Understanding habitat extension that limits species distribution is a crucial tool for management and conservation, in which habitat mapping plays a pivotal role. The mesovoid shallow substratum (MSS) is a type of shallow subterranean habitat with an important conservation value for invertebrate communities, functioning as climatic/reproductive refuge, biogeographic corridor and/or permanent habitat. Methodologies to map the mesovoid shallow substratum (MSS) are currently lacking. We propose a novel method for colluvial MSS habitat mapping, combining geographic information systems, geological maps, and geological knowledge on the habitat genesis. We tested and validated the efficiency of the method using the Arrábida karst area (Portugal) as a model. The method allowed the remote detection of MSS habitats suitable for invertebrate communities ex situ within the study area, and enabled the estimation of habitat extent. The faunal communities sampled in the selected location were dominated by arthropods, especially insects, showcasing the efficacy of this mapping method to detect suitable MSS habitats. The use of this method considerably reduces the in situ scouting area, providing a more efficient way of locating these habitats. The MSS is protected under EU legislation concerning floral communities and geological features, completely neglecting its faunal communities. This method also allows to estimate potential MSS habitat extension in several lithologies, facilitating the implementation of invertebrate prospections, and the establishment of more effective conservation measures.

Arrábida Natural Park, habitat mapping, scree slopes, shallow subterranean habitats, subterranean ecosystems, Portugal

The Mesovoid Shallow Substratum (or Stratum) was described in French as the “Milieu Souterrain Superficiel” (MSS) by Juberthie et al. (1980) in the Pyrenees, as an extensive type of shallow subterranean habitat, “formed by a network of air-filled voids and open spaces in between rock fragments” (for recent review on terminology see Mammola et al. 2016). Like caves, the MSS is characterized by total darkness, stable temperatures, and high humidity levels (Culver and Pipan 2014; Mammola et al. 2016; Nitzu et al. 2018).

Colluvial MSS, one of the four recognized types of MSS (Mammola et al. 2016), is typically composed of homogeneous rock material (depending on the type of bedrock) in various and heterogeneous sizes. The colluvial material consists of unconsolidated free sediments that have been deposited on hillslopes by either rain-wash, slow continuous downslope creep, or a combination of both these processes (Juberthie et al. 1980; Jiménez-Valverde et al. 2015). This type of MSS, when exposed is referred to as a scree slope, but can also be covered by a layer of soil and vegetation (Culver and Pipan 2009).

The MSS can be inhabited by organisms with different adaptations to life in subterranean habitats, from surface ones living in the soil (edaphic), to species with some affinities to the subterranean habitats (troglophile), to animals fully adapted and dependant on the subterranean habitat to survive (troglobiont) (Ortuño et al. 2013; Culver and Pipan 2014; Nitzu et al. 2014). It has been studied in several countries, mainly for its faunal communities, which are dominated by insects and other arthropods (Pipan et al. 2010; Rendoš et al. 2012; Ortuño et al. 2013; Nitzu et al. 2014; Jiménez-Valverde et al. 2015; Eusébio et al. 2021).

This habitat is an ecotone between surface and deeper subterranean habitats (Gers 1998), and can serve three very distinct, but equally important functions: 1) climatic/reproductive refuge for surface species, as in this habitat temperature variations along the year are much less extreme than at surface (Nitzu et al. 2011, 2014, 2018; Mammola et al. 2016); 2) biogeographic corridor to deeper ecosystems, as in many cases the MSS is connected to caves by a system of fissures, especially in karstic landscapes (Ortuño et al. 2013; Jiménez-Valverde et al. 2015); 3) permanent habitat for some species (Culver and Pipan 2009; Pipan et al. 2010; Eusébio et al. 2021).

Habitat mapping is a critical step for the management and conservation of habitats and their faunal communities (McDermid et al. 2005), as it allows an estimation of species distribution and richness patterns (Lecours et al. 2015), leading to the development of better monitoring protocols (Cogan et al. 2009; Lecours et al. 2015). Regarding subterranean habitats, improving habitat mapping methods will aid answering two of the fundamental research questions on subterranean biology: “What is the species richness pattern of subterranean organisms globally?” and “What would be the best monitoring protocols to quantify long-term changes in the distribution and abundance of subterranean invertebrates?” (Mammola et al. 2020).

