Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

Research article Quarterly Journal of Engineering Geology and Hydrogeology Published online July 23, 2019 https://doi.org/10.1144/qjegh2019-018 | Vol. 53 | 2020 | pp. 19–30

Comparison of chemical sediment analyses and field oiling observations from the Shoreline Cleanup Assessment Technique (SCAT) in heavily oiled areas of former mangrove in Bodo, eastern Niger Delta

Matthijs Bonte1, Erich R. Gundlach2, Ogonnaya Iroakasi3, Kabari Visigah3, Ferdinand Giadom4, Philip Shekwolo3, Vincent Nwabueze3, Mike Cowing5 & Nenibarini Zabbey4 1 Shell Global Solutions International B.V., 40 Lange Kleiweg, 2288GK Rijswijk, Netherlands 2 E-Tech International Inc., Boulder, CO 80302, USA 3 Shell Petroleum Development Company of Nigeria Ltd, Port Harcourt, Nigeria 4 University of Port Harcourt, Port Harcourt, Nigeria 5 Independent Consultant, 56 Temple Road, Epsom KT19 8HA, UK MB, 0000-0003-3191-9290;KV,0000-0001-8261-3924; PS, 0000-0001-6012-2884; NZ, 0000-0002-8787-7096 * Correspondence: [email protected]

Abstract: Trial pitting, borehole drilling, and soil, sediment and groundwater sampling are important components of oil spill response and contaminated land assessment. These investigations provide detailed information on the subsurface geology and contaminant occurrence and transport but have disadvantages including worker safety hazards, cost and time required for completion, and may cause cross-contamination among aquifers. An alternative to such investigations applied in oil spill response is the Shoreline Cleanup Assessment Technique (SCAT) approach, which relies heavily on direct visual observations to assess the severity of oil contamination and guide cleanup efforts. Here, we compare SCAT observations of oil type, surface coverage and pit oiling with collected surface and subsurface sediment samples taken concurrently and analysed for a suite of hydrocarbon constituents. Results indicate that although limited sampling and analysis is required to chemically characterize the contamination, SCAT observations can be calibrated using limited sediment sampling and are sufficient to steer physical cleanup methods. This is particularly evident as even closely spaced chemical samples show high variability. A coarser direct visual observation is fit-for-purpose considering the wide variability in contaminant distribution at even local levels. In this contribution, we discuss the limitations of the different methods. Supplementary material: The modified SCAT data collection form, figures showing subsurface versus ground surface total petroleum hydrocarbons (TPH) and a variogram of TPH measured in the ground surface and subsurface samples, and data tables are available at https://doi.org/10.6084/m9.figshare.c.4534682 Thematic collection: This article is part of the Measurement and monitoring collection available at: https://www. lyellcollection.org/cc/measurement-and-monitoring Received 28 January 2019; revised 13 May 2019; accepted 10 June 2019

The study area of Bodo Creek, located in the eastern part of the the Bodo area is the largest documented extent of spill-related Niger Delta (Fig. 1), was exposed to two oil spillages in 2008 damage to mangroves (Duke 2016; Gundlach 2018) and the largest resulting from leaks in the Trans Niger Pipeline operated by The cleanup of oiled mangrove habitat known to the authors. Likewise, Shell Petroleum Development Company of Nigeria Ltd (SPDC). the Shoreline Cleanup Assessment Technique (SCAT) and The affected area consists of low-lying mangrove habitat, numerous chemical sampling programmes described herein are also the tidal channels lined with very soft mud, and harder substrates along largest ever undertaken within oiled mangrove forest habitat. the upper intertidal mainland and island shorelines and along Before the two primary spills of 2008, the number of reported constructed fish ponds no longer in use. In 2015, SPDC agreed to spills in the area was relatively small (17 between 1986 and 2008), remediate 1000 ha of oiled former mangrove areas in the Bodo of which 10 were from illegal activities (Gundlach 2018). Illegal Creek. Phase 1 of the cleanup programme began in September 2017, activities causing spillage include pipeline tapping, connection from lasted for about a year, and focused on reducing sediment oiling in the tap to a vessel by (leaking) hoses, transport of stolen oil by a the top 30 cm by using surface agitation (raking) in lesser variety of open-hulled large (e.g. >30 m) and small wooden vessels, contaminated areas and deeper low-pressure, high-volume water and shore-side refining. After 2010, the number of spills caused by flushing in highly contaminated soft mud areas, most commonly illegal actions increased dramatically (over 30 in 2010–2011). found along intertidal channels. Phase 1 also included a mangrove These illegal activities and associated spillages have continued into replanting pilot programme to assess seedling survival in 2018, even as cleanup and the SCAT–chemistry programmes were contaminated sediments and a coring programme at 30 sites to continuing. Mangrove plant recovery since the primary loss in 2008 determine depth of oiling. Results will be provided in forthcoming is very limited and the need for planting is evident. An aerial view of publications. A more intensive Phase 2 remediation and mangrove part of a central portion of the study area in 2015 and 2018 is shown replanting programme is scheduled to start in 2019. The damage to in Figure 2.

