Groundwater remediation action plan of polluted aquifer by heavy fuel oil, Northern Jordan Open Access

Heavy fuel oil (HFO) spills pose a critical environmental threat, especially to groundwater systems, due to their persistence and bioaccumulative effects (Etkin 2014; Michel & Fingas 2016). HFO spills commonly result from industrial sites, pipelines, and transportation systems, leading to widespread contamination in both surface and subsurface environments (Pan et al. 2018; Sam & Zabbey 2018). The impact of these spills varies with oil type; while crude oil spreads slowly, light non-aqueous phase liquids (LNAPLs) such as HFO migrate rapidly through soil and fractured rock, reaching groundwater (Cassidy 2007; Pan et al. 2012; Koosha et al. 2019).

In karst aquifers, such as those in northern Jordan, hydrocarbons can infiltrate groundwater through dissolution porosity and turbulent flow in conduits, bypassing natural filtration processes. Groundwater velocities in such aquifers range from 218 m/day during base flow to up to 9,500 m/day during peak flow, which intensifies containment and remediation challenges (Quinlan & Ewers 1985; Bakalowicz 2005). Developing a conceptual site model (CSM) that captures contaminant pathways is essential for selecting effective remediation strategies (Kresic & Mikszewski 2012; CL 2014; David et al. 2017).

The Al Hussein Thermal Power Station (HTPS) in northern Jordan, a critical facility situated over a karst aquifer, has experienced groundwater contamination due to HFO leaks. This aquifer is a crucial water source for the region and understanding the movement of LNAPLs such as HFO through porous and fractured media is vital for devising effective remediation strategies (Adelana et al. 2011). To address this, the study employs a combined approach using field investigations, hydrogeophysical data, and pH-REdox-EQuilibrium (PHREEQC) modeling to assess the impacts of HFO biodegradation on groundwater chemistry, particularly redox conditions and mineral dissolution.

The selected pump-and-treat remediation approach at HTPS aims to reduce hydrocarbon concentrations in groundwater and prevent further contamination. The findings contribute to understanding HFO behavior in karst systems and support the development of improved groundwater management strategies for similar hydrogeological settings (Park & Park 2011).

The facility receives fuel oil directly from the Jordan Petroleum Refinery Company and stores it in 12 primary tanks, with a reserve of 35,000 tons. Water needs are met by groundwater from an underlying aquifer, yielding 235 m³/h.

The proposed plant will be constructed on brownfield land leased from CEGCO. It will be separate, both physically and functionally, from the HTPS, and will operate independently. This includes having its own entrance, fuel and water supplies, and security measures. The project company will lease the land from CEGCO, covering a total area of 149,992 m².

The site is bordered by the existing HTPS concrete perimeter wall on its northern, eastern, and western sides. While there were several groundwater wells, pump houses, and water pipelines on the site, these have been capped. Previously, these facilities supplied water for HTPS operations from depths of up to approximately 80 m. An incoming potable water pipeline serving the existing plant and local area enters the proposed site at its extreme eastern point. Additionally, three overhead transmission line pylons, whose bases are within the project footprint, will remain in place for the combined cycle gas turbine (CCGT) project.

The geology near HTPS includes wadi fill deposits composed of gravels, limestone boulders, basalts, sands, and silt (Howard & Humphreys 1983). The outcropping rocks belong to the Amman (B2) and Kharj Limestone Formations, characterized by undulations, fracturing, and jointing of chert beds. The Amman formation is divided into silicified limestone (lower unit) and phosphorite (upper unit).

The hydrogeology features an unconfined upper aquifer (Amman Aquifer, B2) and a lower aquifer (Wadi Sir Aquifer, A7), separated by the Umm Ghudran aquitard (B1). The distinguishing feature of this formation is its undulations in addition to fracturing and jointing of the chert beds. This formation is subdivided into two units: the lower unit is the silicified limestone unit, and the upper unit is the phosphorite unit. The wadi fill deposits are confined to the wadi bed and overlie the Amman Formation (B2) and Wadi Umm Ghudran Formation (B1). They consist of sand gravels with clay and have a variable thickness from 15 to 20 m (Howard Humphreys 1983). The Amman aquifer in the vicinity of HTPS is overlain by wadi fill deposits that are in hydraulic continuity. This aquifer on a local scale is well-jointed and fissured and exhibits solution channels and karstic features. In the northern parts of the area a Tertiary basaltic flow crops out and extends from the Samra treatment plant to the north along the Zarqa River.

According to the geotechnical survey undertaken in 2021, the bedrock underlying the site belongs to the Amman Formation (B2). The bedrock relating to the Amman Formation of the Campanian Age consists of limestone with bands of chert. The bedrock relating to the Quaternary age consists of black basalt. These formations extend to more than 20–30 m below existing ground level, the maximum depths of borings.

