Indicators for Tropical Forests Degradation Assessment: A Multitemporal Perspective Using Lidar Data

Abstract:

Laser scanning technology (lidar) has the ability to capture three-dimensional structural information from forests. This study aimed at assessing forest structure and dynamics of logged and unlogged forests on a ranch in the Eastern Brazilian Amazon (Paragominas Municipality, Pará State) using high resolution (1 m) lidar derived data acquired in 2012 and 2014. We used the 2012 coverage (including areas logged between 2007 and 2010) to estimate canopy and ground damage: canopy gaps (defined as contiguous areas > 10 m², with height < 10 m) were mapped from a canopy height model (CHM) and ground damage was mapped using a relative density metric that calculates the proportion of returns at understory level versus total returns. Additionally, we used 2012 and 2014 coverages (including areas logged between 2007 and 2013) to identify areas of height loss by comparing the CHM from both dates. Forest turnover and gap formation was greater in areas logged between the two acquisition dates (10.4 gaps/ha and 893m² of gaps/ha). Older logged areas (2007 to 2010) presented intermediate values (9.89 gaps/ha and 346m² of gaps/ha), while unlogged areas showed the smallest values (7.10 gaps/ha and 270m² of gaps/ha). The greater amount of gaps in the older logged compared to the unlogged forests suggests that there is an increased tree mortality associated to the post-logging period. We found significant differences in the number of canopy gaps between undisturbed and logged forests (p-value<0.001), but not among logging dates (p-value>0.428). In contrast, we observed differences at the understory level among the logged areas (p-value<0.015), and no difference between unlogged forests and areas logged in 2007 (p-value>0.765). Our findings suggest that canopy damage lasts longer than ground damage, and therefore, add relevant information for long-term degradation assessments.