Ecological connectivity modelling for memory-constrained compute environments

Circuit theory has been successfully applied to ecological connectivity modelling, notably via the Circuitscape software, which is typically run locally on a laptop, or on a dedicated server. However, for some applications, such web apps, automated pipelines, or high-throughput parallel analyses, backend infrastructure is often required. These systems can be costly, complex to operate, and introduce privacy and data governance concerns. Recent developments in WebAssembly (Wasm) now also allow fast C++ or Rust code to be run directly in a sandboxed browser environment. However, these are also contexts in which memory is limited. In order to make connectivity analysis feasible under memory constraints, we developed a Rust/WebAssembly toolset with a geospatial data pipeline and efficient, memory-optimised implementation of connectivity analysis. This approach may be useful for geospatial modelling software where raster sizes and memory are constrained, such as in the browser context. We also demonstrate that lower memory presents opportunities for parallelism, such as in the context of algorithms that require sampling, enabling previously constrained analyses. Our approach has the potential to reduce backend operational complexity, resource requirements, and overall cost.