With global carbon credits valued at over US$100 billion/year, accounting under REDD will drive demand for high quality forest monitoring systems. The choice of system to adopt should be guided by good science. The adequacy and comparability of different national systems for forest carbon measurement under REDD have not been fully evaluated. There is a growing body of scientific and technical information on ground-based and remote-sensing methods of carbon measurement. This extensive, often conflicting, knowledge base has not been systematically reviewed in the transparent, readily-repeatable manner consistent with evidence-based practice. This paper argues that such an approach, regarded in medicine as the gold standard for evidence evaluation, is long overdue in forestry generally and carbon monitoring and assessment specifically. Preliminary findings from an international project set up to scope the potential for a systematic review approach indicate that this evidence-based approach would add value to REDD implementation.