Elucidating stygofaunal trophic web interactions via isotopic ecology.

Subterranean ecosystems host highly adapted aquatic invertebrate biota which play a key role in sustaining groundwater ecological functioning and hydrological dynamics. However, functional biodiversity studies in groundwater environments, the main source of unfrozen freshwater on Earth, are scarce, probably due to the cryptic nature of the systems. To address this, we investigate groundwater trophic ecology via stable isotope analysis, employing δ13C and δ15N in bulk tissues, and amino acids. Specimens were collected from a shallow calcrete aquifer in the arid Yilgarn region of Western Australia: a well-known hot-spot for stygofaunal biodiversity. Sampling campaigns were carried out during dry (low rainfall: LR) and the wet (high rainfall: HR) periods. δ13C values indicate that most of the stygofauna shifted towards more 13C-depleted carbon sources under HR, suggesting a preference for fresher organic matter. Conversion of δ15N values in glutamic acid and phenylalanine to a trophic index showed broadly stable trophic levels with organisms clustering as low-level secondary consumers. However, mixing models indicate that HR conditions trigger changes in dietary preferences, with increasing predation of amphipods by beetle larvae. Overall, stygofauna showed a tendency towards opportunistic and omnivorous habits-typical of an ecologically tolerant community-shaped by bottom-up controls linked with changes in carbon flows. This study provides baseline biochemical and ecological data for stygofaunal trophic interactions in calcretes. Further studies on the carbon inputs and taxa-specific physiology will help refine the interpretation of the energy flows shaping biodiversity in groundwaters. This will aid understanding of groundwater ecosystem functioning and allow modelling of the impact of future climate change factors such as aridification.

Introduction

During recent decades, investigations of trophic webs have become a cornerstone for the interpretation of functional biodiversity in freshwater ecosystems. Within both lentic and lotic environments, macroinvertebrate food web dynamics play a key role in shaping process-level aquatic ecosystem attributes [1]. Aquatic faunal trophic characterization is usually conducted by employing the morpho-behavioural based concept of functional feeding groups (FFGs) [2]. Since its inception, FFGs have been extensively used in ecological assessments and biomonitoring studies, and have allowed detailed assessment of ecological patterns in both natural and disturbed environments [3,4,5].

However, despite the hydraulic and ecological continuum in groundwater dependent ecosystems, the subsurface ecosystem and the study of its food chain interactions have suffered from a conceptual disconnection from surficial aquatic habitats. The main reasons are attributable to methodological limitations [6,7], scarce aquifer accessibility [8] and the lack of interdisciplinary approaches [9]. Moreover, compared to surface freshwater ecosystems, groundwaters are subjected to relatively extreme environmental conditions: sparse organic inputs, lack of light and primary production, and truncated trophic webs [10,11,12,13]. Altogether, these unique conditions shape obligate subterranean aquatic communities (stygofauna) dominated by plastic and opportunistic trophic behaviours [14,15], whose categorization via feeding modes such as FFGs is constantly at risk of misinterpretation. As a result, our knowledge about how food web interactions shape groundwater ecological functioning and community patterns is fragmented [16].

Stygofauna—when present—play a key role in regulating both ecological and hydrological dynamics in aquifers [17,18]: they actively bioturbate the sediment, facilitate nutrient recycling and, in combination with microbial communities, degrade/retain contaminants. In groundwaters, carbon inputs (allochthonous dissolved organic carbon (DOC) and chemoautotrophic production) are mediated by microbes which are then grazed by basal stygofauna [19]. Organic matter (OM) is transferred along the trophic chain via prey-predator interactions. Therefore, OM inputs, microbial communities, and the stygofaunal trophic web, all shape the energy flows sustaining the subterranean biodiversity [20].

The incorporation of biogeochemical approaches (i.e. stable isotopes composition, fatty acids content, radiocarbon analysis) has recently led to re-evaluation of the archetype of poorly structured–and generalist-dominated–trophic dynamics in groundwaters [21]. These designs are leading a vital transition from purely descriptive to functionally-based investigations, providing wider perspectives to the field [22].

Carbon (δ13C) and nitrogen (δ15N) stable isotope analysis (SIA) is a well-established approach enabling quantitative investigation of food webs [23,24]. Since its initial application in groundwater trophic ecology, several studies have benefited from the insights provided by the study of naturally-occurring stable isotopes [25,26]. However, δ13C and δ15N SIA investigations on bulk material are limited by the mixing of tissues and different biochemical pathways [27]. These limitations can be addressed by the complementary or alternative use of compound-specific approaches.

δ13C and δ15N Compound Specific Isotope Analysis (CSIA) on amino acids allows detailed characterization of food web interactions [28], by focusing on compounds created by definable biosynthetic pathways. Single amino acids can be divided into essential (EAA) and non-essential (NEAA). Whilst primary producers (plants, algae and bacteria) biosynthetise de novo EAA from a bulk carbon pool, animals lack these enzymatic pathways and acquire EAA from their diet [29]. As a result, tracking of EAA allows carbon fingerprinting of food sources down to the base of food webs [30]. Concurrently, δ15N CSIA can distinguish between compounds reflecting the source isotopic signal, and that enriched with each trophic step, thus providing crucial information on prey-predator interactions [31]. The application of CSIA in amino acids has allowed a much more thorough understanding of food web dynamics in freshwater [32], marine [33] and terrestrial environments [34], but despite the greater potential than bulk analysis [35], this technique has yet to be applied to food web studies of groundwater environments.

This study is, to our best knowledge, the first based on the combination of carbon and nitrogen CSIA in groundwaters, and focuses on a calcrete stygofaunal community under two contrasting environmental conditions: low rainfall (LR, dry season) and high rainfall (HR, wet season). We hypothesise that different environmental conditions trigger species-specific adaptations that are ultimately responsible for distinct food web interactions. The specific objectives of this work are: 1) unravel OM incorporation trends across the stygofaunal community, 2) decipher the trophic habits of the species and elucidate prey-predator interactions, and 3) provide biochemically-based knowledge about trophic web interactions in arid zone calcrete aquifers.

Methodology

Study area and field work

The field work was carried out at a calcrete aquifer (28°41‘S 120° 58‘E) located on Sturt Meadows pastoral station, Western Australia, ~42 km from the settlement of Leonora (833 km northeast of Perth, Fig 1A).

Fig 1

a) Borehole grid and its location in the Yilgarn region, Western Australia. b) Photos of some specimens from the bore samples (from left to right Paroster macrosurtensis adult, Paroster microsturtensis larvae, Scutachiltoni axfordi, Oligochaeta sp. and Oribatida sp.).

The Sturt Meadows calcrete hosts a very shallow aquifer, located two to four metres below the surface, and accessible through bores characterised by water depths ranging from a few centimetres to ten metres. The bore grid was initially drilled for mineral exploration and comprises 115 bore holes of between 5–11 m in depth forming a 1.4 km by 2.5 km (3.5 km2) area (Fig 1A). The bores are unlined, except for about the upper 0.5 m which are lined with 10 cm diameter PVC pipe for stability, and capped [36]. Three sampling campaigns–two of them corresponding to low rainfall periods (LR) and one during the wet season (high rainfall, HR) [37]–were carried out in July and November 2017, and March 2018. More details about the sampling design, monitoring of water depth and hydrogeological background at Sturt Meadows can be found in Saccò et al. [38].

The high morphologically (Fig 1B) and taxonomically diverse stygofaunal community at Sturt Meadows comprises three sister species of subterranean beetles (Paroster macrosturtensis (Watts & Humphreys 2006), Paroster mesosturtensis (Watts & Humphreys 2006) and Paroster microsturtensis (Watts & Humphreys 2006) and respective larvae)), three species of amphipods (Scutachiltonia axfordi (King, 2012), Yilgarniella sturtensis (King, 2012) and Stygochiltonia bradfordae (King, 2012)), aquatic worms (family Tubificidae (Vejdovský, 1884)) and water mites (order Oribatida; Dugès, 1834). Within the stygobiotic meiofaunal community, two species of harpacticoids (Novanitocrella cf. aboriginesi (Karanovic, 2004), Schizopera cf. austindownsi (Karanovic, 2004) and four species of cyclopoids: Halicyclops kieferi (Karanovic, 2004), Halicyclops cf. ambiguous (Kiefer, 1967), Schizopera slenderfurca (Karanovic & Cooper, 2012) and Fierscyclops fiersi (De Laurentiis et al., 2001)) can be found.

Adult and larval stygofaunal specimens were collected by hauling a small weighted plankton net (mesh 100 μm, [36]) five times from the bottom through the water column of 30 boreholes (Fig 1A) selected by simple random sampling [38]. Stygofaunal abundance data across the boreholes are reported in S1 Table.

All biological samples were kept frozen (–20°C) in darkness until further processing in the laboratory where individual organisms were counted and identified (and consequently separated) to the lowest taxonomic level via optical microscopy and reference to specific taxonomic keys. Roots and sediment samples from the bottom of the aquifer were obtained through the stygofaunal haul netting procedure, and were separated by using sterile glass pipettes during the sorting in the laboratory according to the sampling campaign (LR or HR). Sediment samples were soaked in acid (0.1 N HCl) to remove inorganic carbon and dried at 60°C for 24 hours. Given the delicacy of the hydrological dynamics in shallow calcretes [39], extensive water extractions spread along the bores were avoided and preliminary tests were carried out to quantify the potential risk of dewatering the calcrete.

