Artificial light at night (ALAN) is a widespread alteration of the natural environment that can affect the functioning of ecosystems. ALAN can change the movement patterns of freshwater animals that move into the adjacent riparian and terrestrial ecosystems, but the implications for local riparian consumers that rely on these subsidies are still unexplored. We conducted a 2-year field experiment to quantify changes of freshwater-terrestrial linkages by installing streetlights in a previously light-naïve riparian area adjacent to an agricultural drainage ditch. We compared the abundance and community composition of emerging aquatic insects, flying insects, and ground-dwelling arthropods with an unlit control site. Comparisons were made within and between years using two-way generalized least squares (GLS) model and a BACI design (BeforeAfter Control-Impact). Aquatic insect emergence, the proportion of flying insects that were aquatic in origin, and the total abundance of flying insects all increased in the ALAN-illuminated area. The abundance of several night-active ground-dwelling predators (Pachygnatha clercki, Trochosa sp., Opiliones) increased under ALAN and their activity was extended into the day. Conversely, the abundance of nocturnal ground beetles (Carabidae) decreased under ALAN. The changes in composition of riparian predator and scavenger communities suggest that the increase in aquatic-to-terrestrial subsidy flux may cascade through the riparian food web. The work is among the first studies to experimentally manipulate ALAN using a large-scale field experiment, and provides evidence that ALAN can affect processes that link adjacent ecosystems. Given the large number of streetlights that are installed along shorelines of freshwater bodies throughout the globe, the effects could be widespread and represent an underestimated source of impairment for both aquatic and riparian systems.

Drastic declines in insect populations, 'Ecological Armageddon', have recently gained increased attention in the scientific community, and are commonly considered to be the consequence of large-scale factors such as land-use changes, use of pesticides, climate change and habitat fragmentation. Artificial light at night (ALAN), a pervasive global change that strongly impacts insects, remains, however, infrequently recognised as a potential contributor to the observed declines. Here, we provide a summary of recent evidence of impacts of ALAN on insects and discuss how these impacts can drive declines in insect populations in light-polluted areas. ALAN can increase overall environmental pressure on insect populations, and this is particularly important in agroecosystems where insect communities provide important ecosystem services (such as natural pest control, pollination, conservation of soil structure and fertility and nutrient cycling), and are already under considerable environmental pressure. We discuss how changes in insect populations driven by ALAN and ALAN itself may hinder these services to influence crop production and biodiversity in agricultural landscapes. Understanding the contribution of ALAN and other factors to the decline of insects is an important step towards mitigation and the recovery of the insect fauna in our landscapes. In future studies, the role of increased nocturnal illumination also needs to be examined as a possible causal factor of insect declines in the ongoing 'Ecological Armageddon', along with the more commonly examined factors. Given the large scale of agricultural land use and the potential of ALAN to indirectly and directly impact crop production and biodiversity, a better understanding of effects of ALAN in agroecosystems is urgently needed. (Hallmann et al., 2017), popularly named 'Ecological Armageddon'. The authors used Malaise traps to collect data from 63 sites located within protected areas that are embedded in an agricultural landscape in two regions in Germany and found that biomass of flying insects decreased more than 75% over 27 years. This trend could not be explained by the changes in habitat, climate or land use (Hallmann et al., 2017), suggesting that another large-scale factor must therefore be involved. We suggest