Meteo climatologia Mediterranea Venezia 89 Febbraio 2001

1
Meteo climatologia MediterraneaVenezia 8-9
Febbraio 2001

• Modelli Numerici Oceanografia ENEA-Roma
• MOM Mediterraneano (globale, circolazione
  generale, bassa risoluzione)
• POM Stretto di Gibilterra (locale, alta
  risoluzione)
• POM Canale di Sicilia (regionale alta
  risoluzione biologia)
• MOM Breve presentazione di esperimenti di
  sensitività e di tecniche diagnostiche
  lagrangiane (circolazione generale)
• POM Qualche cenno ad esperimenti in corso
  (Gibilterra e Canale di Sicilia (processi locali)
• V. Artale, A.Bargagli, E. Napolitano, V.
  Rupolo, G. Sannino collaborazioni esterne IFA,
  CASPUR, KNMI, MISU, LPO .

2
MEDMOM
Surface temperature

• Primitive equation Cox Bryan model for Climatic
  studies
• Coarse resolution 0.25 x 0.25 x 19 levels
  (deformation Rossby radius from 5 to 30 Km)
• Rigid lid, surface restoring conditions (SSST and
  MODB SSS)
• Perpetual year forcing (different forcing, 1988)
• Horizontal and vertical mixing scheme (different
  diffusivity parameterization)
• Long integration experiments (equilibrium)
• Lagrangian trajectories and diagnostics

3
Vertical stratification

• Three main water masses (MAW, LIW and W-EDW)
• LIW depth ? sills depth
• DW formation maintains LIW at intermediate depth
  
• LIW at intermediate depth is a precondition for
  WDW formation
• Before going on higer resolution work on
• Surface forcings, tracer diffusivity

4
MEDMOM domain
Model resolution 0.25 x 0.25 x 19 (In the
shaded area tracers are restored to 3D
climatological fileds) Gibraltar 3 points T
and two points u
SST
5
Model equilibrium

• Kinetic energy (? 10 years)
• Tracer drift (? 30 years)
• Equilibrium between surface boundary conditions
  and water masses dynamics (? 100 -150
  years)
• ? ?
• LONG INTEGRATIONS
• (at least 300 years)

6
Sensitivity studies Long integrations (at least
300 years, model equilibrium)
Laplacian operator for diffusion Monthly
forcing restoring surface condition (perpetual
year) BBRS BBRS Parameterization SM
Smagorinsky paramaterization Khprofvertical
profile of diffusivity (constant in time and
horizontally) Keke tracer diffusivity
proportional to eddy kinetic energy
BBRS Parameterization for tracer diffusivity
Biharmonic operator for diffusion (perpetual year
for surface forcing) Daily daily wind and
surface temperature (SST), restoring Monthly
Monthly surface forcing, restoring Sflux1
Salinity surface fluxes from Monthly (E-P ? 80
cm/year) Sflux2 as Sflux1 with E-P ? 20
cm/year
7
Surface forcing (biharmonic operator)

• Monthly Monthly forcing, (SST, ECWM wind, MODB
  salinity)
• Daily Daily forcing for wind (ECWM and SST),
  monthly forcing for salinity

• Sflux1 Monthly forcing for T and wind, E-P
  surface flux from ten year averages (150-160
  year of Monthly ?80 cm/year)
• Sflux2 as Sflux1 with E-P1/4 of Sflux1 (?20
  cm/year)

8
No drifts Lttle diapycnal mixing 100th year
of integration
NWM deep cell
LIW formation
Numerical noise
Gibraltar
Sicily
Pedro de Vreijs KNMI
9
DAILY Experiments, upper branch of the THC
Particles are integrated using the mean averaged
velocity field. Lines are drawn joining 2 10-days
apart points. Colors represent depth. OFF Line
integration
10
DAILY Experiments 10 particles are integrated
using the mean averaged velocity field. Lines are
drawn joining 30 1-day apart points. Color
represent depth. The length of the trajectory
drawn represent 1 month.
11
DAILY EXP Monthly averaged stream function as a
function of density
February to July August to January
12
(No Transcript)
13
MONTHLY
DAILY
14
Effect of HF forcing Western cell more
active Lower branch of THC denser
15
DAILY MONTHLY EXP monthly averaged stream
function contribution to the THC of high
frequency forcing
With high frequency forcing western cell more
active in winter
February to July August to January
16
(No Transcript)
17
Tracer Diffusivity parameterization