© 2019 Shell Global Solutions International BV. This is an Open Access article distributed under the terms of the Creative Commons Attribution 4.0 License (http://creativecommons.org/licenses/by/4.0/). Published by The Geological Society of London. Publishing disclaimer: www.geolsoc.org.uk/pub_ethics Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

20 M. Bonte et al.

Fig. 1. Location of the Bodo Creek study area along the eastern edge of the Niger River delta in Rivers State, Nigeria. Background satellite image is copyright Google, image Landsat/Copernicus.

The method to assess the distribution of oil spills and the preferred undertaken concurrently. It should be noted that this study does methods of cleanup have evolved over several decades: from purely not assess the tolerance of mangroves to crude oil or total petroleum scientific investigations of surface and subsurface spill extent (e.g. hydrocarbons (TPH) levels, nor does it determine crude oil derived Gundlach & Hayes 1978; Gundlach et al. 1978) to using SCAT. contaminants in fish or other aquatic species. Both topics are the SCAT involves collaborative field teams that include the responsible subject of continuing investigations. party, regulator, landholder and representative from the local community and non-governmental organizations. Participants all view the site at the same time and make a consensual recommen- Methods dation regarding the appropriate response action (Owens & Sergy SCAT surveys 2003; Santner et al. 2011). SCAT was also called upon by the cleanup managers who are represented in the Bodo Mediation SCAT is a systematic method for surveying an affected shoreline Initiative (BMI) and SPDC to monitor and confirm when the spill- following an oil spill. The SCAT approach was developed during response contractor successfully completes the Phase 1 cleanup. The the response to the 1989 Exxon Valdez oil spill and has since been BMI was established following a mediation process in 2013 led by applied and further refined (e.g. NOAA 2013). It is a viable and the Dutch Embassy in Nigeria between the Bodo community and practical technique that maximizes the recovery of oiled habitats and SPDC (Zabbey & Arimoro 2017; Sam & Zabbey 2018). resources while minimizing the risk of further ecological deterior- SCAT is in essence a relatively quick, straightforward, robust and ation from cleanup efforts. participatory approach that contrasts with traditional land-based site SCAT field procedures in Bodo initially utilized standard forms assessments, which rely heavily on complex investigations and soil, (e.g. NOAA 2013; IMO-REMPEC 2009) but these were found sediment and groundwater sampling and analyses. Although inadequate to accurately describe the contaminated mud–mangrove drilling boreholes, sediment sampling and laboratory analyses root environments of the area affected by the spills. Principal provide detailed information on the subsurface geology and among the difficulties in using these forms was the time needed to contaminant occurrence and transport, they have disadvantages capture the information when dealing with security issues, boat including worker safety hazards, cost and time required for transport, the diurnal ≥2 m tide that floods the area, very soft muds completion, and may cause cross-contamination in the case of making walking to each site difficult, and the inability to determine deep boreholes that connect different aquifers and are improperly specific horizons of oiling when oil enters a pit. The modified SCAT sealed or drilled (Bonte et al. 2015). data collection form is included in the supplementary material. Although SCAT clearly has advantages over the traditional Specific to this paper, we use the visual estimation of per cent oil on drilling, sampling and laboratory analysis approach, a comparison the surface and per cent subsurface oil as observed on the surface of between SCAT data and chemical sample analyses has to our water in a pit dug to 25–30 cm and after waiting c. 5 min for the knowledge not been reported in the literature to assess the accuracy incoming water and oil quantity to stabilize. Both surface and of the direct field observations collected during SCAT. To bridge subsurface oil estimations were averaged based on at least three this gap, this work compares results of SCAT field observations SCAT members giving their estimation. Oil type and extent of with those of an extensive chemical sampling programme coverage (%) on the surface and subsurface area were categorized as Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

