Annual Joint WMO Technical Progress Report on the Global Data Processing and Forecasting

Annual Joint WMO Technical Progress Report on the Global Data processing and Forecasting System (GDPFS) including Numerical Weather Prediction (NWP) Research Activities March 2010

(A) Contribution of India Meteorological Department

1. Summary of highlights

The main mandate of the Numerical Weather Prediction NWP Group at the Head Quarters (New Delhi) of IMD has been to support for the day to day operational weather forecasts in the time scale of nowcasting to medium range. This involves processing of observations received through Global Telecommunication System (GTS) and automatic preparation of various sysnoptic charts, operational run of NWP models and preparation of various graphics products, dissemination, validation of model performance and R&D, data updates for IMD Web site, archival of all data and forecast outputs.

India Meteorological Department operationally runs number of regional models, namely, Limited Area Model (LAM), MM5, WRF and Quasi-Lagrangian Model (QLM) for short range prediction. The MM5 model is run at the horizontal resolution of 45 km with 23 sigma levels in the vertical and the integration is carried up to 72 hours over a single domain covering the area between lat. 30 o S to 45 o N long 25 o E to 125 o E. Initial and boundary conditions are obtained from the National Centre for Medium Range Weather Forecasting (NCMRWF) T-254 Global Forecast System (GFS). The boundary conditions are updated at every six hours interval. WRF is run for 72 hours forecast at the horizontal resolution of 27 km and 38 vertical levels domain covering the area between lat. 30 o S to 45 o N long 25 o E to 125 o E with NCEP initial and boundary condition. The resolution

1 of inner domain is 9 km covering India. The QLM model is used for cyclone track prediction in case of cyclone situation in the Arabian Sea or Bay of Bengal.

Considering need of farming sector, India Meteorological Department (IMD) has upgraded the Agro-Meteorological Advisory Service from agro climate zone to district level. As a major step, IMD started issuing district level weather forecasts from 1 June 2008 for meteorological parameters such as rainfall, maximum and minimum temperature, relative humidity, surface wind and cloud octa up to 5 days in quantitative terms. These forecasts are generated through Multi-Model Ensemble (MME) system making use of model outputs of state of the art global models from the leading global NWP centres. These forecasts are made available on the national web site of IMD. The method has been further updated for monsoon 2009 from the use of five NWP models namely, (i) NCMRWF T-254, (ii) ECMWF T799, (iii) JMA T859, (iv) UKMO and (v) NCEP GFS T-382.

These NWP products are disseminated to the operational forecasters at various IMD Forecast Centres/Offices through a ftp connectivity.

With the commissioning of High Performance Computing System (HPCS) at IMD H/Q in January 2010, IMD has been in the process of expanding NWP activities to meet the growing operational demands of multiscale forecasts ranging from nowcastng to medium range and extended range. Global, regional and mesosale NWP models with state of the art data assimilation procedure are expected to be made operational at H/Q with in a few moths. At twelve other regional centres, very high resolution mesoscale models will be made operational under the guidance of H/Q. From the ongoing modernization programme of IMD, observations (both conventional and non conventional) are expected to be available on the mesoscale both in space and time by means of Doppler Weather Radar (DWR), Satellites (INSAT Radiance), Wind Profilers, meso-network

2 (Automatic Weather Stations), buoys and aircrafts in the real time mode with the use of advanced telecommunication system.

2. Equipment in use at the Centre

High Performance Computing System (HPCS) with peak speed 14,2 Tera Flop was commissioned in IMD New Delhi in January 2010. High end servers at 12 different locations across the country (Pune; Regional Met. Centres Delhi, Kolkata, Chennai, Mumbai, Guwawati and Nagpur; Met. Centres Ahmedabad, Bangalore, Chandigarh, Bhubaneswar and Hyderabad) are installed (10 completed and 2 under progress).

Computing Racks with peak Power : Peak Speed 14. 4 Tera FLOPS 28 Nodes : POWER-6, 4.7 GHz Processors &128 Giga Bytes Memory per Node Storage : 300 Tera Bytes (100 TB online and 200 TB near online) Archival : 200 Tera Bytes Operating Environment : IBM-AIX 5.3 with Parallel Computation Support Network Bandwidth : 10 Gbps for Switching (Clustering)

4 High End Servers with a total Computing Power (134 GF x 4) = 536 G FLOFS 8 Racks for Storage 1 Rack of Robotic Tape Library

Computer System: (a) Altix- 350 (b) 0rigin 200 and (c) IBM P5/595 (64 processors).

