Application of Remote Sensing in the Assessment of Oil Pollution Impacts on Biodiversity in Rivers State, Nigeria
2020-03-04T15:10:04Z (GMT) by
Biodiversity loss remains a global challenge, and monitoring methods are often limited in their coverage. Rivers State is a biodiversity hotspot because of the high number of endemic species endangered by oil pollution. This thesis investigates the potential of integrating remote sensing tools for monitoring biodiversity in the State using vascular plant species as indicators. Satellite data from Hyperion, Sentinel 2A and Landsat were analysed for their usefulness. Soil samples from polluted and control transects were analysed for total petroleum hydrocarbon (TPH), phosphorus (P), lead (Pb), temperature, acidity, species diversity, abundance and leaf chlorophyll concentration. Field data results showed significant differences in all variables between polluted and control transects. Average TPH on polluted transects was 12,296 mg/kg, and on control transects was 40.53 mg/kg. 163 plant species of 52 families were recorded with Poaceae and Cyperaceae the most abundant. Floristic data ordinated on orthogonal axes of soil parameters revealed that TPH strongly influenced species occurrence (r = -0.42) and abundance (r = -0.39). Similarly, application of the spectral variability hypothesis (SVH) revealed the underlying environmental gradient controlling vegetation composition on polluted transects as TPH and on control transects as P. Models of relationship between spectral metrics and soil properties estimated soil TPH (R2 = 0.45) and P (R2 = 0.62) with marginal errors. Hyperion data provided better insight into vegetation response to oil pollution. Continuum removed reflectance, band depths of absorption maxima, red edge reflectance all significantly differed between polluted and control vegetation. Furthermore, a new index created from TPH sensitive Hyperion wavelengths- normalised difference vegetation vigour index (NDVVI) outperformed traditional narrowband vegetation indices (NBVIs) in models estimating species diversity in Kporghor. R2 and RMSE values for Shannon’s index were 0.54 and 0.5 for NDVVI-based models and 0.2 and 0.67 for NBVI-based models respectively. This research provides evidence of oil pollution effect on vegetation composition, abundance, growth and reflectance and outlines how this information can be used for biodiversity monitoring.