6 Conclusion

Continents are undergoing rapid change. At fine scales, terrestrial landscapes are rapidly changing too, whatever they are, agricultural, forested, urban or arid. In addition, these four main landscapes are continuously meeting together in space and time, shifting from one of the previous types to another, under natural (climate change) and/or human driving forces (Figure 1*). To benefit from a relatively unified theory of such landscape dynamics would help to better understand each of these landscape types as well as each observed landscape type transition. We have discussed here only intrinsic needs to understand the landscape object, but such a review can obviously also serve extrinsic needs: for example, population dynamics, fire and pest spread, human production systems, climatic influence are environmental processes that would all benefit from an improved landscape modelling. Such improvements in dynamic landscape theory and modelling would greatly facilitate the (weak or strong) coupling between terrestrial landscapes and species and would serve climate change and biodiversity assessments.

We are now starting to have powerful models to manipulate various landscape dynamics and to better understand their functioning. Landscape research offers several interesting concepts to understand landscapes, but such concepts are insufficiently adapted to recent modelling approaches. The intrinsic dynamic properties of landscapes, their multiscale properties, the network and patch interactions they inhabit are difficult to manage by the simple use of heterogeneity and fragmentation notions. To improve as well as to harmonise current landscape modelling, we propose in this review and discussion paper to explore several actions: i) to develop explicit process-based models, handling both natural and human driving forces, simultaneously to neutral LMs; ii) to favour the interaction of various landscape type models; iii) to develop patchy, network (of various types) and multiscale modelling; iv) and above all, to develop a coherent theoretical framework of dynamic landscape, on the basis of new and powerful neutral models and mathematical formalisms.