Scaling processes and problems

Authors

  • P. G. JARVIS

    Corresponding author
    1. Institute of Ecology and Resource Management, University of Edinburgh, Darwin Building, Mayfield Road, Edinburgh EH9 3JU, UK
    Search for more papers by this author

P. G. Jarvis, Institute of Ecology and Resource Management, University of Edinburgh, Darwin Building, Mayfield Road, Edinburgh EH9 3JU, UK

ABSTRACT

A personal view is presented of current approaches to scaling processes and variables in the context of better understanding ecosystem function and predicting the consequences of global environmental change. Issues considered include spatial and temporal scales of interest, the scaling process, scaling strategy, scaling problems, heterogeneity, patchiness and non-linearity, aggregation methodology and feedbacks.

Knowledge of processes in plants and vegetation is largely at small scales. The transfer of this knowledge up to larger spatial and longer temporal scales is an open-ended process with potential errors arising from heterogeneity and patchiness in the distribution of processes and non-linearities in the functional relationships between processes and environmental variables. Scaling now covers several orders of magnitude with respect to spatial and temporal scales with attendant risks of propagating errors.

At larger scales the wide diversity of vegetation classes poses a problem, and it is suggested that this can be countered by classifying classes of vegetation (not species) into a small number of ‘functional types’ of vegetation. Scaling through summation of component processes and through derivation of appropriately averaged parameters is considered. However, the increasing role of feedbacks at larger spatial and longer temporal scales is an essential feature of the scaling process. Thus, understanding the feedbacks and including them in upscaling schemes is a major priority.

A scaling strategy is outlined to minimize the propagation of errors. Because the scaling process is open-ended it is essential that good models are used and tested at each increase in scale.

Ancillary