Modelling the effect of vertical mixing on bottle incubations for determining in situ phytoplankton dynamics. II. Primary production

The estimation of in situ phytoplankton primary production is pivotal to many questions in biological oceanography and marine ecology both in a local and global context. Applications range from earth system modelling, the characterisation of aquatic ecosystem dynamics, or the local management of water quality. A common approach for estimating in situ primary production is to incubate natural phytoplankton assemblages in clear bottles at a range of fixed depths and to measure the uptake of carbon ( 14 C) during the incubation period (typically 24 h). One of the main concerns with using fixed-depth bottle incubations is whether stranding samples at fixed depths biases the measured CO 2 fixation relative to the 'true' in situ mixed conditions. Here we employ an individual based turbulence and photosynthesis model, which also accounts for photoacclimation and -inhibition, to examine whether the in vitro productivity estimates obtained from fixed-depth incubations are representative of the in situ productivity in a freely mixing water column. While previous work suggested that in vitro estimates could either over-or underestimate the in situ productivity, we show that the errors due to arresting the incubation bottles at fixed depths are indeed minimal. We present possible explanations for how previous authors could have arrived at contradictory results and discuss whether they might be artefacts related to the particular sampling protocol used. We discuss the errors associated with chlorophyll-based incubation methods for determining in situ phytoplankton growth rates in Ross et al. (