Scaling is a naturally iterative and bi-directional component of problem solving in ecology and in climate science. Ecosystems and climate systems are unquestionably the sum of all their parts, to the smallest imaginable scale, in genomic processes or in the laws of fluid dynamics. However, in the process of scaling-up, for practical purposes thewhole usually has to be construed as a good deal less than this. This essay demonstrates how controlled large-scale experiments can be used to deduce key mechanisms and thereby reduce much of the detail needed for the process of scaling-up. Collection of the relevant experimental evidence depends on controlling the environment and complexity of experiments, and on applications of technologies that report on, and integrate, small-scale processes. As the role of biological feedbacks in the behavior of climate systems is better appreciated, so the need grows for experimentally based understanding of ecosystem processes.
We argue that we cannot continue as we are doing, simply observing the progress of the greenhouse gas-driven experiment in global change, and modeling its future outcomes. We have to change the way we think about climate system and ecosystem science, and in the process move to experimental modes at larger scales than previously thought achievable.