We study how particles — cloud droplets, ice crystals, sediment, microplastics, microswimmers — interact with the flows that carry them. The methods are theoretical and computational: asymptotic analysis, hydrodynamic stability theory, and direct numerical simulation. The questions connect microscale particle physics to macroscale geophysical phenomena — from droplet collisions and rain initiation in warm clouds to wind-wave generation and cyclone intensification over tropical oceans.
Droplet collisions, ice-crystal dynamics, charge transfer and precipitation initiation in warm and mixed-phase clouds.
Jump to section ↓
Inertial particles in vortices, suspensions in falling films, pulsatile flows, anisotropic particles and microswimmers.
Jump to section ↓
Internal gravity waves, surface gravity waves, stratified shear instabilities and viscoelastic flow transitions.
Jump to section ↓Warm cumulus clouds can produce rain within 15–20 minutes of formation, although condensational growth alone predicts timescales of hours. The gap is bridged by collision-coalescence: turbulence, differential sedimentation, electrostatics, and short-range hydrodynamic forces act together to accelerate droplet growth across the condensation-coalescence bottleneck. Our group develops the theory and computation needed to quantify this process from first principles.
Using direct numerical simulations (DNS), stochastic collision models, and pairwise hydrodynamic interaction theory, we have quantified how turbulence, inertia, gravity-induced differential sedimentation, and short-range forces (van der Waals, non-continuum lubrication) influence collisional growth of cloud droplets. Our collision kernels incorporate both hydrodynamic and environmental factors, providing more accurate inputs for models that couple microphysical processes with cloud-scale dynamics.
Cloud droplets carry electric charge, and a naïve point-charge argument suggests that like-charged droplets should always repel each other. Our work has shown that like-charged dielectric spheres can attract each other at close separations: induced-dipole interactions arising from finite droplet size overpower the point-charge repulsion, transforming the dynamics into a near-field attractive force. This result bears directly on collision processes during lightning discharge and in electrified cloud environments. At high charge ratios, the collision efficiency between like-charged droplets significantly exceeds that of uncharged pairs.
A parallel thread of our research focuses on ice microphysics in mixed-phase and convective clouds. Ice particle growth, aggregation, and charging play critical roles in hail formation, lightning initiation, and storm intensity. Unlike fluid droplets, ice particles introduce additional complexities: diverse crystal habits, anisotropic mechanical behaviour, and quasi-liquid layers. We develop models for ice–ice collision outcomes, rebound dynamics, charge transfer, and fluid-mediated reorientation during settling and turbulent motion.
Our earlier work has shown that sedimenting asymmetric rods can achieve nontrivial alignment due to the competition between gravitational and hydrodynamic torques — a transition identified as a supercritical pitchfork bifurcation, validated experimentally. This alignment of ice crystals by inertial torques produces the atmospheric optical phenomenon known as sundogs, and the competition between gravitational alignment and turbulent randomisation is central to understanding electromagnetic scattering from icy clouds — one of the largest sources of errors in global climate models.
We develop multiscale simulation frameworks that combine particle-resolved microscale simulations with mesoscale cloud-resolving or large eddy simulations (LES). This includes developing super-droplet models (SDM) that represent billions of droplets via reduced statistical particles, demonstrating how accurate collision kernels from DNS can be embedded within SDM frameworks to improve rain initiation modelling in LES. This effort is strengthened by ongoing collaboration with the Mesoscale and Microscale Meteorology (MMM) Lab at NCAR, where insights from our group help shape next-generation cloud-resolving models, and with researchers at Microsoft toward physics-informed surrogate models for scalable, cloud-based forecasting systems.
To connect in situ cloud processes to remotely sensed observations, we have developed numerical tools for evaluating electromagnetic scattering by cloud particles. Using a Lattice Boltzmann Method-based electromagnetic solver capable of handling complex particle geometries and irregular orientations, we link microphysics outputs to radar reflectivity and optical signatures for conducting, dielectric and bi-dielectric (Janus) particles.
Flows laden with particles are ubiquitous — from dust devils and atmospheric cyclones to coating processes and the respiratory airways. When particles have inertia, they don't passively follow the fluid; they centrifuge out of vortices, cluster in straining regions, alter the carrier-phase rheology, and can destabilise flows that would otherwise remain stable. Our group investigates these phenomena across scales, from granular media and fibre suspensions to microswimmers, combining theoretical analysis with computational modelling.
Gravity-driven interactions among charged dielectric spheres in a gaseous medium are examined, revealing that non-continuum lubrication interactions drive contact within finite time. Attraction is observed between closely approaching dielectric spheres with similar charges across various size and charge ratios. We have derived the asymptotic interparticle force under lubrication limits for arbitrary size ratios, showing how the attractive electric force diverges as separation approaches zero, overcoming continuum lubrication resistance.
Vortical flows suspended with heavy inertial particles occur widely — from dust devils and atmospheric cyclones to Jupiter's Great Red Spot. Heavy particles centrifuge away from vortical regions and accumulate in straining regions. We model the dynamics of inertial particles in flows generated by isolated elliptic vortex patches and study their clustering behaviour. Particles sample the flow field according to stable and unstable manifolds, and the inclusion of external shear induces chaos in particle trajectories.
Active suspensions — assemblies of self-propelling micro-organisms — exhibit remarkable collective behaviour driven by the interplay between self-generated flows and confinement. We explore how geometric confinement creates flow instabilities that give rise to collective motion. Modelling micro-swimmers as force dipoles, we have uncovered instabilities driven by the interplay between active stress and perturbations in swimmer density. An attractant gradient introduces anisotropy to the orientation field, revealing a novel instability mode driven by interface deformation due to active stress — even in suspensions of chemotactic pullers previously considered stable.