Geological and hydrogeological investigations at the site were utilized to develop a conceptual model for the transport and flow direction of LNAPLs. The main hydrogeological units in the study area include the composite aquifer formed by the Wadi Sir Limestone (A7) and the Amman Silicified Limestone (B2) units, both of which are highly productive with an average thickness of 130 m within the thermal station. The Um Ghudran chalky Limestone (B1) unit, situated between the B2 and A7 aquifers, acts as a semi-aquifer and hydraulically connects the B2 aquifer above with the A7 aquifer below, creating the B2/A7 composite aquifer system. The base of this productive aquifer system is found at a depth of 80 m in the northwestern section of the area, gradually deepening to a maximum of 240 m in the southern part (BGR & WAJ 1997). Within the project site, the base of this unit ranges from 100 to 160 m, resulting in a maximum well depth of 160 m in the southern section of the plant, where the proposed power generation project is located.

In the study area, a few wells have been drilled beyond 200 m in depth to access a second productive aquifer, known as the Hummar Formation (A4) in the middle Ajlun group. The B2/A7 aquifer and the A4 aquifer are separated by the A5/A6 Aquitard unit, which has a thickness ranging from 100 to 108 m. Due to its low permeability, this aquitard prevents any hydraulic connection between the B2/A7 and A4 aquifer systems.

According to records from the Ministry of Water and Irrigation (MWI), groundwater flow maps indicate that the aquifers in the Hashmiyeh area are being extracted at a rate higher than their natural recharge, leading to the formation of a sink that draws groundwater from surrounding areas to compensate for the deficit. Additionally, the discharge from the Samra wastewater treatment plant has created a recharge mound, which infiltrates the upper aquifer system, resulting in water table levels as high as 550 m above sea level (asl).

Table 2 provides details of all existing wells near the HTPS. The inventory of deep wells indicates a significant decline in discharge flow rates, dropping from 531 m³/day to approximately 235 m³/day. Similarly, the total annual water discharge decreased dramatically from 1,162,334 m³ in 1994 to 275,853 m³ in 2010, reflecting a 76% decline over this period. Additionally, groundwater salinity showed a sharp rise during the same timeframe, with total dissolved solid (TDS) levels increasing from below 1,400 to over 2,600 mg/L. This significant increase in salinity is attributed to the over-extraction of the upper carbonate aquifer in the study area.

Due to the presence of a groundwater table at a depth of fewer than 60 m under the site, the estimated time for pollutant spills to reach groundwater is 24.8 days since the average permeability is 2.42 m/days (2.8 × 10⁻⁵ m/s). This means that the potential for shallow groundwater contamination is very high if leakage of oils or untreated wastewater takes place.

The inventory well data were obtained from the MWI such as design flow, well yield at installation, recent pump depth, current well yield, pumping water level, static water level, and salinity. These data were used to estimate groundwater transmissivity and hydraulic conductivity using Driscoll (2007) and Bradbury & Rothschild (1985) methods as presented in Table 3.

Based on Equation (4), the average linear velocity was estimated based on the available data between the DW8 and DW9 wells. Since the distance between wells is estimated to be 270 m, the head loss (dh/dl) is estimated to be (1.04 m), the average hydraulic conductivity is estimated to be 4.0 m/day, and the porosity of the aquifer as estimated from the resistivity survey is 20%, then the average linear velocity is estimated to be 0.077 m/day (8.9 × 10⁻⁷ m/s).

The expected water consumption of the new ACWA Power Project will be 160,000 m³ per annum. The fuel used for the operation of the project will be natural gas, with light distillate oil (LDO) available as a backup. The project design includes three new boreholes to be drilled in locations adjacent to and at the same depth as three of the previous HTPS wells on the site (95, 110, and 220 m depth). The expected water consumption of the proposed CCGT project will be 160,000 m³ per annum, which is therefore expected to significantly reduce water consumption from that of the HTPS, which had operated from the same aquifer with water consumption of approximately 430,600 m³ per annum.

Two deep wells indicated signs of contamination with HFO. Deep well No. 8 (AL-1552), located in the eastern inside of the HTPS border, was drilled in 1983 (Figure 3). The well yield decreased from 85 m³/h in 1983 to 63 m³/h until it went out of service in 2007 due to its pollution with wastewater first (2006) and then with HFO in 2008. The TDS of the water increased from 580 to 2,100 mg/L in 2010. The current static water is 59 m, whereas the pumping water level is 54 m, which means that the drawdown of the well reached 5 m.

