Earth System Modeling for Monsoon Projections

1. National Monsoon Mission Scoping Workshop Indian Institute of Tropical Meteorology 11 – 15 April, 2011 Towards Earth System Modelling for Monsoon Projections Under Changing Climate – Based on CFS R. Krishnan Centre for Climate Change Research Indian Institute of Tropical Meteorology, Pune

2. Increase in Surface Temperature Observations Predictions with Anthropogenic/Natural forcings Predictions with Natrual forcings 1.0º C IPCC 2007

3. Challenges in assessment of future changes in South Asian monsoon rainfall • Wide variations and uncertainties among the IPCC AR4 models in capturing the mean monsoon rainfall over South Asia (eg., Kripalani et al. 2007, Annamalai et al. 2007). • Systematic biases in simulating the spatial pattern of present-day mean monsoon rainfall (eg., Gadgil and Sajani, 1998; Kripalani et al. 2007) • Realism of present-day climate simulation is an essential requirement for reliable assessment of future changes in monsoon

4. South Asia (5-35N, 65-95E)) Source: Kripalani et al. 2010

5. Summer monsoon precipitation IPCC models: 20C3M 1979-1998 Observed rainfall (JJAS) The 20c3m simulations attempt to replicate the overall climate variations during the period ~1850-present by imposing each modeling groups best estimates of natural (eg., solar irradiance and volcanic aerosols) and anthropogenic (eg. GHG, sulfate aerosols and ozone) during this period. Seven 20C3M models (GFDL CM2.0, GFDL-CM2.1, MPI-ECHAM5, MRI, MIROC3-HIRES, HadCM3, NCAR-PCM – Source: J. Shukla)

6. Questions : On Attribution? How much of the observed variability of the mean Indian Summer Monsoon rainfall due to Climate Change? How much of the observed increase in temperature over India been decreased by increasing presence of aerosols? Questions : On Projections of Monsoon • What will happen to the monsoon hydrological cycle 50-100 years from now under different scenarios? In particular, will the quantum of seasonal mean rainfall increase or decrease and if so by how much? • What is the uncertainty in these projections? Can we quantify this uncertainty? • How can we reduce this uncertainty?

7. Some indicators of regional monsoon climate • Observed changes in frequency of monsoon depressions during the last century • Changes in the observed extreme rainfall events during the 20th century • Question: • Attribution: How much of the observed regional monsoon variability is due to global warming?

8. All India summer monsoon rainfall variability Climatological Mean (JJAS) Interannual Variability Goswami et al., Science, 2006

9. Time series of count over CI Low & Moderate events Heavy events (>10cm) V. Heavy events (>15cm)

10. Time series of frequency of monsoon depressions • Decreasing frequency of monsoon depressions during last 2-3 decades (eg., Rajeevan et al., 2000; Amin and Bhide, 2003; Dash et al., 2004) • Recovery in the activity of monsoon depressions during the recent years (2005 – 2007) • Activity of monsoon depressions modulated by low-frequency variability of atmospheric large-scale circulation on inter-decadal time-scales

11. Strategy on Regional Climate Change Research at IITMCentre for Climate Change Research (CCCR)Ministry of Earth Sciences, Govt. of India • To build capacity in the country in high resolution coupled ocean-atmosphere modelling to address issues on Attribution and Projection of regional Climate Change • Earth System Model (ESM) • To provide reliable input for Impact Assessment studies • Dynamic downscaling of regional monsoon climate using high resolution models; quantification of uncertainties • Observational monitoring: Network with other Institutions

12. Earth System Model (ESM) development • Start with an atmosphere-ocean coupled model which has a realistic mean climate – eg. NCEP CFS • Fidelity in capturing the global and monsoon climate • Realistic representation of monsoon interannual variability • Features of ocean-atmosphere coupled interactions • … • Include components of the ESM • Aerosol and Chemistry Transport Module • Biogeochemistry Module (Terrestrial and Marine) • … • … • .

