Collaborative Research on Sunlight

1

• Collaborative Research on Sunlight
• and the Arctic Atmosphere-Ice-Ocean System
• CRREL
• Don Perovich (lead PI) and John Weatherly
• UAF
• Hajo Eicken, Tom Weingartner, and Jeremy Harbeck
• UW
• Bonnie Light, Rebecca Woodgate, Ron Lindsay, Kay
  Runciman
• September 2005 August 2008

2
Motivating questions Evidence of operative
ice-albedo feedback processes? Are mechanisms
that produce polar amplification of global
warming in GCMs commensurate with
observations? Has the AIOS approached a
tipping point as suggested by model simulations
(e.g., Lindsay and Zhang, 2005)? What is the
role of the atmosphere (cloud cover,
precipitation) in modifying the response of the
cryosphere to variations in external forcing?
What is the role of the ocean (through import and
storage of solar heat in the mixed layer) in
transporting and sequestering solar energy?
3

• Goals
• Pan-Arctic maps of shortwave flux variables and
  parameters for 1982 present
• Surface shortwave flux (spatial and temporal
  variability)
• Snow deposition on sea ice
• Surface albedo
• Quantification of oceanic uptake and transport of
  solar heat through the Arctic system from
  oceanographic mooring data over the past 15 years
  (Woodgate et al., 2005) and satellite-derived sea
  surface temperature maps
• Synthesis of these data sets, GCM simulations and
  research findings

4
VARIABLE SOURCES TIME PERIOD
Radiation atmospheric variables Radiation atmospheric variables Radiation atmospheric variables
Incident solar irradiance NCEP reanalysis, satellite data (ERBE, CERES, ISCCP), GCM simulations 1979-2004
Surface albedo Modeling, dataset integration, AVHRR Pathfinder 1979-2004
Cloud cover Satellite data (ISCCP), GCM simulations 1983-2004
Surface air temperature POLES data set, satellite data, NCEP reanalysis 1970s-2004
5
State of the ice cover State of the ice cover State of the ice cover
Ice extent SMMR, SSM/I QuikSCAT 1979-2004 1999-2005
Ice concentration SMMR, SSM/I 1979-2004
Snow depth NCEP reanalysis, field expts., coastal weather stations, GPCP 1970s-2000s
Ice thickness Field expts., submarine sonar data, satellites 1970s-2000s
Ice melt rates Field expts., models 1970s-2000s
Pond fraction Field expts., pond hydrology model, "national asset satellites 1970s-2000s
FY / MY fractions Ice concentration Ice types SSM/I(for freezing season) SMMR, SSM/I (for freezing season) QuikSCAT (for freezing season) 1987-2004 1978-2003 1999-2005
Open water fraction RGPS (summer only) 1998-1999
Ice motion RGPS SSMI (Oct-May) AVHRR 1996-2000 1978-1997 1978-2003
Ice type - surface roughness Field expts. (incl. Russian monitoring flights), ICESat 1970s-1990s 2003-2004
Date of onset of melt freezeup Accurate timing of melt/freeze SMMR, SSM/I QuikScat (1999-2005) 1979-2004 1999-2005
Biogenic/sediment inclusions Field expts., remote sensing 1980s-2000s
6
Optical properties of snow, ice, water biogenic/sediment inclusions Optical properties of snow, ice, water biogenic/sediment inclusions Optical properties of snow, ice, water biogenic/sediment inclusions
Extinction coefficients Field expts., lab expt., modeling
Scattering coefficients Lab expts., modeling
Phase functions Lab expts., modeling
Absorption coefficients Lab expts.
State of the upper ocean State of the upper ocean State of the upper ocean
Influx of heat from Bering Sea Oceanogr. moorings, CTD cruises, satellite SSTs 1990-2004
Influx of heat from Atlantic European VEINS ASOF Programs 1997-2005
Influx of heat from rivers R-ArcticNet, temperature from field expts. 1970s-2000s
Surface hydrochemistry US-Russian Arctic Hydrochem. atlas, field expts. 1970s-2000s
Thermohaline properties of upper 50 m Environmental Working Group Atlas, Polar Science Center Hydrographic Climatology (psc.apl.washington.edu/Climatology.html) 1970s-2000s
7

• So far
• 1. Easy things
• Meetings
• Grid (25km x 25km EASE, 1982-2005)
• web page for status, updates
• http//synthesis.apl.washington.edu
• acquire data (ERA-40 downwelling Fr, SSMI ice
  concentrations)
• define a simplified subproblem assume the ice is
  opaque
• 2. Not-so easy things
• difficult decisions about difficult data (ice
  concentration)
• GRL manuscript (now in review)

