Arctic requirements for high resolution reanalysis · Outline (1) The increasing importance of the...
Transcript of Arctic requirements for high resolution reanalysis · Outline (1) The increasing importance of the...
Norwegian Meteorological Institute
Arctic requirements for
high resolution reanalysis
Harald SchybergThanks to: Jun She (DMI), Malte Müller (MET Norway),
Trond Iversen (MET Norway)
h.schyberg<at>met.no
Outline
(1) The increasing importance of the Arctic: changes under
global warming, new economic activities, governance
(2) Examples of potential users/usage areas for Arctic regional
reanalysis
(3) Related projects and datasets. How can Arctic regional
reanalysis add value to already existing or planned global
reanalysis and other datasets
(4) Thoughts and suggestions on requirements for design of
Arctic reanalysis
What do we mean with «the Arctic»?
Definitions differ – there is no universally
agreed southern border:
The Arctic circle 66° 33ʹ N
From climatology: The July 10°C
isotherm (roughly coincides with N
border for forest)
Here: Will not adhere to a strict
definition, but it could be natural for C3S
to have an interest in
1. key earth system processes
2. a geographical domain
corresponding to European
economical/administrative interests Illustration: Igesund/NPI
Arctic climate – rapid change is seen
Temperatures increasing more rapidly than the global average – the
“Arctic Amplification”
Sea ice – last 20 years:
Approximately half the summer coverage
Satellite and other data indicate a reduction of the order of 50% in sea
ice thickness
Summer sea ice volume roughly reduced to ¼
Permafrost temperatures have increased in most regions since the early
1980s
Impacts on ecosystems, economic activities, climate feedbacks, …
Snow and ice data provided by the National Center for Environmental Prediction/NOAA, NSIDC, U. Bremen
Climate change: The sea ice decline
Climate change: The sea ice decline (Sept.)
Projected and hindcasted September sea ice extent (colors and shading) for
IPPC climate models and observations (black line). The shading indicates the one
standard deviation range in the hindcasts and projections.
Credit: J. Stroeve and A. Barrett, National Snow and Ice Data Center
Governance of the Arctic
Some main elements:
•National states
•International agreements
•International cooperation entities
•Arctic Council
•EU has just adapted a new policy for the Arctic
Climate monitoring is one input for Arctic governance
Governance of the Arctic
International legal frameworks that applies to the Arctic:
•The UN Convention on the Law of the Sea (UNCLOS), which asserts jurisdictional rights of nations in the various maritime zones.
•The Arctic Council is an international, intergovernmental forum - ”the primary body for circumpolar regional cooperation”. (The EU is an ad hoc observer to Arctic Council proceedings, 3 Member States are members of the Arctic Council, Denmark, Finland and Sweden, while seven Member States are permanent observers);
Governance of the Arctic
On more specialized areas:
• The Barents Euro Arctic Council (BEAC) is the forum for intergovernmental and interregional cooperation in the Barents Region. The European Commission is a full member.
• The Northern Dimension is a joint policy between the EU, Russia, Norway and Iceland. It was initiated in 1999 and aims at providing a framework to promote dialogue and concrete cooperation in issues such as economy, culture, environment and transport.
• Transatlantic Ocean Research Alliance (US-Canada-Europe), the Galway agreement on ocean research, includes the Arctic.
• The OSPAR Convention aims to protect the marine environment and ecosystems from emerging threats linked to pollution, maritime activities, together with climate change and increased human presence.
• The International Maritime Organisation (IMO), a specialisedagency of the United Nations with responsibility for the safety and security of shipping and the prevention of maritime pollution by ships. All EU Member States are IMO Members. The European Commission has an observer status.
27 April 2016: An integrated European Union
policy for the Arctic
Background: The European Parliament and the Council in
2014 asked the Commission and the High Representative to
develop an integrated policy on Arctic matters, with a more
coherent framework for EU action and funding programmes.
A policy proposal that will guide the actions of the European
Union in the Arctic region. 3 main policy objectives:
protecting and preserving the Arctic in cooperation with
the people who live there
promoting sustainable use of resources
international cooperation.
39 actions listed
An integrated EU policy for the Arctic - some of
the 39 actions:
• Maintain current funding levels for Arctic research under
Horizon 2020 (on average 20 million per year). Around 40
million has already been earmarked for 2016 and 2017 for
projects on observation, weather and climate change in the
northern hemisphere and permafrost decrease.
