Seeing a wood through trees

According to the Nobel prizewinning chemist Manfred Eigen, a theory has only the alternative of being right or wrong. A model has a third possibility: it can be right, but irrelevant. In this book, David Mladenoff and William Baker have assembled a number of contributors who have each sought to identify relevant (and they hope right) representations of some aspect of forest change at the landscape scale.

Models of forest dynamics are important tools. They enable us to test our understanding of the mechanisms that generate the observed changes, and they also allow us to predict the consequences of managing forests in certain ways.

Forests are complex systems and the questions that each set of authors have asked are such that the investigators have invariably resorted to computer simulations, rather than mathematical analyses. The problems addressed are diverse, but in general each contribution centres on describing, spatially and temporally, explicit simulations of one of two related processes affecting forests: ecological succession or disturbance. For example, the chapter by John Caspersen et al asks how tree species communities are likely to respond to landscape-scale variation in soil moisture.

Typically, the models employed take a decidedly unattractive name (often an acronym) such as Sadie, Zelig or Landsim and each operates by orchestrating a wide variety of sub-routines. Detailed descriptions of how models are put together do not make the most riveting bedside reading, but they are necessary, given the scope of the book. Some models (eg the chapter by Caspersen et al ) take an individual-based approach and account for the position and fate of every plant in the landscape, while other models consider dynamics at the level of populations of plants in patches (eg, the chapter by David Roberts and David Betz). Clearly, there are trade-offs here between the amount of detail and the spatial extent that can be considered, and there is a thoughtful chapter Dean Urban et al) that considers several ways in which fine-scale models can be extended to larger landscapes.

Overall, the book amply fulfils its aim of highlighting the relationships between a variety of modelling approaches, but it also provides a deeper appreciation of the value and inevitable limitations of the models that are presented.

One weakness with this book is that the vast majority of its authors are North American, and this is reflected in the book's emphasis on boreal and temperate forests. That said, there are two interesting chapters: one on the impact of adjacent areas on tropical forest diversity, and another on the factors driving deforestation in the Brazilian Amazon.

Unfortunately, landscape ecology is still in that rather tedious phase of development in which it is trying to define itself as a discipline within ecology, and identify underlying principles and approaches. Readers with a more practical interest may prefer to skip these homilies and look at the problems that each chapter raises, to evaluate for themselves how these problems have been addressed.

The models outlined in this book make useful tools for quantitatively evaluating specific hypotheses: in this respect they are entirely relevant. However, as the editors acknowledge, there is still a long way to go to convince forest managers fully of the value of employing these types of model in practical decision-making.

Tom Sherratt is senior lecturer in ecological modelling, University of Durham.