Proactive management responses to sensor data
Proactive uses of sensor networks and analytics portals will benefit from long-term management planning through research and adaptive management. For example, experimental campaigns can be initiated to optimize management by modifying environmental flows from reservoirs (Figure 2b). At present, reservoir operators release water to support downstream ecosystems, aiming to maintain the quantity and quality of water, based on the taxonomic or behavioural response of targeted biological groups, such as fish and invertebrates (Gillespie et al., 2015). However, these flows additionally modify downstream water quality, such as temperature, which is a strong control on carbon cycle processes (Yvon-Durocher et al., 2011). The water release also alters emissions from previously dry sediments (Perez-Calpe et al., 2022), and transfers dissolved GHG from in-reservoir processing (Shi et al., 2023) to modify downstream emissions (Guerin et al., 2006). With the ability of sensor networks to provide rapid insights into downstream river ecosystem responses to changes in outflow volume, reservoir managers could more effectively balance water supply requirements with minimising downstream ecosystem damage and emissions.
GHGs are known to be emitted from all freshwaters, but the lack of direct accounting for many of these systems, despite studies showing their important role in GHG budgets both naturally and when modified, can now be remedied with enhanced environmental data collection. Without such investments, commitments to develop net-zero adaptation policies by governments and businesses, such as water and hydropower companies, are likely to be hampered if freshwater systems remain understudied. Following recent IPCC inventory refinements (IPCC, 2019), emissions from managed inland waters (e.g. farm ponds, reservoirs, and their outflows) now need to be quantified, which is adding some impetus to data collection, but for river systems it is also necessary to consider business operations effects ‘offsite’ or ‘downstream’, such as lengths of watercourses influenced by upstream contaminant inputs (Hu et al., 2018) or flow modifications (Shi et al., 2023). The integration of sensor networks with ML models will be a key step toward meeting this need for whole catchment understanding and improved management.