3. Data and Products from GTS in use

Data management at IMD H/Q is comprising of four stages namely, (a) Reception of data through the Global Telecommunication System (GTS) from RTH (b) Processing of observations for various operational use (c) Disposal of final products and (d) Archival of data. Reception of data includes of real time global weather observations and NWP outputs of other operational NWP centres, llike ECMWF and JMA (in the GRIB format)

3 (a) Reception: Meteorological observational data received on line at Regional Telecom Hub (RTH) round the clock are being used in Northern Hemisphere Analysis Centre (NHAC) at the H/Q of IMD in two different channels namely: (i) Manual Plotting of operational Synoptic Weather Charts, (ii) Processing of data (Decoding, Quality Control) in the NHAC Computer system for Automatic Plotting of weather charts for synoptic use and ingesting of data in the assimilation cycle of NWP models. Manual synoptic weather charts are preserved (b) Processing of data in NHAC Computer System: Automatic Plotting of synoptic charts are done at every three hours interval where as plotting of AWS observations are done at hourly interval. Synoptic Observations are also used for preparation of various display tables (like Current Weather ) and Graphics as required by the IMD web site, which are made updated on the basis of latest observation. At IMD H/Q, NWP models are run based on 00 UTC and 12 UTC observations. These models are run based on the initial and boundary conditions from NCEP GFS (received online through the Internet)/NCMRWF T-254 (received online through RTH). Based on the model outputs various NWP graphics products are prepared as required for the operational forecasting. ECMWF model outputs are decoded and then graphics product are prepared. (c) Disposal of Final Product: All the automatic plots of weather charts and NWP graphics outputs are used for the operational forecasting at NHAC. Display tables and graphics product (based on observations as well as NWP outputs) are used for IMD web site. Some of the products are kept in IMD ftp site for access to the field forecasters at MC/RMC/MO. (d) Archival data: Following data are being archived at NHAC (Computer): (i) GTS observations (ii) Decoded synoptic and upper air observations (iii) NWP outputs both graphics and digital data and (iv) Automatic Plots of weather charts. Manual synoptic charts are also preserved at NHAC.

Data received from regional telecommunication hub (RTH) to NHAC computer through GTS channel are WMO specific code format. They are further processed for plotting as well as used for model input. Besides these data, initial and boundary fields are regularly downloaded from NCEP/NCMRWF site to run LAM and MM5. NCEP data are also used to run QLM to give cyclone track prediction up to 72hr. JMA and ECMWF outputs are also being achieved. Graphic products of all model output are uploaded to web server of IMD

4 ( www.imd.gov.in) besides other routine observational data in tabular form. Hourly plots of Automatic Weather Station (AWS) data, 3 hourly surface data & 12 hourly upper air plotting are regularly uploaded to ftp server of IMD. Five days forecast products from UKMET office & T254L64 products received from NCMRWF are also uploaded to ftp server of IMD to facilitate field forecasting offices on day to day basis.

4. Forecasting System

(a) The High Performance Computing System (HPCS) Data Flow Diagrams :

The HPCS at IMD HQs receives the entire data including manual and automatic devices from across the globe, processes it and generates global and regional forecasts for the purpose of  Generating forecast guidance for operational offices  Generating the initial conditions for feeding very high resolution models run at regional meteorological centres

5 METEOROLOGICAL OBSERVATIONS x x

HPCS DELHI IMD PUNE GLOBAL/MESOSCALE ANALYSIS & CLIMATE MODELS MODELS FORECAST (1.0 TFlops) (14.4 TFlops)

RMC RMC RMC MC MESOSCALE MESOSCALE MESOSCALE MESOSCALE MODELS MODELS MODELS MODELS (134 GFlops) (134 GFlops) (134 GFlops) (134 GFlops)

ANAL & F/C ANAL & F/C ANAL & F/C ANAL & F/C PRODUCTION

END USER DISSEMINATION NETWORK

(b) The HPCS Connectivity :

 At the incoming end the HPCS is connected to the central message switching computer called “TRANSMET”.

 The products are seamlessly connected to the operational forecasting system of IMD called “SYNERGEE”. It directly flows through the manual value addition stages to product generation platforms which create the dissemination products.

 HPCS server feeds regional servers through automated ftp via VPN circuits. Data and products are exchanged with other national users like Indian Navy, Indian Air force etc.

System run schedule and forecast ranges

6 Medium Range Forecast System (4-10 days)

Implementation of Global Forecast System (GFS)

Global Forecast System (GFS, based on NCEP) at T382L64 resolution has been implemented at NHAC, IMD HQ on IBM based High Power Computing Systems (HPCS). In horizontal, it resolves 382 waves ( 35 Km) in spectral triangular truncation representation (T382), for which the Gaussian grid of 1152 x 576 dimensions are used. The model has 64 vertical levels (hybrid; sigma and pressure). The GFS is running in experimental real-time mode since 15th January 2010. This new higher resolution global forecast model and the corresponding assimilation system are adopted from NCEP, USA. The horizontal representation of model variables are in spectral form (spherical harmonic basis functions) with transformation to a Gaussian grid for calculation of nonlinear quantities and physics. The GFS at IMD Delhi involves 4 steps as given below :

Step 1 - Data Decoding and Quality Control: First step of the forecast system is data decoding. It runs 48 times in a day on half-hourly basis, as soon as GTS data files are updated at regional telecom hub (RTH) of global telecom system (GTS), at IMD, New Delhi.

