Application of Neural Network Technique to improve the location specific forecast of temperature over Delhi from MM5 model

India Meteorological Department (IMD) has been using direct model output (2 meters height temperature) of MM5 model as numerical guidance for forecasting maximum and minimum temperature of Delhi in short range time scale (up to 72 hours). Performance statistics of the direct model outputs of the model for maximum and minimum temperature show that forecast skill of the model is reasonably good, particularly for the minimum temperature. For further improving the model forecast, Neural Network (NN) as well as regression techniques are applied so that the systematic errors of the direct model output of the model for maximum and minimum temperature could be reduced. The study shows that both Neural Network approach and regression technique are capable to improve the forecast skill of maximum and minimum temperature. Daily modified forecasts are found persistently closer to the observations when the method is tested with the independent sample. The methods are found to be promising for operational application.

District level value-added dynamical-synoptic forecast system for rainfall

Lkkj & bl ’kks/k&Ik= esa o"kkZ ls lacaf/kr ftyk Lrjh; ekuksa ij vk/kkfjr xfrdh; flukfIVd iwokZuqeku dh Ik)fr vkSj mlds iz;ksxkRed ifj.kkeksa dks izLrqr fd;k x;k gSA blds igys bl rduhd dk mi;ksx o"kZ 2005 dh ekulwu iwoZ dh _rq] nf{k.kh if’peh ekulwu _rq vkSj ekulwu ds Ik’pkr dh _rq ds nkSjku gqbZ ftysokj o"kkZ dk iwokZuqeku yxkus ds fy, fd;k x;k FkkA mDr 2005 ds iwokZuqeku ds fy, Vh-&80],e- ,e- 5] jk"Vªh; e/; vof/k ekSle iwokZuqeku dsUnz ¼,u- lh- ,e- vkj- MCY;w- ,Q-½ ds- bZ- Vh- ,- vkSj Hkkjr ekSle foKku foHkkx esa dk;Z’khy ,e- ,e- 5 xfrdh; fun’kksZa dk lefUor #Ik ls mi;ksx fd;k x;k FkkA iwokZuqeku esa lefUor #Ik ls ;ksxnku nsus okys lHkh fun’kksZa ds ekuksa dk vyx&vyx ewY;kadu djds muls izkIr gq, o"kkZ ds iwokZuqekuksa dk mi;ksx djds xfrdh; iwokZuqeku rS;kj fd;k x;k gSA ifjpkyu y{k.kksa] m/okZ/kj osx] mixzg ls izkIr lwpuk flukfIVd pkVksZa vkSj tyok;q foKku vkfn tSls vU; fun’kksZa dk vkdyu djds xfrdh; iwokZuqekuksa dks eku vk/kkfjr ¼osY;w ,fMM ½ flukfIVd ekSle iwokZuqekuksa esa ifjofrZr fd;k x;k gSA o"kZ 2005 esa fd, x, iwokZuqeku ds vuqHko ls ;g irk pyk gS fd lefUor #Ik ls ;ksxnku nsus okys fun’kksZa ls vyx&vyx izkIr gq, o"kkZ ds iwokZuqekuksa dh fuiq.krk dh rqyuk esa eku&vk/kkfjr ¼osY;w ,fMM ½ xfrdh; flukfIVd iz.kkyh ds 24, 48, 72 ?kaVksa ds ftyk Lrjh; o"kkZ ds iwokZuqeku dgha vf/kd fuiq.krk ls rS;kj fd, tk ldrs gSaA The paper presents the methodology and trial results of the district level value-added dynamical-synoptic forecast for rainfall. The technique was tried for forecasting districtwise rainfall during Pre-monsoon, Southwest Monsoon and Post monsoon seasons of 2005. The constituent dynamical models were T-80, MM5, ETA of the National Centre for Medium Range Weather Forecasting (NCMRWF) and the MM5 model operational at India Meteorological Department, New Delhi. The dynamical predictions were prepared using the rainfall predictions of the constituent models by assigning different weights. The dynamical predictions were converted into value-added synoptic-weather forecasts by taking into account other inputs like circulation features, vertical velocity, satellite information, synoptic charts and climatology etc. The experience during 2005 has shown that the value-added dynamical-synoptic system can produce 24, 48, 72 hours district level rainfall forecast of greater skill than the skills of the constituent models.

SENSITIVITY TEST OF PARAMETERIZATION SCHEMES OF MM5 MODEL FOR PREDICTION OF THE HIGH IMPACT RAINFALL EVENTS OVER BANGLADESH

