Defining a successful river restoration project

In this post Associate Editor Tadeu Siqueira discusses a paper he recently handled by Daniel Hering and colleagues ‘Contrasting the roles of section length and instream habitat enhancement for river restoration success: a field study of 20 European restoration projects

Streams and rivers are among the most degraded ecosystems in the world. In comparison to oceans, these ecosystems contain a tiny portion of the total amount of water in the planet, and yet they provide a habitat for a disproportional number of species, e.g. one-third of all vertebrate species, and are responsible for important ecosystem services, e.g. water provision. In response to this alarming situation, more developed countries have made significant efforts to promote river restoration by creating specific committees and agencies and financing the implementation of restoration projects – a business that moves millions of dollars around the world.

It would be reasonable, thus, to expect that the practice of river restoration usually reaches its objectives. Unfortunately it is not that simple. Rivers are part of a complex hierarchical dendritic network with strong aquatic–terrestrial linkages. That is: 1) headwaters flow into larger streams that flow into larger rivers and so on; these elements are continuously connected via water flow; 2) microhabitats (e.g. moss on boulders, leaf detritus) are part of riffles and pools, which are part of a river reach, which are part of a larger river section. In addition, all these elements are embedded in a terrestrial landscape that strongly influences what happens in the water body – “in every respect, the valley rules the stream” (Hynes 1975). Thus, human actions in the surrounding landscape alter physical, chemical and biological characteristics of rivers differently at distinctive spatial scales. For example, if one is using aquatic insects to measure the outcome of a restoration project, the heterogeneity of instream substrates may be more important than the area of the landscape occupied by agriculture. On the other hand, if one is using fish, then the minimum area available for spawning and rearing may be of greater importance.

Fall river dam removal. Photo credit: Massachusetts Division of Ecological Restoration (;
Fall river dam removal. Photo credit: Massachusetts Division of Ecological Restoration (;

Given the above, a successful river restoration project may depend on the scale at which restoration actions are made, the size of the restored area, the response variable being analysed (e.g. fish, insects, abiotic variables) and/or the geomorphological features of the river basin, among others. Thus, rigorous scientific tests are essential for river restoration, as the solution to environmental problems must be grounded in science (see Palmer 2009). However, rigorous evaluation of the efficiency of environmental programs seems to be less common than in other fields, such as public health and poverty reduction (Ferraro 2009). Some years ago, a global ecosystem assessment (Millennium Ecosystem Assessment) concluded that “few well-designed empirical analyses assess even the most common biodiversity conservation measures”. Although there has been progress since 2005 – especially in the field of restoration ecology, if we want to elucidate causal relationships we need to use more experimental and quasi-experimental designs to evaluate environmental programs.

The new study by Hering and colleagues overcomes some problems I mention above and provides important evidence from the practice of river restoration. These authors used a quasi-experimental design to contrast the role of the restoration project size vs. the enhanced heterogeneity of instream substrates in restoring river biota across 10 European regions and restoration projects. They were very careful in preparing their study design, as the result of a river restoration project may be affected by confounding variables correlated with the location of the restoration actions. In each of the studied regions, they selected one extensively restored river section, one short restored section and two non-restored, degraded sections, each one upstream of the restored sections. More importantly, to account for the confounding factors I mentioned earlier, they compared restored sections to the corresponding pairs and their degraded control sections within each region. With this design they were able to compare (1) restored vs. degraded sections and (2) long vs. short restored sections. Finally, but also crucial for the success of the study, they measured the response of different biological and abiotic variables, including: hydromorphological variables, three aquatic organism groups (fish, benthic invertebrates and aquatic macrophytes), two floodplain-inhabiting organism groups (ground beetles and floodplain vegetation) and stable isotopes as indicators of land–water interactions and food-web interactions.

Left: Enns short restored section. Right: Ruhr long restored section. Photo credit: Daniel Hering.
Left: Enns short restored section. Right: Ruhr long restored section. Photo credit: Daniel Hering.

Hering and colleagues found a number of important results, but here I highlight three:

  1. When considering only restored vs. degraded sections, the success of a restoration project depends on the indicator you are using to measure success. Ground beetles and floodplain vegetation strongly responded to restoration. These were followed, almost in a linear trend, by aquatic biota, with abiotic variables and food-web interactions responding less.
  2. The most surprising result: considering long vs. short restored sections, none of the biological groups showed a statistical difference. The overall difference seems positive, but even with such careful study design there was too much variation in the response of all metrics, biotic or abiotic.
  3. Now the good news: a third type of comparison was performed by regrouping the river sections based on changes in substrate – i.e. sites with larger changes in substrate and habitat composition vs. those with smaller changes. Except for fish, most variables (biotic and abiotic) indicated a larger restoration effect in sections with larger changes in instream substrate and habitat composition.

These results have direct implications for the practice of river restoration. For strongly degraded sites, even small changes (e.g. wood and boulder placement) may reestablish aquatic-terrestrial linkages, which in turn allow colonization by floodplain biota with good dispersal abilities. For these organisms, instream barriers or minimum area available for spawning and rearing are of lesser importance. For other organisms, especially those that depend on continuous long-lasting heterogeneous river sections for dispersing, colonizing and maintaining viable populations, more profound restoration measures may be needed, including restoring longer river sections and larger substrate changes. Hering and colleagues suggested that even the longer sections they analysed (no longer than 2 km) seemed still too short to cause changes in most aquatic groups. Although the authors do not defend this idea, there might be a threshold size value after which restoration effects are stronger. In that case, the interplay between length and quality of the restoration project might be the key to more successful restoration projects.


Ferraro, P.J. (2009) Counterfactual thinking and impact evaluation in environmental policy. New Directions for Evaluation, 2009, 75–84.

Hynes, H. BN 1975. The stream and its valley. Verh. Internat. Verein. Limnol, 19, 1–15.

Palmer, M.A. (2009) Reforming Watershed Restoration: Science in Need of Application and Applications in Need of Science. Estuaries and Coasts, 32, 1–17.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s