| Agriculture, forestry, and other human activities have transformed natural landscapes into a patchwork of natural habitat surrounded by intensively managed areas. This has many effects on natural biodiversity, including dramatic changes in the local assemblages of pollinating insects. Many studies have documented ecological effects, such as changes in pollination services and in plant community composition across gradients of land-use intensity. Generally, these studies do not consider the recent insight that evolution by natural selection can occur much more rapidly than traditionally thought. Indeed, there is strong evidence that plant mating- and pollination systems can evolve quickly in a changing pollination environment, which need to be incorporated into forecasting models and in management decision-making.
We will contribute to knowledge-based landscape management while explicitly considering the potential for rapid evolution. Our strategy is to collect data on plant-pollinator interactions over several years across many populations of our model plant-pollinator system, Viscaria vulgaris (Tjärblomster), and then combine these data with data on local land-use into predictive models. Our basic data collection will include yearly surveys of plant populations and their pollinators, phenotypic traits, and reproductive success (fruits and seed set). We will then choose a set of populations characterized by distinct pollinator-assemblages and local land-use, and conduct detailed studies of natural selection and evolutionary potential. Evaluating evolutionary potential across many populations is now feasible because recent results from the applicant show that genetic variation in functional traits can be reliably inferred from the amount of standing phenotypic variation.
We will combine the resulting datasets into predictive models, allowing us to make predictions for how each ecological pattern (e.g. the composition of local pollinator assemblages) and eco-evolutionary process (e.g. natural selection and evolutionary potential) varies along land-use gradients. Our analyses will be easily interpretable by quantifying, for example, the percent change in fruit set per percent change in the cover of a land-use category. These gradients could represent, for example, variation in the local cover of a specific crop, say oilseed rape.
By considering evolutionary processes, we will be able to take this exercise one step further. By combining the model-predicted patterns of natural selection and evolutionary potential, we will be able to calculate, for example, how many generations would be required for a specific population to evolve to its new predicted optimum (as determined by the predicted local pollinator assemblage). By doing so, we will contribute to a more predictable, sustainable, and quantitative management science.
We will disseminate our results both to colleagues (through journal articles and conference presentations), and to practitioners involved in landscape management, especially in our core study region (Scania). This way we will ensure that our results can be of use for those directly involved in management decisions.
|