Salinization effects on coastal ecosystems: a terrestrial model ecosystem approach

Abstract

In coastal areas, intrusion/irrigation with seawater can threaten biodiversity along with crop yields, and the leaching of salts from areas affected by these processes can increase the salinity of water bodies nearby. The aims of this study were to evaluate the effects of salinization on coastal soil ecosystems due to saline intrusion/irrigation. Terrestrial model ecosystems were used to simulate two soil salinization scenarios: (i) seawater intrusion and irrigation with distilled water and (ii) seawater intrusion and irrigation with saline water. Three sampling periods were established: T0—after acclimation period; T1—salinization effects; and T2—populations' recovery. In each sampling period, the abundance of nematodes, enchytraeids, springtails, mites and earthworms, and plant biomass were measured. Immediate negative effects on enchytraeid abundance were detected, especially at the higher level of saltwater via intrusion+irrigation. Eight weeks after the cessation of saline irrigation, the abundance of enchytraeids fully recovered, and some delayed effects were observed in earthworm abundance and plant biomass, especially at the higher soil conductivity level. The observed low capacity of soil to retain salts suggests that, particularly at high soil conductivities, nearby freshwater bodies can also be endangered. Under saline conditions similar to the ones assayed, survival of some soil communities can be threatened, leading to the loss of biodiversity.

This article is part of the theme issue ‘Salt in freshwaters: causes, ecological consequences and future prospects’.

1. Introduction

Global mean temperature increases and consequent water expansion along with melting of snow covers and ice stocks are contributing to the rise of sea level [1]. This rise in sea level has severe effects on groundwater, potentiating its salinization [1], although this process is dependent on various factors such as the configuration of the hydraulic aquifer and shoreline [2]. Seawater intrusion effects include the scarcity and reduction in quality of freshwater. A contamination of less than 1% seawater will impair the use of such water for human consumption [2]. Besides this, the consequent use of saline water for irrigation is an emerging problem, since it can lead to the accumulation of salts in soils (salt-affected soils) [3], with severe consequences for soil fertility and yield production [1]. The invasion of seawater in groundwater has been reported worldwide though with different genesis: the main reasons for seawater intrusions in Asia are dissolution, igneous activity, hydrothermal mineral water, combination of evaporation and dissolution, and irrigation; while in northern Europe it is caused by marine transgression [4]. A few examples of regions affected by seawater intrusions are: (i) coastal areas of southern Pakistan, either through natural processes or due to irrigation [4], where, in 2002–2003, 92% of total groundwater was composed of seawater [5]; (ii) the Tiber River delta (Rome, Italy), with seawater intrusions occurring through natural processes or due to the combination of transgression and recent flooding [4], and corresponding to electrical conductivities between 550 and 10 750 µS cm⁻¹ (October 2012, dry season) [6]; and (iii) Almada (Lisbon, Portugal), where 5% of the total groundwater was composed of seawater, corresponding to an electrical conductivity of 6290 µS cm⁻¹ [7].

Soil salinization (accumulation of soluble salts to the extent that fertility is affected) can occur through natural processes like increased evapotranspiration (primary salinization) and/or be induced by human activities like the increased withdrawals from aquifers (secondary salinization) [8]. The occurrence of saline soils is mostly dependent on the geographical location: soils from coastal areas are more affected by groundwater withdrawals and extreme events (tsunamis and floods) [3]. Saline soils are normally characterized by an electrical conductivity equal to or higher than 4000 µS cm⁻¹ (a pH lower than 8.5 and an exchangeable sodium percentage lower than 15%) [9,10]. Globally, the most affected regions for soil salinization are the arid and semi-arid parts of Australia (84.7 million hectares (Mha); but the main reason for soil salinization in Australia is dryland salinity instead of sea-level rise and/or seawater intrusion), Africa (69.5 Mha), Latin America (59.4 Mha), the Middle East (53.1 Mha) and Europe (20.7 Mha) [8,11]. It is estimated that global soil salinization will continue spreading at a rate of up to 2 Mha yr⁻¹ [12]. Salinization effects on agriculture include productivity changes, an increased risk for less tolerant species and an increase of plant pests and/or diseases, and, in a long-term scenario, the loss of biodiversity [1]. Soil organisms are essential to the normal functioning of the soil ecosystems since they participate in key ecological processes like organic matter decomposition, nutrient cycling and maintenance of soil structure [13]. Studies on the salinization effects on soil organisms are scarce, but negative effects of saline conditions on survival and reproduction of soil invertebrate species [14–19] or on avoidance behaviour of earthworms [20] have been reported. Deleterious effects of soil salinization on diverse life stages of plants were also described [21,22], including decreased and/or delayed germination and/or effects on seedling physiognomic state, deficient growth, as well as a decrease of photosynthetic pigments, and of global biomass [23]. This leads to yield reduction of several commercial plant species, for example: 72–100% for rice, Oryza sativa (using soil from Tanjung Karang, Malaysia) [24]; 5–79% for wheat, Triticum spp. (in Utah, USA) [25]; and 26.2% in tomato, Lycopersicum esculentum, (in Negev, Israel) [26]. However, salinization tolerance can be found in some soil inhabitants like the spider Arctosa fulvolineate and the beetle Merizodus soledadinus, which survived salinity levels up to 70‰ (around 70 g l⁻¹ assuming NaCl is the main salt present) [27,28]. Also, some littoral and terrestrial amphipod species can survive salinity levels up to 900 mOsm external concentration (around 26 g l⁻¹ assuming NaCl is the main salt present) [29]. Nevertheless, in a six month exposure scenario to a low salt concentration (which theoretically would induce a 20% reduction of offspring production), the springtail Folsomia candida revealed a higher sensitivity than expected, with a decrease of 95% of the initial population density. On the other hand, the enchytraeid Enchytraeus crypticus presented a higher tolerance after being long-term exposed (six months) than after exposure for only 28 days (normal standard reproduction test duration) [30]. Despite this, several authors have highlighted the need to use more realistic approaches and techniques in the evaluation of both salinization effects and potential recovery of the soil ecosystems [15,18,20,30].