SCAT Observations v. Chemical Sampling, Niger Delta 21

Fig. 2. Overview of the central portion of the Bodo Creek study area showing living and dead (former) mangrove area. (a) photograph taken on 15 September 2015 (V. Imevbore photographer); (b) Google Earth image dated 2 January 2018. The location of the photograph is shown in Figure 4. sheen (silver or rainbow), brown or black oil, or tar or asphalt. oiling was 35% or less and therefore no Phase 1 work was required; Layers of brown oil (probably with some emulsification) and black ‘Pre-Treatment’, where treatment was required (pit oiling 35% or oil (no emulsification) are thicker than sheen (Bonn Agreement greater) but not yet started; ‘During Treatment’, where treatment 2017) and would be expected to show higher concentrations started but pit oiling was still 35% or greater, indicating that measured by chemistry. Figure 3 illustrates the surface and additional work was required; ‘Post-Treatment’, where pit oiling subsurface oil observations for two SCAT sites. was reduced to 35% or less and therefore Phase 1 was completed. A combination of surface and subsurface observations were made Sites were not repeated; for example, sites of ‘Pre-Treatment’ in a to steer the cleanup activities, where surface observations grid were not the same as ‘Post-Treatment’. For this reason, this determined whether raking to remove tar, dead stumps and paper reviews overall trends and is not site specific. garbage was needed, and subsurface pit observations determined An overview of the work grids and former mangrove area (in red) the need for sediment flushing. is shown in Figure 4. This figure also shows the location of the Where the percentage coverof blackor brown oil in the pit was 35% Figure 2 photograph as well as the sites of the two 2008 spills or less, the cleanup Phase 1 treatment was confirmed as completed. discussed above. No treatment was performed in areas within c. 3 m of living mangrove SCAT surveys were undertaken in August 2015 (35 sites) and from as the presence of living mangroves indicates that the residual 18 September 2017 to 30 August 2018 (911 sites). The chemical weathered oil in the soil is not restricting mangrove growth. sampling team participated with SCAT during the 2015 surveys and To undertake cleanup and SCAT field monitoring of the Bodo from September to December 2017. SCAT teams included represen- Creek oil spill site, the delineated 1000 ha was subdivided into 200 tatives from federal government (DPR, NAPIMS and NOSDRA), m × 200 m (4 ha) grids. Because of channels and presence of living state government (RSMENV), SPDC-ORP, BMI, the Bodo commu- mangrove, grid size varied from a full 4 ha to much smaller sizes nity, non-governmental organizations (NACGOND) and the two (e.g. tens of m2). Channels and live mangrove areas are not included cleanup contractors. Participating personnel and organizations, with in the 1000 ha to be treated. Two cleanup contractors were selected full versions for acronyms, are provided in the Acknowledgements. for the Phase 1 cleanup encompassing 535 grids. SCAT was applied first to advise on where and how to clean each grid and then to Chemical sampling monitor and verify when Phase 1 cleanup was adequately performed. As cleanup progressed, SCAT categorized the grid Previous chemical sampling in the area is limited (UNEP 2011; and work to be done into four categories: ‘No Treatment’, where pit Linden & Palsson 2013; Little et al. 2018). For this study, sediment Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

22 M. Bonte et al.

Fig. 3. Overview of SCAT observations at sites L10-1 (top) and I49-1 (bottom) with pits, SCAT surface and subsurface oiling observations, and surface and subsurface − TPH (mg kg 1). BO + BRO indicates black oil + brown oil. The locations of the sites are shown in Figure 4. Photographs by E. Gundlach. sampling for chemical analyses was carried out before, during and Chain of Custody documentation for laboratory chemical analysis after the first phase of cleanup in conjunction with the above- by methods detailed below. Pits were backfilled following SCAT described SCAT site investigations. Sampling sites were selected to data recording and site photography. In total, we utilize data from ensure a wide coverage of habitats, locations and oiling conditions. 317 SCAT sites that have observations of surface and subsurface Figure 5 provides an overview of the location of sites reviewed in oiling and results from chemical analysis. this paper. In addition to the composite sampling, for every 20 samples a At each chemical sample site, separate composites of five grab blind duplicate was collected consisting of a homogenized sample samples were taken from the surface (0–5 cm depth) and subsurface divided into two subsamples. Each duplicate portion was assigned (15–25 cm depth). One of the five pits was used for the SCAT visual its own sample number to be unknown to the analytical laboratory. surface and subsurface observations. The two selected sampling At four sites, the five individual (discrete) samples used for the depths are based on the observations from 2015 surveys, which composite sample were analysed independently for comparison to indicated that the most heavily oiled sediments are located in the top assess small-scale variability. 30 cm and that the associated cleanup method should predomin- This sampling plan follows the ISO 10381 standard for soil antly focus on this depth of oiling. quality sampling (ISO 2002). Laboratory analyses were carried out Figure 6 illustrates field-sampling activities. A clean stainless- using certified methods (MCERTS) with associated laboratory steel spoon was used to take sediment from each of five surface and quality assurance and quality control procedures. The 2015 samples subsurface locations spaced equally within c. 5 m around a centre were analysed by Alcontrol (UK) whereas the 2017 samples were point. The subsamples from surface and subsurface were placed into analysed by I2 (UK). separate and clean stainless-steel bowls. After thorough mixing, a Samples were analysed for total and fractionated TPH, fraction composite sample was packaged and shipped with accompanying banding as defined by the Total Petroleum Hydrocarbon Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

SCAT Observations v. Chemical Sampling, Niger Delta 23

Fig. 4. Phase 1 work area outline in yellow with grids, former mangrove, live mangrove, water and illegal refineries indicated. The yellow arrow indicates location of the photograph in Figure 2. The location of the ‘Spills 2008’ that initiated these cleanup and sampling activities is also shown. The labels on the x- and y-axes of the map represent geographical coordinates (latitude and longitude) in decimal degrees.

