Artificial neural network model for precipitation forecast over Western Himalaya using satellite images

Kalpana satellite images in real time available by India meteorological department (IMD), contain relevant inputs about the cloud in infra-red (IR), water vapor (WV), and visible (VIS) bands. In the present study an attempt has been made to forecast precipitation at six stations in western Himalaya by using extracted grey scale values of IR and WV images. The extracted pixel values at a location are trained for the corresponding precipitation at that location. The precipitation state at 0300 UTC is considered to train the model for precipitation forecast with 24 hour lead time. The satellite images acquired in IR (10.5 - 12.5 µm) and WV (5.7 - 7.1 µm) bands have been used for developing Artificial Neural Network (ANN) model for qualitative as well as quantitative precipitation forecast. The model results are validated with ground observations and skill scores are computed to check the potential of the model for operational purpose. The probability of detection at the six stations varies from 0.78 for Gulmarg in Pir-Panjal range to 0.95 for Dras in Greater Himalayan range. Overall performance for qualitative forecast is in the range from 61% to 84%. Root mean square error for different locations under study is in the range 5.81 to 8.7.

Lidar-derived aerosol concentration and their relationship with horizontal winds over an urban location

Lidar-derived aerosol vertical profiles obtained at Pune, a low latitude tropical station, on about 535 days during a ten-year period (1987 – 96) along with simultaneous pilot-balloon wind (speed and direction) data of India Meteorological Department, Pune have been used in the study to investigate the influence of horizontal winds on the aerosol characteristics in the lower atmosphere. Aerosol column content in the atmospheric boundary layer (surface to 1100 m altitude above ground-level) as well as aerosol number density at the surface level (at 50 m) showed relatively higher values over the lidar site whenever the winds were blowing from the main urban and industrial regions of the city of Pune. This effect was found to be more pronounced during the winter season. Wind speeds also correlate well with increased aerosol loading, but only during selected high wind speed episodes. Thus the study shows that the short- and long-term increases in aerosol concentration/loading over the observation site are, to a large extent, influenced by horizontal winds in the surface layers and this in turn, can be attributed to the increasing human/urban activity around the lidar site over the years.

Possible mechanisms related to internal dynamics for anomalously high rainfall during August 2020 of Indian Summer Monsoon

In the present study, an investigation is made to understand the physical mechanism behind the anomalous high rainfall during August 2020 over the Indian subcontinent using both observation and GFS T1534 weather forecast model. According to India Meteorological Department (IMD), the country receives 27% excess rainfall in the month of August 2020. The excess rainfall is mainly contributed by the 5 well marked low pressure systems which formed over Bay of Bengal and moved west-northwestwards across central India up to Western Madhya Pradesh and Rajasthan. The analysis reveals that the observed anomalous rainfall is distributed over central India region extending from coastal Orissa to central part of Chhattisgarh, Madhya Pradesh and western coast of Gujarat region. It is also found that the August-2020 heavy rainfall is mainly contributed by the synoptic (2-10 days) component of the total rainfall whereas the contribution of the large-scale intraseasonal oscillation (ISO) component (10-90 days) is quite less. Although the present operational Global Forecast System (GFS) T1534 (GFS T1534) is able to predict the anomalous high rainfall with day-1 lead time, it underestimates the magnitude of the synoptic variance. Further, the large-scale dynamical and thermodynamical parameters show anomalous behaviour in terms of strong low level (850 hPa) jet, vertical velocity and associated moisture convergence in the lower level. The GFS T1534 is able to forecast the above large-scale features reasonably well even with day-5 lead time. From energetics analysis, it is found that the mean kinetic energy (MKE) is stronger for August 2020 as compared to climatological value and the strong MKE efficiently transfers the energy to the synoptic scale, and hence the synoptic eddy kinetic energy is higher. Along with that, the ISO scale kinetic energy for August 2020 is less compared to the August climatological value. GFS T1534 model has some fidelity in capturing the energy conversion processes, but it has some difficulty in capturing the magnitude with increased lead time.

Algorithms for computerised monitoring the performance of DCP network

India Meteorological Department (IMD) has been operating a network of one hundred Automatic Weather Stations called Data Collection Paltform (DCP) since 1986. All these stations are unmanned and there is no way to know the working status of DCP equipment except to extract such Information from the data transmitted by them. Hence suitable algorithms were developed to evaluate the working status of various sub-systems of DCP stations by analysing the data received from them, which is essential for their effective and efficient maintenance. The concept used in developing these algorithms is described here.

