1): Colluvial or slope MSS

This is formed on sloping ground [7], and is the result of the erosion and deposition of rock fragments produced by the mechanical fragmentation of rock walls or the erosive action of continental glaciers. The nature of the rock can be diverse, including calcareous, siliceous and even volcanic rocks. These rocky deposits can be covered by evolving soil (with edaphic horizons) of different thicknesses which often supports dense plant cover. In other cases, the upper level of the MSS is bare, appearing as mountain scree.

2): Bedrock MSS

This is formed by the weathering of much of the bedrock [13], a gradual process that occurs almost simultaneously with the formation of edaphic horizons which seal it off from the external environment. This type of MSS is found in valley bottoms or areas with little or no slope. The nature of the rock may be very diverse, although it is more likely to occur in rocks that are easily altered.

3): Volcanic MSS

This is the result of the accumulation of volcanic material [14]. From formation onwards, it contains many interconnected fissures forming a tangled web of micro-spaces, which over time become isolated from the outside due to the development of an edaphic environment [15]. This type of MSS is more ephemeral than others, since deposits of volcanic material are more susceptible to weathering than those composed of more compact rocks such as sedimentary, plutonic or metamorphic rocks.

In this paper, we present a proposal supported by experimental and theoretical arguments for a new type of MSS which we have named “alluvial MSS”, and which corresponds to the crevices (unfilled spaces) which lie beneath the streambeds of temporary watercourses.

Study area

The geographic area selected for our study encompassed the eastern end of the karst reliefs of the Prebetic range, from the centre and north of the province of Alicante to the south of the province of Valencia (Eastern Spain), and was chosen because there are numerous temporary watercourses in the area, thanks to a series of geological, geographical and climatic factors. In the local toponymy, these types of watercourse are termed “rambla” (low, wide channel) and “barranco” or “barranc” (tall, narrow channel), words which reflect cultural and linguistic factors [23,24] rather than a true hydrographic characterisation. Regardless, these dry watercourses are usually dry for most of the year, only carrying water for an average of ten days a year [24,25].

The study area was divided into first-order river basins (Figure 1), following the criterion of IDEJúcar [26], and subsequently, basins in the northern half, the wettest, most mountainous area, were selected. A preliminary selection of promising sites for sampling was made using Sigpac [27] and Google, Earth [28]. Of the 42 sites selected a priori, 16 were considered eligible (Figure 1; Table 1) because the watercourses they contained were dry for most of the year and had thick alluvial deposits with numerous subterranean spaces. The other sites were rejected because they did not present either of the two features mentioned. Many of them overlay inappropriate lithologies, as was the case with the marly series of the Tap [29] which constitute physical barriers to the movement of subterranean fauna [30].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Study area.

(A) Alicante watersheds. (B) Sampling points.

https://doi.org/10.1371/journal.pone.0076311.g001

The permissions for the sampling and capture of invertebrate fauna in the study area were given by the Conselleria d’Infraestructures, Territori i Medi Ambient of the Generalitat Valenciana. No private lands or protected areas were sampled, and neither endangered nor protected species were involved in this study. Specimens collected are deposited in the zoological collection of the University of Alcalá.

Construction and installation of subterranean devices

To avoid mechanical disturbance of the substratum each time the pitfall traps were removed, a multi-perforated cylinder was previously buried underground; the perforations gave fauna access to the inside whilst the cylinder enabled their subsequent collection since the pitfall trap could be removed via the vertical tube. This technique was first invented by Gers [11,13] and later developed by Owen [31]. Several years later, López & Oromí [32] proposed a somewhat more sophisticated trap model. Henceforth, we will refer to these devices as “subterranean sampling devices” (SSD). The design of the SSD used in this study was based on the model described by López & Oromí [32] to which some innovations were added (Figure 2). Cylinder length varied between 50 and 100 cm, and the cylinders were perforated below the upper 20 cm. The pitfall trap was lowered into and retrieved from the bottom of the tube using a nylon cord. The fauna collection chamber contained 1,2-propanediol, which does not evaporate and enables preservation of DNA. It also had a small central compartment which housed a phial, in the form of a renewable cartridge, containing solid bait (very strong-smelling cheese) to attract arthropods [4]. A single SSD was installed in each chosen location. The traps inside each SSD were operated for 12 months and were renewed every three months, from October 2011 to October 2012. Temperature readings were taken both inside the cylinders (at a depth of 20-30 cm) and above ground (at a height of between 1 and 2 m), recording data every hour (in most cases) or every ½ hour. Mean (T_mean), minimum (T_min) and maximum (T_max) temperatures were calculated for the entire cycle, as was the mean of the minimum (T_Mmin) and maximum (T_Mmax) daily (24 hours) temperature and the mean of the daily temperature range (T_Dmean). We applied a filter (moving average) to the time series using a window size of 12 (in most cases) or 24 hours, in order to better visualise the temperature pattern over time and determine the day/night cycle; minimum (T_Fmin) and maximum (T_Fmax) temperatures were calculated after applying the filter. We also measured relative humidity inside the cylinders and calculated the mean (RH_mean), minimum (RH_min) and maximum (RH_max) for the entire cycle.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Subterranean Sampling Device design.