Steps 2 - Preprocessing of data (PREPBUFR) : Runs 4 times a day at 0000, 0600, 1200 & 1800 UTC. List of data presently being pre-processed for Global Forecast System are :

1. Upper air sounding – TEMP, GPS & PILOT 2. Land surface – SYNOP, SYNOP MOBIL & AWS 3. Marine surface - SHIP 4. Drifting buoy - BUOY 5. Sub-surface buoy - BATHY 6. Aircraft observations - AIREP & AMDAR 7. Automated Aircraft Observation - BUFR (ACARS) 8. Airport Weather Observations - METAR 9. Satellite winds - SATOB

7 10.High density satellite winds - BUFR (EUMETSAT & Japan) 11.Wind profiler observations - BUFR (US/Europe) 12.Surface pressure Analysis - PAOB (Australia) 13.Radiance (AMSU-A, AMSU-B, HIRS-3 and HIRS-4, MSU, IASI, SSMI, AIRS, AMSRE, GOES, MHS) 14.GPS Radio occultation 15.Rain Rate (SSMI and TRMM)

Step 3 - Global Data Assimilation (GDAS) cycle :

The Global Data Assimilation (GDAS) cycle runs 4 times a day (00, 06, 12 and 18 UTC). The assimilation system is a global 3-dimensional variational technique, based on NCEP’s Grid Point Statistical Interpolation (GSI) scheme, which is the next generation of Spectral Statistical Interpolation (SSI).

Step 4 – Forecast Integration for 7 days

The analysis and forecast for 7 days is performed using the HPCS installed in IMD Delhi. One GDAS cycle and seven day forecast (168 hour) run takes about 30 minutes on IBM Power 6 (P6) machine using 24 nodes with 7 tasks (7 processors) per node.

Operationally available Numerical Weather Prediction (NWP) Products IMD also makes use of NWP global model forecast products of other operational centres, like NCMRWF T-254, ECMWF, JMA, NCEP and UKMO to meet the operational requirements of day to day weather forecasts in the short to medium range time scale. Under a joint collaborative research project IMD has been receiving global model outputs (in the GRIB format) of ECMWF and JMA. The outputs (GRIB) of NCEP GFS are available freely from the Internet. The model outputs of these models are post processed using GRIB

8 decoder and various graphics products are generated operationally in the real time mode. These NWP products are disseminated to the operational forecasters at various IMD Forecast Centres/Offices through a ftp connectivity. IMD receives NCMRWF T-254 and UKMO model outputs online from NCMRWF, Noida

Operational techniques for application of NWP products.

IMD implemented a Multi-model Ensemble (MME) based district level quantitative forecasts in the operational mode since 1 June 2008, as required for the Integrated Agro-advisory Service of India. The forecasts prepared daily are also made available in the IMD web site (www.imd.ernet.in or www.imd.gov.in).

Five NWP models considered for this development work are: (i) National Centre for Medium Range Weather Forecasting (NCMRWF) T-254, (ii) ECMWF T799, (iii) JMA T899, (iv) UKMO and (v) NCEP GFS T-254. As the model outputs available are at different resolutions, in the first step, model outputs of the constituent models are interpolated at the uniform grid resolution of 0.25oX0.25o lat/long. In the second step, the weight for each model at each grid is determined objectively by computing the correlation co-efficient between the predicted rainfall and observed rainfall. High resolution gridded rain-gauge data produced operationally at National Centre of IMD Pune are used for development and validation of the forecasts. The weight (Wi,j,k) for each member model (k) at each grid (i,j) is obtained from the following equation:

Ci, j,k 5 Wi, j,k , = , i = 1, 2, ….., 161; j=1,2,....,161 ……… (1) Ci, j,k k 1

9 Ci,j,k = Correlation co-efficient between rainfall analysis and forecast rainfall for the grid (i,j) of model (k). For the computational consistency, Ci,j,k is taken as

0.0001 in case Ci,j,k is less than or equal to 0.

The ensemble forecasts (day 1 to day 5 forecasts) are generated at the 0.25ox0.25o resolution. The ensemble forecast fields are then used to generate district level forecasts by taking average value of all grid points falling in a particular district.

4.2.3 Research Performed in this field

Performance skill of forecasts by these models in short to medium range time scale during summer monsoon 2009 will be presented in the AMR meeting of 2010. Some of the important verification results are given below:

 Various need based R & D activities such as, model validation, impact of new conconventional data and model resolution, various diagnostic studies, customization of NWP outputs for operational forecast are being carried out. Recently two significant outcomes of these R & D activities are:  A multi-model ensemble technique has been developed for five days weather forecasts making use of state of the art global model outputs. The method is made operational from 1 June 2008 for district level Integrated Agro-advisory services.  Development and implementation of a Statistical Dynamical method for cyclone genesis and intensity prediction

4.3 Short Range forecasting system (0 – 72 hours) 4.3.2 Model

NWP Models operational at IMD New Delhi are:

10  The Limited Area Model (LAM) forecast is being produced regularly in respect of 00 UTC and 12 UTC observations for day-to-day operational use. The operational forecasting system known as Limited Area Forecast System (LAFS), is a complete system consisting of data decoding and quality control procedures, 3-D multivariate optimum interpolation scheme for objective analysis and a semi-implicit semi-Lagrangian multi-layer primitive equation model. The horizontal resolution of the model is 0.75 ox0.75 olat. /long. with 16 sigma levels in the vertical.  The Quasi-lagrangian Model (QLM) model is run to produce track forecasts based on the initial conditions of each day based on 00 UTC and 12 UTC observations when the disturbance is in cyclonic storm stage. The QLM is a multilevel fine-mesh primitive equation model with a horizontal resolution of 40 km and 16 sigma levels in the vertical. The integration domain consists of 111x111 grid points in a 4440x4440 km2 domain that is centred on the initial position of the cyclone. Very recently, model has been updated (from 36 to 72 hours) to get six hourly track forecasts valid up to 72 hours.  The non hydrostatic mesoscale model MM5 is run at the resolution of 45 km daily with 00 UTC initial and boundary condition of NCEP GFS (National Centre for Environmental Prediction, USA; Global Forecast System).  A multi-model ensemble technique has been developed for five days weather forecasts making use of state of the art global model outputs. The method is made operational from 1 June 2008 for district level Integrated Agro-advisory services.  The mesoscale model WRF has been implemented with the assimilation of local observations.  The storm scale model ARPS (Advanced Regional Prediction System) has been experimented at the horizontal resolution of 9 km with the assimilation of Doppler Weather observations.

11  Development and implementation of a Statistical Dynamical method for cyclone genesis and intensity prediction  For Storm Surge Prediction, Dynamical Storm Surge model of IIT Delhi has been made operational. Graphics product of all these models are available in the IMD web site.

Meso-Scale Assimilation System (WRF-VAR) Recently, the regional mesoscale analysis system WRF-Var is installed on High HPCS at Head Quarter, IMD, Delhi with its all components i.e. preprocessing programs (WPS and REAL), observation assimilation program (WRF-Var), boundary condition updation (update_bc) and forecasting model (WRF).

The pre-processed observational data from GTS and other sources prepared for the Global Forecast System in the BURF format (PREPBUFR of step 2 in GFS) is also used in case of WRF assimilation.

In the WRF-Var assimilation system, all conventional observations over a domain (200S to 450N; 400E to 1150E) which merely cover RSMC, Delhi region are considered to improve the first guess of GFS analysis. Assimilation is done with 27 km horizontal resolution and 38 vertical eta levels. The boundary conditions from GFS forecasts run at IMD are updated to get a consistency with improved mesoscale analysis. WRF model is then integrated for 75 hours with a nested configuration (27 km mother and 9 km child domain) and with full physics (including cloud microphysics, cumulus, planetary boundary layer and surface layer parameterization). The post-processing programs ARWpost and WPP are also installed on HPCS to generate graphical plots and grib2 out for MFI- SYNERGIE system respectively.

12 Outer and inner domain of WRF model at 27 km and 9 km

4.3.3 Operationally available NWP Products 4.3.4 Operational Techniques for application of NWP Products

Various stages of cyclone forecasting are: (a) Genesis, (b) Track, (c) Intensity and (d) Decay after landfall. During 2008-09, IMD used an objective numerical method for the operational cyclone forecasting work. The method comprises of four forecast components, namely (a) cyclone genesis potential parameter (GPP), (b) Multi-model Ensemble (MME) technique for track prediction, (c) cyclone intensity prediction (SCIP) model and (d) predicting decaying intensity after the landfall.

13 Genesis Potential Parameter (GPP)

Genesis Potential Parameter (GPP) is defined as:

 850 xMxI GPP = if 850 > 0, M > 0 and I > 0 S

= 0 if 850 ≤ 0, M ≤ 0 or I ≤ 0

-5 -1 Where , 850 = Low level relative vorticity (at 850 hPa) in 10 s S = Vertical wind shear between 200 and 850 hPa (knots)

= Middle troposphere relative humidity Where, RH is the mean relative humidity between 700 and 500 hPa

I = (T850 – T500) °C = Middle-tropospheric instability (Temperature difference between 850 hPa and 500 hPa). All the variables are estimated by averaging of all grid points over an area of radius 2.5o around the centre of cyclonic systems using model analysis field.

GPP values for developing and non-developing systems are shown in Table 1.

Table 1. Genesis potential parameter (GPP) for Developing Systems and Non- Developing Systems.

GPP (x10-5) 

T.No.  1.0 1.5 2.0 2.5 3.0

Developing 11.1 12.3 13.3 13.5 13.6

Non-Developing 3.4 4.2 4.6 2.7 -

Various thermo-dynamical parameters, which are used for real time analyzing Genesis Potential Parameter (GPP) for cyclonic storms over the Bay of Bengal during 2008-2009, are derived from the operational model analysis of the limited area model (LAM) of India Meteorological Department (IMD), New Delhi.

14 Track : Multimodel Ensemble (MME) Technique

A multimodel ensemble (MME) technique is developed using cyclone data of 2008. The technique is based on a linear statistical model. The predictors (shown in Table 2) selected for the ensemble technique are forecasts latitude and longitude position at 12-hour interval up to 72-hour of five operational models. In the MME forecasts, model-forecast latitude position and longitude position of the member models are linearly regressed against the observed latitude position and longitude position respectively for each forecast time at 12- hours intervals for the forecast up to 72-hour. Multiple linear regression technique is used to generate weights (regression coefficients) for each model for each forecast hour (12hr, 24hr, 36 hr, 48hr, 60hr, 72hr). These coefficients are then used as weights for ensemble forecasts.

