Use of a biophysical NPZ model to
investigate the effect of alongshore vs.
cross-shelf transport in the coastal Gulf of
Alaska on quality of habitat for migrating
juvenile salmon
or
What is the source of nutrients?

S. Hinckley (AFSC), A.J. Hermann (PMEL/JISAO),
E.L. Dobbins (PMEL/JISAO) & G. Blamey (UAF)

Questions:

  • Why is the coastal GOA such a productive area?
  • What is the origin of nutrients fueling this productivity?
  • What is the role of alongshore vs. cross-shelf transport of nutrients?
  • Is cross-shelf surface transport of nutrients due to downwelling (Ekman transport) important for productivity in the coastal Gulf?

Methods:

Model Overview:
  • 10 compartment NPZ model of the coastal area
  • Physical variables derived from the output of a hydrographic model
  • 3D grid representing an idealized, alongshore segment of the continental shelf in the CGOA, intended to simulate the GLOBEC NEP study area
  • Flat bottom at 100m, 20 vertical levels
  • Open horizontal boundary conditions, with information passed by radiation and nudging.
    • Offshore boundary = oceanic regime
    • Upstream/downstream boundary = shelf in 1D (vertical)
    • Inshore boundary = edge of ACC, also represented by shelf conditions
Map of the Gulf of Alaska with the Seward Line and idealized model domain. Flowchart of the 10 compartment NPZ model.

Offshore NPZ boundary conditions were taken from a 1D biological model (Kawamiya et al. 1995) for the deep North Pacific. Two simulations were tried:

  • Diffusivity high above the mixed layer and low below resulted in nutient depletion in the surface waters, similar to that observed onshore.
  • Diffusity from Kawamiya et al. produced HNLC as observed in oceanic areas.
Flowchart of the oceanic model.
Idealized Velocity:
  • Original velocity derived from idealized SCRUM case
  • Modified by 4 parameters controlling variations in alongshore and cross-shelf flow:
    • "Aoff" = alongshore yearly mean velocity
    • "Aamp" = alongshore seasonal range
    • "Coff" = cross-shelf yearly mean
    • "Camp" = cross-shelf seasonal range
  • Stronger downwelling in winter. Possible upwelling in summer, depending on combinations of parameters
  • Alongshore current is constant in space, but varies in time
A) View of surface current velocities from above. The top of the plot is landward. B) Timeseries of alongshore current magnitude. A Monte Carlo optimization and sensitivity analysis of the 4 velocity parameters was used to understand:
  1. what combination of cross-shelf and alongshore transport the biomasses of zooplankton are most sensitive to;
  2. what combination maximizes the amount of food available for migrating juvenile salmon; and
  3. what velocity field produces zooplankton biomass most like those seen at Seward Line in 1999.
A) Cross sectional view of cross-shelf and vertical velocity for a downwelling state. The right of the plot is landward. B) Timeseries of cross-shelf current magnitude. Positive velocity indicates downwelling, and negative velocity indicates upweliing.

Results of Sensitivity Analysis:

Sensitivity analysis indicates that cross-shelf velocities are more influential than alongshore velocities in affecting the biomass of coastal copepods, oceanic copepods, and euphausiids.

Maximum Zooplankton Productivity:

Maximum zooplankton productivity occurs in runs with the low cross-shelf mean downwelling, and high seasonal range, ie. runs with the longest period of upwelling (no surprise here). Supplies of oceanic nutrients supplied at the surface (by downwelling) are quickly (1-2 days) used up by biology, whereas nutrients transported at depth (by upwelling) penetrate farther into the interior.

Productivity was not obviously associated with any particular alongshore current.

Results are similar with either type (low or high diffusivity/HNLC) of oceanic boundary condition.

Annual mean cross-shelf velocity vs seasonal range cross-shelf velocity, for all the parameters sets used in the Monte Carlo. Symbols are color coded by zooplankton productivity (Red = high productivity, Black = low productivity). A) Alongshore seasonal range, B) alongshore annual mean, C) cross-shelf seasonal range and D) cross-shelf annual mean velocity parameters vs zooplankton productivity.

Comparison with Data:

The model run that agrees best with Seward Line data (1999) is characterized by a long period of upwelling between late April and the end of August. However, 1999 was not a year with strong upwelling at the Seward Line (Stabeno et al., in press).

Results are similar with either type (low or high diffusivity/HNLC) of oceanic boundary condition.

Depth averaged coastal copepod (C), oceanic copepod (NC), and euphausiid (E) biomass in time. Red lines and numbers are measured biomass. Black lines are simulations with diapause for the oceanic copepods, and the green lines are simulations without it.
  • Euphausiids are consistently overestimated by the model
  • Model missed the early bloom of coastal copepods, but otherwise they were in good agreement.
  • Oceanic copepods show good agreement in spring, but the model uderestimates them in late summer
    • When diapause removed, the model results are closer to data
    • Some species (e.g. Eucalanus bungii) diapause later
    • These spesies may have been present on the Seward Line later in the summer

Conclusions:

Is productivity on the shelf driven by onshore transport of nutrients at the surface (Ekman transport)?
  • Time scales for Ekman transport (1-2 km/day) are much slower than for biological response to an influx of nutrients (10-15 days between nutrient influx and bloom of copepods)
  • In the model, nutrients advected for the open ocean are used up quickly, and never reach the inner coastal areas
We may need to look for another explanation for the source of nutrients onto the shelf.
  • Influx through cross-shelf canyons?
  • Influxes from offshore via eddies or diversions of the Alaskan Stream?
  • Tides?

Funding support provided by GLOBEC