No abstract

Abstract. Process-based numerical models are a useful tool for studying marine ecosystems and associated biogeochemical processes in ice-covered regions where observations are scarce. To this end, CSIB v1 (Canadian Sea-ice Biogeochemistry version 1), a new sea-ice biogeochemical model, has been developed and embedded into the Nucleus for European Modelling of the Ocean (NEMO) modelling system. This model consists of a three-compartment (ice algae, nitrate, and ammonium) sea-ice ecosystem and a two-compartment (dimethylsulfoniopropionate and dimethylsulfide) sea-ice sulfur cycle which are coupled to pelagic ecosystem and sulfur-cycle models at the sea-ice–ocean interface. In addition to biological and chemical sources and sinks, the model simulates the horizontal transport of biogeochemical state variables within sea ice through a one-way coupling to a dynamic-thermodynamic sea-ice model (LIM2; the Louvain-la-Neuve Sea Ice Model version 2). The model results for 1979 (after a decadal spin-up) are presented and compared to observations and previous model studies for a brief discussion on the model performance. Furthermore, this paper provides discussion on technical aspects of implementing the sea-ice biogeochemistry and assesses the model sensitivity to (1) the temporal resolution of the snowfall forcing data, (2) the representation of light penetration through snow, (3) the horizontal transport of sea-ice biogeochemical state variables, and (4) light attenuation by ice algae. The sea-ice biogeochemical model has been developed within the generic framework of NEMO to facilitate its use within different configurations and domains, and can be adapted for use with other NEMO-based sub-models such as LIM3 (the Louvain-la-Neuve Sea Ice Model version 3) and PISCES (Pelagic Interactions Scheme for Carbon and Ecosystem Studies).

In this paper we review available time series for the Strait of Georgia to identify trends and variability in physical and biogeochemical properties. Change is partly imported from the open ocean and partly results from processes operating at the local scale. The largest component of variation occurs at the seasonal scale, although the timing in annual cycles differs among properties. A second important component of variability is associated with cycles at the decadal (PDO) or sub-decadal scales (ENSO). Long-term trends are superimposed on the variability. Seawater in the Strait has been warming at N1°C/century, as has the freshwater entering from the Fraser River. The number of days when Fraser River temperature exceeds the 18°C threshold for salmon migration has increased over the last 50 years. In the Strait itself, the temperature increases are of the same magnitude in deep water as at the surface, but are probably more significant in the deeper water because of the narrow seasonal range of temperature at depth. The change in annual freshwater discharge from the Fraser River over the period of record is much smaller than the interannual variability, but there has been a notable change in timing, with more of the discharge occurring in spring and less in summer. This, together with warming, may be producing an earlier spring bloom and an altered coupling between phytoplankton and zooplankton. Sea-level rise is occurring within the Strait at rates similar to other locations, but the presence of the large Fraser River delta, undergoing industrial and municipal development, makes this region especially sensitive to sea-level rise and to increased storm activity. Variability in bottom water properties is predominantly forced from outside the basin, depending especially on the timing of coastal upwelling, which delivers water containing high nutrients and low dissolved O 2 and pH. As in the global ocean, pH in the Strait is likely declining, but records remain too short to produce a confident assessment. The timing of geochemical cycles in the Strait of Georgia is delicately poised, with, for example, deep-water oxygen reaching a hypoxic tolerance threshold in the spring, just before deep-water renewal replenishes the oxygen from outside. However, long-term trends in oxygen, temperature and timing of biological activity may lead to the crossing of crucial biological tipping points within this century. Timing is particularly important for monitoring. Relatively long records for basic water properties like temperature and salinity are accompanied by much shorter records for biogeochemical properties like dissolved O 2 , pH, nutrients and vertical flux, making it difficult to assemble a clear picture of the sorts of changes that may be occurring in the latter. A confident assessment of the ecological resilience of the Strait of Georgia will require longer time series of biological and geochemical properties that are collected with consideration for the strong seasonal variability.

Dinitrogen fixation (DNF) provides a large fraction of the 'new' nitrogen supporting upper ocean productivity, and is associated with environmental conditions likely to show substantial change under anthropogenic warming. For example, surface warming induces stronger stratification, weaker nutrient supply and more rapid nutrient depletion. Using six Earth System Models, we have examined spatial patterns and trends of DNF in the CMIP5 historical and RCP 8.5 experiments. Four models (CanESM2, CESM1-BGC, IPSL-CM5R-LR, and UVicESCM) show high DNF rates in warm, stratified waters mostly associated with the western parts of the ocean basins, while GFDL-ESM2M and MPI-ESM-LR show elevated rates near the eastern boundaries because of coupling of DNF and denitrification. Despite a growing body of data, the spatial pattern of DNF is still insufficiently resolved by available observations, and none of the models agrees well with the observations. Modelled and observed rates are mostly in the same general range except for UVicESCM, and frequency distributions are similar, but spatial pattern correlations are weak and in most cases not statistically significant. Only a few models show strong trends in DNF and primary production in a warming climate, and the signs of the trends are inconsistent. Observations of primary production at the benchmark subtropical station ALOHA (22.75°N, 158°W) and proxies for historical DNF from the same region appear to corroborate trends in CanESM2 that are not present in other models. However, the CanESM DNF parameterization does not include any limitation by P or Fe, so modelled future trends may not materialize due to nutrient limitation. Analysis of available models and observations suggests that our understanding of environmental controls on ocean DNF remains limited and future trends are highly uncertain. Long-term global simulations of DNF will only be meaningful if we maintain long-term observations and extend coverage to undersampled regions.Keywords: ocean dinitrogen fixation; Earth system modelling; anthropogenic climate change Riche and Christian: Ocean dinitrogen fixation and its potential effects on ocean primary production in Earth system model simulations of anthropogenic warming Art. 16, page 2 of 18 of phosphorus or iron are available, with opposing effects on primary production (Karl et al., 1995(Karl et al., , 1997Karl, 1999). As diazotrophic phytoplankton thrive in warm, stratified waters (Sohm et al., 2011), investigating the potentially enhanced role of diazotrophs in tropical and subtropical ocean ecosystems in a warming climate is important. Proxy δ 15N measurements of the DNF contribution to the N inventory in the subtropical Pacific suggest that the recent apparent enhancement discussed by Karl (1999) is a long-term trend throughout the 20 th century, possibly associated with anthropogenic warming (Sherwood et al., 2014).We have examined trends in DNF in a variety of Earth System Models associated with the CMIP5 project (Taylor et al., 2012) to determine which if any sh...