Development of 1961-1990 monthly surface climatology of India and patterns of differences of some meteorological parameters with respect to the 1951-1980 climatology

Climate normals are used to describe the average climatic conditions of a particular place and are computed by National Meteorological Services of all countries. The World Meteorological Organization (WMO) recommends that all countries prepare climate normals for the 30-year periods ending in 1930, 1960, 1990 and so on, for which the WMO World Climate Normals are published. Recently, Climatological Normals for the period 1961-1990 have been prepared by India Meteorological Department (IMD) which will change the baseline of comparison from 1951-1980. In this paper, preparation of the 30-year Climatological Normals of India for the period 1961 to 1990 and spatial patterns of differences of annual means of temperatures, relative humidity, clouds, rainfall and wind speed from the previous normals (1951-1980) are documented.The changes from earlier climatological normals indicate increase in annual means of maximum temperature, relative humidity and decrease in annual means of minimum temperature, cloud amount, rainfall, rainy days and wind speed over large parts of the country during 1961-1990. The spatial patterns of changes in dry bulb temperatures and relative humidity are complementary over most parts of the country. Compared with 1951-1980 climatology, there are large scale decreases in annual mean rainfall, rainy days and wind speed over most parts of the country during 1961-1990. The decrease in wind speed may be partly due to changes in exposure conditions of observatories due to urbanization.

Observational aspects including weather radar for tropical cyclone monitoring

Hkkjr ekSle foKku foHkkx ¼vkbZ- ,e- Mh-½ ds iwokZuqeku izn’kZu ifj;pkstuk ¼,Q- Mh- ih-½ ds lanHkZ esa dh xbZ fofHkUu izs{k.kkRed lqfo/kkvksa vkSj rduhdksa dh leh{kk dh xbZ gSA ftudk iz;ksx pØokr ds iFk dk irk yxkus vkSj m".kdfVca/kh; pØokrksa dks le>us ds fy, fd;k tk ldrk gSA izkS|ksfxdh ds laca/k esa fd, x, iz;klksa ds okLrfod ijh{k.k ls izpkyukRed lanHkZ esa gekjs iwokZuqekuksa ds fu"iknu dk irk pyrk gSA bl laca/k esa vko’;d mik;ksa ij bl 'kks/k&i= esa fopkj&foe’kZ fd;k x;k gSA In the context of the Forecast Demonstration Project (FDP) of the India Meteorological Department (IMD), a review is made of the various observational facilities and techniques which can be deployed, for the detection tracking and understanding of tropical cyclones. The real test of the efforts in terms of technology is the performance of our forecasts in an operational context. The paper discusses the steps needed in this regard.

Estimation of intensity of tropical cyclone over Bay of Bengal using microwave imagery

