Docstoc

aero_for_GEWEX_1108

Document Sample
aero_for_GEWEX_1108 Powered By Docstoc
					                aerosol

an ancillary data-set in many applications


  How best to describe global aerosol optical
       properties for the last 30 years?
                       aerosols !
• aerosol (‘small atmos.particles’)
   – many sources
                                         highly variable
   – short lifetime
                                         in space and time !
   – diff. magnitudes in size
   – changing over time
          •   aerosol  clouds                      rapid
          •   aerosol  chemistry               atmospheric
                                                   ‘cycling’
          •   aerosol  biosphere
          •   aerosol  aerosol

               ocean
                                      industry       cities
 desert                 volcano                                forest
                     needs
• global consistent data for all aerosol properties
  – AOD, SSA, g        (amount, absorption, size)
• satellite data … only deliver in part
     • limited spatial coverage (non-polar, darker)
     • usually only AOD … with assumptions
• modeling … can deliver
     • but can it be trusted ? (even if constrained by
       satellite AOD pattern & component treatment)
• data assimilation ? !
     • MACC is not ready yet / hindcasts are too few
                climatology
• the next best thing
• monthly 1x1 climatology for mid-visible aerosol
  properties of AOD, SSA and Angstrom ( g)
  – start with ‘one year’ (year 2000 or ‘current’)
  – median of 15 component models (no extremes)
  – enhance with quality AERONET data (merging)
• extend spectrally
• use simulation to scale back/forward in time
• add altitude information (modeling or CALIPSO)
           start with modeling
                       all relevant aerosol optical properties
                       - aerosol optical depth (AOD)
                       - single scattering albedo
• advantages           - Angstrom  asymmetry factor
     • complete
     consistent
     (no data-gaps)     annual aerosol optical depth at 550nm

• issues
     • as good as
     assumptions
     … the ‘tuning’
                                                        modeling
     to observations
     merge in AERONET data
• merge
  – quality statistics
    of AERONET
• onto
  – consistent               AERONET
    background
    statistics by
    global modeling
     • NO satellite
     • NO grd in-situ        modeling
                    the merging
– combine point data on regular grid (1ox1o lon/lat)
         – extend spatial influence of identified sites
           with better regional representation
– spatially spread valid grid ratios and combine
         – decaying over +/- 180 (lon) and +/- 45 (lat)
– apply ratio field values (A) at site domains with
  distance decaying weights (B)  corr. factors
                  example: correction to model suggested AOD
A                        B                      A*B


    ratio field             site weight        correction factor
         merging results
         model    merged   AERONET
0.55mm


• AOD



• SSA



• Ang
          spectral extension
• use the Angstrom to statify AOD into
  – AOD, f (fine-mode … particles > 0.5mm radius)
  – AOD, c (coarse-mode … part. < 0.5mm radius)
• use SSA to identify dust size
  – when fine mode SSA falls below threshold
• coarse mode spectral dependence is defined
  – sea-salt or dust (4 different sizes)
• fine mode spectral dependence only for solar
  – g (from Angstrom) and SSA drop in near-IR
 spectral samples – annual avg maps
           AOD      SSA         g

• UV
  0.4mm

• VIS
  0.55mm

• n-IR
  1.0mm

• IR
  10mm
           extension in time
• does NOT change with time
  – coarse mode
  – pre-industrial fraction of fine mode
     • use info on what is ‘anthropo’ from modeling
• changes with time
  – stratospheric aerosol
  – anthropogenic fraction of the fine mode
     • back: ECHAM simulation with NIES emissions
     • forward: ECHAM runs with rcp scenarios
                   altitude
• current distributions
  – total aerosol
     • CALIPSO version 2
     • ECHAM
  – Separately for fine and coarse mode
     • ECHAM

• next … CALIPSO version 3
     • use (non-sphere) (=dust) for coarse
     • use total-(non-sphere) for fine
      total AOD (550nm) by layer
• ECHAM 5 / HAM (model)   CALIPSO vers.2 (obs)
looking for trends
aerosol climatology


               S. Kinne
            and colleagues
            MPI-Meteorology

              with support by
          AeroCom and AERONET
                    overview

• aerosol in (global) climate modeling
• (tropospheric) aerosol climatology
     •   concept
     •   column optical properties
     •   altitude distribution
     •   changes in time
     •   link to clouds
     •   some applications
     •   further improvements
                       aerosols !
• aerosol (‘small atmos.particles’)
   – many sources
                                         highly variable
   – short lifetime
                                         in space and time !
   – diff. magnitudes in size
   – changing over time
          •   aerosol  clouds                      rapid
          •   aerosol  chemistry               atmospheric
                                                   ‘cycling’
          •   aerosol  biosphere
          •   aerosol  aerosol

               ocean
                                      industry       cities
 desert                 volcano                                forest
                 aerosols ?
• the radiation budget is modulated by clouds
  – so, what is all the fuss about aerosol ?

