PRELIMINARY APPLICATIONS OF THE ROMS BIOGEOCHEMICAL MODEL TO THE NORTHERN ADRIATIC SEA

1
PRELIMINARY APPLICATIONS OF THE ROMS
BIOGEOCHEMICAL MODEL TO THE NORTHERN ADRIATIC
SEA I. Iermano,a A. Russo,a S. Carniel,b J.
Chiggiato,c A. Coluccelli,a R. P. Signell,d a
DiSMar, Polytechnic University of Marche, Ancona,
Italy, b National Research Council-ISMAR,
Venice, Itlay, cARPA-SIM , Bologna, Italy, d
U.S. Geological Survey, Woods Hole, MA, Usa
Abstract The Adriatic Sea (Figure 1) is a
continental basin of the Mediterranean Sea, and
its northern part is particularly shallow
(average depth about 35 m) and influenced by
large river runoff (Po is the main river, with an
average runoff of 1500 m3 s-1). This portion of
the Adriatic basin is affected by phenomena such
as eutrophication, mucilage, and bottom water
anoxia, which have negative impacts on the
environment, tourism and fisheries. Despite the
buoyancy gain due to the freshwater input, very
dense waters (up to 30 in sigma-theta) may form
in the northern Adriatic during winter due to
Bora (a strong, dry north-easterly wind) events.
These characteristics make the area challenging
to model. To address the environmental concerns
about the northern Adriatic basin, the
biogeochemical module of ROMS has been used for
realistic simulations with the general aim to
improve understanding of the Adriatic
physical-biogeochemical phenomenology. Some of
the current specific objectives are evaluation of
the model ability to provide short term forecast
of hypoxic events and evaluation of the influence
of nutrient input from different rivers on the
eutrophication and hypoxic/anoxic phenomena.
The impact of Bora wind on a hypoxic event in the
Northern Adriatic Sea During summer-autumn
season, the NA basin often presents a very stable
and stratified water column the bottom waters,
if interested by a relevant deposition of organic
matter produced in the upper layer, can be
affected by hypoxic and anoxic events (negatively
influencing environment, tourism and fisheries).
The Bora wind, which blows from the NE sector, is
particularly important in determining changes in
the vertical structure of this shallow water
system. This ROMS simulation allows a more
detailed description of this hypoxic event
evolution, in particular, the response of
dissolved oxygen to a relatively strong Bora wind
event (September 23 to 25, 2002).
Figure 8 shows a sequence of dissolved oxygen
bottom distribution ml/l in the NA
basin. (a)measured data used to initialized ROMS
model (16-19 September 2002). (b-c-d) 3 daily
modeled snapshots from September 21nd till 23nd
2002 (e)measured data interpolated from 22 to 25
September 2002.
Study area Northern Adriatic (NA) can be defined
as that region with relatively homogeneous
physical water properties, extending up to the
100m isobaths in the south (Artegiani et at.
1997). Even though the NA sea represents only a
portion of the total area of the Adriatic Sea, it
is the most productive region because it receives
nutrient rich freshwaters from several rivers and
in particular from the Po river. It is
particularly susceptible to eutrophication by
increases in the anthropogenic nutrient load to
the Po river watershed, leading to hypoxic and
anoxic conditions.
The sequence (Fig. 8) is an example of model/data
comparison simulated bottom distribution of
dissolved oxygen (ml l-1) compared to measured
data. The sequence shows that the hypoxic area
(the shallower western coastal one in particular)
received an important supply of oxygen due to the
vertical mixing generated by Bora, and is also
particularly evident the horizontal displacement
of hypoxic waters due to resulting intensified
bottom currents. The simulation shows that the
Bora blowing was able in few hours to
re-oxygenate fully the shallower hypoxic area,
and partially the deeper one moreover, Bora
enhanced the bottom water circulation, and this
in turn caused an horizontal displacement of the
hypoxic area. From the qualitative point of view,
dissolved oxygen simulated patterns resemble the
observed ones. However, it should be remembered
that observations have not been collected
sinoptically, since they started in the southern
area the morning of 22 September (before Bora
onset) and ended in the northern area on 25
September (during the Bora blowing).
ROMS model description The Regional Ocean
Modeling System (ROMS) is a 3-D,
primitive-equation, hydrostatic, finite
difference hydrodynamic model. In this
application, ROMS model is implemented on a
curvilinear-orthogonal grid with a horizontal
resolution of about 2-3 km in the northern
Adriatic, while 20 levels are used in the
vertical. Surface fluxes are derived from the
LAMI model, in particular heat fluxes are
interactively computed using bulk formulae. 26
rivers are included as mass and biogeochemical
sources, with daily runoff measured values where
available (Po river) and climatological values
elsewhere (biogeochemical values were deduced
from literature). The Fasham-type biogeochemical
model Fasham et al., 1990 is a representation
of nitrogen cycling processes in the water column
and organic matter remineralization at the
water-sediment interface that explicitly accounts
for sediment denitrification (see also Fennel et
al., 2006). The simulation is initialized in
September 2002 and runs for one year to represent
the seasonal cycle of biological properties in
northern Adriatic sea. Model initialization is
derived by MEDAR/MEDATLAS climatological data set
for the biological variables and by measured
field data collected during oceanographic cruises
in mid September 2002 for hydrological
parameters.
