Effects of Prosopis velutina Invasion on Soil Characteristics along the Riverine System of the Molopo River in North-West Province, South Africa

Invasive alien plants are one of the major threats to ecosystems. Many invasive plant species, such as Prosopis species, have been introduced around the world and can alter the soil properties of invaded ecosystems. It is one of the most aggressive invasive plant invaders in the North-West Province of South Africa, but little information is available about their influence on soil properties. -is study was conducted to investigate the effect of Prosopis velutina invasion on selected soil properties at five different sites along the riverine system of the Molopo River in North West Province. At each study site, soil characteristics were measured from soil samples taken under P. velutina canopies, between canopies and in the benchmark stands free of Prosopis species. -e effect on selected soil properties of P. velutina invasion varied between the three stands and between sites. In all the sites, almost all soils collected from under the canopies had a significantly higher soil exchangeable Ca, K, Mg, and Na, organic matter (OM), total nitrogen (TN), available phosphorus (P), Electrical conductivity (EC), and cation exchange capacity (CEC) than the other sample positions, except for the pH which had the high value in intercanopies. Significantly higher (p< 0.05) values of almost all soil properties were found on the densely invaded sites (Tshidilamolomo I and Tshidilamolomo II) compared to lightly invaded sites (Mabule, Black Heat Farm, and Bray). However, it was difficult to generalise as the effects often appear to be site-specific. In addition, the findings also indicated that soils textural classes ranged between sand, silt, and clay in all study sites with a higher proportion of sand in the benchmark than in the soil under the canopies and intercanopies. Soil characteristics differed significantly more between sites than among positions.-e site effects observed in this current study provide evidences that this species may occupy a relatively broad soil niche.


Introduction
Alien plants have consistently caused environmental changes worldwide and represent a frequent and widespread threat to natural ecosystems [1,2] and to biodiversity conservation [3,4]; these impacts are very often irreversible. Alien plant species can influence the plant-soil relationship in the invaded habitats [5,6] modifying both biotic and abiotic elements of soil and diversity [7]. e spread of alien plant species in South Africa is an increasing concern because it leads to decreases in both surface water runoff and groundwater recharge, causing direct habitat destruction, intensifying flooding, and increasing the risk and intensity of wildfires [8]. Among the widespread alien plants in South Africa, Prosopis species cover very large areas of arid and semiarid parts of the country, causing severe threats to native plant communities in the Northern Cape and Western Cape provinces and in the North West Province as well [9,10]. Moreover, these species are causing severe damage to the natural environment and habitats, threatening many plant species. It has been shown that in some cases, invasive plant species are capable of altering the characteristics of invaded soil and native species diversity to facilitate further invasion [11,12]. A number of studies reported that Prosopis species can significantly alter soil properties, which may favour their own development and proliferation [13,14].
Prosopis species, introduced to South Africa as agro forestry trees to provide wood, fodder, and shade, charcoal in the late 1800s [9,15], have since been identified as a serious problem in parts of South Africa and become the second most widespread invasive alien plant after Australian acacias [16]. ey are known to impact negatively on the conservation of biodiversity as well as the livelihoods of rural people that depend heavily on natural resources [17], such as overrunning grazing land, consuming excessive quantities of ground water, and disrupting ecosystem services [10,18,19].
Much work has been undertaken on the influence of invasive plants on soil properties [12,14,20]. Recently, many studies have increased interest in identifying appropriate indicators for describing soil characteristic changes [12,21,22]. Although much is well documented on the ecological impacts of Prosopis species, research exploring the effects of Prosopis velutina invasion on soil properties has received little attention and more particularly in the North-West Province of South Africa.
e paucity of such important studies could be due to the high costs of conducting soil analyses, particularly of multiple soil samples collected across wider geographic areas over time. e documented impacts on soil properties are diverse. According to [20], Prosopis species may have different influences on soil properties, depending on local conditions. erefore, this study was undertaken with the aim to investigate and compare the impact of Prosopis on selected soil properties at five sites with different Prosopis velutina density along the riverine system of the Molopo River in North-West Province of South Africa.
Soil properties are the essential factors that influence plant species distribution [23]. A better understanding of soil properties in different communities is crucial for clarifying the influence of P. velutina on soil characteristics. is study was therefore conducted to analyze and increase the understanding on the variability of soil properties as influenced by P. velutina. For this reason, two major objectives were addressed: (i) to compare the selected soil properties from under Prosopis canopy to soil from between canopies and adjacent benchmark stands and (ii) to examine the alterations in soil properties under P. velutina invasion across the five study sites.

