Additional Calendars
Calendar Views
Conferences and Meetings
Law School
Special Events

Department of Estuarine and Ocean Sciences MS Thesis Defense by Ronni Mak

When: Friday, February 23, 2024
11:00 AM - 12:00 PM
Where: > See description for location
Description: SMAST West 204 and via Zoom

Depth Related Ecoregion Thresholds and Satellite Remote Sensing for Ponds and Lakes in southeastern Massachusetts

by Ronni Mak

Eutrophication is an ongoing issue around the world, particularly in Cape Cod and Plymouth, MA where the economy relies on tourism and an ethos of pristine waters. Approaches to support better pond and lake management, through establishing more realistic prospective standards and applications of satellite remote sensing techniques, were studied using Pond and Lake Stewards (PALS) data from 2001 to 2022 (130-180 ponds sampled per year). Cape Cod and Plymouth are in the Atlantic Coastal Pine Barrens ecoregion, which is defined by sandy soils and near surface groundwater systems. Ecoregion thresholds, or reference conditions, were found to be related to pond depth for all water quality assays considered which included pH, Secchi depth (SD), chlorophyll a (CHL), total nitrogen (TN), and total phosphorus (TP). To better represent pond conditions and improve accuracy of target prospective standards, ponds were separated into three depth categories, ultrashallow (<1.5m), shallow (1.5-9m), and deep (<9m). Linear models of yearly ecoregion thresholds identified significant (p<0.05) relationships with SD, pH, and TP for certain pond depth categories (e.g., shallow vs. deep), consistent with expected enhanced productivity and associated eutrophication in surface waters. A lag time of P transport from the groundwater following the 1980-1990's increases in population in Barnstable County for shallow and deep depth classes was found, but this did not coincide with CHL increases within the timeframe studied. TN did not show significant changes of eutrophication from 2001 to 2022, likely due to the differing modes of transport compared to TP. The application of Landsat-8 and 9 satellite observations for remote sensing of SD and CHL in the study area was found to be feasible, but data-limited. Both the C2X and C2RCC neural network algorithms were evaluated. The strongest correlations between in-situ measured and satellite-derived estimates were found for the C2X algorithm for images taken within 24 hours of in-situ sampling. Measured and C2X-derived CHL results were encouraging when the times between satellite overpass and in-situ sampling were within 24 hours (R2=0.712, RMSE=8.83, p=0.001). Significant relationships between in-situ and satellite-derived values of SD were seen for in-situ samples acquired within 24 and 48 hours of the satellite overpass. However, predictive skill was limited. In general, the C2X algorithm performed better for both shallow and deep ponds underscoring the importance of algorithm selection for satellite image processing.
Contact: > See Description for contact information
Topical Areas: SMAST, Students, Graduate