Spatial analysis of plant species distribution among small water bodies in an agricultural landscape

• Main Authors: Beata Bosiacka, Krzysztof Pacewicz, Paweł Pieńkowski
• Format: Article
• Language: English
• Published: Polish Botanical Society 2012-12-01
• Series: Acta Agrobotanica
• Subjects: midfield water bodies; plant species; distribution; spatial autocorrelation; lattice model; agricultural landscape
• Online Access: https://pbsociety.org.pl/journals/index.php/aa/article/view/1709

Due to their small size (0.02-1.0 ha), distinct boundaries, and conditions highly contrasting with those in the surrounding area, midfield water bodies are regarded as the so-called habitat islands. Their effective conservation calls for knowledge on their inhabitants' dispersal potential. However, direct empirical data are available for very few species only, but dispersal potential of a species may be inferred indirectly, from its distribution. The study addressed the question whether there is spatial autocorrelation in the distribution of plant species in midfield water bodies, or if the distribution is random. Spatial distribution of the midfield water bodies surveyed was analysed using the CrimeStat software, while spatial autocorrelation in distribution of 29 species was explored with the joincount.test routine of R CRAN software. Explorative spatial data analysis (ESDA) involving join-count statistics showed the presence of positive spatial autocorrelation in the distribution of ten hydro- and helophytic species. In their case, ESDA made it possible to reject the random distribution hypothesis, which opens up an avenue for exploring spatial patterns. Activities promoting the occurrence of species with limited dispersal potential should take into account their preferences in terms of shorter distances between neighbouring sites. This should make it possible to plan conservation of midfield water bodies not only as refuges, but also as stepping stone habitats facilitating migrations of wild species growing in an agricultural landscape.