In surface gravity waves, an anisotropic particle exhibits orbital motion whose character depends on its orientation. A negatively buoyant particle settling into a deep flow field couples translational and rotational dynamics: as the particle settles, its orientation evolves in time and reaches a steady state, with different steady orientations leading to qualitatively different trajectories.
The dynamics of inertial particles in pulsatile (Womersley) flows are crucial for understanding biological processes and designing microfluidic devices. In collaboration with Prof. Mahesh Panchagnula (IIT Madras), we study the transport of particulate matter in respiratory airways using Maxey-Riley and Langevin models, and investigate dispersion in axially diverging oscillatory conduits with focus on the interplay of particle inertia, temporal oscillation, and geometric divergence.
Falling liquid films laden with particles appear in industrial coating processes, the tear film protecting our eyes, and many natural settings. We studied the stability of particle-laden gravity-driven falling films and discovered that the viscosity stratification created by suspended particles can solely destabilise the flow — a finding highlighted by the J. Fluid Mech. Focus on Fluids. In parallel, we characterised the instability of a two-dimensional Rankine vortex laden with inertial particles using linear stability analysis, revealing mechanisms driving instability in dusty semi-dilute flows.
Active suspensions display medium-like behaviour arising from the self-propulsion of their constituent organisms. We explore the role of geometric confinement on the creation of flow instabilities that eventually give rise to collective motion. Extending from micron-sized swimmers to virus particles at the nanometre scale, we also investigate the role of convective fluid motion in the aerosol transmission of diseases.
Oceans and atmospheres host a rich spectrum of wave phenomena across length and time scales, from internal gravity waves deep in the stratified ocean to wind-generated surface waves at the air-sea interface. These waves sustain large-scale circulations, mediate energy transfer, and undergo nonlinear interactions and instabilities that bear directly on weather and climate. Our group combines linear stability analysis, non-modal theory, and direct numerical simulation to understand these processes.
In a density-stratified ocean or atmosphere, displaced fluid parcels oscillate and radiate internal waves that contribute to deep-ocean mixing and sustain large-scale circulations. These highly nonlinear waves interact through triadic resonances, exchanging energy and amplifying. We study the underlying nonlinear interactions and resonances amongst internal wave modes, including the effects of background shear.
Surface gravity waves, generated by wind over a disturbed interface, are parameterised into global climate models since they regulate fluxes of heat and momentum between ocean and atmosphere. We perform linear stability calculations to identify parameter regimes of instability, determine growth rates, and validate predictions with DNS. Our work addresses the Miles mechanism (energy transfer from a critical layer in air to the interface), the rippling instability from the critical layer in water, and their viscous counterparts.
Stratified shear flow instability plays a key role in geophysical turbulence and mixing. Our group has studied the linear stability and nonlinear evolution of various stratified instabilities — Kelvin-Helmholtz, Holmboe, and Taylor instabilities. Not all stratified shear flows are linearly unstable, yet they still transition to turbulence. We have identified an algebraic instability in stably stratified shear flows where the eigenspectrum is purely continuous, followed by optimal perturbation analysis to find initial conditions that maximise energy amplification.
Vortex column stability has been studied since Lord Kelvin's classical work in 1880. We have identified the singular eigenfunctions of a vortex column and shown how their inclusion completes the spectrum, enabling a description of how a vortex interacts with external turbulence — a feature previously understood only through the initial value problem. In parallel, we discovered a novel inertio-elastic instability of a vortex in a dilute polymer solution, arising from the resonance of elastic shear waves aided by background shear. This connects to the broader question of turbulent drag reduction: why adding a few parts per million of polymer to a turbulent flow can reduce drag by up to 80%.
Our research draws on two communities that rarely overlap — suspension mechanics and geophysical fluid dynamics. The directions below sit at this interface and are the ones we are actively pursuing.
Physically-based, scale-aware models for droplet collision, coalescence and breakup across turbulence regimes and size distributions — integrating high-fidelity DNS, stochastic frameworks, super-droplet LES, and machine-learning surrogates, in collaboration with NCAR-MMM. (ANRF project on multiscale modelling of warm rain, with IIT Bombay.)
Physically consistent collision and rebound models for realistic ice habits, quantifying charge transfer during ice–ice interactions under turbulent flow, and embedding charge models into cloud-resolving simulations to predict charge stratification and lightning precursors.
Extending Lattice Boltzmann electromagnetic solvers to wet, irregular particle shapes, generating scattering lookup datasets, and coupling them with evolving particle states from LES to support cloud profiling radars and Earth observation missions.
How the droplet-laden boundary layer between ocean and atmosphere modifies surface drag and heat transfer — acting as a lubrication layer that accelerates cyclones. We develop modified Kolmogorov-Prandtl models and two-way coupled DNS. (SERB-SUPRA project, with IIT Bombay & IIT Ropar.)
Classical Stokes drift applies only to tracers, but microplastics carry inertia and anisotropy. We study particle transport in flows combining waves and turbulence, and simulate Langmuir turbulence via the Craik-Leibovich equations to understand vertical transport of finite-sized particles.
Contrail cirrus is the largest non-CO2 contribution of aviation to climate forcing, and its microphysics has been studied almost entirely in mid-latitude airspace. We are extending our parcel models, generative machine-learning tools, and electromagnetic scattering solvers to aviation-induced cirrus in the tropical upper troposphere — ice nucleation on soot in aircraft plumes, habit-resolved growth in ice-supersaturated regions, and first radiative-forcing estimates along Indian flight corridors.
Erosion, deposition and transport in coastal regions form a complex multiphase problem. Having shown that particulate microstructure can destabilise stable flows, we use Euler-Lagrange and Euler-Euler simulations to model particle dynamics and wave-field modulation by the dispersed phase.