Rediscovery of the endangered species Harpalus flavescens (Coleoptera: Carabidae) in the Loire River

The Loire River is one of the last European large rivers with important sediment dynamics and numerous sandbanks. The extraction of sediment from the riverbed during decades and the construction of levees for flood prevention have strongly affected and shaped the biodiversity of the Loire River. Many species from pioneer riverbanks have been impacted with particular consequences for psammophilous insects. The ground beetle Harpalus (Acardystus) flavescens (Piller & Mitterpacher, 1783), is considered to have disappeared from the Middle Loire River for 40 years and is endangered everywhere in Europe. In 2012 and 2013, we recorded two specimens of H. flavescens in Région Centre‐Val de Loire (France), in the course of a survey dedicated to evaluating the impact of fluvial maintenance operations upon sediment and biodiversity dynamics. The presence of H. flavescens may be linked to the interruption of riverbed extractions and the vegetation removal of sandbanks of the Loire River (ecosystem restoration).


Introduction
Alluvial rivers have historically been an attractive source of sediment for economic development activities. But sediment extractions have produced many detrimental effects, including channel incision (sinking of riverbed), loss of riparian habitats and several other ecological and environmental impacts (Rinaldi et al., 2005). With a length of 1012 km, the Loire River is the largest river in France and one of the last wild rivers of Europe (Bastien et al., 2009). The path of the Loire and some of its tributaries often change due to an important morpho-sedimentary dynamics (Claude et al., 2014;Rodrigues et al., 2015). This leads to the transportation and natural accumulation of sediments and the formation of sedimentary bars (sandbars), which constitute one of the Loire River characteristics. These formations have a peculiar dynamics due to the hydraulic characteristics (depth, speed) and shape of the channel (Wintenberger et al., 2015a). Some of these bars are in stable position (forced bars) while others are very mobile in flood period (free bars). Free and forced bars can stabilise and support pioneer vegetation that can evolve to woody stages, where black poplar is prevalent (Wintenberger et al., 2015b), or to more diversified woodland (island stage). Early stages of the succession correspond to pioneer sandy-gravel habitats and constitute a reservoir for a large number of species associated to sand. These species are referred to as 'arenicolous' (Torre-Bueno & Nichols, 1989;Hanson, 2007), 'sabulicole' (Hanson, 2007) or 'psammophilic/psammophilous' (Lewis, 1977;Thiele, 1977;Torre-Bueno & Nichols, 1989). Since the 1950s, the massive gravel extraction from the Loire riverbeds (e.g. 6.4 million tonnes in 1979: see Gasowski, 1994) has caused a significant imbalance between the amount of sediment extracted and the quantity naturally replenished by natural deposition (Claude, 2012). Coupled with containment, it has caused the incision of the Loire River and the lowering of the water line during low-water periods. These events have led to the disconnection of hydraulic attachments (backwaters, islands, sedimentary bars, etc.) and more generally to the decrease in the active bank of the river (Latapie et al., 2014). In addition, the progressive cessation of agropastoralism and bank maintenances and the reduction in exceptional winter floods have led to the gradual plant colonisation of pioneer environments (Grivel, 2008).
At the same time, rivers and adjacent terrestrial habitats are impacted by eutrophication due to agricultural intensification and rising urban pressures (Smith et al., 1999). This has led to the modification of the composition of the flora (Walker & Preston, 2006) and both herbivorous and predatory insects (Haddad et al., 2000). In this context, the Loire River has been considered eutrophic since at least 1980 (Minaudo et al., 2013).
Both these phenomena have contributed to the decline of pioneer sedimentary bars and to the modification of the species communities associated with these environments (e.g. loss of species richness, overgrowth of ordinary species, speeding up of the vegetation succession).
The ground beetle Harpalus (Acardystus) flavescens (Piller & Mitterpacher, 1783) ( Fig. 1a) is an emblematic species of pioneer sand habitats such as continental aeolian sand dunes, river sandbanks, coastal dunes, or sandpits. It is a steno-xerophilous species (Luka et al., 2009) that lives deep in the sand, especially among roots of various grasses, such as Corynephorus, Psamma and Panicum (Lindroth, 1992). Adults appear and breed mostly in autumn. H. flavescens is a macropterous species (Hurka, 1996) but no flight observations have been made (Lindroth, 1992). Larvae live in sand, hibernate before the winter and complete their development during the next spring. Hence, sand exploitation has a direct impact on population dynamics particularly in riverbeds where extractions are conducted in autumn.
