999. If the minimum beach width was larger than 9 meters these sites were considered to have wide enough beaches to prevent wave erosion from being a dominant force throughout the entire hindcast record. The plot below shows the mean runup CWIH versus recession rates correlation with these wide beach omitted.
Cumulative Wave Impact Height (CWIH) versus recession rates with wide beach data omitted.
Omitting the beaches that are wide (>9 m) under the highest water levels improved the correlation, though considerable scatter still exists. Another variable to consider is bluff height. As seen from the Site Characteristics, there is a range of bluff heights. Bluffs of different heights have different mechanisms and cycles. The next adjustment removes the data from the very low bluffs (< 2 m) and the high bluffs (> 17m). The CWIH versus recession rate correlation for the intermediate height bluffs is shown below.
Cumulative Wave Impact Height (CWIH) versus recession rates for intermediate bluff heights with wide beach data omitted.
Again the CWIH versus recession rate correlation has improved by considering bluffs with similar heights. The final adjustment to be made is to consider bluff material. Sand bluffs are expected to erode differently than cohesive (clay) bluffs because their ability to resist erosion and their failure mechanisms are different. Thus, these two bluff material types are separated out to reveal the CWIH versus recession rate correlation for each material type.
Cumulative Wave Impact Height (CWIH) versus recession rates for cohesive bluffs with intermediate bluff heights and wide beach data omitted.
Cumulative Wave Impact Height (CWIH) versus recession rates for sand bluffs with intermediate bluff heights and wide beach data omitted.
The CWIH versus recession rate correlation was relatively affected by analyzing the cohesive bluffs separately. The correlation greatly improved for the sand bluff analyzed separately, but there were only three points left. However, an R2 of 1 is very good as long as more than two points are available. There is considerable scatter in the cohesive bluff correlation and further investigation into the reliability of the recession data and hindcast CWIH data may remove some of the outliers from the correlation.
Based on the trends obtained, the sand bluffs appear to be more dependent on wave erosion than the cohesive bluffs. This makes sense considering the sand bluffs have no cohesion and relatively little resistance (just gravity) to wave erosion.