The main goal of this study was to evaluate the effects of salinity on coastal ecosystems, focusing on the soil compartment. More specifically, this study aimed: (i) to evaluate the effects of seawater intrusion alone and combined with irrigation with saline water, resultant from the mixing of seawater into groundwater, on soil invertebrate and the natural plant communities; and (ii) to assess the recovery potential of those communities after the cessation of saline irrigation.

2. Material and methods

Terrestrial model ecosystems (TME) [31] were used to assess soil salinization effects on soil communities and their potential recovery. These reproducible and controlled semi-field set-ups are accepted tools to evaluate the effects of soil contaminants and other stressors (including climate changes), as well as management practices, on soil communities [32–37]. Indeed, they consist of intact soil cores, containing the natural communities of soil fauna and flora, where long-term stability of the soil system is provided, allowing evaluation of not only the effects but also the communities' recovery potential [32,38]. A pre-incubation period of TME is advisable in order to allow stabilization of the soil communities. This method has its associated constraints, such as: high spatial variability, mimicking that found in the field [32]; choosing the maximum time period; and, also, assessing the time necessary to detect effects and recovery for the different endpoints [32]. Nevertheless, it is a great link between laboratory (single species) and field assays, having some of the advantages of both, such as standardization and controlled conditions (from the laboratory) and natural variability and complex interactions (from field assays). It also diminishes some of the disadvantages of both: simplicity from the laboratory and high amount of man power needed for field assays [32,39].

(a). Extraction of TME

Thirty-one soil cores, from now on referred to as TME, were extracted in an undisturbed area at the Agricultural School in Coimbra, Portugal (40°12′54.7″ N 8°26′59.9″ W). The extraction was made using a stainless steel extractor and the intact soil core was encapsulated in a high density polyethylene tube with 16.5 cm internal diameter and 40 cm in length. A few days before sampling (December 2013), the existing cover vegetation was cut to approximately 5 cm high, in an area of approximately 10 m². In this area, the groundwater level is very close to the soil surface and the soil has a sandy-loam texture (sand—62.4%, silt—21.2%, clay—16.4%), pH = 6.1 (measured in 1 M KCl) and 3.1% organic matter (measured by ‘Direcção Regional de Agricultura de Entre-Douro e Minho, Portugal’ and ‘Agência Portuguesa do Ambiente, Portugal’ according to [40]). After the extraction, the cores were transported to a controlled room (temperature of 20 ± 2°C, photoperiod 16 : 8 light : dark and humidity of 60 ± 3%) where the experiment was conducted. They were placed in buckets lined with a layer of marbles to improve the contact between soil and intrusion water.

(b). Experimental design

After collection, six TME immersed in and irrigated with distilled water for two weeks were sampled (T0) to assess the initial condition of the soil communities. Afterwards, the salinization effects on soil communities and the seawater intrusion with or without saline irrigation were assessed during six weeks (sampling period—T1; defined according to the standard duration of standard ecotoxicological tests—generally four weeks for soil invertebrates and two/three weeks for plants). The irrigation with saltwater was then stopped and replaced by distilled water for eight weeks, after which the recovery of the communities was assessed (T2). The latter (and longer) period was established assuming that the recovery processes take more time to be detected in the organisms than the potential effects. The total length of the test was 16 weeks as advised in order to maintain the stability of the soil communities [32,35]. A summary of the test-design is presented in table 1.

