a synthesis of arctic climate changepolarmet.osu.edu/asr/asr_bromwich_hines_200804.pdf · division...

Division of Global and Environmental Change Brown Bag April 4, 2008

Polar Meteorology Group, Byrd Polar Research Center, The Ohio State University, Columbus, Ohio

A Synthesis of Arctic Climate Change*

*Supported by NSF, NOAA, NASA, and DOE

David H. Bromwich1,2, and Keith M. Hines1

1Polar Meteorology GroupByrd Polar Research CenterThe Ohio State University

2Atmospheric Sciences ProgramDepartment of GeographyThe Ohio State University



OutlineOutlinePrimer on Arctic Climate ChangeGlobal Reanalysis Performance in the ArcticOutline of Arctic Regional ReanalysisPolar WRFAtmospheric Data AssimilationLand Surface Data Assimilation



Figure 3: September ice extent from 1979 to 2007 shows an obvious decline. The September rate of sea ice decline since 1979 is now approximately 10 percent per decade, or 72,000 square kilometers (28,000 square miles) per year.

Arctic Sea Ice Shatters All Previous Record Lows

press release from the National Snow and Ice Data Center (NSIDC)1 October 2007

(http://www.nsidc.org/news/press/2007_seaiceminimum/20071001_pressrelease.html)



Figure 1. Actual observations of September Arctic sea ice (red) show a more severe decline than any of the eighteen computer models (average, dashed line) that the 2007 IPCC reports reference.

NSIDC Press Release:Models UnderestimateLoss of Arctic Sea Ice

30 April 2007

Stroeve, J., M. M. Holland, W. Meier, T. Scambos, and M. Serreze, 2007:Arctic sea ice decline:Faster than forecast.Geophys. Res. Lett., 34,doi:10.1029/2007GL029703



Greenland Melt Extent 2005Konrad Steffen and Russell Huff

Cooperative Institute for Research in Environmental Sciences (CIRES)University of Colorado at Boulder, CO 80309-0216

http://cires.colorado.edu/science/groups/steffen/greenland/melt2005/

Greenland melt extent for 1992 and 2005

Total melt extent area that experiences 1+ melt day April – September. 2005 melt extent exceeds the previous 2002 record. (Steffen et al. 2004; Hanna et al. 2005)



Permafrost distribution in the Arctic. Pink is continuous, blue is discontinuous, green is sporadic. [Romanovsky et. al., 2002, Fig. 1]

Mean annual ground temperatures in Yakust, Siberia (62.1°N, 129.8°E), from 1833-2003. [From V. Romanovsky]

Mean annual ground temperatures at Fairbanks, Alaska, for 1930-2003. The temperature 1 meter below the surface has risen very close to 0°C. [From V. Romanovsky]

Permafrost Melting in the Arctic? (NOAA Arctic Change)

http://www.arctic.noaa.gov/detect/land-permafrost.shtml



What is Atmospheric Reanalysis?What is Atmospheric Reanalysis?• A detailed, quantitative history of the weather and climate

achieved with a fixed analysis scheme.• Combines in-situ and remote observations as input, and integrates

with a numerical weather/climate prediction model for a physically realistic depiction of the atmosphere. Therefore, reanalysis is a synthesis of our efforts to observe and understand the Earth.

• Outputs fields for quantities that can be readily observed (such as pressure and temperature), quantities that are problematic to measure (such as precipitation), and quantities that cannot be directly measured (such as surface fluxes).

• While the fields within a given reanalysis are not sensitive to the assimilation methodology (as it does not change), the fields aresensitive to the availability and quality of observations (the better the observations, the better the reanalysis).

• The first 2 generations of atmospheric reanalyses are dominated by global projects (NCEP/NCAR, ERA-15, NCEP-2, ERA-40, JRA-25, …).

• Reanalyses can also be performed over a limited area (NARR).



FIGURE 1. The vertical structure of the Arctic warming during the 1980s and 1990s, based on the ERA-40 reanalysis. Averaged temperature trends around latitude circles for 1979–2001 plotted versus latitude and height for the four seasons.

