Skip to content

Integration of Spatial Dependencies

Home / Open Questions / Integration of Spatial Dependencies

Background: Forecasting models often evaluate single-site performance, ignoring spatially distributed atmospheric interactions such as wake effects in wind energy systems.

Question / Future Work: Advancements in time-series forecasting require the integration of spatial information, specifically inter-turbine or inter-node dependencies, into architectures beyond single-site models, such as using graph-based representations.

Why It Matters: Spatio-temporal modeling is essential for improving accuracy when observations are highly coupled due to environmental factors.

Evidence: the evaluation is limited to single-site (turbine or farm-level) forecasting without modeling spatial dependencies such as inter-turbine wake effects; integrating spatial information through graph-based representations is a natural extension.

Metadata & Links