Table 1.

Test-design, including the treatments, their durations and number of replicates. (IIDW—intrusion and irrigation with distilled water; SIDWI—saline intrusion combined with irrigation using distilled water; SISI—saline intrusion combined with saline irrigation; SINSI—saline intrusion combined with irrigation using distilled water after the cessation of the saline irrigation).

	initial community	saltwater exposure phase			recovery phase
sampling period	T0	T1			T2
duration (weeks)	2	6			8
treatment	IIDW (control)	IIDW (control)	SIDWI and SISI		IIDW (control)	SINSI
seawater dilution (%)	0	0	8	31	0	8	31
corresponding solution conductivity value (µS cm−1)	9	9	6000	18 000	9	6000	18 000
no. replicates	6	5	4	4	4	4	4
endpoints measured	abundance of microarthropods, nematodes, enchytraeids, earthworms and plant biomassa
test conditions	20 ± 2°C; 16 : 8, light : dark.

^aPlant biomass was not assessed in T0.

(i). Treatments

The 31 soil cores were divided into four treatments: (IIDW) intrusion and irrigation with distilled water (control; sampled in T0, T1 and T2); (SIDWI) seawater intrusion with simultaneous irrigation with distilled water (sampled in T1); (SISI) seawater intrusion and saline irrigation (half sampled in T1 and the other half continuing to T2 as the treatment SINSI); and (SINSI) seawater intrusion and irrigation with distilled water (after the cessation of saline irrigation; only sampled in T2). Seawater intrusion and/or saline irrigation were simulated using two seawater dilutions (8 and 31%) with conductivity values of 6000 and 18 000 µS cm⁻¹ (table 1). The seawater dilutions were prepared with distilled water instead of tap water owing to the presence of salts in the latter, which would probably change the conductivity of the test solutions. For the same reason, distilled water was used for TME irrigation without saline solution.

The scenario of soil intrusion here presented was used to simulate soils of coastal areas in which the groundwater is close to the soil surface. In a real case scenario, the farmers would use the groundwater (saline or not) to irrigate the fields, conditions reflected in the treatment SISI. The last treatment, SINSI, was set up to assess the community recovery if the saline irrigation were to stop. Saline intrusion was maintained since it would be impossible to replace the saline groundwater. The irrigation was performed twice a week with 250 ml of saline or distilled water according to the treatment (value derived from mean annual irrigation in the Vouga, Mondego and Lis river watersheds) [41].

In each TME of treatment SISI, the conductivity value of the water for irrigation and intrusion was the same. The conductivity values of water wells in Almada (Portugal) was assessed and the maximum value recorded was around 6000 µS cm⁻¹ (corresponding to 5% seawater present in the total groundwater) [7]. These observations were the basis for the selection of the lower conductivity value tested. The higher seawater dilution was chosen in order to simulate a scenario of seawater intrusion that could negatively affect soil communities (conductivity of 18 000 µS cm⁻¹ in the seawater dilution is equivalent to ca 850 µS cm⁻¹ in the test soil, which was the EC_50reprod calculated by [18] for F. candida in soil—the lowest value obtained).

(c). Destructive sampling procedure of each TME (in each sampling period)

(i). Plants

No crops were used in this study. However, the aerial part of the natural community, mainly composed of small herbaceous plants, from each TME was sampled. The plant material was collected in T1 and also in T2 and transported to the Soil Ecology and Ecotoxicology Laboratory (Coimbra, Portugal). The samples were dried at 60°C for 24 h and weighed.

(ii). Microarthropods and enchytraeids

Two soil cores (5 cm in diameter and length) were collected per sampling period per TME. One of the soil cores was placed in a Macfadyen extractor for seven days at 45°C to extract microarthropods (mites and springtails), which were then kept in 80% ethanol until being sorted under a stereomicroscope. The other soil core was used for enchytraeid extraction. The soil was mixed with 96% ethanol (in a ratio of 5 ml of ethanol per 25 g of soil fresh weight—FW), filled with water until all the soil was submersed, and stained with six to ten drops of rose Bengal per 25 g of soil FW. The enchytraeids were then quantified under a stereomicroscope.

(iii). Nematodes

Three soil cores (1 cm in diameter and 5 cm in length) were collected per sampling period per TME. The nematodes were extracted using the tray method [42] and the suspension was kept at 4°C, for a maximum of 14 days, during which time they were quantified under an inverted stereomicroscope.