ERA-40 NCEP JRA-25

Nature, 451, 53-56 (3 January 2008) doi:10.1038/nature06502

Vertical structure of recent Arctic warmingRune G. Graversen, Thorsten Mauritsen, Michael

Tjernström, Erland Källén & Gunilla Svensson



Arctic System Reanalysis MotivationArctic System Reanalysis Motivation1. Rapid climate change is observed in the Arctic, as

illustrated by the all-time minimum of summer sea ice extent in September 2007. Comprehensive pictures of Arctic climate and Arctic physical processes are needed.

2. Previous global atmospheric reanalyses (NCEP/NCAR, ERA-15, NCEP-2, ERA-40, JRA-25) encounter many problems at high latitudes. The ASR would use the best available depiction of Arctic meteorological and climatological processes with improved temporal resolution and much higher spatial resolution.

3. The ASR would provide fields for which direct observation are sparse or problematic (precipitation, radiation, clouds, ...) at higher resolution than existing reanalyses.

4. The ASR would provide a convenient synthesis of field programs to observe the Arctic (SHEBA, LAII/ATLAS, ARM, FIRE, M-PACE, ...).



ASR OutlineA physically-consistent integration of Arctic data

Participants:Ohio State University - Byrd Polar Research Center (BPRC)

- and Ohio Supercomputer Center (OSC)National Center Atmospheric Research (NCAR)University of ColoradoUniversity of IllinoisUniversity of Alaska Fairbanks

High resolution in space (~10 km) and time (3 hours)

Begin with years 2000-2010 (EOS coverage)

Supported by NSF as an IPY project


Polar Meteorology Group, Byrd Polar Research Center, The Ohio State University, Columbus, OhioASR Duty RosterPolar WRF Model Development and Optimized Sea Ice Representation

OSU BPRC Polar Meteorology Group, PI

Mesoscale Atmospheric Data Assimilation

NCAR MMM (D. Barker + Y.-H. Kuo)

Land Surface Treatment and Data Assimilation

NCAR (F. Chen, developer of the Noah LSM)

University of Colorado (M. Serreze)

Data Ingest, Data Monitoring, and Quality Control

University of Illinois (J. Walsh) and U. Colorado

Computing

Ohio Supercomputer Center

Arctic Regions Supercomputer Center?

Reanalysis Distribution to the Community

U. Illinois/NOAA CDC?/NCAR?



ASR High Resolution Domain

Outer Grid: 30 km resolution

Inner Grid: 10 km resolution

Vertical Grid: 71 levels

Inner Grid includes Arctic river basins



Weather Research and Forecasting Model (WRF)

Atmospheric Limited Area NumericalModel

Sub-grid scale cloudsClouds from resolvable storms

Terrestrial infrared and solar shortwave

Soil, water, vegetation, snow and ice

Just above the surface

Before Polar WRF: Polar MM51. Polar work began with Penn State/NCAR MM4

2. MM5 was also adapted for Arctic and Antarctic applications(1) Real-time forecasting/Operational uses(2) Synoptic meteorology studies(3) Regional Climate studies(4) Paleoclimate studies

3. Polar Optimizations to MM5 physics(1) Revised cloud / radiation interaction(2) Modified explicit ice phase microphysics(3) Optimized turbulence (boundary layer) parameterization(4) Implementation of a sea ice surface type(5) Improved treatment of heat transfer through snow/ice

surfaces(6) Improved upper boundary treatment

4. Limited future support for MM5

Polar Optimization:Fractional sea ice (ice and water

within the same grid box)Sea ice albedo (specify to vary with

time and latitude)Morrison microphysics (2-moment)Noah LSM modifications

Heat transfer for snow and iceSurface energy balance

Now there is Polar WRF



Testing of Polar WRF

1. Permanent ice sheetsStart with Greenland (Follow Polar MM5 path)

January 2002 (winter) and June 2001 (summer)

Also Antarctic AMPS forecasts (NCAR MMM Division)

Starting Antarctic climate simulations (Elad Shilo at BPRC)

2. Polar pack iceUse 1997/1998 Surface Heat Budget of the Arctic

(SHEBA) observations on drifting sea ice

Selected months: January, June, and August

3. Arctic landSoon to begin



North Atlantic Grid for Greenland Polar WRF Simulations

97 x 139

24 km spacing

28 vertical levels

Begin Polar WRF work with Greenland studies



2 m Temperature at Swiss Camp and Summit

-60

-55

-50

-45

-40

-35

-30

-25

-20

-15

-10

-5

0

1 6 11 16 21 26 31

December 2002

Tem

pera

ture

(C)