13. Plan to include ESM components in the CFS-2 coupled ocean-atmosphere model • CFS-2 coupled ocean-atmosphere model simulations on HPC initiated • Ocean Biogeochemistry Module coupled to MOM4. Runs are ongoing on HPC • Aerosol Transport Module coupled to AGCM. Runs are ongoing on HPC Ongoing efforts towards development of Earth System Model (ESM) to address the Scientific Challenges of Global Climate Change and the Asian Monsoon System Basic structure of ESM

14. Climatological (JJAS) mean monsoon rainfall from CFS model – 100 year free run Climatological (JJAS) mean SST from CFS model – 100 year free run

15. Taylor diagram of spatial pattern of climatological seasonal mean (JJAS) rainfall CFS Model High pattern correlation with observed rainfall over India (IMD gridded Dataset) Source: Seasonal Prediction Group, IITM

16. CMAP precip JJAS (1980-2009) CFSv2 precip JJAS 10 yr mean CFSv1 precip JJAS 100 yr mean CFSv2 runs on PRITHVI by CCCR Source: Roxy Mathew

18. Interannual variability of summer monsoon rainfall in the CFS model – 100 year free run Domain: 70E-90E; 10N-30N Time in years CFS model JJAS climatological mean rain rate = 5.80 mm / day (red line) Standard Deviation of JJAS rain rate = 0.82 mm / day Observed rainfall (IMD) JJAS climatological mean rain rate = 7.5 mm /day Standard Deviation = 0.85 mm / day

19. M O N S O O N A L D R O U G H T S Complex Interactive Mechanisms of Monsoon Droughts Surface Boundary Conditions Land Surface Process SST Eurasian Snow Cover ENSO Cycle Interactive Dynamics Other Possible Causes Low Frequency Intra-seasonal 30-50 day scale Solar Volcanic Anthropogenic North Ward Moving Episodes East Ward Moving Episodes Stratospheric ? Synoptic Scale <One Week Source: Sikka, 1999 S.H. mid Latitudes Indian and West Pacific Ocean N.H. mid Latitudes

20. Monsoon droughts in CFS model • Atmosphere – Ocean coupling in the tropical Indo-Pacific sector • Tropical central-eastern Pacific (El Nino / Modoki) • Eastern equatorial Indian Ocean (Negative IOD) • Monsoon and mid-latitude interactions

21. Monsoon drought composites in CFS model: Rainfall and wind (850 hPa) anomalies SST and wind (850 hPa) anomalies in the CFS model • Anomalous warming of • the tropical Indo-Pacific: • Atmosphere – Ocean Coupled Interaction • Bjerknes-type feedback • El Nino - Modoki • Negative IOD

22. Monsoon-midlatitude interactions during droughts over India Rainfall and 850 hPa winds Droughts emanate from prolonged monsoon-breaks Decreases meridional temperature gradient; dry winds decrease convective instability, suppress convection and weaken monsoon Suppressed monsoon convection over sub-continent forces Rossby wave response extending over sub-tropics and mid-latitudes Winds & temperature: 500 hPa Cold air advection from mid- latitude westerly troughs cools middle and upper troposphere Rossby response: Anomalous Troughs over West-Central Asia and Indo-Pak; a stagnant blocking ridge over East Asia Wind anomalies: 200 hPa Anomaly composites during monsoon droughts: CFS model Monsoon and Mid-latitude Interactions Krishnan et al. J. Atmos. Sci., 2009

23. ESM components • Strong Monsoon Seasonal Cycle: Unique feature of tropical Indian Ocean • Aerosols transport • Terrestrial and Marine Ecosystem • Quantify the effects of aerosol variability on the South Asian Monsoon - Aerosol transport module • Quantify the climate response to variations in the regional ecosystem - Ecosystem and biogeochemical modeling