Bores D13 and W4 host groundwater systems which are representative of the geological conformations of the area—phreatic and vadose calcretes interspersed with clay material—and were finally selected because of their hydrological and biotic stability (lowest risk of drying and representative ranges of Sturt Meadows stygofaunal diversity) [38]. Water samples for POC (particulate organic carbon) analysis were collected using a submersible centrifugal pump (GEOSub 12V Purging Pump) after wells were purged of three well-volumes and stabilisation of in-field parameters was observed. POC samples were obtained by filtering water from the bores D13 and W4 through GF/F filters (pre-combusted for 12 hours at 450°C), washed with 1.2 N HCl to remove any inorganic carbon, and subsequently dried at 60°C for 24 hours. The field site was accessed and samples were collected with permit approval (permit number 08-003150-1) from the Department of Parks and Wildlife of Western Australia.

Sample preparation and study design

All individuals from a single taxon were pooled for each sampling campaign (LR1, LR2 or HR) and subsequently washed with MilliQ water to remove external contaminants. Subsequently, samples were oven dried at 60°C overnight and crushed to a fine powder which was stored at –20°C until further analysis (Fig 2).

Fig 2

Methodological scheme of the study for stygofaunal samples (including copepods for bulk SIA).

EAA: essential amino acids; NEAA: non-essential amino acids; TP: trophic position; TDF: trophic discrimination Factor; β = ratio between δ15NGlu and δ15NPhe values in primary producers; SIMM: stable isotopes mixing models; LC-iRMS: Liquid Chromatography-isotope Ratio Mass Spectrometry; GC-iRMS: Gas Chromatography-isotope Ratio Mass Spectrometry; EA-CF-iRMS: Elemental Analyser-Continuous Flow-isotope Ratio Mass Spectrometry.

Due to sample size constraints, the samples for each taxon from the two low rainfall periods were further combined. Previous metagenomics investigations, together with mesocosm experiments and field observations at Sturt Meadows provided some information about the trophic habits of beetles and amphipods [40]. Adult subterranean beetles had active predatory feeding on epigean amphipods and copepods (including group feeder behaviours) together with scavenger habits (and potential active predatory pressures) on sister species. Beetle larvae (third and last instar) showed opportunistic predatory habits with a range of prey from copepods and amphipods to adult beetles from the three species (inter and intraspecific cannibalism), while amphipods displayed predation of copepods, epilithic biofilm grazing, root shredding and sediment filter feeding.

Bulk stable isotope analysis

δ13C and δ15N SIA on bulk homogenised samples of sediment, roots and stygofauna (respectively 1.28 mg, 0.08–0.14 mg and 0.63–2.79 mg per samples, S2 Table) were performed at the Australian Nuclear Science and Technology Organisation (ANSTO, Sydney). Samples were loaded into tin capsules and analysed with a continuous flow isotope ratio mass spectrometer (CF-IRMS, Delta V Plus, Thermo Scientific Corporation, U.S.A.), interfaced with an elemental analyser (Thermo Fisher Flash 2000 HT EA, Thermo Electron Corporation, U.S.A.) following the procedure of Mazumder et al. [41]. δ13C values are reported in per mil (‰) relative to the Vienna Peedee Belemnite (VPDB), while δ15N values are reported relative to reference N2 of known nitrogen isotopic composition (in ‰), previously calibrated against the AIR international isotope standard. δ13C POC (0.6 mg, S2 Table) was analysed at the Western Australian Biogeochemistry Centre at The University of Western Australia using a GasBench II coupled with a Delta XL Mass Spectrometer (Thermo-Fisher Scientific). Results have a precision of ± 0.10 ‰, and are reported relative to the NBS19 and NSB18 international carbonate standard [42].

Single amino acid carbon and nitrogen isotope analysis

δ13C CSIA

Stygofaunal samples (0.16–2.89 mg per sample, S2 Table) were hydrolysed under vacuum with 0.5 to 1 ml of amino acid-free 6 M HCl (Sigma-Aldrich) at 110 ˚C for 24 h. The protein hydrolysates were dried overnight in a rotary vacuum concentrator and stored in a freezer. Prior to analysis, the samples were resolved in Milli-Q water and 10 μl of 1-mmol solution of 2-aminoisobutyric acid (Sigma-Aldrich) was added as internal standard. The sample stock had a concentration of approximately 8 to 10 mg/ml, which was further diluted as needed. Single amino acid carbon isotope analysis was carried out at the La Trobe Institute for Molecular Sciences (LIMS, La Trobe University, Melbourne, Australia) using an Accela 600 pump connected to a Delta V Plus Isotope Ratio Mass Spectrometer via a Thermo Scientific LC Isolink (Thermo Scientific).

The amino acids were separated using a mixed mode (reverse phase/ion exchange) Primesep A column (2.1 x 250 mm, 100°C, 5 μm, SIELC Technologies) following the chromatographic method described in Mora et al. [43], after Smith et al. [44]. Mobile phases are those described in Mora et al. [45]. Percentage of Phases B and C in the conditioning run, as well as flow rate of the analytical run and timing of onset of 100% Phase C were adjusted as needed. Samples were injected onto the column in the 15 μl—partial loop or no waste—injection mode, and measured in duplicate or triplicate.

δ15N CSIA

CSIA nitrogen analyses were undertaken at the Organic Geochemistry Unit of the University of Bristol, UK. To extract the AAs, crushed samples (2.47–5.19 mg per sample, S2 Table) were hydrolysed in culture tubes (6 M HCl, 2 ml, 100°C, 24 h). A known quantity of norleucine (1 mg mL–1 in 0.1 M HCl) was added to each sample as an internal standard prior to hydrolysis. After heating, the tubes were allowed to cool then after centrifugation (3000 rpm, 5 min) the supernatant containing the hydrolysate from each tube was transferred to a clean culture tube and dried under N2 whilst being heated to 70°C. Once dry, each sample was re-dissolved in 0.1 M HCl and stored in the dark at -18°C until required for analysis.

The derivatisation procedure followed Styring et al. [46] and included isopropylation, with a 4:1 mixture of 2-propanol and acetyl chloride heating to 100°C for 1 hour, the reaction was quenched by rapidly cooling in a freezer. After removing the residual solvents under N2, acetylation of the amino group was achieved by adding a 5:2:1 mixture of acetone, triethylamine and acetic anhydride then heating to 60°C for 10 minutes before being allowed to cool. The derivatised AAs were isolated via liquid-liquid separation, residual solvent being removed by evaporating under N2. Samples were again stored at -18°C until required for analysis.

A Thermo Finnigan Delta Plus XP isotope ratio mass spectrometer (Thermo Scientific, Bremen, Germany) was used to determine the δ15N values of derivatised AAs. The mass spectrometer (EI, 100 eV, three Faraday cup collectors for m/z 28, 29 and 30) was interfaced to a Trace 2000 gas chromatograph via a Combustion III interface (CuO/NiO/Pt oxidation reactor maintained at 980°C and reduction reactor of Cu wire maintained at 650°C), both from Thermo Scientific.

Samples were dissolved in ethyl acetate and 1μl of solution was injected via a PTV injector. Helium at a flow of 1.4 ml min–1 was used as the carrier gas and the mass spectrometer source pressure was maintained at 9 X 10–4 Pa. The separation of the AAs was accomplished using a DB-35 capillary column (30 m X 0.32 mm i.d., 0.5 mm film thickness; Agilent Technologies, Winnersh, UK). The oven temperature of the GC started at 40°C where it was held for 5 min before heating at 15°C min–1 to 120°C, at 3°C min–1 to 180°C, at 1.5°C min–1 to 210°C and finally at 5°C min–1 to 270°C and held for 1 min. A Nafion dryer removed water and a cryogenic trap was employed to remove CO2 from the oxidised and reduced analyte.

All the δ15N values are reported relative to reference N2 of known nitrogen isotopic composition, previously calibrated against the AIR international isotope standard, introduced directly into the ion source in four pulses at the beginning and end of each run. Each reported value is a mean of duplicate δ15N determinations. A standard mixture of derivatised AAs of known δ15N values was analysed every three runs in order to monitor instrument performance.

Data treatment and statistical analysis

Only AAs that returned results for each taxon were considered. EAA and NEAA were separated according to the classification provided by Boudko [47]. EAAs were used in the interpretation of carbon flows—and potential shifts in OM incorporations—because they persist through the trophic chain [48] due to the little fractionation they undergo when incorporated into consumer’s tissue [49]. NEAA, which are subjected to much greater fractionation because of their de novo biosynthesis mainly from intermediates of the Krebs cycle (serine (Ser), glycine (Gly) and alanine (Ala)) and glycolysis (glutamic acid (Glx), aspartic acid (Asx) and proline (Pro) [50], were compared to EAA to investigate taxa-specific carbon isotopic trends (biosynthesis vs assimilation through diet) across the two rainfall periods (LR and HR).