Subterranean habitats, “the most widespread non-marine environments on Earth” (Mammola et al. 2019), are particularly difficult to map due to their challenging access for humans (Mammola et al. 2019; Tanalgo et al. 2022). Caves have been mapped throughout the world since the beginning of speleological studies (Martel 1894), evolving into digital mapping (Trimmis 2018; Jones 2022; Kampolis et al. 2022), which include robots (Chang et al. 2022; Tabib et al. 2022) and remote sensing from space (Sharma and Srivastava 2022), using terrestrial lava tubes as models (Bell Jr. et al. 2022). On the other hand, the network of spaces and voids that composes the majority of subterranean habitats, which is inaccessible to humans, has never been mapped (Mammola et al. 2019).

This inaccessible underground network is the habitat that most subterranean species occupy, placing them “among the least documented fauna on our planet” (Mammola et al. 2019). Despite the different MSS types being properly described (Mammola et al. 2016), the worldwide extent of this habitat is currently unknown, with no available mapping methods proposed so far. Currently, its prospection relies solely on in situ visual verifications.

We propose a methodology to locate and estimate the areas of colluvial MSS habitats, using a combination of geographic information systems, geological maps, and geological knowledge on the genesis of this habitat. The Arrábida massif in Portugal was selected as the study area to develop, test, and validate the mapping method. The work conducted was divided into six steps: 1) defining a geological area to map, 2) defining main geological criteria for habitat genesis, 3) selecting potential locations using satellite imagery, 4) combining the potential sites with geological maps, in order to apply the criteria, 5) performing an in situ verification of the sites that met all the criteria, 6) assessing biodiversity in order to verify if the habitats selected by the method are suitable for typical MSS invertebrate communities.

This new habitat mapping method, combining geographic information systems with geological maps and information on MSS genesis, allows tracking and mapping colluvial MSS habitats, improving the current habitat location technique, which is based on in situ verification.

This method also allows for the estimation of habitat distribution, and when applied to faunal studies allows the estimation of species distribution areas. This knowledge will in turn contribute to improve the establishment of protection and conservation measures for this habitat and its faunal communities.

The application of this method requires: 1) deciding whether the formation of the scree slope is natural or anthropogenic, as this criteria can be biased; 2) selecting appropriate rock types within the lithology of the study site, as long as they allow the formation of scree slopes (this includes non-karst areas) and 3) taking into account vegetation coverage of the area, which can mask scree slopes in the in situ verification stage (Rendoš et al. 2014; Rudy et al. 2018; Jakšová et al. 2019). One method to overcome this last complication is to add an updated vegetation layer to the mapping step, when available, in order to better predict these events. Other approaches can involve long term surveys, as vegetation coverage varies over time, especially in areas prone to wildfires (Bastos et al. 2011).

The initial scan of the study area using satellite imagery resulted in 24 locations for potential MSS habitats. With the use of geographic information systems and geological maps this number was reduced to eleven. After the visual in situ verification, eight locations were confirmed as the intended habitat, i.e., bare scree slopes. Our method has, therefore, proven to be effective in locating potential colluvial MSS habitats with around 75% accuracy after the in situ verification, and also considerably reducing the scouting area in the field.

Our results suggest that the sole use of GIS software to locate MSS habitats is less effective than our mapping method. This can be attributed to the lack of satellite image updating on certain areas of the country, especially natural areas, where vegetation grows fast, drastically changing the landscape in a short time (Clemente et al. 1996; Meira-Neto et al. 2011). The discrepancy between the locations meeting all the criteria, and the ones detected in the field can be explained by vegetation cover. Therefore, the “missing” locations are in fact colluvial MSS habitats covered by the thick vegetation found all over the area of the natural park, mainly Quercus coccifera shrubs (Parque Natural da Arrábida 2000).

The in situ verification for the presence of colluvial MSS in these locations was prevented due to the extremely dense vegetation in the field, which would require cutting down natural vegetation patches. MSS covered with soil and vegetation is common in the margins of bare scree slopes, which are gradually covered by mature soils, leading to more extensive plant coverage over time (Mammola et al. 2016; Giupponi et al. 2023). Therefore, our method also has the potential to detect “hidden” MSS areas, which are covered by soil and vegetation, continuous to the detected bare scree slopes.

