🤖 AI Summary
Sparse spatiotemporal observations—e.g., groundwater storage in Bangladesh—hinder accurate continuous field reconstruction. Method: We propose a hybrid spatiotemporal modeling framework integrating deep learning with geostatistical interpolation. Departing from conventional “aggregate-then-model” or “model-then-interpolate” paradigms, our approach innovatively employs time-series dynamic clustering to guide spatial interpolation, explicitly encoding geological uncertainty’s influence on temporal evolution. It combines grid-to-grid/point deep learning prediction with kriging interpolation and identifies dynamically similar yet non-neighboring observation regions via adaptive similarity metrics. Contribution/Results: Experiments demonstrate significantly higher temporal forecasting accuracy than pure spatial interpolation; proximity does not imply dynamic similarity; and the method exhibits superior interpolation robustness and physical interpretability—particularly for latent driving variables such as groundwater storage.
📝 Abstract
Geospatial observational datasets are often limited to point measurements, making temporal prediction and spatial interpolation essential for constructing continuous fields. This study evaluates two deep learning strategies for addressing this challenge: (1) a grid-to-grid approach, where gridded predictors are used to model rasterised targets (aggregation before modelling), and (2) a grid-to-point approach, where gridded predictors model point targets, followed by kriging interpolation to fill the domain (aggregation after modelling). Using groundwater storage data from Bangladesh as a case study, we compare the effcacy of these approaches. Our findings indicate that spatial interpolation is substantially more difficult than temporal prediction. In particular, nearest neighbours are not always the most similar, and uncertainties in geology strongly influence point temporal behaviour. These insights motivate future work on advanced interpolation methods informed by clustering locations based on time series dynamics. Demonstrated on groundwater storage, the conclusions are applicable to other environmental variables governed by indirectly observable factors. Code is available at https://github.com/pazolka/interpolation-prediction-gwsa.