Due to its ecological specificities, H. flavescens is one of the endangered species that are strongly affected by the loss of sand areas and the degradation of its habitats, particularly in the rivers of Central and Eastern Europe (Boh a c & Jahnova, 2015). Although H. flavescens used to be widely distributed in Europe (Homburg et al., 2013), it has declined and has become very rare in some regions (Kugler et al., 2008). Lindroth (1992) indicated that H. flavescens is very rare in Sweden, absent in Norway and Russian sector, and probably absent in Norway. In Slovak and Czech Republic, it is rare and localised (Hurka, 1996). In Denmark, H. flavescens has not been recorded since 1850 (Hansen & Jorum, 2014), in Belgium it is seriously threatened (Belgian Species List, 2015), in Switzerland and Germany it is an endangered species (Trautner et al., 2005;Luka et al., 2009), and it has been observed only in some northern parts of Italy (Allegro & Sciaky, 2001). In France, the species was historically present in the Tertiary sands of the Paris Basin, in the Loire and Allier basins (Jeannel, 1942;Bonadona, 1971;Velle, 2004), the Seille Valley, in the Bresse and Lyonnais (Sainte-Claire Deville, 1935). It has also been observed in Alsace (Callot & Schott, 1993), Puy-de-Dôme and Pyr en ees (Tronquet, 2014) and Picardie, Nord, Hautes-Pyr en ees and Landes (Valemberg, 1997). H. flavescens is declining (Coulon et al., 2011) or endangered (Tronquet, 2014) in all areas where it exists. In the Loire Valley, H. flavescens has been considered as rare (Favarcq, 1876). In Loiret department, it was reported around Gien (Victor Pyot collection) and Orl eans (Henry-Pierre Sainjon collection) until the early 20th century (Secchi et al., 2009). Currently, H. flavescens is considered to be extinct in R egion Centre-Val de Loire (Binon et al., 2012).
The data reported in the present paper were collected in the course of a study of pioneer sediment dynamics and the associated biodiversity conducted in the Middle Loire between 2012 and 2015 (Villar, 2015;Wintenberger et al., 2015a,b). Ground beetles constitute an ideal biological model to assess the dynamics of these habitats as the group is diversified with 1000 species in France (Coulon et al., 2000) and its taxonomy well known (Coulon et al., 2011). Ground beetles inhabit diversified habitats, their ecology is well documented (Thiele, 1977;Valemberg, 1997;Desender et al., 2010) and robust sampling methods are available (Work et al., 2002). They are good biological indicators of soil or sediment, humidity, coverage, density and type of vegetation, trophic level (Rainio & Niemela, 2003;Paillet, 2007;Lambeets et al., 2008Lambeets et al., , 2009) and their disruption (Avgn & Luff, 2010;Kotze et al., 2011). Many ground beetles are capable of flying and can disperse widely (Thiele, 1977) and are thus adapted to rivers with natural dynamics. As such, ground beetles have been used in numerous biodiversity studies in forest (Magura et al., 2001;Richard, 2004), in poplar plantations (Allegro & Sciaky, 2003;Denux et al., 2007;Elek et al., 2010), in agricultural areas (Liu et al., 2010;Sonoda et al., 2011;Holland et al., 2012), and in alluvial areas (Lambeets et al., 2008;Januschke et al., 2011). In alluvial areas, sand and gravel sediment bars are important for ground beetles (Lachat et al., 2001), where they are good bio-indicators for the management and restoration of river ecosystems (Gerisch et al., 2006;Januschke et al., 2011;Januschke & Verdonschot, 2016) and hydrological conditions (Gerisch et al., 2006). Finally, they colonise all pioneer riparian habitats, including nude sediment bars where few organisms are usually present. A large sampling campaign was carried out in various habitats of the R egion Centre-Val de Loire and offered a unique opportunity to re-examine the distribution of rare or extinct ground beetle species such as Harpalus flavescens.