Criteria Working Group (TPHCWG 1998–1999), by gas chroma- (EGASPIN 2018). EGASPIN provides a tiered approach to tography–flame ionization detector (GC-FID) for non-volatile determine corrective requirements based on the Standard Guide hydrocarbons and by gas chromatography–mass spectrometry (GC- for Risk-based Corrective Action Applied at Petroleum Sites MS) for volatile hydrocarbons as well as individually reported prepared by ASTM (2015). The first tier comprises intervention benzene, toluene, ethylbenzene and xylene (BTEX) components. of ≥5000 mg kg−1, which is derived for a long-term residential Sixteen USEPA (2003) priority polynuclear aromatic hydrocarbons exposure scenario. The higher tier assessments (tiers 2 and 3) are (PAH) were analysed individually and as a sum parameter. more complex in that they consider receptors such as persons living The statistical significance of the difference in TPH concentra- near or using the area of contamination. Tier 2 and 3 assessments are tions between different treatment classes (i.e. No Treatment, During being carried out to derive site-specific target levels, which will Treatment and Post-Treatment, respectively) and the Pre-Treatment provide risk-based criteria that are considerate of local circum- concentration was determined with the nonparametric Mann– stances and are the topic of a subsequent publication. Some of the Whitney U test. The statistical significance of the differences in shorelines, particularly in areas close to Bodo, are used by fishers on TPH concentrations for the different SCAT observation groups (i.e. a regular basis. Most of the other damaged areas are visited oil sheen, brown oil and black oil) was tested with a one-way relatively infrequently. ANOVA test. Statistical calculations were performed using the Python SciPy module (Anonymous 2018). The spatial data structure was investigated by constructing variograms using the Python Results and discussion Pykrige module (Anonymous 2019). Distribution of TPH concentrations Nigerian regulations relating to oil spill cleanup and remediation are described in Environmental Guidelines and Standards for the Figure 7 and Table 1 present results of total TPH concentrations of Petroleum Industries in Nigeria issued in 1992 and updated in 2018 surface and subsurface samples as box–whisker plots for each of the Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

24 M. Bonte et al.

Fig. 5. Sampling locations (n = 317) with colours indicating TPH concentrations for surface (downward pointing triangle) and subsurface (upward pointing triangle) concentrations. The labels on the x- and y-axes of the map represent geographical coordinates (latitude and longitude) in decimal degrees. four Phase 1 treatment conditions: No Treatment, Pre-Treatment, found throughout the remaining area, probably owing to the spread During Treatment and Post-Treatment. Median and mean TPH of the initial spills as well as continued illegal activities. The concentration before Phase 1 cleanup (Pre-Treat) for the ground (G) absence of a spatial structure in the data is confirmed by the surface samples are c. 35 000 and 50 000 mg kg−1, respectively, variograms that were constructed for TPH analyses from ground with the 25%-ile and 75%-ile at 17 700 and 68 000 mg kg−1. Mean surface and subsurface samples and the exponential variogram values are typically higher than the median (50%-ile) because of model that was fitted to the data (supplementary material, Fig. SI-2). ‘hotspots’, which bias the mean upward. Surface concentrations are The fitted variogram models plot as horizontal lines with a nugget- considerably higher than subsurface values, with the latter having to-sill ratio of close to unity, which confirms the absence of any median and mean concentrations of 12 250 and 30 300 mg kg−1 spatial structure in the data (Cambardella et al. 1994). with the 25%-ile and 75%-ile at 2448 and 39 250 mg kg−1.Itis BTEX constituents and light aliphatic constituents are absent or interesting to note that although subsurface sediment concentrations significantly depleted from all samples as compared with analyses are consistently lower than surface concentrations, the correlation of unweathered Bonny Light crude oil. The 16 USEPA PAH sum between surface and subsurface collected at the same location is concentrations of all but one sample are below the Nigerian very poor (R2 = 0.12; Fig. SI-1 in supplementary material), high- regulatory (EGASPIN) threshold of 40 mg kg−1 (which is actually lighting the high degree of heterogeneity. The results of vibracoring for a subset of 10 PAH and not all 16 PAH defined by USEPA) and and analyses undertaken in September–October 2018 at depths of are not further discussed. 2–4 m reveal that the vast majority of petroleum hydrocarbon impact Post-Treatment TPH values show a statistically significant (P < is restricted to depths of <50 cm (data not reported here). 0.01) reduction for the subsurface samples with the median There is no clear spatial pattern present in the TPH concentrations concentrations decreasing from 12 250 to 7250 mg kg−1. (Fig. 5), other than lower concentrations present in the far south However, the median concentrations for the surface samples show even though that area is close to the location of one of the 2008 spills an increase from 35 000 to 39 500 mg kg−1, which was, however, (shown as a star at lower left in Fig. 4). High concentrations are not statistically significant. The No Treatment sites have statistically Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

SCAT Observations v. Chemical Sampling, Niger Delta 25

Fig. 6. Photographs showing sampling in grid cell P09 (4.62582N, 7.26295E) near a live mangrove. (a) Collection of field grab samples using a steel shovel; these samples are composited in a stainless-steel bowl and homogenized. (b) Close-up of the pit (c. 0.3 m across) showing droplets of oil. (c) Homogenization of grab samples. (d) Transfer of composite sample to laboratory supplied sample containers, which were subsequently transferred to ice boxes and transported to the laboratory. Photographs by M. Bonte.

