About

Anand Gnanadesikan trained as an oceanographer with primary expertise in how the ocean circulates and influences marine life, climate, and atmospheric chemistry. He earned his PhD from the MIT/Woods Hole Joint Program in physical oceanography and has spent more than 15 years working at Princeton University and NOAA’s Geophysical Fluid Dynamics Laboratory developing computer models of planetary systems. Since joining the Johns Hopkins University faculty in January 2011, he has used numerical models and large datasets to explore the interactions between the ocean and the biosphere, atmosphere, and cryosphere. His research interests include the role of mixing in setting the large-scale circulation and chemical properties of the ocean, with particular focus on the impact of internal wave breaking and oceanic weather systems on vertical exchange. He investigates ocean biology and chemistry, especially oceanic tracers such as radiocarbon and open-ocean hypoxia, and how biological cycling can feedback on physical climate through alterations in the carbon cycle and ocean color. Additionally, Gnanadesikan studies variability in the Earth system, aiming to understand climate change on timescales from annual to multimillennial, with a focus on the Southern Ocean, tropical variability, and the North Atlantic.

Research topics

• Geology
• Environmental science
• Oceanography
• Atmospheric sciences
• Climatology
• Biology
• Ecology
• Meteorology
• Evolutionary biology
• Geography

Selected publications

• Deep Learning Surrogates for Emulating Stochastic Climate Tipping Dynamics

  arXiv (Cornell University) · 2026-05-20

  preprintOpen accessSenior author

  This work explores a dynamics-informed Temporal Fusion Transformer (TFT) as a data-driven surrogate for computationally intensive Earth system simulations. Focusing on multivariate time series describing global ocean transport, we demonstrate the surrogate's ability to forecast tip events across thousands of time steps. The data involve up to 21 non-stationary time series in addition to static covariates describing free parameters and initial conditions. Modifications to the architecture and objective function yield a surrogate that anticipates the timing of Atlantic and Pacific collapses to high fidelity and captures the stochastic uncertainty in transition timing across ensemble predictions. The learned surrogate achieves a 465x computational speedup over the numerical simulator while maintaining differentiability with respect to parameters and initial conditions.

  PublisherDOI

• Deep Learning Surrogates for Emulating Stochastic Climate Tipping Dynamics

  ArXiv.org · 2026-05-20

  articleOpen accessSenior author

  This work explores a dynamics-informed Temporal Fusion Transformer (TFT) as a data-driven surrogate for computationally intensive Earth system simulations. Focusing on multivariate time series describing global ocean transport, we demonstrate the surrogate's ability to forecast tip events across thousands of time steps. The data involve up to 21 non-stationary time series in addition to static covariates describing free parameters and initial conditions. Modifications to the architecture and objective function yield a surrogate that anticipates the timing of Atlantic and Pacific collapses to high fidelity and captures the stochastic uncertainty in transition timing across ensemble predictions. The learned surrogate achieves a 465x computational speedup over the numerical simulator while maintaining differentiability with respect to parameters and initial conditions.

  PublisherOA PDF

• Using Machine Learning to Uncover Ecological Mechanisms Controlling Abundance of Phytoplankton Size Classes From Large‐Scale Observations

  Global Biogeochemical Cycles · 2026-03-28

  articleOpen accessSenior author

  Abstract Phytoplankton size classes (PSCs) and Phytoplankton functional types (PFTs) determine many fundamental biogeochemical processes including nutrient uptake, energy transfer through marine food webs, ocean carbon export, and gas exchange with the atmosphere. Discerning the causes of spatio‐temporal variability of PSCs is a scientific priority for understanding the ocean's role in and response to climate change. This study intends to decipher the relationships between the abundance of PSCs and environmental predictors using explainable machine learning (XAI) techniques. The target variables were PSCs obtained using three different satellite products: size‐resolved phytoplankton carbon from Kostadinov, Milutinović, et al. (2016), https://doi.org/10.5194/os‐12‐561‐2016 , chlorophyll from MODIS divided according to the algorithm of Hirata et al. (2011), https://doi.org/10.5194/bg‐8‐311‐2011 , and a third product from the Copernicus Marine Service. The environmental predictors were nutrients, light, mixed layer depth, salinity, sea surface temperature (sst), and upwelling. The ML algorithm used was the Random Forest Regressor (RFR). XAI techniques were used to discern the relationship between predictors and PSCs abundance. About 85%–95% of the variability of the size classes in the observational data sets was accounted for by environmental variables known to influence phytoplankton biomass. Although different size classes responded similarly to the environmental drivers (with the exception of Copernicus picoplankton) their scale of response varied. The dominant predictors were found to be shortwave radiation, ammonia, dissolved iron and sea surface temperature. The different satellite products show sensitivity to iron, shortwave radiation, sst and ammonia across the same range of values, but with different magnitudes. Copernicus picoplankton is the only product which is positively related to sst.

