Date of Award

Spring 2015

Project Type


Program or Major


Degree Name

Master of Science

First Advisor

Thomas C Lippmann

Second Advisor

Larry Ward

Third Advisor

Semme Dijkstra


Seafloor classification and environmental assessment in shallow marine waters are crucial to habitat mapping, coastal management policies and maintaining navigational waterways. There are existing methods for remotely estimating some bottom properties, but the large variety of desired measured sediment properties frequently leads to insufficient quantifiable data to support marine policy decisions. This problem is exacerbated by the highly variable bottom composition of typical coastal and estuarine environments. In this work, field observations from an Odom Echotrac vertical-incidence echosounder with a 200 khz transducer were used to estimate seafloor sediment characteristics in regions with variable bottom types. Observations were obtained in water depths ranging 0.5-24 m of the Little Bay, New Hampshire, during February and March, 2013. Backscatter waveforms (the acoustic return representing the first interaction with the bottom) were analyzed and their properties compared to sediment grain size distributions. These comparisons showed varied degrees of predictive capability and require subjective a priori selection. In an effort to better capture the collective effects of seafloor sediment's composition on acoustic returns, empirical orthogonal functions (EOF's) were computed from an ensemble of seven waveform properties and compared with observed surficial sediment size fractions, bulk density, and porosity. A simple logarithmic model relating first mode EOF spatial variability to observed mud fractions explained 43% of the variability and well estimated the spatial pattern of mud across the bay (RMS errors in mud fraction of 10-15%) from deep channels (with no mud) to high concentrations of mud on the shallower flats near the sides of the estuary. This method produced greater coverage and higher resolution predictions of mud fraction than could be obtained using traditional sediment measuring techniques. Deviations from the model are shown to be correlated with lower sediment porosity most likely due to river inflow from the Bellamy River draining into the Bay. Application of the model coefficients to new data obtained in the Great Bay in 2014 with the same sonar and acoustic settings, showed similar predicted mud fractions with RMS errors of 11.9 and 13.2% along two surveyed lines. This empirical analysis provides a first order objective means to interpret acoustic backscatter, an important step towards a widespread quantitative assessment of shallow water seafloor sediments.