Description: Oceanic uptake and long-term storage of atmospheric carbon dioxide (CO2) are strongly driven by the marine “biological pump,” i.e., sinking of biotically fixed inorganic carbon and nutrients from the surface into the deep ocean (Sarmiento and Bender, 1994; Volk and Hoffert, 1985). Sinking velocity of marine particles depends on seawater viscosity, which is strongly controlled by temperature (Sharqawy et al., 2010). Consequently, marine particle flux is accelerated as ocean temperatures increase under global warming (Bach et al., 2012). Here we show that this previously overlooked “viscosity effect” could have profound impacts on marine biogeochemical cycling and carbon uptake over the next centuries to millennia. In our global warming simulation, the viscosity effect accelerates particle sinking by up to 25%, thereby effectively reducing the portion of organic matter that is respired in the surface ocean. Accordingly, the biological carbon pump's efficiency increases, enhancing the sequestration of atmospheric CO2 into the ocean. This effect becomes particularly important on longer time scales when warming reaches the ocean interior. At the end of our simulation (4000 A.D.), oceanic carbon uptake is 17% higher, atmospheric CO2 concentration is 180 ppm lower, and the increase in global average surface temperature is 8% weaker when considering the viscosity effect. Consequently, the viscosity effect could act as a long-term negative feedback mechanism in the global climate system.

Description: The ocean's potential to export carbon to depth partly depends on the fraction of primary production (PP) sinking out of the euphotic zone (i.e., the e-ratio). Measurements of PP and export flux are often performed simultaneously in the field, although there is a temporal delay between those parameters. Thus, resulting e-ratio estimates often incorrectly assume an instantaneous downward export of PP to export flux. Evaluating results from four mesocosm studies, we find that peaks in organic matter sedimentation lag chlorophyll a peaks by 2 to 15 days. We discuss the implications of these time lags (TLs) for current e-ratio estimates and evaluate potential controls of TL. Our analysis reveals a strong correlation between TL and the duration of chlorophyll a buildup, indicating a dependency of TL on plankton food web dynamics. This study is one step further toward time-corrected e-ratio estimates

Description: A global Earth System model is employed to investigate the role of direct temperature effects in the response of marine ecosystems to climate change. While model configurations with and without consideration of explicit temperature effects can reproduce observed current biogeochemical tracer distributions and estimated carbon export about equally well, carbon flow through the model ecosystem reveals strong temperature sensitivities. Depending on whether biological processes are assumed temperature sensitive or not, simulated marine net primary production (NPP) increases or decreases under projected climate change driven by a business-as-usual CO2 emission scenario for the 21st century. This suggests that indirect temperature effects such as changes in the supply of nutrients and light are not the only relevant factors to be considered when modeling the response of marine ecosystems to climate change. A better understanding of direct temperature effects on marine ecosystems is required before even the direction of change in NPP can be reliably predicted.

Description: Gravitational sinking of photosynthetically fixed particulate organic carbon (POC) constitutes a key component of the biological carbon pump. The fraction of POC leaving the surface ocean depends on POC sinking velocity (SV) and remineralization rate (Cremin), both of which depend on plankton community structure. However, the key drivers in plankton communities controlling SV and Cremin are poorly constrained. In fall 2014, we conducted a 6 weeks mesocosm experiment in the subtropical NE Atlantic Ocean to study the influence of plankton community structure on SV and Cremin. Oligotrophic conditions prevailed for the first 3 weeks, until nutrient‐rich deep water injected into all mesocosms stimulated diatom blooms. SV declined steadily over the course of the experiment due to decreasing CaCO3 ballast and – according to an optical proxy proposed herein – due to increasing aggregate porosity mostly during an aggregation event after the diatom bloom. Furthermore, SV was positively correlated with the contribution of picophytoplankton to the total phytoplankton biomass. Cremin was highest during a Synechococcus bloom under oligotrophic conditions and in some mesocosms during the diatom bloom after the deep‐water addition while it was particularly low during harmful algal blooms. The temporal changes were considerably larger in Cremin (max. 15‐fold) than in SV (max. 3‐fold). Accordingly, estimated POC transfer efficiency to 1000 m was mainly dependent on how the plankton community structure affected Cremin. Our approach revealed key players and interactions in the plankton food web influencing POC export efficiency thereby improving our mechanistic understanding of the biological carbon pump. Key points Sinking velocity was higher during oligotrophy than during blooms which is linked to ballast, porosity, and phytoplankton size structure Remineralization was highly variable but tended to be higher during Synechococcus or diatom blooms and lower during harmful algal blooms Plankton community structure had a considerably larger influence on particle remineralization rate than on sinking velocity