ScicomP9 Bologna, 23

1
ScicomP9Bologna, 23 26 March 2004

• Wind driven circulation in the Gulf of Trieste a
  numerical study in stratified conditions

Querin, S., Crise, A. Istituto Nazionale di
Oceanografia e di Geofisica Sperimentale (OGS) Tr
ieste (Italy)
2
OGS - Dept. Of Oceanology (OGA)ECHO
group(Ecologic and Hydrodynamic Modeling)
Numerical studies and research activity

• applied hydrodynamics and turbulence modeling
• ecosystem modeling in open sea and coastal areas

• operational oceanography

Short term forecasts for coastal waters

• environmental protection
• coastal ecosystem monitoring and marine
  protected areas control

• water quality problems
• pollutants transport, oil spill, sewage
  discharge

3
Adricosm (ADRIatic sea integrated Coastal areaS
and river basin Management system pilot project)

• implementation of an integrated coastal zone
  management system in the Adriatic Sea
• predictive circulation module prediction of
  coastal current variability in Near Real Time
  (NRT)
• river basin and wastewater management module
  monitoring and modeling of a river basin and
• wastewater system in a test site
• shelf forecasting system for the Northern
  Adriatic Sea ? weekly
• forecasts of current dynamics
• key technical issues
• model (one-way) hierarchical nesting
• asynchronous coupling of coastal forecasting
  model and river basin management model
• prediction optimized by in situ and satellite
  data assimilation
• partnerships Italian, Slovenian, Croatian and
  French institutions
• project supported by the Italian Ministry for
  the Environment and
• Territory
• companion of MFSTEP (mediterranean sea) and
  MERSEA projects

4

• Our goals
• studying and modeling the dynamics of the Gulf
  of Trieste under realistic forcings for short and
  medium-range forecasts
• model validation comparing calculated fields
  with buoy measurements (MAMBO buoys (OGS), NIBs
  buoys) and satellite data
• embed the code into ADRICOSM structure (improved
  resolution 250 m)

5
The Gulf of Trieste (GoT)

• extension
• bathymetry
• prevalent winds
• effects on Adriatic and Mediterranean sea

6
Main features of the dynamics in the Gulf of
Trieste

• GoT is a ROFI (Region Of Fresh water Influence)
  area (Simpson, 1997)
• The dynamics are governed by
• wind stress
• buoyancy fluxes
• river input
• tides seiches
• remote control

7
The numerical model MITgcm Ocean General
Circulation Model (Marshall et al, 1997)

• MITgcm novel aspects
• designed to study both atmospheric and oceanic
  phenomena
• includes non-hydrostatic capability ? small and
  large scale processes
• adopts a finite volume technique ? treatment of
  irregular geometries using orthogonal curvilinear
  grids and shaved cells
• developed to perform efficiently on a wide
  variety of computational platform including MPI
  parallelizing directives

8
Non-hydrostatic primitive equation model
Adjustable momentum equation
9
Pressure integration
Hydrostatic pressure (vertical pressure gradient,
2D)
Pressure due to surface elevation (2D)
Non-hydrostatic pressure (3D)
10
Finite volume (FV) technique

• FV can accommodate any type of grid
• the grid defines only the control volume
  boundaries and need not be related to a
  coordinates system
• the method is conservative by construction
  surface integrals which represent convective and
  diffusive fluxes are the same for the CVs sharing
  the boundary

• Bottom topography discretization
• full cells
• piece-wise constant (partial) cells
• piece-wise linear (shaved) cells
• (not yet implemented)

full
partial
shaved
11
Adopted numerical schemes
Horizontal explicit quasi second order Adams
Bashford Vertical implicit backward
method/Crank-Nicholson
12
Vertical Mixing KPP profile for ocean interior
Reference Large, W.G., McWilliams, J.C. and
Doney, S.C., 1994 (Rev. Geophys., 32, 363 - 403)

• Processes
• resolved shear instability mixing
• unresolved shear instability due to internal
  wave field
• double diffusion

Parameterization of diffusivity and viscosity
parameterized in terms of gradient Richardson
number
constant (Peters et al., 1988 Large et al., 1994)
function of double diffusion density ratio
13
Vertical Mixing KPP profile for surface boundary
layer
Reference Large, W.G., McWilliams, J.C. and
Doney, S.C., 1994 (Rev. Geophys., 32, 363 - 403)
Surface boundary layer thickness minimum depth
at which bulk Richardson number exceeds the
critical Richardson number Ric0.3

• Processes
• wind driven mixing
• surface boundary fluxes
• convective instability

Parameterization of diffusivity and viscosity
G polinomial function hb boundary layer
thickness
14
Parameterization
Ah 4 40 m2/s Av constant (or KPP
profile) 10 m2/s Kh 4 40 m2/s Kv constant
(or KPP profile) 4 10 m2/s
15
Software Architecture

• Goals
• wide range of phenomena
• wide range of platforms
• high computational performance
• Architecture
• core numerical code
• pluggable packages
• support framework WRAPPER (WRappable
  Application Parallel Programming Environment
  Resource)

16
Parallel implementation
Communication primitives
Domain decomposition (2D)
Domain decomposition (3D)
17
Performance (IBM SP4 computer at CINECA)
Use of up to 16 processors with a domain
decomposition characterized by 4 x 4 spatial boxes
3,5 with 4 processors 5,8 with 8 processors 9,8
with 16 processors
Speed-up
18
Design of the numerical experiments for the Gulf
of Trieste

• f-plane, 250 m spatial resolution
• 30 rotation for domain optimization
• integration domain 88 x 128 x 25
• 25 (1 m thick) levels
• bathymetry
• linear kriging interpolation
• no filtering procedures
• small manual adjustments
• I.C.
• derived from MAMBO buoy profiles
• B.C.
• adiabatic
• lateral free-sleep
• quadratic bottom friction
• open/closed boundary conditions

19
Initial conditions MAMBO (Monitoraggio
AMBientale Operativo) buoy data set
Time/depth Hovmoller diagram of parameters
measured by MAMBO buoy
Depth dependent temperature and salinity profiles
20
Averaged wind regime in Trieste during the period
1991-2000 (after Stravisi)
Layout of numerical experiments
21
Wind driven upwelling in the GoT during a Bora
episode (28-29 July 2003)
ETA 1/6 LAM Model 29/07/03 06.00 GMT U10m
NOAA-12 AVHRR Sea Surface Temperature
22
Temperature after 3 days of bora
Constant bora 14 m/s wind stress curl 0
Vertical section
Meridionally sheared bora 14 - 7 m/s wind stress
curl gt 0
23
KPP profile - 8 days (4 Bora 4 no wind)
Constant bora
Sheared bora
24
Wind driven circulationCoexistence of mixing and
upwelling phenomena a proof using a Large-Eddy
Simulation
Buoy measurements
18 C
Model results
22 C
25
Results

• application of the model to a real case study
• good qualitative match with experimental
  observations
• successful implementation on a parallel
  environment

26
Ongoing work

• nesting with larger model (active BC)
• realistic wind forcing (2.5 km spatial
  resolution)
• Isonzo river input dynamics
• buoyancy fluxes

Real-time oceanographic forecasts