All the statistical analyses were performed in R software version 3.6.0 (Development-Core-Team, 2016). Analysis of variance (ANOVA) coupled with Tukey’s HSD pairwise comparisons (R-package ‘stats’) were employed to inspect significant differences between δ13CEAA (Val, Phe and Arg) and δ13CNEAA (Krebs (Ser, Gly and Ala) and glycolysis (Asx, Glx and Pro) cycles) within the different rainfall conditions (LR and HR). Principal component analyses (PCA, R-package ‘vegan’) and Linear Discriminant Analysis (LDA, R-package ‘vegan’) among EAA was performed to explore sample distribution in the multi-dimensional space. Determination of EAA driving sample variability in the PCA was carried out via function fviz_contrib (R-package ‘factoextra’).

Trophic positions (TP) were calculated using the methodology reported by Chikaraishi et al. [33]: where δ15NGlu = δ15N of glutamic acid, δ15NPhe = δ15N of phenylalanine, β = ratio between δ15NGlu and δ15NPhe values in primary producers, and TDF = the trophic discrimination factor at each shift of trophic position.

Incorporation of source carbon from terrestrial vegetation has previously been reported at Sturt Meadows, with roots from surficial saltbush vegetation (C3 metabolism) frequently found in the groundwater [40]. β was accordingly assigned the value of +8.4 ± 1.6 ‰, which is the established value for aquatic food webs involving C3 plants [31]. Although other carbon sources are possible in groundwaters, as they are not established in this system, a conservative approach has been taken in using the value of an evidenced source. TDF was assigned the value of 7.6 ± 1.2‰, based on Steffan et al. [51] who showed it did not vary across trophic levels one to four in multiple controlled-feeding experiments, and for trophic levels one to five in a natural food chain, using terrestrial arthropod species [28].

Pairwise comparisons for δ15N were carried out with the same approach as for the carbon CSIA data. Robustness and consistency between CSIA and SIA data from beetles and amphipods were inspected using Pearson correlations (function rcorr in R-package ‘Hmisc’). SIMM (Stable Isotope Mixing Models, R-package ‘simmr’) were then applied to establish dietary proportions of the key ecological taxa (Fig 2). Since a specific trophic discrimination factor has not been calculated for stygofauna, we used the widely accepted values of 3.4 ± 2 ‰ for nitrogen and 0.5 ± 1 ‰ for carbon [52]. Markov chain Monte Carlo (MCMC) algorithms were used for simulating posterior distributions in SIMM, and MCMC convergence was evaluated using the Gelman-Rubin diagnostic by using 1.1 as a threshold value for analysis validation.

Results

Stygofaunal carbon fluxes

During LR, δ13C average values of AAs (δ13CNEAA[LR] and δ13CEAA[LR]) spanned from -31.52 ‰ (Phe) to -5.72 ‰ (Gly). Similar values were found under HR conditions (δ13CNEAA[HR] and δ13CEAA[HR]), ranging from -31.55 ‰ (Phe) to -4.92 ‰ (Ser) (Table 1).

Table 1

Low (LR) and high rainfall (HR) carbon amino acids spectrum (δ13C values) for stygofaunal specimens separated by non-essentials (NEAA: aspartic acid (Asx), serine (Ser), glutamic acid (Glx), glycine (Gly), alanine (ala), and proline (Pro)), and essentials (EAA: valine (Val), phenylalanine (Phe), and arginine (Arg)).

Average values (and standard deviation) for the analytical replicates are shown. P values for ANOVA Tukey’s HSD pairwise comparisons between NEAA and EAA are also illustrated.

Taxon	ID	NEAA						EAA			NEAA2 vs EAA3	NEAA4 vs EAA3
		Asx	Ser	Glx	Gly	Ala	Pro	Val	Phe	Arg	Krebs cycle	Glycolysis
LR
Paroster macrosturtensis	B	-17.64±0.59	-10.43±0.1	-18.74±0.53	-10.01±0.12	-20.01±0.12	-17.85±0.53	-24.39±0.55	-24.44±0.13	-19.56±0.55	P < 0.05	P < 0.05
Paroster mesosturtensis	M	-18.85±0.6	-11.32±0.16	-20.31	-12.22±0.6	-21.23±0.52	-15.10±0.63	-24.55±0.27	-26.24±0.23	-20.42±0.07	P < 0.05	P < 0.05
Paroster microsturtensis	S	-22.42±0.18	-12.39±0.62	-22.421	-14.84±0.64	-24.24±0.35	-18.67±0.62	-27.73±0.64	-28.98±0.07	-23.19±0.54	P < 0.05	P < 0.05
Paroster macrosturtensis larvae	Blv	-20.56±0.11	-9.22±0.18	-20.201	-14.17±0.6	-22.01±0.58	-19.31±0.6	-25.97±0.17	-26.48±0.12	-20.59±0.24	P < 0.05	P = 0.0618
Paroster mesosturtensis larvae	Mlv	-19.09±0.04	-6.70±0.18	-20.05±0.43	-16.12±0.65	-21.12±0.37	-16.38±0.3	-24.92±0.09	-27.08±0.1	-19.59±0.3	P < 0.05	P < 0.05
Paroster microsturtensis larvae	Slv	-19.1±0.5	-8.41±0.05	-17.75±0.03	-14.20±0.63	-21.57±0.15	-20.38±0.2	-25.34±0.07	-27.02±0.38	-18.59±0.21	P < 0.05	P < 0.05
Scutachiltonia axfordi	AM1	-16.1±1.3	-5.87±0.8	-14.09±1.4	-5.72±1.01	-20.42±2.88	-16.68±0.2	-24.55±2.01	-21.94±0.28	-15.31±0.74	P < 0.05	P < 0.05
Yilgarniella sturtensis	AM2	-19.09±0.08	-8.31±0.41	-21.35±3.77	-6.81±0.42	-20.29±0.09	-16.63±0.12	-25.48±0.17	-25.6±1.29	-19.56±1.39	P < 0.05	P = 0.0511
Stygochiltonia bradfordae	AM3	-21.7±0.1	-9.15±0.64	-24.65±4.38	-9.00±0.33	-24.08±0.19	-22.76±0.64	-28.84±0.25	-28.27±0.1	-23.54±0.38	P < 0.05	P = 0.0952
Tubificidae sp.	OL	-21.7±0.24	-16.01±0.44	-24.33±0.31	-20.931	-26.36±0.1	-25.78±0.36	-31.46±0.1	-31.52±0.23	-27.58±0.04	P < 0.05	P < 0.05
Oribatida sp.	OR	-20.44±0.63	-11.99±0.23	-17.771	-14.12±0.47	-21.18±0.07	-19.31±0.61	-26.36±0.03	-24.33±0.55	-18.85±0.2	P = 0.0955	P = 0.0826
HR
Paroster macrosturtensis	B	-18.67±0.45	-11.62±0.29	-18.19±0.59	-11.31±0.11	-19.648±0.45	-16.83±0.56	-25.44±0.64	-26.3±0.61	-20.40±0.6	P < 0.05	P < 0.005
Paroster mesosturtensis	M	-23.881	-18.061	-23.561	-16.641	-25.961	-21.991	-29.7681	-31.081	-26.051	P < 0.05	P < 0.05
Paroster microsturtensis	S	-20.871	-11.461	-20.81	-13.941	-22.621	-17.821	-26.7731	-29.041	-22.631	P < 0.05	P < 0.005
Paroster macrosturtensis larvae	Blv	-21.64±0.55	-12.581	-20.631	-13.561	-24.22±0.1	-19.431	-28.12±0.51	-28.22±0.57	-22.35±0.58	P < 0.05	P < 0.05
Paroster mesosturtensis larvae	Mlv	-23.881	-18.061	-23.561	-16.94±0.42	-25.971	-21.991	-29.38±0.55	-31.081	-26.051	P < 0.05	P < 0.05
Paroster microsturtensis larvae	Slv	-24.15±0.35	-18.03±0.59	-24.24±0.01	-18.741	-26.11±0.11	-22.191	-30.0741	-31.551	-26.59±0.2	P < 0.05	P < 0.005
Scutachiltonia axfordi	AM1	-23.73±0.6	-12.36±0.21	-23.55±0.35	-12.89±0.26	-24.13±0.47	-21.68±0.46	-29.52±0.52	-29.56±0.43	-23.80±0.54	P < 0.05	P < 0.05
Yilgarniella sturtensis	AM2	-23±0.01	-13.67±0.25	-23.60±0.42	-14.92±0.32	-26.28±0.65	-24.03±0.45	-30.87±0.13	-30.31±0.63	-24.99±0.58	P < 0.05	P < 0.05
Stygochiltonia bradfordae	AM3	-23.39±0.2	-12.37±0.62	-22.68±0.01	-13.19±0.01	-23.8465±0.14	-21.41±0.51	-28.7±0.22	-28.28±0.52	-22.94±0.11	P < 0.05	P < 0.05
Tubificidae sp.	OL	-20.251	-14.111	-20.251	-15.61	-23.941	-20.471	-27.931	-28.791	-21.621	P < 0.05	P = 0.0951
Oribatida sp.	OR	-20.42±0.64	-4.921	-22.11±0.61	-5.9±0.11	-22.19±0.09	-19.69±0.01	-27.52±0.11	-27.411	-21.02±0.39	P < 0.005	P < 0.05

1Unique run

2Calculated as average value of Ser, Gly and Ala

3Calculated as average value of Val, Phe and Arg

4Calculated as average value of Asx, Glx and Pro

δ13CNEAA[LR] average values varied from -26.36 ‰ (Ala) to -5.72 ‰ (Gly), similar values to δ13CNEAA[HR] spanning from -26.28 ‰ (Ala) to -4.92 ‰ (Ser). Overall, δ13CEAA showed trends towards more negative values than δ13CNEAA, which is involved in both Krebs and glycolytic cycles (ANOVA, P <0.005). This is consistent with the enrichment of NEAA during biosynthesis in the organism. With the exception of water mites (OR), P. macrosturtensis larvae (Blv), Y. sturtensis (AM2), S. bradfordae (AM3) under LR, and oligochaetes (OL) under HR, pairwise comparisons between δ13CEAA and δ13CNEAA confirmed a shift towards more negative values across the stygofaunal community (Table 1).

