Figure 1. The global distribution of river monitoring stations with sensors suitable for developing metabolism estimates and carbon emissions has a strong spatial bias . Examples of the most dense nationwide networks are: (a) StreamPULSE sites in the continental USA; (b) Spanish Environmental Department Water Quality Automatic Information System (SAICA); (c) a catchment-scale monitoring network in the Connecticut River, NE USA (Hosen et al., 2021)
In this review, we evaluate the potential for using sensor data and machine-learning to advance river carbon cycle processes and emission management, both responsively to stressor events and proactively to enhance the management of both water resource security and downstream river system services. Taking into consideration the key drivers of river carbon processes and emissions, we demonstrate how recent technological advances in the development and implementation of sensor networks for river catchment management can be harnessed to improve knowledge of aquatic processes. We examine how sensor and analytics advances offer new opportunities to develop strategic monitoring networks that can capture impacts resulting from a range of catchment processes and human modifications. We illustrate the benefits of incorporating emerging, affordable sensor technologies, and novel robotic sensor deployment technologies, which allow for high-resolution monitoring, and explain how a variety of water quality parameters can be used to develop causal relationships between drivers and response variables. We then assess the most promising analytical approaches and methods for processing, modelling, and visualising high-resolution river system data, demonstrating how novel applications of sensor networks coupled with artificial intelligence (AI) solutions could be developed.