Understanding Climate Understanding Climate Variability and Predictability Variability and Predictability using using Models and Observations: Models and Observations: Research at CGAM Research at CGAM

(1)

Understanding Climate Understanding Climate

Variability and Predictability Variability and Predictability

using using

Models and Observations:

Research at CGAM

(2)

NERC: Natural Environment NERC: Natural Environment

Research Council Research Council

• UK lead responsibility for developing, funding and delivering environmental science

• To prioritise and deliver world-class environmental sciences to understand the Earth System.

• Supports research, training and environmental

observations across all components of the earth system

• Uses a ‘whole system’ approach to earth system

science to find sustainable solutions to environmental

problems.

(3)

NERC Principle Science Areas NERC Principle Science Areas

• Earth’s Life-Support Systems: Water, biogeochemical cycles and biodiversity

• Climate Change: Predicting and mitigating the impacts

• Sustainable Economies: Identifying and providing sustainable solutions to the

challenges associated with energy, land

use and hazard mitigation

(4)

NERC supports scientists within Universities NERC supports scientists within Universities

and at its major core research centres:

• British Antarctic Survey (BAS)

• British Geological Survey (BGS)

• Centre for Ecology and Hydrology (CEH)

• Proudman Oceanographic Laboratory (POL)

• Southampton Oceanography Centre (SOC)

• Centre for Terrestrial Carbon Dynamics (CTCD)

• NERC Centres for Atmospheric Science (NCAS)

• Environmental Systems Science Centre (ESSC)

• Data Assimilation Research Centre (DARC)

• Tyndall Centre for Climate Change Research

• National Institute for

Environmental e-Science (NIEeS)

• Centre for Polar Observations

and Modelling (CPOM)

(5)

• Delivers and coordinates NERC’s core

strategic research in atmospheric science

• Supports centres of excellence and facilities, distributed across many UK universities and related institutions

• Works closely with Meteorological Office/Hadley Centre and the UK Environment Agency

NERC Centres for NERC Centres for

Atmospheric Science

(6)

Centres and Facilities Centres and Facilities

• Centre for Global Atmospheric Modelling (CGAM): Climate processes, variability and change

• Atmospheric Chemistry Modelling Support Unit (ACMSU): Chemistry- Climate modelling and interactions

• Universities’ Weather and Environment Research Network (UWERN):

High impact weather

• Distributed Institute for Atmospheric Composition (DIAC): Laboratory and field studies of processes in the chemical and physical environment.

• British Atmospheric Data Centre (BADC)

• University Facilities for Atmospheric Measurements (UFAM)

• Facility for Airborne Atmospheric Measurements (FAAM)

(7)

Centre for Global Centre for Global

Atmospheric Modelling Atmospheric Modelling

• To understand and simulate the highly non-linear dynamics and feedbacks of the global climate system

• To exploit the revolution in seasonal to interannual prediction as a test bed of climate models and as a vehicle for fostering integrated applications

• To capitalize on and develop NERC expertise in earth system science

• To harness the expected increases in computer power and the

opportunities provided by e-science to perform higher resolution,

more comprehensive integrations of the earth system.

(8)

 Will there be an El Nino this year and how severe will Will there be an El Nino this year and how severe will it be?

it be?

 Will we have autumn floods like 2000 again? Will we have autumn floods like 2000 again?

 Will the milder winters of the last decade or so Will the milder winters of the last decade or so continue?

continue?

 Is there a risk of rapid climate change associated with Is there a risk of rapid climate change associated with THC shutdown?

THC shutdown?

 Can we reproduce and understand past abrupt changes Can we reproduce and understand past abrupt changes in climate?

in climate?

Typical questions that CGAM addresses are:

(9)

To address these questions CGAM uses:

1. Observations of the climate:

Re-analyses of the global circulation

Satellite

Observations

In situ

Measurements

Essential for (i) describing variability of the current climate,

(ii) finding associations between different parts of the climate system,

(iii) evaluating climate model simulations.

(10)

2. Models:

Since the long-term memory of the system, required for predictability, resides in the oceans and land, CGAM uses state-of-the-art models of the complete system:

Models are our laboratory. We use them to (i) investigate predictability,

(ii) to explore forcing and feedbacks in the climate system, and (iii) to

test hypotheses.