(A) Cover and interlocking tube sections to accommodate different substrate thicknesses. (B) Assembled tube sections. (C) Longitudinal tube section showing internal devices (data collection and data loggers). (D) Detail of the capture devices (pitfall trap with float to avoid submersion in the case of a temporary rise in the phreatic level) and the digital temperature and humidity sensor (hourly data recording) protected inside a plastic hood. (E-F) Operation of the auto-save device should the watercourse become active again due to a gradual increase in the phreatic level.

https://doi.org/10.1371/journal.pone.0076311.g002

Relative humidity and temperature

Due to unpredictable floods, most of the data loggers were lost or spoilt and no data was obtained for a complete annual cycle; hence, the results given here are partial and they are thus impossible to generalise. Partial data for both the SSDs and the external environment were recovered for three locations, whilst for a fourth location, only SSD data were recovered. RH values inside the SSDs almost always showed complete saturation, with RH_mean values which were always above 80% and presented little variation (Table 2). With regard to temperature, T_mean values inside the SSDs and above ground were similar (Table 2), and SSD temperatures reflected both the daily and the seasonal cycle of the external environment, although the temperature cycle was somewhat out of phase since maximum and minimum temperatures were reached a few hours later in the substratum (Figure 3). The range of temperature variation, both for the entire cycle ([T_max-T_min] and [T_Fmax-T_Fmin]) and the daily cycle (T_Dmean), was much lower in the SSD than above ground (Figure 3 and Table 2). As regards minimum temperature values, although T_min values in the MSS occasionally fell slightly below 0°C in February, T_Mmin values rarely fell below 8°C (Table 2). T_Mmin values in the MSS were always above the values recorded above ground and showed much less variation. In contrast, T_Mmax values in the MSS were always below those above ground, and also showed less variation. The maximum temperature in the SSDs did not exceed 31.8°C even though temperatures above ground reached as high as 44.3°C, whilst the maximum T_Mmax value recorded for the MSS was 26.55°C.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Temperature series in Barranco de la Cueva de los Corrales (Subterranean Sampling Device 1).

Data for the interior of the Subterranean Sampling Device (solid line) and above ground (dotted line) after applying linear filtering (moving average) with a 12-hour window size. Horizontal lines represent mean values (solid line, Subterraenan Sampling Device; dotted line, above ground). The graph shows how the temperature inside the SSD was not as extreme as above ground and how the daily cycles were slightly out of phase.

