Using Unmanned Aerial Vehicles and Carbon Assimilation Rate Measurements to Estimate Carbon Capture for Red Sea Mangroves

  • Mariana Elias Lara (King Abdullah University of Science and Technology (KAUST) (Creator)

Dataset

Description

To meet ambitious net-zero greenhouse gas emission targets by 2050, large-scale CO₂ reduction and removal are required. Nature-based solutions have been proposed as a potential aid to this process. Mangrove ecosystems, as well as their conservation and restoration, have the potential to make significant contributions in Saudi Arabia and other coastal regions. While field measurements of carbon assimilation rate and leaf area index (LAI) in mangroves provide important insights into carbon fluxes, they are time-consuming, labor-intensive, and limited when covering large areas. To address this issue, multispectral images captured by unmanned aerial vehicles (UAV) are used to generate spectral vegetation indices, which can then be used to build regression models for estimating mangrove LAI and carbon capture capabilities. The carbon assimilation rate measurements in the field for studying both diurnal and sub-seasonal fluxes revealed that Avicennia marina has a high carbon assimilation rate peak in the morning, which decreases thereafter, and a smaller peak in the afternoon. Furthermore, comparing all the studied sites, the KAM site (June) had the highest morning overall carbon assimilation rate values, ranging from 15- 20 μmol CO₂ m⁻² s⁻¹, followed by Island (October) ranging from 10- 17 μmol CO₂ m⁻² s⁻¹, and finally Rheem (February) ranging from 5- 15 μmol CO₂ m⁻² s⁻¹. Moreover, the acquired multispectral images were used to generate spectral vegetation indices, which were then used as input to build a random forest algorithm for estimating the LAI of mangroves. Following an evaluation of each mangrove site, the Rheem site dataset yielded the best Random forest algorithm (R²= 0.88 and RMSE= 0.39), so this model was used to create high resolution spatially distributed LAI-based maps for all of the mangrove sites studied. Knowing the carbon uptake per leaf area as well as the total leaf area (based on UAV-derived LAI estimates) within a mangrove site enabled us to create carbon capture maps (kg C yr⁻¹ per pixel) for all of the sites studied. To enable a more complete carbon accounting of mangrove ecosystems, future research should explore remote sensing approaches for inferring carbon assimilation in both belowground biomass and soils.
Date made available2022
PublisherKAUST Research Repository

Cite this