3D Weather — Real-Time Simulation of Localised Weather

3D Weather was an early real-time simulation project exploring how weather could be represented from the observer’s point of view, rather than through abstract symbols, charts, or satellite imagery alone.

The project investigated how live and recorded weather data, terrain information, and real-time 3D graphics techniques could be combined to create believable, spatially grounded simulations of local weather conditions. The emphasis was not on spectacle, but on understanding, rehearsal, and trust. It was about how people perceive weather as they might experience it, rather than how it is typically abstracted for broadcast.

Unlike traditional top-down weather visualisations, 3D Weather focused on first-person and near-ground perspectives, allowing users to view weather as it unfolds within a recognisable environment. The system explored how volumetric clouds, wind, fog, and atmospheric perspective could be simulated in real time and linked to real-world data sources.

What the project involved

- Real-time 3D simulation of localised weather conditions

- Volumetric cloud rendering using real-time graphics techniques

- Integration of satellite imagery and infrared temperature data

- Terrain derived from LIDAR data to provide geographic specificity

- Wind vectors and atmospheric parameters driving cloud motion

- Exploration of probabilistic and uncertain outcomes, rather than deterministic prediction

A core theme of the work was the relationship between visual realism and trust. Highly realistic simulations can imply certainty where none exists, particularly in forecasting contexts. This project examined how simulation can remain believable and informative while still communicating uncertainty which is a challenge shared by many training, rehearsal, and decision-support systems.

Why this work remains relevant

Although developed in an academic research context, 3D Weather directly reflects concerns common to applied simulation and training systems:

- representing complex, dynamic systems in real time

- grounding simulation in real-world data and geography

- supporting understanding rather than prediction

- balancing fidelity with performance and reliability

The thinking and technical approaches explored in this project later informed larger-scale real-time simulation systems, including rehearsal and validation tools used for physical deployment and live operations.

(Originally developed at Bath Spa University; project documentation and prototype images shown below.)