https://doi.org/10.1371/journal.pone.0076311.g003

Fauna results

The initial taxonomic studies of the samples highlighted the ecological importance of the dry watercourse substratum for terrestrial fauna (Table 3, Table S1). The specimens collected were unevenly distributed among different phyla, both quantitatively and qualitatively. In cases where the SSDs had been affected by a rise in the phreatic level, or the watercourse had recovered its ephemeral hydrological activity, specimens of aquatic or semi-aquatic life were collected, belonging to the phyla Rotifera, Plathelminthes (class Turbellaria), Mollusca (class Gastropoda), Tardigrada, Arthropoda (classes Maxillopoda, Ostracoda, Malacostraca and Insecta) and Chordata (class Amphibia). Despite obtaining data for aquatic fauna, the majority of the fauna collected was terrestrial since these watercourses are dry for most of the year. These terrestrial specimens belonged to the phyla Mollusca (Class Gastropoda), Annelida (class Oligochaeta), Nematoda and Arthropoda, with a predominance of the latter group. Among the terrestrial Arthropoda in the substratum of these watercourses, the four current subphyla were well represented. Of note among the Chelicerata, and more specifically the Arachnida, was the presence of the Acari, especially of groups such as the Oribatida and the Gamasida, most of which can be considered edaphic species [33] which invade subterranean spaces. The Opiliones contributed several species, among which can be highlighted Trogulus lusitanicus Giltai, 1931 and Dicranolasma soerensenii Thorell, 1876. The greater abundance of the latter species in the substratum of the watercourses may be due to its troglophile lifestyle [34], as it has often been observed in caves where it appears to complete its life cycle [35] and has also been found in the colluvial MSS in the study area (unpublished data). Other Arachnida that appeared regularly in alluvial substratum samples were the Pseudoscorpiones, a group which typically includes numerous species specialising in the subterranean environment of the eastern reliefs of the Prebetic System [36,37]. Especially conspicuous among the Araneae were the Linyphiidae family, and in particular, the genus Lepthyphantes Menge, 1866 (sensu lato), a group that numbers many species closely associated with the subterranean environment [38,39]. The Linyphiidae Lessertia dentichelis (Simon, 1884) was also well represented, a species with a high preference for damp areas and a habitual inhabitant of both natural and artificial cavities in the Mediterranean region [40]. The species Pardosa cf. tatarica (Thorell, 1875) (Lycosidae) is closely associated with sandy places and streams edges [41]. Although Pardosa C. L. Koch, 1847 is a clearly epigean genus [41,42], the sampling results suggest that an appreciable population of P. cf. tatarica inhabits the deeper levels of the streambed, but are rarer on the surface since the watercourses remain completely dry for most of the year (unpublished data). Something similar has been observed with spiders of the genus Dysdera Latreille, 1804 (Dysderidae). Above ground, whether dwelling under rocks or in plant detritus [43], these spiders are extremely rare in this Mediterranean area. However, they are much more abundant in the substratum, having been observed not only in alluvial substrata but also occasionally in caves in the Eastern Prebetic range [44,45] and in many colluvial MSS in the study area (unpublished data). This tendency towards a certain level of hypogean activity coincides with the abundance of terrestrial Isopoda (suborder Oniscidea) in this type of habitat. The Myriapoda also had a strong presence in the substratum of the dry watercourses, with Diplopoda from the Polydesmida and, to a lesser extent, the Glomerida and Callipodida orders being particularly numerous. This behaviour may be due to the lucifugous and hygrophilous nature of these arthropods [46] which, similar to the Crustacea Isopoda, find the conditions of epigean environments more hostile. The Chilopoda appeared frequently in the alluvial substratum, among which we found specimens of the genus Lithobius Leach, 1814, which is also common in other subterranean spaces, with a well-documented presence in caves in the Iberian Peninsula [39]. Other Chilopoda captured in the alluvial substratum belonged to the genera Cryptops Leach, 1814, Scolopendra Linnaeus, 1758, and two very characteristic species of Western Mediterranean areas, Theatops erythrocephala (C.L. Koch, 1847) and Scutigera coleoptrata (Linnaeus, 1758). The Hexapoda presented a wide diversity of species in the alluvial substratum samples, and we observed specimens of the following orders: Collembola, Archaeognatha, Zygentoma, Diplura, Dyctioptera, Dermaptera, Orthoptera, Embioptera, Psocoptera, Hemiptera, Coleoptera, Diptera, Lepidoptera and Hymenoptera. Of these, two orders will be discussed in detail here, for which some interesting data were obtained. The Orthoptera contributed species from the genus Petaloptila Pantel, 1890 to the ensemble of fauna inhabiting the substratum of these watercourses. These Gryllidae exhibit troglophilic habits [39], which would explain their presence in these subterranean environments of fluvial origin as well as in other subterranean spaces in the region, such as caves [47]. The Coleoptera present a wide diversity of species and, as in other subterranean environments, the Carabidae are one of the most representative families of hypogean fauna in the Iberian Peninsula [48]. Of particular note was the presence of specimens of Trechus Clairville, 1806 belonging to the lineage T. martinezi (sensu Ortuño & Arillo, 2005), previously only described in a few cave environments [49] and on isolated occasions in the bedrock MSS of the Bernia mountain range in the province of Alicante [50]. We also found Porotachys bisulcatus (Nicolaï, 1822), a stenohygrobic and lucifugous species that in much of the Iberian Peninsula behaves as a true troglophile [51], and Ocys harpaloides (Audinet-Serville 1821), which is eurytopic and mesothermal, living in a variety of damp, dimly lit or completely dark habitats and capable of developing troglophilic habits [52]. In addition, we regularly collected numerous specimens of a Carabidae species of the genus Bembidion Latreille, 1802 which, in the dry watercourses in this geographic area, has not been observed either as a riparian sub-lapidicole or moving over the surface of the streambed. It is also worth noting the presence of Platyderus Stephens, 1827 in the alluvial substratum, a genus containing geophilic and lucifugous species with sub-lapidicole habits which also exhibit a clear tendency to occupy subterranean spaces [53]. The Leiodidae is another family that is well established in hypogean environments [22,54], and the genus Speonemadus Jeannel, 1922 was particularly numerous in the samples collected. In general, the species of this genus are known to inhabit subterranean environments, especially caves and the MSS [21]. Also striking was the recurrent presence of Lampyridae larvae (genera Nyctophila Olivier, 1884 and Lamprohiza Motschulsky, 1853) which, given their helicophagy [55] must have been on the tracks of the Mollusca Gastropoda that inhabit the damp interstices of the substratum. Although Lampyridae larvae have been observed in entrance zone of caves [30], their presence in the MSS has not been described for any Lampyridae species in the Iberian Peninsula [56,57].