• Support the transnational access to research infrastructures
in the Arctic and the open access to data resources. The
EU’s Copernicus space programme is to support
international research on climate change in the Arctic.
• Enhance coordination between EU funding programmes
relevant for the Arctic, identify key investment and research
priorities as well as facilitate capacity building of
stakeholders to maximise financial support for the region.
AMAP
Arctic council has six Working Groups. One of them is Arctic
Monitoring and Assessment Programme (AMAP).
Project under AMAP:
AACA - “Adaptation Actions for a Changing Arctic” – series of reports
Example – some key messages from the draft Barents area
AACA assessment
• The Arctic will warm faster than the average global warming, and
temperature projections suggest a winter-time increase of the order 3-10 ºC
between 2015 and 2080 if the future follows the path of RCP4.5. The
RCP8.5 scenario may push the warming up to 20 ºC.
• A higher proportion of the precipitation is expected to fall as rain in the
future, amplified by the sea-ice retreat and increasing the risk of rain-on-
snow events
• A number of natural hazards are connected to synoptic storms, such as rain-
on-snow events, avalanches, and extreme wave heights, but the current
projection do not provide robust indications of a change other than a
poleward shift in the storm tracks
• Polar lows are small and violent storms that represent a risk to activities in
the Arctic, and projections for the future suggest less favourable
environment for their occurrence
13
Example – some key messages from the draft Barents area
AACA assessment (2)
• The snow cover extent has decreased most at high latitudes (60-70°N), and
the decline of snow cover in Eurasia over 2007-2014 has accelerated
compared to earlier periods
• The duration of the snow season has decreased, and the melt onset date in
spring has advanced about 1-2 weeks in the 1979-2012, and the duration of
snow-cover in 2050 will be about 30-40 % less than in 2011
• Observations of the snow quality suggests an increase in very hard snow
layers from 1961 to 2009, with harder snow in early winter, more moist snow
during spring, and future warming may bring more rain-on-snow events
• The permafrost is thawing because of the Arctic warming, and the projected
warming and increases in snow thickness will result in near-surface
permafrost degradation over large geographic area
14
Some potential use areas for reanalysis
• Economic sectors: natural resource exploitation, fisheries,
tourism, transportation
• Climate system process studies
• Climate model validation
• Climate change effect studies
• Ecosystem change studies: permafrost degradation,
increasing runoff, coastal erosion, reduced ice thickness
• Arctic Populations/communities commonly distributed
along or dependent on coastal waterways and river
systems for access and subsistence
• Input to governance and resource management
• For instance Arctic Council and its “Arctic Monitoring
and Assessment Programme”
Managing the risks in economic activities:
Reanalysis can contribute
Icing from seaspray
Sea ice, icebergs
Significant wave height
Storms
Extremely low temperature
High wind and low temperature
Currents
Sea states
Darkness 24/7
Environment
The
Northern
sea route
Arcticportal.org
Lloyds (2012): «Arctic Opening: Opportunity
and Risk in the High North» (1)
Some points from analysis of a leading international insurance company:
•“Rapid and disruptive change in the Arctic environment presents uneven prospects for investment and economic development. All across the Arctic, changes in climate will create new vulnerabilities for infrastructure and present new design challenges.
•The Arctic is likely to attract substantial investment over the coming decade, potentially reaching $100bn or more. However, given the high risk/potentially high reward nature of Arctic investment, this figure could be significantly higher or lower.
•Arctic conditions will remain challenging and often unpredictable. Many of the operational risks to Arctic economic development – particularly oil and gas developments, and shipping – amplify one another. At the same time, the resilience of the Arctic’s ecosystems to withstand risk events is weak, and political and corporate sensitivity to a disaster is high.”
18
Lloyds (2012): «Arctic Opening: Opportunity
and Risk in the High North» (2)
•“The environmental consequences of disasters in the Arctic are likely to be worse than in other regions.
•The challenges of Arctic development demand coordinated responses where viable, common standards where possible, transparency and best practice across the north. These frameworks need to be in place to enable sustainable development and uphold the public interest.
•Companies operating in the Arctic require robust risk management frameworks and processes that adopt best practice and contain worst case scenarios, crisis response plans and full-scale exercises.”
19
Economic activities in the ArcticCredit:
http://arkgis.org/
Norwegian Meteorological Institute
Icing (ice accretion) conditions 1958-2013.
Norwegian Meteorological Institute
Polar lows 2000-2013
Gunnar Noer, VNN, MET Norway
Norwegian Meteorological Institute
Trend in number of mild and rainy days dec.-feb.
