Response of ground spiders to local and landscape factors in a Mexican coffee landscape

• We studied ground spiders in coffee and forest sites in Chiapas, Mexico.
• Ground spider diversity tended to be higher in coffee (low-shade) than in forest.
• Forest sites had distinct spider species composition, mainly in the dry season.
• Local factors had stronger influence on spider diversity than landscape factors.
• Seasonality had a strong effect on the ground spider community.

In order to secure the provisioning of ecosystem services, detailed analyses of the relationship between biodiversity and agriculture are required. We studied ground spider diversity in a 52 km2 coffee landscape in Southern Mexico, and asked the following questions. (1) How do coffee management variables and local microhabitat variables change among coffee agroecosystems and forest sites and across seasons? (2) How does coffee management affect ground spider richness, abundance, and composition? (3) How do local and landscape factors influence ground spider richness and abundance? and (4) What role does seasonality play in shaping ground spider communities? During the dry season and rainy season of 2011 we sampled ground active spiders using pitfall traps from high and low shade coffee agroecosystems (27 sites) and from forest (10 sites). On local scale, for each 20 m × 20 m site we measured leaf litter variables, invertebrate dry biomass, slope of the terrain and elevation, and management variables such as canopy cover, shade tree richness, shade tree density and proportion of Inga trees. At the landscape scale, we measured distance to the nearest forest and percent of forest in buffers of 500 m. Results show that agricultural management had a strong influence on spider richness and abundance. Across seasons, local spider richness and abundance had or tended to have higher values in the low-shade coffee. Spider richness and abundance were strongly influenced by physiographic and local predictors and weakly by landscape predictors. Furthermore, predictors varied with seasonality, with slope of the terrain being the strongest predictor in the dry season and canopy cover being the strongest predictor in the rainy season. We conclude that ground active spiders in this coffee landscape are greatly influenced by coffee management and local characteristics.