River light pollution can reset fish activity

A streetlight does more than create a glow above a river. Its output may reach the surface and shallow margins where fish feed, rest and move. A 2026 field study found that mosquitofish exposed to artificial light at night for years were much more active after dark and showed altered daytime behaviour. The result establishes neither universal harm nor a harmless adaptation. It shows that darkness is an ecological variable to assess before lighting beside a riverbank, bridge, car park or fish passage is installed or renewed.

Ten populations across five urban drainages

Jenkins and Langerhans studied the eastern mosquitofish, Gambusia holbrooki, a small diurnal freshwater fish. Near Raleigh, North Carolina, they selected five pairs of sites in five separate drainages. Each pair included one artificially lit site and one dark site. The permanent, pond-like waters were connected to streams and surrounded by similarly developed urban landscapes.

Maximum illuminance near the water ranged from 0.34 to 30.51 lux at lit sites, compared with less than 0.10 lux at dark sites during observations. Lit populations had experienced artificial night lighting for three to 25 years. The team characterised light, water variables and population structure, then surveyed 400 shoreline quadrats during the day and about two hours after sunset. Locomotion and daytime behaviour were also quantified, with 430 fish observed for the latter assessment.

This comparative design is valuable because patterns were tested repeatedly across drainages. It is not a randomised lighting experiment, however, and unmeasured site differences may remain. The focal fish is also a resilient, widely introduced species that persists in urban habitats. Its response cannot be transferred directly to salmonids, eels or threatened fishes.

Lit nights became an activity period

At four of the five dark sites, researchers observed no active fish at night during quadrat sampling. Across all ten sites, night-time activity increased strongly with maximum illuminance. At the brightest and longest-lit site, fish no longer showed the usual night-time reduction in swimming speed: a clearly diurnal activity pattern appeared to have been lost.

Changes extended beyond locomotion. Females at lit sites used more offshore habitat at night and often showed lower daytime foraging and feeding rates. At sites with stronger lighting, males showed more daytime aggression and fewer sexual behaviours. The authors discuss several non-exclusive explanations: feeding shifted into the night, altered circadian timing, behavioural plasticity and evolutionary change over generations.

The study cannot separate these mechanisms. Lighting duration and intensity partly covaried, making exposure history difficult to distinguish from dose. Body size and condition did not change consistently, and no general demographic decline was demonstrated. It would therefore be inaccurate to claim either that “light kills fish” or that fish “adapt without consequence”. The evidence supports behavioural reorganisation in this species under these field conditions.

Direction of response depends on species and context

A European experiment adds a useful warning. Tarena and colleagues tested passage over a hydraulic-flume obstacle by European gudgeon, Gobio gobio, and Italian riffle dace, Telestes muticellus, under darkness, daylight and artificial light at night. Sixty fish of each species underwent repeated trials, allowing individual experience to be considered.

Artificial night lighting reduced passage success in gudgeon and, at higher intensity, increased the time needed to cross. Italian riffle dace moved in the opposite direction: under intense artificial light, they passed faster than under darkness. The authors interpret these results cautiously. Light can attract, repel or freeze a fish, or prompt it to leave an exposed area rapidly. Species, normal diel activity, intensity, hydraulics and perceived predation risk all matter.

Managers should not turn one colour temperature or experimental lux value into a universal ecological threshold. The operational questions are broader: where does light reach the water, when is it present, how deep does it penetrate, which life-history period overlaps with it and which ecological function occurs at that location?

Audit light as a connectivity pressure

Assessment should begin after sunset. Map every relevant luminaire and its actual purpose, then measure illuminance at the water surface and along riparian habitat. Satellite radiance and photographs can describe a glow, but not necessarily the light received by a fish below a bank or inside a fishway.

Measurements should be connected to biological timing: migration, spawning, insect emergence, foraging or temporary works. An annual average may hide a short exposure during the most sensitive period. Bridges, hydraulic structures, paths, car parks, sports facilities and construction sites beside aquatic corridors are common locations where a lighting audit can reveal conflicts.

Before–after monitoring needs a short, repeatable set of variables: intensity and spectrum, operating times, illuminated bank length, fish activity or passage, temperature and discharge. Behavioural categories should be defined explicitly for the target species and observation method. Failing to see fish during one survey is not evidence that lighting has no effect; detectability and natural activity must be incorporated into the design.

Remove unnecessary light from the water first

The first mitigation principle is broadly transferable: do not send light where it is not needed. In France, the national order of 27 December 2018 generally prohibits direct lighting of rivers, lakes, ponds and other water bodies, while defining specific exceptions for some safety and operational functions. This legal framework is nationally specific, but the design principle applies internationally: secure the required area without illuminating the aquatic habitat.

Where lighting remains necessary, mitigation can combine shielding and orientation, lower output, curfews or dimming, presence detection and shorter operating periods. Warmer correlated colour temperature can reduce some spectral components, but it does not make unnecessary illumination harmless. Switching lights off, shortening exposure and preventing spill onto water are more dependable interventions than changing a lamp while retaining the same lit footprint.

The French Biodiversity Agency’s “Dark Infrastructure” guidance recommends identifying conflict points, mapping pressure, selecting model species or functions and evaluating the intervention. For aquatic environments, this connects blue corridors with nocturnal continuity. A physically open route can remain behaviourally difficult to cross when a beam divides the dark habitat.

Avoid three common interpretation errors

First, activity is not automatically a welfare gain. A fish feeding at night may exploit a new opportunity, but altered rest, social behaviour or predation exposure may carry costs that the field study did not quantify. Second, a reduction in observed passage does not by itself identify the mechanism; attraction, avoidance and delayed approach can produce different management needs. Third, a response in a tolerant invasive poeciliid cannot be assumed for native species with different sensory systems and life histories.

Mitigation trials should therefore state the intended outcome in advance. A project may aim to restore an unlit bank, improve passage during a migration window or reduce direct irradiance over a spawning habitat. Matching the indicator to that objective makes it possible to distinguish a successful lighting change from a simple energy-saving measure.

Conclusion: make darkness measurable and manageable

The 2026 study does not provide a universal protective threshold. It shows that a resilient fish can reorganise its daily activity after several years of exposure without allowing that response to be labelled automatically as benefit or damage. Together with the flume experiment, it confirms that artificial night lighting is a site- and species-specific behavioural pressure.

Managers can act now by checking whether light reaches water directly, inventorying luminaires around aquatic corridors, prioritising feasible curfews and adding night-time measurements to ecological monitoring. Vetofish can support fish-focused risk assessment, behavioural indicator design, timing of field surveys and before–after evaluation of mitigation plans. Protecting darkness does not require claiming that every fish responds in the same way; it requires measuring exposure, reducing avoidable pressure and verifying the result.