caxky dh [kkM+h esa o"kZ 2008&2010 esa ,Q- Mh- ih- vof/k ¼15 vDrwcj ls 30 uoEcj½ ds nkSjku vk, ik¡p pØokrksa ds lw{e rjaxh; es?k fcEckofy;ksa rFkk 85 fxxkgV~tZ vko`fÙk esa izkIr fd, x, mRiknksa dh tk¡p dh xbZ gS ftlls rkieku nhfIr] rkieku nhfIr esa vfu;ferrk] dsUnz dk LFkku] lrg ij vuojr cgus okyk vf/kdre iou ¼,e- ,l- MCY;w-½ rFkk pØokrksa ds fHkUu&fHkUu fLFkfr;ksa esa muds rhozhdj.k ls lacaf/kr djdksa tSls% vonkc ¼Mh-½] xgu vonkc ¼Mh- Mh-½] pØokrh; rwQku ¼lh- ,l-½] rhoz pØokrh; rwQku ¼,l-lh-,l-½] vfr rhoz pØokrh; rwQku ¼oh-,l-lh-,l-½ vkfn dk vkdfyr dsUnzh; nkc ¼bZ- lh- ih-½ dk vkdyu fd;k tk ldsA izf{kr fd, x, nhfIr rkieku vfu;ferrkvksa dh rqyuk lS)kafrd :i ls bZ-lh-ih- ds csLV VªSd vkdyu ij vk/kkfjr nhfIr rkieku vfu;ferrk ,oa bu pØokrksa ds ckgjh nkc ds lkFk Hkh dh xbZ gSA dsUnz ds LFkku] bZ-lh-ih- ,oa lw{erajxh; fcEckoyh ds vk/kkj ij vkdfyr ,e- ,l- MCY;w- dh rqyuk csLV VªSd ,oa Hkkjr ekSle foKku foHkkx ds Mh- oksjkWd ds vkdyu ls dh xbZ gS vkSj mldk fo’ys"k.k fd;k x;k gSA pØokrh; fo{kksHk ¼lh- Mh-½ ds dsUnz ds LFkku esa varZ tSlkfd lw{erjaxh fcEckofy;ksa rFkk csLV VªSd vkdyu ds }kjk vkdfyr fd;k x;k gS] fo{kksHkksa ds rhozhdj.k ds lkFk&lkFk de gksrk tkrk gS vkSj vonkc ¼Mh-½ dh fLFkfr esa yxHkx 25 fd-eh- ls vfr rhoz pØokrh; rwQku ¼oh-,l-lh-,l-½ dh fLFkfr esa 18 fd- eh ds chp cnyrk jgrk gSA tcfd ;g varj Mh oksjkWd ds vkdyu ls dkQh vf/kd gSA lw{erjaxh; vkdyuksa ij vk/kkfjr ,e- ,l- MCY;w- vkdyu oh-,l- lh- ,l- ds nkSjku csLV VªSd vkdyuksa ls yxHkx 28 ukWV~l vf/kd vkdfyr fd;k x;k gS vkSj vonkc ¼Mh-½@pØokrh; rwQku ¼lh-,l-½@rhoz pØokrh; rwQku ¼,l- lh- ,l-½ dh fLFkfr esa ;g 6&8 ukWV~l vkdfyr fd;k x;k gSA csLV VSªd vkdyuksa ls lkisf{kd varj dks ns[kus ls irk pyk gS fd lh-,l- vkSj ,l-lh- dh fLFkfr esa lw{e rajx esa ,e-,l-MCY;w- yxHkx 12&15 izfr’kr vkSj oh-,l-lh-,l- dh fLFkfr esa yxHkx 30 izfr’kr vf/kd vkdfyr gqvk gS tcfd Mh- oksjkWd dk ,e- ,l- MCY;w- vkdyu lh- ,l-] ,l- lh- ,l- vkSj oh- ,l- lh- ,l- dh fLFkfr;ksa esa 15&18 izfr’kr de gks x;k gSA caxky dh [kkM+h ds Åij 230 dsfYou dk nhfIr rkieku vonkc ds cuus ds fy, vuqdwy gksrk gS] 250 dsfYou dk rkieku bldks pØokrh rwQku esa 260 dsfYou rhoz pØokrh rwQku esa vkSj 270 dsfYou vfr izpaM+ pØokrh rwQku esa cny nsrk gSA nhfIr rkieku ds nsgyheku ¼FkszlksYM osY;w½ ds vfHkKku ¼fMVSD’ku½ ls bl iz.kkyh ds rhoz gksus dk iwokZuqeku nsus ds fy, iz;kIr vfxze le; fey ldrk gSA blh izdkj nhfIr rkieku folaxfr 3 dsfYou ls vf/kd gksus ij pØokrh; rwQku rhoz pØokrh; rwQku esa cny tkrk gS vkSj 8 dsfYou dk rkieku bls caxky dh [kkM+h esa vfr izpaM pØokrh; rwQku ds :i esa cny nsrk gSA Microwave cloud imageries and derived products in the frequency of 85 GHz have been examined for five cyclones that occurred during FDP period (15 October- 30 November) of 2008-2010 over the Bay of Bengal to estimate the brightness temperature, brightness temperature anomaly, location of centre, maximum sustained wind (MSW) at surface level and estimated central pressure (ECP) associated with cyclones in their different stages of intensification like depression (D), deep depression (DD), cyclonic storm (CS), severe cyclonic storm (SCS), very severe cyclonic storm (VSCS), etc. Also the observed brightness temperature anomalies are compared with theoretically derived brightness temperature anomalies based on the best track estimates of ECP and outermost pressure for these cyclones. The location of centre, ECP and MSW based on microwave imagery estimates have been compared with those available from the best track and Dvorak’s estimates of India Meteorological Department and analyzed. The difference in location of the centre of cyclonic disturbance (CD) as estimated by microwave imageries and best track estimates decreases with intensification of the disturbances and varies from about 25 km in depression (D) stage to 18 km in VSCS stage whereas the difference is significantly higher in case of Dvorak estimate compared to best track estimate. The MSW based on microwave estimates is higher than that of best track estimates by about 28 knots during VSCS and 6-8 knots during D, CS, SCS stage. Considering relative difference with respect to best track estimates, the MSW is overestimated in microwave by about 12-15% in case of CS and SCS stage and by about 30% in VSCS stage while Dvorak’s MSW overestimation reduced to 15-18% during CS, SCS and VSCS stages. Brightness temperature of the order of 230 K is favourable for genesis (formation of D), 250K for its intensification into CS, 260 K for intensification into SCS and 270K for its further intensification into VSCS stage over the Bay of Bengal. Detection of threshold value of brightness temperature may provide adequate lead time to forecast intensification of the system. Similarly, when brightness temperature anomaly exceeds 3K, CS intensify into SCS and 8K, it intensifies into a VSCS over Bay of Bengal.

