Bipin Kumar
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Bipin Kumar has a total of 111 co-authors. Here are the top co-authors by number of shared publications:
Bhupendra Bahadur Singh
Indian Institute of Tropical Meteorology, Ministry of Earth Sciences, Government of India, Pune, India
6 shared publicationsRajib Chattopadhyay
Indian Institute of Tropical Meteorology, Ministry of Earth Sciences, Government of India, Pune, India
6 shared publicationsManmeet Singh
Indian Institute of Tropical Metereology, Pune, India
5 shared publicationsSuryachandra A. Rao
5 shared publications- Wojciech W. Grabowski
MMM Laboratory, NSF National Center for Atmospheric Research, Boulder, CO, USA
4 shared publications Neethi Suresh
3 shared publications
Latest Publications
- 27 August 2026
- Published by Springer Nature in Journal of Earth System Science
- Vol. 135 (3) , 213
Advancing oceanographic forecasts in the Bay of Bengal: an AI-based spatiotemporal prediction for SST and MLD
- 27 August 2026
- Published by Springer Nature in Journal of Earth System Science
- Vol. 135 (3) , 213
Advancing oceanographic forecasts in the Bay of Bengal: an AI-based spatiotemporal prediction for SST and MLD
- Open Access
- 12 July 2026
- Published by Bilingual Publishing Group in Journal of Atmospheric Science Research
- Vol. 9 (3) , 68-82
Role of Resolution on Atmospheric Dynamics during Fani Cyclone in the Indian Subcontinent
A persistent discrepancy between mathematically modeled atmospheric circulation and observed precipitation has become increasingly evident over recent decades. High-resolution numerical weather prediction (NWP) models offer a pathway to reduce this gap, yet their practical deployment is constrained by limited data availability, high computational cost, and reliance on traditional model verification measures. This study proposes an efficient NWP framework implemented on small-scale parallel computing systems using an adaptive grid-resolution strategy within the well-known Nonhydrostatic Icosahedral Atmospheric Model (NICAM), enabling global weather analysis. By systematically varying horizontal grid spacing, the approach identifies the minimum resolution required to reliably capture mesoscale variations in wind, temperature, and precipitation while maintaining computational efficiency for targeted forecasting objectives. Global cloud-resolving simulations using NICAM are conducted under low-pressure conditions characteristic of cyclone formation over the Indian subcontinent, with specific focus on Cyclone Fani. Results show that even at reduced resolution, the model reproduces key atmospheric dynamics, including temperature, pressure, wind velocity, and precipitation at both near-surface and upper-atmospheric levels. These findings indicate that reasonable coarse-resolution NICAM simulations executed on small-scale parallel computing systems can effectively capture the essential atmospheric circulation associated with Cyclone Fani, providing a computationally efficient alternative for high-resolution numerical weather prediction.
- 12 July 2026
- Published by Bilingual Publishing Group in Journal of Atmospheric Science Research
- Vol. 9 (3) , 68-82
Open AccessRole of Resolution on Atmospheric Dynamics during Fani Cyclone in the Indian Subcontinent
A persistent discrepancy between mathematically modeled atmospheric circulation and observed precipitation has become increasingly evident over recent decades. High-resolution numerical weather prediction (NWP) models offer a pathway to reduce this gap, yet their practical deployment is constrained by limited data availability, high computational cost, and reliance on traditional model verification measures. This study proposes an efficient NWP framework implemented on small-scale parallel computing systems using an adaptive grid-resolution strategy within the well-known Nonhydrostatic Icosahedral Atmospheric Model (NICAM), enabling global weather analysis. By systematically varying horizontal grid spacing, the approach identifies the minimum resolution required to reliably capture mesoscale variations in wind, temperature, and precipitation while maintaining computational efficiency for targeted forecasting objectives. Global cloud-resolving simulations using NICAM are conducted under low-pressure conditions characteristic of cyclone formation over the Indian subcontinent, with specific focus on Cyclone Fani. Results show that even at reduced resolution, the model reproduces key atmospheric dynamics, including temperature, pressure, wind velocity, and precipitation at both near-surface and upper-atmospheric levels. These findings indicate that reasonable coarse-resolution NICAM simulations executed on small-scale parallel computing systems can effectively capture the essential atmospheric circulation associated with Cyclone Fani, providing a computationally efficient alternative for high-resolution numerical weather prediction.
