Changes in Plants and Invertebrate Animals Communities along Gradients of Petroleum Products Pollution

In this work, we investigated the effects of petroleum products contamination, caused by petrol stations rejects, on the vegetation and soil invertebrate animal communities. Three sites near two petrol stations were chosen, and the vegetation was inventoried during each season of the year. The invertebrates were sampled in the third site using Pitfall traps. Results indicated an influence of the petrol stations rejects on soil properties (TPH, N, C, pH and moisture). It was the same for plant species and invertebrate communities that were influenced through the abundance, species richness, biological diversity (H’) and species evenness (J), especially in the most contaminated zones. Some weed plant species showed a tolerance to the contamination, as Xanthium strumarium, Avena fatua, Conium maculatum and Oxalis pes-caprae, and their establishment may have been facilitated by the disturbances created by the petrol stations rejects. Also, some invertebrate orders, as Isopoda and Panpulmonata were mainly present in the contaminated zones. High levels of petroleum products contamination and the changes they can induce in soil properties, can affect the vegetation and soil invertebrates, by influencing the abundance and species richness and by facilitating the establishment of some species, like weed species in the vegetation case.


Introduction
Petroleum products are mixtures composed mainly of hydrocarbons which can cause soil contamination (Henry, 1998;Pinedo, 2013). Some of them are persistent, whilst others are significantly reduced through time by means of many processes (Rhbal et al., 2010), including biodegradation, which depends on many factors (Solano-Serena et al., 2001). Hydrocarbons can cause toxic effects on plant species and soil living organisms when they reach a certain level of concentration (Tang et al., 2011). Hydrocarbon contamination can either positively or negatively affect different environment components, like soil properties (pH, total nitrogen and carbon for example) (Adam and Duncan, 2002;Akpan and Udoh, 2013), germination and growth of plant species (Luhach and Chaudhry, 2012;Fatokun et al., 2015) and macrofauna diversity and abundance (Erstfeld and Snow-Ashbrook, 1999). Transects have already been used to study the effects of pollution gradients on vegetation communities (Dazy et al., 2009) and macrofauna communities (Erstfeld and Snow-Ashbrook, 1999;Grzes´, 2009). Ecological analyses can be used to understand changes in vegetation communities (Dazy et al., 2009) and macrofauna communities (Grzes´, 2009) along pollution gradients in contaminated sites.
Petroleum products are an indispensable energy source for most countries in the actual context, and their use represents a source of many environmental contaminations and problems. Thus, the biotechnologies are proposed as alternative green solutions for the restoration of polluted sites (Kathl and Khan, 2011;Mandal et al., 2014). Vegetation sampling through pollution gradients can highlight some petroleum products tolerant plant species, which can be proposed for further research in order to determine their potential for the remediation of contaminated sites.
Petrol stations are potential sources of petroleum product soil contaminations. Even if they are subject to several environmental regulations, there are still significant soil contaminations in the environs of petrol stations, due mainly to fuel spills and fuel tanks leaks (Borowiec et al., 2008;Moschini et al., 2005;Parvizi Mosaed et al., 2015). We know that fuel spills from petrol stations can affect the surrounding environment and people's health (Alharbi et al., 2018;Pearson and Oudijk, 1993); consequently, an investigation about the effects of those petrol station wastes on their surrounding environment will provide some interesting information about the changes they may cause in the soil and the response of living organisms to this situation. In this study, we selected three sites near two petrol stations, and we investigated the effects of the petroleum product contamination on the vegetation and the invertebrate communities, and tried to identify some potential tolerant species to such contaminations.

Sites description
This work was performed near two petrol stations, in Freha (36°44,10'40"N, 4°18,24'83"E) and Idjeur (36°39'30.58"N, 4°27'54.51"E), two regions located in the middle-north of Algeria. The climate is Mediterranean continental with cold winters and hot dry summers. Altitudes in Freha and Idjeur are 124 m and 470 m, respectively. Three study sites were selected: the first site is located near the Idjeur petrol station and two other sites near the Freha petrol station. The study sites were subject to liquid discharges, contaminated with petroleum products emanating from the petrol stations. The wastes are mainly composed of fuels (petrol and diesel), waste oils, lubricating oils and water.

Sampling design
Based on the vegetation structure, it was possible to identify four vegetation gradients that were supposed to be related to the contamination. Then, two transects (T1 and T2) were established from the rejection point (low vegetation cover) to a distant area supposed to be unpolluted (important vegetation cover), and four zones were delimited in each transect. The first zone is located near the rejection point and the fourth zone is located in the unpolluted zone. This sampling design was used for soil, vegetation and invertebrates sampling (Figure 1 and 2).  synthesis, in the Faculty of Hydrocarbons and Chemistry (Boumerdes University), using the Mid-IR laser spectroscopy method. TPH analyses were only performed for site 3. Macherey Nagel qualitative oil test paper was used to determine the polluted and the unpolluted zones through the transects of sites 1 and 2.

