Tuesday, August 25, 2009

In Search of the Asian Clam...

Researchers at the UC Davis Tahoe Environmental Research Center invited Gavia operators from the University of British Columbia, University of New Hampshire, and University of Delaware to Lake Tahoe in search of the invasive Asian clam that is becoming more prevalent in these waters. In a true collaborative effort between UBC and UDel/UNH AUV teams, the Gavia was run with the UBC control, props, and nose module and the UDel/UNH positioning system and Geoswath, with the battery modules of each being swapped in to extend run time. Due to the transparency of lake waters of up to 70m, our primary objective was to photograph the lake bed to determine clam distributions. The clams threaten the clarity of the lake by increasing nutrient levels in the lake, encouraging a thick algae growth to carpet the bottom. Consequently we also collected water quality measurements including CDOM, backscatter, chlorophyll concentration, and CTD during these missions.

The Gavia ready for night swimming in the quiet waters of Lake Tahoe, is trimmed with incompressible foam needed to provide lift at the nose in the freshwater.

Working on the boat from dusk until dawn had its rewards: beautiful sunrises on the lake!

Our first week of missions entailed circumnavigating the lake along the 5m contour at an altitude of 2.5m. In order to optimize the clam detection algorithm we wrote, which targets the white color of dead clams, we ran missions at night to avoid sunlight reflections on the bottom, which would be classified as clams. We also avoided boat traffic by executing missions at night. Night work in clear water made tracking the Gavia easy, as we just followed the strobe/tower lights. This also allowed us to precisely time when the camera card was full by having a stopwatch that started when the strobe started and stopped when it turned off. We captured 2 hours of photos before pulling the Gavia back aboard and transferring the photos to the Toughbook via LAN.

An image of the bed reveals sandy sediments with an abundance of clams and algal growth.
Our image-processing algorithm is shown below with the original image in the top right, initial clam detections in the top left, and advanced clam detection in the bottom right.
Mission planning was the most challenging task as the perimeter of the lake is full of erratics. Our most successful technique for mission planning included using Fledermaus to locate waypoints on the 5m contour from the multibeam data Mayer et al. published, Control Center to plot the points, and GoogleEarth to check for erratics. Though tedious this has allowed us to run 2-hour missions (to fill the camera card) without having to stop and restart due to obstacles. This afforded us time to watch the amazing Perseid meteor showers.

A glance at the Lake Tahoe shoreline by day reveals the dangers of flying the Gavia through shallow waters!
At the end of the first battery pack our team transformed from Gavia-/star- gazers to a professional pit-crew able to download data, change the Gavia battery (which entailed removing a CTD, buoyant trim, cleaning o-rings, and carefully aligning the modules) with great efficiency.

The circumnavigation was completed by the start of our second week and yielded important information on clam presence around the lake. The remainder of our missions was spent surveying areas with high clam concentrations using two strategies: 1. Contour following at 10-15m depth intervals by running shorter survey lines parallel to contours as in the circumnavigation. 2. Running survey lines perpendicular to contours from about 5m to 70 m depth or to the edge of the shelf break. Bottom-tracking proved difficult during these missions, the Gavia would often abort missions because she was unable to reach a waypoint in time. Another difficulty was maximizing the number of useable images we could collect in missions that cut across contours. Flying across contours was challenging because the Gavia has trouble bottom-tracking along steep areas, which are prevalent along the basin walls. Slopes of greater than 10 to 15 degrees were common and because of this we ran more, shorter missions. We would send Gavia perpendicular to the contours at 1.5 m altitude and then make the return leg constant depth of 2 m to avoid ploughing the AUV into the steep slopes. Because of the many sites of interest and the relatively time-consuming nature of this mode of operation, we ran about 2-4 lines per site.

Project PI, Marion Wittmann, and boat captain, Brant Allen of UC Davis enjoying the night on Tahoe.
Throughout our two weeks the Gavia covered more ground than on any previous research trip. We found clams at greater depths (80 m) than previously discovered and also imaged algal mats and the pond liner used to control clam populations along the south shore. The only disappointment is that we did not have the time to run survey lines with sufficient overlap to obtain quality Geoswath data. Leaving Tahoe with the knowledge that Gavia has collected an abundance of data and successfully contour-followed and run across steep contours, two seemingly impossible survey goals, is extremely rewarding. This was a spectacular end to the Gavia’s and my first full field season with Team CSHEL!Andrew, me, and Alex keeping warm in the cold dawn hours on Lake Tahoe.

No comments: