observations. High resolution seismic reflection profiling, 

 discussed in detail by Belknap et al. (1986, 1987a), involved 3.5 

 kHz Raytheon tuned transducer and ORE-Geopulse boomer systems. 

 The data has been interpreted in terms of intensity and geometry 

 of reflections to form a consistent local stratigraphic model 

 (Belknap et al., in press). Sediment process indicators, 

 interpretable from the seismic data, include slumping, bedforms, 

 topset/foreset contacts, natural gas deposits, and rhythmic 

 bedding. The nature of the sediments could be estimated based on 

 correlation with cores, grab samples, submersible observations 

 and outcrops. The seismic reflection data were used to produce a 

 seismic facies map (Figure 4) , as well as interpreted cross 

 sections. 



A second major remote sensing tool is the side-scan sonar, 

 an EG&G SMS 260 Sea Floor Mapping System. Backscattered sonar 

 signals are digitally rectified into a true-ratio scale that is 

 compensated for slant range and ship speed. The pixels are 

 printed in a broad-band gray scale, producing an image analogous 

 to an aerial photograph. These images are relatively easy to 

 interpret and to form into mosaic maps. The side-scan images are 

 used to identify sediment and bedrock features, process markers 

 such as bedforms and slumping, and to locate submersible dive 

 sites. 



Investigation of sedimentation rates and seasonality of 

 sedimentation is essential to the study of the nearshore 

 environment. Two techniques were used, shipborne sampling, with 

 a Smith-Maclntyre grab sampler (Kelley et al., 1987a, b) , and 

 submersible-manipulated sampling techniques. Smith-Maclntyre 

 grab samples were collected during 1986 and 1987. Samples were 

 described for grain size and fauna, and then stored for 

 laboratory analysis. Over 1000 samples have been collected from 

 southwestern and central Maine (Kelley et al . , 1987a, b; Kelley 

 and Belknap, 1988) . Only a small percentage of these have yet 

 received a complete grainsize and compositional analysis. No 

 cores are yet available from the present series of studies. 

 Submersible manipulated sampling included scoop samples with the 

 manipulator arm, and a few experimental manipulations with box 

 cores and tube cores. Neither of these techniques were used 

 extensively. Preliminary sampling of suspended sediments was 

 accomplished in 1987 using sediment traps of our own design 

 (Figure 5) . Two pairs of sediment traps were deployed on the 

 Kennebec River paleodelta (Belknap et al., 1986) in Sheepscot 

 Bay, and two pairs were emplaced in Saco Bay (Figure 1) . 



The sediment trap array bases are triangular, with a 2 m 

 base, constructed from 5 cm PVC pipe, and filled with concrete 

 ("Figure 5) . The hypotenuse leg was marked into decimeter 

 segments to facilitate measurement of bedforms and photographic 

 monitoring of their movement. Vertical legs were installed on 

 the two short sides of the triangle to support sediment traps. 

 The traps themselves are removable, for emplacement and removal 

 by submersible. Each was constructed of 10 cm diameter by 30 cm 



150 



