Transcript Diapositivo 1

Development of the
MOHID consortium
Presentation of 4 PhD
thesis works
Menu
Thesis menu
•High resolution tridimensional hydrodynamical
model of the Portuguese coast (Guillaume
Riflet)
•Operational modelling in Estuaries (Angela
Canas)
•Bidirectional coupling of atmospheric and
oceanic regional circulation models (Ana Rosa
Trancoso)
•Climate change, fisheries pressure and coastal
upwelling ecosystems (Ricardo Lemos)
Guillaume
Thesis main goals



Development of a high resolution
hydrodynamical 3D model of the
Portuguese coast with MOHID.
Coupling the model to a large scale Atlantic
ocean model (ex: HYCOM, MERCATOR,
MOHID). Nested models
Use of remote-detection surface data
(buoys and satellite temperature surface
and altimetry data) to validate the model’s
results.
1. Developping the hydrodynamical 3D model
Guillaume
•
•
•
•
•
•
•
•
•
3 nested levels: Atlantic scale,
regional scale and local scale.
Implementation and testing of several
vertical discretization schemes
(sigma, cartesian, hybrid).
Definition of initial condition from
objective analysis of available
datasets.
Creation of coherent and stable
climatologic fields (ex:Levitus?).
Quantification of the pressure
gradient error in the oceanic
circulation.
Implementation and testing of several
vertical turbulence parameterizations
(such as GOTM).
Study for the pertinence and
implementation of a wave-induced
vertical mixing scheme.
Fiddling with the bathymetry from
ETOPO, fiddling with the tide
components, fiddling with the
boundary conditions.
Do it and make it work!
2. Downscaling physics from large scale models to
regional models
Guillaume
•
•
•
Development and/or
Implementation of
software code for format
integration to import data
delivered by large scale
models.
Testing nesting
techniques between the
large scale and local
hydrodynamic properties
(ex: adaptive open
boundaries, flow
relaxation scheme,
radiation schemes, free
surface elevation).
Identification of sources
for imposing the physical
properties in the open
boundary (ex: MW, Gulf
of Cádiz, Azores Current,
Wind).
3. Develop the Ocean-Atmosphere coupling with high
resolution atmospheric forcing
Guillaume
•
•
•
Test the Water-Air
interface module of the
model
Use of established
methodologies for
imposing atmospheric
fluxes in the coastal
model
Couple the model to
an ocean-atmosphere
modelling system fed
by high-resolution
observations
Guillaume
4. Use of experimental data to validate the model’s
results
•
•
Adjust model parameters and coefficients to improve
agreement between observed data and the model’s
results.
What is the best methodology to validate the model from
satellite-data? How reliable this is?
Guillaume
Work done so far
Mohid Hydrodynamic’s technical manual:
Physics (Hydrodynamic’s equations, Tracer
equations, Turbulent mixing)
Numerical Discretization
Test cases (1D vertical model, Freshwater cylinder,
Rossby solitary wave, Isolated Seamount ...)
Application cases: Ria de Vigo, Tagus operational
model, circulation off the portuguese coast, internal
waves in submarine canyons (Nazaré)
Ângela
P.h.D work: Operational Modelling in estuaries
Outline
• Objective: Improve the results of existing
prediction system for Tagus Estuary;
• Methodology:
– State of the art: Operational models, Data
assimilation;
– Study of existing system and new approaches
(boundaries, advection, turbulence): validation and
error quantification;
– Implementation of data assimilation module:
Kalman Filter techniques, Optimal Interpolation;
– Cost-benefit assessment of improvement
options facing operational framework.
Ângela
P.h.D work: Operational Modelling in estuaries
Work developed so far
Data
Assimilation
1D linear water level model: Kalman Filter


c
0
k
x
time
level
space
1D hydrodynamic model (shallow water equations): EnKF
New configurations
Tagus Estuary
Prediction System
Bathymetry:
Level 2 (Tagus Estuary)
Boundary conditions:
Level 1 (Portuguese Coast)
Level 0 (North Atlantic)
Validation
Water level measurements
(tide gauges) - 1972
Future... measurements for model validation and data assimilation
Rosa
Thesis objectives
•
Bidirectional coupling of the mesoscale
atmospherical model MM5 and of the
oceanographical model MOHID.
Identification of the oceanic-atmosphere
interaction’s effects:
•
–
–
•
•
•
Compare with the results of the independent
models
Quantifiy the required resolution degree
Influence of the coupling in the wind
waves and in the oceanic mixed layer.
Improve the Upwelling forecasts in the
portuguese coast.
Improve rainfall forecasts.
Rosa
MM5 Operacional
•
An operational system of meteorological forecasting for the Iberian
peninsula, with MM5: http://meteo.ist.utl.pt
x = 81 km
x = 27 km
x = 9 km
Rosa
WaveWatch III
•
•
•
•
Wind waves
Changes in the superficial
oceanic layer
Runs operationally with ARPS
model’s data (MeteoGalicia)
http://maretec.mohid.com/ww3
Rosa
Modelos Envolvidos
Future:
Now:
Global
Forecast
System
Atmosphere :
MM5
Radiation;
Heat and mass
flux
Mass, momentum and
energy flux
Heat and mass
flux;
SST
Shear Stress
and superficial
rugosity
Ocean:
ARPS
Applied
shear
stress
Superficial
rugosity
WaveWatchIII
Surface currents
Ricardo Lemos
Climate Change, Fisheries Pressure and Upwelling Ecosystems
Objectives:
•Identify the causes for the evolution of the Portuguese coastal upwelling system in the
20th century
•Assess the responses of primary productivity (seasonality, annual production, spatial
distribution, etc) to the observed changes in upwelling
•Estimate the impacts of changes in primary productivity on the food web
•Estimate, in a virtual, stable ocean, the impacts of fisheries on the structure of
Portuguese coastal ecosystems
•Combine the two impact pathways into a single model
H. sapiens
Top down
Fisheries Resources
Bottom up
Phytoplancton
Climate Change
Ricardo Lemos
Climate Change, Fisheries Pressure and Upwelling Ecosystems
The evolution of the upwelling regime off west Portugal, 1941-2000
Porto
Lisbon
Cape Carvoeiro
Evolution of annual mean ocean temperature
(ºC) in region C2, at different standard
depths. Linear trend estimates (x0.01ºC/yr)
are displayed on the right.
Evolution of the April-September v-wind
component at Porto, Cape Carvoeiro and
Lisbon (9 yr running-means and long-term
trend)
Ricardo Lemos
Work underway
Bayesian Dynamic Linear Models
Data
yi
Spatial Process
=
=
S(xi, b2)
+
+
Noise
ei
Time 1
Time 2
Time 3
S(., b1)
S(., b2)
S(., b3)