Mixed_layers - Center for Ocean

Report
HYCOM 2.1 Development at RSMAS
George Halliwell
• HYCOM Development
– Included several vertical mixing algorithms
– Refine the hybrid vertical coordinate adjustment algorithm
– Miscellaneous algorithms
• Ongoing HYCOM Evaluation and Analysis Efforts
– Evaluate the vertical mixing algorithms
– Evaluate other algorithms
– Provided 1.4-degree Atlantic model to C. Thacker for data
assimilation studies
– Climate studies
• Upper limb of Atlantic meridional overturning circulation
• Modes of Atlantic Ocean response to NAO atmospheric forcing
• Planned HYCOM Evaluation and Analysis Efforts
– West Florida shelf (with R. Weisberg and RSMAS HYCOM
group)
– Hurricane response (with D. Jacob and N. Shay)
– Coupled hurricane model (with S. Chen)
– Florida Current (with C. Mooers and RSMAS HYCOM group)
Vertical Mixing Choices in HYCOM 2.1
• Vertical Mixing Models (surface to bottom)
– KPP
– Mellor-Yamada level 2.5 turbulence closure
• Mixed Layer Models
–
–
–
–
Kraus-Turner A (hybrid)
Kraus-Turner B (hybrid, simplified, from Bleck global HYCOM)
Kraus-Turner C (MICOM mode, from MICOM 2.8)
Price-Weller-Pinkel dynamical instability model
• Interior Diapycnal Mixing Models (for use with the mixed
layer models)
– Explicit algorithm A (hybrid, MICOM-like)
– Explicit algorithm B (MICOM mode, from MICOM 2.8)
– Implicit algorithm (hybrid, KPP without surface b.l. model)
HYCOM Atlantic Ocean Simulations
• Domain
– Atlantic Ocean, 20S to 62N
– Resolution: 2.0 degrees horizontal, 22 layers vertical
• Forcing / Boundary Conditions / Initial Conditions
–
–
–
–
Surface forcing by COADS climatology
Relaxation to Levitus climatology at N/S boundaries
Relaxation to Levitus surface salinity
Initialized from p(latitude)
• Model Runs
– 30-year spinup
– One-year analysis run
– Winter (mid-February) and summer (mid-August) fields saved
• Eight Cases Summarized Here to Illustrate Different Vertical
Mixing Choices
s2
Latitude
Latitude
Winter and summer cross-sections along 29W. The central
sigma-0 value in the alternating white and gray bands of
the shading bar are the isopycnic reference values. Some
of the layer numbers are shown inside the shading bar. The
thick line is the diagnosed mixed layer base.
Winter mixed layer thickness
(m).
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Winter barotropic streamfunction (Sv)
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Ekman spirals
produced by the
KPP and MY
mixing models
in the Westerlies
(top) and the
trade Winds
(bottom).
Produced from
32-layer experiments
Winter SST (C) for the KPP
experiment (top left). Otherwise,
differences between the other
seven experiments and the KPP
experiment)
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Summer SST (C) for the KPP
experiment (top left). Otherwise,
differences between the other
seven experiments and the KPP
experiment)
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Winter SSH (m) for the KPP
experiment (top left). Otherwise,
differences between the other
seven experiments and the KPP
experiment)
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Winter sigma-0 for the KPP
experiment (top left) along
A 29W cross-section.
Otherwise, differences
between the other seven
experiments and the KPP
experiment)
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Winter SST differences (C)
between the eight experiments
and Levitus climatology.
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Summer SST differences (C)
between the eight experiments
and Levitus climatology.
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Winter sigma-0 differences
along 29W between the eight
experiments and Levitus
climatology.
Experiments:
KPP (k-profile parameterization)
MY (Mellor-Yamada 2.5)
PWP (Price-Weller-Pinkel)
KTAI (K-T A m.l., implicit diap.mix.)
KTAIP (same with penetrating sw. rad.)
KTAE (K-T A m.l., explicit diap. mix.)
KTB (K-T B m.l., implicit diap. mix.)
MIC (K-T C m.l., MICOM mode)
Figure: Preliminary simulation for the Gilbert
case using HYCOM including three non-slab
Schemes: (a) mixed layer depth, (b) mixed
layer temperature, (c) zonal current,
(d) meridional current, and e) surface fluxes
to the atmosphere. The differences in the
directly forced region are much larger
between these schemes compared to those
used in MICOM simulations. A surface flux
difference of up to 500 Wm2 is seen at the
point of closest approach of storm center.
Red, Green, Blue and Cyan lines represent
quantities simulated using KT, KPP, MY
and PWP schemes respectively.
Hurricanes and analysis regions for HYCOM simulations

similar documents