EWWA – Changes in winter warming events in the Nordic Arctic Region
From Vikhamar-Schuler et al (2013): Snow model
assessment of grazing conditions for reindeer
Norwegian Meteorological Institute
Existing Arctic reanalyses/hindcasts – atmos/ocean/ice
(probably uncomplete list)
Producer/
Dataset
Variables Area Period/ resolution Obs. Used in DA Model and DA method
DMI Atmosphere Pan-
Arctic
1999-2004, hourly,
15km
Synoptic data 3DVAR with ERAI as BC and
background; HIRLAM
DMI Ocean-sea ice Pan-
Arctic
1994-2013, hourly,
10km
SST, Sea ice
concentration
Nudging, HYCOM-CICE (incl.
tide)
NERSC
(Copernicus
Marine Service)
Ocean-sea ice Pan-
Arctic
1991-2014, daily,
12.5km
Currents, T/S profiles,
SST, SIC, SLA
EnKF, Hycom/TOPAZ+NERSC
ice model (no tide)
MET Norway Atmosphere
hindcast “NORA
10”
Regional 1957-present, 10km Hindcast only ERA40 BC and background,
HIRLAM
MET Norway Atmosphere
hindcast
Local/reg
ional
Not completed,
2.5km
Hindcast only HARMONIE/AROME
SMHI Ocean-sea ice-
atmosphere
Pan-
Arctic
IPY (2005-2007),
22km
AMSU-A, SST, SIC, in-situ
atm.&ocean
HIRLAM (4D-Var), HIROMB
(OI)
USA/Ohio Univ.:
Arctic System
Reanalysis
Atmos-ocean-
ice
Pan-
Arctic
2000-2012, 3hourly,
15 and 30 km
Polar WRF (3D-Var)
US: The Arctic System Reanalysis (ASR)
Polar Meteorology Group, Ohio State Univ. (Dave Bromwich), funded by NSF and NASA
Polar WRF with 3D-Var:
• 15- and 30-km (inner domain) horizontal resolution versions
• Version 1: 30 km version is complete for 2000-2012
• Version 2: 15 km version with some model updates: available soon
Bromwich et al, QJ, 2016 – comparison of version 1 with ERA interim:
Higher resolution terrain and detailed land-surface description useful
Some verification scores better, some worse than ERAI
Improvements over ERAI in near-surface fields
Improvements in depiction of mesoscale processes
http://polarmet.osu.edu/ASR/
Hindcast partly covering Arctic: «NORA10»
Hindcast produced at MET Norway, NWP model HIRLAM is used to downscale ERA-
40 data to 0.1° resolution over Norway and adjacent sea areas incl parts of Arctic
Originally covered the ERA-40 period,until 2002, has since been extended using
ECMWF operational analyses (Reistad et al. 2011, Haakenstad et al. 2012.)
Below figure from Reistad et al, 2009:
Wind fields benefit from more resolved orography
Summary: Need for high-resolution Arctic
reanalysis vs existing datasets
Existing datasets for the atmosphere:
• Resolution down to 10-15 km
• Limited time ranges
• Limited satellite data usage
Possible added value from a new reanalysis:
• An open European dataset
• Higher horizontal resolution would add value by
• Capability to describe important mesoscale convective
phenomena in cold air outbreaks over ocean (extreme surface
heat fluxes, polar lows)
• Forcing from lower boundary (orography, sea ice, snow cover,
…)
Major challenges for Arctic atmospheric reanalyses
• Good datasets and strategies for the lower boundary:• Sea-ice with varying extent, ice concentrations and
properties that influence the surface heat and moisture fluxes.
• Snow cover/properties, glaciers (Greenland)• Coupling with sea ice/ocean model could come later
• Very stable boundary layers connected to strong surface cooling
• Vigorous convection over open ocean when air masses arrive from the sea-ice or cold land (Major Cold Air Outbreaks)
• Mesoscale phenomena in cold air outbreaks over ocean:Extreme surface heat fluxes, convective systems, polar lows
• The observing system: Use of satellite data will be important
Vigorous convection: Polar lows
Satellite image of the polar low situation 6 March 2013
Norwegian Meteorological Institute
Polar lows
(An unusual polar low 8 January
2010: First time a polar low has
been observed North of
Spitsbergen. Possible with
retreat of winter sea ice. Polar
lows form only over open
ocean.)