Fig. 7. Box–whisker plots of TPH concentrations (in mg kg−1; log scale) for surface (G) and subsurface (S) samples under four different conditions: No Treatment, Pre-Treatment, During Treatment and Post-Treatment. The number above the top x-axis represents the number of samples in each class. The box extends from the lower to upper quartile values of the data, with an orange line at the median (or 50%-ile). The whiskers extend from the box to show the range of the data between the 5- and 95%-ile. The green triangle shows the mean value. Flier points are those past the end of the whiskers. An asterisk above a box plot indicates that the median is significantly different at a 99% confidence level (P < 0.01) from the Pre- Treatment value. Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

26 M. Bonte et al.

− Table 1. TPH (mg kg 1) for surface and subsurface sediment samples

No Treatment Pre-Treatment During Treatment Post-Treatment Ground surface Mean 40 300 49 500 93 200 62 600 25% percentile 38 000 17 700 37 500 13 975 50% percentile (median) 38 000 35 000 55 000* 39 500 75% percentile 41 500 68 000 82 750 87 750 Number of samples 3 121 26 163 Subsurface Mean 600 30 300 21 700 17 400 25% percentile 450 2448 1915 2223 50% percentile (median) 690* 12 250 6800 7250* 75% percentile 810 39 250 19 800 23 700 Number of samples 3 122 24 169

*The median is significantly different at a 99% confidence level (P < 0.01) from the Pre-Treatment value. significant lower TPH concentrations for subsurface samples consisting of a mixture of detritus (e.g. dead mangrove roots) and compared with the Pre-Treatment samples whereas for surface clumps of clay-to-sand-sized sediments. Additionally, contam- samples no statistical difference is found. ination was present in the form of isolated pockets of residual free The fractionated TPH data show that relatively heavy TPH oil droplets scattered through the sediment, which combined with fractions (>C16) dominate both the Pre- and Post-Treatment the problematic homogenization may cause large differences conditions (Fig. 8). Compared with unweathered Bonny Light between different parts in the composite sample (see photograph crude oil, the sediments are enriched in heavier fractions as a result in Fig. 6b). of weathering (combined volatilization and biodegradation) that To assess small-scale variability from the same sample site, the preferentially removed the more volatile and biodegradable five individual grab samples used for the composite sample were compounds (Brown et al. 2017a). In particular, the heavier aromatic analysed for four sites. Samples were collected within 5 m of a fractions have been enriched, consistent with findings for land centre point. Results show that TPH is highly variable on a small farming trials showing that although biodegradation decreases scale (Fig. 10). The mean of the individual discrete samples deviates concentrations of all TPH fractions, the rate of depletion is lower for considerably from that of composite sample from the same site. This heavier fractions compared with lighter fractions (Brown et al. is in line with the previously discussed results of blind duplicate 2017b). The distribution of different TPH fractions is similar both samples, further suggesting that the high RPD in blind duplicate between surface and subsurface samples as well as for Pre- sampling was a result of incomplete homogenization. The relative Treatment and Post-Treatment conditions. The physical cleanup standard deviation (ratio of standard deviation and mean) ranged method applied mobilizes free oil but does not preferentially remove between 29 and 111%, and 33% and 96% for the surface and the lighter TPH fractions (unlike biodegradation). subsurface samples, respectively. This is in the same order of magnitude as the average RPD for the set of blind duplicate Data reproducibility and variability samples. The small-scale variability in the presence of crude oil in For TPH, an average relative percentage difference (RPD) sediments that is observed visually through the SCAT process was between primary and blind duplicate samples was determined to investigating by replicate pit oiling observations in five SCAT pits be 63% (Fig. 9), above generally accepted criteria ranging within a 6 m radius (data shown in Table SI-3 in supplementary between 40 and 50% (e.g. NJDEP 2014; IDEQ 2017). The likely material). Similar to the TPH observations, these data demonstrate a explanation for the large RPD is that homogenization of the high variability in the observed oil coverage in pits with values composite samples was complicated by the texture of sediment ranging between 5 and 95% coverage. The relative standard

Fig. 8. TPH fraction distribution for fresh Bonny Light crude oil and ground surface and subsurface sediment samples, before and after cleanup. Ali, aliphatics; Aro, aromatics. Bonny #1 shows crude oil fractions from Brown et al (2017b). G, surface samples; S, subsurface samples; PreT, Pre-Treatment conditions; PostT, Post-Treatment conditions. Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

SCAT Observations v. Chemical Sampling, Niger Delta 27

Fig. 9. Results from blind duplicate sampling and analysis (n = 28). deviation of these observations ranges between 5 and 183%, with an (including during and after the spills in 2008), the dynamic tidal average RSD of 62%. environment causing variable oil settlement on intertidal sediments, Overall, the blind duplicate and discrete samplings, and replicate and the nature of the mud-dominated sediments inhibiting internal SCAT pit observations, show that TPH concentrations are highly oil movement with depth. The small-scale variability should, variable over short distances. The highly heterogeneous nature of however, also be placed in the context of the overall variation in TPH concentrations is probably the result of both the continued and TPH concentrations, which ranges over three orders of magnitude as spatially variable impacts to the area over the past 10 years discussed further below.

− Fig. 10. TPH (in mg kg 1) from five discrete samples taken within 5 m of a central point at each site. Grey bars show TPH, the red line shows the mean of the five discrete samples and the green line shows the composite sample result from the same site. Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

28 M. Bonte et al.

Fig. 11. Box–whisker plots of TPH concentrations (in mg kg−1; log scale) for different oil types as delineated by SCAT surveys for surface and subsurface samples. The box extends from the lower to upper quartile values of the data, with an orange line at the median (or 50%- ile). The whiskers extend from the box to show the range of the data between the 5- and 95%-ile. Flier points are those past the end of the whiskers. The number above the upper x-axis indicates the number of samples in each group.