Description of the Study Sites.
e study was conducted in five sites in Tshidilamolomo I, Tshidilamolomo II, Mabule, Black Heat Farm, and Bray along the riverine system of the Molopo River in North-West Province of South Africa (Figure 1). Each site is more than 1500 m 2 in extent and was selected for the comparison of soil characteristics. e five study sites were selected using stratified purposive sampling [24]. ese sites were paired to include one stand with Prosopis velutina and another without. In the present study, the uninvaded stands called benchmark stands were selected to be in close vicinity of the invaded stands. In addition, the criteria used to select the five sites were also based on the availability and abundance of P. velutina in invaded stands and the homogeneity of sites in terms of soil type under both the benchmark stands and invasive Prosopis stands. To clarify the difference in the soil characteristics between invaded and benchmark stands, the density of dominant vegetation in the five sampling sites was recorded (Table 1). e riverine system of the Molopo River in North-West Province of South Africa has a semiarid to arid climate and situated between the latitudes 23°20′ S and 28°30′ S and longitudes 20°15′ E and 26°10′ E [25], with an average annual precipitation of about 550 mm [26]. In the North-West Province, the daily temperature is about 42°C during summer months and often −9°C during winter months [26]. According to [27], soil in this province is sandy loam and is classified as a Hutton form [28] or Rhodic Ferralsol [29] or Chromic Luvisol [30].

Soil Sampling.
At each sampling site, soil from the upper 15 cm in three replicates was collected from the under canopy (UC) and the intercanopy (IC) of five selected P. velutina in invaded patches located in the centre of the sites, mostly where the species were more dense; while in adjacent benchmark (BM) stands, soil from the upper 15 cm was collected in three selected 1 m × 1 m plots. All soil samples from UC of the five selected P. velutina in each site were pooled homogeneously into a single bulk sample as well as all soil samples from IC and BM.
e effect of pooling samples on the efficiency of experiments was meant to detect the difference in soil properties between UC, IC, and BM in the same site and between sites. e pooling of samples is that it also reduces financial cost of analysis. ese soil samples from UC, IC, and BM were air-dried separately at room temperature, passed through a 2 mm sieve to remove large particles and for homogenization before analyzing for various contents. For analysis, about 2 kg of soil samples from the same position and from each site were carried in brown paper bags to the laboratory at the Agricultural Research Council (ARC) in Pretoria (South Africa).

Soil Analyses.
A total of 13 soil variables were estimated on each sample. e soil pH was determined using a glass electrode in a 1 : 2.5 soil-water suspension following equilibration for 16 h [31]. Soil organic matter (OM) content was determined using the WalkleyBlack method [32]. Total nitrogen (TN) was determined by the micro-Kjeldhal technique [33]. Exchangeable cations were extracted with 1 N ammonium acetate, and the concentration of potassium (K) was determined using a flame photometer and that of calcium (Ca) and magnesium (Mg) was determined with an atomic absorption spectrophotometer, while atomic emission spectrophotometry was used for sodium (Na). Available phosphorus (P) was measured using the Bray analysis method [34]. Soil texture (Sand %, silt % and clay %) was determined using hydrometer method [34]. Cation exchange capacity (CEC) was estimated by ammonium acetate extraction at pH 7 [35], and electrical conductivity (EC) was performed on a saturated extract of the soil [36].

Statistical Analysis.
One way analysis of variance (ANOVA) was used to test and compare the differences among different soil parameters in different positions and at different sites. Statistical significance was determined at p < 0.05, and differences among soil properties were separated using Tukey's HSD (honestly significant differences) post hoc test. All statistical methods were computed using Excel 2013.