MET Norway forecasters saw a
significant improvement in
Polar Low NWP modeling
capability when resolution went
to 10-15 km and below.
Convection resolving models
an advantage.
28. april 2016Danmarks Meteorologiske Institut
Veðurstofa Íslands
Side 31
Glaciers need to be accounted for:
• No glacier surface type was defined in SURFEX
caused large drift in NWP output
• Remedies were needed to remove drift, and
turned out to make HARMONIE is able to
simulate melt area well
It is important to use correct and updated
physiography data for NWP modelling & reanalysis
• Work is done to update Greenland ice sheet
extent as included in the EcoClimap-II data set
• Data from Greenland Mapping Project used to
replace GMTED2010 topography
Some experience regarding
Greenland (DMI, IMO)
The Arctic observing system
Upper air conventional observations:
Example radiosonde (left) and aircraft coverage (right) (13 Sep 2013)
Observing system relies on remote sensing in the Arctic: Good use of remote
sensing data is important for reanalysis quality
Measuring of model quality with sea-level pressure forecasts
We chose to use pressure forecasts at coastal stations:
• strongly connected to capturing the main weather systems
• not as strongly influenced by local conditions at the measurement stations as for instance wind and temperature would be
• is thus more comparable between different observing stations as a measure of the general quality of the large-scale forecast fields
Of interest to know atmospheric pressure variability which needs to be described for reference
Left: Mean absolute day-to-day observed pressure differences in hPa
Norwegian Meteorological Institute
Model accuracy decrases towards the North (1)
The figures shows the RMS errors in pressure (vs observations) for forecasts in the range from 18 to 42 hours
· Left: Operational ECMWF global model (~16 km resolution)
· Right: Operational HIRLAM 12 km regional model
34
The Arctic Atmospheric observing system
Relies heavily on satellite data
A preliminary gap analysis
• gap in pressure observations over sea ice and ocean areas
• almost no coverage of near-surface wind observations over sea ice
• gaps in conventional upper air data is compensated by data from satellite
sounding instruments
• data from temperature sounding in the lower troposphere is difficult to
exploit the signals have surface contributions which are generally not
well modelled (sea ice surface)
Both wind and temperature information in the lower Arctic
troposphere is missing in the remote ocean and ice areas
(Schyberg and Randriamampianina, EU ACCESS project)
35
Some onging frameworks for enhancement and
coordination of the Arctic observing system
SAON – Sustaining Arctic Observing Networks
EU Horizon 2020 calls under «Blue Growth» (under evaluation):
· BG-9-2016: Integrated Arctic Observing System
· BG-10-2016: Impact of Arctic changes on weather/climate of Northern Hemisphere
ESA Polaris program: «… development of the next generation of space infrastructure to support both scientific and operational information needs in the rapidly evolving Polar Regions»
As well as planned polar orbiters from ESA and EUMETSAT (for instanceADM-Aeolus)
… sets the scene for increased quality and usefulness of Arctic reanalysis
36
Operational regional NWP systems in the Arctic
Upper: «AROME Arctic» of MET Norway
Lower: DMI/IMO joint project
Both are HARMONIE/AROME:
• At 2.5 km horizontal resolution
«AROME Arctic» at MET Norway:
• 3D-Var data assimilation, conventional
obs + AMSU/MHS, IASI, ASCAT
• Ongoing work on AMV implementation
• Sea ice and SST from daily OSI SAF
products. Ongoing work on updated sea
ice module in SURFEX.
37
Requirements for Arctic reanalysis could
include:
• Convection resolving?
• Not fully pan-Arctic, but cover some domains of interest for
economic activities and change processes
• Make good use of the main components of the Arctic
observing system (satellite datasets needed)
• Make good use of complementary data on surface properties
relevant in the area, such as ocean and sea ice, snow, …
• Length of time period: similar to approach for other regional
reanalysis datasets
Summary
Arctic reanalysis can contribute to
• Provide important input to studies of climate change and earth
system processes in the area
• Managing risk in emerging economic activitis in the area
• Provide knowledge input to governance towards sustainable
development of the area
Some suggestions:
• Needs a good strategy for providing a good description of the
lower boundary forcing: Sea ice, ocean, snow, glaciers
• (A fully coupled atm.-ocean-seaice reanalysis a task for the future)
• Higher resolution than global reanalysis will help description of
convective phenomena in Arctic climate, effects of small scale
orography and surface forcing
Thank you!