Vegetation sampling
Four surveys were performed (in November 2017, January 2018, March 2018and May 2018. For each zone, we delimited a quadrat of 50 X 50 cm, in which we performed the vegetation inventory and the estimation of plant species abundance-dominance developed by Braun-Blanquet et al. (1932). Species richness (S) and abundance (the number of individuals, noted as N), in each zone were estimated. Shannon-Weaver's diversity (H'), Pielou's evenness (J) and Sorensen-Dice coefficient (DSC) were calculated according to Shannon and Weaver (1964), Pielou (1966), Sorensen (1948) and Dice (1945), respectively.

Invertebrates sampling
Soil invertebrates sampling was performed in the third site using pitfall trapping during ten days in August 2018. We used plastic pitfall traps of 9 cm in diameter and 17 cm in height. In each zone of the two transects, one pitfall trap was placed and was third party filled with water, saturated with salt, and a small amount of liquid detergent was added. Pitfall traps were checked every two days (five collections), and trapped individuals were placed in plastic bottles containing 70° alcohol. Invertebrates were identified until the order, and the approximate number of species was determined without the species being identified. Order abundance, number of orders and species richness (S) were estimated for each zone. Shannon-Weaver's diversity (H'), Pielou's evenness (J) and Sorensen-Dice coefficient (DSC) were calculated.

Data analyses
ANOVA followed by the Newman-Keuls test was performed to highlight the significant differences (p < 0.05). The Pearson linear correlation method was used to investigate the relationships between soil properties and ecological indexes. All tests were performed under Statistica 7.1 software (©StatSoft).
Soil sampling was performed in May 2018. For each zone, soil samples were collected at a depth of 20 cm, put in plastic bags and conserved at 4°C before being analysed. Soil physical-chemical analyses concerned the granulometric composition, pH, electrical conductivity (EC), total Kjeldahl nitrogen (N), carbon (C) and moisture. These analyses were performed at the National Institute of Forester Research (INRF) based in Algiers. Total petroleum hydrocarbons (TPH) analyses were performed at the laboratory of petrochemical