Relation between chemistry and visual field observations (2) Sheening and the presence of droplets of oil in a SCAT pit can occur over a range of TPH concentrations. This implies that Comparisons between oil type and TPH concentrations are SCAT observations are not expected to provide an exact proxy for presented in Figures 11 and 12. In addition, Tables SI-1 and SI-2 TPH analyses, but rather that a certain SCAT observation can in the supplementary material summarize the key statistics for TPH provide a bandwidth of TPH concentrations. concentrations grouped by oil type. The comparison between SCAT described oil type and TPH concentrations indicates a good correlation for subsurface samples (Fig. 11, lower panel). The The ground surface samples do not show a correlation between different SCAT groups have statistically different TPH concentra- SCAT field observations and chemical analysis results (Fig. 11, tions, which is confirmed by an ANOVA test. Observations of upper panel); this is confirmed by the ANOVA test, which showed ‘silver sheen’ in the waters of a 25–30 cm test pit have a median no statistical difference in TPH concentrations for the different TPH value of slightly less than 2000 mg kg−1 and a 75%-ile value SCAT groups. The difference between ground surface and of c. 4000 mg kg−1, whereas brown and black oil have higher subsurface TPH correlations with SCAT observations is probably median TPH values of 8000 and 20 000 mg kg−1, respectively. The the result of the black organic-rich type colour of the sediment relative standard deviation (RSD) of TPH in the classes for the (believed to be a result of the high organic matter content and subsurface shown in Figure 11 range between 115% for black oil sulphate-reducing conditions) on which it is hard to differentiate and 219% for silver sheen (Table SI-1 in supplementary material), weathered oil from sediment. Field reports indicate that only when which is higher than that of the small-scale sampling discussed in oil was seen pooled on water or where oiling was fresh (black) could the previous section (33–96%). The higher RSD of TPH it be clearly seen on the sediment surface, whereas the oiling of pit concentrations within each SCAT category can be caused by a waters could be discerned more easily and more consistently. In – combination of the following factors. addition, the surface sample included sediments 0 5 cm deep, whereas the visual observation was solely on the sediment surface (1) SCAT observations were made in one pit whereas the (0 cm deep) and therefore does not include oil incorporated into the analysed composite sediment sample consisted of five grab sediment. samples. For future work, it is recommended to take the average Pit oiling of ≤25% brown or black oil may be one criterion to of three SCAT pits to account for the high degree of variability. provide a visual means to confirm completion of Phase 2 treatment Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

SCAT Observations v. Chemical Sampling, Niger Delta 29

Fig. 12. Box–whisker plots of TPH − concentrations (in mg kg 1; log scale) of Post-Treatment samples for different types of SCAT surface and subsurface oil observations classified according to a proposed close-out criterion for Phase 2 under which SS plus BO ≤25%. SS, silver sheen; BO, black oil; BRO, brown oil.

(compared with the performance criterion of <35% applied in Phase in heavy hydrocarbons compared with unweathered Bonny Light 1). This is believed to represent an oil level that can practicably be crude oil and with low BTEX and PAH components. The high achieved with flushing and physical agitation. In parallel to SCAT spatial variability that is present on both small and large scales, and cleanup activities, pilot replanting of mangrove seedling is however, limits the applicability to directly steer and confirm being carried out to determine if this cleanup target is adequate. A completion of cleanup activities using a single chemical-based comparison of SCAT data and chemical data shows that the visual analytical criterion. Given this variability, SCAT pit observations criterion corresponds to a median subsurface concentration of c. can be considered sufficiently reliable to guide confirmation of 4000 mg kg−1 (Fig. 12). For 75% of the sites sampled Post- Phase 2 cleanup in Bodo in conjunction with a risk assessment Treatment (upper box limit), a TPH concentration <10 000 mg kg−1 according to EGASPIN requirements. Another alternative being was found. The sites that do not meet the criterion have a 50%-ile reviewed is to reduce the level of contamination in the sediments and 75%-ile concentration of 20 000 and 35 000 mg kg−1. TPH sufficient that mangroves can be replanted and survive, aiding the concentrations measured in ground surface samples are much further degradation of remaining oil through phytoremediation higher, reflecting the inability of SCAT observations to reliably processes. Results from seven planted sites after 1 year show 82% detect surface oiling as compared with chemical values taken from survival of 346 seedlings planted and good growth (mean: +46% 0–5 cm depth. plant height) in spite of high TPH levels (2210–87 000 mg kg−1 surface, 420–59 000 mg kg−1 subsurface). Continued spillages, Conclusions however, will affect and kill replanted mangroves. We conclude that for the relatively large affected area under study Results show that the assessed cleanup area is affected by here, with challenges pertaining to site access and sample shipment, hydrocarbons that are widely spread and highly variable both visual SCAT observations can effectively be calibrated with a spatially and with depth. Spatial variability is both on a broad scale limited set of chemical field data. As such, the study provides an over the entire work area (several kilometres in length) and on very example of the application of SCAT as a speedy and reliable model small scales (within 10 m). Chemical sampling provides important to assess and manage cleanup endeavours in mangrove and tidal flat data to characterize the hydrocarbons present, demonstrating that environments. The comparison between SCAT observations and the they are primarily composed of highly weathered crude oil, enriched sediment chemical characterization did, however, reveal that SCAT Downloaded from http://qjegh.lyellcollection.org/ by guest on September 25, 2021

