ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks

Page created by Wanda Greene
 
CONTINUE READING
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
Boundary layers in the Bay of Bengal
                         Amit Tandon                 ASIRI
                       UMass Dartmouth

                                   Summer
                                 Monsoon winds

                       Arabian              Bay of
                        Sea                 Bengal

                                 Northern
                                  Indian
earth.nullschool.net              Ocean
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
2.5 billion people depend on the Monsoon for water, food, economy

                                                            Bay of Bengal

                                                            •   Highest per
                                                                capita
                                                                impact of
                                                                oceanic
                                                                influence
                                                            •   Largest
                                                                human
     World Population as a function of Longitude   Bay of       susceptibility
                                                   Bengal       to climate
                                                                change.
                       Atlantic
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
Motivation for ASIRI
•   Almost a third of the world’s population depends on the
    South Asian Monsoon for water
•   Climate models have difficulty in predicting the monsoons,
    particularly their sub-seasonal variability (active/break
    periods).
•   The ocean supplies heat and moisture for the monsoon.
    Role of oceans is important, but in coupled models SST is
    too cold.
•   Upper ocean structure, physical processes, and air-sea
    interaction affect the SST.
•   The Bay of Bengal is strongly affected by freshwater from
    the monsoons. How does this modify oceanic processes?
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
While forecast skill for precipitation is good for El-Nino
         region, it is very poor for the Asia-Western Pacific!

            El Nino region (10oS-5oN, 80oW-180oW)
                  WNP (5-30oN, 110-150oE)
            Asian-Pacific MNS (5-30oN, 70-150oE)
                                                                         Area averaged correlation coefficient
Correlation Coefficient between the observed and Multi-Model-Ensemble   between model and observations (skill)
                  hindcasted June-August precipitation

                                                     (Bin Wang, et al. 2005)

 ➡    Coupling of upper ocean with lower atmosphere is key to better
                prediction of tropical storms in this region.
 ➡      ONR motivation: Safety of ships at Sea, Active-Break periods
                   of the Monsoon impact wave forecasts
 ➡       India MoES motivation: Better characterization of air-sea
           interaction is important to improve Monsoon forecasts
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
Monsoons: A challenge for prediction
                      Active & break periods
                                                                      Goswami (2016)              1986
                                                                                                 drought
                                                                                                   year

          Why Bay of Bengal?
                                               Genesis and tracks of depressions
                                                                            Goswami (2016)
            Why North Bay?

Weakness of coupled forecast
                                                     Ocean: What is affecting the monsoon forecasts?
models                                               (1) Poorly known air-sea fluxes in the Bay.
•   Ocean is poorly represented                      (2) Unique ocean response, mixing and boundary
•   Sea surface too cool, stratification weak        layer physics in the region is not well understood and
                                                     not captured by GCMs.
•   Intra-seasonal variability not captured          (Collection of papers by IITM Pune 2016)
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
Monsoon rainfall

Bay of Bengal: An unusually fresh ocean

                     Sea surface salinity
                      Aquarius satellite
                           (Oct)

                                                Bay of Bengal

                                            Salinity: NRL Model
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
Bay of Bengal - anomalous ocean
• Winds & Circulation reverse seasonally                    India       Mahanadi
                                                                                   Ganges-Brahmaputra

• Extremely fresh
                                                                                                        Irrawaddy
                                                            Godavari

                                                             Krishna

                                              Arabian Sea              Bay of Bengal

• Highly stratified, poorly ventilated
• Warm surface - SST responds quickly
• Air-sea fluxes ! FW runoff
• Oxygen minimum zone
                                      Surface salinity from the Aquarius satellite, Fall 2013

                                                                Monsoon ISV - “active-break” cycle
                                               Salinity
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
What sort of program ?
                       Toward improving the monsoonal forecast
                                                                                           ASIRI
                                       Modeling and Prediction
                                            Coupled models
                                             Assimilation
                                                                                    MISOBOB

