ABSTRACT
At the landscape scale, controlled experiments are often logistically or –nancially dif-–cult, and precision landscape measurement tools such as remotely sensed images and yield monitors may provide observational data that support scienti–c interpretation. Nevertheless, if it is possible to carry out a controlled, replicated experiment under the same conditions as an observational study, then the controlled experiment provides more powerful results. Often the results of observational studies and controlled experiments can be used in a complementary manner to address a question from different perspectives. The controlled experiment in this ideal scenario provides precise, readily interpretable answers to speci–c questions, and the observational study permits those answers to be extended to more general conditions than those under which the controlled experiment is conducted.