Materials and methods
The study was conducted in France, in the R egion Centre-Val de Loire on four sites of Loire River between Châteauneuf-sur-Loire (45), upstream, and Blois (41), downstream (see Fig. S1), with a more important sampling effort on the islands of National Nature Reserve of Saint-Mesmin (see Fig. S1, sector b [47°51 0 54.0″N, 001°46 0 56.3″E]). We studied other Loire River sections, located upstream in the areas of Sandillon and Saint-Denis-en-Val (see Fig. S1 Ground beetles species abundance was surveyed from 2012 to 2014 in five pioneer habitats of the Middle Loire River: sandy formations, gravelly formations, mudflats and cracked soil, grassland and poplar coppice. Ground beetles were sampled using two methods. We used pitfall traps filled one-third full by 20% Mono-Propylene Glycol (also called Propanediol) because it is the most common technique to sample ground-moving species (Work et al., 2002). Mono-Propylene Glycol, combined with saturation of salt and a wetting agent (to decrease the surface tension of the water and thus lead to the drowning of the insects) allows a very good conservation of the specimens between two sampling periods (2 weeks lag). It also limits the evaporation of the liquid in the pitfall trap (Denux, 2005) which is an important issue when sampling in xero-thermophilic habitats where sunshine and heat can dry out a pitfall trap in a few days.
Among riparian and mudflat ground beetles, very common along the Loire River, many species are generally difficult to catch in the pitfall trap (e.g. small species). Quadrat samples were therefore used to complete the pitfall traps and to optimise our sampling (Bigot & Gautier, 1981;Dajoz, 2002). The quadrat samples were used with a square metal quadrat (0.25 m 2 ) applied to the soil to limit invertebrate escape (Andersen, 1995). Individuals were caught following three complementary actions: (i) visible individuals were immediately caught, if necessary using a mouth aspirator equipped with a recipient, (ii) The surface of the quadrat was watered to cause a phenomenon of escape of individuals, which were immediately captured, and (iii) Stones, gravel and debris were removed to extract hidden individuals.
Pitfall traps were used from early July to late September, between 2012 and 2014, and checked every 2 weeks. The sampling by quadrat was carried out when the climatic conditions (heat and sunshine) were favourable: on 31 July 2012, 08 August 2012, 14 July 2013, 02 September 2013, 15 September 2013 and 03 September 2014.

Results and discussion
During the study period, 754 pitfall trap samples and 223 quadrat samples were collected for a total of 99 species and 8743 individuals (pitfall traps: 97 species and 8055 individuals, quadrats: 29 species and 679 individuals). Among these specimens, two females of H. flavescens were collected with pitfall traps. No individuals of H. flavescens were captured with quadrats, in spite of the removing of stones, gravel and debris in the upper soil layer. In most cases of active capture of H. flavescens, individuals were hidden below stones and debris.
The first individual was captured on 03 September 2012 on an island of the National Nature Reserve of Saint-Mesmin (Mareau-aux-Pr es, 45; Fig. 1b). The pitfall trap was located on a dry sand/gravel substrate with little vegetation (about 10% herbaceous cover) and at 10 metres from the river (see Fig. S2a). In the pitfall trap, H. flavescens was found with two other Carabidae, Amara fulva (M€ uller, 1776) and Lionychus quadrillum (Duftschmid, 1812).
The relative abundance of H. flavescens was 0.025%, for all pitfall traps and habitats sampled in our study. Considering the ecological criteria that appear to be the most suitable to this species (Table 1) and filtering our pitfall trap sampling data on these criteria (≥50% sand; 'dry vegetation' ≤50%, sampling period between early August and late September), the relative abundance of H. flavescens in favourable pitfall trap samples was 0.4% (for 76 pitfall trap samples).
Harpalus flavescens had not been observed for at least 40 years in the R egion Centre-Val de Loire and had been considered extinct (Binon et al., 2012). Our study provides a proof of the recent presence of H. flavescens in the Middle Loire. Hence, it will be necessary to change the protected status of this species in the R egion Centre-Val de Loire, from disappeared species to threatened species . Several hypotheses may explain the rediscovery of this species in the Middle Loire.
(1) H. flavescens is a rare and autumnal species (Luka et al., 2009), and inhabits habitats that are rarely studied by entomologists. This could explain the low probabilities of capture, without specific studies of this species. Our important sampling effort (754 pitfall trap samples and 223 samples per quadrat for a total of 8743 ground beetles identified) may explain the rediscovery of this rare species erroneously considered to be extinct in the R egion Centre-Val de Loire. Other studies, however, including beetles were carried out in the Middle Loire, particularly in the National Nature Reserve of Saint-Mesmin, where two surveys have been conducted: Pratz and Roger (1998) used active insect collection and Jaulin (2004) with 166 samples and several methods (traps and actively collect insects). In a legislative framework, a number of inventories were also conducted by entomologists in this National Nature Reserve: in the period 1997-2016, 482 species of beetles including 70 ground beetles species were identified (extraction from the National Nature Reserve of Saint-Mesmin database). In all of these cases, H. flavescens was not observed. This suggests that sampling power may not be the sole explanation behind the rediscovery of H. flavescens in the Middle Loire.
(2) The cessation of sand mining in the Loire Riverbed (Dambre, 1996) did not prevent the river incision in some locations (Latapie et al., 2014), nor did it hamper the revegetation of the bed (Braud, 2012). As the current hydrology of the river does not curb this process, the State services have chosen to carry out mechanical interventions on the river since the 1990s, with levelling, scarification and devegetation actions. In Europe, the effects of hydromorphological river restoration were studied in 20 rivers, using a standardised monitoring and sampling design Muhar et al., 2016). Januschke and Verdonschot (2016) indicated that river restoration had a significant positive effect corresponding to an increased richness of specialist riparian ground beetles. Another study showed that river re-braiding measures have increased the habitat diversity and the number of ground beetle species (J€ ahnig et al., 2009). These management operations contributed to the restoration of pioneer habitats such as sandy habitats with little vegetation favouring the return of species such as H. flavescens. H. flavescens may thus have benefited from these restoration operations to gradually recolonise the Middle Loire River from refuge sites (e.g. head of river).
(3) Winter floods are able to carry plant debris and insects in winter state of torpor over substantial distances (Colas, 1974), which become stranded on the banks and islands of the Loire, far from their area of origin. Such dispersion was reported for several ground beetles (Paillet, 2007). In our case, this would correspond to the transportation of individuals of H. flavescens from the upstream of the Loire or Allier River to our sampling sites. Finally, although no active flight has ever been observed, H. flavescens has full wing development (Lindroth, 1992). Hence, it is not strictly speaking impossible that individuals could have flown from refuge areas to the Loire River.
The rediscovery of H. flavescens raises questions about the origin of individuals observed in the Middle Loire River (local population or individuals from refuge areas), the stability of the populations (possibility of full life cycle, population size) and the anthropogenic impacts  Luka et al. (2009) (end of sediment extractions and hydromorphological river restoration). New inventory based on intensive sampling targeting its specific habitats are necessary to confirm the presence of H. flavescens and better document its abundance and spatial distribution. The complete cessation of sand exploitation in the Loire riverbed in 1995, the restoration engineering of shores and islands and more generally, the integration of biodiversity in the riverbed management strategies contributed to the restoration of pioneer habitats, hence allowing the recolonisation of the river banks by rare or extinct species such as H. flavescens.

Acknowledgements
We thank Aminata NDhyae, Vanessa Imbault, Audrey Bras and Alexis Bernard for their technical support in the field. Many thanks to Michel Chantereau and Damien H emeray, Conservator and technician of the National Nature Reserve of Saint-Mesmin, for entomological data and field support. We also thank Michel Binon, Jacques Coulon, Bernard Lemesle and Ren e Pupier for Harpalus flavescens distribution data. We express our gratitude to St ephane Rodrigues (UMR CITERES, University of Tours, France) for information and corrections on the morpho-sedimentary aspects of the Loire River in the manuscript. We thank the pilot of the ULM, Franck Robinet, who helped us to take the aerial photographs (Fig. 1b, c). This study was included in BioMareau project, coordinated by Marc Villar (INRA, UR588 AGPF, France). It was supported financially by the R egion Centre-Val de Loire and European Union. European Union is engaged on the Loire Basin with the European regional development fund (ERDF).

Supporting Information
Additional Supporting Information may be found in the online version of this article under the DOI reference: doi: 10.1111/ icad.12228: Figure S1. Location of sampling sites: Sandillon (a 1 ), Saint-Denis-en-Val (a 2 ), Mareau-aux-Pr es (b), Baule and Lailly-en-Val (c), Su evres and Saint-Dy e-sur-Loire (d). Green points represent pitfall trap sampling and red stars represent quadrat sampling. Figure S2. Habitats where Harpalus flavescens was caught: (a) dry sand with sparse vegetation; (b) cracked soil with silt/clay deposit awaiting vegetation.