Subtidal: In the dataset we truncated the size of the smallest measured invertebrates at 14 mm (mesh diagonal dimension). Shallow subtidal sampling targeted established clam beds throughout the central and lower parts of Glacier Bay (Bodkin et al. 2002). From 2001-2002, 13 subtidal sites were chosen based on three criteria: 1) proximity to areas occupied and not occupied by sea otters; 2) spatial separation from other sites; and 3) relatively high clam densities at the study's onset. To approximate clam densities, divers counted clam siphons on drop camera transects with visible shell litter. Due to the logistical constraints of working underwater in an area with poor visibility and substantial tide swings, sites were restricted to maximum depth of 12 m at high water even though sea otters are capable of diving much deeper. When a location was found to meet the three criteria, GPS coordinates from the center of the clam bed were recorded to mark the site. Using scuba and a diver-operated suction dredge, divers randomly placed and excavated, twenty 0.25 sq meter quadrats within a circular 314 sq meter area. While one diver used the dredge to remove sediment from inside the quadrat, the other diver collected large invertebrates as they became exposed and placed them into a mesh bag. Sediments containing smaller infauna were deposited into a mesh bag connected to the end of the dredge exhaust hose and later sieved through 10 mm screens (Weitzman et al. 2012). All bivalves, crabs, and urchins were collected, identified to the lowest possible taxon, and measured to the nearest millimeter. In 2011, 11 of the original 13 subtidal sites were resampled using the same methods.
Intertidal: Intertidal sampling for sea otter prey occurred from both random and selected sites spread across the entirety of Glacier Bay National Park and Preserve (GBNPP). The coastline of GBNPP was divided into 5,545 200 meter-long shoreline segments. Every 23rd segment (241 segments) was selected for an aerial survey to categorize slope, substrate, and percent coverage of major biota (Irvine 2010). After eliminating segments that were too steep, had too much bedrock, or had been selected for inclusion in a monitoring protocol development study, we chose a random starting segment and every 3rd segment for sea otter prey sampling. Several selected sites were eliminated due to snow avalanche danger, consolidated substrate, or excessive mud resulting in a total of 48 randomly selected sites being sampled. The purpose of this random sampling procedure was to provide data that would be representative of unconsolidated habitat throughout all GBNPP so results of analyses could be extrapolated to all of GBNPP.
These 48 sites were supplemented with 12 selected sites referred to as 'Preferred Clam Habitat' [PCH] sites, identified based on presence of clam siphons and squirts at low tide, along with obvious shell litter deposition. Our intent for sampling PCH sites was to provide information from areas with potential for large changes in prey abundance.
Some sites were sampled multiple times to provide data on annual and short-term variation and eventually, examine the impacts of sea otter predation and long-term variation. In 2004 we randomly selected a subset of the 60 intertidal sites (48 random plus 12 PCH) sampled in 1999-2000 to examine intra-annual variability within sites. Based on observations of foraging sea otters and the presence of sea otters during surveys, sites that were likely to have been impacted by sea otters were excluded from the random sample we drew. The remaining sample included 15 sites, 9 random and 6 preferred clam habitat (PCH) sites that had been previously sampled. In order to facilitate future resampling, 2 – 3 ft rebar stakes were installed at the starting end of the 200 m transect at 0 MLLW. Additionally, GPS readings were collected for the starting and ending points and tracklines saved for the transect. It should be noted that although the sites to be resampled were selected based on presupposed lack of sea otter influence, field observations indicated evidence of otter impacts such as foraging pits and/or clam debris (shell litter with characteristic otter-preyed cracking patterns and siphons that had not been consumed) at some sites.
All 60 intertidal sites were sampled in the early period (1998-2000), 19 sites were resampled in the mid time period (2004-2007), and all 12 PCH and 33 (out of 48) random sites were resampled in the late period (2010-2011). Access restrictions prevented resampling at several sites; in the lower Bay these restrictions were due to National Park regulations; whereas in the upper bay, some areas were so physically altered due to landslides and glacial activity, that the original sites were inaccessible.
Protocols developed in Glacier Bay were later published and used to monitor mixed-sediment beaches in the Gulf of Alaska (Weitzman et al. 2017) except that transects extended for 200 m of contiguous beach in the Glacier Bay protocol and 10 quadrats were excavated at each intertidal site in the Glacier Bay protocol. All sampling was conducted during a low tide series when the water level was below 0 mean lower low water (0 MLLW).
To determine 0 MLLW, daily tide tables based on NOAA tide stations were printed in 6-minute intervals for the tide prediction station nearest each site. No corrections were made for distance from tide prediction station or geographic features affecting flow. At each random site, a handheld GPS was used to navigate to the specific segment coordinates. The starting point was set as the left-hand end of the segment as the observer’s back was to the water. At the predicted time for 0 MMLW, two 100 m transect tapes were laid in series, parallel to the waterline.
Beginning at a randomly selected starting location within the first 20 m, ten 0.25 m2 quadrats were placed 20 m apart along the water-side of the transect and excavated to 25 cm depth or until clams were no longer being unearthed, whichever was deeper. Before removal of any sediment, all sea urchins were collected from the surface layer. Unearthed sediment was passed through a 10 mm mesh sieve to obtain the associated infauna. All recovered clams and sea urchins were identified to the lowest possible taxon, counted, and measured at their widest dimension to the nearest millimeter using dial calipers. We truncated the size of the smallest measured invertebrates at 14 mm (mesh diagonal dimension). Field crew noted observations of signs of site use by sea otters such as otter-cracked shell litter, foraging pits, presence of foraging otters or large rafts nearby).