Climate Simulation and Modelling at Ministry of Earth Sciences - - PowerPoint PPT Presentation

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Indian Institute of Tropical Meteorology (IITM)

A.K.Sahai

Climate Simulation and Modelling at Ministry of Earth Sciences

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Major Objectives of MoES

To provide the country best possible weather forecast (short range ) and climate prediction (long range ) To conduct the R & D required to improve the skill of both weather and climate forecasts To conduct regional climate change research to provide reliable projection of monsoon under climate change

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Weather Climate

Climate: A statistical description of weather

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Global climate Modeling

F = Friction (turbulent dissipation) Q = Non Adiabatic Heating = Net Radiation + Latent heat (clouds) + Sensible heat

.

p p

C Q p T C RT T V t T p V p RT p F V k f Dt V D

.

. . ˆ                                         

Basic equations for Weather and Climate Models in pressure coordinate system

• -------------(1)
• -------------(2)
• -------------(3)
• -------------(4)

Complexities involved in a Climate Modelling System

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Calculation of Heating Distribution

AGCM/OGCM

NS Eqs. Global 3-D + time

Radiation

Incoming SW Outgoing LW

Clouds

• Convective
• Startiform

Land-Surface Processes

• Vegetation
• Soil moisture

Boundary Layer Turbulence

• Fluxes
• Mixing
• Dissipation

Stratospheric Chemistry

• Heating
• Stability

Aerosols

• Direct Rad Eff
• Indir eff thr clouds

Key Uncertainties for Climate :

High Clouds: Dominant effect is that they Trap heat (warm) More Clouds=Warming Fewer Clouds=Cooling

Source: Schär

Source: Schär

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

All physical processes involved in heating arise from complex small scale processes, that need to be parameterized in the model Accuracy of paramereization determines heating distribution and hence weather and climate More complex paramerization requires more computation Improvement of parameterization need R & D

Scale Interaction and Parameterization

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Horizontal Resolution of the Contemporary AGCM/OGCM 500 km 300 km 75 km 150 km

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Simulations at 200km and 50 km

Resolution is of key importance for the representation of hydrological Cycle and extreme rainfall events

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Source: Kinter

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Weather & Climate Prediction Chaotic System; Probabilistic prediction

Large number of ensemble of each prediction

Initial value problem; 4-D data Assimilation

 Variational Assimilation ; Adjoint of the model; Extremely computation intensive;  It is found that preparation of the I.C. for

• perational weather prediction at high

resolution takes more computer time than actually making the prediction!

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

The Butterfly Effect

• Discovery of the

“butterfly effect” (Lorenz 1963)

• Simplified climate

model… When the integration was restarted with 3 (vs 6) digit accuracy, everything was going fine until…

Time

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

• Solutions began

to diverge Solutions diverge Time

The Butterfly Effect

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

• Soon, two

“similar” but clearly unique solutions Solutions diverge Time

The Butterfly Effect

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

• Eventually, results

revealed two uncorrelated and completely different solutions (i.e., chaos)

Solutions diverge Time Chaos

The Butterfly Effect

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

• Ensembles can be

used to provide information on forecast uncertainty

• Information from the

ensemble typically consists of… (1) Mean (2) Spread (3) Probability

Ensembles useful in this range! Solutions diverge Time Chaos

The Butterfly Effect

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

• Ensembles extend

predictability…

• A deterministic

solution is no longer skillful when its error variance exceeds climatic variance

• An ensemble remains

skillful until error saturation (i.e., until chaos occurs)

Solutions diverge Chaos Time

Ensembles extend predictability

Ensembles especially useful in this range!

The Butterfly Effect

Data Assimilation

Univariate SI Multivariate SI 3DVAR 4DVAR 4DVAR/EnKF 4DVAR/EnKF

• Adv. sounders
• Adv. sounders
• Adv. sounders

IR/MW sounders IR sounders Scatterometer Scatterometer Scatterometer Scatterometer TRMM TRMM TRMM Rainfall assimilation Rainfall assimilation Mesoscale assimilation Chemical species

1975 1985 1990 1997 1999 2005 -

Increasing complexity Vast increase in data

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Success story of Numerical Weather Forecasting! Great Improvement in medium-range forecast skill.

12-month running mean

• f

anomaly correlation (%) of 500 hPa height forecasts

Note the convergence

• f skill in NH

and SH

From ECMWF

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Improving the Forecasts

Must improve the MODEL and Data Assimilation Ex: Consider seasonal prediction with a CGCM

100 yr integration for testing mean climate Hindcast experiments to test prediction skill; 25 member ensemble x six month prediction x 20 years = 250 year integration Must turn around within a few days so that

• ther improvement could be tested!

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Evolution of Climate Models in last 5 decades at renowned climate centers

India’s Present status Leading climate centers’ status

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Computational requirements of climate models.  Range: Assumed efficiency of 10-40%. 0 - Atmospheric General Circulation Model (AGCM; 100 vertical levels) 1 - Coupled Ocean-Atmosphere-Land Model (CGCM; ~ 2X AGCM) 2 - Earth System Model (with biogeochemical cycles) (ESM; ~ 2X CGCM)]

[Range: Assumed efficiency of 10-40%. 0 - Atmospheric General Circulation Model (AGCM; 100 vertical levels) 1 - Coupled Ocean-Atmosphere-Land Model (CGCM; ~ 2X AGCM) 2 - Earth System Model (with biogeochemical cycles) (ESM; ~ 2X CGCM)]

How a code in a coupled model works?

