North West Shelf (WA, 2004-2005)

NWS JEMS Project

Primary production was studies as part of the North West Shelf Joint Environmental Management Study ( Details can be found in:

Herzfeld, M., Parslow, J., Sakov, P., Andrewartha, J. (2006) Biogeochemical modelling of Australia’s North West Shelf. North West Shelf Joint Environmental management Study, Technical Report No. 8.

Key elements of this study are described below.

Research into the dynamics of primary and secondary productivity on the North West Shelf (NWS) has been hampered by the lack of high frequency long term data records. Existing studies indicate that the standing crop of phytoplankton is nitrogen limited on the NWS, with the source of nitrate being nutrient rich water residing on the continental slope below the surface mixed layer. There exists very little nitrate in surface waters and terrestrial inputs are negligible. Enhanced productivity in the mixed layer is attributed to sub-pycnocline fluxes of nitrogen resulting from several processes, including upwelling on time scales of the semi-diurnal tide to lower frequencies (> 35 hours), horizontal and vertical mixing due to the barotropic tide and internal tide activity, and the effect of tropical cyclones. It is rare for nitrate rich slope water to intrude onto the shelf further than the 50m isobath as a result of these processes. These fluxes of nitrogen are roughly constant throughout the year.

The lack of a highly resolved temporal and spatial data set for the NWS shelf region leaves many questions about nutrient cycling and primary productivity within the region unresolved. A numerical biogeochemical model is used here to investigate the primary productivity and nutrient cycling dynamics of the NWS region and provide insight into the key processes controlling the spatial and temporal patterns of primary production.

The model used to investigate the primary productivity and nutrient dynamics on the NWS is a general dynamical process model of biogeochemical and ecological processes developed by CSIRO for estuaries and coastal waters. This model is based on the National Estuarine Audit Model (Baird et al., 2003), which in turn represents an extension of the Port Phillip Bay model (Murray and Parslow, 1999). The model represents the cycling of nitrogen, phosphorous and carbon through both pelagic and benthic ecosystems. The ecological model has three modules: water column, sediment (1 layer), and epibenthos.

The water column module describes a simple planktonic food web. The model currently includes two phytoplankton functional groups: small phytoflagellates and large bloom-forming phytoplankton with nominal cell diameters of 5 mm and 20 mm respectively. These classes can be considered to represent flagellates and diatoms. There are in turn two size classes of zooplankton (0.025 and 1 mm diameter) which graze respectively on small and large phytoplankton. The model represents a range of forms of nonliving particulate and dissolved organic matter, as well as inorganic nutrient species, dissolved inorganic carbon (DIC) and dissolved oxygen.

The sediment module represents the breakdown of particulate and dissolved organic matter through microbial and detritivore activity that consumes oxygen and releases DIC and inorganic nutrients. The sediment and water column modules include the processes of nitrification and denitrification.

The epibenthic module represents two functional classes of attached macrophytes: macroalgae, which take up nutrients from the water column, and seagrass, which take up nutrients from the sediment pore water. A schematic view of nitrogen cycling through water column, sediment and epibenthic components is shown in Figure 1.

A schematic of nitrogen cycling in the NWS model through pelagic, benthic and epibenthic components.

Figure 1: A schematic of nitrogen cycling in the NWS model through pelagic, benthic and epibenthic components.

The area of the NWS modelled focuses on the region from North West Cape to just north of Port Hedland. The continental shelf and slope are resolved throughout the domain (Figure 2). The shelf is wide in the north-eastern section of the domain but becomes very narrow and steep near the southern boundary. The biogeochemical model is coupled directly to the hydrodynamic model, hence the spatial discretization of the study region is identical to that used in the hydrodynamic model (the hydrodynamic modelling of the region is described in Condie et. al, 2006). In order to achieve realistic run-times for the coupled model, a relatively coarse grid size of 20 km alongshore by 10 km cross-shore was used. The time-step used for the biochemical model is different to the hydrodynamic model, and is typically longer. The equations representing the biochemical processes are integrated over the chosen time-step using an adaptive integration scheme suitable for stiff systems, which adjusts the number of sub-steps to achieve prescribed accuracy. While there are no stability restrictions on this time-step, time-scales attributed to certain processes in the model must be resolved, which in practice places limitations on the maximum allowable time-step. The biogeochemical time-step used is 1 hour, as compared to 13 minutes and 1 minute for the baroclinic and barotropic hydrodynamic time-steps respectively.

Computational grid used for hydrodynamic and biogeochemical models.

Figure 2: Computational grid used for hydrodynamic and biogeochemical models.

Primary productivity on the NWS is maximum in a distinct band lying below the mixed layer (the sub-surface chlorophyll maximum, SCM, Figure 3). Chlorophyll aconcentrations within the SCM are approximately 1 to 1.5 mg Chla m-3, and less than 0.5 mg Chla m-3 at the surface. There exists very little nitrogen in the surface mixed layer, but below the SCM nitrate increases steadily up to ~250 mg N m-3 at depths of around 300m. The model phytoplankton composition in the SCM is predominantly the larger size classes, e.g. diatoms. Nanophytoplankton dominate within the mixed layer with large turnover rates but low biomass, utilizing predominantly recycled nitrogen (ammonia). Large zooplankton graze on the larger phytoplankton classes, and consequently large zooplankton biomass is also maximum in the SCM. Microzooplankton graze on nanophytoplankton and have relatively larger biomass in the mixed layer than the large size class.

