Estuarine Temperature Variability: Integrating Four Decades of Remote Sensing Observations and In-Situ Sea Surface Measurements

Abstract

Characterizing sea surface temperature (SST) variability is a critical aspect of studying long-term changes in estuarine environments. However, the scales of estuarine variability and change can be quite small (10 m-10 km). In this study, we present the first combined analysis of an estuary using the 39-year-long SST evolution from the multi-satellite Landsat data ($\sim18$ day average sampling), over a decade of in-situ buoy records (15 min. sampling), and tide gauges (60 min. sampling). We retrieved the seasonal-to-decadal sea surface and tidal temperature variabilities and trends over four decades in Narragansett Bay and its arm, Mt. Hope Bay. The seasonal solar heating, river run-off, and resulting salinity stratification, and bathymetry determine the dominant ($\sim80\%$) temperature variance in the bay. The warming trend of the annual mean SST is \sipm{0.057}{0.024}{\celsius\per\year} for Narragansett Bay and \sipm{0.015}{0.018}{\celsius\per\year} for Mt. Hope Bay. We classified each Landsat image by tidal phase using tide gauge measurements in order to produce composite SST anomaly maps corresponding to each tidal phase, but non-tidal noise made the signal trustworthy in only a few regions. High-frequency measurements reveal that tidal temperature changes are detectable and consistent at buoy sites but secondary to the temperature changes by season in the bay. The shallower, fresher upper bay shows greater SST variability than the lower bay, whose temperature approaches the more oceanic, less seasonal temperatures at the mouth. Importantly, our study represents the synergistic advantages of utilizing Landsat and in-situ buoy data to offer new and deeper insights into the changing conditions of global estuaries.

Keywords: Sea surface temperature, remote sensing, Narragansett Bay, Tide, Landsat, EOF