River fish kills: preserve the evidence

A river fish kill often triggers an immediate search for a single culprit: pollution, oxygen depletion, disease or a harmful algal bloom. That early conclusion can undermine the investigation. Water conditions may change within hours, carcasses deteriorate, and detecting a contaminant does not establish causation. The professional priority is to notify, secure, document and sample through a coordinated response. A protocol prepared before an incident offers the best chance of distinguishing environmental, toxic, infectious and multifactorial causes.

Why the first hours matter

The Australian Government’s National Investigation and Reporting Protocol for Fish Kills stresses that evidence deteriorates quickly and calls for site inspection as soon as possible. Visible deaths may be the endpoint of a short-lived event: overnight oxygen depletion, an intermittent discharge, a storm after prolonged dry weather, a sudden flow change or disturbed sediment.

Response structures differ by jurisdiction, so local reporting routes must be confirmed in advance. A French departmental example from Savoie directs observers of dead fish to contact the French Biodiversity Agency without delay and describes information-sharing among water, health, environment, emergency and local government bodies. The transferable lesson is not the local telephone number. It is that a field team should not conduct alone an incident that may have animal-health, environmental and enforcement implications.

The initial report should capture the discovery time, exact location, apparent extent, affected species or groups, rough numbers, condition of the animals and unusual observations. Dated overview and close-up photographs, with coordinates where possible, preserve a situation that may change before response officers arrive.

Make the site safe and describe it before handling

Risk assessment comes before sampling. Suspected chemicals, unstable banks, traffic, depth, current, heat, structures and potentially hazardous atmospheres can all threaten responders. Appropriate protective equipment and access rules follow from that assessment. No one should enter the water or handle animals and unknown containers without suitable training and controls.

Site notes should separate observation from interpretation. Investigators can map the mortality front, carcass accumulations, live or moribund fish, apparently spared taxa and size classes, tributaries, outfalls, drains, barriers, stagnant areas, foam, odours, surface films, discolouration and blooms. Time, recent weather, apparent flow and witness descriptions of pre-mortality behaviour complete the record.

A kill restricted to one reach or a subset of species differs from widespread mortality involving fish, invertebrates and amphibians. That pattern guides sampling but does not prove a cause. Fish crowding near the surface may indicate respiratory distress, yet it cannot identify what reduced oxygen or whether a gill-active toxicant was involved.

Measure water at the right place and time

Field measurements should be taken promptly with maintained, calibrated instruments and recorded with time, depth and coordinates. Temperature, dissolved oxygen and percent saturation, pH, conductivity and, where relevant, turbidity, oxidation-reduction potential or ammonium provide an initial profile. A single daytime reading cannot exclude an overnight minimum or an event that has already passed.

The Australian protocol recommends comparing the kill site with upstream and downstream locations, then adding stations around any suspected source. This spatial design is more informative than a bottle filled only where current or wind accumulated carcasses. Reference stations should be sampled as close together in time as possible using consistent equipment and procedures.

Laboratory water analyses must be chosen with the observed scenario in mind. Nutrients, oxygen demand, metals, pesticides, phytoplankton, toxins and other contaminants require different containers, volumes, preservatives and holding times. A generic bottle of water is rarely suitable for every test. Field blanks, replicates and transport conditions should be agreed before deployment when the stakes warrant them.

Collect fish with laboratory guidance

Freshly dead or moribund fish usually provide more diagnostic value than decomposed carcasses. The official protocol favours whole fish when practical and distinguishes storage by test: chilled fresh material for several examinations, fixative for histopathology, dedicated media for some molecular tests and targeted freezing for toxicology. These are not universal instructions. Species, body size, delay, transport and laboratory capability all affect the plan.

Animal-health and environmental laboratories should therefore be contacted before collection. They can specify numbers, matrices, containers, temperatures, fixatives, priority tests and shipping conditions. Specimens intended for microbiology, toxicology and histology are not interchangeable and should not be pooled or preserved in the same way.

Every item needs a unique identifier linked to location, time, collector, species, condition, matrix, preservation method and requested analysis. A signed chain-of-custody record is essential when pollution or an offence is suspected. It also protects scientific interpretation by preventing confusion between upstream reference, kill zone and downstream samples.

Combine evidence before assigning a cause

Two Mediterranean river investigations illustrate why restraint is necessary. In the Bogdanas River, Petriki and colleagues found no organic poisons, pesticides or abnormal heavy-metal concentrations in the fish tissues examined. High biochemical and chemical oxygen demand, nitrogen, conductivity, warm conditions and low flow instead supported an upstream organic-load hypothesis. Their interpretation depended on converging evidence, not one result.

Following flash storms on the River Tiber, Carere and colleagues combined targeted chemical analysis, effect-based methods and fish-assemblage examinations. Several contaminants occurred at low concentrations while bioassays detected some effects. The authors described a multiple-stressor scenario involving hydrology, urban pressures and chemical mixtures rather than assigning the entire kill to one compound.

A well-run investigation may therefore remain inconclusive. A negative result does not prove that every toxicant or pathogen was absent; it only reports what was sought in the tested matrices, places and time window. Conversely, finding a microorganism or contaminant does not show that it caused death. Temporal, spatial, clinical, pathological and environmental evidence must be interpreted together.

Build a kit and a decision chain before the incident

Preparedness is easiest before a crisis. Managers can maintain a contact sheet, notification form, access and outfall map, verified meters, laboratory-supplied containers, camera equipment, durable labels, safety information and custody forms. Expiry dates, batteries and calibration status need scheduled checks.

A short exercise can allocate responsibility for coordination, safety, water measurements, animal survey, sampling, laboratory liaison, records and public communication. Clear roles reduce omissions and prevent several responders from collecting at the same location while leaving no comparative stations.

The operational sequence is straightforward: notify promptly, freeze the scene through traceable observations, measure before conditions change, call laboratories before sampling, compare upstream–kill zone–downstream locations, and preserve custody records. Vetofish can help river managers, local authorities and fisheries organisations prepare response plans, coordinate animal and environmental sampling, interpret combined results and review incidents—without promising that every fish kill will yield one definitive cause.