Utilizing Empirical Eigenfunctions and Neural Network to Describe and Model RI Coastal Morphology
Southern Rhode Island’s microtidal, sandy beaches have been monitored using stadia-style profiling techniques in bi-weekly time intervals during the spring, fall, and winter, and monthly during the summer since the early 1960s. This dataset provides a time-series of cross-sections based on which vol...
| Main Author: | |
|---|---|
| Format: | Text |
| Language: | unknown |
| Published: |
DigitalCommons@URI
2017
|
| Subjects: | |
| Online Access: | https://digitalcommons.uri.edu/theses/1125 https://doi.org/10.23860/thesis-davis-sierra-2017 https://digitalcommons.uri.edu/context/theses/article/2134/viewcontent/Davis_uri_0186M_11811.pdf |
| Summary: | Southern Rhode Island’s microtidal, sandy beaches have been monitored using stadia-style profiling techniques in bi-weekly time intervals during the spring, fall, and winter, and monthly during the summer since the early 1960s. This dataset provides a time-series of cross-sections based on which volumetric changes can be inferred. Early studies utilized these profile volume calculations for spectral analyses, which revealed high-frequency cycles of 1 year and 1.5-5 years attributed to seasonal trends and longshore sediment transport, respectively. Additionally, varved sedimentary records in southern Rhode Island provide locally-derived proxies that indicate North Atlantic climatic drivers such as North Atlantic Oscillation (NAO) influence local weather patterns. Currently, with nearly fifty-five consecutive years of surveying, these lower frequency climatic cycles (5-15 years) can be resolved. This work presents statistical analyses using empirical orthogonal eigenfunctions to describe variations in profile shape as well as spatial and temporal patterns within the time-series dataset. Dominant cycles within the beach volume time-series are identified through spectral analysis techniques. With these methods, links between those aforementioned Northern Hemisphere climatic cycles and their impact on coastal geomorphology are investigated. Additionally, using nearshore wave climate data derived from a 35-year long dataset (1980-2014) from the nearest United States Army Corps of Engineers’ Wave Information Study (WIS) buoy, we attempt to explain the higher-frequency cycles in beach volume change through a correlation analysis for this period. In an effort to model and predict beach volume, methods of Neural Networking, a form of Artificial Intelligence, are applied using wave climate data, mean sea level, and the NAO index as input parameters. |
|---|