Data-Driven Compute Overlays for Interactive Geographic Simulation and Visualization

TL;DR

This paper tackles the challenge of delivering real-time, interactive overlays for 3D geographic maps with large data volumes. It introduces data-driven compute overlays implemented as a multi-pass GPU compute workflow using WebGPU, enabling dynamic overlays directly on the client. The authors present weBIGeo, an API and open-source tool, and demonstrate two use cases—snow cover and avalanche simulations—with performance benchmarks showing orders-of-magnitude speedups over CPU references (FlowPy). The results indicate that GPU-accelerated, client-side overlays can enable real-time parameter exploration and visualization on commodity hardware, with potential extensions to broader geospatial analyses.

Abstract

We present interactive data-driven compute overlays for native and web-based 3D geographic map applications based on WebGPU. Our data-driven overlays are generated in a multi-step compute workflow from multiple data sources on the GPU. We demonstrate their potential by showing results from snow cover and avalanche simulations, where simulation parameters can be adjusted interactively and results are visualized instantly. Benchmarks show that our approach can compute large-scale avalanche simulations in milliseconds to seconds, depending on the size of the terrain and the simulation parameters, which is multiple orders of magnitude faster than a state-of-the-art Python implementation.

Figures (3)

• Figure 1: A node graph generated from the workflow specification for the avalanche use case (\ref{['sec:avalanches']}) This workflow generates a multi-resolution Avalanche Overlay Texture (right) from terrain model tiles in a given Region AABB (left).
• Figure 2: Simple snow cover simulation using a fragment shader on top of the base map (top) or using a data-driven compute overlay (bottom). When zooming out (right), the fragment shader approach shows too much snow on top of forests and other locally steep terrain (top right). Zoom in for more details.
• Figure 3: weBIGeo avalanche simulation results on the parabola (left) and vals (right) datasets. Release points are highlighted in blue. Contour lines represent the terrain, spaced at 50m intervals. Color encodes the maximum vertical displacement per step $Z^{\delta}_{max}$, which can be interpreted as an indicator for velocity. Zoom in for more details.