AI Paper Atlas

Research map / Probabilistic and scientific modeling / Variational and generative sampling

Spatiotemporal Pyramid Flow Matching for Climate Emulation

Jeremy A. Irvin, Jiaqi Han, Zikui Wang, Abdulaziz Alharbi, Yufei Zhao, Nomin-Erdene Bayarsaikhan, Daniele Visioni, Andrew Y. Ng, Duncan Watson-Parris

CVPR 2026 · Probabilistic and scientific modeling › Variational and generative sampling · approach: Flow matching and flow sampling · 3 citations (OpenAlex)

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
SPF vs autoregressive Pyramidal Flow at yearly scale (200M)SPF · ClimateBench yearlyCRPS: 0.327 -> 0.222Improved
parallel SPF instead of autoregressive Pyramidal Flow, runtimeSPF · ClimateBench yearly 10-year samplingruntime (s): 19 -> 3Improved
SPF vs PixelFlow at yearly RMSE (100M)SPF · ClimateBench yearlyRMSE: 0.549 -> 0.565Hurt
SPF vs Multi-Monthly Flow at 200M, monthlySPF · ClimateBench monthlyCRPS: 0.443 -> 0.453Hurt
use 3-stage pyramid instead of 5-stage variants (ablation)SPF · ClimateBench monthlyCRPS: 0.474 -> 0.453Improved
alternate temporal and spatial transitions (DYYMM vs DYMMM) (ablation)SPF 5-stage · ClimateBench monthlyCRPS: 0.474 -> 0.463Improved
add multi-timescale training (DYM-Any vs DYM-Monthly) (ablation)SPF · ClimateBench monthlyRMSE: 1.064 -> 1.060Improved
scale to 600M parameters without pretraining (ablation)SPF · ClimateBench yearlyCRPS: 0.222 -> 0.229Hurt
scale 200M to 600M (ablation)SPF · ClimateBench monthlyCRPS: 0.453 -> 0.442Improved
add ClimateSuite pretraining (ablation)SPF 600M · ClimateBench yearlyCRPS: 0.229 -> 0.216Improved
add ClimateSuite pretraining (ablation)SPF 600M · ClimateBench monthlyCRPS: 0.442 -> 0.432Improved
SPF vs UNet multi-model emulationSPF 600M · ClimateSuite held-out SAI UKESM1-0-LLCRPS: 0.466 -> 0.315Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 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