- Open Access
- 21 May 2026
- Published by Elsevier in SSRN Electronic Journal
Multiplicative Layer Corrector PINN for 2D Singularly Perturbed Reaction–Diffusion Problems
The singularly perturbed problems exhibit sharp layer behavior, which pose major challenges for both traditional numerical methods and standard PhysicsInformed Neural Networks (PINNs). To address this class of problems, the Singular Layer Physics-Informed Neural Network framework incorporates asymptotic layer corrections to enhance the learning of steep solution gradients which was introduced by Gie et al[2]in 2026.The SL-PINN fails to solve 2-D singularly perturbed reaction diffusion problems with multiple layer interactions.In this work, we propose and analyse a layer-aware physics-informed neural network (PINN) that employs a multiplicative layer corrector, with the decay rate of the corrector introduced as a learnable parameter. The decay rate of the layer corrector allows the network to adaptively infer the effective layer thickness during training.
- 21 May 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessMultiplicative Layer Corrector PINN for 2D Singularly Perturbed Reaction–Diffusion Problems
The singularly perturbed problems exhibit sharp layer behavior, which pose major challenges for both traditional numerical methods and standard PhysicsInformed Neural Networks (PINNs). To address this class of problems, the Singular Layer Physics-Informed Neural Network framework incorporates asymptotic layer corrections to enhance the learning of steep solution gradients which was introduced by Gie et al[2]in 2026.The SL-PINN fails to solve 2-D singularly perturbed reaction diffusion problems with multiple layer interactions.In this work, we propose and analyse a layer-aware physics-informed neural network (PINN) that employs a multiplicative layer corrector, with the decay rate of the corrector introduced as a learnable parameter. The decay rate of the layer corrector allows the network to adaptively infer the effective layer thickness during training.
- 18 March 2026
- Published by Springer Nature in Theoretical and Applied Climatology
- Vol. 157 (4) , 223
Estimation of location-specific precipitation using Deep Neural Networks
- 18 March 2026
- Published by Springer Nature in Theoretical and Applied Climatology
- Vol. 157 (4) , 223
Estimation of location-specific precipitation using Deep Neural Networks
- Open Access
- 13 November 2025
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 52 (22)
Spatially Inhomogeneous Phase Change, and Its Effect on Cloud Turbulence
We ask whether cloud turbulence differs from typical fluid turbulence, and answer in the affirmative: small‐scale turbulence is significantly enhanced in clouds by inertial droplets. This enhancement stems from a mismatch in time scales between vortex evacuation and condensation. Inertial droplets are centrifuged out of vortical regions much faster than condensation can occur. They then grow by condensation, releasing latent heat locally. Our simulations reveal that vortical regions co‐locate with cooler temperatures and higher moisture, leading to sharp density gradients at their boundaries. These gradients drive baroclinic torque, generating small‐scale turbulence. The time scales involved align with our scaling estimates. While our earlier two‐dimensional model anticipated this effect, it underestimated its intensity and limited it to sub‐Kolmogorov scales. The enhancement of turbulence over inertial scales that we find here could drive rapid raindrop growth.
- 13 November 2025
- Published by American Geophysical Union (AGU) in Geophysical Research Letters
- Vol. 52 (22)
Open AccessSpatially Inhomogeneous Phase Change, and Its Effect on Cloud Turbulence
We ask whether cloud turbulence differs from typical fluid turbulence, and answer in the affirmative: small‐scale turbulence is significantly enhanced in clouds by inertial droplets. This enhancement stems from a mismatch in time scales between vortex evacuation and condensation. Inertial droplets are centrifuged out of vortical regions much faster than condensation can occur. They then grow by condensation, releasing latent heat locally. Our simulations reveal that vortical regions co‐locate with cooler temperatures and higher moisture, leading to sharp density gradients at their boundaries. These gradients drive baroclinic torque, generating small‐scale turbulence. The time scales involved align with our scaling estimates. While our earlier two‐dimensional model anticipated this effect, it underestimated its intensity and limited it to sub‐Kolmogorov scales. The enhancement of turbulence over inertial scales that we find here could drive rapid raindrop growth.
