Published data from tropical forest sites with data on forest biomass and at least one component of NPP, with documentation of the methods underlying these data (and associated references)

It is clear that the currently available data on net primary production in tropical forests are extremely limited and that even our best estimates for this biome can only be thought of as rough approximations within wide bounds. Nevertheless, this study has provided a basis for evaluating the quality and utility of the data generated by past studies and of the NPP estimates that have been reported in the literature for tropical forests. Clearly, the most important knowledge gap is with respect to the belowground components of NPP in these ecosystems. In particular, fine root losses are likely to be a substantial proportion of NPP in many tropical forests, and they as yet remain unquantified. It is also clear that there have been many of the previous NPP studies in this biome have involved methodological problems and poor documentation. Henceforward, it is paramount that such studies be carefully designed and reported so as to avoid these pitfalls. A third issue that needs to be addressed is that of sampling bias and the lack of replication in either time or space for NPP studies in these forests. Given the very small and subjectively located plots that are the basis for most of the existing data, it is possible that the existing understanding of NPP processes in tropical forests is highly skewed toward the highest biomass patches (flat sites with large trees, no gaps) and is unrepresentative of the larger forest landscapes.