Anirban Middey

Bio: Anirban Middey is an academic researcher from National Environmental Engineering Research Institute. The author has contributed to research in topics: Thunderstorm & Environmental science. The author has an hindex of 11, co-authored 30 publications receiving 291 citations. Previous affiliations of Anirban Middey include University of Calcutta.

Track and intensity forecast of tropical cyclones over the North Indian Ocean with multilayer feed forward neural nets

Abstract: Precise forecasting of the track and intensity of tropical cyclones remains one of the top priorities for the meteorological community. In the present study multilayer feed forward neural nets with different architectures are developed to identify the best neural net for forecasting the track and intensity of tropical cyclones over the North Indian Ocean (NIO) with 6, 12 and 24 h lead time. Forecast errors are estimated with each neural net. The result reveals that the neural net architecture 1 (NNA 1) with 10 input layers, 2 hidden layers, 5 hidden nodes and 2 output layers provides the best forecast for both the track and the intensity of the tropical cyclones over NIO. Two cyclones of the same category in Saffir Simpson Hurricane Scale, namely Nargis and Phet, that occurred over the Bay of Bengal and the Arabian Sea of the NIO basin are considered in the present study for validation. The result reveals that the prediction errors (%) with NNA 1 model in estimating the intensity of the cyclones Nargis and Phet during the validation are 3.37, 8.29 and 9.74 as well as 6.38, 11.26 and 18.72 with 6, 12 and 24 h lead time, respectively. The mean track errors for 6, 12 and 24 h forecasts are observed to be 45, 69 and 89 km for cyclone Nargis and 54, 87 and 98 km for cyclone Phet. NNA 1 model is observed to perform better than NNA 2 and NNA 3 models and the existing numerical models.

Adaptive neuro-fuzzy inference system to forecast peak gust speed during thunderstorms

Abstract: The aim of the present study is to develop an adaptive neuro-fuzzy inference system (ANFIS) to forecast the peak gust speed associated with thunderstorms during the pre-monsoon season (April–May) over Kolkata (22°32′N, 88°20′E), India. The pre-monsoon thunderstorms during 1997–2008 are considered in this study to train the model. The input parameters are selected from various stability indices using statistical skill score analysis. The most useful and relevant stability indices are taken to form the input matrix of the model. The forecast through the hybrid ANFIS model is compared with non-hybrid radial basis function network (RBFN), multi layer perceptron (MLP) and multiple linear regression (MLR) models. The forecast error analyses of the models in the test cases reveal that ANFIS provides the best forecast of the peak gust speed with 3.52% error, whereas the errors with RBFN, MLP, and MLR models are 10.48, 11.57, and 12.51%, respectively. During the validation with the 2009 observations of the India Meteorological Department (IMD), the ANFIS model confirms its superiority over other comparative models. The forecast error during the validation of the ANFIS model is observed to be 3.69%, with a lead time of <12 h, whereas the errors with RBFN, MLP, and MLR are 12.25, 13.19, and 14.86%, respectively. The ANFIS model may, therefore, be used as an operational model for forecasting the peak gust speed associated with thunderstorms over Kolkata during the pre-monsoon season.

Prediction and examination of seasonal variation of ozone with meteorological parameter through artificial neural network at NEERI, Nagpur, India

Abstract: The present study focused on seasonal relations and predictions of the ozone (O3) coupled with NO2 and meteorology. Monitoring of ozone concentration throughout year shows an increasing trend during summer and a decreasing trend in the winter season. A comparison between three types of ANN; multilayer perceptron trained (MLP) with back-propagation, radial basis functions (RBF) and generalized regression neural network (GRNN) for short prediction of ozone are conclusively demonstrated. The model results are validated with observations from next monsoon. Based on the model's performance, the MLP back propagation model gives the best correlation between observed and predicted ozone concentrations than other models. Performance assessment parameters considered in the study also indicates that MLP is the best-fit model for prediction of ozone concentration throughout the year.

Intensity forecast of tropical cyclones over North Indian Ocean using multilayer perceptron model: skill and performance verification

Abstract: The coastal regions of India are profoundly affected by tropical cyclones during both pre- and post-monsoon seasons with enormous loss of life and property leading to natural disasters. The endeavour of the present study is to forecast the intensity of the tropical cyclones that prevail over Arabian Sea and Bay of Bengal of North Indian Ocean (NIO). A multilayer perceptron (MLP) model is developed for the purpose and compared the forecast through MLP model with other neural network and statistical models to assess the forecast skill and performances of MLP model. The central pressure, maximum sustained surface wind speed, pressure drop, total ozone column and sea surface temperature are taken to form the input matrix of the models. The target output is the intensity of the tropical cyclones as per the T—number. The result of the study reveals that the forecast error with MLP model is minimum (4.70 %) whereas the forecast error with radial basis function network (RBFN) is observed to be 14.62 %. The prediction with statistical multiple linear regression and ordinary linear regression are observed to be 9.15 and 9.8 %, respectively. The models provide the forecast beyond 72 h taking care of the change in intensity at every 3-h interval. The performance of MLP model is tested for severe and very severe cyclonic storms like Mala (2006), Sidr (2007), Nargis (2008), Aila (2009), Laila (2010) and Phet (2010). The forecast errors with MLP model for the said cyclones are also observed to be considerably less. Thus, MLP model in forecasting the intensity of tropical cyclones over NIOs may thus be considered to be an alternative of the conventional operational forecast models.