(iv). Earthworms

Earthworms were searched for in the remaining soil in two layers inside the TME: 0–30 and 30–40 cm. Such sorting between soil layers was intended to evaluate the abundance of earthworms inhabiting the layers of the soil closer to the surface (epigeic and endogeic; 0–30 cm) and earthworms that live in deeper layers of the soil (anecic; 30–40 cm). The individuals collected were divided into adults and juveniles.

(v). Soil conductivity

Soil conductivity was measured as a surrogate of soil salinity. The measurements were determined in three parts of each TME. Samples from 0–5 cm (top layer), 5–30 cm (intermediate layer), and 30–40 cm (bottom layer) were collected to evaluate the influence of the intrusion in the bottom layer combined with the influence of irrigation in the top layer. The water movement inside the TME was assessed in the intermediate layer. Each sample (5 g FW) was mixed with distilled water (1 : 5 soil : water) and the conductivity measured with a conductivimeter.

(d). Data analysis

Data analysis of total abundance of mites and springtails, and the number of organisms/g of soil (FW) for enchytraeids and nematodes was carried out. The results were compared through generalized linear models (GLM; Gaussian family and link identity, chosen owing to the lowest Akaike information criterion). In order to assess the effects of seawater exposure, comparisons between the treatments (SIDWI and SISI) and the control of the sampling period T1 were performed using the statistical test mentioned above. To assess the community recovery, comparisons between treatment SINSI and the control in T2 were also investigated through GLM (Gaussian family and link identity). All the analyses mentioned above were performed in the R v. 3.5.1 software using the interface RStudio (R available at https://cran.r-project.org/bin/ and RStudio available at https://www.rstudio.com/products/rstudio/download/).

3. Results

(a). Soil conductivity

An increase in soil conductivity values was observed from IIDW to SIDWI and continuing in SISI at T1 (table 2). As expected, higher values were observed in the bottom (30–40 cm) and intermediate layers (5–30 cm) with the SIDWI 18 000 µS cm⁻¹ level as compared with the 6000 µS cm⁻¹ level of the same treatment, indicating an upwards movement of salt by capillarity. The increase in conductivity values in the SISI levels, when compared with the corresponding levels of the SIDWI, especially in the topsoil layer (0–5 cm), was also expected owing to the irrigation with saline water. At T2, the conductivity values from SINSI (after the cessation of the saline irrigation) were higher for the bottom layers (5–30 and 30–40 cm) when compared with the SISI treatment at T1. The conductivity measured in the top layer (0–5 cm) at the 18 000 µS cm⁻¹ level decreased owing to irrigation with distilled water. This could have also caused the leaching of salt to deeper layers as observed by the increase in conductivity values in both the intermediate (5–30 cm) and the bottom (30–40 cm) layers. Thus, leaching occurred for both conductivity levels (table 2).

Table 2.

Mean conductivity values measured in each treatment for soil at three soil depths in the TME (for treatment or set-up detailed information see table 1).

treatment (number of replicates)	conductivity of the seawater dilution (µS cm−1)	soil depth in the TME (cm)	mean soil conductivity (µS cm−1) ± standard deviation
IIDW (5)	9	0–5	103 ± 52
		5–30	116 ± 254
		30–40	32 ± 16
SIDWI (4)	6000	0–5	142 ± 58
		5–30	109 ± 58
		30–40	163 ± 14
	18 000	0–5	179 ± 118
		5–30	352 ± 115
		30–40	569 ± 82
SISI (4)	6000	0–5	652 ± 91
		5–30	250 ± 31
		30–40	240 ± 53
	18 000	0–5	1629 ± 436
		5–30	566 ± 163
		30–40	529 ± 132
SINSI (4)	6000	0–5	806 ± 136
		5–30	454 ± 59
		30–40	323 ± 74
	18 000	0–5	1090 ± 89
		5–30	696 ± 64
		30–40	791 ± 66

(b). Effects of seawater exposure on plant biomass and abundance of soil invertebrates

Throughout the 16 weeks of the experiment and only taking into account the control results, earthworms and plants increased their total abundance and biomass, respectively. For total abundance of earthworms, statistically significant differences were found between the initial community (T0) and the organisms found at the end of the test (T2; p = 8.75 × 10⁻⁵; figure 1). Plant biomass recorded for T1 and T2 was also significantly higher than in T0 (p = 0.035 and 6.56 × 10⁻⁶, respectively; figure 1). Springtails and nematodes also showed an increase in total abundance between T1 and T0, followed by a small decrease in T2 (figure 1). Nevertheless, for springtails these differences were not statistically significant (p = 0.106 and 0.059, respectively for T1 and T2; figure 1), while for nematodes differences between total abundance in T1 and T2 in relation to T0 were statistically significant (p = 0.003 and 0.006, respectively; figure 1). The total abundance of mites gradually decreased along the experiment period, although with no statistically significant differences (p = 0.775 and 0.526, respectively for T1 and T2; figure 1). In enchytraeids, similar abundance values were found for the initial (T0) and the first sampling period (T1), but a substantial and statistically significant decrease was observed in the last sampling period (T2, p = 0.001; figure 1).

