Ecosystem engineering is a ubiquitous process where species influence the physical environment and thereby structure ecological communities. However, there has been little effort to synthesise or predict how ecosystem engineering may impact the structure and stability of interaction networks. To assess the current scientific understanding of ecosystem engineering impacts via habitat forming, habitat modification, and bioturbation on interaction networks/food webs, we reviewed the literature covering marine, freshwater, and terrestrial food webs, plant-pollinator networks, and theory. We provide a conceptual framework and identify three major pathways of engineering impact on networks through changes in resource availability and energy flow, habitat heterogeneity, and environmental filtering. These three processes often work in concert and most studies report that engineering increases species richness. This is particularly marked for engineers that increase habitat heterogeneity and thereby the number of available niches. The response of network structure to ecosystem engineering varies, however some patterns emerge from this review. Engineered habitat heterogeneity leads to a higher number of links between species in the networks and increases link density. Connectance can be negatively or positively affected by ecosystem engineer impact, depending on the engineering pathway and the engineer impact of species richness. We discuss how ecosystem engineers can stabilize or destabilize communities through the changes in niche space, diversity, network structure, and the dependency on the engineering impact. Theory and empirical evidence need to inform each other to better integrate ecosystem engineering and ecological networks. A mechanistic understanding how ecosystem engineering traits shape interactions networks and their stability will be important to predict species extinctions and can provide crucial information for conservation and ecosystem restoration.

To link engineering impact to network or food web responses, we searched the literature using a set of search strings in Web of Science.

We searched Web of Science on the 22nd of May 2023 with two search strings that complemented each other. The first one used the keywords "ecosystem engineer" AND "food web" OR network which resulted in 258 studies. The second search used "ecosystem engineer" OR facilitation OR "foundation species" OR "non-trophic" OR bioturbation AND ("food web" OR "trophic web" OR "interaction network*" OR "ecological network" OR "pollination network" OR "seed*dispersal network") and retrieved an initial number of 544 studies. This second search widened the review to studies beyond the word “ecosystem engineering”.

We only included primary research that measured the response of the community or network structure to the impact of an engineer species, resulting in 36 studies (with two studies reporting each two separate engineer impact cases). For each study, we extracted information about engineer identity, whether it was native or exotic, the community/food web/network, the engineer classification (habitat forming, habitat modification, bioturbation), the main outcomes for diversity, and the structure of the network or food web, such as network metric(s) and functional composition that have been measured.