(11)

CGAM Principal Research Areas CGAM Principal Research Areas

 Euro-Atlantic climate variability and predictability Euro-Atlantic climate variability and predictability ( ( Rowan Sutton, Brian Hoskins Rowan Sutton, Brian Hoskins ) )

 Mechanisms of climate variability and predictability using Mechanisms of climate variability and predictability using a model hierarchy (

a model hierarchy ( Brian Hoskins, Mike Blackburn Brian Hoskins, Mike Blackburn ) )

 Paleoclimate modelling ( Paleoclimate modelling ( Paul Valdes Paul Valdes ) )

 Tropical climate variability and predictability ( Tropical climate variability and predictability ( Julia Julia Slingo

Slingo ) )

(12)

Understanding the UK floods of Autumn 2000: Were they predictable?

Autumn 2000 experienced heaviest rainfall

on record Associated with

persistent weather pattern

driven from the tropics?

Atlantic Ocean was also much warmer than usual.

Did the ocean play a role?

But the extreme rainfall was not captured

by the seasonal forecasts.

We need to understand

why.

Courtesy: Mike Blackburn, Brian Hoskins, CGAM

(13)

Does El Nino/La Nina influence UK climate?

• El Nino/La Nina affects the seasonal climate of the Atlantic and the UK in a potentially

predictable manner.

• The forcing from the Pacific dominates when El Nino/La Nina is

strong.

Global ocean forcing

Without Atlantic forcing

1997:El Nino 1998: La Nina

Implied

effect of

Atlantic

(14)

So what role for the Atlantic?

Consider another year (1999) when El Nino/La Nina was weaker:

• In the absence of strong Pacific forcing, the state of the Atlantic Ocean is important for seasonal predictability.

• Strong evidence that the Atlantic Ocean affects the climate of the UK

and western Europe.

(15)

The Thermohaline Circulation (THC) and UK climate

The THC describes the transport of heat by the global ocean circulation.

The release of heat to the atmosphere over the

Atlantic gives our

relatively mild winters.

Recent CGAM research has reinforced the association between changes in the strength of the THC and those in the NAO. We have shown that variations in the THC lead those in sea surface

temperatures and the NAO by ~ 3 years.

(16)

See: http://www.nerc-essc.ac.uk/~kih/TRACK/Track.html

Example of tracks and intensities (coloured dots) of cyclones during typical winter season (DJF)

Note two major storm tracks over Pacific and Atlantic.

Also secondary storm track over Mediterranean and extending eastwards.

By producing a statistics of the tracks and intensities, various aspects of storm track behaviour can be diagnosed.

Variations in the tracks and intensity of cyclones

depending on the phases of the NAO, ENSO, for

example, can also be studied.

(17)

Mean track density for DJF: 1979-95

ERA-15

HadAM3

Pacific storm track is too strong. Atlantic storm track too weak in HadAM3, possibly related to lack of systems coming off the eastern side of the Rockies.

(18)

Mean intensity of the cyclones for DJF: 1979-95

ERA-15

HadAM3

Patterns quite well simulated, although peak intensity generally underestimated in HadAM3. Resolution?

(19)

Mean genesis (source) regions for cyclones in DJF: 1979-95

ERA-15

HadAM3

Cyclones generated mainly by orography and in the baroclinic zones off the eastern seaboards.

Note also secondary genesis over mid-ocean. Regions of genesis well captured by HadAM3, although the birth of cyclones along the eastern edge of the Rockies is inadequate.

(20)

2002: A SUMMER OF DROUGHTS AND FLOODS Drought in India:

India experiencing worst monsoon for over 30 years. Major drought conditions for NW India (Figure 1)

Associated with major break in monsoon during July (Figure 2), linked to suppressed phase of an exceptionally active MJO season (Figure 3)

Figure 1: Accumulated All India Rainfall (upper) and % departures from normal up to 21 August (lower).

Indian Rainfall courtesy of ‘Monsoon On Line’: http://tropmet.res.in/~kolli/MOL/

Figure 2: Daily All India Rainfall

Figure 3: Hovmoller diagram of equatorial 200hPa velocity potential anomalies showing eastward propagation of the MJO. Orange is suppressed phase

(21)

Floods in Eastern Europe:

A link to the Indian Monsoon?

Rodwell and Hoskins (1999, QJRMS) identified a key link between the Asian Summer Monsoon (ascent) and dry weather over the Eastern Mediterranean (descent; Fig.3). Could a significant break in the monsoon influence Eastern Europe?

Figure 3: Vertical motion at 500hPa generated by an idealised model in response to monsoon heating centred on 25⁰N, 90⁰E.

