Abstract |
The oceans are a diverse soup of organic life and are major regulators of the earth's many systems. Tracking ocean systems is necessary for the regulation of healthy habitats, maintaining clean recreational environments, and monitoring pollutants. Using satellite sensors, we have access to tons of real-time public data of the world's surface, which can be used to do all these things. In this project, a statistical approach to remote sensing called varimax--rotated, principal component analysis (VPCA) was utilized to identify the suspended matter in the water. This approach takes the derivative of the reflectance spectra and unmixes it to give us a more accurate reading of materials that influence the reflected light. Mainly we are looking for any color-producing agents (CPAs) suspended in the water i.e., phytoplankton, detritus, and dinoflagellates or sediment or sediment minerals. By comparing the values to an existing spectral library, we can identify the components. In 2017, Hurricane Irma struck the US Virgin Islands leaving behind a wake of destruction. By comparing images before and after the hurricane, we can track how pigment distribution changed after the event. We observe that all the same components were identified between both dates, but that their distributions vary. Possible further applications to this project include creating seasonal time series to understand distributions year-round and validating our data with samples collected in the USVI from around these dates.
|