Data for model testing and              Improve representation of             Parameterize processes
       improvement                          Boundary layers                        and fluxes
                                             Air-sea fluxes
                                           Subgrid processes
             Observations                                                Process Studies
          Mooring, floats, gliders,                                      Multi-scale modeling
          drifters, ships, satellite                                     Process experiments

      Measurements - upper ocean                                    Understand physical processes,
      structure, fluxes, phenomena                                  explore parameter dependence
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
10 USA Institutions (about 20+ US PIs) and 8 Indian
    institutions (10+ Indian PIs) collaborating

              SAC,ISRO#Ahmedabad#

                                 INCOIS#and#TIFR#Hyderabad#

                                                   IIT#Bhubaneswar#
                  NIO#Goa#                     NIO#Vishakapatnam#

                                           NIOT#Chennai#and#IIT#Madras#

                             IISc#Bangalore#

                                           EBOB#
ASIRI Boundary layers in the Bay of Bengal - Amit Tandon UMass Dartmouth - KITP Online Talks
ASIRI – OMM: Observations Begin!                               ASIRI

• Nov-Dec 2013.
  – Survey and Process Experiment 1: R/V Revelle with US and Indian
    teams; Cruise on Nidhi with Indian and US scientists

• 2014:
  – Survey and Process Experiment 2: Pre-Monsoon observations with
    Indian scientists in international waters (June, R/V Revelle Chennai
    Port call)
  – July: Summer Workshop on Upper Ocean Physics in the Bay of
    Bengal; August 2014.
  – August: R/V Nidhi Monsoon cruise with Indian and US scientists
  – Nov.-Dec.: Flux mooring deployment from R/V Nidhi

• August-September 2015. Survey and Process Experiment 3: SW
  Monsoon Intensive Observational Period on US R/V Revelle and Indian
  ships.
ASIRI: OBJECTIVES
 What is the role of freshwater and seasonality of forcing on the upper ocean
 structure and circulation of the Bay of Bengal?

1. Boundary forcing: What is the variability in air-sea flux and boundary inputs? How are they
   affected by (and affect) freshwater distribution? Will accurate boundary fluxes improve forecasting
   of upper ocean structure and evolution?
2. How are physical processes, such as 1-d wind-driven mixing, Ekman transport, mesoscale
   circulation (nominally already included in models) playing out in unusual ways here?
3. What sets the stratification (submesoscale mixing and/or re-stratification, episodic events like
   strong cyclones, internal wave breaking, or others) ?
ASIRI: OBJECTIVES - 1

            Boundary forcing: What is the
            variability in air-sea flux and
            boundary inputs? How are they
            affected by (and affect)
            freshwater distribution? Will
            accurate boundary fluxes
            improve forecasting of upper
            ocean structure and evolution?
Need for benchmark measurements
          of air-sea fluxes

                            Lack of agreement in air-sea fluxes from
                                 different reanalysis products.
JJA                                                                                      AVISO
     Wind-                        ASIRI: OBJECTIVES - 2                                       SSH
     SST                                                                                      (MAM)
        How are physical processes, such as 1-d wind-driven mixing,
      Model – Buoytransport,
        Ekman        SST/SSS Temporal Change
                                 mesoscale   circulation (nominally
                                                               Modelalready
                                                                     – Satellite Comparison
      18 N 89.5 E : (OMM-Mooring)
        included in models) playing out in unusual ways here?
                                                              GHRSST                                  Model

     JJA                                                      SSH
     Wind-                                                                                    AVISO
     Winds
     SSS                                                                                      SSH (JJA)

e   Winter                        Model – Satellite Comparison

                GHRSST                   SST              Model       Surface Salinity
           Strong Coupling                                        SMAP
                                       Is the air-sea coupling linked to the weak/strong eddy         Model
                 Salinity gradients areactivity???
                 weak in model
                  ASIRI-OMM Science Meeting: 9-10 Jan, 2017

                                                                               Sarma et al.
Observational Focus- Mapping the full-scale of
   Physical Processes at work in the Bay of Bengal

  Instrumental &
  Observational
    Challenge
   Our approach
integrating moored
    time series,
    autonomous
      assets, &
 shipboard surveys
with state of the art
  instrumentation
ASIRI: OBJECTIVES - 3

                                T

    What sets the
Salinity along transect
    stratification
    (submesoscale mixing
    and/or re-stratification,
                                S
    episodic events like             13
    strong cyclones,
    internal wave
    breaking, or others) ?
                                N2

                                          Distance along track
ASIRI
 Science Results: What we have learned?
What improvements are possible for weather
         and ocean forecasting?
A. Large pools of sub-surface warm water in Barrier layers.

