Conservation of interacting species in network‐constrained environments

Abstract

Aim:We apply a novel approach in a spatial network context to identify factors influencing a parasite-host system and to distinguish focal areas for conservation of interacting species.Location: Twenty river networks in southern Europe (Spain) Methods: Spatial stream network (SSN) models were applied to analyse and predict density and biomass of interacting species in river networks. Density of an endangered freshwater mussel (parasite) and biomass data of its host fish were response variables for models with fixed-effect biotic and abiotic predictors and three random effect models for spatial covariance. Universal kriging with the SSN models was used to obtain predictions for parasite density and host biomass for the entire study region.Results: Spatial stream network (SSN) models fitted separately for parasite density and host biomass explained 75% and 77% of the variance, respectively. Predictors explained 5% of the variance for the parasite and 14% for the host. Host biomass was the most important predictor of parasite density. The tail-up and tail-down spatial covariances fitted to the residual variance explained more than a half of the total variance in both models. Significant biotic and abiotic predictors differed between the parasite and host models. Combining model predictions for parasite density and host biomass identified areas where different strategies might be employed to conserve biotic interactions.Main conclusions: Conservation of biotic interactions requires consideration of respective ecologies of the interacting species. In a network environment, connectivity of habitats can be an important determinant for occurrence of biotic interactions.Spatially explicit analyses, such as the SSN, can identify focal areas for conservation of biotic interactions. Conservation focused on biotic interactions, as opposed to a single species, could yield benefits to the focal species, the biotic community with which they occur and the ecosystem that supports them. K E Y W O R D Sfreshwater mussels, geostatistics, parasite-host biotic interactions, Salmonidae, spatial stream network models, universal kriging

Combining geostatistical and biotic interaction model to predict amphibian refuges under crayfish invasion across dendritic stream networks

“…In the full mixed models, spatial dependencies accounted for a far greater proportion of variation in species occurrences than the fixed component, as observed in other studies using geostatistical tools (Filipe et al, ; Lois et al, , Lois & Cowley ; Quaglietta et al, ). This was probably because the fixed component was specified using variables that are spatially structured, either Euclidean (e.g.…”

“…Incorporating predictors describing spatial autocorrelations in invasive species occurrences to account for unmeasured dispersal and colonization processes may help mitigating, albeit not solving, problems associated with non‐equilibrium conditions in SDMs (De Marco et al, ; Václavík et al, ; Filipe, Quaglietta, Ferreira, Magalhães, & Beja, ). For alien species invading river ecosystems, SDMs can be improved using geostatistical models accounting for spatial dependencies in physical and ecological processes across stream networks (Filipe et al, ; Lois et al, ; Lois & Cowley ). These models are similar to conventional mixed models, with species occurrence modelled in relation to environmental variables using a logistic function, and spatial autocorrelation considered in the random errors (Peterson & Ver Hoef, ; Peterson et al, ; Ver Hoef & Peterson, ; Ver Hoef, Peterson, & Theobald, ).…”

“…distances along the waterlines) representing flow‐connected relations (tail‐up models); and (c) hydrologic distances irrespective of flow connection (tail‐down models) (Ver Hoef & Peterson, ). This approach can easily incorporate biotic interactions by including predictors describing the occurrence or abundance of potentially interacting species in the fixed component (Lois et al, , Lois & Cowley ). Another possibility is to develop a geostatistical model for the invasive species itself, and then use the fitted response (i.e.…”

Beta diversity and specialization in plant–pollinator networks along an elevational gradient

“…Recently, conservation management has shifted its focus from strictly preserving species diversity to also embracing the maintenance of community structure, function or ecological processes like pollination (Lois & Cowley, 2017;Tylianakis, Laliberté, Nielsen, & Bascompte, 2010). Plants and pollinating animals form complex interaction networks.…”

Different Conceptualizations of River Basins to Inform Management of Environmental Flows

“…The network describes the interactions between parts of the basin mediated by the patterns of connectivity. The precise nature of the network model will depend on the biota being considered with linear networks being appropriate for macroinvertebrates (Clarke et al, 2008) and fish (Lois and Cowley, 2017;Radinger et al, 2017), while a dispersed network may be more appropriate for waterbirds (Kingsford et al, 2010;Pedler et al, 2014) and their passengers (Reynolds et al, 2015).…”