We simulated documented behaviors using the Autonomous Behaving Lagrangian Explorer (ABLE). It can be programmed to maintain depth or vertically migrate in response to in-situ variables, like the larvae under study. It can reveal quasi-Lagrangian transport of vertically migrating plankters that swim between water parcels at different depths. ABLE weighs 3 kg and is 36 cm tall, topped by a 15 cm antenna mast. It necessarily integrates water motions at and below its own scale. Consequently, it cannot mimic transport of individual plankters, nor diffusive processes at scales smaller than its own. ABLE best simulates the transport of the centroid of a cloud of plankters that is large relative to its own dimensions.

ABLE dynamically calculates its target depth from measurements of its immediate microenvironment and a behavioral model for the organism being simulated. It moves toward the new target depth at a biologically realistic velocity, permitting it to show transport consequences of adaptive behaviors in response to actual (not average) conditions and actual (not modeled) water movements. Because behavioral patterns are under the experimenter’s control, ABLE can reveal effects of either known or hypothetical behavior patterns. ABLE has no structures outside the parcel of water in which it is embedded, hence no extraneous drag that would cause drift errors. Use of ABLE (unlike modeling) requires no a priori characterization of the system before the first data can be collected; immediately upon deployment it begins yielding information on how water and organisms in the system move.

Although ABLE has no extraneous drag, hence no drift errors, while embedded in the tracked water parcel, it must periodically leave that parcel and make excursions to the surface to obtain and transmit GPS fixes. A drift error is created by velocity differences (relative to the target parcel) at other depths multiplied by the time ABLE spends transiting each during a pop-up, which cannot be simply estimated in heterogeneous systems. A rule of thumb analogous to that for suspended-drogue drifters would be that ABLE must spend <1/40 of the time making excursions to the surface. As target (operating) depth increases, transit time to the surface increases, and hence allowable fix frequency decreases.

To facilitate tracking, it has an ultrasonic beacon that provides bearings and telemeters depth during operation at depth; when at the surface it obtains fixes from its GPS receiver and transmits the fix data by VHF radio (short range) and satellite modem (global range). The GPS fix obtained at each surface interval is logged in ABLE’s data memory, even if it is not received by the Globalstar satellite system. To facilitate recovery at the sea surface, it transmits updated fixes continuously by VHF and periodically via satellite while blinking high-brightness LED beacons for visual fixes. We also command ABLE to surface for recovery by decoding ultrasonic signals while rejecting noise from surf and biota. It senses the bottom and swims up a programmed distance above the substrate.

When deployed, it uses measurements of in-situ variables (depth, T, S, PAR, time of day, vertical speed relative to water). It subtly adjusts buoyancy (by < 1g) to "swim" toward that target depth, maintaining a rate realistic for the organism being simulated (0 to >10 cm/s). It periodically pops to the surface to obtain a GPS fix and transmit it by VHF, ultrasonic pinger and satellite (or cell phone) modem. Along its entire trajectory, it logs in-situ measurements; the suite of variables and frequency of logging are user-selectable. On the bench, ABLE communicates by wireless Bluetooth with a host computer or smart phone and presents a menu for downloading logged data, testing and calibrating sensors, altering data logging parameters, or even rewriting the entire program. Endurance during deployments is about 2 wk with 7 NiMH "D" cells, depending on frequency of excursions to the surface and pumping of ballast to hoist antennas.