12-hourly forecast latitude (LATf) and longitude (LONf) positions by multiple linear regression technique is defined as:

f lat lat lat lat lat LAT t = ao+ a1ECMWFt + a2NCEP t +a3JMAt + a4MM5t + a5QLMt

f ’ ’ lon ’ lon ’ lon ’ lon LON t = a o+ a 1ECMWFt + a 2NCEPt +a 3JMAt + a 4MM5t +

’ lon a 5QLMt for t = forecast hour 12, 24, 36, 48, 60 and 72

The dependent variable latitude (LATf) in °N and longitude (LONf) in °E. The detailed of model predictors are given in Table 3.

Table 3. Model Parameters

S.No. Member models Symbol of Predictors Latitude Longitude position position 1. European Centre for Medium- ECMWFlat ECMWFlon Range Weather Forecasts (ECMWF), 2. GFS of National Centers for NCEPlat NCEPlon Environmental Prediction (NCEP)

15 3. Japan Meteorological Agency JMAlat JMAlon (JMA) 4. MM5 Model MM5lat MM5lon 5. Quasi-Langrangian model (QLM) QLMlat QLMlon

Intensity prediction

A Statistical Cyclone Intensity Prediction (SCIP) model for the Bay of Bengal for predicting 12 hourly cyclone intensity (up to 72 hours), applying multiple linear regression technique using various dynamical and physical parameters as predictors. The model equation is given as:

dvt = ao+ a1 IC12 + a2 SMS +a3 VWS+ a4 D200+ a5 V850+a6 ISL+ a7 SST+ a8 ISI for t= forecast hour 12, 24, 36, 48, 60 and 72

dvt = Intensity change during the time interval t

The detailed of model predictors are given in Table 4.

Table 4 Model parameters S.No. Predictors Symbol of Unit Predictors 1. Intensity change during last 12 IC12 Knots hours 2. Vorticity at 850 hPa V850 x 105 s-1 3. Storm motion speed SMS ms-1 4. Divergence at 200 hPa D200 x105 s-1 5. Initial Storm intensity ISI Knots 6. Initial Storm latitude position ISL °N 7. Sea surface temperature SST °C 8. Vertical wind shear VWS Knots

Decay of intensity after the Landfall

16 The forecast of inland wind after the landfall of a cyclone is of great concern to disaster management agencies. To address this problem, an empirical model for predicting 6-hourly maximum sustained surface winds (intensity) was developed. The maximum sustained surface wind speed (MSSW) after the landfall at time t is given by:

Vt+6 = Vb+(Vt-Vb)*R1, for t=0

= Vb+(Vt-Vb)*R2, for t=6,12,18 and 24

Where, reduction factors

R1 = exp(-a1*6.0)

and, R2 = exp(-a2*6.0)

Decay constant a1 for the first six hours after the landfall (for t= 0 to 6) is given by: a1 = [ln {(Vo –Vb)/(V6-Vb))}]/6

The decay constant a2 for the remaining 12 hours (for t= 6 to 18 hours) is taken as:

a2 = [ln {(V6 –Vb)/(V18-Vb))}]/12

Regression equation relating R1 and R2 as given below:

R2 = 0.982*R1 –0.081

Where, V0 is the maximum sustained surface wind speed at the time of landfall, Vt is the wind speed at time t after the landfall and Vb is the background wind speed. After landfall, tropical cyclone decays to some background wind speed.

4.4 Nowcasting and very short range forecasting systems (0-6 hours) For nowcasting purposes, application software called “Warning Decision Support System Integrated Information (WDSS-II)”, developed by National Severe Storm Lab, USA has been used in experimental mode. For mesoscale forecasting, radar data has

17 been assimilated into the ARPS mesoscale model. With the ingesting of Indian DWR observations, the application software is capable of detecting and removing anomalous propagation echoes. The application software could successfully track storm cells and meso-cyclones through successive scans. Radar reflectivity mosaics are created for the recent November 2009 Bay of Bengal cyclone “Khaimuk” using observations from three DWR stations namely, Visakhapatnam, Machilipatnam and Chennai. Positive impact of the radar observations in a very high resolution NWP model (ARPS) have been demonstrated for land falling cyclones.

arps_ref arps_vel A A

c c B Figure : displays a sequence of mosaic images of the tropical cyclone Khaimukh of 14 November 2008, which was tracked by the three radars at Chennai, Machhilipatnam and Visakhapatnam. B

(a) (b)

arps_both

arps_con A

A c c B B

(c) (d) 18 Fig. 8 A(a-e Fig.: Inter-comparison of reflectivity fields of various simulation experiments against the observed field valid c at 0600 UTC of 27 November 2008 for Bay of Bengal B (e ) Cyclone Nisha: (a) Reflectivity by arps_ref experiment (b) Reflectivity by arps_vel experiment (e) (c) Reflectivity by arps_both experiment 4.5 Specialized numerical prediction (d)( seaReflectivity wave, by storm arps_con surge experiment etc.) (e) Observed field from the radar station For the storm surge prediction IMD has been using a dynamical model developed by Indian Institute of Technology, Delhi. IMD also uses nomogrrams in conjunction with the dynamical model for storm surge prediction.

4.6 Extended range Forecast (10-30 days) (Model, ensemble, Methodology)

19 Extended range forecast products generated from NCEP (CFS) and ECMWF (Ensemble) are currently used for extended range forecasts over the Indian region

4.7 Long range forecast

IMD uses a statistical model for the long range forecasting of Indian monsoon. Dynamical model outputs generated from other national meteorological institutes are also used in conjunction with the statistical model for the long range forecast of Indian monsoon.

