Modeling below-ground biomass to improve sustainable management of Actaea racemosa, a globally important medicinal forest product
Abstract
Non-timber forest products, particularly herbaceous understory plants, support a multi-billion dollar industry and are extracted from forests worldwide for their therapeutic value. Tens of thousands of kilograms of rhizomes and roots of Actaea racemosa L., a native Appalachian forest perennial, are harvested every year and used for the treatment of menopausal conditions. Sustainable management of this and other wild-harvested non-timber forest products requires the ability to effectively and reliably inventory marketable plant components. However, few methods exist to estimate below-ground biomass (rhizomes and roots) based on above-ground metrics. To estimate the relationship of above-ground vegetation components to below-ground biomass, data from a long-term sustainable harvest study of A. racemosa was used to develop a predictive model for rhizome mass. Over 1000 plants were extracted from two sites in the Central Appalachian Mountains of Virginia. Measurements of plant height and canopy dimensions were matched with corresponding green weights of rhizomes and roots. A multi-staged process was used to fit a mixed effects model. A random effects structure was selected using Akaike’s Information Criterion, while the fixed effects structure was simplified through backward selection using likelihood ratio tests. Over 500 plants were harvested from three neighboring sites to evaluate the effectiveness of the model in predicting below-ground biomass based on above-ground metrics. The relationships between above and below-ground biomass of plants from the sustainability study sites and the validation study sites were similar, indicating effectiveness of the model. Predicted values for the validation data were, on average, slightly larger than the observed values, indicating a small bias. The 95% prediction intervals computed from the model, however, covered the true values more than 95% of the time. This study demonstrates that estimating marketable rhizome biomass of native medicinal plants is feasible at a stand level. The model will serve as a valuable tool for inventorying forest products, allowing estimation of below-ground biomass based on above-ground metrics. Use of this tool will aid in developing effective inventory and management strategies for wild-harvested medicinal plants. Adaptation of this model to other species will encourage efforts toward sustainable use of non-timber forest products worldwide.