AWS Swiss Camp AWS SummitPMM5 Swiss Camp PMM5 Summit

10-m Wind Speed at Swiss Camp and Summit

-10

-5

0

5

10

15

20

25

30

35

1 6 11 16 21 26 31

December 2002

Spee

d (m

/s) a

t Sw

iss

Cam

p

0

5

10

15

20

25

30

35

40

45

AWS Swiss Camp PMM5 Swiss Camp

AWS Summit PMM5 Summit

Spee

d (m

/s) a

t Sum

mit

2 m Temperature at Swiss Camp and Summit

-60

-55

-50

-45

-40

-35

-30

-25

-20

-15

-10

-5

0

1 6 11 16 21 26 31

December 2002

Tem

pera

ture

(C)

AWS Swiss Camp AWS SummitPMM5 Swiss Camp PMM5 SummitWRF Swiss Camp WRF Summit

10-m Wind Speed at Swiss Camp and Summit

-10

-5

0

5

10

15

20

25

30

35

1 6 11 16 21 26 31

December 2002

Spee

d (m

/s) a

t Sw

iss

Cam

p

0

5

10

15

20

25

30

35

40

45

AWS Swiss Camp WRF Swiss CampPMM5 Swiss Camp AWS SummitWRF Summit PMM5 Summit

Spee

d (m

/s) a

t Sum

mit

Polar MM5

Correlation 0.84

Bias -2.3

RMSE 5.6

Polar WRF

Noah + MYJ + WSM5

Correlation 0.80

Bias 3.0

RMSE 6.0

Polar MM5

Correlation 0.87

Bias 2.5

RMSE 3.1

Polar WRF

Correlation 0.85

Bias 1.5

RMSE 2.4



Incident Shortwave Radiation Summit June 2001

0

100

200

300

400

500

600

700

800

900

0 3 6 9 12 15 18 21 24

Local Standard Time

Flux

(w/m

**2)

Observed Min

Observed Average

Observed Max


0

100

200

300

400

500

600

700

800

900

0 3 6 9 12 15 18 21 24

Local Standard Time

Flux

(w/m

**2)

Observed Min

Observed Average

Observed Max


0

100

200

300

400

500

600

700

800

900

0 3 6 9 12 15 18 21 24

Local Standard Time

Flux

(w/m

**2)

Observed MinObserved AverageObserved MaxPolar MM5 MinPolar MM5 AveragePolar MM5 Max


0

100

200

300

400

500

600

700

800

900

0 3 6 9 12 15 18 21 24

Local Standard Time

Flux

(w/m

**2)

Observed MinObserved AverageObserved MaxPolar MM5 MinPolar MM5 AveragePolar MM5 MaxPolar WRF MinPolar WRF AveragePolar WRF Max

Incident Longwave Radiation at Summit June 2001

100

120

140

160

180

200

220

240

260

280

300

0 3 6 9 12 15 18 21 24

Local Standard Time

Flux

(w/m

**2)

Observed Max Observed Average Observed Min


100

120

140

160

180

200

220

240

260

280

300

0 3 6 9 12 15 18 21 24

Local Standard Time

Flux

(w/m

**2)

Observed Max Polar MM5 Max Observed AveragePolar MM5 Average Observed Min Polar MM5 Min


100

120

140

160

180

200

220

240

260

280

300

0 3 6 9 12 15 18 21 24

Local Standard Time

Flux

(w/m

**2)

Observed Max Polar MM5 Max Polar WRF MaxObserved Average Polar MM5 Average Polar WRF AverageObserved Min Polar MM5 Min Polar WRF Min

Polar MM5 shows a deficit for incident longwave radiation.

Polar WRF is much improved and shows a small bias at Summit Greenland

Incident Longwave Radiation at Summit Greenland June 2001

Incident Shortwave Radiation at Summit Greenland June 2001

Local Standard Time

Both Polar MM5 and Polar WRF reasonably represent the diurnal cycle of incident shortwave radiation.