Chla distribution for April 1999 on the NWS.

Figure 3: Chla distribution for April 1999 on the NWS.

The SCM is maintained by a balance between phytoplankton nitrate uptake fed by fluxes into the SCM and nitrate export in particulate form. The SCM exists at a depth where light and nitrogen limited phytoplankton production is such that SCM phytoplankton uptake, organic N export as sinking particulates, balance vertical nitrate fluxes. A decrease in light availability or increase in nitrogen flux will move the SCM closer to the surface, and vice versa. At equilibrium the production of phytoplankton balances consumption by zooplankton. Any change in the grazing of zooplankton results in a change in production, regulated by light limitation via vertical displacement of the SCM in the water column. A schematic of these processes is presented as Figure 4.

Primary productivity schematic.

Figure 4: Primary productivity schematic.

The supply of nitrate into the SCM appears to be primarily due to vertical processes. The largest (temporal) mean flux of nitrate, or largest constant background flux, is due to vertical diffusive processes. Vertical advection is capable of locally increasing nitrate flux into the SCM. The horizontal advective fluxes of nitrate into the SCM (i.e. above ~80m depth) do not appear to contribute much to sustaining the SCM but are important for maintaining the deep pool of nitrate. A net supply of nitrate is delivered to the SCM through the north-eastern and offshore open boundaries, but this is largely lost through the south-western open boundary.

The SCM and surface chlorophyll signature undergo variability on time-scales of the neap-spring tidal cycle. Greater turbidity during the spring can decrease light availability, allowing the SCM to become shallower. Mixing to the surface under these conditions pushes the zone of maximum surface biomass further offshore. During the neap tide the SCM lies at constant depth and intersects the continental slope at depths 70 – 80m, acting as a barrier that strips any nitrate advected into the shallow coastal zone (Figure 5). As the SCM shallows during the spring tide a pathway is created for nitrate to enter the coastal zone (under favourable advective conditions). The bypassing of the SCM under these conditions can elevate nitrate in depths less than 40m (Figure 6).

Cross-shelf sections originating from Port Hedland for a neap tide.

Figure 5: Cross-shelf sections originating from Port Hedland for a neap tide.

Cross-shelf sections originating from Port Hedland for a spring tide.

Figure 6: Cross-shelf sections originating from Port Hedland for a spring tide.

The region where top and bottom boundary layers overlap creates a well mixed water column from the surface to sea floor; the largest vertical mixed zone possible. Where the SCM intersects this zone, mixing to the surface results in the largest surface Chlaconcentration. During spring tides the mixing in this zone is more vigorous and the surface Chla concentration is consequently greater  (Figure 7). The north-eastern side of the domain is associated with the least steep bottom slope, and is thus the region of greatest variability of boundary overlap zone, hence of cross-shelf surface phytoplankton distribution. Generally, wide flat shelves will be associated with this characteristic.

Cross-shelf sections originating from Port Hedland showing chlorophyll concentration.

Figure 7: Cross-shelf sections originating from Port Hedland showing chlorophyll concentration.

Seasonal variability of biomass is evident, where the SCM is more distinct with a surface signature closer inshore in the wet season, and the SCM more dispersed with surface signature offshore in the dry (Figure 8). These differences stem from changes in mixed layer depth, and a corresponding shift offshore of the boundary layer overlap zone. The mixed layer deepening originates from processes captured by the global hydrodynamic model into which the biogeochemical model is nested, and is attributed to atmospheric forcing, particularly zonal wind stress and the surface heat flux. The summer and autumn months are characterised by upwelling favourable winds, which in the absence of the pycnocline lowering effect of the Leeuwin Current acts to bring nutrients into waters of depth 40-60m and enhance productivity. Larger variability on time-scales of the semi-diurnal tide and neap-spring cycle are observed in summer and autumn, probably due to the combined effects of a shallower mixed layer and upwelling favourable conditions.

Surface Chla distributions for wet and dry seasons.

Figure 8: Surface Chla distributions for wet and dry seasons.

The passage of tropical cyclones in the model increases primary productivity at the surface and in the SCM by small amounts. Mixing increases in strength and penetrates deeper under the cyclone, and elevation rises as a result of the inverse barometer effect, with an associated baroclinic adjustment of the pycnocline downwards. This downward adjustment also moves the nutricline downward, beyond the zone of increased mixing which results in little new nitrate being mixed to the surface. After the cyclone has passed, however, vertical motions result in significant upward fluxes of nitrate which leads to increases in nitrate concentration and productivity above and within the SCM (Figure 9). These elevated levels oscillate at the near inertial period and Chla reverts to pre-cyclone levels within the SCM on time-scales of about one month. Nitrate concentrations in the SCM and surface Chla remain slightly elevated in comparison to pre-cyclone levels after one month.

Nitrate and Chla time series at 85 m during the passage of cyclone Bobby, February 1995. The red and blue curves are 25 km and 55 km respectively from the shelf.

Figure 9: Nitrate and Chla time series at 85 m during the passage of cyclone Bobby, February 1995. The red and blue curves are 25 km and 55 km respectively from the shelf.

Condie, S.A, Andrewartha, J, Mansbridge, J., Waring, J. (2006) Modelling circulation and connectivity on Australia’s North West Shelf, Technical Report No.