- Open Access
- 5 September 2025
- Published by Wiley
Spatially inhomogeneous phase change, and its effect on cloud turbulence
We ask whether cloud turbulence differs from typical fluid turbulence, and answer in the affirmative: small-scale turbulence is significantly enhanced in clouds by inertial droplets. This enhancement stems from a mismatch in time scales between vortex evacuation and condensation. Inertial droplets are centrifuged out of vortical regions much faster than condensation can occur. They then grow by condensation, releasing latent heat locally. Our simulations reveal that vortical regions co-locate with cooler temperatures and higher moisture, leading to sharp density gradients at their boundaries. These gradients drive baroclinic torque, generating small-scale turbulence. The time scales involved align with our scaling estimates. While our earlier two-dimensional model anticipated this effect, it underestimated its intensity and limited it to sub-Kolmogorov scales. The enhancement of turbulence over inertial scales that we find here could drive rapid raindrop growth.
- 5 September 2025
- Published by Wiley
Open AccessSpatially inhomogeneous phase change, and its effect on cloud turbulence
We ask whether cloud turbulence differs from typical fluid turbulence, and answer in the affirmative: small-scale turbulence is significantly enhanced in clouds by inertial droplets. This enhancement stems from a mismatch in time scales between vortex evacuation and condensation. Inertial droplets are centrifuged out of vortical regions much faster than condensation can occur. They then grow by condensation, releasing latent heat locally. Our simulations reveal that vortical regions co-locate with cooler temperatures and higher moisture, leading to sharp density gradients at their boundaries. These gradients drive baroclinic torque, generating small-scale turbulence. The time scales involved align with our scaling estimates. While our earlier two-dimensional model anticipated this effect, it underestimated its intensity and limited it to sub-Kolmogorov scales. The enhancement of turbulence over inertial scales that we find here could drive rapid raindrop growth.
- Open Access
Towards Location-Specific Precipitation Projections Using Deep Neural Networks
Accurate precipitation estimates at individual locations are crucial for weather forecasting and spatial analysis. This study presents a paradigm shift by leveraging Deep Neural Networks (DNNs) to surpass traditional methods like Kriging for station-specific precipitation approximation. We propose two innovative NN architectures: one utilizing precipitation, elevation, and location, and another incorporating additional meteorological parameters like humidity, temperature, and wind speed. Trained on a vast dataset (1980-2019), these models outperform Kriging across various evaluation metrics (correlation coefficient, root mean square error, bias, and skill score) on a five-year validation set. This compelling evidence demonstrates the transformative power of deep learning for spatial prediction, offering a robust and precise alternative for station-specific precipitation estimation.
Open AccessTowards Location-Specific Precipitation Projections Using Deep Neural Networks
Accurate precipitation estimates at individual locations are crucial for weather forecasting and spatial analysis. This study presents a paradigm shift by leveraging Deep Neural Networks (DNNs) to surpass traditional methods like Kriging for station-specific precipitation approximation. We propose two innovative NN architectures: one utilizing precipitation, elevation, and location, and another incorporating additional meteorological parameters like humidity, temperature, and wind speed. Trained on a vast dataset (1980-2019), these models outperform Kriging across various evaluation metrics (correlation coefficient, root mean square error, bias, and skill score) on a five-year validation set. This compelling evidence demonstrates the transformative power of deep learning for spatial prediction, offering a robust and precise alternative for station-specific precipitation estimation.
- 16 April 2024
- Published by Elsevier in Physics and Chemistry of the Earth, Parts A/B/C
- Vol. 135
Utilizing deep learning for near real-time rainfall forecasting based on Radar data
- 16 April 2024
- Published by Elsevier in Physics and Chemistry of the Earth, Parts A/B/C
- Vol. 135
Utilizing deep learning for near real-time rainfall forecasting based on Radar data
- 12 July 2023
- Published by Springer Nature in Modeling Earth Systems and Environment
- Vol. 10 (1) , 927-941
Harnessing deep learning for forecasting fire-burning locations and unveiling $PM_{2.5}$ emissions- 12 July 2023
- Published by Springer Nature in Modeling Earth Systems and Environment
- Vol. 10 (1) , 927-941
Harnessing deep learning for forecasting fire-burning locations and unveiling $PM_{2.5}$ emissions