Description of a New Subterranean Habitat: The “alluvial MSS”

For the most part, the fauna communities found in the alluvial substratum consisted of species belonging to four groups with well-defined ecophysiological characteristics, being geophilic, hygrophilous, mesothermal and/or lucifugous species. These species exhibit a geophilic nature, inhabiting environments closely associated with soil (edaphic horizons or different epiedaphic habitats and biotopes). They can be either totally geophilic, i.e. their entire life cycle occurs in these habitats/biotopes, or partially geophilic, in which only the larval stages inhabit this type of environment, as in the case of Lampyridae species of the genera Nyctophila and Lamprohiza, both frequently observed in the alluvial substratum. The second ecophysiological characteristic is the strongly hygrophilous nature of the species which were well represented in the alluvial substratum. In areas such as that studied, with a Mediterranean climate, these highly hygrophilous species are extremely rare or absent in edaphic/epiedaphic environments. The third characteristic is the mesothermal nature of many of these species, a circumstance that forces them to seek refuge in favourable environments which mitigate the sudden daily and/or seasonal changes in temperature. The last ecophysiological characteristic was the lucifugous nature which, however, only a portion of the ensemble of alluvial substratum fauna exhibited. This characteristic may be expressed to a lesser or greater degree, and thus some species may show a preference for shaded environments (sciophily) whilst others thrive in aphotic environments (a diversity of subterranean environments).

Different species presented different combinations of these characteristics. There were strongly hygrophilous, lucifugous and mesothermal species, all found in caves and other subterranean spaces, such as species belonging to the lineage Trechus martinezi, Lepthyphantes spp. and certain Isopoda Oniscidea, among others, which are considered troglobiont forms. Then there were strongly hygrophilous, mesothermal and sciophilic species (not exclusive to aphotic environments), such as Dysdera spp., Lithobius spp., Porotachys bisulcatus, Ocys harpaloides, Platyderus spp., Speonemadus spp. and some Oniscidea Isopoda, among others, which constitute an ensemble of species which are typically troglophilic. There were also moderately hygrophilous, mesothermal and sciophilic species which may sometimes behave like true troglophiles, such as Dicranolasma soerensenii and Petaloptila spp., and lastly, there were species which were markedly hygrophilous, may or may not be mesothermal and were not lucifugous, as is the case of fauna exhibiting ripicolous behaviour such as Pardosa spp. and some of the Carabidae Bembidiinae species, among others.