5. Verification of prognostic products

Before one uses outputs of a NWP model in preparation of final operational forecast, adequate knowledge on the performance skill of the model is a pre- requisite. Towards this direction, continues efforts are being made by the NWP group of IMD to document performance skill of operational NWP models. These documents would contribute in updating knowledge of operational forecaster for judicious use of NWP products for delivering improved operational forecasts. For example, track forecast error (km) of the multi-model ensemble forecast and member models during the year 2009 is given below.

HOUR ECMWF GFSDAY-5 :MON-2009JMA MM5 QLM MME 12 hr 72 83 86 153 77 O70 R I S S A 241 hr 111 191 167 234 124 R90 A J A S T H A N 0.936 hr 114 193 142 320 143 M147 A H A R A S T R A K E R A L A 0.848 hr 93 117 86 246 242 199 G U J A R A T 60 hr 168 126 85 351 447 242 0.7 M A D H Y A P R A D E S H 72 hr 217 151 152 415 577 293 0.6

D 0.5 O P The State-wise0.4 performance of district level rainfall forecasts for some selective states for day0.3 5 forecasts has been presented in a Fig. given below.

0.2

0.1 20 0 NO RAIN LIGHT RAIN MOD RAIN HEAVY RAIN 6. Plans for the future

Under the modernization programme, IMD is in the process of commissioning a state of the art High Performance Computing (HPC) system with a peak performance of 15 TF at IMD HQ., 1 TF at IMD Pune along with high end servers of 100 GF capacities to each in major meteorological centers viz. Delhi, Mumbai, Chennai, Nagpur, Kolkata, Guwahati, Ahmedabad, Bangalore, Bhubaneswar, Chandigarh, Hyderabad and Pune for global and regional NWP modeling, particularly for the regional database management, mesoscale data assimilation and high resolution local area model. From the ongoing modernization programme of IMD. observations (both conventional and non conventional) are expected to be available on the mesoscale both in space and time by means of Doppler Weather Radar (DWR), Satellites (INSAT Radiance), Wind Profilers, meso-network (Automatic Weather Stations), buoys and aircrafts in the real time mode with the use of advanced telecommunication system.

21 In view of growing operational requirements from various user agencies, there is a need for a seamless forecasting system covering now-casting to medium range user specific forecasts. There is also need for the improved extended range and long range forecasts, particularly for the agricultural requirements. Future Weather Forecasting System of IMD would be as briefly given below:

(a) Now-casting and Mesoscale Forecasting System (valid for half hour to 24 hours)  Processing of Doppler Weather Radar (DWR) observations at a central location (NHAC) to generate 3 D mosaic and other graphics products for nowcasting applications.  Enhancing mesoscale forecasting capability of local severe weather by providing 3 hourly area specific rainfall and wind forecasts (up to 24 hours) at the resolution of 3 km from ARPS (Advanced Regional Prediction System) with the assimilation (hourly intermediate cycle) of DWR, AWS, Wind profilers and other conventional and non-conventional observations.  UKMO based nowcast system

Nowcast and mesoscale forcast system would be expanded for major cities of India

(b) Regional Models for Short Range Forecasting System ( valid up to 3 days)  72 hours forecasts from WRF model with 3 nested domains (at the resolution of 27 km, 9 km and 3 km). The nested model at the 3 km resolution would be operated at the Regional/State Met Centres at 6 hours interval with 3 DVAR data assimilation.  For Cyclone Track Prediction, 72 hours forecast from Quasi Lagrangian Model (QLM) at 40 km resolution at six hours interval; WRF (NMM) at 27

22 km resolution with assimilation package of Grid Statistical Interpolation (GSI).  For Cyclone track and intensity prediction: multimodel ensemble technique and application of dynamical statistical approach for 72 hours forecasts, forecast would be updated at 12 hours interval.  Development of multimodel ensemble technique for probabilistic forecasts of district level heavy rainfall events.

(c) Global model for Medium range Forecasting (valid up to 7 days)  Global Data Assimilation System (GDAS), six hourly cycle with GSI (Grid Statistical Interpolation).  Global Forecast System (GFS) T-382  Global Ensemble Prediction System (d) Extended range forecast for rainfall and temperature  To implement a statistical dynamical model

7. References

Roy Bhowmik S.K. and Durai V.R., 2010, Application of multi-model ensemble technique for real-time district level forecasts over Indian region in short range time scale, Meteorl. Atmos. Phy., 106, 19-35

23 Kotal, S.D., Roy Bhowmik S.K. and Mukhopadhaya, B, 2010, Real-time forecasting of Bay of Bengal Cyclonic Storm Rashmi of October 2008 – A statistical dynamical approach, Mausam, 61, 1-10

Sen Roy Soma, Roy Bhowmik, SK, Lakshmanan, V, and . Thampi S.B., 2010, Doppler Radar-based Nowcasting of the Bay of Bengal Cyclone – Ogni of October 2006, J., Earth SCI. Sys (to appear)