24. Marine phytoplankton absorb sunlight within the 350 - 700 nm spectral range and thereby modulate heat flux in the upper ocean Ecosystem response in the tropical Indian Ocean is linked with the seasonal cycle of monsoon winds and Indian Ocean circulation Variations in ocean biology influenced by climate phenomena - El Nino, IOD Marine ecosystem and biogeochemistry modelling Phytoplankton – Zooplankton – Detrius – Nutrient food web model (Doney et al. 1996, Fasham et al. 2001, Moore et al. 2001, 2004)

25. SST and currents) SST and currents) Chlorophyll Chlorophyll MOM4p1 forced ocean simulation – 120 year spin up Physical and Biogeochemical Parameters for Tropical Indian Ocean January July January July Source: Aparna, Swapna

26. Source: John Dunne, GFDL

27. Verification of MOM4p1 forced runs for Tropical Pacific: CORE forcing (1959 – 2004) Reproducibility of the simulated SST and Chlorophyll variability Analysis of MOM4P1 runs by Raghu Loop 5: (1959 – 2004) SST Anomaly, EOF1, (59.89 %): (1948 – 1992) MOM4p1 runs and analysis made at CCCR – Aparna and Swapna

28. Analysis of MOM4P1 runs by Raghu MOM4p1 runs and analysis made at CCCR – Aparna and Swapna

29. Modeling the effects of aerosols on the South Asian monsoon • GHG have contributed to a total of 2.7 Wm-2 to the greenhouse effect since 1850. The contribution of aerosols & other non-greenhouse factors is less certain (IPCC AR4, 2007) • Contrasting views on the impact of anthropogenic aerosols on South Asian monsoon climate • Weakening of monsoon circulation via., solar dimming (eg. Ramanathan et al. 2005) • Elevated Heat Pump (EHP) theory: Aerosol heating over Tibetan Plateau during pre-monsoon months would intensify the monsoon Hadley circulation (eg. Lau et al. 2006) • EHP hypothesis is not supported in CALIPSO Lidar satellite observations (Kuhlmann and Quass, 2010)

30. Desert air and aerosol incursions during dry Indian monsoon spells - TN Krishnamurti et al. 2010 Vertically integrated lower tropospheric (950 – 700 hPa) specific humidity (kg / kg) Dry monsoon spell: 10 – 19 June, 2009 Wet monsoon spell: 14 – 20 July, 2009 Dry air 01 Aug 2004 18 June 2009 14 Aug 2005 16 July 2002

31. 10-19 June 2009 Winds 700-300 hPa Advection of dry air & aerosols from extra-tropics into Indian region: TN Krishnamurti et al. 2010 Composite of AOD @ 550 nm during monsoon breaks – MODIS Terra / Aqua West Asia Blocking High V. Ravi Kiran, M. Rajeevan, V.B. Rao and N.P. Rao, GRL, 2009 Break monsoon Active monsoon High AOD over Arabian Sea High AOD Indo Gangetic plains

32. Modeling the effects of Aerosols on the South Asian Monsoon HAM (Hamburg Aerosol Module) Predicts evolution of an ensemble of 7 interacting internally and externally-mixed aerosol modes. Compounds considered are Sulfate, Black Carbon, Organic Carbon, Sea Salt, Mineral Dust Main Components • Microphysical core (coagulation of aerosols, condensation of gas-phase SO2 on aerosol surface) • Aerosol radiative properties & sink processes • Aerosol wet deposition • Emissions of mineral-dust (based on 10 m winds) • Sea salt emissions (based on 10 m winds) • Sulfur cycle (Inputs monthly oxidant fields ) • Emissions of DMS (Inputs DMS sea-water concentrations; calculations using 10m winds, SST) • Terrestrial biogenic DMS emissions are prescribed • AeroCom inventory for all other compounds

33. AGCM with aerosol transport Rainfall & low level winds (JJAS) Aerosol minus Control Very slight increase in precipitation over monsoon trough region Simulations of HAM coupled to an AGCM – on PRITHVI ,CCCR, IITM

