Summer Research Project Grant Report
NOTE: Completed as a joint UWA DI/ICRAR summer project, this report meets dual requirements and is therefore longer than others.
Cross-disciplinary Domains: Data science, Environmental sciences and Software engineering
Research Project Objective
The primary objective of this project is to create a multi-source visualisation platform that brings together different data sources into a single, user-friendly analytical environment. By combining ground-truth DPIRD observational station data with global predictive ECMWF operational forecasts, the tool acts as a diagnostic bridge for understanding Western Australian meteorology.
The specific contributions and functional goals of this project are outlined below:
Integrated Data Synthesis
The central deliverable is a visualisation architecture that performs real-time overlaying of DPIRD’s local station network against ECMWF forecast grids. This integration allows for a direct, side-by-side comparison between theoretical atmospheric models and the localised “ground truth,” facilitating a level of scrutiny that is impossible through numerical spreadsheets alone.
Enhanced Data Analysis Capabilities
By transforming raw data into easy-to-understand visuals, the tool empowers users to achieve rapid identification of systemic and isolated errors, specifically:
- Station integrity & Quality control: The system enables the immediate detection of spurious data points or failing hardware. By visualising outliers against neighbouring station trends, users can distinguish between genuine localised weather phenomena and mechanical sensor drift.
- Spatial and Temporal divergence: Users can pinpoint the exact geographical coordinates and timeframes where global model predictions begin to decouple from observed conditions. This “forensic” view of model breakdown is essential for understanding the limitations of the ECMWF within the unique Western Australian climate.
