Gains and losses in ecosystem services and disservices after converting native forest to agricultural land on an oceanic island

Abstract

Habitat conversion to agricultural land is one of the main threats to terrestrial biodiversity and can affect ecosystem processes and cause changes in ecosystem services (ESs) and disservices (EDs). Yet, studies often rely only on the abundance and diversity of the service providers; the effects on ecological processes of habitat conversion are rarely directly monitored. In this study, we used the sentinel approach to evaluate how habitat conversion from native forest to agricultural land affected ESs and EDs on an oceanic island. We quantified herbivory on lettuce plants, invertebrate and vertebrate predation rates on artificial caterpillars, pollination on strawberry plants, seed predation on wheat and mustard seeds, and leaf decomposition rates in native forests, maize fields and pastures on Terceira Island, Azores (Portugal). Herbivory, invertebrate predation rates, and pollination service were not significantly different between habitats. Vertebrate predation rates in native forests (mean 6.1% d−1) were significantly higher than that in pastures (0.3% d−1), or high-elevation maize fields (0.5% d−1), and marginally higher than in low-elevation maize fields (2.2% d−1). Overall seed predation after 48 h was significantly higher on wheat (mean 16.8%) than mustard seeds (5.6%). High-elevation maize fields also had higher seed predation (27.8%) than low-elevation ones (0.6%) or pastures (3.6%), but did not differ from the native forest (12.9%). Decomposition after 90 days was highest in pastures (78.4% and 45.9%, for tea and rooibos, respectively); although no significant differences between habitats were detected, except for low-elevation maize fields (64.4% and 33.6%). Conversion from native forest to cultivated land did not cause a clear decrease in the intensity of the studied ESs/EDs except for vertebrate predation. Using direct monitoring tools to simultaneously and consistently quantify multiple ecological processes is not only possible but needed, as ecological processes can respond differently to landscape changes.