Source: Anne Roches & Piccinali

International Centres HPC Current Capacity

NERSCC, USA Cray XT5 ~1.17PF (peak) UKMet Office IBM P6 IBM P7 ~150TF ~900TF (by 2011) NCAR, USA IBM P5/P6 ~80TF NCEP, USA IBM P6 ~90TF German Met Office IBM P6 ~165TF ECMWF IBM P6 IBM P7 ~300TF ~1PF (by 2011) JAMSTEC Earth Simulator ~131TF KMA Cray XT5 ~600TF National Supercomputing Center in Tianjin China NUDT ~4.7PF Oak Ridge National Laboratory USA Cray XT5 Supercomputer(JA GUAR) ~2PF

These centres are also having additional HPC for operational/other usage.

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Phase 1 Number of Clusters 2 compute clusters (272 nodes each) Compute Nodes 272 x 32-core POWER6 (SMT) Peak Performance ~300TF (Total) Sustained Performance ~20TF HPC at ECMWF

National Weather and Climate Centres HPC Current Capacity NCMRWF, Noida IBM P6 ~23TF IMD, New Delhi IBM P6 ~15TF INCOIS, Hyderabad IBM P6 ~7TF IITM, Pune IBM P6 ~70TF

2010 July Ranking 94th 2010 November Ranking 137th

2011 November Ranking 403rd

In India where do we stand?

• No. of Systems in Top 500 from Different countries

India:2 (0.4%)

Source: Top 500 list, Nov. 2011

China:74 USA:263

T62L64 T126L64

Dynamical Seasonal Prediction of Indian Monsoon JJAS Rainfall – 2010 (CFS V1.0) Issues in April

Central Indian drought predicted by CFS model Above normal rainfall over southern peninsular India IITM CFS T62 IITM CFS T126 IMD

T62L64 T126L64

Dynamical Seasonal Prediction of Indian Monsoon

With Initial Conditions generated within India at (INCOIS & NCMRWF)

JJAS Rainfall – 2011 (Issued in March)

Central Indian above normal rain predicted by CFS model Below normal rainfall over southern peninsular India IITM CFS T62 IITM CFS V2.0 T126 IMD

Upto 10th September

Monsoon Performance = 100±4.5 %

Predictions for 2012: Predicted Vs. Observed

Monsoon Performance = 92 %

Actual Rainfall Departure (IMD)

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Extended range Prediction of Active Break Cycles of Monsoon

• Forecast of seasonal mean rainfall may not be very useful and

meaningful when the mean is close to normal (70%). The regional/temporal distribution of rainfall anomalies is very

• inhomogeneous. Therefore, in addition to the seasonal mean All

India rainfall, we need to predict some aspects of monsoon 3-4 weeks in advance on a relatively smaller spatial scale that will be useful for farmers.

• The extended range prediction refers to a meteorological

forecast more than 10 days in advance which is the normal predictability range of weather systems (storms, cyclones etc.)

• In two slides to follow some efforts for extended range

prediction during summer monsoon season 2012 over India is presented.

Evolution of daily rainfall and wind at 850 hPa From 29th Aug, 2012 IC from

20-25 June

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Climate Change Simulation

Proposed System of MoES Computing Facilities (2012-2015)

IITM, Pune 1250 TF NCMRWF, Delhi 700 TF INCOIS, Hyderabad 80 TF IMD, Delhi 350 TF

NKN Connectivity 1 GBPS 10/40 GBPS Earth Science Grid Internet 2 US-India-EU (Monsoon Mission)

Universities & Academic Institutes

Data Transfer among MoES Institutes/Day (Present requirement)

IITM, Pune (16 TB) NCMRWF, Delhi 4 TB INCOIS, Hyderabad 1 TB IMD, Delhi 2 TB

NKN Connectivity 1 GBPS 10/40 GBPS Earth Science Grid

5 TB 1 TB

0.5 TB

Data Transfer among MoES Institutes/Day (Projections)

IITM, Pune (16 TB) NCMRWF, Delhi 4 TB INCOIS, Hyderabad 1 TB IMD, Delhi 2 TB

NKN Connectivity 10/40 GBPS Earth Science Grid

10 TB 4TB

2 TB

Universities & Academic Institutes

Estimated Data to be generated by planned MoES HPC Systems (2012-2015)

IITM, Pune 30 Petabytes NCMRWF, Delhi 10 Petabytes INCOIS, Hyderabad 1 Petabyte IMD, Delhi 5 Petabytes

NKN Connectivity 1 GBPS 10/40 GBPS Earth Science Grid Internet 2 US-India-EU (Monsoon Mission)

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Data Transfer requirements at MoES

 Initial Conditions Required for Making Seasonal and Extended range predictions are being prepared at INCOIS and NCMRWF, however, the real predictions are made at IITM super computer. The data required for the prediction is required to be transferred from INCOIS and NCMRWF to IITM.  Similarly, the observed data being collected at IMD is transferred to NCMRWF to prepare initial data for prediction.  Out of the Estimated MoES HPC Data of 36 Petabytes approx 25 % are Model outputs which are required to transfered to and fro between MoES Institutes for research collaborations, validations and forecasting purposes.

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Plans through NKN

 Connecting MoES HPC clusters and to consolidate the computational resources thus forming an Earth Science HPC grid.  To share HPC data through common shared file systems in enabling deduplication of Model data as well as computational time for similar runs.  Extending MoES HPC facilities to other Academic and Research Institutes through NKN.  Other IP based Applications like VOIP, MoES Intranet based applications, automations, training..etc

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Enabling Earth Science NKN Grid for MoES. More number of Internet Public IP’s for International Collaborations (At least 256) per Institute. Onsite support from NIC/NKN team for flawless, seamless migration on to NKN .

Requirements through NKN

First annual NKN workshop, IIT-B, 31 Oct. -2 Nov., 2012

Thank you