Neither PCAs nor LDAs on EAA distinguished different clusters within taxa or main groups (adult and larval beetles, and amphipods) nor among different rainfall periods (LR and HR). All three EAA correlated positively and significantly (P < 0.005), with phenylalanine and valine being the most informative AAs explaining the isotopic variability across stygofauna (~70%). δ13C values of valine (δ13CVal) and phenylalanine (δ13CPhe) show that, with the exception of P. microsturtensis (S) and S. bradfordae (AM3), the entire stygofaunal community experienced a significant shift towards more 13C-depleted values under HR (MANOVA, P < 0.005) (Fig 3, Table 2).

Fig 3

Biplot of δ13CPhe values vs. δ13CVal values for a) beetles (B, M, S, Blv, Mlv and Slv) and b) amphipods (AM1, AM2 and AM3), water mites (OR) and aquatic worms (OL). Red arrows indicate significant decreasing trends between LR and HR, while green arrows indicate increasing trends within rainfall periods. Refer to Table 1 for taxa IDs.

Table 2

Tuckey’s post hoc pairwise comparisons between phenylalanine and valine values under low (LR) and high (HR) rainfall conditions.

In bold significant results.

Taxon	ID	Phe			Val
		d.f.	T-ratio	P	d.f.	T-ratio	P
Paroster macrosturtensis	B	28	-4.497	< .0005	28	-2.163	< .005
Paroster mesosturtensis	M	28	-4.846	< .0001	28	-3.297	< .005
Paroster microsturtensis	S	28	-0.149	0.8829	28	1.967	0.0592
Paroster macrosturtensis larvae	Blv	28	-4.218	< .0005	28	-4.42	< .0005
Paroster mesosturtensis larvae	Mlv	28	-9.657	< .0001	28	-9.16	< .0001
Paroster microsturtensis larvae	Slv	28	-10.933	<0.001	28	-9.73	< .0001
Scutachiltonia axfordi	AM1	28	-18.4	< .0001	28	-10.2	< .0001
Yilgarniella sturtensis	AM2	28	-11.383	< .0001	28	-11.067	< .0001
Stygochiltonia bradfordae	AM3	28	-0.037	0.9704	28	0.282	0.7797
Tubificidae sp.	OR	28	-7.418	< .0001	28	-2.389	<0.05
Oribatida sp.	OL	28	6.594	< .0001	28	7.252	< .0001

Within the significant trends, P. macrosturtensis adults and larvae (B and Blv) showed the smallest change in carbon values (B: δ13CVal+Phe = -2.91; Blv: δ13CVal+Phe = -3.89) between rainfall regimes, while amphipods S. axfordi and Y. sturtensis (AM1 and AM2) showed the largest depletion (AM1: δ13CVal+Phe = -12.59 ‰; AM2: δ13CVal+Phe = -10.10 ‰), suggesting differential carbon incorporations under HR conditions.

δ15N and trophic levels

δ15NGlu average values varied between 15.4±0.4‰ (AM3[HR]) and 22.31±0.29‰ (M[HR]), while δ15NPhe values ranged from 10.67±0.45‰ (AM3[HR]) to 14.53±0.06‰ (M[HR]). When converted to trophic positions, the stygofaunal community at Sturt Meadows shows a truncated trophic chain, clustering around the secondary consumer level (Fig 4).

Fig 4

Calculated trophic positions (TP) of the stygofaunal specimens studied under LR (a) and HR (b) conditions.

Under LR conditions, P. macrosturtensis larvae (Blv) show the highest trophic position (TP = 3.33±0.02), while water mites (OR) sit at the lowest (2.78±0.09). Under HR conditions, P. microsturtensis adults (S) have the highest trophic position (3.27±0.01), whilst S. bradfordae (AM3) show the lowest value (TP = 2.73±0.01). Due to the low abundances it wasn’t possible to analyse biochemical fingerprints from water mites (OR[HR]: 37 individuals) and P. microsturtensis larvae (Slv[HR]: 10 individuals) during the wet season (HR) (S1 Table).

Overall, adult beetles (B, M and S) revealed higher trophic levels (TP>3) than amphipods (AM1, AM2 and AM3, TP<3). However, B, M and S did not show statistically higher values than AM1 under LR, the same pattern seen in P. mesosturtensis (M) under HR.

S. bradfordae (AM3) and P. macrosturtensis larvae (Blv) are the only organisms to show a statistically significant change in their TP values between LR and HR (Table 3), with both decreasing trends.

Table 3

δ15NGlu, δ15NPhe and TP values (±SD) during LR and HR regimes.

Pairwise comparisons within taxa from the same rainfall conditions and between rainfall periods (in bold significant patterns) for the same taxa are also illustrated. Taxa sharing the same letter do not differ significantly (Tukey’s HSD test, P < 0.05).

	δ15NGlu (‰)		δ15NPhe (‰)		TP		TP pairwise comparison
	LR	HR	LR	HR	LR	HR	LR	HR	LR vs HR
B	20.99±0.79	20.93±0.23	12.23±0.34	12.16±0.30	3.26±0.06	3.26±0.01	de	e	0.9712
M	22.29±0.89	22.31±0.29	14.25±0.26	14.53±0.06	3.17±0.08	3.13±0.03	bcde	cde	0.5415
S	22.12±0.23	20.69±0.08	13.47±0.13	11.87±0.03	3.25±0.01	3.27±0.01	de	e	0.6698
Blv	20.61±0.5	20.77±0.14	11.35±0.68	13.79±0.80	3.33±0.02	3.03±0.09	e	bcd	< .0001
Slv	21.99±0.55	Na	13.98±1.06	Na	3.16±0.21	Na	bcde	Na	Na
AM1	20.84±0.62	18.19±0.1	14.44±0.83	11.57±0.13	2.95±0.03	2.98±0.01	abcd	bc	0.6193
AM2	19.38±0.01	20.45±0.08	13.85±0.7	14.31±0.52	2.84±0.09	2.92±0.08	abc	b	0.135
AM3	20.04±0.7	15.4±0.4	14.24±0.13	10.67±0.45	2.87±0.07	2.73±0.01	ab	a	< .05
OR	16.11±0.85	Na	11.05±0.19	Na	2.78±0.09	Na	a	Na	Na

Food web dynamics

CSIA-based TP correlated significantly with SIA δ15N and δ13C values both under LR (P<0.05 in both cases) and HR conditions (P<0.01 and P<0.05 respectively). Under the latter conditions, δ13CVal values correlated significantly with CSIA-based TP (P<0.05). Copepods are generally thought to sit at the base of the food web [53,54]. However, these were analysed only via bulk SIA due to organism and sample size, and so could not be included in the TP analysis. They showed more 13C-depleted δ13C (cyclopoids: δ13CLR = -20.5‰, δ13CHR = -21.9‰; harpacticoids: δ13CLR = -20.6‰, δ13CHR = -23.5‰) and enriched δ15N (cyclopoids: δ15NLR = 13.9‰, δ15NHR = 14.5‰; harpacticoids: δ15NLR = 11.9‰, δ15NHR = 15.8‰) values under HR (Fig 5A and 5B), indicating that the change in rainfall regimes could play a role in stygobiotic meiofaunal biochemical incorporations.

Fig 5

SIA biplots of adults P. macrosturtensis (B), P. mesosturtensis (M), P. microsturtensis (S), P. macrosturtensis larvae (Blv), S. axfordi (AM1), Y. sturtensis (AM2), Cyclopoida sp. (C) and Harpacticoida sp. (H) under low rainfall (a) and high rainfall (b). AM1a and AM2a (in red): taxa showing the biggest depletion in δ13C values for essential amino acids (phenylalanine and valine) across rainfall conditions; Blvb (in green): taxon showing the biggest drop in trophic position value (TP) between LR and HR. Refer to S3 Table for δ13C and δ15N values of the taxa.

Amphipods AM1 and AM2 sat at the base of the trophic web under both rainfall conditions (TPs always below three, Table 3), and SIA carbon values (δ13C) confirmed a shift–already pinpointed via CSIA—towards more 13C-depleted carbon sources under HR. AM3, the smallest and rarest amphipod species in the calcrete, did not allow bulk SIA analyses due to the low abundances detected (LR (average value between LR1 and LR2): 27; HR: 19, S1 Table).

With respect to dietary preferences, for the amphipod S. axfordi (AM1), mixing models suggest that roots (and hosted microbial flora) contributed the greatest proportion (50%) during low rainfall conditions (LR) (Fig 6). The remaining diet was composed of POC (16.1%, derived from allochthonous carbon incorporations, a potential organic source for microbes), copepods (harpacticoids (13.9%) and cyclopoids (11.9%)) and sediment (8.1%) (i.e. OM laying at the bottom of the aquifer or epilithic biofilms).

