Abstract: Beneficial arthropods that provide biological control of aphids or weed seeds use avariety of habitats in agricultural landscapes. Information on the movement behaviour of thesearthropods between these habitats is needed to develop conservation strategies that sustain pestsuppression in agricultural landscapes. Models for movement behaviour may help to understandand explore biocontrol functions. As measurements of behaviour at the landscape scale aretechnically difficult to make, measurements are often made at smaller scales. It is then necessaryto upscale to larger scales, using movement models. Here we present a case study on suchupscaling. The first results indicate that upscaling from small scales to large scales, using acorrelated random movement model, may result in errors. An alternative approach, to be tested infurther work, is to fit the movement model directly to the large scale data.