Sustainability

Food waste has global economic, environmental and social implications, and hence tackling food waste provides significant opportunities to achieve several Sustainable Development Goals (SDGs). The literature has already documented several promising technological opportunities to help food firms reduce waste in their supply chains. Real-time monitoring of food supply chains with Internet of Things (IoT) sensors is now well documented, particularly for storage and transport, yet the data these systems generate are mostly used for simple threshold alerts rather than learning. Evidence on artificial intelligence (AI) for food waste reduction is growing but remains concentrated in hospitality and retail and is largely conceptual. With binding EU food waste targets for 2030 now in law, food businesses need practical routes from monitoring data to AI-supported decisions. This paper explores the potential of artificial intelligence (AI) to address food waste challenges, using action research and drawing on extensive collaboration with industry partners across the agri-food supply chain in North-West Europe. Our data were collected over a period of 2 years using observations and interactions with food supply chains during the execution of a large-scale European-funded project. This paper describes several opportunities for using AI in detail. Our research illustrates the multi-disciplinary challenges and the promise of AI approaches in reducing food waste, thereby achieving several SDGs. We develop several research objectives to support future researchers.

Water circulation and exchange in shallow lakes can be restricted by complex bathymetry and human activities. However, the interaction between dynamic water replenishment strategies and complex bathymetry remains unclear. This study developed a two-dimensional hydrodynamic model of the highly fragmented Baiyangdian Lake. Using the spring water replenishment process in 2024 as the baseline, it simulated the effects of six control scenarios covering total replenishment volume, duration and the temporal distribution of inflow on the flow field and water exchange percentage. The results showed the following: (1) Increasing replenishment volume to 1.5 times the baseline raised the main-lake active water area fraction (flow velocity > 0.60 cm s⁻¹) from 31.37% to 40.71%. The main-lake water exchange percentage reached 49.34% on day 32. (2) The water exchange percentage exhibited strong spatial heterogeneity, and the relative performance of the tested scenarios depended on local topographic resistance. Short-duration or concentrated replenishment may promote water exchange in topographically constrained areas, whereas open-channel areas may benefit from longer-duration replenishment. (3) This study characterized the simulated responses of shallow lakes to water replenishment regulation under complex microtopography, providing hydrodynamic evidence for sustainable, spatially differentiated replenishment planning in fragmented wetlands.

Source separation is central to Sustainable Development Goal targets 11.6 and 12.5, yet behavioural interventions are often implemented without establishing what proportion of the waste stream users can realistically separate. This study presents a 2026 diagnostic of waste management at the Universidad de Monterrey (UDEM), Mexico. An exploratory, constraint-based classification framework indicated that 8.8% of campus waste was amenable to user-level separation at the point of disposal (Type A), 19.1% could be separated through staff protocols (Type B), and 72.1% was practically non-separable under the conditions assessed (Type C). Two field studies examined factors associated with Type A separation: bin configurations and the temporary reintroduction of classroom bins. In an observational comparison (N = 50), units with restricted apertures (circular, bottle-sized openings) showed median item-based compliance of 60–82%, compared with 18–19% for open-aperture units. However, the configurations also differed in signage, location, and user populations, precluding attribution of the observed differences to aperture geometry alone. Under normal operating conditions, classrooms have no bins. Temporarily reintroducing an unsorted bin in 20 classrooms revealed no statistically detectable difference in visible litter items per classroom-day (p = 0.513). Recyclables deposited in the reintroduced bins entered the residual waste stream, although the fate of comparable items under bin-free conditions was not tracked. Practical implications for campus waste management are discussed.