Real-Time Snow Cover Simulation (Computers & Graphics)

Title: Real-time GIS-based Snow Cover Approximation and Rendering for Large Terrains

Journal: Computers & Graphics

Authors: Benjamin Neukom, Stefan Müller Arisona, Simon Schubiger

Abstract: Various terrain visualization techniques based on geographic information system (GIS) data already exist. One major drawback of existing visualizations is that they do not capture seasonal variations well. Besides vegetation variations, in colder areas this particularly also applies to snow cover. In this paper, we propose a real-time multi-scale snow cover approximation and visualization for large terrains. The computation runs on a large grid, calculates the snow/water equivalent based on precipitation data from a GIS and snowmelt based on a physically-based solar radiation calculation combined with a degree-day snowmelt approach using level of detail (LOD). The snow visualization is divided into two parts: Zero thickness snow cover textures are generated for distant views. For close up views the terrain’s height field is modified using displacement maps and tessellation to produce thick snow covers. The GPU-based data-parallel computation and the visualization run on the GPU in real-time on a modern desktop computer. The implementation is tested using a real area in the Swiss Alps, with a size of 14.16 by 12.88 km, a grid resolution of 222 × 206, and a time step of 1 h. We compare the rendered results spanning several months with a time series of photographs from webcams for visual accuracy.

Comparison of webcam images (left) of Zermatt with our simulation (right). Trees were placed by hand and the sun position was approximated from the webcam image. (A) Start of the computation on 12.10.2015 with only the initial snow visible. (B) The results on 18.10.2015 after the first snowfall. The snowline was accurately computed. (C) The results on 16.01.2016 after the first heavy snowfall with the whole scene covered in snow. (D) The results on 21.04.2016 with snow beginning to melt as the temperature rises.

Code for Unreal Engine (open source): https://github.com/bneukom/UnrealSnow

Title: Visualizing Waypoints-Constrained Origin-Destination Patterns for Massive Transportation Data
Authors: Benjamin Neukom, Stefan Müller Arisona, Simon Schubiger
Journal: Computers & Graphics
Publisher: Elsevier
Year: 2018
Link: https://doi.org/10.1016/j.cag.2017.10.003