Fig 6

Modelled contributions to the diet of amphipod S. axfordi (AM1) under a) LR and b) HR conditions. POC: particulate organic carbon, C: Cyclopoida sp.; H: Harpacticoida sp. Medians and quartiles of each prey category are represented in the boxplot, see S3 Table for SIA δ13C and δ15N data. AM2 illustrated the same dietary preferences as AM1 under both rainfall conditions.

Under HR conditions, the POC dietary contribution reached 66.1%, while roots plummeted to 3.3% (Fig 6). Overall, amphipod Y. sturtensis (AM2) showed the same dietary patterns as AM1.

Adult beetles P. macrosturtensis (B) and P. mesosturtensis (M) show only slight depletions in their isotopic values during HR in bulk δ13C and δ15N SIA, in contrast to the larger changes seen in the CSIA data. P. microsturtensis (S), which showed an isotopic enrichment in CSIA, counter to the rest of the community, shows similar behaviour to P. macrosturtensis (B) and P. mesosturtensis (M) in the SIA (S3 Table). All the three species show similar dietary preferences in mixing models across the rainfall periods (S4 Table). While diets were dominated by amphipods AM1 and AM2 during the LR period (B: 39.9%, M: 49.3% and S: 47.9% (Fig 7)), predation/scavenging of sister beetle species accounted for the biggest dietary proportions during the wet season (B: 52.9%; M: 49.4%; S: 41.9% (Fig 7)).

Fig 7

Contributions of P. microsturtensis adults’ diet for a) LR and b) HR. Diet sources: P. macrosturtensis (B), P. mesosturtensis (M), S. axfordi (AM1), Y. sturtensis (AM2), Cyclopoida sp. (C) and Harpacticoida sp. (H). Medians and quartiles of each prey category are represented in the boxplot, see S3 Table for δ13C and δ15N bulk data. P. macrosturtensis (B) and P. mesosturtensis (M) illustrated same trends of dietary contributions across rainfall periods (S4 Table). In these analyses, sister species P. mesosturtensis (M) and P. microsturtensis (S) were considered as Paroster prey items for diet reconstruction of P. macrosturtensis (B), while contributions from Paroster diet sources P. macrosturtensis (B) and P. microsturtensis (S) were used for P. mesosturtensis (M).

Mixing models indicate that P. macrosturtensis larvae (Blv), which showed the biggest shift in trophic position, has a preference for amphipods S. axfordi (AM1) and Y. sturtensis (AM2) under LR conditions (accounting for 52% of the diet contributions), but also consumes a range of other organisms (Fig 8). During HR, Blv’s diet is dominated by the two amphipod species, accounting for 79.6% of food sources (Fig 8). Overall, these results indicate changes in amphipods (AM1 and AM2) diet preferences linked with different OM inputs, coupled with enhanced species-specific predatory pressures from Blv under HR conditions.

Fig 8

Stygofaunal contributions to the diet of P. macrosturtensis larvae for a) LR and b) HR. Diet sources: P. macrosturtensis (B), P. mesosturtensis (M), P. microsturtensis (S) S.axfordi (AM1), Y. sturtensis (AM2), Cyclopoida sp. (C) and Harpacticoida sp. (H). During HR, diet source P. macrosturtensis (B) was discarded as the Gelman-Rubin diagnostic reported values exceeding the corresponding upper confidence limits at the 95% confidence level. Medians and quartiles of each prey category are represented in the boxplot, see S3 Table for δ13C and δ15N bulk data.

Discussion

Shifts in basal OM assimilation

δ13CEAA data suggest that the stygofaunal community at Sturt Meadows experienced a seasonal shift in carbon flows during the wet season (HR). The overall tendency towards more 13C-depleted δ13CVal and δ13CPhe values indicate general stygofaunal discrimination against 13C sources in that season.

Three groups (OL, AM3 and S) showed a counter trend of 13C enrichment in the EAAs during HR. Of these, the easiest to account for are the oligochaetes (OL) which also showed increased abundances (χ2 = 6.7698, P < 0.05) of individuals ranging from 2 (LR1) and 1 (LR2) to 21(HR) (S1 Table), indicating ideal conditions for the taxon during the wet season. As detritovores, oligocheates may be expected to preferentially consume more degraded, and so 13C-enriched, OM. The enrichment in S. bradfordae (AM3) and P. microsturtensis (S) is harder to explain at this stage. The low abundance of S. bradfordae (AM3) means that, like the oligochaetes, it was not included in the SIA analysis and less data is available. This taxon, together with P. macrosturtensis larvae (Blv), Y. sturtensis (AM2) and water mites (OR), lacked statistically significant differences when δ13C values of EAA during LR are compared with those of NEAA involved in the glycolytic cycle. Newsome et al. [50] indicated that δ13C values of NEAA from diets of omnivorous animals reflect de novo synthesis but also dietary incorporations. Differential routing of macromolecules by consumers [55] are one possible contributor to our results. However, to date isotopic routing hypotheses have been tested only in vertebrates [30,56], with the study of metabolic pathways in aquatic invertebrates largely unexplored. Further CSIA investigations involving species-specific bio-assimilation processes within the stygofaunal community are needed to provide a more accurate understanding of the biochemical dynamics regulating this system.

In line with our general data trends, Hartland et al. [15], who reported consistent depletion in δ13C stygofaunal values within OM-enriched groundwaters via sewage contamination, concluded that stress-subsidy gradients in groundwaters trigger profound changes in stygofaunal assemblages and have the potential to trigger shifts in feeding habits. Rainfall events trigger OM inflows which constitute high quality carbon sources for aquatic biota in groundwaters [16, 57].

Reiss et al. [26] demonstrated a strong link between nutrient inputs (mainly DOC) and groundwater microbe functional and metabolic richness after a major flooding event. Unfortunately, their methodology did not allow for corresponding macrofaunal trends to be identified. Nonetheless, microbially-derived OM incorporation by stygofauna has been reported in a number of groundwater ecology studies [16,21,58], and the biochemical importance of this linkage is widely accepted.

A key role in the observed trends at Sturt Meadows is played by amphipods which, together with copepods, are recognised as crucial actors in transferring OM to the upper stygofaunal trophic levels [16]. Specialized trophic habits in amphipods include epigean predation [10], detritivory [59], parasitism [60], biofilm grazing [61] and necrophagy [62]. Several studies have reported high degrees of trophic opportunism [54] and plasticity [63], allowing amphipods crucial shifts in feeding modes. Concurrently, niche partitioning has been addressed as a key mechanism to reduce intraspecific competition in ecosystems shaped by scarce nutrient availability [64]. However, our results do not show any conclusive evidence of epigean amphipod niche partitioning, with amphipods S. axfordi (AM1) and Y. sturtensis (AM2) showing the same dietary patterns. Overall, the isotopic data support the concept of opportunistic behaviours linked with changes in resource availability as a result of different rainfall regimes.

The HR event triggered substantial changes in the dietary proportions of S. axfordi (AM1) and Y. sturtensis (AM2), with notable decreases in root input and increases in POC. The extent of direct plant matter consumption by stygobionts reported in the literature–particularly by amphipods, which are facultative shredders–is both site and species-specific. Jasinska et al. [65] found that aquatic root mats were a key food source for a biodiverse cave fauna hosted by a shallow groundwater stream in Western Australia. Conversely, Navel et al. [66], reported the widely distributed amphipod species Niphargus rhenorhodanensis having preferential OM collector/gatherer feeding habits. In another study, Simon et al. [67] suggested that wood inputs played a role as indirect source of OM consumed by the ephilitic microbial mats which were ultimately targeted by common Gammarus amphipods.

At Sturt Meadows, a plausible explanation for the patterns observed is that during the dry season epigean amphipods rely on a more omnivorous diets where roots falling from the surface, and associated microbial and fungal biota, provide a substantial food source. Conversely, the wet season triggers inflows of replenished carbon (13C-depleted POC) that might fuel biological turnovers in microbiological activity, and POC-attached microflora may be ultimately targeted by epigean amphipods. These assumptions are in line with the finding reported by Brankvotis et al. [16], and support the concept that grazers play a crucial role in sustaining the functional diversity in groundwaters. The importance of plant matter input during at least part of the year is supported by a previous bulk SIA investigation at Sturt Meadows [40], which also suggested that terrestrial sources of carbon, mainly DOC, reached the aquifer via percolation and play a crucial role in energy flows within the system. It is worth noting that our δ13C values of decarbonated sedimentary fractions (referred above as ‘sediment’) were less 13C-depleted than those in other groundwater investigations ([68,69,70], and had ranges close those for dissolved organic carbon (DIC) in the region ([71,72,73]. Portillo et al. [74] reported karst microbial growth induced by both carbonate precipitation and dissolution, suggesting the inclusion of inorganic carbon within the estimation of global carbon budgets in groundwaters. In line with this work, Chapelle [75] reported in situ DIC production as a result of microbial metabolism involved in the dissolution of carbonate material in the black Creek aquifer (California, USA). Our results suggest that carbonate assimilation and/or dissolution processes are likely to occur in the sedimentary deposits of the aquifer, transferring an inorganic carbon isotopic fingerprint into the decarbonated and organic fractions. This can be tested in future by further functional studies on the microbial community [76] at the site.

