1-D results for the Nutrient-Phytoplankton-Zooplankton Model
In a typical run of the 1-D model, the large
phytoplankton bloom first, followed by a larger bloom of the small
phytoplankton fraction. The larger phytoplankton species grow
at lower temperatures, but are very sensitive to nutrient depletion,
whereas small phytoplankton species prefer warmer temperatures and are
less sensitive to lower nutrient concentrations.
Nutrients become
depleted in the surface waters following the large bloom of small
phytoplankton, as is seen in data from the Shelikof Strait region (Napp
et al., 1996).
Heterotrophic dinoflagellates show higher densities than
tintinnids due to their ability to eat both large and small
phytoplankton, however tintinnids begin to comprise a larger fraction of
the microzooplankton after the small phytoplankton begin to grow,
perhaps accounting for the second bloom of tintinnids in June, after the
May bloom of small phytoplankton.
Neocalanus, spp. increase until they
begin to be removed from the upper water column by diapause in early to
mid-summer. Euphausiid concentrations increase gradually over time
until the late summer when their density begins to decrease. No
vertical migration is included for Neocalanus, spp. or euphausiids.
|
|
Timeseries of vertically integrated variables: |
NPZ variables in Depth and Time:
The purpose of these experiments was to
look at how different elements of the biological model would respond to
mixing in different parts of the water column, with mixing above the
pycnocline representing wind mixing, and mixing below the pycnocline
representing tidal mixing. The mixing schemes tested included: high
mixing throughout the water column, ie. high wind and tidal mixing; low
mixing throughout the water column (low wind and tidal mixing); and high
mixing in the top of the water column and low mixing at depth (high wind
mixing and low tidal mixing).
In water columns which were well
mixed from surface to bottom, nutrients were depleted in the entire
water column by late June. Several blooms were seen, with large
phytoplankton blooming first, followed by repeated blooms of small
phytoplankton. Microzooplankton and small copepods responded to each
bloom in a cyclic manner, as nutrients were resupplied from deep waters,
causing a bloom, which was then brought under control by grazers.
When levels of mixing were low throughout the water column, nutrients
gradually became depleted in the upper layer, but remained abundant at
depth. Only small, early blooms were seen, and consequently densities
in the upper trophic levels remained low. With high wind mixing and low
tidal mixing, nutrients also became depleted in the upper layers due to
reduced resupply from deep waters. Several blooms were seen, including
a late deep bloom of large phytoplankton. Microzooplankton responded to
the early blooms of small phytoplankton, but not to the late deep bloom,
whereas small copepods responded to all blooms. Euphausiids declined in
the upper layers, but high concentrations continued at depth.
A sensitivity analysis was performed on the
1-D model to determine the most important parameters affecting model
behavior. A Monte Carlo analysis was done, using a Latin Hypercube
Sampling scheme (Hinckley, 1999).
Parameters Most Significantly Affecting Output Variables
PARAMETER |
DESCRIPTION |
NO. OUTPUT VARIABLES AFFECTED |
% |
slope1k |
Light absorption as a function of Phytoplankton density |
17 |
47.2 |
pmld2 |
MLD=f(date) parameter |
15 |
41.7 |
powD |
Doubling rate parameter |
13 |
36.1 |
mpredE |
Predation mortality on Euphausiids |
13 |
36.1 |
gammaMD |
Assimilation efficiency of Dinoflagellates |
12 |
33.3 |
mC |
Density-independent mortality of Coastal Copepods |
12 |
33.3 |
NcritPS |
Critical nutrient level below which mortality for Small Phytoplankton is nutrient dependent |
12 |
33.3 |
kE |
Euphausiid metabolized fraction |
11 |
30.6 |
alphaPS |
Initial slope of Small Phytoplankton production vs. Irradiance curve |
10 |
27.8 |
Q10MD |
Dinoflagellate Q10 for ingestion |
8 |
22.2 |
pmld4 |
MLD=f(date) parameter |
8 |
22.2 |
mpredC |
Predation mortality on Coastal Copepods |
8 |
22.2 |
The parameter relating light
absorption to phytoplankton density was the input variable which
affected the most output variables. Variables which significantly affected
more than 30% of the output variables examinated include the doubling rate
of small phytoplankton, predation mortality on euphausiids, assimilation
efficiency of heterotrophic dinoflagellates, the density independent
mortality of small copepods, the critical nutrient concentration below
which mortality for small phytoplankton is nutrient dependent, and the
metabolized fraction for euphausiids.
Output Variables Sensitive to the Most Parameters
OUTPUT VARIABLE |
DESCRIPTION |
NO. PARAMS SIGNIFICANTLY AFFECTING OUTPUT VAR. |
% |
NO may1 |
Nitrate concentration on May 1st |
22 |
31.4 |
NO jun1 |
Nitrate concentration on June 1st |
20 |
28.6 |
NO aug1 |
Nitrate concentration on August 1st |
20 |
28.6 |
NO apr15 |
Nitrate concentration on April 15th |
18 |
25.7 |
E jun1 |
Euphausiid concentration on June 1st |
16 |
22.9 |
E aug1 |
Euphausiid concentration on August 1st |
15 |
21.4 |
C apr15 |
Coastal Copepod concentration on April 15th |
15 |
21.4 |
C may1 |
Coastal Copepod concentration on May 1st |
13 |
18.6 |
E apr15 |
Euphausiid concentration on April 15th |
12 |
17.1 |
E may1 |
Euphausiid concentration on May 1st |
11 |
15.7 |
C jun1 |
Coastal Copepod concentration on Jun 1st |
10 |
14.3 |
The output variables sensitive to the most input parameters were
the concentrations of nitrates at different times during the modeled period.
Other sensitive variables include euphausiid densities in mid-summer,
early small copepod densities and early euphausiid densities.
|