Influence of Prosopis velutina on selected soil properties in
three stands at five study sites. All selected soil properties were compared between soils under the canopies (UC) of P. velutina, with soil in intercanopies (IC) and soil in an adjacent benchmark (BM) in order to assess the difference between the five sites. Almost all soil properties measured were statistically different (p < 0.05) at different positions or stands ( Table 2). While significant differences were found in selected soil properties in all the five sampling sites (Table 3). Soil characteristic differed significantly more between sites than among positions (

e concentration of exchangeable cations (calcium, potassium, magnesium and sodium).
Overall, the concentration of exchangeable cations (Ca, K, Mg, and Na) was higher under canopies (UC) of Prosopis velutina than in the intercanopies (IC) and the benchmark (BM) ( Table 2). e concentrations of exchangeable Ca, Mg, K, and Na in almost all the study sites showed a decreasing trend with increasing distance from tree canopies to nearby benchmark stand. In contrast, there was no regular pattern of exchangeable Ca, Mg, K, and Na with distance between IC and BM ( Table 2). All concentrations of exchangeable Ca, K, Mg, and Na recorded significant differences (p < 0.05) between UC, IC, and BM in all sites, except for exchangeable Ca and Na in Bray and Black Heat Farm where the p value was (p � 0.44) and (p � 0.37), respectively (Table 2). e mean concentration of exchangeable Ca, Mg, K, and Na tended to be in the order of Ca > Mg > K > Na in all the five sites (Table 3). Among all sites, the greater concentration of exchangeable Ca, Mg, and Na was observed in Tshidilamolomo II, while the higher exchangeable K was recorded in Bray site. e range of exchangeable Ca, Mg, K, and Na was between 1.76 and 10.53 mg kg −1 , 0.77 and 3.35 mg kg −1 , 0, 22 and 0.68 mg kg −1 , and 0.02 and 0.05 mg kg −1 , respectively. e lowest concentration of exchangeable Ca, Mg, K, and Na was recorded in Black Heat Farm. Most particularly, the analysis of Variance (ANOVA) showed significant differences in soil exchangeable cations of Ca, Mg, K, and Na at all the five sites (Table 3).

Soil Organic Matter (OM).
e soil organic matter content was significantly different (p < 0.05) between soils under canopy, intercanopies, and benchmark stands in all the five sites (Table 1). In all the five sites, soil OM declined with distance from under canopy (UC) to intercanopy (IC) and then increased from IC to the benchmark (BM) ( Table 2). e maximum range of OM was recorded in Tshidilamolomo II which was from 0.36 to 7.29%, and the minimum range was recorded in Black Heat Farm between 0.2 and 1.5%. In addition, the overall mean value of OM proportion among the five different study sites varied between 0.30 and 3.82%. Among the study sites, the highest proportion of OM was recorded in Tshidilamolomo II, whereas the lowest percentage was observed in Black Heat Farm (Table 3). However, a one-way ANOVA test showed that OM was statistically significantly different (p < 0.05) between sites (Table 2).

Total Nitrogen (TN).
Like soil organic matter, total nitrogen (TN) tended to be higher under Prosopis velutina canopies in all study sites except in Black Heat Farm where it was higher in the benchmark stand. e highest TN content in all sites was recorded under the canopies, which ranged from 0.02 to 0.11% followed by the benchmark, from 0.03 to 0.08% and then by intercanopies from 0.02 to 0.04% (Table 2). However, significant differences (p < 0.05) existed in TN percentage between the UC, IC, and BM. e overall mean value of TN percentages in all different sites varied between 0.02 and 0.07%. Among the five sites, the minimum mean values of TN percent were recorded in Black Heat Farm (0.02%), and the maximum mean value (0.07%) was recorded in Tshidilamolomo I and Bray. However, the difference among the different sites was significant (p < 0.05) ( Table 3).

Soil pH.
e maximum range of pH in all the study sites was recorded under the intercanopy (IC) which was between 7.4 to 8.3, and the minimum was recorded in the benchmark (BM) ranging from 6.5 to 8.3. However, no significant differences (p < 0.05) in pH were observed between under canopy UC, IC, and BM in Tshidilamolomo II, whereas there was a significant difference (p < 0.05) between UC, IC, and BM Tshidilamolomo I, Mabule, Black Heat Farm, and Bray (Table 2). e soil pH under the three positions (UC, IC, and    (Table 3). However, the difference in pH was significant (p < 0.05) between the different sites ( Table 3).