Summary of Greenland SimulationsSummary of Greenland SimulationsFollowing the path of development for Polar MM5, WRF has being optimized for polar applications beginning with Greenland domains.Polar WRF is at least as successful as Polar MM5 for simulations of the Greenland winter surface layer.Polar WRF simulations of the Greenland summer surface layer are comparable to those of Polar MM5 when verified with AWS observations, and surface energy balance for Polar WRF is better.



SHEBA Location (from Perovich et al. 2007)

Arctic Ocean north of Alaska

Test Polar WRF for Arctic Ocean/sea ice with selected SHEBA case studies (1997/1998)



141 x 111 25 km spacing 28 levels

Western Arctic Domain for Comparison with SHEBA observations

Noah LSM + MYJ PBL + Morrison et al. microphysics

January

August

June



Temporal evolution of sea ice surface

Apr 1 May 1 Jun 1 Jul 1 Aug 1 Sep 1 Oct 10.0

0.2

0.4

0.6

0.8

1.0

Alb

edo

Leads

Ponds

Bare iceSnow covered ice

Snow, leads, bare ice constant – ponds always changing



January Surface Pressure at Ice Station SHEBA

10001005101010151020102510301035104010451050

1 6 11 16 21 26 31

Pres

sure

(hPa

)

ObservationsPolar WRF

Correlation: 0.98Bias: 0.5 hPaRMSE: 2.2 hPa

Agreement between simulated and observed surface pressure demonstrates that Polar WRF is capturing the synoptic variability at Ice Station SHEBA during January 1998

Similar results are seen for June and August 1998

Correlation 0.98

Bias 0.5 hPa

RMSE 2.2 hPa



January Surface Temperature at Ice Station SHEBA

-45

-40

-35

-30

-25

-20

-15

1 6 11 16 21 26 31

January 1998

Tem

pera

ture

(C)

Polar WRFObservations

aCorrelation: 0.83Bias: -2.20°CRMSE: 4.29° C

June Surface Temperature at Ice Station SHEBA

-5

-4

-3

-2

-1

0

1

2

3

4

5

1 6 11 16 21 26 31

June 1998

Tem

pera

ture

(C)


b

Correlation: 0.38Bias: 0.42°CRMSE: 1.05°C

August Surface Temperature at Ice Station SHEBA

-7-6-5-4-3-2-1012

1 6 11 16 21 26 31

August 1998

Tem

pera

ture

(C)


c

Correlation: 0.46Bias: 0.20°CRMSE: 1.15°C

Surface Temperature at Ice Station SHEBA January 1998

Surface Temperature at Ice Station SHEBA June 1998

Surface Temperature at Ice Station SHEBA August 1998

August 1998



Figure 9 10-m Wind speed (m s-1) from observations and the Polar WRF simulation at Ice Station SHEBA for January, June and August 1998

August 10-m Wind Speed at Ice Station SHEBA

0

2

4

6

8

10

12

14

1 6 11 16 21 26 31

August 1998

Spee

d (m

/s)


cCorrelation: 0.83Bias: -0.6 m/sRMSE: 1.6 m/s

June 10-m Wind Speed at Ice Station SHEBA

0

2

4

6

8

10

12

1 6 11 16 21 26 31

June 1998

Spee

d (m

/s)


bCorrelation: 0.71Bias: -0.6 m/sRMSE: 1.5 m/s

January 10-m Wind Speed at Ice Station SHEBA

0

2

4

6

8

10

12

14

1 6 11 16 21 26 31

January 1998

Spee

d (m

/s)


aCorrelation: 0.88Bias: -0.6 m/sRMSE: 1.5 m/s



PWRF Summary/What is next?Polar WRF works well over the Greenland Ice SheetPolar WRF captures the synoptic variability over the Arctic pack ice

small biashigh-frequency errors due to clouds

• Test over Arctic land nextis there a warm bias during winter?

snow coverinitial soil temperature and moisturestable boundary layer + topography

• Antarctic climate testsAMPS Antarctic real-time forecasts





Dale BarkerNCAR MMM



WRF-Var Observations for ASRWRF-Var Observations for ASRIn-Situ:

- Surface (SYNOP, METAR, SHIP, BUOY).- Upper air (TEMP, PIBAL, AIREP, ACARS).

