Date of Degree

2-2021

Document Type

Dissertation

Degree Name

Ph.D.

Program

Earth & Environmental Sciences

Advisor

Rebecca Boger

Committee Members

Brett Branco

Chester Zarnoch

Jamie Vaudrey

Subject Categories

Environmental Monitoring | Environmental Sciences

Keywords

Jamaica Bay, eutrophication, macroalgal blooms

Abstract

The main objective of this study involves the development of a habitat suitability model for the Ulva genus in Jamaica Bay, New York. This incorporates several steps that were initiated by the selection of the most suitable water quality parameters that facilitate the successful growth of Ulva. These water quality parameters include dissolved oxygen, pH, temperature, salinity, nitrate + nitrite, ammonium, phosphates, dissolved organic nitrogen, dissolved organic carbon, depth and Secchi depth. This water quality data was generated by the Department of Environmental Protection. The Secchi depth and Jamaica Bay bathymetry data were necessary for the calculation of the % light to bottom that has been vital to the development of this model. For model development, inverse distance weighted interpolation was used to generate water quality surfaces. Because Jamaica Bay possesses islands, a modelling challenge is presented. In order to take into account the presence of these islands, polyline data was included in the creation of the IDW surfaces so that hard lines can delineate the water column from the islands. This allowed better water quality analyses to be carried out. After the development of the IDW surfaces, scored ranges and weights were applied so that the more influential and important parameters for Ulva growth such as light, temperature and nutrients were highlighted and given higher weights than the other parameters. After the assignment of the scored ranges and weights using the reclassify and weighted sum tool in ArcGIS, these surfaces were summed to create habitat suitability models. These models were then validated using Ulva biomass data and subsequently, composite bands and iso cluster analysis using ArcGIS Pro. Ulva biomass data were collected in 2012, 2015 and 2017. The 2017 sampling sites that were used in both biomass and satellite imagery analyses were Marine Park, Plumb beach, Big Egg, Cross bay bridge and Norton basin. In the iso cluster and composite band analyses, several band combinations were applied to visualize the algal/phytoplankton content of the bay. The most effective visualizations were obtained from 12-8-3, 12-11-4 and 4-8-11 based on the combined comparisons for both random and non-random analyses for biomass, composite bands and iso clusters. Additionally, for the biomass-model prediction comparisons, there was a 40.6% match rate. However, when biomass data comparisons were combined with that of the iso clusters and the composite bands, the model assessment was increased to 73.4% for 12-8-3 and 70.3% for several other combinations that includes 11-8-2, 12-8-4, 8-3-2, 12-11-4 and 4-8-11. However, for the random point model assessment, there was a 62.4% overall model accuracy for band combination 12-11-4. Overall, the model assessment has shown acceptability based on Holmes et al. (2008): 67-84%, Renken and Mumby (2009): 55-100%, and Zavalas et al. (2014): >70% acceptability scales.

Share

COinS