34. AOD and 850 hPa winds JJAS temperature anomaly at 925 hPa Aerosol minus Control

35. With aerosols Mean = 12.25 mm / day Stdv = 0.95 mm / day No aerosols Mean = 12.09 mm / day Stdv = 0.85 mm / day Time series of the JJAS seasonal mean monsoon rainfall averaged over the region (70E – 100 E; 10N – 30N) based on the 22 member ensemble realizations for the two AGCM experiments (a) With aerosols (b) No aerosols

36. Aerosols No Aerosols Frequency distribution of rainfall over monsoon region (70E – 95E; 10N – 30N) Frequency Rainfall (mm / day) Spatial map of difference in frequency of rainfall events ( < 50 mm / day) between the Aerosol and Control simulations. The daily rainfall at each grid point covers the June to September monsoon season for 22 years

37. Break monsoon anomaly composites Rainfall No Aerosols Rainfall With Aerosols V. Ravi Kiran, M. Rajeevan, V.B. Rao and N.P. Rao, GRL, 2009 Active Break AOD High AOD Arabian Sea High AOD Indo Gangetic plains With Aerosols Summary: Conducted two sets of 22 member ensemble AGCM runs with and without aerosols. Both runs use prescribed monthly climatological SST The frequency of low & moderate rainfall events over the Indo-Gangetic plains shows very slight increase in Aerosol run as compared to no-aerosol simulation Interannual variability of monsoon rainfall in the 2 experiments is uncorrelated AOD changes during break monsoons in the GCM are broadly consistent with the observed AOD anomaly pattern Monsoon internal dynamics is dominantly seen. Role of aerosols in affecting the monsoon rainfall and circulation is being investigated.

38. High resolution regional climate change scenarios and quantification of uncertaintiesProvide reliable inputs for impact assessments and contribute to IPCC AR5 • High resolution dynamic downscaling of monsoon: Baseline climate runs using WRF, RegCM and LMDZ partially completed. Future climate scenario runs to be initiated in January 2011. • Two member 19 year (1989 : 2007) run of WRF (50 km) model completed. ERA Interim LBC • One member 19 year (1989 : 2007) run of RegCM (50 km) model completed. ERA Interim LBC • One member 10 year (1979 : 1988) run of LMDZ (50 km) model completed

39. LMDZ global atmospheric model: Variable resolution with zooming capability Source: Sabin, CCCR

40. High resolution monsoon simulations: Global model with zoom over monsoon domain JJAS rainfall JJAS SLP and winds 850 hPa LMD model 1 degree (Global) LMDZ model 1/3 degree zoom for Monsoon Domain (40-110E; 15S-30N) & 1 degree outside Initial runs made at CCCR on PRITHVI, IITM Source: Sabin

41. High resolution ( ~ 35 km) dynamical downscaling simulations using LMDZ over South Asia: Proposed at CCCR Historical (1890-2005): Includes natural and anthropogenic (GHG, aerosols, land cover etc) climate forcing during the historical period (1890 – 2005) ~ 106 years Historical Natural (1890 – 2005): Includes only natural climate forcing during the historical period (1890 – 2005) ~ 106 years RCP 4.5 scenario (2006-2100): Future projection run which includes both natural and anthropogenic forcing based on the IPCC AR5 RCP 4.5 climate scenario . The evolution of GHG and anthropogenic aerosols in RCP 4.5 scenario produces a global radiative forcing of + 4.5 W m-2 by 2100

