Geo-Spatial Analysis of Oil Spill Distribution and Susceptibility in the Niger Delta Region of Nigeria

Abstract

Oil spill occurrence during exploration, production and distribution can cause deleterious impact on the environment. Contamination of local streams/rivers, farmlands, forest resources and biodiversity in oil producing areas presents strong significant possibility of significant harm to human health. Geo-information technologies present new opportunities for assessing stress environment and ways of determining exposure susceptibility in such areas. The study assesses the geographical distribution of oil-spills cluster and pattern using three geospatial techniques with ground data at 443 oil-spill incident sites from 1985-2008. The places with high (high-volume/ large impact/close proximity to communities) and low incident (low-volume/less impact/fardistance) are related to the quantity of oil-spills identified within those communities considered susceptible to spill impact and possible exposure. While the average nearest neighborhood analysis showed a probability that oil-spill distribution in the area is clustered (ratio < 1 with index value 0.19), the Getis-Ord General G test indicated that the oil-spill with high quantities (volume) discharge are significantly clustered within every 400 m. The Moran's I index indicted that there is <1% likelihood that the clusters are as a result of random chance. These findings will help to combat the environmental problems and risks of prolong exposure to petroleum hydrocarbons by addressing future incidents or relocating oil facilities/communities and positioning of rapid response strategies.

A Geospatial Appraisal of Oil Spill Health Impacts: A Niger Delta Case Study

“…A geospatial analysis of the distribution of oil spills in the Niger Delta defines a significant clustered pattern (p-value of ≈0.0004 and z-score of ≈3.53), suggesting that incidence rates are non-random and closely associated with petroleum extraction and refining activities in the region (Figure 5). These results are consistent with previous investigations and further supported by the spatial autocorrelation statistics given in Table 1 [34]. Subjecting infant diarrhea incidence to a spatial autocorrelation analysis found a significantly clustered pattern of cases for the region (p-value = 0.02, z-score 2.27).…”

Mapping terrestrial oil spill impact using machine learning random forest and Landsat 8 OLI imagery: a case site within the Niger Delta region of Nigeria

“…Non-polluted sample sites are necessary in this study for two main reasons: first, for the identification of oil-free (non-polluted) landcover types within the study area and secondly for an effective discrimination between pixels of oil-free and oil spill–impacted landcovers. Proximity analysis as suggested by (Obida et al 2018; Park et al 2016; Whanda et al 2016) provided the basis for the selection of the polluted and oil-free vegetation pixels. The minimum rule was set that all non-polluted sites must be located at least 600 m away from all polluted sites based on the maximum area of spill recorded.…”

“…Table 1 shows the distribution of the polluted spill sites and oil-free sites according to their respective landcover classification schemes. To this end, 30 m buffer ring polygons were established around all the training sites to ensure that only adjacent pixels within the high consequence area close to the point of impact are selected specially for the polluted sites (Alexakis et al 2016; Whanda et al 2016).…”

Consistently unreliable: Oil spill data and transparency discourse

“…Squaring with this, prior to the establishment of NOSDRA the above referenced 1995 Shell Environment Brief recorded that between 1989-1994, 71% of oil spilt by volume was due to corrosion and operational problems whereas only 28 percent was due to sabotage (ibid, p 11). Whanda et al also demonstrate the considerable proportion of spills attributable to corrosion and production error from 1989 -2004 based on Department of Petroleum Resources data predating NOSDRA's establishment (Whanda et al, 2016). But this causal attribution changes markedly in the subsequent data sets (Obida et al, 2018).…”