     • aerosol has a role in cloud formations
  – many clouds exist because of avail. aerosols

     • aerosol has an anthropogenic component
  – the impact on climate is highly uncertain and
    at times speculative (for indirect effects)
     complex aerosol modules
• they do exist - even in Hamburg ECHAM/HAM
• they are useful … but often too expensive

 HAM calculations
 carry 27 tracers !




                                              size 
optically active
and relevant sizes
              simplifications ?
• simplified methods have been tried in the past
     •   ignore aerosol
     •   compensate with a radiation correction
     •   simplify by ignoring fine-mode absorption
     •   improve on the Tanre-scheme used in ECHAM
         IPCC 4AR simulations
          – a highly absorbing dust bubble over Africa
          – lack in seasonal feature (e.g. biomass)
• now better simplifications are possible
     • matured complex aerosol module output
     • improved obs. constrains by remote sensing
    new aero climatology goal
• provide a simplified aerosol representation in
  global models for rapid estimates on aerosol
  direct & indirect effects on the radiation budget

     • establish aerosol optical properties (as they
       required for atmospheric radiative transfer) at
       sufficient resolution … in order to resolve
       regional variability and seasonality
     • whenever possible, tie climatology data to
       (quality) observations
     • establish via CCN (cloud condensation nuclei)
       or IN (ice nuclei) links to cloud properties
              data sources

• the available observables

     • ground based in-situ samples
     • satellite guesses
     • sun-/sky- photometry

• the modeling world

     • data-constrained simulations
     • ensemble output
            ground samples ?
• issues
     • ground composition is not that of the column
     • samples are usually heated (less or no water)
  – monthly statistics example

                        size 




                                                       absorption 
        up quartile
          median
        low quartile
              satellite data ?
• issues
     • ancillary data needed
        – indirect reflection info requires information
          on aerosol absorption (usually assumed) and
          reflection contribution by other sources
          (such as surface or clouds) at high accurcay
     • limited information content
        – only AOD is retrieved and via AOD spectral
          dependency some information on size
     • limited coverage
        – land?, not with clouds, daytime, not polar, …

• positives
     • spatial distribution statistics
seasonal AOD by satellites
               combining MODIS, MISR, AVHRR data
ground-based sun-photometry
• advantages
  – direct AOD measurements
     • AOD and fine-mode AOD
  – other aerosol properties with sky-radiance data
     • size distribution (0.05mm < size bins < 15 mm)
     • single scatt. albedo (noise issue at low AOD)
     •


• issues
  – local … may fail to represent its region
  – daytime and cloud-free ‘bias’
            global modeling
                       all relevant aerosol optical properties
                       - aerosol optical depth (AOD)
                       - single scattering albedo
• advantages           - Angstrom  asymmetry factor
     • complete
     consistent
     (no data-gaps)     annual aerosol optical depth at 550nm

• issues
     • as good as
     assumptions
     … the ‘tuning’
                                                        modeling
     to observations
          climatology concept
• merge
  – quality statistics
    of AERONET
• onto
  – consistent                  AERONET
    background
    statistics by
    global modeling
     • NO satellite
     • NO grd in-situ           modeling
          aerosol climatology

• resolution and data content
• processing steps
• essential properties
• temporal variations
• link to clouds
• first applications
• further improvements
         climatology products
• global monthly 1x1 lat-lon maps
   – fine mode (radius < 0.5mm) properties
      • column AOD, SSA, g              (… for each l)
      • altitude distribution
      • anthropogenic AOD fraction      (1850 till 2100)
      • pre-industrial AOD fraction
   – coarse mode (radius > 0.5mm) properties
      • column AOD, SSA, g              (… for each l)
      • altitude distribution
   – cloud relevant properties
      • CCN concentration           (kappa approach)
      • IN concentration            (coarse mode dust)
     climatology details / steps
• define global monthly maps of mid-visible aerosol
  column prop (AOD, SSA & Ang)

