One-by-one, the bright yellow, finned cylinders, bristling with sensors are lowered from the ship's stern into the ocean. Then, following internal control programs, they sink slowly, in unison, to about 200 m, where they pause then slowly ascend back toward the surface with sensors firing to measure temperature, depth, salinity, currents, and turbulence. When each 90-minute excursion ends with a return to the surface, the robot’s location is fixed by GPS, then the profile data and position information are transmitted back to the ship by a satellite uplink.

Investigators aboard the R/V Sikuliaq used swarms of EM-APEX floats, shipboard sensors, and towed instruments to survey the North Pacific Subtropical Front north of Hawaii for one month in early 2017. Here, they were looking for regions of active de/re-stratification — the upper ocean adjusting in response to mixing induced by strong storms — during the Submesoscale MIxed-layer Eddies (SMILE) experiment funded by the National Science Foundation.

During SMILE, the team completed deployments and recoveries of three EM-APEX swarm arrays, with 9, 16, and 23 floats profiling the upper ocean for 5, 6, and 7 days, respectively. Programming the swarm to profile in sync addresses a common problem for observational oceanographers — space-time aliasing. Measurements from the profiling swarm help them to separate the observed spatial features associated with the ship moving from one water mass to another from temporal features. This enables a rendition of the actual three-dimensional evolution of the upper ocean's structure.

Based on early analyses, the SMILE data collection strategy was a success. The ship surveys and float swarms captured spatial snapshots at regular intervals, and data from each can be mapped independently and compared. "Our challenge is to understand the advection of horizontal property gradients in the upper ocean, and we know we were sampling the gradients and velocities with EM-APEX swarms," states Principal Oceanographer James Girton. He is confident that they captured the submesoscale (1–10 km) dynamics well, which was the primary goal. This is no small challenge because at this small scale, the rules oceanographers apply to larger-scale motions in the ocean can fall out of balance. Girton adds, "This is one reason why we needed to measure rapidly and with dense coverage."