Available Phosphorus (P).
e analysis of variance (ANOVA) for available phosphorus revealed that there were significant differences (p < 0.05) among the three positions under canopy (UC), intercanopy (IC), and the benchmark (BM) across the five study sites (Table 1). e maximum range of P concentration was recorded under the BM, which ranged from 7.31 to 79.61 mg kg −1 followed by the IC from 7.17 to 74.24 mg kg −1 and then followed by UC from 7.84 to 57.40 mg kg −1 ( Table 2). It is clear that concentrations of soil available phosphorus increase with increasing distance from under Prosopis canopies to open areas benchmark. e mean available phosphorus concentration under different sites was observed ranging from 9.21 mg kg −1 in Tshidilamolomo I to 65.82 mg kg −1 in Bray (Table 3). However, significant differences (p < 0.05) regarding the concentration of available phosphorus were recorded between the five study sites.

Electrical Conductivity (EC).
e value of soil electrical conductivity in Tshidilamolomo I, Tshidilamolomo II, Mabule, and Bray ranges from 0.23 to 1.09 mS m −1 ; 0.54 to 1.26 mS m −1 ; 0.29 to 1.12 mS m − ;1 and 0.89 to 2.01 mS m −1 , respectively. e results showed that these values decrease from under canopy (UC) to intercanopy (IC) and then to benchmark (BM). Soil EC was higher in UC followed by IC and BM. However, significant differences (p < 0.05) in EC were found between UC, IC, and BM, but no differences (p > 0.05) in EC were recorded between IC and BM in these sites except in Black Heat Farm (Table 2). In Black Heat Farm, EC ranged from 0.22 to 1.40 mS m −1 and tended to increase from UC to BM (Table 2). e mean values of soil EC distribution under different sites varied and were recorded in the range of 0.232.04 mS m −1 . Among all the sites, the minimum and maximum mean values of soil EC were observed in Black Heat Farm and Bray, respectively (Table 3). ere was a significant (p < 0.05) difference in EC between all sites.

Soil Cation Exchange Capacity (CEC).
Soil cation exchange capacity (CEC) was significantly higher under the canopies (UC) of Prosopis compared to soil CEC in intercanopies (IC) and benchmark (BM). It ranged from 3.16 to 13.18 cmolc kg −1 ; 3 to 7.5 cmolc kg −1 ; and 3.76 to 4.68 cmolc kg −1 in Tshidilamolomo I, Mabule, and Black Heat Farm, respectively. Whereas, in Tshidilamolomo II and Bray, the higher value of CEC was recorded in the BM and IC (Table 2). However, across all sites, the CEC values were significantly (p < 0.05) affected by different positions or distance (UC, IC, and BM). e mean values of CEC distribution under different sites varied from 4.50 to 10.59 cmolc kg −1 . Among all sites, the minimum mean value of soil CEC was recorded in Black Heat Farm, while the maximum concentration was observed in Tshidilamolomo I (Table 3). e analysis of variance (ANOVA) for CEC revealed a significant variation (p < 0.05) between the sites.

Soil Texture (Sand, Silt, and Clay).
e results of this study showed that sand had the dominant proportion in soils in all the sites ( Table 2). Analysis of variance for soil sand content revealed that it was significantly affected by the position, i.e., by under canopy (UC), intercanopy (IC), and benchmark (BM) (p < 0.05) in all sites except in Black Heat Farm where the soil sand content was not affected by the position (p � 0.07). However, maximum sand content was recorded in IC between the range of 86 and 90% followed by UC from 75 to 90% and by BM from 67 to 90%. e values for clay content were higher UC and tended to decline from UC to IC and to the BM (Table 2). Across sites, significant differences (p < 0.05) in clay percentage were recorded  1). e results also revealed that the proportion of silt was slightly higher UC than that of IC and BM. Silt soil texture decreases with distance from the trunk of Prosopis trees towards the open area benchmark, but its proportions did not significantly vary (p > 0.05) between the IC and the BM in Tshidilamolomo I, Mabule, Black Heat Farm, and Bray, while no significant difference was recorded between UC and IC in Tshidilamolomo I site (Table 2). However, the textural analysis from different sites was classified as followed sand >clay >silt (Table 3). Significant difference (p < 0.05) between the mean values of sand clay and silt were recorded between the sites.