Figure 1.

Over-time variation of the dry weight of plants and total abundance of five invertebrate groups in the control treatment (IIDW): mean (+ standard deviation) abundance of nematodes, springtails, mites, enchytraeids and earthworms, and dry weight of plants at the beginning of the test (T0), and sampling periods T1 and T2. Asterisks mean statistically different from control in T0, Generalized linear model, Gaussian family (link identity), p < 0.05.

Focusing on the seawater treatments, a decrease in enchytraeids' abundance was detected in both treatments and related to the increase in conductivity values. This decrease was higher in the SISI treatment, where statistically significant differences were found between treatment SISI at both conductivity levels (6000 and 18 000 µS cm⁻¹) and the control (p = 0.010 and 0.000, respectively; figure 2). Despite some trends found in organisms' responses, namely mites, earthworms and plants (decreased in the treatments SIDWI, SISI, or SISI at the higher conductivity level, respectively, when compared with the control), no significant differences were found (p = 0.799 and 0.785 for mites at respectively SIDWI 6000 and 18 000 µS cm⁻¹, p = 0.293 and 0.137 for earthworms at respectively SISI 6000 and 18 000 µS cm⁻¹, and p = 0.088 for plants at SISI 18 000 µS cm⁻¹; figure 2). For nematodes, a similar number was found in the control and in both treatments (SIDWI and SISI), with no significant differences (p = 0.497, 0.392, 0.957 and 0.713, respectively for SIDWI 6000 and 18 000 µS cm⁻¹ and SISI 6000 and 18 000 µS cm⁻¹; figure 2). Despite the higher average abundance of springtails in the control compared with salinization treatments, no statistically significant differences were found (p = 0.315, 0.223, 0.131 and 0.587, respectively for SIDWI 6000 and 18 000 µS cm⁻¹ and SISI 6000 and 18 000 µS cm⁻¹; figure 2). Such results were probably due to the high variability within each treatment, visible in the range of the standard deviation.

Figure 2.

Effects of saltwater exposure on soil communities: mean abundance (+ standard deviation) of nematodes, springtails, mites, enchytraeids and earthworms, and mean dry weight of plants in the control (black bars) and both treatments (grey bars) in the first sampling period (T1). Treatment codes as in table 1. Asterisks mean statistically different from control, generalized linear model, Gaussian family (link identity), p < 0.05.

(c). Recovery potential of the soil organisms after the cessation of saltwater irrigation

In the last sampling period (T2), earthworm abundance at both conductivity levels in SINSI was statistically significantly higher than the control (p = 0.001 and 0.000, respectively for SINSI 6000 and 18 000 µS cm⁻¹; figure 3). Since, per TME replicate, the collected earthworms were separated into adults and juveniles found in two layers (0–30 and 30–40 cm), further data analysis was performed to investigate the salinization effects on the different life stages and their vertical distribution (the latter with the calculation of the ratio between the number of organisms found in each layer). Statistically significant effects were seen for the abundance of juveniles, especially those found in the 30–40 cm layer (p = 0.020 and 0.005, respectively for the juveniles exposed to 6000 and 18 000 µS cm⁻¹ in the layer 0–30 cm, and p = 7.51 × 10⁻⁵ and 3.84 × 10⁻⁵, respectively for the juveniles exposed to SINSI 6000 and 18 000 µS cm⁻¹ in the layer 30–40 cm; figure 4a). The ratios calculated revealed a statistically significant upward movement inside the TME among the juveniles inhabiting the 30–40 cm layer (p = 0.001 and 0.000, respectively for SINSI 6000 and 18 000 µS cm⁻¹; figure 4b).

Figure 3.

Abundance of soil organisms after the recovery period, i.e. after cessation of saltwater irrigation: mean abundance (+ standard deviation) of nematodes, springtails, mites, enchytraeids and earthworms, and mean dry weight of plants in the control (black bars) and treatment SINSI (grey bars). Treatment codes as in table 1. Asterisks mean statistically different from control, generalized linear model, Gaussian family (link identity), p < 0.05.

Figure 4.

Discrimination of the delayed effects on earthworms after the cessation of saltwater irrigation: mean number (+ standard deviation) of earthworms divided into adults and juveniles and, also, juveniles in the two sampled layers of the TME (0–30 and 30–40 cm) (a); ratio between the number of individuals collected from the layers 30–40 cm and 0–30 cm (b). Treatment codes as in table 1. Asterisks mean statistically different from control, generalized linear model, Gaussian family (link identity), p < 0.05.

