Quantifying forest carbon stocks and changes in support of the Kyoto Protocol

This thesis brings together research conducted on field based, remote sensing and modelling approaches to meet reporting requirements set by the Kyoto Protocol. Parties are given the option to meet part of their greenhouse gases reduction requirements through the conservation and enhancement of the carbon stored in forest ecosystems. Two contrasting forests (Monks Wood, UK, 52°24' N, 0°14' W and Thetford UK, 52°30' N, 0°30' E) were selected for the development and assessment of the selected methods. Field-based measurements were used to quantify carbon stocks in Monks Wood, providing the first exhaustive assessment of the carbon content held in a UK semi-natural woodland. The total carbon content of the stands varied from 346 to 616 tonnes per hectare (t ha-1) and highlighted the importance of broadleaved woodlands as carbon stores in the UK. A quantitative appraisal of remote sensing methods was also provided. For land cover discrimination, both optical and radar remote sensing have been successful. For forest carbon stock estimation, LiDAR approaches may provide the only viable remote sensing tool for this purpose. As a result, a LiDAR-based method was developed and the results compared to field-based estimates. At the stand level, the agreement between the field-based and the LiDAR estimates was r=0.85. At the woodland level, due to the enhanced capability of LiDAR to monitor the natural variability of carbon across the woodland, the estimates were nearly 24% lower than those from the ground. Remote sensing of field-based approaches are unsuitable alone for quantifying below-ground carbon content and can be resource intensive. Process-based models enable an estimation of below-ground components to be made. Much uncertainty however arises from the lack of information available on model parameter values. The 3-PG model was used to simulate forest production in Thetford forest and a Bayesian calibration was applied. The results showed that this statistical approach could provide an overall framework for integrating and quantifying the uncertainty in the combined field based, remote sensing and modelling datasets, a result highly relevant in the context of the Kyoto Protocol.