AI Paper Atlas

Research map / Probabilistic and scientific modeling / Differential equations and PDEs

DeepPrim: a Physics-Driven 3D Short-term Weather Forecaster via Primitive Equation Learning

Jiawei Chen, Weiqi Chen, Rong Hu, Peiyuan Liu, Haifan Zhang, Liang Sun

ICLR 2026 · Probabilistic and scientific modeling › Differential equations and PDEs · approach: Physics-informed operator learning

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add 3D modules (cross-pressure self-attention and pressure-level derivatives) (ablation)DeepPrim · ERA5 global forecastingRMSE: average degradation of 12.3% when removedImproved
add temporal variation and 3D spatial derivative of initial conditions to initialization network (ablation)DeepPrim · ERA5 global forecastingRMSEImproved
add the 3D spatial derivative term in initialization (ablation)DeepPrim · ERA5 global forecastingRMSEImproved
add source-sink network (ablation)DeepPrim · ERA5 global forecastingRMSEImproved
train on only five key output variables instead of all variablesDeepPrim · ERA5 5.625 degRMSE: DeepPrim† generally outperforms full-output-variable DeepPrimImproved
use DeepPrim instead of ClimODEDeepPrim · ERA5 5.625 degRMSE: outperforming ClimODE by 38.6%Improved
use DeepPrim instead of pre-trained ClimaXDeepPrim · ERA5 5.625 degRMSE: by 10.7% on averageImproved
use DeepPrim instead of pre-trained ClimaXDeepPrim · ERA5 1.40625 degRMSE: by 36.11% on averageImproved
use DeepPrim instead of GraphCastDeepPrim · ERA5 0.25 deg z500RMSE: 44.8 vs 41.0Hurt
use DeepPrim instead of second-best baselineDeepPrim · regional forecasting (3 regions)RMSE: by 35.8% on averageImproved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy