Monitoring tropical forest degradation and restoration with satellite remote sensing: A test using Sabah Biodiversity Experiment

Abstract

Selective logging has been so extensive that harvested forest now exceeds unlogged areas in most tropical forest regions outside of the Amazon. In response, in Southeast Asia, enrichment planting with dipterocarp tree species is carried out in an attempt to accelerate restoration of forest structure and functioning. However, assessing the impacts of degradation (from selective logging and other causes) and subsequent restoration with field measurements is expensive and time-consuming. There is therefore a need to develop methods for the assessment of forest quality using remote sensing. Here, we use high spatial resolution satellite imagery and advanced remote sensing products to monitor the pattern and dynamics of estimated vegetation cover, Leaf Area Index (LAI), and the biomass of plots within a field-scale (500 ha) replicated and randomized manipulation that compares different forest restoration treatments with naturally regenerating controls within the Sabah Biodiversity Experiment (SBE). We also compare the biodiversity experiment plots with the surrounding area of the Malua Forest Reserve that was selectively logged for the second time in 2007. In general, satellite remote sensing detected differences in degradation between the once- and twice-logged areas as well as between the different experimental restoration treatments. We found that approximately 70% of the Malua Forest Reserve experienced a decrease of vegetation cover after the selective relogging in 2007, while the Sabah Biodiversity Experiment area that was not relogged showed increasing vegetation cover. Within the experiment, we found that plots restored using Enrichment line planting, had higher remote sensed vegetation cover (Mean ± SE: 66.90 ± 0.06 vs. 61.96 ± 0. 16) and LAI (Mean ± SE: 5.09 ± 0.03 vs. 4.61 ± 0.11) than that of unenriched plots. Among the enrichment planted plots, those planted with mixtures of (4 or 16) species exhibited higher vegetation cover (Mean ± SE: 67.72 ± 0.06 vs. 65.35 ± 0. 09) and LAI (Mean ± SE: 5.29 ± 0.04 vs. 4.82 ± 0.06) than that of monoculture plots. Overall, when our test case of the Sabah Biodiversity Experiment was viewed through the lens of remote sensing indicators, satellite imagery was able to detect changes in forest quality due to selective logging and restoration enrichment planting. Furthermore, our results suggest that diverse mixtures of planted tree seedlings enhance restoration of forest canopies compared to planting with single species. Confirmation with ground data will be needed to validate these results and to better understand the biological processes determining tropical forest degradation and restoration.