The sensitivity test of parameterization schemes for prediction of summer monsoon high impact rainfallevents (HIRE) over Bangladesh has been performed using the Fifth-Generation PSU/NCAR Mesoscale Model(MM5) conducting six historical HIRE cases. The MM5 model was run on triple-nested domains at 45, 15, 5 kmhorizontal resolutions using Anthes-Kuo (AK), Grell (Gr), Kain-Fritsch (KF), Betts-Miller (BM) andKain-Fritsch2 (KF2) cumulus parameterization schemes (CPS) with Medium Range Forecast (MRF) andBlackadar planetary boundary layer (PBL).The model predicted rainfall was compared both spatially andquantitatively with Tropical Rainfall Measuring Mission (TRMM) rainfall. While parameterization options ofMM5 model have been investigated spatially for Bangladesh, Anthes-Kuo CPS with both MRF and BlackadarPBL (AKM & AKB) options of MM5 have found suitable. Quantitatively, Anthes-Kuo CPS with MRF PBL (AKM)option has calculated the better average rainfall over Bangladesh. By this way, AKM has found suitable in bothspatial and quantitaive comparisons. Thus, Anthes-Kuo CPS with MRF PBL (AKM) has considered as the bestMM5 option for prediction of summer monsoon HIRE cases over Bangladesh.The sensitivity test of parameterization schemes for prediction of summer monsoon high impact rainfall events (HIRE) over Bangladesh has been performed using the Fifth-Generation PSU/NCAR Mesoscale Model (MM5) conducting six historical HIRE cases. The MM5 model was run on triple-nested domains at 45, 15, 5 km horizontal resolutions using Anthes-Kuo (AK), Grell (Gr), Kain-Fritsch (KF), Betts-Miller (BM) and Kain-Fritsch2 (KF2) cumulus parameterization schemes (CPS) with Medium Range Forecast (MRF) and Blackadar planetary boundary layer (PBL).The model predicted rainfall was compared both spatially and quantitatively with Tropical Rainfall Measuring Mission (TRMM) rainfall. While parameterization options of MM5 model have been investigated spatially for Bangladesh, Anthes-Kuo CPS with both MRF and Blackadar PBL (AKM & AKB) options of MM5 have found suitable. Quantitatively, Anthes-Kuo CPS with MRF PBL (AKM) option has calculated the better average rainfall over Bangladesh. By this way, AKM has found suitable in both spatial and quantitaive comparisons. Thus, Anthes-Kuo CPS with MRF PBL (AKM) has considered as the best MM5 option for prediction of summer monsoon HIRE cases over Bangladesh. DOI: http://dx.doi.org/10.3329/jme.v44i1.19496

A model-based study of the wind regime over the Corinthian Gulf

The Corinthian Gulf is a narrow sea-level passage surrounded by a steep complex topography that consists of high mountains as well as elevated and sea-level gaps and straits. The complex terrain is expected to affect the wind flow in the area that often experiences high winds, with important consequences on the commercial and recreational activities over the gulf's maritime area. For that reason, a model-based study of the wind regime over the Corinthian Gulf has been created, as observational data over the area are recent and spatially sparse. Analysis of 5 yr of data from the fifth-generation Penn State/NCAR mesoscale model (MM5) model reveals that the wind regime of the gulf is greatly influenced by the topography. Easterly winds occur more frequently and are stronger in the maritime area in the western edge of the gulf, with a frequency of occurrence on the order of 70%. Moreover, the most intense wind events at this area occur during the winter season (December, January, and February). Finally the paper also provides a discussion on the synoptic patterns, which lead to the strongest wind events in the studied area.

A model-based study of the wind regime over Corinthian Gulf in Greece

The Corinthian Gulf is a narrow sea-level passage surrounded by a steep complex topography that consists of high mountains, elevated and sea-level gaps/straits. The complex terrain is expected to affect the wind flow in the area that often experiences high winds with important consequences on the commercial and recreational activities over the Gulf maritime area. For that reason, a model-based study of the wind regime over the Corinthian Gulf has been built, as observational data over the area are recent and spatially sparse. The analysis of 5yr data from MM5 model reveals that the wind regime of the gulf is greatly influenced by the topography. Easterly winds occur more frequently and are stronger at the maritime area in the western edge of the gulf with a frequency of occurrence of the order of 70%. Moreover, the most intense wind events at this area occur during the winter season (December, January, and February). Finally the paper also provides a discussion on the synoptic patterns, which lead to the strongest wind events in the studied area.

Forecasting mesoscale precipitation using the MM5 model with the four-dimensional data assimilation (FDDA) technique

The two basic forms of multi-scale data assimilation procedures (FDDA), based on Newtonian relaxation, of analysis and observations nudging have been applied for precipitation event period occurred over Portugal during summer season, using the Fifth Generation Mesoscale Model (MM5) developed and maintained by the Pennsylvania State University and National Center for Atmospheric Research (PSU/NCAR). The model has been configured for three nested grid domains covering part of the Eastern part of North Atlantic region evolving the Portugal, with 35 vertical levels, from surface up to 100 hPa top level. The model forecasting have been conducted employing daily available data from surface observational network, radio-sounding from Lisbon/Portugal and NOAA-16 polar orbiting satellite retrieved vertical profiles data. The three integration domains of MM5 model have been processed using, as boundary and first guess fields, the global atmospheric forecast NCEP-NWS/AVN model data gathered through the Unidata Local Data Manager (LDM)/Unidata Internet Data Distribution (IDD) system. All daily forecasting, with FDDA and with no FDDA, have been run for 60 hours forecast, with 30 minutes interval model data output to provide enough timely detailed results. The FDDA analysis presented a quite reasonable data ingesting volume of almost all available satellite data, with the exception of humidity data retrieved for high levels, above around 500 hPa. The obtained results indicate that, even using weak FDDA constraint coefficient values, presents a significant improvement in the numerical prognosis in the precipitation field, on both space and time integration levels. The results also presented an enhancement of the physics of the convective mesoscale system development, particularly over mountain region, indicating that it would be interesting to conduct an experiment with a dense data collecting platform coverage focused on events which occur in some prevailing mountain region of Portugal.