Organic matter remineralisation and biogenic silica dissolution in a deep fjord in British Columbia, Canada: a regression analysis of upper ocean sediment-trap fluxes

Remineralisation of sinking particulate matter can be inferred from sediment-trap studies where fluxes decrease with depth. However, during a 4-year experiment of monthly sediment-trap deployments at two stations in Jervis Inlet, British Columbia, Canada, fluxes often increased with depth, a common observation in coastal and some oceanic settings. A regression analysis that simultaneously estimates the composition of ‘additional’ fluxes to deep sediment traps and inverse decay length (IDL) scales of the primary (operationally termed ‘anticipated’) flux is applied to the data from Jervis Inlet. IDL scales for organic carbon (OC), nitrogen (N) and biogenic silica (BSi) decreased with depth in a manner well described by the power function, and are used to generated curves of flux versus depth in the absence of additional fluxes. The curve for OC flux is comparable to previously published models, but in Jervis Inlet remineralisation was somewhat higher in the upper water column, and lower in deep waters, than implied from open-ocean curves. The first upper-ocean estimates for dissolution rates of sinking BSi are presented; conversion to time-dependent rate constants shows that BSi dissolution in Jervis Inlet was similar to results from other studies where dissolution was rapid. A fitting of the flux curves to the reactive continuum model demonstrates that settling organic matter in Jervis Inlet had 3–4 reactive components, while BSi had 6–7 reactive components and over-all was less susceptible to decay. onal fluxes to deep sediment traps in Jervis Inlet appear to have resulted partly from resuspended sediment increasingly winnowed with distance from topographic boundaries. Resuspended fluxes were particularly abundant during deepwater renewals, and another source supplied additional fluxes to mid-depth and deep traps on a continual basis. Included in the analysis is an error term, which is interpreted as the degree to which anticipated fluxes were over- or under-represented in deep sediment traps. Although for the deep depth intervals in Jervis Inlet errors of regression analyses were negligible, for the more shallow intervals they were not. An enigma of the results from Jervis Inlet is that it appears OC, N and Al were generally over-collected at mid-depth sediment traps, while BSi was under-collected.