Thus, as happens in caves, the alluvial substratum constitutes a habitat occupied by fauna which is exclusively hypogean (troglobionts), frequently hypogean (troglophiles), and where numerous edaphic species occur (Acari and Collembola, among others) to a different degree, in addition to ripicolous and trogloxene fauna. The latter group consisted of species with a small number of individuals belonging to different orders (Diptera and Hymenoptera, among others), for which living conditions in the substratum are not optimal. This group of specimens reached the substratum by accident, either actively, as part of their daily scavenging activities in the edaphic/epiedaphic environment, or passively, usually as a result of hydrochory and also zoochory. Despite being considered species which are not adapted to this environment, they make a vital contribution to the flow of energy in these subterranean habitats [11,58]. Therefore, there is fauna with different ecological roles: some of them shared with caves and other MSSs (troglobionts and trogloxenes), which characterizes this habitat as a true MSS; some others are typical of soil (edaphobionts), which are also frequent in other MSS [59]; epigean fauna (trogloxenes), due to its proximity to the surface and the conditions of temperature and humidity; and finally ripicolous fauna, whose presence would be particularly characteristic of this habitat.

We propose naming these subterranean spaces located in the alluvial substratum “alluvial MSS”, a hypogean habitat composed of cracks and crevices in the substratum that form among the gravel and variously sized pebbles that constitute the alluvial deposits of dry watercourses (Figure 4). To date, all approaches to the study and recognition of fauna in the substratum of watercourses have focused on the hyporheic zone and, collaterally, with certain adjacent biotopes, some of which might fall under the classification of MSS. The hyporheic zone [60,61] is limited to the substratum of active watercourses, through the interstices of which water circulates and provides a habitat for aquatic species, many of which can be described as stygobionts. In addition, the dynamics of the recent past of these rivers has led to the formation of river terraces beneath which groundwater circulates, constituting aquifers (phreatic biotopes) that may be in contact with the waters of rivers and streams; this forms another type of subterranean water environment [61,] where invertebrate stygobionts are also frequently present. With regard to the MSS that is subject to a fluvial influence and whose origin is unrelated to the dynamics of river currents, Uéno [62] revealed the importance of colluvial deposits and slopes that are more or less directly in contact with the waters of rivers with a moderate current.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Alluvial MSS.

(A) Dry watercourse (Barranc de la Vall de Gallinera). (B, C, D) structure of the substratum.

https://doi.org/10.1371/journal.pone.0076311.g004

The “alluvial MSS” appears in dry watercourses, which appear as scars on the terrain, through which water only flows in times of flood, creating torrential type hydrological regimes. This hydrological dynamic is associated with the Mediterranean climate characteristic of the area, in which maximum rainfall occurs in autumn with a smaller peak in spring, whilst the summer is characterised by a very marked minimum rainfall accompanied by an increase in temperature [24]. Floods occur after irregular, heavy rainfall which starts suddenly and sometimes violently [23,]. However, there are many types of flood, of a diverse nature and intensity [64]. It is of interest to note that those known as flash floods [65] whilst not the only type to occur in the areas studied, have a very significant effect since they can even push solid blocks of a diameter of up to twice the height of the water [24,]. These phenomena may be due to different causes, such as summer convective storms, weather fronts and orographic precipitation [67], and often also to the phenomenon known as “cut-off low” [68-70], responsible for numerous catastrophic episodes of flooding.

Since the watercourses in these areas are dry for most of the year, the fissures in the substratum are not permanently waterlogged. However, they retained sufficient water even when water levels were at their lowest for the stones removed during the installation of the SSDs to be damp (unpublished data), and RH_mean levels of 80% and RH_min levels of 46.5% were recorded for some of the SSDs (Table 1, Barranco de la Cueva de los Corrales). This circumstance facilitates the survival of plants growing in the streambed, provides a refuge for hygrophilous terrestrial fauna and supports the presence of both exclusively and frequently hypogean species, rendering this set of spaces among the stones of alluvial debris a new type of MSS.

Formation and structure of the “alluvial MSS”

No single factor determines the presence of temporary watercourses in the Eastern Iberian Peninsula. In other regions of the Mediterranean, the presence of temporary watercourses has been described as the result of a synergy of geological, geomorphological and climatic factors [71]. However, in the case of the study area, there is no clear relationship between the presence of such courses and a given geological substrate or a particular rainfall pattern [23]. Furthermore, it should be noted that not all of the Eastern Iberian watercourses are subject to a seasonal regime, as some have a constant flow of water throughout the annual hydrological cycle, either along the entire course or only a part of it, in which groundwater plays an important role in maintaining a base flow [72]. It should also be noted that the presence of dry watercourses is not a feature unique to the Iberian Peninsula. Rather, they appear in numerous regions of the world with very different climates and geology, receiving different names: the “Chapp” in the Gobi, “Laagate” in the Kalahari, “Wadi” in the Maghreb, “Donga” in South Africa, “Nullah” in India, “Fiumare” in Italy, “Arroyo” in the U.S.A. and Mexico, and “Dry Valley” in England [24,].