The Applicability of Bipartite Graph Model for Thunderstorms Forecast over Kolkata

Abstract: Single Spectrum Bipartite Graph (SSBG) model is developed to forecast thunderstorms over Kolkata during the premonsoon season (April-May). The statistical distribution of normal probability is observed for temperature, relative humidity, convective available potential energy (CAPE), and convective inhibition energy (CIN) to quantify the threshold values of the parameters for the prevalence of thunderstorms. Method of conditional probability is implemented to ascertain the possibilities of the occurrence of thunderstorms within the ranges of the threshold values. The single spectrum bipartite graph connectivity model developed in this study consists of two sets of vertices; one set includes two time vertices (00UTC, 12UTC) and the other includes four meteorological parameters: temperature, relative humidity, CAPE, and CIN. Three distinct ranges of maximal eigen values are obtained for the three categories of thunderstorms. Maximal eigenvalues for severe, ordinary, and no thunderstorm events are observed to be , , and , respectively. The ranges of the threshold values obtained using ten year data (1997–2006) are considered as the reference range and the result is validated with the IMD (India Meteorological Department) observation, Doppler Weather Radar (DWR) Products, and satellite images of 2007. The result reveals that the model provides 12- to 6-hour forecast (nowcasting) of thunderstorms with 96% to 98% accuracy.

A review of artificial neural network models for ambient air pollution prediction

Abstract: Research activity in the field of air pollution forecasting using artificial neural networks (ANNs) has increased dramatically in recent years. However, the development of ANN models entails levels of uncertainty given the black-box nature of ANNs. In this paper, a protocol by Maier et al. (2010) for ANN model development is presented and applied to assess journal papers dealing with air pollution forecasting using ANN models. The majority of the reviewed works are aimed at the long-term forecasting of outdoor PM10, PM2.5, and oxides of nitrogen, and ozone. The vast majority of the identified works utilised meteorological and source emissions predictors almost exclusively. Furthermore, ad-hoc approaches are found to be predominantly used for determining optimal model predictors, appropriate data subsets and the optimal model structure. Multilayer perceptron and ensemble-type models are predominantly implemented. Overall, the findings highlight the need for developing systematic protocols for developing powerful ANN models.

Forecasting lightning threat using cloud- resolving model simulations

Abstract: Two new approaches are proposed and developed for making time and space dependent, quantitative short-term forecasts of lightning threat, and a blend of these approaches is devised that capitalizes on the strengths of each. The new methods are distinctive in that they are based entirely on the ice-phase hydrometeor fields generated by regional cloud-resolving numerical simulations, such as those produced by the WRF model. These methods are justified by established observational evidence linking aspects of the precipitating ice hydrometeor fields to total flash rates. The methods are straightforward and easy to implement, and offer an effective near-term alternative to the incorporation of complex and costly cloud electrification schemes into numerical models. One method is based on upward fluxes of precipitating ice hydrometeors in the mixed phase region at the-15 C level, while the second method is based on the vertically integrated amounts of ice hydrometeors in each model grid column. Each method can be calibrated by comparing domain-wide statistics of the peak values of simulated flash rate proxy fields against domain-wide peak total lightning flash rate density data from observations. Tests show that the first method is able to capture much of the temporal variability of the lightning threat, while the second method does a better job of depicting the areal coverage of the threat. Our blended solution is designed to retain most of the temporal sensitivity of the first method, while adding the improved spatial coverage of the second. Exploratory tests for selected North Alabama cases show that, because WRF can distinguish the general character of most convective events, our methods show promise as a means of generating quantitatively realistic fields of lightning threat. However, because the models tend to have more difficulty in predicting the instantaneous placement of storms, forecasts of the detailed location of the lightning threat based on single simulations can be in error. Although these model shortcomings presently limit the precision of lightning threat forecasts from individual runs of current generation models,the techniques proposed herein should continue to be applicable as newer and more accurate physically-based model versions, physical parameterizations, initialization techniques and ensembles of forecasts become available.

Dynamical downscaling: Fundamental issues from an NWP point of view and recommendations

Abstract: Dynamical downscaling has been recognized as a useful tool not only for the climate community, but also for associated application communities such as the environmental and hydrological societies. Although climate projection data are available in lower-resolution general circulation models (GCMs), higher-resolution climate projections using regional climate models (RCMs) have been obtained over various regions of the globe. Various model outputs from RCMs with a high resolution of even as high as a few km have become available with heavy weight on applications. However, from a scientific point of view in numerical atmospheric modeling, it is not clear how to objectively judge the degree of added value in the RCM output against the corresponding GCM results. A key factor responsible for skepticism is based on the fundamental limitations in the nesting approach between GCMs and RCMs. In this article, we review the current status of the dynamical downscaling for climate prediction, focusing on basic assumptions that are scrutinized from a numerical weather prediction (NWP) point of view. Uncertainties in downscaling due to the inconsistencies in the physics packages between GCMs and RCMs were revealed. Recommendations on how to tackle the ultimate goal of dynamical downscaling were also described.