B. Microstructure and scalar mixing in the BoB.

C. What air-sea flux variations prediction models are missing - and why.

D. Interplay of mesoscale eddies, wind and the freshwater from rivers
  sets the multiple stratification layers.

E. Unprecedented near-surface process observations near fronts
ASIRI 2013-2017:
                                     Outcomes
                   with International Collaboration in the
                                      region
                             MISO-BOB 2017-2021
                      • Air-sea flux measurements + solar input
Long term observatory • Test bulk formulae, improve models
                      • Upper ocean vertical structure
                        • Relate to air-sea fluxes and stratification
Measure mixing rates    • Test and improve parameterizations
                        • Improve upper ocean structure in models
                       • Understand freshwater dispersal mechanisms
  Process modeling     • Understand mixing processes
                       • Estimate advection - test for seasonal variations
Central
         J O U Bay  Process
               RNAL O F P H Y S I C A LStudy:
                                       O C E A N O GNov,
                                                    R A P H Y 2013                              Aquarius
                                                                                                 VOLUME 48 salinity

                                                                                                        +
                                                                                                 CCAR SSHA

                                                                                                Mesoscale strain
                                                                                               (OSCAR) ~ 0.15 f

                                                                                              0-10 m averaged
                                                                                              potential density

                                                                                              UCTD resolution: 0.7-1.0 km
                                                                                              Duration of transects: 4-6 hrs
Ramachandran,               Tandon et al. March 2018 JPO
   FIG. 1. CCAR SSHA (color) for the week starting 18 Nov 2013 and salinity contours (black
                         21
 solid and dashed; g kg ) for the week starting 12 Nov 2013 from the level 3.0 Aquarius
Ramachandran, Tandon et al. March 2018 JPO

   ADCP                                       Lucas

                                                          Pinkel
                  Waterhouse
                                 Wirewalker
                    Analysis             Tandon
 Mackinnon
                Ramachandran                            Shroyer

                   UCTD
Weller

                                                      Met. data

         Nash        Farrar
                                       Mahadevan
• StandardMeasurements
           shipboard throughflow T&Sused !     in this study
• Underway CTD!
•  Underway CTD
• Thermistor+CTD     bow-chain!
                                                           •   Wirewalker
• Spar buoy: freely drifting, upward looking
• ADCPs
    Standard       shipboard
            at 20 and            through flow T & S
                       xx meters below
   surface, regularly spaced thermistors !
                                                           •   Shipboard met.
                                                                  data
•• Hull-mounted
    Permanent         sonars
                  ADCPs        on the ship (4)
                           + HDSS!

   • Side-pole
• Side-pole   mounted ADCP
                        ADCP + Pipestring ADCP
Surface Forcing
JOURNAL OF PHYSICAL OCEANOGRAPHY   VOLUME 4

         22
T and S sections for each
MARCH 2018                 R A M A of
                                   CHAthe
                                       NDRtransects
                                           AN ET AL.        485

                                                       21
Pycnostads

             Transect B
Large ⇣ ⇡ @v/@x

                  Transect B
⇣/f ⇡ 35

           Transect C
Vorticity is O(f) with more structure emerging as scales get finer.
                            Positively skewed
    JOURNAL OF PHYSICAL OCEANOGRAPHY                        VOLUME 48

                                                   FIG. 6. Vertical vorticity, z ’ ›y/›x, scaled with the Coriolis parameter, for transect C, s
                                                (a) 2 km, (b) 4 km, (c) 8 km, and (d) 16 km. The color bar on top is for (a) and (b) while that
                                                The solid lines are isopycnals obtained from the UCTD measurements, spaced 0.05 kg m2
                                                the same scale as the vorticity field in each panel. (e) Skewness of z ’ ›y/›x at 8-km scale, ob
                                                from the five transects at each depth.