With regard to the dry watercourses studied, these are known to have a considerable carrying capacity and thus it is very common to see large boulders, pebbles, gravel and sand on the streambed, forming a very heterogeneous group of rocks from different sources [72]. The loss of the water’s inertia results in the rapid and massive sedimentation of the transported materials [74]. This circumstance converts the lower reaches of these dry watercourses (the low, wide channel known as the “rambla”) into areas where “alluvial MSS” is unlikely to exist because the interstices between the boulders are generally filled (Figure 5a, b). Our field work revealed the existence of “alluvial MSS” in the upper and middle reaches of these watercourses, where the alluvial debris formed by calcareous stones contained a network of empty spaces which had not been filled by sediments, allowing the movement of fauna between these interstices (Figure 6). However, we cannot assert that dry watercourses in different regions of the world behave in the same way.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 5. Watercourse with colmated river-bed.

(A, B) Rambla de la Torre (Sax, Alicante). Sampled watercouse sections at different moments in time: (C, D) Barranc de la Vall de Gallinera (Subterranean Sampling Device 15); (E, F) Barranc de Famorca (SSD 6).

https://doi.org/10.1371/journal.pone.0076311.g005

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 6. Schematic graph illustrating the connections between hypogean environments, including the “alluvial MSS”.

By way of example, three groups of arthropods are shown (Trechus: Carabidae, Dysdera: Dysderidae, Petaloptila: Gryllidae), present in all the subterranean environments in the area (caves, colluvial MSS and “alluvial MSS”).

https://doi.org/10.1371/journal.pone.0076311.g006

Although the “alluvial MSS” is not covered by true soil, it is noteworthy that it may have some degree of insulation from the surface as a result of one phenomenon which occurs with relative frequency in these watercourses. This is the phenomenon known as “armouring” [24,], which occurs when flooding ceases and the carrying capacity of the watercourse decreases. At this point, the coarser materials are deposited, forming protective films or “armour”, and the finest materials filter between and rearrange the interstices. Hence, a thin surface layer of silt, sand and gravel is created between the larger stones close to the surface, which contributes to insulating the MSS to a certain extent (Figure 4b, c, d).

In addition, the vegetation that grows not only on the margins of these dry watercourses but also in their channels, plays an important geomorphological role, proffering mechanical resistance to the water and anchoring the alluvial deposits with their roots. Hence, they provide the “alluvial MSS” with greater structural stability against the drag force of torrential waters and favour sedimentation of the particles carried by the water, thus helping to increase the thickness of these deposits [76]. The dry watercourses studied were mostly colonised by oleander formations, shrubs which are characteristic of dry Mediterranean environments and are adapted to withstand both hot drought and the force of floodwaters. Although the oleander (Nerium oleander L.) predominated in these formations, it was often accompanied by various thorny bushes (mostly of the genera Rubus L, Rosa L. and Crataegus L.) and by nitrophilous plants whose growth was favoured by the contribution of organic detritus periodically deposited by floods. When the floods are not very violent or very frequent, it is common for other plants from surrounding plant communities (forest species, mainly) to appear in these streambeds.

The thickness of the alluvial deposits was highly variable, as was the size of the rock fragments of which they were composed, both characteristics being related to the slope and, in general, the orographic and lithological configuration of the site. In some cases, the “alluvial MSS” was no more than 50 cm deep before hitting the bedrock, whilst in other cases its depth exceeded one or one and a half metres (Figure 4).