A gradual decrease in plant biomass was observed with increasing conductivities, but statistically significant differences in relation to the respective control were only found for the conductivity level of 18 000 µS cm⁻¹ (p = 0.009; figure 3).

When comparing SINSI with the respective control, no statistically significant differences were found for enchytraeid abundance (p = 0.555 and 0.718, respectively for 6000 and 18 000 µS cm⁻¹; figure 3) or nematode abundance (p = 0.826 and 0.990, respectively for 6000 and 18 000 µS cm⁻¹; figure 3). A high variability in the responses of both springtails and mites persisted in T2 and no significant differences were found between SINSI and the control (p = 0.528 and 0.772 for springtails at SINSI 6000 and 18 000 µS cm⁻¹, and p = 0.159 and 0.812 for mites at SINSI 6000 and 18 000 µS cm⁻¹; figure 3).

4. Discussion

(a). Fluctuations of the soil invertebrates in the control scenario

Two limitations inherent to the TME methodology are the definition of the maximum exposure period as well as of the necessary time to detect effects and recovery for the endpoints previously established, without compromising the communities' stability [32]. During the 16 weeks of the TME assay, the soil cores were kept under controlled conditions that would be optimal for the growth and reproduction of soil organisms [38]. Therefore, a constant or even an increase in abundance and/or biomass of the organisms sampled in IIDW should be observed. Actually, no temporal variability was detected for microarthropods (springtails and mites), which might have been masked by the high spatial variability. On the other hand, enchytraeid total abundance presented a huge decrease in the last sampling period while nematode total abundance increased in T1 and decreased in T2, both presenting a high temporal variability. This decrease in T2 might be due to scarcity of food or the presence of predators (e.g. mites), which could have limited the expansion of both nematodes and enchytraeids. As for enchytraeids, their sensitivity to the location of food, intra- or interspecific attraction or repellence, and competition has been described in field experiments [31], which could explain the decrease in the abundance observed after 16 weeks. Enchytraeids can also be susceptible to changes in soil moisture [43,44]. Although the irrigation in the present study was performed twice a week, some moisture fluctuations might have occurred. Temporal variability was also found for earthworms and plants, which had a gradual increase over time. These two groups of organisms had no predators, theoretically, therefore, allowing the free growth of their populations over time.

A high spatial variability is also associated with TME experiments, mimicking that found in the field, already described for microarthropods and microbial communities [32,45]. This aspect can sometimes impair the detection of differences when comparing the treatments with the control. In the present study, a high spatial variability was found for microarthropod (springtail and mite) total abundance, contrarily to nematodes, enchytraeids, earthworms and plants (represented in the graphs by the low standard deviations, i.e. low variance between replicates).

(b). Effects of salinization on coastal ecosystems

In a real scenario of saline intrusion (combined or not with saline irrigation), under similar exposure conditions to the ones here tested, we should have expected to observe hazard effects mostly for two groups of soil mesofauna, enchytraeids and springtails (although for the latter group no statistically significant differences were observed, mostly owing to the high spatial variability found in IIDW—control treatment). With respect to plants, soil microfauna (i.e. nematodes), macrofauna (i.e. earthworms) and other mesofauna groups such as mites, the results obtained suggest no hazard effects are foreseen. Actually, no effects were expected in nematodes' abundance since it is known that hatching is not influenced by sodium chloride (along with other salts) in considerable concentrations—up to 850 mM sodium chloride and up to 266 mM calcium chloride [46]. In fact, the perfect solution for hatching [42] and the exsheathing of nematodes before the parasitic juvenile stage [47] is contains a small percentage of sodium chloride. Mites were also not expected to be affected as an EC_50reprod of 6028 µS cm⁻¹ was calculated for the mite Hypoaspis aculeifer [18], assuming that the mite community has a similar sensitivity, and for the conductivity values measured in soil in the present study. This high tolerance of mites could be due to the presence of a rigid exoskeleton [48], and/or specialized organs responsible for osmoregulation, like the coxal glands [49], and sclerotized rings in the cuticle [50]. The actual limit to define saline soils—4000 µS cm⁻¹—relies on the effects on plants [9]. Therefore, no negative effects for plants were expected at the conductivity values measured in the soil (the large majority under 1000 µS cm⁻¹ in T1; table 2). Strategies like the reduction of stomatal density and transpiration have already been described in plant species under saline conditions [51].