Early results suggest that a link may exist (Fig. 4) between monsoon breaks and disturbed weather over eastern Europe, particularly in extreme events.

Also noteworthy that extreme events tend to occur in years with El Nino conditions e.g. 1972, 1987?

Figure 4: Timeseries of anomalies in All India Rainfall and subsidence over Eastern Europe (0⁰-30⁰E, 35⁰N – 45⁰N) during July. Correlation of r=0.39. Note monthly mean subsidence of ~0.05hPa and All India Rainfall of ~5mm/day Courtesy: Mike Blackburn (CGAM)

(22)

The MJO and coupling with the ocean: Observations

(Woolnough et al., 2000: J. Clim., 13, 2086-2104)

Observations show a

coherent relationship between

convection and SST. Warm SSTs precede convection by 5-10 days and are the result of weaker winds, reduced LH flux and increased SW flux during suppressed phases of the MJO.

(23)

The MJO and coupling with the ocean: Modelling

(Inness and Slingo, 2002, J. Clim. In press)

CGCM has a propagating

convective signal compared with standing oscillation in AGCM.

Coherent variations in SST in CGCM

Coupling with the upper ocean is important for the

MJO

(24)

BUT intraseasonal SST variations in CGCMs are too small and the MJO signal is still weak:

Is the representation of the upper ocean adequate?

Large freshwater flux sets up a salt stratified barrier layer so that a shallow mixed layer forms which can respond rapidly to flux

variations, such as the diurnal cycle in solar radiation. The

presence of this barrier layer can potentially provide much stronger local coupling in the warm pool region than is currently found in coupled models which do not resolve the detailed structure of the warm pool upper ocean.

Schematic showing formation of salt barrier layer

(From Anderson et al., 1996: J. Clim)

(25)

Typical window brightness (K) images showing scales of convective organization

Note tendency for cloud clusters to congregate together to form super-clusters with multi-day life cycles e.g. Madden Julian Oscillation Self organization

(26)

Temporal behaviour of convection around the equator from window brightness temperature for Jan.-Feb. 1992

Note evidence of coherent propagation.

(27)

Kelvin

Rossby

Inertio-gravity Inertio-gravity

Mixed Rossby-gravity

Space-time spectra showing the organization of convection in association with theoretical equatorial waves.

Anti-symmetric Symmetric

MJO

From Wheeler and Kiladis 1999: J. Atmos. Sci.

(28)

Space-time spectra from R30 version of GFDL model

Note lack of organization, an error common to many GCMs.

Lack of self-organization mechanism?

(29)

Evolving grid approach Example of application of adaptive mesh refinement (AMR) to tropopause fold event.

AMR places the

resolution where the situation demands it, in this case around PV filaments.

Courtesy Dr. N. Nikiforakis, DAMTP

(30)

Probability distribution functions (PDF) of monthly mean SST and precipitation over the tropical Pacific:

DJF (upper panels), MAM (lower panels)

CMAP HadAM3 HadAM3-CMAP

Note tendency for HadAM3 to overestimate precipitation over warm SSTs. PDF is also too tight, following closely the exponential relationship implied by the Clausius-Clapeyron equation for saturated vapour pressure.

(31)

HadAM3 ENSO Simulations

• 6 member ensemble of HadAM3 forced with observed SSTs: 1870-1998

• Compare with NCEP Reanalyses

• Composite El Nino events which have similar strength, evolution and seasonality. 5 events chosen: 1957/58, 1965/66, 1972/73, 1982/83 and 1997/98.

Ref: Spencer and Slingo, 2002: Journal of Climate (Submitted)

(32)

Composite SST anomalies for El Nino events

DJF at peak of El Nino

MAM after peak of El Nino

Note global patterns of coherent changes in SST. These are forced by the atmospheric response to the primary SST anomalies in the tropical Pacific.

(33)

Composite PMSL anomalies for DJF at peak of El Nino

NCEP Reanalyses

HadAM3

Note good simulation of tropical anomalies – the Southern Oscillation. Anomalies over N.

Pacific show major errors with an in situ deepening of the Aleutian Low in HadAM3 rather than a shift eastwards and the development of a ridge over the north west Pacific.

(34)

Composite precipitation for DJF at peak of El Nino

NCEP Reanalyses

HadAM3

Note eastwards shift of precipitation maxima in NCEP reanalyses with reduced rainfall over the Maritime Continent. HadAM3 retains the precipitation maxima over the West Pacific, leading to the lack of eastwards shift in the Aleutian Low.

