Spatiotemporal optimization of groundwater monitoring networks using data-driven sparse sensing methods

Abstract. Groundwater monitoring and specific collection of data on the spatiotemporal dynamics aquifer are prerequisites for effective groundwater management determine nearly all downstream decisions. An optimally designed network (GMN) will provide maximum information content at minimum cost (Pareto optimum). In this study, PySensors, a Python package containing scalable, data-driven algorithms sparse sensor selection signal reconstruction with dimensionality reduction is applied to an existing GMN in 1D (hydrographs) 2D (gridded contour maps). The algorithm first fits basis object training then applies computationally efficient QR that ranks wells (for 1D) or suitable sites additional 2D) order importance, based state tailored basis. This procedure enables be reduced extended along Pareto front. Moreover, we investigate effect choice performance by comparing three types typically used (i.e., identity, random projection, SVD, respectively, PCA). We define gridded function extension case penalizes unsuitable locations. Our results show proposed approach performs better than best randomly selected wells. optimized makes it possible adequately reconstruct removed hydrographs highly subset low loss. With 94 %, average absolute accuracy 0.1 m achieved, addition 0.05 69 % 0.01 18 %.