Srivastava Kuldeep, Roy Bhowmik, S.K., Sen Roy, S., Thampi S.B. and . Reddy Y.K., 2010, Simulation of high impact convective events over Indian region by ARPS model with assimilation of Doppler Weather Radar radial velocity and reflectivity , Atmosfera, 23, 53-74

Roy Bhowmik S.K.,Shakar Nath, Mitra, A and Hatwar, H.R., 2009, Application of neural network technique to improve the location specific forecast of Delhi from MM5 Model, MAUSAM, 60, 11-24

Srivastava Kuldeep, Roy Bhowmik S. K., Hatwar H.R., 2009, Evaluation of different Convective schemes on simulation of thunderstorm event over Delhi by ARPS Model, Mausam, 60(2), 123-136

Kotal, S.D., Kundu P.K. and Roy Bhowmik S.K. , 2009, An analysis of cyclo- genesis parameter for developing and non-developing low pressure systems over the Indian Sea, Natural Hazards, 50,389-402

Roy Bhowmik, S.K. and Prasad K, 2008 Improving IMD operational limited area model forecasts , Geofizika, 25(2), 87-108

Roy Bhowmik S.K. and V.R. Durai, 2008, Multi-model Ensemble Forecasting of rainfall over Indian monsoon region, Atmosfera , 21(3), 225-239

Kotal, S.D., Roy Bhowmik S.K.. P.K. Kundu and Das Ananda K., 2008, A Statistical Model for Cyclone Intensity Prediction, Earth Sc. System, 117(2), 157-168

Kotal, S.D., Roy Bhowmik S.K. and Kundu Prabir, 2008,“Application of Statistical - Dynamical scheme for real time forecasting of the Bay of Bengal very severe cyclonic storm SIDR of November 2007, Geofizika, 25 (2), 139-158

Srivastava Kuldip, Roy Bhowmik, S.K., Hatwar, H.R. Ananda K. Das and Kumar Awadsesh, 2008, “ Simulation of mesoscale structure of thunderstorm using ARPS model, Mausam, 59 (1), 1-14

24 Roy Bhowmik S.K., Joardar, D. and Hatwar H.R. 2007, An evaluation of precipitation prediction skill of IMD operational NWP system, Meteorl Atmos Phy, 95(3) ,205-221

Roy Bhowmik S.K. and Das Ananda K., 2007, Rainfall Analysis for Indian monsoon region from merged dense raingauge observations and satellite estimates – Evaluation of monsoon rainfall features, Journal of Earth System Science System, 116(3), 187-198

Roy Bhowmik, S.K., Joardar, D, Das, Ananda .K., Rama Rao Y.V and Hatwar H.R., 2006, Impact of KALPANA-1 CMV data in the analysis and forecast of IMD operational NWP system, Mausam, 57, 319-331

Lal, B., Singh, O.P., Prasad, O., Roy Bhowmik, S.K., Kalsi, S.R. and Subramanian, S.K., 2006, District Level value added dynamical Ensemble Forecast, Mausam,57, 209-220

7.2 Meteorological Monograph/Science Reports

Kotal S.D., Roy Bhowmik S.K. and Mukhopadhaya B., 2009, Performance of IMD NWP based Objective Cyclone Forecast System during 2008-2009, IMD Met Monograph No. Cyclone Warning 4/2009

Roy Bhowmik S.K., Durai, V.R., Das Ananda K and Mukhopadhaya B., 2009, Performance of IMD Multi-model ensemble based district level forecast system during summer monsoon 2008, IMD Met Monograph No. 8/2009

Roy Bhowmik S.K., Durai, V.R., Das Ananda K and Mukhopadhaya B., 2009, Evaluation of Prediction skill of ECMWF forecasts over Indian monsoon region in medium tange time scale during summer monsoon 2008, IMD Met Monograph No. 7/2009

Roy Bhowmik S.K. and Hatwar H.R., 2008, Performance of operational NWP short-range forecast - Monsoon 2007 Report, p. 78-92, IMD Met Monograph No. Synoptic Meteorology 6/2008.

Kalsi et al. (Roy Bhowmik as co-author), 2007, Probable maximum storm surge heights for the maritime districts of India, IMD. Met Monographh No. Synoptic Meteorology No. 5/2007

25 (B) Contribution of National Centre for Medium Range Weather Forecasting (NCMRWF), Noida (UP), India

1. Summary of Highlights :

The operational forecast suite at NCMRWF is the Global Forecast System (GFS, adapted from NCEP, USA) forecasting suite at T254L64 resolution The Grid point Statistical Interpolation (GSI) Analysis Scheme has been made operational in the GFS from 1 Jan 2009.

26 2. Equipment used in the centre :

Cray X1E (64 Processor 1.1TF PARAM PADMA (IBM-p5) (64 proc 0.5TF) LAN (1000mbps backbone) Internet Leased Line (8 mbps)

3. Data and products in use from GTS :

Surface observations: SYNOP ( ~40,000), SHIP(~2500), METAR(~10,000), BUOY (~30,000) Upper-air observations: TEMP(~1230) , PILOT(~300), Wind profiler (BUFR) (~90) Air-craft observations: AIREP, AMDAR, ACARS(BUFR) (total ~10,000) Satellite winds (in BUFR): METEOSAT -7, 9, GMS (total ~2,50,000) from NESDIS site (through ftp)

Satellite Radiance : AMSU-A/B, HIRS, MHS – from NOAA-17/18, METEOP QSCAT winds, GOES winds (BUFR) from NCEP site ((through ftp) : Analysed SST , snow etc.