• spread data spectrally to define AOD, SSA & g needed
  for radiative transfer simulations

• add altitude information using active remote sensing
  from space and/or modeling

• account for decadal (anthrop.) change using scaling
  data from aerosol component modeling

• establish a (needed) link to cloud properties by
  extracting associated CCN and IN
    column optical properties
• concept
  combine
       – for mid-visible aerosol properties     (0.55 mm)
       – for monthly (multi-annual) data
       – for a (1ox1o lon/lat) horizontal resolution

    • AeroCom model median maps
       – consistent and complete maps
       – median of a 15 model ensemble removes
         extreme behavior of individual models
    • quality data by sun-/sky-photometry networks
       – data of ~400 AERONET and ~10 Skynet sites
                    the merging
– combine point data on regular grid (1ox1o lon/lat)
         – extend spatial influence of identified sites
           with better regional representation
– spatially spread valid grid ratios and combine
         – decaying over +/- 180 (lon) and +/- 45 (lat)
– apply ratio field values (A) at site domains with
  distance decaying weights (B)  corr. factors
                  example: correction to model suggested AOD
A                        B                      A*B


    ratio field             site weight        correction factor
         merging results
         model    merged   AERONET
0.55mm


• AOD



• SSA



• Ang
merging




          note: low ssa at low
          AOD are unreliable
           merging results
           model       merged       AERONET
0.55mm

• AOD


• SSA


• Ang

                     fine-mode
other ..             anthropogenic
            kappa    fraction         dust size
         spectral extension (1)
• concept
  – split AOD contribution by small & large sizes
        – fine-mode size (radii < 0.5um)
        – coarse-mode size (radii > 0.5um)
     • apply the Angstrom for total aerosol … with
        – pre-defined fine-mode Angstrom at function
          of ‘low clouds’: Ang,f = [1.6 moist … 2.2 dry]
        – fixed coarse-mode Angstrom: Ang,c = 0.0

  – coarse mode (spectr. dep. properties) prescribed
     • apply the single-scatt. albedo for total aerosol
        – sea-salt (re=2.5mm / no vis. absorption: SSA = 1.0)
        – dust (re=1.5, 2.5, 4.5, 6.5, 10mm / SSA =.97, .95, .92, .89, .86
       spectral extension (2)
• concept
  – fine-mode spectral dependence by property
       – only adjustments for solar spectrum needed
    • AOD spectral dependence is defined by the
      fine-mode Angstrom parameter [1.6-2.2] range
    • SSA spectral dependence considers lower
      values in the near-IR (Rayleigh scattering limit)
    • asymmetry-factor and its spectral dependence
      is linked with fine mode Angstrom parameter
     (sharing common information on aerosol size)
       g = 0.72 - 0.14*Ang,f *sqrt[l(mm) -0.25] g,min=0.1
 spectral samples – annual avg maps
           AOD      SSA         g

• UV
  0.4mm

• VIS
  0.55mm

• n-IR
  1.0mm

• IR
  10mm
          altitude distribution
• concept … tied to the mid-visible AOD

  – define altitude distribution for total aerosol
     • use CALIPSO (vers.2) multi-annual statistics
       or use ECHAM5 / HAM simulation data
  – define altitude distribution separately for fine-
    mode and for coarse-mode aerosol
     • use ECHAM5 / HAM simulation data

        – pre-defined layer boundaries
          0 - 0.5 - 1.0 - 1.5 - 2.0 - 2.5 - 3.0 - 3.5 - 4.0 - 4.5
          - 5 - 6 - 7 - 8 - 9 - 10 - 11 - 12 - 13 (- 15 - 20) km
          (if surface height > 0.5km  less atmos. layers)
      total AOD (550nm) by layer
• ECHAM 5 / HAM (model)   CALIPSO vers.2 (obs)
        temporal change
• concept
  – assume that the coarse mode is all natural
  –


