Prediction of Indian Summer Monsoon on Seasonal and Intraseasonal Timescales: IITM Initiatives

1. Prediction of Indian Summer Monsoon on Seasonal and Intraseasonal Timescales: IITM Initiatives B N Goswami Indian Institute of Tropical Meteorology

2. There is great deal of interest on the long range forecast of the Seasonal mean due to its policy implications and dependence of economy on it. • A severe drought still influences the GDP by 2-5% • Hence, it is most important to predict the extremes, the droughts and floods! • Also, during extremes like severe droughts and floods, the rainfall anomaly over the country is homogeneous and hence even the All India mean Rainfall is useful for agricultural planning The Need and What Can be Predicted

3. On seasonal mean time scale, only large spatial scales are predictable. Hence, we expect to have skill in predicting only the All India mean rainfall. • However, during ‘normal’ monsoon years All India mean is not a meaningful quantity. • And 70% of the time, the Indian monsoon is ‘normal’. • Hence, a prediction of a ‘normal’ all India rainfall may have a comfort factor, but not useful for agricultural planning. • 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

4. Seasonal rainfall anomaly in a flood, normal and a drought years (Xavier and Goswami, 2007)

5. Active-break spells (cycles) Daily rainfall (mm/day) over central India for three years, 1972, 1986 and 1988 The smooth curve shows long term mean. Red shows above normal or wet spells while blue shows below normal or dry spells

6. A two pronged strategy • To set up a Dynamical Seasonal Prediction System at IITM • To develop a system for extended range prediction of Active-Break phases of the monsoon

7. During 7 years (including 2009) error is ≥ 10% with highest during 2002 (20%) and 1994 (18%). Error during 2009 was 15%. Average Abs Error of Op. forecasts (1988-2009) = 7.5% Performance of Operational Forecast for All India Summer Rainfall (1988-2009)

8. Tends to predict ‘normal’. Can not predict extremes • No improvement of skill in 30 years Gadgil et al, 2005, Curr. Sci

9. El Nino region(10oS-5oN, 80oW-180oW) WNP (5-30oN, 110-150oE) Asian-Pacific MNS (5-30oN, 70-150oE) Skill of Various Models in Predicting the Asia-Pacific Monsoon System Models have become good in predicting rainfall over the El Nino region, but they fail over the Asian monsoon region!

10. Correlation bet. Prediction and observation of Precipitation Current Dynamical models have little skill in predicting Indian monsoon There is a great need to improve this!!

11. Wang et al, Climate Dynamics, MME based on predictions from 14 models

12. Why all models have poor skill of monsoon prediction? • Bias in simulating the climatological seasonal mean • Teleconnection with ENSO • ‘Internal’ Interannual variability • Local air-sea interactions over warm pool

13. Development of Multi Model Ensemble Seasonal Prediction System for Indian Monsoon Model Outputs from the following models will be obtained from respective agencies for MME The core model system on which developments will be carried out will be the CFS system of NCEP. Initially we start with the CFS1.1 and replace it by CFS2.0 by the end of 2010.

14. Multi Model Ensemble approach to be used in IITM • As a starting step the following MMEs will be implemented this year • Simple Composite MME scheme – SCM • Probabilistic MME scheme – GAUS • Model weights inversely proportional to the errors in forecast probability • associated with the model sampling errors • A parametric Gaussian fitting method for the estimate of tercile-based • categorical probabilities i.e Above Normal(AN), Near-Normal(NN) and • Below-Normal(BN). • Coupled models that will be used in MME: POAMA (BMRC), • NCEP-CFS T61 (NCEP), • NCEP-CFS T126 (IITM) • CCSM3 (APCC)

15. Why CFS Model System? Through the NOAA-MoES MoU Institutional support from NCEP will be available. For predicting monsoon rainfall, skill of no coupled model is good. However, amongst the existing model systems, skill of CFS seems to be on the better side. It also has a reasonable monsoon climatology Appears to be a system upon which future developments could be built

16. CFS T62L64 ISMR and IMD corr (April IC and IMD)= 0.56 corr (March IC and IMD)=0.42 corr (May IC and IMD)=0.43

20. CFS T126L64 • The NCEP CFS Components • T126/64-layer version of the CFS • Atmospheric GFS (Global Forecast System) model • – Model top 0.2 mb • – Simplified Arakawa-Schubert convection (Pan) • – Non-local PBL (Pan & Hong) • – SW radiation (Chou, modifications by Y. Hou) • – Prognostic cloud water (Moorthi, Hou & Zhao) • – LW radiation (GFDL, AER in operational model) • GFDL MOM-3 (Modular Ocean Model, version 3) • – 40 levels • – 1 degree resolution, 1/3 degree on equator