Transect C
                                       as follows. Along each transect, we first locate the po-              frontal axis consistently, we
                                                                         Are these fronts in balance?
                                       sition where the magnitude of the lateral buoyancy                    fronts closer to the northe
                                       gradient at the surface is maximum (Johnston et al.                   The locations of the fronta
                                       2011). This procedure yields a location that alternates               A to E, xFR 5 (28, 28, 28
Geostrophy

  For a linear regression model:
                ig
       ez = a
       u           r⇢
              f ⇢0

  the smallest residual occurs when
       if ⇢0 huz r⇢⇤ i
    a=
         g hr⇢r⇢⇤ i

For thermal-wind balance,   a=1
                                      Rudnick and Luyten (1996)
Low-pass: 8 km

•   What types of submesoscale
    instabilities is this weakly
    geostrophic flow unstable to?
Cooling
                                    Winds   Frontogenesis

                           PV < 0

                            Subduction

Creation of zero
PV by SI, inertial
   instability

                                            Transect B
PV forcing and budget
pt for a portion of                                                1         ›b    ›b
                                                     EBF 5                 ty 2 tx    .                               (7)
 dix B), where the                                               r0 f        ›x    ›y
 ccur during
           492 a pe-                       JOURNAL OF PHYSICAL OCEANOGRAPHY                                                    VOLUME 48

   The gradients in          Equation (7) shows winds with a downfront component
  ignificant in this         result in a positive EBF and thus extract PV from the
   B).                       surface. In using (7), we neglect the term containing
r-zero or negative           ›b/›y (justification in appendix B). The EBF is also some-
 ral extent of these         times expressed as an Ekman heat flux (W m22; Fig. 2b)
   respectively. The         EHF 5 [r0Cp/(ag)]EBF, where a is the thermal expan-
  y low’’ or ‘‘vorti-        sion coefficient for seawater and Cp 5 3984 J kg21 8C21 is
 results principally         the specific heat of seawater at constant pressure (Thomas
 , resulting in low q        2005). In using (5), we replace z with f as we lack in-
1 f is the absolute          formation about the vorticity field at the surface. Evalu-
w PV are seen in             ating (5) with z at z 5 28 m reduces slightly the injection
   with PV , 0, lo-          of PV into the mixed layer during transects C and D. For
 sopycnals (Fig. 8),         the other transects, the rate of PV extraction is virtually
  hese contrast the          indistinguishable from that obtained by setting z 5 f in (5).
  sitive, q and weak            Integrating (4) vertically over the mixed layer,
  absence of strong             FIG. 9. (a) Ertel PV averaged over the mixed layer corresponding to an increment of
                                                 ð                                                 $
                             Dr 5 0.2 kg m23 from 0a depth of 1 m. (b) Net buoyancy flux at the ocean0 surface (W kg21; line
ve z.                        ›        1                 ›q    1                                    $         1
                                hqi 5 H
                                                           5 2 (J            DIAB fromFRIC
                             with circles), Ekman buoyancy flux (W kg21) obtained
                                                                             z      1J      z    )$$     2
                                                                                              15-min winds and the buoyancy
                                                                                                               ( . . . ),
                             ›t       H            2H   ›t    H                                              H
                             field at a depth of 1 m, smoothed to 8 km. A positive flux indicates cooling of the ocean.
                                                                                                     2H averaged over the
surface fluxes               (c) Contribution (s24) from the diabatic and frictional flux to the PV budget
                             mixed layer.                                                                             (8)
 PV q is governed
      The
        Nmchanges
 r J (Marshall    and      in   PV are
            is the bulk stratification
                            where      h!i H
                                            large,
                                       between
                                              denotesand
                                                1 m and
                                                  DIAB
                                                         the cannot
                                                             mixed allbe
                                                          averaging
                                                             FRIC
                                                                        five explained
                                                                       over                 byestimates
                                                                             transects. These
                                                                               H.  The           surface
                                                                                         parentheses         forcing.
                                                                                                         describe         These changes
                                                                                                                   averages
        layer depth. The combined effect of Jz         and Jz     is between the mixed layer base and 8 m. The magni-
        happen       over
        mostly to extractinPV,a   longer
                                 theexcept   time
                                        finalwhen
                                               term  scale     than
                                                    solaronheating,   the side
                                                              the right
                                                                     tudes survey
                                                                           range  of (advected
                                                                                  from (8)
                                                                                        O(10represent
                                                                                            210        into
                                                                                                ) s23 (for A, the   sampled
                                                                                                               B, and D) to   volume)
          through JzDIAB , injects PV into the mixed layer (Fig. 9c).         O(10 29 ) s 23 (for C and E). Thus, the lateral advective
            Integrating 2(JzDIAB 1 JzFRIC )/H over the duration of            flux, as approximated here, is also incapable of gener-
          transect A yields 27.57 3 10210 s23, the PV extracted               ating O(1028) s23 variations in hqiH over the duration
          from the mixed layer. Bulk of the PV removal (69%)                  of a transect.
Potential instabilities when f q < 0