  PublisherDOI

• Mechanistic Modeling of Aedes aegypti Mosquito Habitats for Climate-informed Dengue Forecasting

  2025-02-06

  preprintOpen access

  The mosquito-borne disease dengue is sensitive to climate, in part because of the influence climate has on breeding habitats of dengue’s Aedes mosquito vectors. Dengue risk assessment models currently leverage climate-dengue statistical associations, yet what remain understudied are the mechanistic pathways that yield different statistical relationships in different locations. We hypothesize that elucidating the mechanisms by which spatiotemporal variability in climate influences dengue incidence will improve dengue dynamics predictions across climatically distinct locations and beyond dengue’s well-known seasonal cycles. We test this hypothesis by investigating a key pathway in the climate-dengue process chain: climate impacts on Aedes breeding habitats. We have implemented a mechanistic modeling pipeline that simulates climatic influence on habitat water dynamics and thereby on the relative population size of the vector. We use this modeling pipeline, driven by meteorological data, to simulate monthly Aedes populations for three climatically distinct cities in Sri Lanka. We find that simulated vector abundance is plausibly associated with climate conditions and that climate drivers of vector abundance vary among locations. Moreover, tercile-tercile comparisons of dengue incidence against model variables indicate that risk assessments based on predicted vector abundance perform similarly to those based on meteorology alone—the signal of weather variability and its relationship to dengue propagates through the modeling pipeline. These results justify future assessment of this modeling pipeline’s utility within a dengue risk assessment framework, where its process-based structure may be leveraged to inform intervention strategies in high-risk years and to simulate impacts of future climate conditions on dengue dynamics.

  PublisherOA PDFDOI

• Machine Learning Methods Suggest That Large Regional Changes in Phytoplankton Biomass Produced by Earth System Models Do Not Reflect Realistic Responses to Changing Climate

  Global Biogeochemical Cycles · 2025-09-26 · 1 citations

  article1st authorCorresponding

  Abstract Surface phytoplankton biomass (measured in mol C m −3 ) represents a critical parameter within the Earth System that is measured from space and simulated in Earth System Models. Under climate change, the current generation of Earth System Models agrees that low‐latitude biomass will decline and high‐latitude biomass will increase. However, on a regional scale, the magnitude, phenology and spatial pattern of these changes are highly inconsistent across models. We use machine learning to investigate the sources of the divergence and evaluate the realism of the simulations. We train Random Forests driven by environmental drivers to simulate surface phytoplankton biomass under both pre‐industrial control and SSP5‐8.5 scenarios. Outside the Arctic, the bulk of the changes in biomass are driven by rearrangements in the spatiotemporal distribution of environmental predictors. Large regional changes in models, however, are associated either with unrealistically low pre‐industrial levels of macronutrients or unrealistically strong responses to those macronutrients. Within the Arctic, relationships between environmental predictors and biomass change under global warming. While increased light drives increased biomass, the effect is largest in models with a high nutrient bias. Feeding inputs from an ensemble of models to an emulator trained on observations predicts observed biomass better than the ensemble of the models does, highlighting the fact that models do not produce the correct relationships between environmental predictors and biomass. However, this technique does not yield mechanistically consistent predictions of biomass under climate change. Skepticism of large regional changes in surface phytoplankton biomass produced by individual models is warranted.

  PublisherDOI

• Mechanistic Modeling of <i>Aedes aegypti</i> Mosquito Habitats for Climate‐Informed Dengue Forecasting

  GeoHealth · 2025-09-01

  articleOpen access

  Abstract The mosquito‐borne disease dengue is sensitive to climate, in part because of the influence climate has on breeding habitats of dengue's Aedes mosquito vectors. Dengue risk assessment models currently leverage climate‐dengue statistical associations, yet what remain understudied are the mechanistic pathways that yield different statistical relationships in different locations. We hypothesize that elucidating the mechanisms by which spatiotemporal variability in climate influences dengue incidence will improve dengue dynamics predictions across climatically distinct locations and beyond dengue's well‐known seasonal cycles. We test this hypothesis by investigating a key pathway in the climate‐dengue process chain: climate impacts on Aedes breeding habitats. We have implemented a mechanistic modeling pipeline that simulates climatic influence on habitat water dynamics and thereby on relative population size of the vector. We use this modeling pipeline, driven by meteorological data, to simulate monthly Aedes populations for three climatically distinct cities in Sri Lanka. We find that simulated vector abundance is plausibly associated with climate conditions and that climate drivers of vector abundance vary among locations. Moreover, tercile‐tercile comparisons of dengue incidence against model variables indicate that risk assessments based on predicted vector abundance perform similarly to those based on meteorology alone—the signal of weather variability and its relationship to dengue propagates through the modeling pipeline. These results justify future testing of this modeling pipeline within a dengue risk assessment framework, where its process‐based structure may be leveraged to guide proactive dengue control efforts in high‐risk years and to simulate impacts of future climate conditions on dengue dynamics.