Discussion
In this study conducted along the riverine system of the Molopo River in North-West Province of South Africa, the selected soil properties from each site were compared between under canopy (UC), intercanopy (IC), and benchmark (BM) in order to assess the difference in soil characteristic among the five sites. e results of this study revealed a strong effect of Prosopis velutina on the properties of the soil. ese outcomes are generally consistent with other studies on invasive species, which indicate significant effects of invasive species on different soil properties [13,37,38]. In this research, soil UC of P. velutina consistently had higher value of soil properties compared to soil in IC and BM stands (Table 2), which is in agreement with many earlier report where the soil characteristics decline from the base of the tree to adjacent open areas [39,40]. Table 3 shows that the site effect was highly significant (p < 0.05) for all the soil parameters analysed, indicating that P. velutina can adapt to and thrive in sites with a significant variation in soil characteristics. is is in line with earlier studies reported by [38]. Furthermore, the results of this study reveal that P. velutina has a significant effect on all the soil properties, indicating that the influence on soil characteristic varied widely depending on the existing site conditions. Soil properties differed significantly across sites and this could reveal that the five study sites covered a wide range of different soil conditions. e present study showed varied effects of P. velutina on soil exchangeable Ca, Mg, K, and Na with significant differences between UC, IC, and BM stands in all the five sites. Most exchangeable Ca, Mg, K, and Na were higher under P. velutina than in IC and BM. Similar to this result, in [41,42], significant decreases in exchangeable Ca, Mg, K, and Na in soils from under canopies to open areas of P. juliflora and P. laevigata have been reported. Likewise, in [43], significant differences in exchangeable Ca, Mg, K, and Na between soil under and outside Celastrus orbiculatus canopies have been reported. e high value of exchangeable cation under the canopy of P. velutina could be ascribed to the high accumulation of litter under the tree canopies as the cations are released when the accumulated litters from the canopies of the trees undergo microbial decomposition followed by mineralization. Unlike to this study, in [44], no significant differences are shown in the levels of Na and Ca between soil within the canopy P. juliflora and the open areas in the deserted rangelands of Bahrain. e relative higher accumulation of soil organic matter (OM) under P. velutina canopies than in inter canopy and benchmark in all the sampling sites indicated the soil enriching capacity under this tree. As reported by [45], mesquite (P. juliflora) trees enrich the soil under their canopies at the expense of the soil nutrient capital in the open areas. e higher percentage of OM under tree canopy could be attributed to the leaf litter deposition, higher organic matter production by trees and its slower rate of mineralization, root turnover, and animal defecation as they stay under shade for browsing. is result is in line with [13], who also reported that OM was significantly higher under than outside P. flexuosa canopies. Similarly, many authors, such as [46,47], have also reported gradual decreases in soil organic matter with increasing distance from tree base. e lower soil OM content in adjacent open areas could be attributed to the fact that the main source of organic matter there is grass. Similarly to this study, in [48], higher OM under canopies of Prosopis juliflora than outside canopies is reported. Contrary to the findings of this study, in [49], higher soil organic matter stocks in the perennial grasses than in the shrubs are observed; this might be attributed to both greater inputs of poor quality litter that is relatively resistant to decay and the lower ability of microorganism to decompose these organic matter.
Soil total nitrogen (TN) was significantly higher under P. velutina canopies than the outside of the canopy in all the five sites. e higher percentage of total nitrogen under this species may be attributed to the accumulation of leaf litter within the crown limits and also to the N-fixing capabilities of the Prosopis species. However, the results of this study were in accordance with those of [13] who also observed higher value of TN in under than outside P. flexuosa canopies. is is consistent with the previous studies that have reported increased fertility under the canopy of P. glandulosa [50] and P. cineraria [51]. Furthermore, for other trees species, [52] found higher value of TN in soil under the canopies of Hieraciurn species than in stands outside. Similar trends were also reported for Millettia ferruginea and Cordia africana on the topsoil properties in agroforestry practices in Sidama, Southern Ethiopia [53].