42. Hydrologic Impacts of Climate Change Source: Deepashree Raje • Large scale hydrologic impacts • Continental water balance projections • Water resource projections for India • Large scale water mass variation • River discharge projections • Runoff computation from GCMs with global river channel network • Lateral movement of water • River flow model for river routing • Runoff and discharge projections using macroscale hydrologic models – Eg. Water Gap Global Hydrology Model (WGHM) • Global datasets for landuse, soil types, vegetation cover • River routing linear models • Validation with global datasets: Global Runoff Data Center (GRDC) • Uncertainty modeling and decision support • Multimodel ensemble of GCM simulations • Uncertainty quantification of key hydrologic parameters at large scales • Exploring adaptation and mitigation measures

43. Macroscale hydrologic modelingHydrologic Impacts of Climate Change • Variable infiltration capacity (VIC) macroscale model (Liang et al., 1994) at 1 deg. by 1 deg. resolution over Indian region • subgridvariability in land surface vegetation classes • subgrid variability in the soil moisture storage capacity, which is represented as a spatial probability distribution • subgridvariability in topography through the use of elevation bands • spatial subgrid variability in precipitation • Simulation of water balances for 1 deg x 1 deg grids • Inputs are time series of daily or sub-daily meteorological drivers (e.g. precipitation, air temperature, wind speed) • Land-atmosphere fluxes, and the water and energy balances at the land surface, are simulated at a daily or sub-daily time step • Daily runoff and baseflow routed using independent routing model (Lohmann et al., 1996) • Calibration and validation of model using current meteorologicforcings and observed discharge data in three river basins (Narmada, Ganga, Krishna)

44. Data sources • Digital elevation data at 1 km (USGS) • Land cover classification (UMD) at 1 km (UMD) • Soil texture mapping at 0.0833 deg (FAO) • Hydrologic data analysis using GIS • Basin delineations • Flow directions and accumulations at fine and coarse resolution (1 deg) grids • Area-weighted derived soil property maps at 1 deg • Area-weighted landcover maps at 1 deg • Discharge validation with global datasets : Global runoff data center (GRDC)

45. DEM for study region River routing for study basins

46. Soil Mapping Unit (SMU) classification Land cover classification

47. Preliminary results : Calibration Observed discharge m3s-1 Calibration using data from four discharge stations : Farakka (Ganga), Jamtara (Narmada), Garudeshwar (Narmada), Vijayawada (Krishna) for years 1965-1970 Source: Deepashree Raje

48. Preliminary results : Testing Observed discharge m3s-1 Testing for years 1971 - 1975

49. Features of Dynamic Climate Data Portal CCCR Climate Data Web Portal http://www.cccr.res.in CCCR Visualize data with on-the-fly graphic Easy and user friendly analysis of climate data through graphical display on the browser with one click Example : IMD daily rainfall (1951 to 2009) URL: http://cccr.hpc:8080/CCCR Step 1: Click on the above URL Centre for Climate Change Research Indian Institute of Tropical Meteorology, Pashan , Pune – 411 008 Ministry of Earth Sciences, Govt. of India

50. Summary • Long term plans (~ 3 years) to develop an Earth System Model (ESM) • A global atmosphere-ocean coupled model (CFS) is operational. A century long simulation and several other runs have been performed • Aspects of global and regional monsoon climate are realistically captured by CFS model • Realistic features of monsoon interannual variability is seen from the CFS simulations (e.g., Atmosphere-ocean coupling over tropical Indo-Pacific, Monsoon and mid-latitude interactions, etc) • Plans to improve the simulation of present day monsoon climate in the CFS model. Need to reduce model systematic biases. • Ongoing efforts to include ESM components in CFS model (ie., Aerosol transport module, Marine and Terrestrial Ecosystem and Biogeochemistry module, Sea-Ice module, etc). • Dynamic downscaling of regional monsoon climate using high resolution models • Downscaling simulations of present day and future monsoon climate scenarios will be completed by early 2012 • Contribute to IPCC AR5 report through its activity • Quantify uncertainties in regional monsoon projections using results from multiple models • Share model data and conduct inter-disciplinary collaborative research towards impact assessment, vulnerability and adaptation. • Hydrological Modeling recently started