Indigenous forest wood volume estimation in a dry savanna, Zimbabwe : exploring the performance of high - and - medium spatial resolution multispectral sensors

Continuous, accurate estimation and mapping of forest dendrometric characteristics such as wood volume using remote sensing is fundamental for better understanding the role of forests in the carbon cycle and for informed management strategies for forests and woodlands. In this study, we tested whether and to what extent spectral vegetation indices derived from new generation high-spatial resolution multispectral sensors (WorldView-2 and GeoEye-1) estimate indigenous forest wood volume when compared to medium-spatial resolution broadband sensors (Landsat 5 Thematic Mapper) based on two savanna woodland types in Zimbabwe. Subsequently, the best regression model relating wood volume and vegetation indices was applied to map the indigenous forest wood volume of two study sites. Our results showed that spectral vegetation indices derived from new generation multispectral sensors of high resolution yielded plausible indigenous forest wood volume estimates when compared to spectral vegetation indices derived from medium resolution broadband Landsat 5 TM. The findings of our study demonstrated that accurate estimates of indigenous forest wood volume can be derived using vegetation indices derived from high spatial resolution images. Furthermore, these findings emphasise the importance of vegetation indices derived from high resolution satellite images for estimating forest structural attributes, such as wood volume, among others.