Conversely, effects of salinity on reproduction and survival of two species of earthworms (Eisenia fetida and Aporrectodea caliginosa) have been found [15]. There was no reproduction at conductivity values of 520 µS cm⁻¹ and higher, and effects on survival were detected at conductivity values of 920 and 1310 µS cm⁻¹, respectively for E. fetida and A. caliginosa. In the present study, the absence of negative effects on earthworm abundance in T1 suggests no impairment of reproduction and survival in the native community being studied. Indeed, when eight earthworm species from Pakistan were exposed to a gradient of salt contaminated soil for four weeks, three of them survived salinity levels up to 10 480 µS cm⁻¹ [52]. These results reinforce the idea that the community tolerance to salt also depends on its composition, since different species might have different tolerance/sensitivity thresholds to contaminants [52]. The springtail species Folsomia candida was also used in an ecotoxicological standard test with natural saline soil (electrical conductivity values ranging from 80 to 1620 µS cm⁻¹) and salinity effects were not found on survival but rather on reproduction, at conductivity values starting from 1030 µS cm⁻¹ [15]. After a long exposure to a saline stress, springtails showed a higher sensitivity than springtails exposed for 28 days only [30]. Therefore, the effects observed in the present study on springtail abundance were expected, especially at the higher conductivity level (18 000 µS cm⁻¹). The decrease in abundance was even observed at the lower conductivity level (6000 µS cm⁻¹) of both treatments, indicating a higher sensitivity of the natural springtail community as compared with the one previously found for F. candida. When testing a community, the range of sensitivities or range of effects gets wider, making the comparison with data obtained from standard species difficult. In a real case scenario, some species can actively avoid saline conditions, moving to more favourable microhabitats [53]. Physiological avoidance mechanisms (i.e. absorption of water vapour; reduction of water loss increasing the haemolymph osmolality, by the abnormal increased production of sugars and polyols; and production of chemicals turning the cell membranes fully functional by changing their fluidity and functionality) can be activated [54,55]. Nevertheless, in the case of salt-sensitive species dominance, a decrease in abundance can be observed.

Enchytraeids, as expected, had a significant reduction in abundance in the treatment with both saline intrusion and saline irrigation, which is in accordance with previous findings [15,18]. Indeed, the effects were observed at conductivity values in the top soil layer of SISI higher than previously derived EC₂₀ and EC₅₀ for reproduction [18]. On the other hand, lower conductivity levels such as the ones measured in the saline intrusion treatment did not impair survival or reproduction of the enchytraeid community. Enchytraeids possess osmoregulation mechanisms, like the release of hypotonic urine, impermeabilization of their membranes, which is observed for salinity levels of up to 40‰ seawater (around 40 g l⁻¹ assuming NaCl is the main salt present) [56], and transport of active amino acids and sugars by means of an ATPase-dependent Na-gradient which is activated in the presence of sodium and inactivated at lower than 15‰ salinity levels (around 15 g l⁻¹ assuming NaCl is the main salt present) [57]. In any case, these mechanisms could have worked in low salt concentration but not in higher concentrations, as observed in the topsoil layer of SISI.

In regard to the soil conductivity values measured in T1, the increase observed for deeper soil layers (SISI) suggests that the soil could not retain the salts. In a real case scenario of salt-affected soils, if there was an episode of rainfall, salt lixiviation could occur, affecting adjacent water bodies. The maximum conductivity value in soil observed in the effects phase was 1628 µS cm⁻¹, and freshwater bodies are not supposed to be affected by salt leaching up to 1000 µS cm⁻¹. Nevertheless, negative effects on less tolerant aquatic communities have previously been observed [58–61], with a change in the community structure (substitution of the less tolerant taxa to more tolerant) [62,63].

Most of the groups described above seem not to be affected at the conductivity levels tested, which represent seawater dilutions up to 31%. In scenarios of seawater intrusion higher than the ones here presented (e.g. 92% seawater intrusion [5]), the effects can be more severe. Also, other salt sources, such as the release of salts accumulated in deeper soil levels, owing to the seawater intrusion [62], can exacerbate or provoke negative effects on the ecosystem. Besides, the avoidance behaviour of some of these organisms has already been described for salinity, which can impair coastal ecosystems functioning at lower conductivity levels than those described here [20,62]. In future studies, it will be advisable to increase the range of conductivity levels and to tackle aquatic and terrestrial compartments simultaneously, combining methodologies to assess the effects on soil and aquatic communities (for example, a combination of aquatic and terrestrial mesocosms).