Copepods (C and H) showed high δ15N values compared to the rest of the stygofaunal community (S3 Table), suggesting alternative nitrogen sources linked to different microbial baselines. Copepods act as energy drivers in recycling nitrogen via ingestion of sediment and attached bacteria [12], with ammonia (NH3), together with nitrates (NO3-), being an essential nutrient and energy source for subterranean microorganisms [77]. At Sturt Meadows aquifer, where ammonia levels are considerably higher than the natural concentrations [38], proliferation of selectively grazed ammonia-oxidising bacteria (AOB) might have played a key role in triggering the enriched δ15N values in copepods.

The present study is constrained by its focus on stygofauna and therefore cannot provide direct evidence of microbially-derived ecological shifts. Future research needs to combine stygofaunal and microbial investigations to create a complete picture of the ecosystem. CSIA and functional genetic studies on microbes and copepods would also help define transitions from microflora to stygofauna, a process that has been understudied so far. Recent promising investigations in surface terrestrial [34] and aquatic [78] environments suggest a design—carbon fingerprinting—based on the incorporation of isotopic data into multifactorial mixing models that allow specific elucidation of bacterial sources in diets. Overall, despite the methodological challenges posed by groundwaters, isotopic data on stygofaunal carbon fluxes provides baseline knowledge that help untangle the intricate biochemical dynamics regulating subsurface environments.

Trophic interactions

Our data on nitrogen CSIA pinpointed two main trophic levels marked by a small but clear separation between the top predators—adult beetles (B, M and S)—and primary consumer amphipods (AM1, AM2 and AM3) under both rainfall conditions. Compared to other ecosystems [29], the Sturt Meadows aquifer shows a very simple and truncated trophic web dominated by omnivorous habits. This is consistent with previous assumptions [67] due to the lack of primary producers [9] and scarce nutrient availability [79].

Within subterranean beetles, the smallest species P. microsturtensis (S), together with P. macrosturtensis (B), sit at the top of the trophic chain during HR (Table 3). Under those conditions, increased oxygen levels [38] may play a role in shaping changes in stygofaunal niche occupation. Subterranean beetles’ body size has been found to drive differential physiological responses to increased exoskeleton respiration rates (inversely proportional to the body size) which ultimately affect the ability to allocate energy for breeding and foraging [80]. As the smallest species P. microsturtensis (S) can adapt their metabolism more quickly than direct competitor sister species P. mesosturtensis (M) under favourable conditions—such as HR regimes–they are more likely to show shifts in ecological niche occupation [38]. This trend, combined with the group feeding tendency of P. microsturtensis (S) beetles [40], indicates higher efficiency in activating more intensified predatory strategies when compared to P. mesosturtensis (M).

Dytiscidae beetle larval stages—commonly referred as ‘water tigers’—are ferocious carnivores [81] with extremely opportunistic feeding behaviours involving scavenging and cannibalism [82]. At Sturt meadows, the third instar of blind P. macrosturtensis larvae (Blv) has a considerably bigger head capsule—paired with elongated mouthparts—than adult stages (Fig 9). These morphological features are likely to provide ethological advantages for non-visual predacious habits within light-less environments such as groundwaters [83]. This is consistent with stable isotope data from LR conditions positioning Blv at the top of the trophic web (Table 3).

Fig 9

Comparisons between adult and larvae (whole body and head capsule) of P. macrosturtensis (a and b), P. mesosturtensis (c and d) and P. microsturtensis (e and f).

Overall, modelled dietary contributions of P. macrosturtensis larvae (Blv) indicated a preference for amphipods (AM1 and AM2) coupled with residual cannibalism/scavenging (B, M and S) and predation of copepods (C and H) (Fig 8). Under HR, P. macrosturtensis larvae (Blv) showed the biggest drop in TP compared to LR (TPLR-HR = -0.3), which can be explained by an increased predatory focus on amphipods, and reduced feeding on secondary consumer sister species.

Previous work on surficial Dytiscidae larval stages published by Inoda et al. [84] stressed the importance of prey recognition through smell. According to their results, prey density was found not to be a key factor in shaping feeding behaviours, and self-other recognition played a role in group feeding. Overall, these findings indicated prevention of cannibalism through scent recognition. In groundwater, with total darkness and high influence of OM inputs on population dynamics [85,86], these patterns are likely to be strengthened. We suggest that the shifts in Blv predation seen in our results are dictated by a combination of chemical recognition and increased likelihood of encountering prey (amphipods) driven by enhanced resource availability (OM) during HR periods.

The role of bottom-up vs top-down forces in natural communities has been a cornerstone issue in the field of trophic ecology since the first empirical investigations [87]. Despite the controversy generated by the debate, there is now consensus that both forces act simultaneously on populations. This reinforces the need for whole system studies considering the interaction between heterogeneous (biotic and abiotic) forces and their effect on communities [88,89,90].

Our study, in line with a number of other investigations in the field [26,91] confirms that rainfall events via water advection are key drivers in defining energy flows and ecological patterns within resource-limited environments. We suggest that OM-driven bottom-up regulations, increasingly accepted as driving factors shaping population dynamics in aquifers [92], shaped the shifts in feeding behaviours among amphipod taxa in the calcrete. However, despite the beneficial conditions for primary consumers triggered by increased nutrient availability (i.e. microbial biofilms) and better environmental settings (i.e. increased oxygen, [38]), a decrease in amphipod populations under HR indicates the existence of additional ecological factors.

Top-down forces (i.e. natural predators), widely studied in surface aquatic ecosystems [93,94], have hardly been addressed in groundwater. Previous genetic investigations at Sturt Meadows pinpointed predatory pressures from beetles on amphipods and copepods [95] and reported a lack of trophic niche partitioning among the Paroster species. In another study, Hyde [96] reported evidence from metagenomics data suggesting that subterranean blind beetles at Sturt Meadows feed on both prey invertebrates and their sister species. Our isotope results support these hypotheses, indicating opportunistic predaceous habits in the calcrete, mixed with scavenging/cannibalism. However, substantial uncertainty remains about the magnitude of interspecific predatory pressures among Paroster sister species, and further species-specific lab experiments are needed to investigate these ethological aspects.

Biochemical functional role interpretation coupled with abundance data suggests that bottom-up population dynamics are counterbalanced in the system by top-down forces. Increased numbers of top predators (adult beetles B, M and especially S) were paired with a decrease of key prey items (amphipods AM1, AM2 and AM3) when HR is compared with the dry season (LR1 and LR2) (Fig 10).

Fig 10

Bar chart graphs comparing dry season abundances (as the average value of LR1 and LR2) with HR conditions for top predators (beetles B, M and S) and key prey items (amphipods primary consumers AM1, AM2 and AM3).

See S1 Table for detailed abundance data. None of the abundances of these taxa changed significantly between LR and HR. See Saccò et al. [38] for detailed statistical analyses across LR1, LR2 and HR.

In light of the dynamics shown by our isotope data, we suggest that the reported shift in amphipods (AM1 and AM2) carbon incorporation during HR might have triggered changes in their ecological behaviour, exposing them to increased predatory pressures from the top predator Paroster beetles (B, M and S). However, given the high degree of opportunistic behaviour reported by stygofauna [14], further investigations on species-specific ethological dynamics would be helpful to infer community dynamics.

The number of third instar dytiscidae larvae ‘Blv’ did not vary across sampling campaigns, suggesting differential ecological niche occupations. Previous investigations on Paroster larvae detected three instars before pupating, with the first two occupying a reduced proportion of their lifetime [83]. Future investigation of early stages of larval developments are needed to establish if potential population blooms (i.e. mass reproduction) are linked with contrasting recharge periods.

Conclusions

The application of CSIA and SIA allows elucidation of the trophic dynamics shaping stygofaunal communities in an arid zone calcrete aquifer. Rainfall acts as a driver in regulating both top down and bottom up changes in dietary habit. Subterranean invertebrate population dynamics are notoriously hard to investigate due to sampling obstacles and a current lack of knowledge around stygofaunal biological cycles [7,36]. However, our isotopic results allow a greater insight into the food web dynamics and the biogeochemical forces that shape them than has previously been possible. Further investigations involving higher numbers of samples from more biodiverse systems or complex trophic assemblages (i.e. alluvial aquifers) will help refine the approach. The incorporation of qualitative analyses such as DNA metabarcoding would also complement quantitative isotopic methods to provide crucial insights into processes (i.e. cannibalism) and key driving forces (i.e. niche partitioning) that are hard to detect via one method alone. Lastly, investigation of nitrogen sources and their isotopic changes would open up the nitrogen data collected to interpretation beyond trophic position.

Groundwater environments are fundamentally important to ecosystems, communities and industry, and a robust understanding of their ecosystem dynamics is essential to accurately assess environmental impacts, whether anthropogenic, or climatic. Isotopic data, especially if combined in multidisciplinary studies with other parameters [22] has a key role to play in elucidating previously hard to investigate function within these cryptic systems.

Abundance data of the different stygofaunal taxa at Sturt Meadows detected during the sampling campaigns LR1, LR2 and HR.

(XLSX)

Click here for additional data file.

Sample weights (mg) for SIA and CSIA analyses.

n/a: not available.

(XLSX)

Click here for additional data file.

δ13C and δ15N values of stygofauna, roots, sediment and POC during LR and HR.