30 M. Bonte et al. observations for surface oiling are a poor indicator for surficial central Iowa soils. Soil Science Society of America Journal, 58, 1501–1511, sediment TPH concentration. This contrasts with subsurface (or pit) https://doi.org/10.2136/sssaj1994.03615995005800050033x Duke, N.C. 2016. Oil spill impacts on mangroves – supplementary data. Marine SCAT observations, which showed a much better correlation with Pollution Bulletin, 109, 700–715, https://doi.org/10.1016/j.marpolbul.2016. TPH concentrations. Calibration of SCAT observations (pure 06.082 product or sheen type) to sediment chemical analyses is also EGASPIN 2018. Environmental Guidelines and Standards for the Petroleum Industry in Nigeria, 3rd edn. Department of Petroleum Resources, Lagos. probably effective for rapid field assessment of other hydrocarbon Gundlach, E.R. 2018. Oil related mangrove loss east of Bonny River, Nigeria. In: products (e.g. different crude oils or gasoline spills), but this Makowski, C. & Finkl, C.W. (eds) Coastal Research Library (CRL): Threats requires further work given that the solubility and visual appearance to Mangrove Forests: Hazards, Vulnerability and Management Solutions. – will vary. Coastal Research Library, 25. Springer, Cham, 267 321, https://link.springer. com/chapter/10.1007/978-3-319-73016-5_13 Gundlach, E.R. & Hayes, M. 1978. Investigation of beach processes. In: Hess, Acknowledgements We would like to thank many people who W.N. (ed.) The Amoco Cadiz Oil Spill, A Preliminary Scientific Report. – participated in this programme for their persistence and good humour in spite NOAA EPA, Environmental Research Laboratories, Boulder, CO, pp. – of very difficult circumstances. These include: from the Bodo community: 85 196, https://www.oil-spill-info.com/Publications/1978_NOAA-EPA_ P. K. Lenu, N. N. Vite and F. Zorbilade; from the Bodo Mediation Initiative Amoco_Cadiz_Report_4_Beach_Processes.pdf (BMI) Directorate: P. Kpenu and O. Nwokoro; from the Department of Petroleum Gundlach, E.R., Ruby, C.H., Hayes, M.O. & Blount, A.E. 1978. The Urquiola oil Resources (DPR): B. Atiyegoba; from the National Oil Spill Detection and spill, La Coruña, Spain: impact and reaction on beaches and rocky coasts. – Response Organisation (NOSDRA): S. Oladipo, S. Ogwuche and A. Ismail B.; Environmental Geology, 2, 131 143, https://doi.org/10.1007/BF02430667 from National Petroleum Investment Management Services (NAPIMS): T. Etok; IDEQ 2017. Standard Operating Procedure for the Waste Management and from Rivers State Ministry of Environment (RSMENV): H. Chukwu, Mankie- Remediation Division Data Validation of Third-party Petroleum Data Tanen B. and C. Atteng, from The Shell Petroleum Development Company of Submittals SOP WST-2014-6, Version 1, State of Idaho Department of Nigeria Ltd (SPDC): A. Chibuzor, D. Essien, F. Igbodo, I. Olere, J. Edeh, Environmental Quality, https://www.deq.idaho.gov/media/60180050/rp- O. Sowagbein, T. Mzaga and V. Chukwudi; from National Coalition on Gas third-party-petroleum-data-validation-sop.pdf Flaring and Oil Spills in the Niger Delta (NACGOND): N. Zabbey; from cleanup IMO-REMPEC 2009. Mediterranean Guidelines on Oiled Shoreline Assessment. contractor Giolee-Lamor: A. McArthur, A. Dumbari, C. Alozie, S. Rickaby, Mediterranean Action Plan, Response Centre for the Mediterranean Sea, http:// L. Deeko, F. Beubizua, P. Parigas, K. Ekweozor and A. Abudun; from contractor www.rempec.org/admin/store/wyswigImg/file/Information%20resources/Guide Inkas: A. Cooke, S. Johnson, F. Nwabueze, J. Shekwolo, B. Kpenu, C. Dogalah, lines/RIS%20D13/EN/RIS%20D13-Mediterranean%20Guidelines%20on% P. M. Baabel and Z. K. Noble; and from the chemical sampling crew 20Oiled%20Shoreline%20Assessment%20(EN).pdf (GeoTerrain): K. Adesida, D. Adebayo, P. Ugbehe, J. Olanrewaju, D. Yisa, ISO 2002. Soil quality Sampling, Part 2: Guidance on sampling techniques. S. Ademiluyi, N. West (née Aburo), S. Babatunde, B. Ismail, E. Amadi, Standard nr. ISO 10381-2. F. IhedIohanma, I. Emefo, K. Briggs, F. Uche (née Uruakpa) and M. Taiwo. Linden, O. & Palsson, J. 2013. Oil contamination in Ogoniland, Niger Delta. – Thanks also go to I2 laboratory and ALcontrol laboratory (UK) for undertaking Ambio, 42, 685 701, https://doi.org/10.1007/s13280-013-0412-8 the chemical analysis. Finally, we thank J. Smith and G. Devaull of Shell Global Little, D.L., Holtzman, K., Gundlach, E.R. & Galperin, Y. 