Biomass and volume models for different vegetation types of Tanzania

Abstract

Climate change and high rates of global carbon dioxide (CO2) emissions have increased the attention paid to the need for high-quality monitoring systems to assess how much carbon (C) is present in terrestrial systems and how these change over time. The choice of a system to adopt relies heavily on the accuracy of the method for quantifying biomass and volume as important primary variables for computing C stock and changes over time. Methods based on ground forest inventory and remote sensing data have commonly been applied in the recent decade to estimate biomass and volume in the tropical forests. However, regardless of the method, accurate tree level biomass and volume models are needed to translate field or remotely sensed data into estimates of forest biomass and volume. Therefore, the main goal of this study was to develop biomass and volume models for the forests, woodlands, thickets, agroforestry systems and some selected tree species in Tanzania. Data from destructively sampled trees were used to develop volume and above- and below-ground biomass models. Different statistical criteria, including coefficient of determination (R2), relative root mean square error (RMSE %) and Akaike Information Criterion (AIC), were used to assess the quality of the model fits. The models selected showed good prediction accuracy and, therefore, are recommended not only to support the ongoing initiatives on forest C Measurement, Reporting and Verificatio (MRV) processes but also for general forest management in Tanzania.