(XLSX)

Click here for additional data file.

Dietary proportions of P. macrosturtensis (B), P. mesosturtensis (M) and P. microsturtensis (S) under LR and HR conditions.

(XLSX)

Click here for additional data file.

4 Sep 2019

[EXSCINDED]

PONE-D-19-21542

Elucidating stygofaunal trophic web interactions via isotopic ecology

PLOS ONE

Dear Mr. Sacco,

Thank you for submitting your manuscript to PLOS ONE. After careful consideration, we feel that it has merit but does not fully meet PLOS ONE’s publication criteria as it currently stands. Therefore, we invite you to submit a revised version of the manuscript that addresses the points raised during the review process.

Please pay attention specially to explain or provide a reason for the high sediment d13C values in your study.

We would appreciate receiving your revised manuscript by Oct 19 2019 11:59PM. When you are ready to submit your revision, log on to https://www.editorialmanager.com/pone/ and select the 'Submissions Needing Revision' folder to locate your manuscript file.

If you would like to make changes to your financial disclosure, please include your updated statement in your cover letter.

To enhance the reproducibility of your results, we recommend that if applicable you deposit your laboratory protocols in protocols.io, where a protocol can be assigned its own identifier (DOI) such that it can be cited independently in the future. For instructions see: http://journals.plos.org/plosone/s/submission-guidelines#loc-laboratory-protocols

Please include the following items when submitting your revised manuscript:

A rebuttal letter that responds to each point raised by the academic editor and reviewer(s). This letter should be uploaded as separate file and labeled 'Response to Reviewers'.

A marked-up copy of your manuscript that highlights changes made to the original version. This file should be uploaded as separate file and labeled 'Revised Manuscript with Track Changes'.

An unmarked version of your revised paper without tracked changes. This file should be uploaded as separate file and labeled 'Manuscript'.

Please note while forming your response, if your article is accepted, you may have the opportunity to make the peer review history publicly available. The record will include editor decision letters (with reviews) and your responses to reviewer comments. If eligible, we will contact you to opt in or out.

We look forward to receiving your revised manuscript.

Kind regards,

Jose M. Riascos, Ph.D.

Academic Editor

PLOS ONE

Journal Requirements:

When submitting your revision, we need you to address these additional requirements.

Please ensure that your manuscript meets PLOS ONE's style requirements, including those for file naming. The PLOS ONE style templates can be found at

http://www.journals.plos.org/plosone/s/file?id=wjVg/PLOSOne_formatting_sample_main_body.pdf and http://www.journals.plos.org/plosone/s/file?id=ba62/PLOSOne_formatting_sample_title_authors_affiliations.pdf

1. In your Methods section, please provide additional information regarding the permits you obtained for the work. Please ensure you have included the full name of the authority that approved the field site access and, if no permits were required, a brief statement explaining why.

[Note: HTML markup is below. Please do not edit.]

Reviewers' comments:

Reviewer's Responses to Questions

Comments to the Author

1. Is the manuscript technically sound, and do the data support the conclusions?

The manuscript must describe a technically sound piece of scientific research with data that supports the conclusions. Experiments must have been conducted rigorously, with appropriate controls, replication, and sample sizes. The conclusions must be drawn appropriately based on the data presented.

Reviewer #1: Yes

Reviewer #2: Yes

**********

2. Has the statistical analysis been performed appropriately and rigorously?

Reviewer #1: I Don't Know

Reviewer #2: Yes

**********

3. Have the authors made all data underlying the findings in their manuscript fully available?

The PLOS Data policy requires authors to make all data underlying the findings described in their manuscript fully available without restriction, with rare exception (please refer to the Data Availability Statement in the manuscript PDF file). The data should be provided as part of the manuscript or its supporting information, or deposited to a public repository. For example, in addition to summary statistics, the data points behind means, medians and variance measures should be available. If there are restrictions on publicly sharing data—e.g. participant privacy or use of data from a third party—those must be specified.

Reviewer #1: Yes

Reviewer #2: Yes

**********

4. Is the manuscript presented in an intelligible fashion and written in standard English?

PLOS ONE does not copyedit accepted manuscripts, so the language in submitted articles must be clear, correct, and unambiguous. Any typographical or grammatical errors should be corrected at revision, so please note any specific errors here.

Reviewer #1: No

Reviewer #2: Yes

**********

5. Review Comments to the Author

Please use the space provided to explain your answers to the questions above. You may also include additional comments for the author, including concerns about dual publication, research ethics, or publication ethics. (Please upload your review as an attachment if it exceeds 20,000 characters)

Reviewer #1: This article shows a great deal of field and laboratory work, where a multi-disciplinary approach is used to elucidate a complex series of questions regarding trophic interactions and their shifts within stygofauna species comparing two distinct environmental seasons. Bulk and compound specific stable isotope analysis coupled with environmental factors are linked together to explain a series of changes in stygofauna composition, feeding behavior and trophic shifts. The authors do a good job justifying their research and stating the importance of groundwater ecosystems and the stygofauna. They present a clear and concise background on the role that stygofauna play in the carbon cycle and how it incorporates into the trophic web, setting the stage for the methods the authors selected to answer the research questions. The objectives of this work are stated clearly.

It would make this paper a lot easier to read if the authors would include a broader description of precisely how d13C CSIA on EAA and NEAA allows the identification of source carbon signatures (lines 81-86), giving a hint to what the reader should expect to find and thus, would make the reader fully understand the differences pointed out in the results. As a suggestion, including what the authors expected to find on CSIA EAA and NEAA would also be nice.

Their data supports a change in feeding habits of several stygofaunal components from different levels of trophic web, thus demonstrating the importance of carbon input in these subterranean ecosystems.

Overall, the text is well structured, the results are very interesting and I believe this article would serve as a good baseline for future work in Australian calcretes. However, to my understanding, some of the assertions in the discussion are not fully supported by the data and some of the results are not fully discussed. Thus, with some adjustments, I would recommend this article to be published.

Annotations and suggestions:

Line number: Comment.

Methods: Some specifications would help the reader to have a better understanding of exactly how the sampling took place and would be vital for replication: How deep below the surface is the water table in this region? Does it vary throughout the year? Are 11m enough to get deep into the groundwater? How far is the distance from the water table to the bottom? how much does the net go underwater?

78: It is not clear if the authors will use SIA to refer to stable isotope analysis or to bulk material Stable isotope analysis.

127: The authors should state how the boreholes were selected (randomly or selected and the selecting criteria). Could there be any bias in the boreholes selected?

131: In an effort to describe a replicable method, please comment on the methods used to obtain such samples: were the sediments collected from the bottom of the borehole, how were they obtained and what instruments were used?

133: The authors should mention why they consider that these 2 boreholes are representative of the rest and how they were chosen.

145: Figure caption requires modification of EEA to EAA as is referred to in the figure. Initials in figure need to be explained: LC-iRMS (liquid chromatography- Isotope Ratio Mass Spectrometry), GC-iRMS (Gas), EA-iRMS (Elemental analysis)

161: Possibly the authors would use the above defined abbreviation?

164-166: This procedure is missing or unclear in Fig. 2

228: It would help the reader know the number of specimens (n) the authors used in the different sets of analysis.

234: Tukey

272-277: I would suggest the authors to be consistent with the abbreviations, probably using NEAA and EAA would eliminate the need for a new abbreviation in this table.

Check for misspellings in Table 1.

Table 1. Annotation 2: using only Ser, Gly and Ala is not explicitly explained or justified in the text by the authors and is suggested.

279: The results shown here should be further explained and clarified to suggest patterns of predators or prey so the readers understand exactly what the authors are pointing out.

286-293: This section of results was very nicely explained.

303: Table 2: The readers would appreciate the taxon names along with the ID´s

320: the authors should identify which is LR and HR within a and b in the caption of fig4.

325: I would suggest the authors be consistent with the description style and include the taxon name before the abbreviation ID, as they did in 232-235.

326: A sample size number (n) would help the readers have an idea of what is a low abundance in these sites, it could help the authors to remember it may be very different in other subterranean environments.

336: table 3 caption: It would help the reader to know the n size in each of the samples, the authors should also specify what the range after each value is referring to (SD, error, variance, etc). The code description under the TP pairwise comparison column is missing in the caption.

350: the authors should use formal English.

356: VPBD and AIR are not explained in caption of figure. I also suggest the authors to include 13C and 15N values of sources in their figures, it can give a better understanding to the reader of how these sources relate to each of the stygofaunal elements and how these may also shift during LR and HR.

367: I would suggest the authors include the number of specimens obtained.

368: LR has been defined as dry season previously.

368-377: It appears as this paragraph has been separated or is unclear and the authors need to revise it.

370: AM1 is mentioned in the text and figure caption but AM2 is shown in the figure. Are the differences shown significantly different?

377: It would help the reader if this was mentioned along with the reference to fig 6.

379-381: it would help the reader if the authors would include the taxon name again.

381: The authors are suggested to include the n values in table S1.

382: table S2 has not been mentioned and the authors are invited to consider including it into their results, considering it will be discussed further ahead and the abundance data (and figure) provide interesting information on the community assemblage.

387: misspelling

388: C and H are not used in this figure.