2018. Sediment Solutions for providing a thorough review of the draft paper. The views expressed hydrocarbons in former mangrove areas, southern Ogoniland, Eastern Niger are those of the authors and may not reflect the policy or position of SPDC or its Delta, Nigeria. In: Makowski, C. & Finkl, C.W. (eds) Threats to Mangrove Joint Venture partners. Forests: Hazards, Vulnerability and Management Solutions. Coastal Research Library, 25. Springer, Cham, 267–321, https://www.springerprofessional.de/ en/sediment-hydrocarbons-in-former-mangrove-areas-southern-ogonilan/157 Funding The Shell Petroleum Development Company of Nigeria Ltd 09984 (SPDC) funded the study and cleanup activities described in this paper. NJDEP 2014. Data quality assessment and data usability evaluation. Technical guidance. New Jersey Department of Environmental Protection, https://www. Scientific editing by Stephen Buss; Birgitta Beuthe nj.gov/dep/srp/guidance/srra/data_qual_assess_guidance.pdf NOAA 2013. Shoreline Assessment Manual, 4th edn. National Oceanic and Atmospheric Administration, Emergency Response Division, Office of References Response and Restoration, https://response.restoration.noaa.gov/sites/default/ Anonymous 2018. SciPy Reference Guide Release 1. 1.0.dev-659017f, p. 2481, files/manual_shore_assess_aug2013.pdf https://docs.scipy.org/doc/scipy/scipy-ref-1.1.0.pdf [last accessed 5 December Owens, E.H. & Sergy, G.A. 2003. The development of the SCAT process for the 2018]. assessment of oiled shorelines. Marine Pollution Bulletin, 47, 415–422, Anonymous 2019. PyKrige Documentation Release 1.4.1, p 52, https:// https://doi.org/10.1016/S0025-326X(03)00211-X buildmedia.readthedocs.org/media/pdf/pykrige/latest/pykrige.pdf [last Sam, K. & Zabbey, N. 2018. Contaminated land and wetland remediation in accessed 16 April 2019]. Nigeria: opportunities for sustainable livelihood creation. Science of the Total ASTM 2015. Standard Guide for Risk-Based Corrective Action Applied at Environment, 639, 1560–1573, https://doi.org/10.1016/j.scitotenv.2018.05. Petroleum Release Sites. Active Standard ASTM E1739. American Society for 266 Testing and Materials, West Conshohocken, PA, https://www.astm.org/ Santner, R., Cocklan-Vendl, M.C., Stong, B., Michel, J., Owens, E. & Taylor, E. Standards/E1739.htm 2011. The Deepwater Horizon, MC 252-Macondo Shoreline Cleanup Bonn Agreement 2017. Aerial Operations Handbook, Part 3, Annex A. The Bonn Assessment Technique (SCAT) program. In: International Oil Spill Agreement Oil Appearance Code, https://www.bonnagreement.org/site/ Conference, Portland, Oregon, 23–26 May 2011, abstract 270, http:// assets/files/1081/special_on_volume_calculation_20160607.docx ioscproceedings.org/doi/pdf/10.7901/2169-3358-2011-1-270 Bonte, M., Zaadnoordijk, W.J. & Maas, K. 2015. A simple analytical formula for TPHCWG 1998–1999. Total Petroleum Hydrocarbon Criteria Working Group the leakage flux through a perforated aquitard. Groundwater, 53, 638–644, Series, Volumes 1–5. Amherst Scientific Publishers, Amherst, MA, https:// https://doi.org/10.1111/gwat.12239 hero.epa.gov/hero/index.cfm/reference/details/reference_id/3381245 Brown, D.M., Bonte, M., Gill, R., Dawick, J. & Boogaard, P.J. 2017a. Heavy UNEP 2011. Environmental Assessment of Ogoniland. United Nations hydrocarbon fate and transport in the environment. Quarterly Journal of Environment Programme, Nairobi, https://postconflict.unep.ch/publications/ Engineering Geology and Hydrogeology, 50, 333–346, https://doi.org/10. OEA/UNEP_OEA.pdf [last accessed October 2018]. 1144/qjegh2016-142 USEPA 2003. Federal Register (43 FR 4108). US Environmental Protection Brown, D.M., Okoro, S. et al. 2017b. Comparison of landfarming amendments to Agency, Washington, DC, https://www3.epa.gov/region1/npdes/permits/ improve bioremediation of petroleum hydrocarbons in Niger Delta soils. generic/prioritypollutants.pdf Science of the Total Environment, 596–597, 284–292, https://doi.org/10.1016/ Zabbey, N. & Arimoro, F.O. 2017. Environmental forcing of intertidal benthic j.scitotenv.2017.04.072 macrofauna of Bodo Creek, Nigeria: preliminary index to evaluate cleanup of Cambardella, C.A., Moorman, T.B., Novak, J.M., Parkin, T.B., Karlen, D.L., Ogoniland. Regional Studies in Marine Science, 16,89–97, https://doi.org/10. Turco, R.F. & Konopka, A.E. 1994. Field-scale variability of soil properties in 1016/j.rsma.2017.08.004