Similarly, deep well No. 9 (AL1553), located on the eastern side inside of the HTPS border, was drilled in 1983. The well yield at installation decreased from 85 m³/h in 1983 to 45 m³/h till it went out of service in 1998 due to its pollution with HFO. The static water is 59 m, whereas the pumping water level is 57 m (i.e., the drawdown of the well is 2 m). The TDS of the well at the beginning was about 450 ppm but gradually increased to reach approximately 2,200 mg/L.

To understand the behavior of LNAPL petroleum hydrocarbons in the aquifer system, it is required to develop a hydrogeological CSM) for the study area. Based on onsite geological and hydrogeological investigations, the upper aquifer in the study area exhibits the solution channels due to karstic features; therefore, the pathway for the HFO (LNAPL) can emerge through jointing and fracturing of Amman formation, which is the target aquifer. This aquifer consists mainly of chert limestone, which is highly fractured, therefore the hydraulic conductivity and the transmissivity values of the wells DW8 and DW9 are considered to be high. Moreover, the LNAPL (HFO) with a density less than water may enter karst aquifers through soil percolation and/or directly through preferential pathways such as open sinks, or by a combination of both routes.

According to the results from the time-domain electromagnetic inversion sections, the aquifer displays three distinct ranges of resistivity values, as follows: low (3–15 Ohm·m), located at the lower boundaries of the known fractured aquifer; moderate (30–50 Ohm m); and high (100–1,000 Ohm·m) (Abu Rajab et al. 2018). The very low resistivity values could be attributed to the lowest resistivity measured from groundwater samples, which is around 30 Ohm-m. Conversely, the high resistivity values indicate the presence of cavern structures (Abu Rajab et al. 2018).

The electrical resistivity tomography profiles similarly support the presence of cavern structures, identifying shallow caverns at approximately 10 m depth with resistivity of about 100 Ohm-m at BH-9, and deeper caverns around 30 m depth with resistivity of 200 Ohm-m at BH-8 (Abu Rajab et al. 2018). The cavern structures are distinctly observed in the area between DW-9 and DW-8, where resistivity ranges from 100 to 10,000 Ohm-m (Abu Rajab et al. 2018). This notable resistivity contrast in the cavern structures suggests the influence of other components beyond the lithological units, as the geology between DW-9 and DW-8 is generally uniform except for variations in the thickness of the alluvium layer (Abu Rajab et al. 2018).

After identifying potential exposure pathways at the site, various remedial options were evaluated and screened. The selected approach involved pumping and treating the groundwater using the on-site oil/water separator. This method quickly removed visible contamination from the groundwater. Pumping continued for three months, operating for 2 h each day with a pump capacity of approximately 90 m³/h, resulting in a targeted removal of about 180 m³/day from the contaminated wells. Daily chemical analyses were conducted at the on-site laboratory, and all results consistently showed hydrocarbon concentrations below detectable limits.

Additionally, groundwater samples from deep wells #8 and #9 were tested at a local accredited laboratory in Jordan for total petroleum hydrocarbons (TPH), oils and grease, and polyaromatic hydrocarbons. All results returned concentrations below detectable limits, confirming the successful remediation of the groundwater contamination. Following the remediation, the existing wells were capped and are no longer in use. However, three new boreholes will be drilled at a proposed site adjacent to the former HTPS wells, accessing the same shallow aquifer at depths of 95, 110, and 220 m. In the worst-case scenario during winter, when the plant operates on LDO fuel, a volume of 117.2 m³/h. will be required.

Records from the MWI indicate that the average depth of groundwater wells in the area ranges from 86 to 280 m, with static water levels between 480 and 490 m. Over the past 40 years of HTPS operation, baseline groundwater quality from this aquifer has been considered suitable, although total dissolved solids (TDS) levels have increased over time.

The presence of hydrocarbons in aquifers is a particular issue because their degradation and natural attenuation change the redox chemistry of the aquifer. These changes yield other changes in the overall geochemistry, which change the conditions for degradation, which in turn change geochemistry. The objective of this model is to evaluate the use of the PHREEQC computer code for the modeling of the degradation of carbon species in groundwater environments and its effects on groundwater chemistry.

PHREEQC is a program for thermodynamic-based equilibrium calculations of hydrogeochemistry. However, the degradation of organic matter and the associated reduction reactions to the oxidation of organic matter can be modeled with this code. The thermodynamic-based calculation of the degradation of organic matter is used to model bacterial degradation of a contaminant. The redox chemistry of natural water depends strongly on the abundance of oxygen and nitrate and, as a matter of course, on the abundance of organic matter. The degradation of hydrocarbons (i.e., HFO) leaking from the drains was modeled using PHREEQC.

The keyword REACTION in PHREEQC enables to add or remove elements or compounds to the defined solution. Representing organic matter, the 0.07 mole of CH₂O(NH₃) and 0.02 mole of CH₄ (g) is added incrementally in log_n (n = 1–3) steps, at different temperatures of 20, 25, and 30 °C, respectively. The water sample is chosen to determine how the chemical composition infers redox changes due to hydrocarbon degradation.

