Intercomparison of Extended-Depth SBE41CP, SBE61, and RBRargo|deep6k CTDs for Deep-Argo Application Using Three- and Two-Headed Deep-Arvor Floats

Abstract Within the international OneArgo program, a global array of autonomous profiling floats monitoring seawater properties, the Deep-Argo mission aims to provide temperature, pressure, and salinity measurements down to the seabed with accuracy targets of ±0.001°C, ±3 dbar, and ±0.002, respectively. One of Deep-Argo’s main challenges is to achieve this level of accuracy. Three different conductivity–temperature–depth (CTDs) are available for Deep-Argo applications: the extended-depth SBE41CP, SBE61, and RBR argo |deep6k. We evaluated their performance at sea down to 4000 dbar using four Deep-Arvor floats equipped with two or three of these CTDs. Pressure differences between the sensors ranged from about 0–1 dbar near the surface to a maximum of 5 dbar at 4000 dbar. Temperature differences were within ±0.002°C below 1000 dbar. A time lapse of up to 0.5 s between sensor acquisitions occasionally led to sensor-independent temperature differences greater than ±0.002°C at the shallowest levels. Pressure differences as small as 1 dbar can induce temperature differences greater than 0.01°C in a large temperature gradient. Independent shipboard-calibrated CTD observations were used to correct, within an uncertainty of ±0.004, a pressure-dependent salinity bias found on all CTDs, as well as a salinity offset on two of them. After correction, salinity differences between the sensors were less than 0.004 below 500 dbar. They increased to 0.01 at shallower depths, as any remaining pressure-dependent errors were projected onto the surface layers. To achieve Deep-Argo’s target accuracy through the intrinsic quality of the sensors, pressure sensor accuracy and compressibility coefficient of the conductivity sensor estimates need to be improved.