4. Forecasting System :

4.1 System run schedule and forecast ranges:

The GFS forecasting suite at T254L64 resolution runs up to 7 days based on 0000UTC initial condition of every day. (Starts at 0500 UTC everyday, takes about 4hours for the whole suite including data assimilation)

WRF model (Nested 27 km resolution, 38 levels) runs everyday based on 0000UTC for 72 hours. (Starts at 1000 UTC everyday, takes about 2 hours for the whole suite, including high resolution runs at 9 km resolution for select locations )

Medium range forecasting system :

4. 2.1 Data Assimilation :

27 4.2.1.1 In Operation: Six hourly intermittent 3D-VAR assimilation system based on Grid point Statistical Interpolation (GSI) Analysis Scheme, along with GFS (T254L64)

4.2.1.1 Research performed: Observation impact studies using various conventional as well as non-conventional observations

4.2.2. Model :

4.2.2.1 In Operation: GFS at T254L64 resolution

4.2.2.2 Research Performed: Sensitivity experiments with different physical parameterisation scheme

4.2.3 Operationally available NWP products

Analysed and predicted (up to 168 hr) wind, geopotential height, at various atmospheric levels and precipitation based on every day 0000 UTC initial condition are available on the web in real time. In addition to these fields, various other anaylsed and predicted fields are provided to the users as per specific demands.

Location specific predictions for surface winds, maximum, minimum temperature, cloudiness, humidity, mean sea level pressure, rainfall etc. for 70 major cities of India.

4.2.4 Operational techniques for application of NWP products

4.2.4.1 In operation : PPM 4.2.4.2 Research : KF

4.2.5 Ensemble Prediction

4.2.5.1 In operation: Multi-model Ensemble (MME) forecasts of rainfall during the monsoon season using predicted rainfall from four models ( viz. NCMRWF, NCEP, UKMO and JMA)

4.2.5.2 Research: Experimental Ensemble Prediction System (EPS) with 8 members of T80L18 model, with perturbed initial state using breeding method

4.2.5.3 Operationally available EPS products : NIL

4.3Short-range forecasting system (0-72hr)

28 4.3.1 Data Assimilation method 4.3.1.1 In operation : WRF-3DVAR 4.3.1.2 Research: Assimilation of Indian Radar observations, Impact of background error covariance etc.

4.3.2 Model 4.3.2.1 In operation : WRF-ARW at 27km resolution with 38 vertical levels over Indian region 4.3.2.2 Research Preformed : Very high resolution (9km and 3km) nested WRF model integrations for exclusive case studies of severe weather, such as tropical cyclone ,very intense rainfall etc.

4.3.3 Operationally available NWP products Analysed and predicted (up to 72 hrs) wind, geopotential height at various atmospheric levels and precipitation based on every day 0000 UTC initial condition are available on the web in real time. In addition to these fields, various other anaylsed and predicted fields are provided to the users as per specific demands.

4.3.4 Operational techniques : NIL 4.3.5 EPS : NIL

4.4 Nowcasting : NIL

4.5 Specialised NWP :

4.5.2 Specific Models :

4.5.2.1 In Operation: Wave model: WAVEWATCH -III Model (Version 2.22) The model is run for 00z cycle only, and starts with a 6-hr hindcast to assure continuity of swell. Spatial resolution is 10 x 10 longitude-latitude grid extending from 77.50S to 77.50N.

4.6 Extended Range :

4.6.1 In operation: The real-time seasonal prediction for monsoon rainfall is being carried out at NCMRWF using a two-tier approach. In this approach, the predicted Sea Surface Temperatures (SST) from NCEP CFS are provided as input to the global atmospheric model.

4.7 Long range :

29 4.7.1 In operation: The real-time seasonal prediction for monsoon rainfall is being carried out at NCMRWF using a two-tier approach. In this approach, the predicted Sea Surface Temperatures (SST) are provided as input to the global atmospheric model. These SST data are obtained from IRI, USA and these contain multi-model ensemble monthly SST predictions and an uncertainty factor. Three SST scenarios are prepared based on these datasets and are used as input to the model. Several ensemble integration runs for the monsoon season are made and a probabilistic prediction is prepared by taking into account the bias of the model. First set of predictions are made in mid- April and predictions are updated in mid-May.

5. Verification

Objective verification scores against the analysis and observations are computed every day valid for 00UTC at standard pressure levels for different areas as recommended by the CBS, WMO. Monthly averages are then computed from the daily values of all forecasts verifying within the relevant month. The scores are shared with other operational NWP centres.

6. Plans for future :

6.1.1 Next year

It is proposed to increase the resolution of the current operational GFS to T382L64 and assimilate the Indian satellite observations (INSAT-3d) and OCEANSAT)

6.1.2 Next four years

It is planned to implement the U. K. Met Office, Unified Model and 4-D Var assimilation system and make it operational by next year. A coupled ocean- atmosphere assimilation-forecast system will also be implemented for development of unified prediction suite for different space-time scales (up to a season), with emphasis on predictions from days to weeks in advance initially.