Remotely sensed retrievals:- Atmospheric Motion Vectors (e.g. MODIS).- Ground-based GPS Total Precipitable Water.- SSM/I oceanic surface wind speed and TPW.- Scatterometer oceanic surface winds.- Wind Profiler.- Radar radial velocities and reflectivities.- Satellite temperature/humidities (e.g. TOVS, AIRS?).- GPS refractivity (e.g. COSMIC).

Radiance Assimilation:- Microwave: AMSU, SSM/I, SSMI/S(?)- Infrared: HIRS, AIRS(?), IASI(?).



WRF-Var Radiance Assimilation StatusWRF-Var Radiance Assimilation StatusBUFR 1b radiance ingest.

RTM interface: RTTOV8_5 or CRTM

NESDIS microwave surface emissivity model

Range of monitoring diagnostics.

Quality Control for HIRS, AMSU, AIRS, SSMI/S.

Bias Correction (Adaptive, Variational in 2008)

Variational observation error tuning

Parallel: MPI

Flexible design to easily add new satellite sensors

QuickTime™ and aTIFF (Uncompressed) decompressor

are needed to see this picture.

DMSP(SSMI/S)

Aqua (AMSU, AIRS)

NOAA (HIRS, AMSU)



DATC Antarctic TestbedDATC Antarctic Testbed

Testbed Configuration (from MMM/AMPS):Model: WRF-ARW, WRF-Var (version 2.2).Namelists: 60 km (165x217), 31 levels, 240 s timestep.Period: October 2006.Suite: NoDA, 3D-Var (6-hourly full cycling).

SondeCoverage

COSMIC Coverage

Hui Shao, DATC



Land Component for Arctic System Reanalysis

Land Component for Arctic System Reanalysis

Fei Chen and Michael Barlage Research Applications Laboratory (RAL)

The Institute for Integrative and Multidisciplinary Earth Studies (TIIMES)National Center for Atmospheric Research



Noah LSM includes Noah LSM includes Land ice (glacier) Sea ice– The above two will be treated by

components in Polar-WRF. The routines in Noah to deal with these two processes will not be used for ASRLand-vegetationLand-bare soil



High-Resolution Land Data Assimilation System (HRLDAS) for ASR

High-Resolution Land Data Assimilation System (HRLDAS) for ASR

Blending atmospheric and land-surface observations and land surface modelTo provide land state variables for driving the coupled Polar WRF/Noah modeling system– Soil moisture (liquid and solid phase) – Soil temperature– Snow water equivalent and depth– Canopy water content – Vegetation characteristics

To provide long-term evolution of the above variables plus surface hydrological cycle (runoff, evaporation) and energy cycle (surface heat flux, ground heat flux, upward long-wave radiation)



Satellite data - MODIS & AVHRRSatellite data - MODIS & AVHRR

fPAR/LAI: MODIS 8-day, 1km, 2000-currentAlbedo: MODIS 8-day, 1km, 2000-currentLand Skin Temperature: MODIS 8-day and daily, 1km and 6km, 2000-currentGreen Vegetation Fraction: MODIS-based, 1km, 16-day, 2000-currentGreen Vegetation Fraction: AVHRR, weekly, 0.144deg, 1982-2005



MODIS 1km Vegetation DataJune 18, 2002

MODIS 1km Vegetation DataJune 18, 2002



Summary of ASR StatusSummary of ASR StatusASR grew out of Antarctic NWP. Development of enhanced components are proceeding, and will soon be merged. Coupled atmosphere-land DA, but not atmosphere-ocean. Arctic ocean DA being done by others that offers the prospect of enhanced ocean conditions (e.g., sea ice thickness).WRF (and Noah LSM) physics are being optimized for polar applications beginning with Greenland and Arctic Ocean domains. Arctic land is next.Atmospheric data assimilation advances at NCAR. Start with 3DVAR, but transition to EnKF anticipated.HRLDAS will provide high-resolution land surface variables on the same grid as WRF-3DVAR.Timeline: Completion of 2000-2010 by 2011. Second phase is anticipated to cover 1985-present in a climate monitoring capacity with major NOAA participation likely.

a synthesis of arctic climate changepolarmet.osu.edu/asr/asr_bromwich_hines_200804.pdf · division...

Documents