  – assume that the coarse mode does NOT
    change with time
  – assume that the pre-industrial fine-mode
    aerosol fraction does NOT change with time
  – the only aerosol to change over time is the
    anthropogenic fraction of the fine-mode (which
    is by definition zero before 1850)
     from the past – from 1860
• concept
   – use results ECHAM/HAM simulations with NIES
     historic emissions
       – aggregate simulated fine-mode monthly AOD
         patterns in 10 year steps for local monthly
         decadal trends
       – linearly interpolate between 10 year steps to
         get data for individual years
       – (at this stage) only changes to AOD are
         allowed … no changes to composition
           » could be an issues as there was a higher
             BC before WWII and a higher sulfate
             fraction after WWII
     into the future – until 2100
• concept
  – simulate with ECHAM/HAM AOD pattern
    changes by modifying suggested future
    emissions according to
     • RCP 2.6 IPCC% scenario
     • RCP 4.5 IPCC 5 scenario
     • RCP 8.5 IPCC 5 scenario
       – simulate changes for sulfate and carbon
         emissions stratified by selected regions
       – quantify the associated changes to fine-
         mode AOD pattern
       – sum impacts from all regions
              link to clouds
• aerosol provide CCN and IN to clouds

  – changes to aerosol concentration can modify
    cloud properties and the hydrological cycle

  – critical parameters are
     • aerosol concentration    from climatology
     • aerosol composition      from modeling
     • super-saturation
     • temperature
      define concentrations
• aerosol concentrations
  …are defined by the AOD vertical distribution
   and assumed log-normal size-distributions

     • coarse-mode log-normal properties for sea-
       salt or dust are prescribed (see above)
     • fine-mode log-normal width is fixed (std.dev 1.8)
       and the mode radius is linked to the Angstrom
       of the fine-mode (which depends on low cloud cover)
        – larger aerosol radii at moister conditions
        – smaller aerosol radii at drier conditions
              define CCN / IN
• all coarse mode particles are CCN … plus
• a fraction of fine-mode particles are CCN
  – for fine-mode concentrations:
     • a critical cut-off size is determined as function
        – super-saturation
        – kappa (‘humidification’) - maps are based on
          ECHAM /HAM simulations for fine-mode AOD
           » kappa weights: SS =1.0, SU =0.6, OC =0.1, DU/BC =0)

     • only sizes larger that the cut-off are
       considered as fine-mode CCN

• coarse mode dust particles are IN
     critical fine-mode radius
          1km     3km       8km

• SS
  0.1%

• SS
  0.2%

• SS
  0.4%

• SS
  1.0%
     critical fine-mode radius
         natural CCN   anthrop.CCN   (dust) IN

• SS
  0.1%

• SS
  0.2%

• SS
  0.4%

• SS
  1.0%
                applications

• aerosol direct forcing
  – in global numbers      (vs +2.8 W/m2 by greenhouse)
  – global direct forcing pattern         (highly uneven)
  – strong seasonality       (snow, clouds, aerosol type)
  – sensitivity to input        (define uncertainty range)
  – temporal change           (explore climate sensitivity)


• CCN application
  – first tests for impacts on clouds
 direct effects by the numbers
• global averages for direct radiative forcing
     • - 0.5 W/m2 anthropogenic at ToA            all-sky
           » + 0.3 W/m2 for anthropogenic BC
        – compare to +2.8 W/m2 by greenhouse gases

     • - 1.1 W/m2 anthropogenic at ToA            clear-sky
     • - 4.7 W/m2 solar at ToA                    clear-sky
           » as seen by satellite sensors
     • - 8.6 W/m2 solar at surface                clear-sky
           » as seen by ground flux radiometers
direct forcing/effects                 global patterns

            solar + IR       solar             anthrop. solar
• clear
  ToA
                         ‘seen’ by satellite
• clear
  surface                ‘seen’ by ground
                          flux-radiometer

• all-sky
  ToA                                            ‘climate’ impact


• all-sky
  surface

                             cooling           warming 
    direct ToA forcing – seasonality
•




                   cooling    warming 
direct forcing - sensitivity tests
• ToA all-sky forcing
  uncertainty           no              lower          AOD
                        AERONET         low clouds     uncertainty
  – examine impact
    by changing         AOD             higher
                                        low clouds
                                                       SSA
                        by satellite                   uncertainty
    individual prop.
    to aerosol or       fine-mode =     higher         g
    environment         anthropogen     surf. albedo   uncertainty


