Project Title: An Autonomous Profiler for Estuarine Research and Monitoring
Principal Investigator(s): Rocky Geyer, Daniel Frye, and Kenneth Doherty

Accomplishments
The objective of this development project is to design, build and deploy a general purpose estuarine winch that can perform hourly profiles of the water column with a standard sensor suite for several months. The tasks completed in this final reporting period included verifying the performance of the winch, winch controller and sensor module during two deployments in the Hudson River. The initial trial was conducted in about 5 m of water between April 23 and May 13, 2002. Profiles were collected hourly during this period with a few gaps due to strong tidal currents. The sensor module attachment point was modified for better high current performance and the system was redeployed in about 7.5 m of water on May 21, 2002. Three hundred and ninety one profiles were collected before the winch was recovered on June 6. Figure 1 (see Figure 1, The history of the Profiling vehicle versus depth for the second Hudson River deployment) shows the time history of the sensor module during the second test deployment. The RF link (an Argos PTT) was integrated into the sensor module and tested successfully during the Hudson River deployments. It was programmed to transmit a small subset of the measured profile data as a way to verify that the winch was working properly and that it was still on station.

Anticipated Success in Meeting Project Objectives in Scheduled Project Period
The profiler vehicle (see Figure 2, Sensor Module) operates autonomously and is moved through the water column by the bottom-deployed winch unit. The profiler vehicle contains a Seabird SBE19 CTD with turbidity sensor, an Argos PTT and a system controller based on an Onset Tattletale8. The winch unit and profiler vehicle are completely self-contained and do not communicate during operation. The vehicle controller has a serial interface that is used to program the vehicle and winch unit simultaneously. Prior to deployment, the user selects a schedule that includes start and stop times, time between ascents and the number of minutes at the surface. Once the schedule is entered, the vehicle controller synchronizes both system clocks and downloads the profile information. The vehicle monitors time with its internal clock and turns on the CTD one minute before a profile begins and starts logging the CTD data to internal flash memory. Independently, the winch unit releases its brake and allows the vehicle to ascend. The vehicle controller monitors the pressure information from the CTD to determine when it has reached the surface. At the surface, a sub-sampling of the data from the previous profile is transmitted through the Argos PTT. The vehicle's location is determined from the Argos transmissions. After the surface interval, the winch retracts the vehicle while the CTD continues to collect information. The CTD is turned off after a few minutes and goes into low power mode.

Preliminary Results
The controller/data logger is an Onset Model 8 TattleTale. It serves three functions: 1. Profile scheduling, i.e., experiment start time, surface interval, data transmit interval, etc. 2. Motor control, as explained below, and 3. Data logging The motor control circuitry, which interfaces with the Model 8, consists of several computer-controlled switches, two current sinks, current and voltage signal conditioning electronics and an analog voltage limiter. The high current sink is set to limit at 3.2 Amps. The low current sink limits at 400 mA and the voltage limiter cuts in at 40 Volts. When a profile is begun, the controller releases the motor's brake by energizing a solenoid. As buoyancy drives the profiling vehicle to the surface, motor voltage is monitored by the voltage limiter and clamped to 40 Volts, thus limiting the maximum ascent rate to approximately 50 cm/s. The controller also monitors this voltage and when it drops below a set point for over 5 seconds, it is assumed that the profiler has reached the surface. The battery voltage at the beginning of an ascent is logged as well as the voltage on ascent. Once the profiler is on the surface, the controller connects the motor to the 48 Volt supply so that the winch can respond rapidly to wave action. The wave response feature is designed to minimize slack in the winch line and the possibility of the winch line getting tangled or snagged. This feature has worked very well in the field trials and is a key feature in the reliability of the system under a variety of wind, wave and current conditions. When the scheduled surface time has elapsed, the controller connects the winch motor to the 15 Volt supply to recover the profiler. This voltage provides a retrieval rate of approximately 15 cm/s. The controller monitors and logs the drive current during the descent. When this current rises past a set point, it is assumed that all the line has been retrieved and that the profiler is now at the bottom. Battery voltage is logged and the controller enters a very low power sleep mode awaking once a second to monitor the time. At the next scheduled ascent time, the cycle repeats. Figures 3 (see Figure 3, Time history of profiler motion during part of the initial Hudson River deployment), 4 and 5 are examples of data collected by the profiler during the first Hudson River deployment. They illustrate the temporal and vertical resolution achievable with the estuarine winch system, which is far more detailed than is possible with discrete instruments or even using ship lowered equipment. Figure 4 (see Figure 4, Profiler turbidity and salinity versus depth and time during the initial Hudson River deployment. Spring tide conditions.) shows the water column properties during typical spring tides and Figure 5 (see Figure 5, Profiler turbidity (upper) and salinity (lower) versus time during the initial Hudson River deployment. Neap tide conditions.) is the analogous plot for neap tide conditions.

Tasks and activities for next reporting period
We have designed, built and tested an estuarine winch system that achieves our original design goals. It can perform hourly profiles in 20 m of water for up to 2 months. It is reliable even in areas with high turbidity and high-suspended sediment loads, high waves and strong currents. The system is lightweight and can be deployed and recovered from a small boat. It is easy to program and operate and is reasonably inexpensive to build. We have had discussions with McLane Research Laboratories of Falmouth, MA about producing the winch system commercially and they are planning to pursue SBIR funding to develop a production version of the system. One of the PIs (Geyer) has plans to use the winch system on an upcoming field program in the Mediterranean early next year. Enhancements of the system that we hope to pursue include: 1. Profiler vehicle designed for lower drag (and possibly more buoyancy) so that the system will operate in higher currents and deeper water. We think it will be possible to operate the system in depths of 50 to 100 m using more line on the winch and a larger winch drum. 2. Improved RF link so that we can transmit all of the data collected. The Argos link presently implemented is a very limited data link. We have designed an Iridium link for the profiler that has the bandwidth to transmit all of the data collected, but this will require some redesign on the profiler body and antenna mount.