Source Themes

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.

We introduce CAMulator version 1, an auto-regressive machine-learned (ML) emulator of the Community Atmosphere Model version 6 (CAM6) that simulates the next atmospheric state given the prescribed sea surface temperatures and incoming solar radiation. CAMulator explicitly conserves global dry air mass, moisture, and total atmospheric energy while remaining numerically stable over indefinite climate integrations. It successfully reproduces the annual CAM6 climatology and key modes of climate variability, including the El Niño-Southern Oscillation, the North Atlantic Oscillation, and the Pacific-North American pattern, with slightly muted variability. When forced with sea surface temperature (SST) outside the training distribution, CAMulator exhibits a systematic cold bias in high-latitude regions, particularly in boreal winter, likely due to the absence of interactive land and sea ice. Nonetheless, CAMulator achieves these results with a 350 times speedup over CAM6, making it an efficient alternative for generating large ensembles.

Accurately quantifying air-sea fluxes is important for understanding air-sea interactions and improving coupled weather and climate systems. This study introduces a probabilistic framework to represent the highly variable nature of air-sea fluxes, which is missing in deterministic bulk algorithms. Assuming Gaussian distributions conditioned on the input variables, we use artificial neural networks and eddy-covariance measurement data to estimate the mean and variance by minimizing negative log-likelihood loss. The trained neural networks provide alternative mean flux estimates to existing bulk algorithms, and quantify the uncertainty around the mean estimates. Stochastic parameterization of air-sea turbulent fluxes can be constructed by sampling from the predicted distributions. Tests in a single-column forced upper-ocean model suggest that changes in flux algorithms influence sea surface temperature and mixed layer depth seasonally. The ensemble spread in stochastic runs is most pronounced during spring restratification.

Reliable dynamic sea level forecasts are hindered by numerous sources of uncertainty on daily-to-seasonal timescales (1-180 days) due to atmospheric boundary conditions and internal ocean variability. Studies have demonstrated that certain initial states can extend predictability horizons; thus, identifying these initial conditions may help improve forecast skill. Here, we identify sources of dynamic sea level predictability on daily-to-seasonal timescales using neural networks trained on CESM2 large ensemble data to forecast dynamic sea level. The forecasts yield not only a point estimate for sea level but also a standard deviation to quantify forecast uncertainty based on the initial conditions. Forecasted uncertainties can be leveraged to identify state-dependent sources of predictability at most locations and forecast leads. Network forecasts, particularly in the low-latitude Indo-Pacific, exhibit skillful deterministic predictions and skillfully forecast exceedance probabilities relative to local linear baselines. For networks trained at Guam and in the western Indian Ocean, the transfer of sources of predictability from local sources to remote sources is presented by the deteriorating utility of initial condition information for predicting exceedance events. Propagating Rossby waves are identified as a potential source of predictability for dynamic sea level at Guam. In the Indian Ocean, persistence of thermosteric sea level anomalies from the Indian Ocean Dipole may be a source of predictability on subseasonal timescales, but El Niño drives predictability on seasonal timescales. This work shows how uncertainty-quantifying machine learning can help identify changes in sources of state-dependent predictability over a range of forecast leads.

Due to the wide range of processes impacting the sea surface height (SSH) on daily-to-interannual timescales, SSH forecasts are hampered by numerous sources of uncertainty. While statistical-dynamical methods like Linear Inverse Modeling have been successful at making forecasts, they often rely on assumptions that can be hard to satisfy given the nonlinear dynamics of the climate. Here, we train convolutional autoencoders with a dynamical propagator in the latent space to generate forecasts of SSH anomalies. Learning a nonlinear dimensionality reduction and the prediction timestepping together results in a propagator that produces better predictions for daily- and monthly-averaged SSH in the North Pacific and Atlantic than if the dimensionality reduction and dynamics are learned separately. The reconstruction skill of the model highlights regions in which better representation results in improved predictions: in particular, the tropics for North Pacific daily SSH predictions and the Caribbean Current for the North Atlantic.

AI emulators for forecasting have emerged as powerful tools that can outperform conventional numerical predictions. The next frontier is to build emulators for long climate simulations with skill across a range of spatiotemporal scales, a particularly important goal for the ocean. Our work builds a skillful global emulator of the ocean component of a state-of-the-art climate model. We emulate key ocean variables, sea surface height, horizontal velocities, temperature, and salinity, across their full depth. We use a modified ConvNeXt UNet architecture trained on multidepth levels of ocean data. We show that the ocean emulator - Samudra - which exhibits no drift relative to the truth, can reproduce the depth structure of ocean variables and their interannual variability. Samudra is stable for centuries and 150 times faster than the original ocean model. Samudra struggles to capture the correct magnitude of the forcing trends and simultaneously remains stable, requiring further work.

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Duis posuere tellus ac convallis placerat. Proin tincidunt magna sed ex sollicitudin condimentum.