The Global Change and Terrestrial Ecosystems (GCTE) project has twin o
bjectives: to predict the feedback effects of changes in terrestrial e
cosystems on the atmosphere and climate; and, at a finer scale, the ef
fects of global changes on the structure and function of natural and a
gro-ecosystems. The main feedback effects are via exchanges of energy,
water and momentum, and changes in biogeochemistry. All these process
es are affected by ecosystem composition; to obtain a quantitative und
erstanding of the interactions involved, it is necessary to describe v
egetation in terms of its functional characteristics, based on a gener
ally applicable classification of plant functional types (PFTs). Globa
l-scale simulations of induced changes in ecosystem composition have p
rogressed from non-dynamic, top-down models to mechanistic, ecophysiol
ogical models that predict PFTs and leaf-area index or biomass using p
lant responses to environmental conditions. Two such models are now be
ing incorporated into general circulation models (GCMs), as first atte
mpts at coupled atmosphere-biosphere models. Parallel developments inc
lude global-scale process models, which use fixed vegetation types to
predict net primary production and nitrogen cycling, and a bottom-up m
odelling approach. The latter begins with detailed, mechanistic patch-
scale models, extending to spatially variable versions involving lands
cape processes and gradually incorporating simplifying routines to all
ow for large data sets. The many implications of global change for agr
iculture are illustrated by two examples: cereal crop-effects; and pes
ts and diseases.