As already mentioned, the loss of many of the data loggers makes it impossible to generalise the results of the temperature and humidity parameters. Furthermore, the measurements taken corresponded to a depth of between approximately 20 and 30 cm, and it is well known that the physical conditions of temperature and humidity in the substratum change rapidly the further down one goes [77]. However, we can conclude that the “alluvial MSS” is an extremely damp environment, with a RH which is almost constantly in the vicinity of saturation levels. It should be mentioned, however, that the plastic hood in which the data loggers were placed may have produced an overestimation of the RH due to the condensation of water inside. Compared with the external environment, the “alluvial MSS” is very stable as regards temperature [59]. Although the temperature in the MSS showed a pattern (cycle) similar to that of the external environment [59,], values for the mean daily variations in temperature above ground ranged between 10.79-14.78°C, whereas at 20-30 cm deep in the MSS, they ranged between 1.12 and 4.96°C, representing an average of 10°C lower (Figure 3 and Table 2). This is not surprising if we consider that below 50 cm in the substratum, daily temperature variations are usually practically negligible [77], although it may be assumed that in MSS without soil cover, such as the “alluvial MSS”, this depth must be greater [78]. The difference in values between the external environment and the MSS for T_max and T_min was around 15°C and 5°C on average, respectively. These data show that although there was an appreciable temperature cycle, the MSS had a more modest variation range than the external environment, and suggest that the MSS was a more effective buffer against maximum temperatures than against minimum temperatures [79,].

Therefore, the “alluvial MSS”, is not only similar in structure to other MSSs (a network of subterranean spaces among rocks and stones), but also in its degree of isolation from surface. However, it has some particularities that differentiate it from other MSSs. The origin of the deposit of stones is alluvial, and the isolation from the surface is not due to a soil layer, but to a layer of silt, sand and gravel (Figure 4b, c, d), and it may eventually have disturbances due to floods of variable intensity.

Ecological importance of the “alluvial MSS”

It is well known that plant and animal species associated with temporary watercourses in arid and semi-arid regions have developed adaptive strategies that allow them to survive water stress [81,]. Watercourses subject to natural spatial and temporal environmental disturbances have been described by Margalef [82] as dynamic evolutionary elements that favour the dynamic processes of species colonisation and expansion [,]. In this paper, we show that this happens not only on the surface and banks of the fluvial network (which includes the edaphic environment), but also in the hypogean environment closely associated with alluvial deposits. In line with the above, the alluvial plains subject to periodic flooding are increasingly being studied as a whole, rather than from an exclusively terrestrial or aquatic perspective [87] As with the “alluvial MSS”, this periodic flooding constitutes a “pulse” that disrupts the dynamic equilibrium these systems had reached before the flood. It is precisely because of these disturbances that it seems difficult to conceive of the existence of terrestrial fauna living in the “alluvial MSS” when, periodically, these channels are completely flooded (Figure 5c-f). However, these apparently catastrophic episodes do not appear to represent an obstacle to the survival of subterranean communities, whose members show very different degrees of adaptation to hypogean life. The proof of this lies in the terrestrial life forms collected in the MSS only days after heavy rainfall. The high speed at which the water flows in the upper and middle sections of these channels may facilitate the retention of large pockets of air within the spaces inside the “alluvial MSS”, and thus, the survival of the terrestrial fauna. The high carrying capacity of these watercourses suggests the possibility that hydrochory events could occur, “seeding specimens” in other sections of the channel suitable for hypogean life [88].

The fauna results obtained show that despite the structural instability of this MSS, the underground spaces in the area studied constituted a habitat that was sufficiently suitable for colonisation both by exclusively hypogean fauna (species only known caves in the area) and frequently hypogean fauna (“troglophile” species sensu Sket [89]). This habitat also acts as a “refuge” for epiedaphic, hygrophilous fauna that are capable of occupying these spaces when necessary to shelter from the seasonal fluctuations in temperature and humidity. Therefore, this subterranean environment must be considered an important habitat due to the role it plays in the conservation of fauna. Furthermore, it represents a new source of information for determining the real biodiversity of sites which, until now, had only been explored above ground and provides data which contribute to improving knowledge of the real distribution of hypogean and epigean species. Lastly, the possibility must also be raised that this new type of MSS may act as an ecological corridor for both hypogean and epiedaphic, hygrophilous fauna.

Synthesis of characteristics of the “alluvial MSS”

The “alluvial MSS” is a type of terrestrial shallow subterranean habitat that consists of a network of spaces which forms among the alluvial deposits of temporary watercourses, and which may be covered or not by a more or less evolved soil on which some herbaceous or woody plants may manage to grow.