                                       Thomas et al. (DSR, 2013)

                ✓              ◆                            ✓         2
                                                                          ◆
            1       (⇣ + f )                            1       |rh b|
c   = tan                               RiB   = tan
                       f                                         f 2N 2
Inertial
                           instability

Stable   Unstable
                             Symmetric
                              instability

                    Cushman-Roisin and Beckers
Vortically low

                        Baroclinically low

Where is the flow unstable?

Ertel PV at
                 Unstable to FSI
8 km scale
Normalized PV and balanced Ri = f2N2/M4
JOURNAL OF PHYSICAL OCEANOGRAPHY                          VOLUM

                             The presence of these regions
                              with flow conditions close to
                                neutral stability for FSI (Ri
                             slightly larger than unity) in the
                             vicinity of other regions where
                                  FSI might be active, is
                                suggestive of the neutrally
                             stable Ri resulting from earlier
                                        FSI events.
Transect A            @Mabs /@x ⇥ 103        (8 km scale)

                    SI

                                               SI + Inertial instability
Lateral extent of alignment in
agreement with length scale for
 most unstable hydrostatic SI
 mode
                                  Symmetric instability:
                               Alignment of contours of
Mabs = v + f x                Mabs and density (2 km scale)

@Mabs /@x < 0 =) Inertial instability

                 Contours are absolute momentum
         Color: Gradient of absolute momentum (8km fields)
Confluent flow

                    Isentropic PV
                      f + ⇣⇢ ⇢
                PV⇢ =
                         P ⇢

                          @v
                     ⇣⇢ ⇡
                          @x       ⇢

               We would like to see the
               imprint of lateral velocity
               gradients in the spatial
               variability of N2

Pollard and Regier (1992).
regions of
                          constant IPV
                            where the
                            thickness
                             between
                           isopycnals
                          and vorticity
                           on density
                             surface
                             mutually
                          compensate
                           each other
t   = 20.7
             t   = 20.8

             t   = 20.9

             t   = 21
Cooling
                                    Winds   Frontogenesis

                           PV < 0

                            Subduction

Creation of zero
PV by SI, inertial
   instability

                                            Transect B
Summary
 Fronts in the northern Bay are sharp, shallow and
 stratified.
 Observations at a salinity front forced with downfront winds
 and moderate mesoscale frontogenesis show a range of
 submesoscale processes at play
  •  Negative Ertel PV, O(f) vorticity
  •  Symmetric instability, inertial instability
Below the top 10 m, isopycnal thickness and isentropic
vorticity vary in lockstep
  • Pycnostads with O(1-10 km) lateral scales.
 Similar dynamical markers missing at a second process study
in winter forced by upfront winds and weak frontogenesis.

2015 data analysis shows intriguing results (ongoing)
Ertel PV     Was CBPS special?

    Upfront winds and weak strain at   17.3N
You can also read