Ecosystem restoration planning requires quantitative rigor to evaluate alternatives, define end states, report progress and perform environmental benefits analysis (EBA). Unfortunately, existing planning frameworks are, at best, semi-quantitative. In this paper, we: (1) describe a quantitative restoration planning approach based on a comprehensive, but simple mathematical framework that can be used to effectively apply knowledge and evaluate alternatives, (2) use the approach to derive a simple but precisely defined lexicon based on the reference condition concept and allied terms and (3) illustrate the approach with an example from the Upper Mississippi River System (UMRS) using hydrologic indicators. The approach supports the development of a scaleable restoration strategy that, in theory, can be expanded to ecosystem characteristics such as hydraulics, geomorphology, habitat and biodiversity. We identify three reference condition types, best achievable condition (A BAC), measured magnitude (MMi which can be determined at one or many times and places) and desired future condition (ADFC) that, when used with the mathematical framework, provide a complete system of accounts useful for goal-oriented system-level management and restoration. Published in 2010 by John Wiley & Sons, Ltd.