An unusual upwelling event An unusual upwelling
event along the western Adriatic coast happened
in summer 2003 between 11 and 23 July, documented
by SST and drifter data (Poulain et al. 2004).
This peculiar situation happened as a combination
of dominant sirocco winds (blowing from S-E in
the Adriatic Sea) and reduced river discharge
rates due to a prolonged dry season. The ROMS
biogeochemical model is able to reproduce the
upwelling event with a very high resolution.
Surface temperature and nutrient concentration
are correlated, indicating that the cold-upwelled
water is richer in nutrients.
Figure 9 shows two daily snapshots of (A)
surface ammonium concentration mmol/ m3 on 13th
July 2003 and on 15th July 2003 (B).
Figure 7 Time series plot of monthly Po river
runoff m3 /s for 2003 (red line) and from
climatological data set (blue line)
Figure 10 shows two daily snapshots of (A)
surface temperatureC distribution on 13th July
2003 and on 15th July 2003 (B).
Model results A typical seasonal biogeochemical
cycle in the water column is evident in the
northern Adriatic. In general, Dissolved
Inorganic Nitrogen and Chlorophyll follow typical
distributions (i.e. values decreasing from
coast to open sea and from surface to bottom).
For DIN and Chlorophyll, the higher concentration
at station B compared to the stations A and C is
likely due to proximity to the Po Delta (note the
different value scales between the three
stations). Nutrient profiles (Figure 4) are
generally characterized by high surface
concentrations, vertically decreasing down to a
depth of about 5-10m, a trend opposite to that of
the vertical distribution of salinity (Figure 6),
which clearly indicates the influence of the
river input on the nutrient concentrations. For
DIN, the autumn-winter season shows the highest
content in the whole water column, with a peak in
December. From winter to summer, concentrations
show a general decrease, with a minimum in
autumn before recovery toward winter values. This
behavior is mainly related to the different
nutrient assimilation by phytoplankton (maximum
in spring, minimum in autumn). Chlorophyll
(Figure 5) shows a well-defined seasonal cycle.
Highest phytoplanktonic biomass are found in
autumn, primarily in the western areas, which are
under direct freshwater influence. A secondary
peak is found in February at station A and C,
reaching the bottom increased light availability
and water column vertical homogeneity are clear
factors. A spring bloom is observed at station B
and A, clearly related to freshwater input. As
can be seen in figure 5, this limited
phytoplankton bloom is not supported by the same
increments in the nutrient concentrations (Figure
4) in the same period this is due to the already
mentioned fast nutrient assimilation in this
season and to the anomalously low Po river runoff
(figure 7 evidences the particularly dry season
of Po river runoff in 2003 compared to the
historical one). The described characteristics
are in good agreement with the known nutrient and
phytoplankton dynamics in the Adriatic Sea.
Conclusions First results of the biogeochemical
fluxes model implementation in the Adriatic Sea
appear very encouraging and are a powerful tool
to study interections between the physical and
ecosystem dynamic at high spatial and temporal
resolution. Based on this success, an
operational forecasting system at short time (2-3
days) for the evolution of hypoxic events in the
south of Po area is being developed.
Figure 3 time series plots of surface salinity
PSU, dissolved inorganic nitrogen mmol/ m3
and chlorophyll mg/ m3 from mid September 2002
to August 2003 in the three stations.
REFERENCES Cushman-Roisin,B.,M.Gacic,P.M.Poulain,a
nd A.Artegiani (Eds) (2001), Physical
Oceanography of the Adriatic Sea. Springer, New
York Degobbis,D. et al. (2000),Long-term
changes in the northern Adriatic ecosystem
related to anthropogenic eutrophication.Int.J.Envi
ronment and Pollution, Vol.13,Nos. 1-6 Fasham et
al,1990.A nitrogen-based model of plankton
dynamics in the oceanic mixed layer.J.Mar.Res.,48.
591-639 Fennel,K. Et al. (2006), Nutrient cycling
in the Middle Atlantic BightResults from a
three-dimensional model and implementation for
the North Atlantic nitrogen budget.Global
biogeochemical cycle,Vol.20,GB3007 Shchepetkin,
A. F., and J. C. McWilliams (2005), The Regional
Ocean Modeling System A split-explicit,
free-surface, topography following coordinates
ocean model, Ocean Modelling, 9, 347-404.
ACKNOWLEDGEMENTS We wish to thank Katia Fennel
(IMCS-Rutgers University, USA) and Emanuele Di
Lorenzo (Georgia Institute of Technology, USA)
for useful discussion and suggestions