21. HPC Facilities at IITM

22. Model comparison with TRMM 0.25 deg. Rainfall dataset TRMM rainfall (cm) Realistic simulation of rainfall over Western Ghats. Spreading of rainfall Into eastern Arabian Sea still remains in T126 NCEP-CFS rainfall (cm)

23. Model comparison with TRMM 0.25 deg. Rainfall dataset ISO Variance in the model is reasonably well simulated.

24. 9 Member Ensemble Hindcast for 17 years with the CFS T126L64 System has been completed at IITM CC=0.44

25. The Monsoon Mission Focused mission on ‘Seasonal and Intra-seasonal Monsoon Forecast’ has been launched. Support focused research by national and international research groups with definitive objectives and deliverables to improve CFS model. Support some specific observational programs that will result in improvement of physical processes in climate models.

26. Improving Prediction of Seasonal Mean Monsoon It is important that all development work should be done a specified model Coupled Model CFS 1.1 or 2.0 Model Development & Improvement in Physical Parameterization Basic Research Data Assimilation

28. Model Development Improvement of Physical Parameterization Diagnostics of Model Biases Resolution Super Parameterization ?

30. Proposed modalities to achieve mission objectives IITM to coordinate the effort. Proposals to be invited from National as well as international Institutes on very specific projects and deliverables through which improvement of the CFS model are expected. Provisions for funding the National partners as well as the international partners will be year marked. The Proposal partners will be allowed to use the HPC facility at IITM which will be suitably enhanced for this purpose. Funding for students, post docs and some scientists time (consultancy) and some minor equipments may be provided.

31. Up gradation Plan for Computing Facility at IITM Existing HPC System is being upgraded with additional 101 IBM P575 nodes with 60.2 TF peak power to achieve more than 70TF peak performance. Additionally 4 high end servers, 10 workstations, 144 ports IB switch! Internet bandwidth is being upgraded to 100 Mbps!

32. Extended Range Prediction of Monsoon ISOs

33. The ISO signal is much larger than signal in IAV of monsoon Amplitude of (s.d.) of interannual variability of JJAS precipitation (mm/day) , (middle) Amplitude of intraseasonal variability (s.d. Of 10-90 day filtered anomalies during June 1 – Sept. 30) and (bottom) climatological mean JJAS precipitation (mm/day).

35. How Predictable are the Monsoon ISOs? • A simple procedure is described to make such an estimate of potential predictability for active and break conditions from observations Goswami and Xavier, 2003, GRL

36. Regions over which potential predictability of precipitation is examined

40. Conclusion: ISO Predictability The transition from break to active conditions is intrinsically more chaotic than transitions from active to break conditions. A fundamental property of monsoon ISOs. Why? Break Active – convective instability— growth of error governed by fast conv. Instability ActiveBreak – dying convection -- slow error growth due to slow oscillation Consequence, The potential predictability limit for monsoon breaks is about 20 days while that for monsoon active conditions is only about 10 days

41. A new Analogue model for Extended Range prediction of the summer Monsoon ISOs useful for real time prediction (Xavier and Goswami, 2007, MWR , 135, 4149-4160) • Data Used and methodology: • Penrad OLR data from 1979 to 2004 are used. No filtering is involved. • To isolate the ISO, OLR anomalies are reconstructed with the first 10 EOFs • First spatial analogues are identified and temporal analogue for each PC is found from the cases selected for spatial analogues.

42. Spatial analogs Initial condition Identify spatial analogues over the entire period (25 years) for the OLR field reconstructed with first 10 EOFs at to Identify temporal analogues for all 10 PCs going 5 pentads backward from each of the spatial analogues Prediction average evolution of the identified temporalanalogues

43. Illustration of a few temporal analogues of PC2 and PC3 for a particular initial condition.

44. t0 t1 Modelling period 2000 Hindcast period 1979 2005

45. Anomaly hindcasts • Reconstruct the OLR data with 2-10 EOFs • ( removes the seasonal cycle) • Look for spatial and temporal analogues as before • Make predictions based on average evolution of these analogues

47. This model is currently being used by the India Meteorology Department to make experimental Extended Range prediction of monsoon ISO this year. Please see the following website www.imdpune@gov.in

48. A Nonlinear Model for Real-Time Extended Range Prediction of Active/Break Cycle over Central India Chattopadhyay,Sahai,Goswami, 2008, JAS

49. We propose that the dominant monsoon ISO is nonlinear Convectively coupled oscillation • Thus a unique relationship between a number of dynamical fields representing a nonlinear phase of the oscillation should be uniquely related a phase of the rainfall oscillation

50. RF data used for verification Model 1 Model 2