Global models of plant and ecosystem processes have now existed for a quarter of a century – and it shows. Although such models can now be found as fully embedded components of advanced Earth System models, their theoretical and empirical bases remain just as flimsy as when they were first designed. This is the fundamental reason why current models disagree in their future projections of land-surface and carbon-cycle changes, and routinely fail independent tests. To make matters worse, such models are continually being “improved” by adding new (usually unimportant) processes, rather than re-visiting the central ones. This approach is creating a downward spiral, whereby ever-increasing complexity is accompanied by a proliferation of uncertain parameters and an inevitable decline in predictive power.
In this talk, I will argue that quantitative models should, above all, be a tool to promote scientific understanding – the role they normally play in other, less politicized areas of science. I will first give a few examples of the spectacularly poor performance of today’s “frankenmodels” against some sets of observations. I will then show how new theory, founded on the general principles of competition and natural selection, leads to relatively simple and tractable mathematical representations of key processes that our new “data-rich” world permits us to test. I will demonstrate the success of this approach in predicting community-level values of some of the key quantites required in lage-scale ecosystem models, including photosynthetic capacity, air-to-leaf CO2 drawdown, leaf nitrogen content, gross and net primary production, leaf mass per area, and green vegetation cover. In some respects, these findings are consistent with a long-standing strand of ecological thought concerning the efficient use of multiple, scarce resources by plants. In other respects, however, they are proving controversial – because they require the replacement of certain pervasive (but incorrect) ecological “laws” with others.
Iain Colin Prentice FRS is Chair of Biosphere and Climate Impacts and Director, Leverhulme Centre for Wildfires, Environment and Society, Imperial College London