In this study, there was a variation (p < 0.05) in soil pH between soil under canopies, between canopies, and the benchmark stands. e measured soil pH was alkaline resulting in soil pH above 7.0. e higher value of pH was recorded in intercanopy than that under the canopy. e lower pH value under the canopy could be attributed to accumulation of organic matter under the trees through litter fall and root decay [54]. According to [55,56], the high soil pH may be attributed to the invasive plants having high nitrate uptake rates and may play also a critical role in the regulation of nutrient cycles [57]. In line with this finding, several studies have also found significant differences in pH when comparing soil from under canopies and open areas of invasive plants. Kahi et al. [37] found significant differences in pH between the soils within and outside the canopies of P. juliflora and Acacia tortilis. However, this work was International Journal of Ecology inconsistent with the findings of [58] who found pH values not statistically different when comparing the soil samples in the areas with and without Prosopis. Witkowski 59 also found no significant differences in pH when comparing soil from under Acacia cyclops canopies with soils from outside their canopies. However, it is clear that both high and low pH have also been reported following plant invasions [60]. According to [56], the variation of soil pH is also dependent on the degree of invasion. e results of this study showed significant differences in available phosphorus (P) between UC, IC, and BM and among the different sites. Furthermore, no clear trend in available phosphorus was detected by [61] [13,37], and Cordia africana [62]. In contrast to these findings, [63] found an increase in p value as the distance increased from the canopies of Faidherbia albida to outside of the canopies. ese differences may be due to specific factors at the different sites.
e results pertaining to the high soil electrical conductivity (EC) content under Prosopis canopies in almost all the sites compared to intercanopies and benchmarks are in agreement with that of [12] who also observed increasing beneath rather than outside Prosopis juliflora canopies on the native flora and soils of the UAE. In a similar study, Mussa et al. [64] [66]. e value of CEC was lower in IC and BM stands as compared to UC tree.
Based on the result of the study, variations in soil texture were observed between the three positions of all the study sites. ese variations may be attributed to the increasing of biological activity which may have improved weathering process and favorable moisture provision under the tree canopy [67]. is study revealed that silt content was slightly higher under canopy than open area whereas the sand and clay content was higher in open area than under the canopy. is finding is in agreement with the study of [68] who reported the same trend in Faidherbia albida and Cordia africana studies in different sites. Furthermore, Sharma and Gupta [69] in India revealed increases in silt proportions under the tree canopies and decreases in sand proportion. e comparisons for the soil properties between sites showed significant (p < 0.05) differences. e results of this study are important because they show the variation in selected soil characteristics on five different sites invaded by P. velutina. e differences in the soil characteristics between the five sites could be the result of various factors such as anthropogenic activities, grazing which alters soil properties and to difference in other plant species occurring in the sites. Descriptive statistics showed that the variable of all soil properties in the densely invaded sites (Tshidilamolomo I and Tshidilamolomo I) was relatively higher, while the variation in lightly invaded sites (Mabule, Black Heat Farm, and Bray) was relatively limited, but with some exception in sand proportion, P and Na values. e observed variations soils characteristics are similar to studies on other invasive plant species that suggest that invasive species may modify soil properties [70,71]. e higher values of soil properties obtained in the densely invaded sites of P. velutina had also been reported in many other invasive species. For example, Mandal and Joshi [72] reported that soil TN, OM, P, and K levels increased with increase in Lantana density.

Conclusion
Based on the result of this study, variations in soil properties were observed not only between the three stands (under canopy, inter canopy, benchmark) but also between different sites invaded by Prosopis velutina. is study demonstrates that the three stands showed significant effects of P. velutina on soil properties, and most of the soil properties were higher under canopies but the lower values did not show consistent trends between inter canopies and the benchmark stands. However, the highest value of pH was observed in inter canopies. e results highlight the importance of sites as a most important source of variation in the impacts of P. velutina on soil properties. Due to variability of soil properties values among different sites, it was difficult to confirm the significant impact of the invader P. velutina on the selected soil properties. Despite a number of studies on Prosopis species, there is no consensus on the predictability of the impact of this species on soil characteristics. As P. velutina invasions are still increasing in different sites, more attention should be paid to studying this problem over a wider spectrum of invaded sites.
is study provides a valuable pattern for future studies in order to achieve successful control of P. velutina.

Data Availability
All data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.