8
The simplified subproblem Increasing solar
heating of the Arctic Ocean and adjacent seas,
1982-2005 Variability and ice-albedo feedback
(GRL, submitted) - not a budget, just a
term - estimate solar heating into the
ocean - Frw Fr (1-a) (1-C) -
assumption that ice is opaque ? minimum
estimate - problems with C and melt ponds ?
overestimate - a is constant, and appropriate
only for open water - every day, every grid
cell - only areas that have ice during the
period of observation - instantaneous and
cumulative
9
Map of mean total annual solar input averaged
over 1982 2005 (units are in MJ m2)
10
Maps showing the anomaly of solar heat input into
the ocean for 1984, 1995, 1996, and 2005 (units
are in MJ m2).
11
Maps showing the anomaly of solar heat input into
the ocean for 1984, 1995, 1996, and 2005 (units
are in MJ m2).
12
Map of the linear trend of annual solar heat
input to the ocean, with units of percent per
year.
13
Figure 4. Time series of solar heat input for a
50 x 50 km area centered on 75 N, 165 W.
14
Whats next? ICE COVER snow albedo during
melt Advective terms, specifically Bering
Sea Putting all the pieces together . . .
15
Interface with PP project PAR? plan use
ERA-40 broadband data and Perovich obs. (in prep)
Spatial coverage? Only ice covered areas
leaves out parts of Norweigan, Barents, E.
Greenland? Temporal coverage?
16

