Aim(s):Methods for estimating carbon sequestration in riparian forest biomass are needed to qualify riparian restoration and conservation projects for funding through carbon credits. As part of a proposed accounting methodology for California's capand-trade system, we described five riparian vegetation associations and created a tool to predict their rates of carbon accumulation under different restoration and land-use scenarios.Location: California, USA. Methods:We assembled a database of 654 forest inventory plots of known age (3-108 years) from replanted and naturally recruiting California riparian forests. We then used detrended correspondence analysis and agglomerative hierarchical clustering to detect vegetation structure in the plots and delineate distinct forest types.We used allometric equations to estimate biomass from individual trees' diameters and calculate total biomass for each plot. Next, we fitted sigmoid growth functions to the plot biomass data to generate expected values for live-tree biomass over a 100year period for each forest type. Finally, we created an algorithm to match userinputted data for an intended restoration project with a look-up table that predicts carbon accumulation appropriate to the user's scenario.Results: California riparian vegetation was divided into willow scrub, cottonwoodwillow, mixed riparian, upland riparian and riparian woodland types. Total carbon sequestration in riparian vegetation and soil varied from 75.7 to 137.4 Mg C/ha at 30 years post-restoration and from 95.1 to 175.8 Mg C/ha 100 years after stand initiation. Vegetation types dominated by taller, more shade-tolerant trees developed more slowly than scrubby and pioneer stands, but came to a higher maximum biomass. Our models put riparian biomass accumulation in the range of the few extant literature values for mediterranean systems, but uncertainties are high, and soil carbon may be underestimated. Conclusions:Our workflow and methods should be transferable to the development of carbon accounting tools for any other woody vegetation type. However, our difficulty in finding appropriate published data for the analysis suggests a critical need for field surveys appropriate to biomass estimation in woodland and forest | 585

No abstract