• - BBRS K ? lt ?2gtltZ-1/2gt
• SM K ? h2 (N2S2)1/2
• Keke K ? lt ?2gtT_L
• - Khprof KK(z)

18
Temperature vs. time in the Tracer diffusivity
experiments
SM Keke
Khpr BBRS
The three space and time dependent diffusivity
parameterizations decrease tracer drift. Mean
temperature are almost constant after 30 years of
integration
19
Ten year averaged (150-160) salinity field
Khprof
BBRS
SM
Keke

20
Existence of LIW very important for deep water
formation
BBRS
Khpr
SM
Keke
BBRS Keke
Khpr SM
MODB
Salinity sections isolines of convection
depth
21
Levantine Intermediate water (colors represent
depth)
22
Khprof BBRS SM Keke
InBBRS the equilibrium between surface forcings
and water mass structure and dynamics is reached
only after 100-120 year of integration
23
Salinity relative Maximum in the NWM (left) and
its depth (right)
Khprof BBRS SM Keke
24
Salinity (left) Temperature (center) and density
profiles in the NWM before, during and after a
conection event
Khprof BBRS SM Keke
25
February diffusivity and
density field in the Lyon Gulf
80 m.
850 m.
BBRS SM Keke
26
TRACMASS U.E .project Lagrangian diagnostics to
study the North Atlantic and Mediterranean
circulation as they result from GCM

• Off line integration, general circulation
  framework
• Overview of the methods and some results
  fromMediterranean

27
OFF LINE INTEGRATION FROM VELOCITY MODEL
OUTPUTAriane code by Bruno Balnke (LPO)
http//www.ifremer.fr/lpo/blanke/ARIANE

• Individual trajectories The calculation on a
  C-grid of 3D streamlines, for periods over which
  the velocity is assumed to be constant, defines a
  way to derive trajectories within any model
  gridcell.
• Given a particle position, the minimum crossing
  time of the cell together with its final position
  on the edge of the gridcell are easily computed
  from the three relations dx/dtU
• Such a way to compute is fast (for a given
  gridcell, it only calculates exit position),
  accurate (it fully respects the 3D nondivergence)
  and flexible (can be adapted to isopycnic models
  and B grid level models)
• Time dependency Strategy for the time sampling
  of the Eulerian model velocity output. Compromise
  between memory storage problems and
  representation of time variability of the
  velocity field (?T?T_forcing is a good compromise
  for statistical analysis)

28
Quantitative diagnostics

• Huge number of particles (?105)
• Each particle conserves its infinitesimal mass
  with the associated transport the transport of a
  given water mass can be calculated from its own
  particles and their associated transport.
• Stream function from Lagrangian particles
• Following a set of particles and summing
  their algebric transport at each point, a 3D
  field corresponding to the flow of the water mass
  is obtained.

T1 T2 T3 T4 T5 T6
Note that considering only particles starting and
ending in two given sections, a stream function
representing the section to section mass
transport is obtained
-T1-T2-T3-T4 -T5-T6
0 T3T4 T3-T6
-T1-T2 -T4-T5-T6
-T5T6
-T-T5 T2-T5 T4T6
0 -T2
-T4-T3
29
Transport section to sectionQuantitative
analysisWater path Tracers analysis Time
statistics

• Lagrangian diagnostics in a OGCM

2
1
3

• Particles are released from section 1 with some
  criterium (e.g. ugt0, SltSo)
• Each particle is integrated till up it reaches 2,
  3 or recirculates in 1
• The transport function relative to the transport
  section to section is computed summing the
  contribution of each particle
• Note that the transport from 1 to 3 is different
  with or without 2