Spectra of seismic noise at selected Indian stations

India Meteorological Department (IMD) is operating five digital seismograph systems at New Delhi (NDI),.Shillong (SHL), Pune (POO), Kodaikanal (KDK) and Dhamlsala (DHM) since last few years. The details pertaining to instrumental characteristics and software for data retrieval and processing are presented in this paper, Through PC based algorithms. noise pectra are computed and interpreted for these five stations. It is found that the maximum peak occurs at about 6Hz for Pune. Shillong and Kodaikanal while at New Delhi and Dharmsala, it is noted at about 2 Hz. The spectral peak at Shillong as deduced from the SRO system shifts to about I Hz which is in agreement with a similar observation reported at Gauribidanur seismic array.

A satellite based study of pre-monsoon thunderstorms (Nor’westers) over eastern India and their organization into mesoscale convective complexes

Nor’wester studies have a long history of climatological, synoptic and radar observations. These studies have been briefly mentioned and the field programs for the study of Nor’westers implemented in India Meteorological Department (IMD) from 1931-1941 have been touched upon. Indian atmospheric science community organized a multi-year STORM program during 2007-2010 to understand the formation of these severe local storms and also understand their dynamics through modeling. An attempt is made to use INSAT Infrared and Visible imageries to document the convective cells which developed over Eastern and North-East (NE) Indian states and adjoining countries of Bangladesh, Bhutan and Nepal for the year 2009. Also convective cells which organized themselves into Mesoscale Convective Complexes (MCCs) for the four years period 2007-2010 have been studied. It is found that by and large Eastern India (Jharkhand, Orissa, Sub Himalayan West Bengal and Bangladesh) is responsible for the initiation of convection. Development occurs as the cells propagate over the neighbouring areas of Bangladesh and NE India. Important observations with regard to initiation, maturity and dissipation etc. of the MCCs are provided. It is suggested that half hourly to hourly monitoring of convection can be accomplished by using INSAT imagery, along with multiple overlapping radar coverages, which could help in nowcasting of convective cells. Synoptic and thermodynamic forcing can help as broad guidance. The only effective way for effective warning is nowcasting using satellite and multiple radar coverage.

Recent models for long range forecasting of Southwest monsoon rainfall in India

India has a long tradition of scientific work or, long range forecasting of the southwest monsoon ever since the times of Blanford and Walker In the early Parts of this century, the recent decades have witnessed increased research in regard to the development of new long range forecast models in the India Meteorological Department which have, given correct long range (seasonal) forecasts of southwest monsoon rainfall, over the country as a whole, during, the successive four year~, 1988 to 1991, Presently, four models namely, Para-metric Power Regression, Dynamic Stochastic Transfer and Improved Multiple Regression models are being used for formulating the seasonal forecast of monsoon rainfall over the country as a whole, The forecast is issued in two stages, In the first stage a tentative inference which is qualitative in nature is issued before the middle of April based mainly on the Parametric model which utilizes signals from 16 regional and global parameters that are related to land, ocean and atmospheric forcing and show physical linkages with monsoon. In the second stage, a firm quantitative forecast is issued towards the end of the May and is based on the remaining three models mentioned above, although higher weight age is given to the Power Regression Model which has shown encouraging performance during the last four years. In this paper, these recently developed models and the scientific basis underlying these are discussed, Data on validation of these Operational models, used for the long range forecast during the past four years (1988-91) are also presented.