Cascading effects of dryland agriculture on animals
Research Background
Extensive land conversion to agriculture in drylands and associated resource use have wide-ranging impacts on desert ecosystems globally. Incorporating the impacts of human-social aspects is thus imperative in examining ecological interactions. The provision of agricultural inputs in these resource-scarce regions supports invasive and pest species, negatively impacting both agricultural productivity and native desert ecosystems. Understanding the spatial dynamics of invasive and pest species requires analyzing both bottom-up resource availability factors underlying animal distributions and top-down biological controls. Here, we evaluate the social-ecological cascading effects of dryland agriculture on vertebrate pest communities in dryland agricultural communities of Israel.
Our study region is characterized by 18 agricultural cooperatives with distinct crop regimes due to contrasting social decision-making and resource allocation schemes (i.e., communal kibbutzim vs. privatized moshavim). Crop choices further affect land management (e.g., enclosed vs. open farm systems) and resource intensity. This system is ideal for studying trophic mechanisms underlying animal assemblages between agricultural regimes. We examine the role of agricultural land-use practices on pest spatial distributions based on multiyear vertebrate pest observations with agricultural data sets.
Research Objectives
We use structural equation modeling (SEM) to quantify the relative importance of added agricultural resources underlying bottom-up and top-down trophic processes regulating vertebrate pest assemblages. Results reveal that crop choices determine pest distributions through bottom-up processes directly, while simultaneously driving pest competitive interactions through indirect top-down cascades impacting pest communities. For example, due to the indirect negative effect of wolves on mesopredators (foxes and jackals) mediated by livestock, the total positive effect of livestock on the abundance of mesopredators is reduced.
Results
We developed binomial GLMs for pest species not incorporated into the structural equation model (and goodness-of-fit measures) due to low abundance and distributional records (e.g., recently invasive boars) or limited interspecies interactions—that is, hyena and avian pests. Geographic variation in pest occurrences is mostly explained by agricultural variables. For example, the distribution of brown-necked ravens, which occur in high abundance around northern moshavim, is positively correlated with vegetables and livestock and negatively affected by dates, whereas Indian house crows, which are restricted mainly to the southernmost kibbutzim (representing an invasion frontier; Yosef et al., 2019), are negatively associated with most variables. Both corvid species are most often observed around livestock enclosures within settlements. Spectacled bulbuls and bee-eater pests, which have wide distributions but occur with higher frequency around northern moshavim, are positively associated with vegetables in these agricultural landscapes.
TABLE. Binomial generalized linear model coefficients of agricultural land variables of rare and avian species.
Our study illustrates the social-ecological cascading effects of agricultural regimes on pest community assemblages mediated by contrasting agricultural land-use practices. Considering the expansion of dryland agroecological systems globally, understanding the intricate cascading pathways of predator- and prey-pest communities has important implications for agricultural management, biological invasions in drylands, and fragile desert environments.