1
30
Quantitative analysis. Lagrangian stream
function transport Alborean Sea to
Levantine (COLOR AND ISOLINES REPRESENT TRANPORT
IN Sverdrup)
31
AW (Atlantic Water) From Alborean Sea to the
North Western Mediterranean Sea
AW (Atlantic Water) From Alborean Sea to the
Corsica Strait
32
LIW From Levantine to the Alborean Sea
LIW From Sicily directly to the Alborean Sea
33
LIW From Sicily to the Corsica Strait
LIW From Sicily to the North Western
Mediterranean Sea
34
ADW (Adriatic deep Water) From Otranto to Sicily
ADW From otranto to Levantine
35
Tracers analysis density on the Lagrangian
stream function Cadiz to Levantine (colors
represent density and isolines volume transport)

36
From where does Mediterranean outflow
come? Results from backward integration from
Cadiz. Top left particle position on a
meridional section in the Ionian for a backward
integration from Cadiz. Top center vertical
distribution of the mass flux top right
cumulated vertical distribution of mass flux
middle left horizontal mass flux
distribution middle center particles density
distrbution Bottom left cumulated horizontal
transport distribution bottom center cumulated
density distribution.
37
Time statistics Cadiz to Strait of Sicily Top
left Arrival time distribution Middle left
arrival time cumulative Middle right particles
concentration Bottom left Monthly initial
transport (particles are released continuosly
during te first year of integration) Bottom
right Monthly transport on the final section.
38
Decomposition based on the time statistics of
deep water pathways
Lagrangian stream function Aegean to Sicily (?
gt 29.1, color represent depth)
From the mean path (a) it has been extracted
the b) mode (peak) path c) path of the
slowest particles (?(t0ltTlt?) gt0.75) d) fast
particles (?(t0ltTlt?) lt0.05).
? is the cumulative function of the
arrival time distribution
39
1-cumulative!
Renewal basin time scales

• The histograms of the exit times in general does
  not display an univocal exponential behavior with
  one only slope For large time scales in many
  cases the curves were found to have tails longer
  than an exponential.
• Here we show the particles concentration for the
  whole Mediterranean basin as a function of time.
  The particle concentration has an exponential
  behavior for time scales shorter than 100
  years(??60 years). For larger times, a clear
  exponential behavior is observedwith a time scale
  of 250 years.
• The two different behaviors could be associated
  to the main branch of the THC (the intermediate
  return flow) and the deep layers renewal.

40
Sub-basin water masses transformation Particles
can be integrated backward in time from the
subbasin entrances toward the interior and
stopped when they exit the subasin. Partitioning
in different classes the particles in the final
and initial section properties, the lagrangian
approach allows to compute the water mass
transformation occurred in the basin.

• Class partitioning of Aegean exchange from
  lagrangian experiments
• Class I are surface freswaters (zlt250m and
  salinity lt38.85)
• Class II are water with salinity gt38.85 a proxy
  for the LIW in the area)
• Class III are waters fresher than 38.85 and
  deeper (i.e. most of the dense water outflow).

41
Arrival times (pollutant dispersal,
intermittency studies)

• Particles are released at one section (Gibraltar
  in the next figures)
• In each grid point is stores the arrival time for
  each particle
• Tave(x,y,z) mean arrival time
• T1 (x,y,z) arrival time of the first particle
• TR (x,y,z) Tave(x,y,z)/ T1(x,y,z)
• Tave different from ventilation time

42
Mean Lagrangian arrival time (Tave) at 50 m.
43
First Lagrangian arrival time (T1) at 50 m.
44
Mean Lagrangian arrival time (Tave) at 700 m.
45

• TRACMASS codes permit quantification of water
  mass pathways and transformation that other
  approach cannot achieve
• Compared with other methods, these codes are
• - fast and inexpensive (thosuands of
  particles can be integrated for hundred of years
  in few seconds)
• -flexible (time statistics, transformation
  of water masses )
• -adatable (isopycnal and level models)
• The main goal of the project is the study of
  the global general circulation but in limited
  basins like the Mediterranean Sea high frequency
  time variability can be introduced and processes
  study can be attempted.