  PublisherDOI

• Using Random Forests to Emulate and Understand the Phytoplankton Dynamics in High Latitudes in 2 Earth System Models under Climate Change

  2025-02-04

  preprintSenior author
  PublisherDOI

• Deep Learning of Systematic Ocean Model Errors in a Coupled GCM from Data Assimilation Increments

  2025-12-04

  articleSenior author

  We present a novel, data-driven approach to predict systematic model errors in the ocean component of a coupled general circulation model leveraging deep learning and data assimilation. We examine the skill of the proposed scheme in learning systematic model errors, including their spatial patterns, variance, scales, and test its sensitivity to different predictors and neural network architecture. The scheme utilizes local state variables such as ocean temperature, salinity, velocities, and surface fluxes to predict corrections to temperature tendency for the upper 1000 meters in the ocean on daily timescales. The performance is evaluated on the withheld test dataset and compared against the empirical climatological temperature corrections that are geographically dependent. The performance is depth-dependent, with significant improvements over the benchmark in the upper 20 meters in the ocean. It degrades rapidly with depth but remains comparable to the climatology benchmark. Neural networks can capture up to 40-50% of the daily variance in temperature increments in the upper 20 meters relative to the benchmark’s 20%. The improvements are associated with networks predicting finer spatiotemporal scales than the benchmark. They are expected to perform better in reducing surface ocean mixed layer bias than previously used techniques. Despite being column-local without geographical inputs, networks can sufficiently reproduce spatial patterns on daily and longer timescales. The patterns consist of corrections to regional dynamical features such as western boundary currents, equatorial undercurrents, bathymetry-related corrections in the Southern Ocean, and warm surface increments over subtropical and midlatitude belts.

  PublisherDOI

• Simulating Aedes Mosquito Habitats for Climate-informed Dengue Forecasting

  2025-08-13

  preprintOpen access

  The mosquito-borne disease dengue is dependent on climate, in part because of the climate sensitivities of its Aedes mosquito vectors. These climate–dengue associations are leveraged in dengue risk assessment models, which incorporate statistical relationships between meteorological variables (e.g., rainfall, air temperature) and dengue cases. However, the mechanistic pathways that result in different statistical relationships in different locations remains understudied. A mechanistic understanding of how interannual variability in climate translates to interannual variability in dengue incidence may aid in predicting dengue dynamics across climatically distinct locations and beyond dengue’s well-known seasonal cycles. In this work we investigate interannual variability within a key pathway in the climate-disease process chain: climate impacts on the Aedes vectors’ breeding habitats. We have implemented a process-based modeling pipeline that simulates climatic influence on habitat water dynamics (using the energy balance container model WHATCH’EM) and thereby on the relative population size of the vector (using models of environment-dependent vector survival and development rates for eggs, larvae, and pupae). We use this modeling pipeline, driven by meteorological data from 2001 to 2020, to simulate monthly relative population sizes of the vector for three climatically distinct locations in Sri Lanka (Negombo, Jaffna, Nuwara Eliya). We find that modeled vector population sizes are plausibly associated with climate conditions (e.g., in the cold, mountainous city of Nuwara Eliya, years of high 2-m air temperature tend to yield high populations) and that the climate drivers of population vary among locations. Moreover, tercile-tercile comparisons of dengue against model variables support the idea that dengue’s connection to temperature is strongly affected by the effect of temperature on vector population size. These results justify future work on rigorously assessing this modeling pipeline’s utility within a dengue risk assessment framework, where its process-based structure may be leveraged to inform intervention strategies in high-risk years and simulate impacts of future climate conditions.

  PublisherOA PDFDOI

• Tools in Harmony: Integrating Observations and Models for Improved Understanding of a Changing Ocean

  Oceanography · 2025-01-01 · 1 citations

  articleOpen access
  PublisherOA PDFDOI

Recent grants

• Collaborative Research: Southern Ocean Convection in climate models: controls and Impacts

  NSF · $388k · 2018–2022

Frequent coauthors

• Marie‐Aude Pradal

  Johns Hopkins University

  51 shared

• John P. Dunne

  36 shared

• Jorge L. Sarmiento

  34 shared

• Stephen M. Griffies

  NOAA Geophysical Fluid Dynamics Laboratory

  33 shared

• Richard D. Slater

  Royal Alexandra Hospital

  27 shared

• Larry W. Horowitz

  National Oceanic and Atmospheric Administration

  26 shared

• H. S. Chen

  25 shared

• Antony K. Liu

  Goddard Space Flight Center

  25 shared