(35)

Understanding the influence of remote ocean response on ENSO teleconnections

HadAM3 experiments: 10 realisations of each phase of El Nino/La Nina cycle

Courtesy: Hilary Spencer

(36)

POGA

IPOGA

TOGA

GOGA DJF at peak El Nino JJA after peak El Nino

•At peak of El Nino, Indian Ocean SSTs have a significant effect on ENSO; tropical

Atlantic and extra-

tropical SSTs are less important.

•After peak of El Nino, warming of tropical remote oceans, both Indian and Atlantic, significantly affects atmospheric response.

•Remote ocean

response to ENSO may lead to extended

predictability.

(37)

Example of El Nino in Hadley Centre coupled models

Note tendency, common to many coupled models, for El Nino to occur too often and to be too regular. Also the temporal behaviour of El Nino (but not amplitude) appears to be insensitive to the ocean model used.

Normalised power spectra of Nino3 SST anomalies.

Numbers in

brackets

indicate the

standard

deviation of

the Nino3

timeseries

(38)

Modular Earth System Modelling:

A new approach for understanding the coupled system A new approach for understanding the coupled system

BIO- GEO- CHEMISTRY

COUPLER COUPLER

LAND BIOSPHERE

OCEAN BIOSPHERE ATMOSPHERIC

CHEMISTRY

LAND SURFACE ATMOSPHERE

CRYOSPHERE OCEAN

SOCIO-ECONOMIC DATA

ASSIMILATION SYSTEMS REGIONAL

CLIMATE MODEL

•The core of the model is the coupler which exchanges

information between different components of the earth system.

•CGAM has pioneered the use of such a structure in the UK and has demonstrated its value by interchanging ocean and

atmospheric modules.

DATA ASSIMILATION

SYSTEMS REGIONAL

CLIMATE MODEL

(39)

An Infrastructure Project for Climate Research in Europe

• Involves current state-of-the-art atmosphere, ocean, sea-ice,

atmospheric chemistry, land-surface and ocean-biogeochemistry models

• 22 partners: leading climate

researchers and computer vendors

• Ultimate objective: Distributed

European network for Earth System Modelling

• See http://prism.enes.org PRISM will:

• Coordinate European Climate Modelling efforts:

– Create a European service and management infrastructure for European wide, multi-institutional climate and Earth System simulations

• Develop a European Climate Modelling System:

– Portable, efficient and user-friendly + based on state-of-the-art models + diagnostics and visualisation

(40)

Application of modular approach:

Understanding coupled GCMs

SINTEX HadOPA HadCM3 HadCEM GloSea … HadAM3

HadOM3 HadOM3 HadGOM OPA

ECHAM

T30/T42/T106

LMDz

Common ocean

Common atmosphere

Intercomparison identify origin of errors Different

resolutions

(41)

What controls El Nino in Coupled Climate Models?

Exchanging ocean

models suggests that:-

•atmosphere controls the periodicity

• ocean controls the strength of El Nino

• El Nino tends to be too regular and occur too often.

Observations Model

BUT the use of a high resolution atmosphere (1

⁰

) dramatically

improves the temporal behaviour of El Nino and for the first time

provides a more realistic simulation of the lower frequencies.

(42)

UK-HIGEM UK-HIGEM

A National Programme in ‘Grand Challenge’

High Resolution Modelling of the Global Environment High Resolution Modelling of the Global Environment

 To develop a high-resolution version (~ 1⁰atmosphere, 0.33⁰ ocean) of the Hadley Centre Global Environment Model (HadGEM).

 To evaluate the model by stringently testing it against observations and more sophisticated, very high-resolution models of the component parts.

 To improve our understanding and predictive capabilities in global environmental variability and change, with particular reference to extreme events, interactions between different components of the climate system, and the potential for climate ‘surprises’.

 To provide the modelling framework in which new developments in numerical methods and key processes in the atmosphere, ocean, cryosphere and land can be efficiently incorporated, leading to the creation of the next generation global environment model

 To provide the background models required for the synthesis and interpretation of the wealth of in situ and satellite observations of the global environment,

 To deliver more robust estimates of the regional impacts of climate change required to guide government policy.

(43)

Image from 1/8⁰ version of OCCAM Ocean GCM

showing salinity jets at a depth of 100m in the South Pacific where the South equatorial Current is blocked by a series of island groups.