• The reviews for the sunlight paper are below. The
  short story is major revisions. As you will see
  the reviewers kind of liked it, but wanted more.
  The good news is that they didn't really
  challenge the SSMI ice concentrations. The bad
  news is that they do want more discussion
  regarding the ERA-40 incidents. Some of the
  suggestions are fairly straightforward, but there
  are a few that will be a little complicated. In
  particular, we need to1. Discuss uncertainties
  in ERA-40 incident.2. Discuss possible issues in
  transition between ERA-40 and ECMFW.3.
  Demonstrate temporal behavior of incident and ice
  concentration. Establish the cause of the
  increase (i.e. changes in concentration). To do
  this I suggest dropping the current Figure 2 and
  replacing it will a figure like the attached
  (only with more curves) that demonstrates that
  the incident fluctuates, but hasn't shown a
  strong trend, while concentration does should a
  trend.4. Further consideration and discussion of
  heat input through Bering Strait.5. Do we add
  1979-1981? Do we have those years?
• Reviewer 1 (Formal Review) This is an
  interesting and well written paper. The
  connection between the changing Arctic sea ice
  and solar heat input is an important aspect of
  the polar climate system. I must admit, though,
  that I am a little disappointed with their
  analysis. The authors say in the abstract "A
  synthetic approach was taken, combining
  satellite-derived ice concentrations, incident
  irradiances....and field observations of albedo
  over the Arctic Ocean and the adjacent seas". But
  then a constant value for the albedo was used
  and, more importantly, they did not distinguish
  between the variability and trends in incident
  solar radiation and the variability and trends in
  ice concentration. If incident solar radiation is
  not really changing or varying then all trends
  and patterns are perfect reflections of just
  changes in ice concentration (something that has
  been studies before). They say in the discussion
  that the "increase in heat input was primarily
  due to reductions in summer ice extent" p.9,
  line 180f but do not quantify this. All the
  figures show solar heat input, but it would be
  interesting to see how much of those trends and
  patterns results from trends inincident solar
  radiation (mainly from changes in cloud cover I
  would assume) and how much results from trends in
  sea ice concentration. There is probably also a
  correlation between incident solar radiation and
  ice concentration. Less sea ice may lead to
  increased cloud cover. The lack of seprating
  variations in solar radiation and in ice
  concentration is, in my opinion, a major weakness
  of this manuscript but it should be something
  that could easily be done with the data set they
  collected. A minor point is that I am wondering
  whether there is an offset between the ECMWF and
  ERA-40 solar radiation data. People may have
  looked at this already so a reference may be
  sufficient. Figure 2 shows heat input "from
  four selected years" p.7, line 137 ."Selected"
  always sounds suspicious. If, as they say, heat
  input was generally below average in the 80s and
  above average in the 2000s, I suggest to show the
  actual averaged 80s and 2000s data (if this makes
  sense). In summary, I suggest the paper to be
  accepted with some revisions. As said above, a
  separation of the effects of incident solar
  radiation and sea ice concentration should make
  the manuscript significantly more valuable.
• Reviewer 2 (Formal Review) General Overall
  I think that this is a potentially useful paper
  which adds to our understanding of Arctic sea ice
  loss. However, I think that it needs further work
  before publication can be recommended, primarily
  to clarify some of the techniques, assumptions
  and uncertainties, and to address what I feel are
  some mis-interpretations. Specific 1)
  Abstract and introduction The authors are
  glossing over the issue of sea ice thinning. That
  sea ice extent is decreasing is clear. However,
  the direct observational evidence of systematic
  thinning is not as convincing. Submarine sonar
  data certainly provide some evidence of thinning,
  but these data are spotty. The cited paper by
  Rothrock et al. 2003 was in part modeling
  based, and as I recall, provided some evidence to
  at least some thickening after the late 1990s. It
  certainly makes sense that reductions in extent
  have been accompanied by thinning, especially due
  to the observed loss of perennial ice, but the
  authors need to be very careful of making blanket
  statements. 2) Introduction, para 1 The ACIA
  citation should be replaced by, or attended by
  reference to the study of Zhang and Walsh 2006,
  J. Climate, who looked at sea ice trends over
  the period of observations from the IPCC-AR4
  models. The ACIA study made use of an earlier
  generation of models. 3) Introduction, para. 1,
  sentences 3 and 4 The discussion here is
  muddled. It is argued that a concurrent reduction
  in summer sea ice and increased hemispheric
  warming is the results of increased heating of
  open water. This makes no sense to me. Is the
  intended argument that general warming (from
  greenhouse gas loading) leads to ice loss, and
  that ice loss is then further enhanced by
  increased solar heating? The next sentence,
  arguing that ice loss then "drives further
  warming at the global scale", doesn't really make
  much sense unless the authors invoke the "Arctic
  amplification" of surface air temperature change,
  which is more of an indirect effect of increased
  solar heating in that with less ice at summer's
  end, and more heat in the upper ocean, there will
  be larger heat losses from the ocean to the lower
  atmosphere in autumn and winter. 4)
  Introduction, para. 1, near bottom Be careful
  with terminology. The Stroeve et al paper looked
  at ice extent (the region with at least 15 ice
  cover), not ice area. Extent and area are
  different things. Also, be more specific and
  state that the Stroeve et al. paper looked at the
  IPCC-AR4 models. 5) Introduction, para. 2 For
  the non specialist, briefly define "leads" and
  "polynyas". 6) Introduction, last para. It is
  stated that use is made of "measurements of the
  optical properties of water and ice during
  summer". Ho so? All I see in the Methods section
  that follows is the use of an assumed open ocean
  albedo of 0.07. 7) Methods, para. 1 Equation 1
  has been corrupted in the .pdf version I have.
  Double check that it is correct. While mention is
  made that only solar energy incident on the open
  ocean is being considered, and that no attempt is
  made to address penetration of radiation through
  ice the ice, it would might help the reader to
  have another sentence to make it crystal clear
  that nothing is being said here about changes in
  the albedo of the sea ice itself. 8) Methods,
  para. 2 There is discussion of the impacts of
  different sea ice data sets on solar heating.
  However, unless I've missed something, the ERA-40
  downwelling solar radiation data seem to be
  accepted without any discussion of possible
  biases. All reanalyses have problems with Arctic
  cloud cover, which in turn will strongly impact
  on the surface shortwave flux. I recall that John
  Walsh (who the authors know) has done some work
  on this. A climate "jump" might also be
  introduced when piecing together the ERA-40
  records (1982-2001) with fluxes from the ECMWF
  operation model. This needs to be addressed.
  Finally, why the odd starting year of 1982? The
  modern satellite data stream for data
  assimilation in ERA-40 began in 1979, and is also
  the start of the SMMR record. 9) Results, para.
  2, bottom Why is the year to year variability in
  the incident solar flux modest? Presumably this
  is because cloud cover shows little variability?
  This goes back to questions raised above
  regarding the quality of the ERA-40 (and also
  operational ECMWF) fields. 10) Discussion,
  para. 4 I'm all for the idea that lateral
  melting in summer is important in the ice-albedo
  feedback. On the other hand, I think the authors
  could be a little more clear in discussing
  seasonal aspects of the feedback mechanism. The
  way I might frame this is that even in a
  greenhouse-warmed world, there is little/no solar
  radiation over the Arctic ocean in autumn/winter.
  Hence, much of the heat that is picked up by the
  ocean in summer is just lost right back to the
  atmosphere (and to space). The autumn/winter
  ocean heat loss to the lower atmosphere is
  actually the climate model "fingerprint" of
  Arctic amplification of surface air temperature.
  However, the fact that it represents ocean heat
  LOSS could itself be viewed as a negative
  feedback as far as the sea ice cover is
  concerned. On the other hand, by warming the
  lower atmosphere, there is more downward longwave
  radiation to the surface as well, which works the
  other way. Which process wins? The point is that
  toperpetuate a strong feedback, you have to hang
  on to some to the extra heat gained in spring and
  summer through the long autumn and winter season.
  Maybe what one needs to do to hang onto some of
  the ocean heat is to actually grow some ice to
  shut down the heat loss? 11) Discussion, page
  5 The author's need to take a closer look at the
  evidence for increased heat transport through
  Bering Strait. My read of the Woodgate et al.
  paper is that the more obvious signal is high
  variability, rather than some general trend. As I
  recall, the heat flux seems to have increased
  between 2001-2004, but fluxes in 2001 were the
  lowest of the record. In turn, it was the fairly
  weak evidence of increased heat flow through
  Bering Strait that (at least in part) led Shimada
  et al. to argue instead for a link with
  redirection of Pacific Surface Water from the
  shelf slope into the Arctic Ocean. Mention should
  also be made of the potential role of "pulses" of
  warm water through into the Arctic Ocean from the
  Atlantic side see Polyakov et al., Geophys. Res.
  Lett, 2005, vol. 32 and the unresolved issue of
  how one gets this heat up to the surface through
  the strong halocline There is a recent review
  paper in "Science" Serreze et al., 2007, vol.
  315 that summarizesthese issues.