The goal of this paper is to give a detailed description of the coupled physical-biogeochemical model of the Gulf of St. Lawrence that includes dissolved oxygen and carbonate system components, as well as a detailed analysis of the riverine contribution for different nitrogen and carbonate system components. A particular attention was paid to the representation of the microbial loop in order to maintain the appropriate level of the different biogeochemical components within the system over long term simulations. The skill of the model is demonstrated using in situ data, satellite data and estimated fluxes from different studies based on observational data. The model reproduces the main features of the system such as the phytoplankton bloom, hypoxic areas, pH and calcium carbonate saturation states. The model also reproduces well the estimated transport of nitrate from one region to the other. We revisited previous estimates of the riverine nutrient contribution to surface nitrate in the Lower St. Lawrence Estuary using the model. We also explain the mechanisms that lead to high ammonium concentrations, low dissolved oxygen, and undersaturated calcium carbonate conditions on the Magdalen Shallows.

Abstract. The ocean biogeochemistry components of the Canadian Earth System Model v. 5 are presented and compared to observations and other models. CanESM5 employs the same biogeochemistry module as CanESM2 whereas CanESM5-CanOE (“Canadian Ocean Ecosystem model”) is a new, more complex biogeochemistry module developed for Sixth Coupled Model Intercomparison Project (CMIP6), with multiple food chains, flexible phytoplankton elemental ratios, and a prognostic iron cycle. This new model is described in detail and the outputs compared to CanESM5 and CanESM2, as well as to observations and other CMIP6 models. Both CanESM5 models show gains in skill relative to CanESM2, which are attributed primarily to improvements in ocean circulation. CanESM5-CanOE shows improved skill relative to CanESM5 in some areas. CanESM5-CanOE includes a prognostic iron cycle, and maintains high nutrient / low chlorophyll conditions in the expected regions (in CanESM2 and CanESM5, iron limitation is specified as a temporally static ‘mask’). Surface nitrate concentrations are biased low in the subarctic Pacific and equatorial Pacific, and high in the Southern Ocean. Export production in CanESM5-CanOE is among the lowest for CMIP6 models; in CanESM5 it is among the highest, but shows the most rapid decline after about 1980. CanESM5-CanOE has relatively low concentrations of zooplankton and detritus relative to phytoplankton, and a high and relatively constant living phytoplankton fraction of total particulate organic matter. In most regions, large and small phytoplankton show decoupled seasonal cycles with greater abundance of large phytoplankton in the productive seasons. Cumulative ocean uptake of anthropogenic carbon dioxide through 2014 is lower in both CanESM5 models than in observation-based estimates or the model ensemble mean, and is lower in CanESM5-CanOE (122 PgC) than in CanESM5 (132 PgC).

No abstract

Abstract. Numerical models are a useful tool for studying marine ecosystems and associated biogeochemical processes in icecovered regions where observations are scarce. To this end, CSIB v1 (Canadian Sea-ice Biogeochemistry version 1), a new seaice biogeochemical model has been developed and embedded into the Nucleus for European Modelling of the Ocean (NEMO) modelling system. This model consists of a three-compartment (ice algae, nitrate, and ammonium) sea-ice ecosystem and a twocompartment (dimethylsulfoniopropionate and dimethylsulfide) sea-ice sulfur cycle which are coupled to pelagic ecosystem 5 and sulfur-cycle models at the sea ice-ocean interface. In addition to biological and chemical sources and sinks, the model simulates the horizontal transport of biogeochemical state variables within sea ice through a one-way coupling to a dynamicthermodynamic sea-ice model (LIM2). This paper describes technical aspects of implementing sea-ice biogeochemistry into NEMO and provides discussion on the results of several model experiments. Results of the reference simulation were evaluated by comparing the model outputs to observations and previous modelling studies. Additional simulations were conducted to 10 assess the model sensitivity to 1) the temporal resolution of the snowfall forcing data, 2) the representation of light penetration through snow, 3) advective and eddy-diffusive horizontal transport of sea-ice biogeochemical state variables, and 4) light attenuation by ice algae. The sea-ice biogeochemical model has been developed within the generic framework of NEMO to facilitate its use within different configurations and domains, and can be adapted for use with other NEMO-based submodels such as LIM3 and PISCES.