46
Toward Gibraltar By G.M. Sannino Need of
high resolution model
47
MEDMOM particles at Gibraltar (BBRS) flux betwen
0.4 and 0.9 Sv in the different experiments.
colors represent salinity
48
MEDMOM results
Alborean Sea
Shallow convection in the NWM Deep convection
in the NWM
Gulf of Cadiz
T/S diagrams of the outflowing Med water in the
Alborean Sea and in the Gulf of Cadiz
No obvious relations, probably numerical noise
determine hydrological properties
49
MODB MED5 yearly averaged salinity field in the
Alborean Sea
50
Salinity Alborean Section
BBRS Keke
Khprof SM
51
Salinity Alborean section
Daily Sflux2
Monthly Sflux1
52
MEDMOM Salinity at 500 m in the Gulf of Cadiz -
BBRS
53
Gibraltar model (POM)

• High Resolution (300 m in the Strait)
• Parallel version
• Sigma levels (20-30)
• Realistic topography
• Advection scheme
• Pressure gradient
• Tides

• Ideal Experiments (Dum break)
• Realistic experiment (MODB and Levitus
  initialization in the Med and Atlantic)
• MEDMOM initialisation in the Mediterranean
• Role of the tides

54
(No Transcript)
55
(No Transcript)
56
(No Transcript)
57
(No Transcript)
58
(No Transcript)
59
POM GRID (301 x 53 x 20) In the Strait
resolution O(300 m)
60
MEDMOM high resolution Gibraltar Model
Stationary (10 years averaged) T/S fields from
different MEDMOM experiments used as initial
condition in the Med domain of the high
resolution Gibraltar (POM) model. Levitus
initial condition in the Atlantic MEDMOM initial
condition from BBRS KHPR DAILY SFLUX SFLUX2 (6-8
months of integration each) Sponge open boundary
condition. Results not yet analized
Gibraltar POM model
MEDMOM (0.25x0.25)
61
BBRS initialization in the Mediterranean Longitud
inal section, i23
62
Velocity
Froude Number
63
BBRS initialisation in the Mediterranean Deepest
sigma level, k24
64
BBRS Longitudinal section, i34
65
Altre di queste riunioni Siamo vicini alla massa
critica Basta vederci più spesso La prossima
volta organizziamo noi A Venezia ?
66

• PERSPECTIVES (Long terms)
• MEDMOM POM (Mediterranean Gibraltar)
• MOM3 Simple geometry to test and compare
  parameterizations (all current parameterizations
  are available)
• MEDMOMPOMMOM3/OPA Mediterranean Gibraltar
  North Atlantic

67
Future work

• Particles in POM (sigma coordinate) for TRACMASS
  diagnostics
• Couple tne model more actively
• Boundary condition
• Extend the Domain in the Atlantic to study the
  Mediterranean outflow

• Analyse the results
• Sensitivity of the Mediterranean outflow to
  Mediterranean climate
• Meddies ?
• Tides

68
MONTHLY EXP Monthly averaged stream function as
a function of density
February to July August to January
69
MEDMOM Salinity Temperature diagram of the
westward flowin the Alborean Sea
70
MEDMOM Salinity Temperature diagram of the
Mediterranean outflow in the Gulf of Cadiz
71
MOM Gibraltar outflow vs density of the lower
layer for the different experiments.
72
Horizontal Tracer Diffusivity parameterization

• - BBRS K ? lt ?2gtltZ-1/2gt
• SM K ? h2 (N2S2)1/2
• Keke K ? lt ?2gtT_L
• - Khprof KK(z)

73
1-cumulative !
Renewal sub basin time scales

• About one half of the computed histograms of
  arrival time do not show an exponential behavior,
  and the best fit was found to be a lorentzian
  function with 1.5lt\alpha lt3.5 and c varying from
  2 to 200.This fit is in general robust up to the
  95\ of the histogram.
• .An example of this behavior is shown here is
  relative to the time scales of the path of the
  intermediate and deep water from the Levantine
  basin to the Strait of Sicily. The values of the
  parameters of the fitted curves are 128.78 and
  2.68 for c and alpha , respectively.
• In general, functional forms different from the
  exponential one could be the results of
  correlations still presents in the particles
  motions as well as the convolution of multiple
  characteristic time scales of the renewal making
  difficult to extract useful information from
  statistical analysis without dynamical knowledge
  of the current paths.