(c). Recovery potential after the cessation of saltwater irrigation

At T2, after eight weeks without saline irrigation, no differences of total abundance in relation to the control were found in nematodes, springtails, mites or enchytraeids, while delayed effects were found for earthworm abundance and plant biomass. For enchytraeids, if we compare the data for treatments SINSI and IIDW from T2, the effects observed after the exposure at T1 were completely overcome, showing a full recovery of the community abundance. This full recovery of enchytraeids had already been observed in relation to drought regimes [64,65]. A natural community of enchytraeids in a copper-contaminated field site in Denmark was exposed to summer drought conditions [64]. Despite the negative effects of copper on diversity and abundance, summer drought effects were not visible. This indicates a rapid recovery, already found in a natural enchytraeid community that recovered (in terms of density and biomass) from a drought stress in two months [65]. Additionally, it has been shown that the recovery is not due to vertical migration, but is related to survival of these organisms in a drought tolerant stage.

The detection of delayed effects on earthworm abundance and plant biomass could be in agreement with the increase in conductivity values particularly in the intermediate and bottom soil layers in the SINSI treatments (T2) when compared with the SISI treatments (T1). These delayed effects were not expected for plants (for reasons described above) but were expected for earthworms [15]. Earthworms also displayed vertical avoidance as has been observed when exposed to other stressors like biochars, drought conditions, seasonal attrition and pesticide—carbendazim [26,66–69]. The delay in effects or the migration of organisms, particularly in these two organisms groups, suggests a potential long-term impairment of the soil functions mediated by them.

The conductivity values in soil after the recovery phase, eight weeks after the end of irrigation with saltwater, were higher in all the soil layers for the lower seawater dilution (6000 µS cm⁻¹) and in the lower soil layers for the higher seawater dilution (18 000 µS cm⁻¹). In a real scenario, continuous salt leaching to nearby water bodies (at approx. 1000 µS cm⁻¹) can persist even two months after cessation of saltwater irrigation, which will probably increase the salinity of such water bodies. Under natural conditions, the recovery of an aquatic ecosystem is fully visible only after three years of the cessation of disturbances due to, e.g., chemicals, drought and flooding [70,71]. In the case of constant input of salt, recovery seems unlikely to occur. The cessation of such a process seems impossible owing to the difficulty of groundwater ‘desalinizing’. It would be interesting to test whether salts in groundwater can travel upwards in the soil as seen at the lower seawater dilution or after some irrigation with non-saline water the amount of salts diminishes as seen at the higher seawater dilution. It would also be interesting to assess for how long salts can be leached from the soil, the effects of both the duration and the continuous input of salt on aquatic biota, and the potential for recovery after these conditions.

5. Conclusion

After a six week exposure to saline intrusion and irrigation, only enchytraeid abundance presented a significant reduction, which was overcome after an eight week recovery period. On the other hand, plant biomass and earthworm abundance showed delayed negative effects, with the latter presenting a vertical migration of the juveniles from the deeper soil layers, probably to avoid the saline conditions. Looking at the conductivity values measured, the transport of salts from deeper soil layers subjected to continuous saline intrusion (30–40 cm) to the top soil layer (0–5 cm) might have happened. Therefore, under real scenarios, salt leaching to water bodies can occur.

The results suggest that soil organisms might be slightly impacted by seawater intrusion combined or not with saline irrigation, at least in the range of the conductivity levels here studied. Nevertheless, values of seawater intrusions higher than the ones used here have been recorded, suggesting that stronger effects than those found in this work can be observed on soil fauna. The results also show that the sensitivity of a community is highly dependent on its composition, and that comparison with standard (one generation) laboratory assays should be made with caution. Therefore, the use of more realistic approaches that simulate the complexity of the soil system, such as TME, is advised. In a real scenario of salinization of a coastal ecosystem, the cessation of the saline irrigation coupled with leaching of the salts from the root zone to deeper soil layers would most probably reverse the effects observed. Nevertheless, the leaching of the salts could lead to an increase in groundwater salinity and affect the water bodies (and their inhabitants) nearby. Future studies on effects of soil salinization should include both aquatic and terrestrial compartments.

Data accessibility

This article has no additional data.

Authors' contributions

C.S.P., J.P.S. and S.C. conceived and planned the methodology. C.S.P. and S.C. carried out the experiment. I.A. and C.S.P. collected the data. I.L. supervised the project. All authors discussed the results and contributed to the final manuscript.

Competing interests

We declare we have no competing interests.

Funding

This study was funded by FEDER funds through the COMPETE programme, by Portuguese national funds through FCT-Fundação para a Ciência e Tecnologia, under the projects SALTFREE (PTDC/AAC-CLI/111706/2009), SALTFREE II (POCI-01-0145-FEDER-031022) and ReNature (SAICT 000007 ReNature), and the CESAM strategic programme (UID/AMB/50017/2013) by a post-doctoral grant to S.C. (SFRH/BPD/84140/2012) and a contract to I.L. (IF/00475/2013) and IATV-Intstituto do Ambiente Tecnologia e Vida funds.