Fig. 1: The scales of human and natural groundwater systems mismatch with remote sensing and models.
From: High resolution US water table depth estimates reveal quantity of accessible groundwater
There is an inherent mismatch between the scales at which national policy is set (~8 M km2), the extent of major aquifers and irrigation projects (~50 K km2), and the scale at which local irrigation decisions are made (~1 km2). Existing models and remote sensing data products are unable to capture this seven orders of magnitude range of scales, and while much progress is being made, gaps remain at important scales for agriculture. The machine learning approach presented here can seamlessly bridge these scales.
