Imagine opening every door in a building while leaving its rooms flooded, overheated, or empty of food. Access would improve, but the building would not necessarily become a good place to live. Something similar happens when river restoration is described entirely through the distance a fish can travel.
Passage matters profoundly. Blocking it can separate feeding grounds from spawning sites and fragment populations. Yet a barrier also changes the habitat on either side: water velocity, sediment movement, depth, and the relation between a river and its surroundings. An open route and a functioning river are related achievements, not identical measurements.
The distinction becomes clearer when small barriers, impounded habitat, and observed recolonization are considered together.
The small structures missing from the map
Belletti and colleagues assembled European barrier records and checked inventories against walkover surveys. Their 2020 investigation of river fragmentation estimated at least 1.2 million instream barriers across thirty-six countries. Most structures were less than two metres high, and existing records omitted many of the barriers discovered on surveyed reaches.
This was an estimate combining harmonized records, field checks, and modeling. It was not an individually verified list of every structure in Europe. Its central insight survives that distinction: the most visually impressive dams are a poor proxy for the full pattern of river fragmentation.
A small structure can be consequential without being an absolute barrier to every species at every flow. Passage depends on the animal, its life stage, hydraulic conditions, and timing. A barrier inventory identifies where questions need to be asked; it does not by itself measure every organism's experience of the network.
The missing structures also create a planning problem. A river reach can look well connected in a large-dam database while being interrupted by culverts, weirs, or bed-control structures. A restoration assessment built on that database might underestimate the work remaining or misidentify the benefit of removing one obstacle.
A route can remain a different habitat
A later European study asked about upstream habitat alteration caused by impoundments. Its Conceptual Habitat Alteration Model for Ponding combined river settings, barrier information, and fish habitat-use guilds. After correcting for incomplete barrier records and missing height data, the authors estimated that more than two hundred thousand kilometres of river habitat had been altered by impounding.
That number is model-based. It depends on extrapolation and on assigning species to groups with shared habitat preferences. It is not a direct census of fish lost from each kilometre. The authors explicitly showed that barrier type and geographical setting affect predicted impacts.
Its useful contribution is to widen the target. A fish pass might allow movement past a structure while leaving a flowing reach converted to pond-like conditions. Species needing swift, oxygenated habitats and particular substrates may experience losses that a passage metric alone does not capture.
This also explains why counting removed barriers can become a weak performance measure. Ten small removals and one larger removal may reconnect different amounts and types of habitat. Their ecological benefits depend on location in the network, neighboring barriers, existing populations, and conditions beyond the removed structure. A tally is an activity record; the outcome requires an ecological comparison.
The Elwha demonstrates access, with qualifications
On Washington's Elwha River, removal of two large dams created a rare opportunity to observe reconnection at substantial scale. Duda and colleagues combined passage observations with continuous snorkeling surveys before and after removal. Their study of spatial fish responses documented migratory fishes entering previously inaccessible areas and major upstream extensions in the observed distribution of adult Chinook salmon and summer steelhead.
The surveys covered sixty-five kilometres of mainstem river, with comparable summer surveys in two years before and two years after removal. That design is particularly valuable because it examines where fish occurred rather than assuming that theoretical access equals use.
The findings do not mean that every population had completed recovery. Passage, spawning, juvenile survival, local adaptation, and a self-sustaining population are different stages. The study's focus and observation periods support particular statements about distribution and density, not a final verdict on every ecological goal.
The Elwha also has a distinctive context, including a largely protected upper watershed and extensive collaborative monitoring. Its responses cannot simply be assigned to a heavily polluted or otherwise altered river elsewhere. It demonstrates an ecological possibility and a way to measure it, rather than a guaranteed timetable.
Sediment belongs in the recovery story
Reconnection can move material as well as animals. Research summarized by the US Geological Survey examined the Elwha's nearshore community during and after sediment released by dam removal. The coastal-ecosystem investigation described substantial changes in seabed conditions and communities, including creation of sandy habitats used by some species.
This is a reason to follow the river past its mouth. A barrier's influence does not stop at the concrete, and neither do removal effects. Sediment can disturb one assemblage while creating conditions for another. Calling all change either damage or recovery without naming the habitat and timescale hides the actual process.
For evaluating a project, I would therefore keep several questions separate. Can organisms pass? Is the habitat they reach suitable? Are sediment and flow processes functioning in the intended way? Do populations persist through subsequent generations? This is an organizing framework drawn from the studies, not a universal checklist that dictates the same intervention everywhere.
The framework also leaves room for negative findings. If passage improves while recruitment remains poor, that result helps locate the next constraint. If flowing habitat returns but a source population is absent, physical restoration and biological recolonization may need different timescales.
The most persuasive restoration claim is not that a river has been given an open door. It is that the doors, rooms, and connecting processes have been investigated together. Fish moving upstream are an extraordinary sign of renewed opportunity. The longer task is to find out what kind of life that opportunity can sustain.
A thought worth keeping?