In fact, the lithology of the area, previous observations, especially historical events related to wildfires (Clemente et al. 1996; Cruz and Viegas 1997), and knowledge of the natural park landscape (Parque Natural da Arrábida 2000) indicate the presence of MSS covered by soil and vegetation in those “missing” locations.

The mapping method also proved to be efficient in locating habitats suitable for invertebrate fauna. Scree slope habitats can host both surface and subterranean-adapted species (Pipan et al 2010; Nitzu et al. 2011, 2014; Jiménez-Valverde et al. 2015). The communities found in the sampled location have a similar composition to communities typically found in the MSS in other European countries, i.e., the absolute dominance of arthropods, more specifically insects (Rendoš et al. 2012; Ortuño et al. 2013; Jiménez-Valverde et al. 2015; Eusébio et al. 2021), and few troglobiont species (Pipan et al. 2010; Jiménez-Valverde et al. 2015). However, we found an unusual pattern, the clear dominance in number of individuals of Diptera over Collembola. This can be linked with the lower altitude of our study (300 m a.sl.), in contrast with the high mountain location of the previous studies (Rendoš et al. 2012; Jiménez-Valverde et al. 2015), as the abundance of several Diptera are known to be affected by elevational gradients and other climatic variables (Hodkinson 2004; Rhoner et al. 2015), but also by the lack of soil and vegetation coverage of bare scree slopes, that may facilitate the colonization of this habitat by flying insects.

This method allows a habitat extent estimation, which in the case of the colluvial MSS within the Arrábida Natural Park is approximately 30,500 m². This value corresponds to the surface area of habitat. However, like in other habitats, such as coral reefs (Dustan et al. 2013), the real extension of the MSS habitat is higher than its surface area, due to its three-dimensionality. Future studies to quantify the total MSS 3D area and volume will have to consider habitat depth and slope, rock fragment size, and rock occupied volume versus air filled spaces. It is also important to consider the area of the habitat that is covered by vegetation.

In Portugal, MSS habitats are protected by EU legislation in the Habitat Directive as “Western Mediterranean and thermophilic deposits”, in the “Limestone scree slopes (8130pt1)” subcategory, regarding plant communities and geological features (ICNB 2000). However, this protection completely disregards faunal communities, and, in a recent study, the Portuguese MSS was considered a habitat with high conservation value for invertebrate communities (Eusébio et al. 2021).

The use of this mapping method might facilitate the detection of MSS, which can later be studied and contribute to overcome the main shortfalls of subterranean biodiversity (Mammola et al. 2019) in shallow subterranean habitats. It will enable the increase of knowledge about habitat distribution (Racovitzan shortfall), which will facilitate new faunal community studies (Mammola et al. 2019).

This new mapping method has the potential to be used worldwide, paying special attention to integrate topography, geomorphology, lithology and tectonic structures of the studied location, improving work efficiency for studies that deal with colluvial MSS, a more and more important habitat in the context of global changes as a climatic/reproductive refuge (Eusébio et al. 2021).

The possibility to survey the area of shallow subterranean habitats worldwide will help define conservation perimeters, and estimate the extent of occurrence (EOO) and area of occupancy (AOO) according to the IUCN Red List criteria (IUCN 2021) for species that inhabit shallow subterranean habitats. Moreover, our method allows for the estimation of the extension of this type of habitat around the world, which will certainly contribute to establish more effective conservation measures for land-use at surface and for faunal conservation.

This work was supported by the Portuguese National Funds through “Fundação para a Ciência e a Tecnologia” (FCT) within the cE3c Unit funding UIDB/00329/2020, CESAM (UIDP/50017/2020, UIDB/50017/2020, LA/P/0094/2020), PhD grant (2021.04868.BD), and by the VILLUM FONDEN (research grant 15471). All specimens were collected under the Instituto de Conservação da Natureza e das Florestas (ICNF). We are also grateful to the Arrábida Natural Park (PNA) for kindly providing accommodation, and to the park rangers for logistic support during fieldwork.