Backup slides follow
40
Norwegian Meteorological Institute
Norwegian Meteorological Institute
Regional downscaling - HARMONIE Climate branchCooperation MET Norway with SMHI (and KNMI)
Norwegian Meteorological Institute
UERRA: Uncertainties in Ensembles
of Regional ReAnalysesObjectives
• To produces ensembles of European regional meteorological
reanalyses of Essential Climate Variables for several decades
• To estimate the associated uncertainties in the data sets
• To provide more observations for reanalyses
• To provide data services and user information
Pre-operational
Copernicus Climate Change Services
Norwegian Meteorological Institute
44
UERRAProvide data sets and uncertainty estimates,
building from FP7 EURO4M and
extends in several dimensions:
More models and analysis systems (also gridded obs)
Ensembles
Much longer time period – 50+ years or 30 for ensemble
Higher resolution than ever before
More digitised observations rescued and provided
Quality and uncertainty measures
Comprehensive web based data and visualisation service
User guidance on user oriented products
Norwegian Meteorological Institute
ensemble ~1978 -2015
1 Control 12 km
70 levels
10 (20) members 24 km
ensemble
Hybrid 4D-Var,
Ensemble of 4D-VARs
Conventional obs,
satellite data, precip?
Met Office SMHI/MF
deterministic ~1961-2015
5 years ensemble
1 member 11 km
65 levels
2 members physics (5 y)
HARMONIE
3D-VAR
Conventional obs,
Large scale constraint
from ERA
Uni Bonn DWD
ensemble ~5 years
1 Control 12 km
40 levels
20 members 12 km
ensemble
Conventional obs
Radar, new in LETKF
LETKF and
Ensemble Nudging
boundary forcing from global ERA reanalyses
(ERA-40, ERA-Interim,ERA -5, incl. Ensembles, EDA)
3-D model level fields u,v,T,q,clouds, 2D fields
precipitation, surface p, T, RH, snow and more
Norwegian Meteorological Institute
2-D surface fields analyses driven by 3D reanalyses
SMHI
MESAN
~1982 - 2013
5 km Europe
Cloud fraction
hourly
2D advanced
Statistical
interpolation
AVHRR, METEOSAT
SEVIRI and
MVIRI
Downscaled
3D HIRLAM model
Climatological
adaptation background
MF/SMHI
MESCAN
1961 - ~2013
5 km Europe
T2m, RH, 24 h
precipitation
2D advanced
Statistical
Interpolation
Surface and climate
stations
T, Td,
precipitation
Downscaled
ALADIN model
background
SMHI
HYPE
~1979 - 2010
River discharge
35000 catchments
Europe, median
215 km2
Hydrological
physical
model
No input observations
Validation against
discharge data
ERA, EURO4M and
UERRA reanalyses
Precipitation and
temperature forcing
MF SURFEX
and TRIP
~1981 - 2010
River discharge
25 km -> rivers
Surface flux model
Hydrological physical
model
No input observations
Validation against
discharge data
MESCAN
atmospheric
variables and
precipitation
Norwegian Meteorological Institute
Uncertainty estimations
Evaluate ensemble reanalyses and downscaled reanalyses through
comparison to independent ECV datasets that were derived independently
Establish a consistent knowledge base on the uncertainty of reanalyses
across all of Europe through a common evaluation procedure
o Extremes, Climate Indices and Indicators of user interest, scales of
variablity
Statistically assess the provided information by applying the common
evaluation procedure on all data
User dependent parameters and language adaptation
Sea Ice Concentration Climate Data Record
1979
(sept)
1978 - 2015 : Climate Data Record 2015 - … : Interim CDR
1979
2012
2015
SIC Algorithm, Processing, and Operationality
Some unique features of the algorithm:
Dynamic adjustments of the algorithm tie-points (=> Consistency) ;
Maps of uncertainties;
Rooted in the OSISAF, consolidated by the ESA CCI Sea Ice R&D.
Processing chain:
Engineered in the OSISAF, consolidated by ESA CCI Sea Ice.
Scalable data re-processing system at MET Norway (but data volumes are small).
Operational in the sense of EUMETSAT SAF and CMEMS:
Review cycles (Req. Review, Product Consolidation Review, Delivery Readiness Review);
Traceability of requirements, CORE-Climax Maturity Matrix, CEOS ECV Inventory,...
Sea Ice Type, Sea Ice Drift, ...
S. Kern (2016)
March 2015
M. Tschudi (2014) US NSIDC
June snow cover 2012 relative to 1971-2000
Credit: James Overland