MS dissertation · 2023
GEDI mangrove canopy height and biomass
Wall-to-wall canopy height and above-ground biomass in the arid mangroves of the Indus Delta, assessed against field plots.
MS research at the Geospatial Research and Education Lab, Institute of Space Technology, Islamabad, supervised by Dr Hammad Gilani, with Dr Adeel Ahmad (University of the Punjab) as external examiner. Field data came from a research collaboration between the Institute of Space Technology and WWF.
Published abstract
Mangrove ecosystems are of critical importance, providing valuable ecological benefits such as coastal protection, biodiversity conservation and carbon sequestration. Accurate mapping of canopy height (CH) and aboveground biomass density (AGBD) of these forests is essential for understanding their structural characteristics and carbon storage potential. We develop a wall-to-wall mapping approach for CH and AGBD estimation in the threatened Indus Delta mangroves (Pakistan) while assessing the Global Ecosystem Dynamics Investigation (GEDI) Light Detection and Ranging (LiDAR) data which provides high-resolution three-dimensional insights into the structure of forested lands, improving our ability to assess overall ecosystem health. GEDI mission provides footprints spaced apart along the satellite track, resulting in a sampling approach that captures a representative subset of the forests. Using Random Forest (RF) Machine Learning (ML) models, representative samples from GEDI L2A RH98 and L4B AGBD data are separately predicted against remotely sensed features and indices calculated and selected from a fusion of ALOS PALSAR-2, Sentinel-1 SAR and Sentinel-2 optical imagery. The resulting continuous surfaces are assessed against in-situ plot-level CH averages and AGBD measurements, i.e., min, max and mean of 1.63, 8.93 and 3.42m, and 0.71, 134.73 and 22.13 Mg/ha respectively. The validated maps after applying regression yield R2, RMSE, min, max and mean of 0.662, 0.5m, 1.68m, 9.06m and 2.84m respectively for CH, and 0.57, 17.47 Mg/ha, 3.78 Mg/ha, 208.23 Mg/ha and 17.35 Mg/ha respectively for AGBD which are comparable with insitu measurements. However, the initially produced wall-to-wall AGBD map based on GEDI observations is shown to significantly underestimate AGBD in our study area, with min, max and mean of 0.82, 17.81 and 1.84 Mg/ha compared to the respective in-situ measurements. Nonetheless, these results indicate the usefulness of GEDI mission for cost-effective large-scale mapping by exploiting the potential of geo-computing cloud platform of Google Earth Engine (GEE) and contributes towards national inventorial mapping, facilitating efforts to protect and restore these endangered habitats.
Ammar, A. (2023) Assessing GEDI LiDAR data for mangrove canopy height and aboveground biomass mapping in Indus Delta, Pakistan. 21st Swiss Geoscience Meeting, Mendrisio. Abstract (PDF) · Conference page
Current work
A fuller study with Dr Gilani is under way, comparing more LiDAR sources and published canopy height maps with stricter validation. The code and text stay private until publication.