                        higher max.
     • +/- 0.3 W/m2     for fine mode   higher         25% more
                                        surf. albedo   absorption
                        absorption
            anthrop. direct forcing in time
    • changing pattern
       – 1940: -.16 W/m2
       – 1945: -.17 W/m2
       – 1950: -.19 W/m2
       – 1955: -.21 W/m2
       – 1960: -.24 W/m2
strongest
increase




       – 1965: -.27 W/m2
       – 1970: -.31 W/m2
       – 1975: -.35 W/m2
       – 1980: -.38 W/m2
       – 1985: -.41 W/m2
       – 1990: -.43 W/m2
       – 1995: -.44 W/m2
       CCN – climatology in use
• CCN /ccm3         at 1km 0.1% super-saturation



          climatology              ECHAM5.5/HAM



• liquid water path comp.
       • stronger aerosol
         indirect effect with
         the climatology
by U.Lohmann ETHZ
        future improvements
• update AOD vertical distribution with CALIPSO
  version 3 data (multi-annual averages)
• include compositional change over time for
  anthropogenic aerosol (BC vs sulfate content)
• allow responses to relative humidity changes in
  the boundary layer
• explore smarter ways for data merging of
  trusted point data onto background data
• allow natural aerosol (sea-salt and dust) to vary
  according to met-data and/or to soil-data
• extras
 data access                  ftp ftp-projects.zmaw.de

• cd aerocom/climatology/2010/yr2000 year 2000
     • look at ‘README_nc’           properties

  – ss-properties for the 14 solar-bands of ECHAM6
     • g30_sol.nc         total aerosol              (yr 2000)
        – g30_coa.nc      coarse mode aerosol
        – g30_pre.nc      fine-mode pre-industrial   (yr 1850)
        – g30_ant.nc      fine-mode anthropogenic    (yr 2000)

  – ss-properties for the 16 IR-bands of ECHAM6
     • g30_fir.nc   total (=coarse mode) aerosol

  – CCN and IN fields
     • g_CCN_km20.nc for supersat. of 0.1, 0.2, 0.4 and 1%
        – anthrop. CCN is changing along with anthrop. AOD
 data access                    ftp ftp-projects.zmaw.de

                                   anthropogenic change
• cd aerocom/climatology/2010/ant_historic
  – based on decadal change by ECHAM/HAM simulations
    for fine mode aerosol with NIES historic emissions
     • aeropt_kinne_550nm_fin_anthropAOD_yyyy.nc
            » simulations by Silvia Kloster, prep. by Declan O’Donnell

• cd aerocom/climatology/2010/ant_future
  – based on sulfate and carbon emission related regional
    changes to fine-mode AOD patterns with ECHAM/HAM
    for IPCC RCP 2.6, RCP 4.5 and RCP 8.5 scenarios till 2100
     • aeropt_kinne_550nm_fin_anthropAOD_rcp??_yyyy.nc
            » simulations by Kai Zhang, concepts by Hauke and Bjorn
 data access                ftp ftp-projects.zmaw.de

                                                  altitude
• cd aerocom/climatology/2010/altitude
  – based on CALIPSO version2 multi-annual data
  – fine/coarse mode altitude differences via ECHAM/HAM
     • calipso_fc.nc
           » CALIPSO data via Dave Winker (NASA_LaRC)
           » ECHAM/HAM altitude data from Philip Stier

• cd aerocom/climatology/2010/ccn_historic
  – applying anthropogenic change to CCN concentrations
     • CCN_1km_yyyy.nc                at 1km above ground
     • CCN_3km_yyyy.nc                at 3km altitude
     • CCN_8km_yyyy.nc                at 8km altitude
                 extra plots

• assumed environmental data
• mid-visible aerosol ss-properties
• current anthropogenic AOD … by month
• solar spectral ant aerosol properties variations
• vertical spread of fine mode AOD by ECHAM
• vert. spread of coarse mode AOD by ECHAM
        environmental data
albedo (MODIS + ice, snow)           clouds (ISCCP)


                     snow fraction




                      ice fraction




                     temperature
AOD / SSA / ASY   at 550nm
looking for trends

				
DOCUMENT INFO
Shared By:
Categories:
Tags:
Stats:
views:0
posted:4/21/2013
language:Unknown
pages:75