?
Challenges in Building and Visualizing Isolines on Tile-Based Data Quickly and Accurately
Modern web-based geospatial applications increasingly demand client-side processing of large scientific datasets. Web Tensor DBMSs (WT-DBMS) such as WebArrayDB enable browsers to receive and process raw tensor data through tile-based delivery mechanisms (such as WMTS with GeoTIFF tiles), shifting computational load from servers to clients. However, generating isolines from tile-based data presents fundamental challenges in maintaining topological consistency and geometric accuracy. This paper explores these challenges in the context of ERA5 climate reanalysis data, examining how traditional isoline construction algorithms must be adapted for incremental tile-by-tile processing. We describe the original CONREC-based algorithm that operates on complete grids, analyze why it cannot be directly applied to streaming tile data, and investigate four distinct approaches to tile-based construction: LineString-based, Polygon-based, Strip-based, and Buffer-based methods. Through systematic experimentation on ERA5 mean sea level pressure data (181×360 grid processed as 64×64 tiles), we demonstrate that each approach exhibits distinct trade-offs between boundary continuity, geometric accuracy, and visual artifact generation. Our analysis reveals that achieving both topological consistency and geometric fidelity in tile-based isoline generation remains an open challenge requiring further research.