Results and Discussion
Soil analyses results are represented in Table 1. TPH analyses on site 3 showed a gradual decrease in the contamination through the different zones, from zones 1 (most polluted zones, with 336.35 mg/kg and 333.62 mg/kg) to zones 4 (no pollution detected). Qualitative tests indicated the presence of contaminated zones in sites 1 and 2, and carbon rates in these two sites suggest a clear differentiation between zones 1 and the other zones. For each site, the granulometric composition was fairly similar across the different zones. However, we observed some variations, in particular for sand rates on site 3. Carbon, TPH and moisture rates decreased progressively from zones 1 to zones 4 while total nitrogen showed an opposite tendency.
No particular tendency was found for the electrical conductivity. In site 3, the Pearson linear correlation showed a very strong positive correlation (p < 0.05) between TPH and C/N ratio ( Table 2). The hydrocarbon contamination caused an increase in carbon rates and moisture in all sites, and potentially a decrease in pH values for sites 1 and 2. Similarly, Akpan and Udoh (2013) reported an acidification of the soil and an increase in carbon rates after a contamination by diesel oil. The petrol stations wastes are composed partly of water, which explains the increase in moisture rates. The C/N ratio was altered by the contamination, resulting in an increase of C/N ratios in zones 1 and a decrease when moving to zones 4. It was reported by Adam and Duncan (2003) that changes in the C/N ratio may negatively affect the vegetation development. Parvizi Mosaed et al. (2015) have reported high levels of TPH in the soils of many petrol stations and have observed some effects on physical-chemical properties of the soils. The homogeneous groups (indicated in Table 1 by different letters) highlight the significant differences through the different zones. Zones 1 (most contaminated zones) are often represented in a separate group, in opposition to the other 3 zones, suggesting an important influence of the petrol stations wastes on soil properties in zones 1.
The total species abundances in each zone presented a general significant tendency (p < 0.05), showing a progressive increase from zones 1 to zones 4 ( Table   3). The unpolluted zones presented the biggest abundances. This tendency is not systematically observed when we look at the species richness, but we can see that zones 1 always present the lowest species richness (Table 3). These observations suggest a negative influence caused by the contamination and the alteration of soil properties. The homogeneous groups (indicated in Table 3 by different letters) highlight the significant differences through the different zones. Zones 1 are often represented in a separate group, in opposition to the other 3 zones, suggesting that the high contamination levels in zone 1 have an important influence on its vegetation.
Abundance-dominance of plant species is represented in Tables 4, 5 and 6. Three species present a good distribution through the different zones in all sites, establishing even in zones 1. Those species are Avena sp, Conium maculatum and Oxalis pes-caprae. However, there are some species that present a good distribution, but they are station specific, as Elymus repens in site 3, Vicia sativa, Cerinthe major and Senecio vulgaris in site 1, and Galactites tomentosa in site 2. Xanthium strumarium showed a preference for zones 1, especially in sites 2 and 3, presenting a tendency to dominate the zones 1; thus, it is rare in zones 3 and absent in zones 4. There are some plant species that avoided the polluted zones, as Beta vulgaris, Senecio vulgaris, Malva sylvestris, Sonchus oleraceus, Galactites tomentosa and Scolymus hispanicus.
The Sorensen-Dice coefficient showed that the highest similarities were between zones 2 and 3, and zones 3 and 4. The highest differences were between zones 1 and 4, and zones 2 and 4 ( Figure 3). Thus, there is a clear differentiation between the more polluted zones and the less polluted and unpolluted zones.
Pearson linear correlation (Table 2) revealed that the C/N ratio, moisture and TPH are the parameters influencing mainly the different ecological indexes. For site 1, pH is also an important parameter. The abundance and H' seem to be the most influenced indexes; thus, there are no significant correlations for S in site 2 and for J in site 3. Consequently, the vegetation structure was influenced by the contamination and the changes in soil properties.
Concerning soil invertebrates, we identified 10 invertebrate orders, spread over the four zones (Table 7). Hymenoptera was the most abundant order throughout the different zones, followed by the Coleoptera and Araneae orders. The response of invertebrate species to a soil contamination depends on many factors, like the specific behaviour of each species under the stress caused by the pollution and its ability to tolerate it; hence, there is a differentiation between the invertebrate communities living in polluted and unpolluted soils (van Straalen, 2004). Therefore, we can differentiate two lots of invertebrates. The first lot is composed of Isopoda, Panpulmonata, Orthoptera, Diptera and Mantodea orders, which are mostly present in the most polluted zones. Hymenoptera abundances did not show a clear tendency; we observed the lowest abundances in zones 1 and the highest in zones 2. It was observed that Isopoda and Hymenoptera can take advantage of high levels of contamination (Faulkner and Lochmiller, 2000), and according to Gospodarek et al. (2019)survival and body mass change of P. scaber reared in investigated soils were observed. The delivered food was not contaminated with PDSs. P. scaber reveals a considerable resistance in a short (4 weeks, Isopoda could have good capacities for the biotransformation of petroleum-derived substances and a low vulnerability to the penetration of these substances through the integuments. Panpulmonata are known for their ability of bioaccumulation and are used as bioindicators (Oehlmann and Schulte-Oehlmann, 2003). In their study, García-Segura et al. (2018) reported a decrease of Gastropods populations with an increase in TPH levels. The second lot is composed by Parasitiformes, Hemiptera, Araneae and Coleoptera orders, which showed a preference for the less polluted and unpolluted zones. Rusin and Gospodarek (2016) observed in their study that petroleum-based substances in most cases did not have a negative effect on the number of spiders; on the other hand, Gospodarek et al. (2019)survival and body mass change of P. scaber reared in investigated soils were observed. The delivered food was not contaminated with PDSs. P. scaber reveals a considerable resistance in a short (4 weeks observed a negative correlation between the Acari and some hydrocarbons.  Order abundance seems to be influenced by the soil contamination. There is a general tendency, showing a clear differentiation between zones 1 and the other zones. Zones 1 present the lowest abundance values ( Table 8). The number of orders were not negatively influenced by the contamination; thus, zones 4 presented the lowest values (Table 8). However, when it comes to the species richness (Figure 4), there is an obvious tendency demonstrating that zones 1 are the poorest zones. H' and J values present a general tendency, showing higher values for zones 3 and 4 compared with zones 1 and 2 (Table 8). Sorensen-Dice coefficient values for transect 1 did not show any tendency, but in transect 2, the lowest value was between zones 1 and zones 4 ( Figure 5). Therefore, there is a differentiation between the most polluted zones and the less polluted and unpolluted zones. Furthermore, there are some other parameters to consider, as the relationships between the vegetation structure and the invertebrates. It was shown that some invertebrate groups like Coleoptera and phytophagous insects have a close relationship with vegetation structure and plant species diversity (Chenchouni et al., 2015). The distribution of some groups, as Coleoptera and Hemiptera individuals could be partly explained by the vegetation cover, which is more important in zones 4, followed by zones 3, 2 and 1, respectively. Araneae distribution could also be related to the Hemiptera distribution, since they are predators of these insects. So, invertebrate communities can be affected directly by the contamination and indirectly by the changes in the vegetation structure produced by the contamination.

Conclusions
Petrol station wastes affected some soil properties as the TPH, N, C, pH and moisture. The vegetation and soil invertebrates were influenced as regards the abundance, the species richness, the biological diversity (H') and the species evenness (J) by the contamination and the changes in soil properties. Some weed plant species showed a tolerance to the contamination, as Xanthium strumarium, Avena sp, Conium maculatum and Oxalis pes-caprae, and some invertebrate orders, as Isopoda, Panpulmonata and Orthoptera were mostly present in the contaminated zones.