01PhD · 2025—Present
◉
Location-specific forecasting of monsoon-induced extremes
My doctoral research aims to develop a location-specific forecasting system by integrating high-resolution numerical weather prediction with machine learning. The work focuses on understanding the physical and dynamical characteristics of monsoon-induced extremes, quantifying model errors, and improving forecast skill through data-driven approaches.
The broader workflow links WRF-ARW simulations, observational and reanalysis datasets, extreme-event diagnostics and ML/DL-based error characterization to move from physical understanding toward more reliable local forecasts.
WRF-ARWML / DLForecastingExtreme events
01 / 07
02Extreme rainfall · Climate variability
∿
Extreme rainfall, monsoon variability & ENSO/IOD teleconnections
Applied Generalized Pareto Distribution methods and estimated return levels for extreme rainfall events, while developing JJAS rainfall climatology datasets to investigate spatial and temporal variability across India.
This work combines extreme-value statistics with long-term monsoon rainfall analysis and investigates how major climate modes, including ENSO and the Indian Ocean Dipole (IOD), are associated with Indian monsoon variability.
GPDReturn levelsJJASENSOIODTeleconnection
02 / 07
03WRF simulations
△
Microphysics & extreme rainfall
Simulated extreme rainfall cases using WRF and evaluated multiple microphysics configurations, including SBM and bulk schemes.
SBMThompsonNSSL
03 / 07
04Deep learning · 100 Indian cities
≈
Analysis and forecasting of temporal rainfall variability over hundred Indian cities
Analysed temporal rainfall variability across 100 Indian cities and developed deep-learning approaches for rainfall forecasting, comparing sequence-learning models to capture temporal patterns in rainfall.
The study demonstrates the application of LSTM, BiLSTM and GRU models for city-level rainfall prediction and highlights the value of deep learning for understanding and forecasting rainfall variability across diverse Indian locations.
100 citiesLSTMBiLSTMGRURainfall forecasting
04 / 07
05Extreme-event process analysis
↕
Dynamics & thermodynamics of extreme rainfall events
Analysing the dynamical and thermodynamical processes associated with extreme rainfall events to understand the atmospheric conditions that favour their development and intensification.
The analysis brings together moisture transport, atmospheric stability, vertical motion and large-scale circulation to distinguish the key dynamic and thermodynamic drivers of extreme precipitation.
DynamicsThermodynamicsMoisture transportExtreme rainfall
05 / 07
06Urban climate · Deep learning
▱
Analyzing urban footprints over four coastal cities of India
Analyzed urban footprints over four coastal cities of India and investigated their association with rainfall and temperature using deep learning models.
The study examines how urban characteristics relate to rainfall and temperature variability, using deep learning to explore these relationships across selected coastal Indian cities.
Smart citiesUrban climateDeep learningCoastal cities
06 / 07
07North Indian Ocean
↯
Tropical cyclone recurvature in the North Indian Ocean
Studying tropical cyclone recurvature over the North Indian Ocean and the associated atmospheric and oceanic conditions that influence changes in track direction and evolution.
NIORecurvatureTropical cyclones
07 / 07
