No adequate community-wide mechanism or set of standards currently exists to ensure the long-term reproducibility of results from numerical modeling in biogeochemical research. Various investigators maintain copies of code and supporting materials through ad hoc methods, and some models and modeling studies are not archived by any mechanism other than the published manuscripts describing specific research results. This state of affairs invites trouble.
To illustrate this situation, one of the authors of this article was recently involved in producing a state-of-the-art review on the topic of the effects of climate change on forests. The literature on this topic contains a rich array of results from dynamic models of global or regional vegetation produced by forcing these models with climate change scenarios.