Conclusions

Multiscale modeling presents a number of challenges beyond standard statistical modeling. In particular, there is the need for consistent modeling across different scales. The methods presented in this book represent a wide variety of such approaches, having been grouped into three general categories: methods for building a model at the finest levels that can be simplified for coarser scales, methods for explicitly modeling the relationship between scales, and methods that implicitly link the scales. We expect that additional new approaches will continue to be developed in the future.