A predictive occurrence model for ‘Elatine gussonei’ based on environmental factors

Abstract

The occurrence of the amphibious plant Elatine gussonei in Mediterranean Temporary Ponds (EU priority habitat 3170* and Natura 2000 site) is patchy, with a strong stochastic component. The specific environmental conditions and tolerance ranges that determine its presence in a pool are not known. This represented the knowledge gap that this study aimed to address. The chemical, morphometric and ecological conditions in a sample of rockpools were collected and used to construct a preliminary occurrence model. Throughout this study, ca. 170 pools from 10 pool landscapes were surveyed from the Maltese Islands. Data collected and used in analyses included species lists, water quality (pH, Electrical Conductivity and Oxidation Reduction Potential), and basin morphometry (dimensions, surface area, maximum water and sediment depths of the basin and depths at which E. gussonei was present). Analyses carried out included correlation plots, CCA, RDA, t-tests, linear regressions (lm) and binomial logistic regressions (generalized linear models – glm). Elatine gussonei occurrence in the model was based on dichotomous presenceabsence data of the species. Therefore, binomial glms were carried out for environmental factors. Only zwm and surface area (both negatively correlated with occurrence) were statistically significant (p<0.05) and were used to model the species occurrence in a given pool. Once the presence of the species was confirmed via glm, lm were used to model the specific depths at which it occurs. The dependent depths (zwe and zse) were significantly positively correlated with independent maximum basin depths (zwm and zsm). Constraining the lms to pass through the origin, however, increased model efficiency by increasing R2 (0.54 to 0.72 and 0.44 to 0.84 for water and sediment depths), indicating better model fit. Literature states that its phenotypic plasticity and rapid response to environmental changes make it a good sentinel species on which to model climate change and predict further environmental changes and habitat status. The species and its habitat are both protected, entitling them to monitoring and conservation.