For this purpose, the values for the hydrochemistry of the initial solution were taken from the deep well DW9, which represents the polluted groundwater with the HFO. The redox potential of the water samples collected from DW9 was measured on 25 April 2015, the Eh ranges from 165 to 201.6 mV with an average value of 187.1 mV indicating reduced oxygen conditions. The p_e (electron activity) can be estimated based on the E_h value ⁠, then p_e is equal to approximately 11.07 mV. A low electron density (E_h greater than 50 mV) indicates oxidizing, aerobic conditions; a high electron density (E_h less than 50 mV) indicates reducing, anaerobic conditions. High positive E_h values (400–800 mV) indicate well-aerated conditions that are optimal for aerobic biodegradation. Values for E_h of 100–400 mV generally indicate reduced oxygen conditions but are acceptable for aerobic biodegradation.

Through the running process using the PHREEQC program for a couple of runs for organic matter, the final saturation indices’ result of the geochemical model indicates that the groundwater from the deep aquifer is capable of dissolving minerals as well as CH₄(g), H₂S (g), NH₃(g), and sulfur at different temperatures (Appendix B). This geochemical model confirms that the biodegradation of the organic matter represented by the HFO can be biodegraded and promote the dissolution of minerals from the aquifer and hence increase the salinity of groundwater by increasing the TDS of the groundwater. This explains both increases in sulfate and other ions in the HTPS groundwater along with the change in E_h values from oxidizing to reducing environment. The solubility of gases depends on the pressure, temperature, and salinity of groundwater. The solubility of gases increases with increasing hydrostatic pressure (i.e., depth) but decreases with increasing temperature and salinity.

This study successfully identified and remediated groundwater contamination by HFO at the Al HTPS in northern Jordan. Comprehensive hydrogeological and geophysical investigations pinpointed a leaking oily water drain from the boilers as the primary contamination source. The LNAPLs were found to migrate through sediment layers and into the karstified limestone aquifer, which is highly permeable due to its karstic features. This contamination was confined to a specific area beneath the HTPS at a depth of approximately 50 m, covering an area of around 0.5 km². Additional boreholes confirmed that contamination did not extend beyond the site boundaries. The hydrogeological conceptual model developed indicated that the LNAPL plume traveled primarily through karstic pathways and was driven by the groundwater hydraulic gradient toward Wells DW9 and DW8. The model demonstrated that the hydraulic gradient varied between 0.4 and 3%, which accelerated the plume migration within the fractured limestone. The remediation plan not only addressed immediate contamination but also focused on preventing future leaks, including plans to cap the existing wells and drill three new boreholes to maintain water supply while mitigating further environmental risks.

The pump-and-treat remediation approach effectively reduced TPH concentrations from over 10 mg/L to below detectable limits (<0.5 mg/L) within a 3-month period, removing approximately 16,200 m³ of contaminated water at a rate of 90 m³/h, twice daily. PHREEQC modeling showed that HFO biodegradation led to significant shifts in groundwater chemistry, with TDS levels increasing from 1,400 mg/L to over 2,600 mg/L and elevated sulfate concentrations indicating increased mineral dissolution due to redox shifts.

The study utilized PHREEQC, a hydrogeochemical modeling tool, to evaluate the degradation of HFO in groundwater at the HTPS site. The degradation process, modeled by adding organic matter (CH₂O(NH₃)) and methane (CH₄) at varying temperatures demonstrated significant changes in the redox chemistry of the aquifer, shifting conditions from oxidizing to reducing. These changes affect the overall geochemistry, leading to increased dissolution of minerals and elevated levels of TDS and sulfates in the groundwater. The results confirm that biodegradation of HFO promotes mineral dissolution and increases groundwater salinity, driven by factors such as pressure, temperature, and salinity. This modeling approach underscores the complex interplay between hydrocarbon degradation and aquifer geochemistry, highlighting the importance of understanding redox conditions to optimize remediation strategies for contaminated groundwater systems.

The study's novelty lies in the combined use of field data and PHREEQC modeling to understand the complex hydrogeochemical interactions in a karst aquifer contaminated by HFO. This approach demonstrated how biodegradation of hydrocarbons can alter aquifer chemistry, which is critical for designing effective remediation strategies in similar settings. The findings provide valuable insights into HFO behavior and contamination pathways in karst systems, supporting the development of more resilient groundwater management and remediation strategies for HFO-contaminated sites.

This research contributes to advancing hydrogeological understanding and offers practical methods for managing LNAPL contamination in karst aquifers, with implications for broader applications in regions facing similar hydrogeological challenges.

This paper is not supported by any type of funding

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.