Let it Snow! Animating Static Gaussian Scenes With Dynamic Weather Effects

TL;DR Our method animates static 3D Gaussian scenes with physically-based dynamic weather effects.

Above we illustrate our method with different weather effects applied to a static 3D Gaussian scene. Our hybrid framework enables animation of physically-based weather effects with realistic motion and appearance.

Abstract

3D Gaussian Splatting has recently enabled fast and photorealistic reconstruction of static 3D scenes. However, introducing dynamic elements that interact naturally with such static scenes remains challenging. Accordingly, we present a novel hybrid framework that combines Gaussian-particle representations for incorporating physically-based global weather effects into static 3D Gaussian Splatting scenes, correctly handling the interactions of dynamic elements with the static scene. We follow a three-stage process: we first map static 3D Gaussians to a particle-based representation. We then introduce dynamic particles and simulate their motion using the Material Point Method (MPM). Finally, we map the simulated particles back to the Gaussian domain while introducing appearance parameters tailored for specific effects. To correctly handle the interactions of dynamic elements with the static scene, we introduce specialized collision handling techniques. Our approach supports a variety of weather effects, including snowfall, rainfall, fog, and sandstorms, and can also support falling objects, all with physically plausible motion and appearance. Experiments demonstrate that our method significantly outperforms existing approaches in both visual quality and physical realism.

Examples of Dynamic Effects

Select a scene and then choose one of the dynamic effects to view the animation enabled by our physics-based approach.
As illustrated in this interactive visualization, our framework animates 3D Gaussian scenes with realistic dynamic effects that naturally interact with scene geometry and evolve over time.

How does it work?

🌍 Given a set of multi-view images, we first reconstruct a static scene using 3D Gaussian Splatting (3DGS).

💡 We map the static Gaussians to the particle world and initialize dynamic particles and simulate them using the Material Point Method (MPM).

🔍 Finally, we map the static and dynamic particles back to the Gaussian world and render with effect-specific appearance parameters and mesh-based collision handling.

📋 See our paper for more details on our hybrid Gaussian-particle framework and how we apply it to enable realistic animation of physically-based weather effects in static 3D scenes.

BibTeX


                @misc{fiebelman2025letsnowanimatingstatic, 

                     title={Let it Snow! Animating Static Gaussian Scenes With Dynamic Weather Effects}, 

                     author={Gal Fiebelman and Hadar Averbuch-Elor and Sagie Benaim}, 

                     year={2025}, 

                     eprint={2504.05296}, 

                     archivePrefix={arXiv}, 

                     primaryClass={cs.GR} 

                }