A third of the forest area in Central Africa has been granted to logging companies. Logging is highly selective in the region, with an average of 0.7 to 4.0 trees harvested per ha, but its direct impact on forest structure and the spatial variation of this impact remain understudied. . 
Here, we investigated the direct impact of logging on forest structure, we related this impact to logging intensity and canopy opening. We compiled unique datasets collecting field measurements and aerial observations in four FSC certified concessions. Our data includes pre- and post-logging inventory of forest plots covering 38 ha, records of over 6,000 harvested trees, and drone RGB images covering over 6,000 ha.
In average, logging activities reduced forest above-ground biomass by 8.8%, stem density by 6.5%, basal-area by 8.5% and canopy cover by 4.4%. Strong relationships were found between the reduction in biomass, stem density, or basal area with logging intensity, canopy opening and the number and volume of harvested trees (rRMSE between 0.128 and 0.164). Additionally, we demonstrated that canopy opening can be a good indicator to monitor and upscale logging intensity (rRMSE between 0.0005 and 0.0022).
This study is the first covering extensive inventory plots and UAV (uninhabited aerial vehicle) images before and after logging in different locations in Central Africa, providing a valuable reference to evaluate the impact of logging on forest structure. It demonstrates how canopy opening can be used to estimate measurements usually collected in the field and provides to the remote sensing community a unique dataset that will help improving monitoring systems (https://hdl.handle.net/2268/323683). These findings also have significant implications to control and manage logging activities, especially for certification standards, forest administrations, and European regulations.