396: There is no sustaining evidence from the authors to claim a change in amphipod behavior in HR; the authors should explain what the amphipod ethology is in LR and HR (and if it differs among species), and further explain how these changes -coupled with increased OM inputs- trigger the species specific predatory response from the P. mesosturtensis larvae (Blv).

398: There appear to be several things to note in figures 7 & 8: S is not present in figure 7, which could mean P. microsturtensis has no cannibalistic behavior (although mentioned earlier) and that other beetles are not feeding on S either (if "B and M illustrated same trends" as noted in fig 7.

B is not present in fig8.b, possibly meaning that Blvs do not feed on adults of their same species during HR. I would suggest the authors to mention these differences in the results section to avoid misinterpretation from the readers.

420-421: POC and DOC would hint the readers to think of carbon instead of matter (and they were also defined earlier in the text).

416-427: Apparently reference [50] says the epilithic biofilms are consumed preferentially from the OM that enters from the surface, and refers to the “high quality” carbon from outside as an assumption, but I could not find their quality measurements in this reference. I would suggest the authors to consider looking at J. Pohlman´s work (1997, 2011) and Brankovits et al 2017 [16] for additional information on "quality" carbon and other production sources in flooded caves. An alternative, would be to consider in situ production, such as it has been observed in other subterranean ecosystems and would probably correspond to a “groundwater microbe functional and metabolic richness” increment after OM recharge event, as demonstrated by [27].

435-437: Although it is not this papers scope, is it possible there was an abundance difference within these species among your monitored sites? If so, could these differences be suggesting intraspecific competition?

449-451: Including the sources 13C and 15N in the figures shown above, could help the reader follow this description.

452: This statement could confuse the readers. Would the authors please expand their explanation on how 13C depleted POC (sources with a greater abundance in 12C) is fresh carbon? Did the authors mean 14C, the replenishment of C sources or something else?

481: Authors should clarify whether these results are from another study or include their environmental data in the supporting information.

483-484: I would suggest the authors to mention that these amazing beetles actually breathe through their exoskeleton. Some readers are unfamiliar with the Australian stygofauna.

479-489: I suggest the authors are consistent throughout the document when referring to the taxon or ID´s of the species.

500: which species are the authors referring to?

501: This statement is confusing to the reader, it appears to suggest that feeding on another predator is less nutritious than feeding on a primary consumer, is there supporting evidence?

502: please include the ID code in the taxon box within figure 8, since the text refers to the code instead of the name.

502-504: this sentence is a little confusing, it would help the reader if the authors would be more specific in their description of when the increased predatory focus occurred (Hr or LR); Consistency of abbreviations would be clearer: "under HR, Blv...compared to the LR"

505-510: In this paragraph it is not clear when the authors are referring to data from [66] and when they are addressing their own findings.

510-513: The authors have not made a statement of any estimation of prey density in HR or LR and there is no reference to this data. This is unsupported. -Figure 10 and table S2 need to be mentioned earlier; probably in methods or results. Statistical analysis should be provided for this claim.

519-527: As a reader who potentially does not know the particular conditions of this site, is it possible that the authors could consider an autochtonous (in situ) primary production that could fuel the trophic web throughout the LR as observed in other studies (16)? –probably a good place in the manuscript to discuss?-

528: Do the authors refer to natural predators?

529: References 77 and 78, though doctoral dissertations, have not been through a peer reviewed process and thus should be used carefully. -78 reference year is 2018, not 2010.

547: Being this the only plausible explanation exposed, I would suggest the authors expand their explanation and present or point to supporting information that leads to their hypothesis.

Questions that strike me are tied to their biological reproductive cycle (for both prey and predators), such as discussed below.

569: authors should revise the sentence.

Reviewer #2: Sacco et al. use stable isotopes of carbon and nitrogen (both bulk and amino acids) to unravel trophic structure of a stygofaunal community. They find clustering of organisms around trophic level 3 and report a change in source use between a low and high rainfall period. The paper is generally well written and the topic is interesting. I have a few comments I hope will improve the manuscript.

1. I think the intro or methods could use a bit more description of the conditions in the boreholes. How far does light penetrate? Should we expect some primary production in the surface waters that could sink to the bottom, or is it entirely runoff from the surface? Are the copepods that were sampled from near the surface, or are they also in permanent darkness? What about POC samples (surface or dark)? How were the roots and sediment samples obtained, and what does the latter represent? Sediment is a mixture rather than a source. Adding detail on the source pathways will help contextualize the isotope data that come later.

2. The number of organisms captured should be expressed as per sampling unit (e.g. # of organisms per tow), since five tows were used at each site at each time. Comparing abundances at different times relies on standardized effort across sites and times. Also, the abundance data should probably be reported in the results section so it doesn’t come as a surprise later (i.e. before line 526 and Figure 10).

3. The sediment has unusually high d13C values, suggestive of carbonate contamination. It is noted that POC samples were acid washed. What about the sediments? Were they also acid washed? Without acid washing, I don’t trust the reported value and it could skew the mixing model very strongly (see Phillips et al. 2014 Can. J. Zool. 92: 823-835).

4. I find it interesting that none of the organisms analysed for CSIA came back as primary consumers. All were near TL = 3. So who are the primary consumers? This points to microbial conditioning of detrital sources as the likely pathway leading to metazoans. This should be highlighted in the discussion.

5. The mixing model did not resolve the diets of adult beetles particularly well (lots of overlap in 95% CIs), so this shouldn’t be stated as a significant difference between seasons.

Specific comments

Line 25 and elsewhere – when using the term “depleted”, always refer to the heavier isotope i.e. “13C-depleted”.

Line 78 – delete “SIA”

Line 84 – change “signatures” to “isotopes” since signature refers to large well-buffered reservoirs

Line 140 – delete “the”

Line 157 – I was surprised to hear the biofilm referred to as “shredding”. Is that an error? Do microbes shred?

Line 162 – are the copepods not considered to be stygofauna?

Line 320 – add (a) and (b) to the figure caption

Line 348 – it is worth noting here that copepods had very high d15N relative to the rest of the food web (not just in HR), despite being assumed to be primary consumers. This implies a distinct N source for these organisms (i.e. different baseline), and it is unfortunate that they weren’t analysed for CSIA. If they truly are primary consumers, then the CSIA would have revealed it.

Line 370 – add “mean and 95% credible interval” to the figure caption

Line 378 – are the depletions here referring to 13C or 15N or both?

Line 397 – mean proportions were the same in both seasons so this statement is rather speculative

Line 407 – suggest adding “in that season” after “sources”

Line 479 – this paragraph is very speculative and could be deleted without affecting the paper’s interpretations.

Line 496 – it might be worth noting that Blv were not at the top of the trophic web when bulk data were considered, perhaps owing to the different baseline d15N value of the copepods

Line 515 – field “of” trophic ecology

Line 569 – quantitative isotopic “methods”

**********

6. PLOS authors have the option to publish the peer review history of their article (what does this mean?). If published, this will include your full peer review and any attached files.

If you choose “no”, your identity will remain anonymous but your review may still be made public.

Do you want your identity to be public for this peer review? For information about this choice, including consent withdrawal, please see our Privacy Policy.

Reviewer #1: Yes: Efrain M. Chavez Solis

Reviewer #2: No

[NOTE: If reviewer comments were submitted as an attachment file, they will be attached to this email and accessible via the submission site. Please log into your account, locate the manuscript record, and check for the action link "View Attachments". If this link does not appear, there are no attachment files to be viewed.]

While revising your submission, please upload your figure files to the Preflight Analysis and Conversion Engine (PACE) digital diagnostic tool, https://pacev2.apexcovantage.com/. PACE helps ensure that figures meet PLOS requirements. To use PACE, you must first register as a user. Registration is free. Then, login and navigate to the UPLOAD tab, where you will find detailed instructions on how to use the tool. If you encounter any issues or have any questions when using PACE, please email us at figures@plos.org. Please note that Supporting Information files do not need this step.

26 Sep 2019

Elucidating stygofaunal trophic web interactions via isotopic ecology

Ms. Ref. No.: PONE-D-19-21542

Title: Elucidating stygofaunal trophic web interactions via isotopic ecology

Journal: PLOS ONE

Please accept our sincere gratitude for providing challenging but highly constructive comments and suggestions, which, we believe, have greatly improved the paper.

Below, we provide a detailed, point-by-point account of how we responded to each comment

In yellow in the MS: changes according to suggestions from reviewer 1

In light blue in the MS: changes according to suggestions from reviewer 2

In green in the MS: authors’ contributions to an improved version of the manuscript

Please pay attention specially to explain or provide a reason for the high sediment d13C values in your study.

Thank you for your comment. Sediment samples (decarbonated sedimentary fractions) were hydrolysed to remove the inorganic component. δ13C values of sediment are close to the ones of DIC in water, suggesting a linkage to carbonate metabolisms as indicated by Portillo et al., 2009. Naturwissenschaften, 96(9), 1035-1042. Further details on the explanation of the observed patterns are reported in the response to reviewer 2 (major comment 3.). We also improved grammar and composition along the manuscript.

1. In your Methods section, please provide additional information regarding the permits you obtained for the work. Please ensure you have included the full name of the authority that approved the field site access and, if no permits were required, a brief statement explaining why.

Thank you for your note. We included the permit number and the full name of the authority for the fieldwork activities (Lines 159-160).

Reviewer #1: This article shows a great deal of field and laboratory work, where a multi-disciplinary approach is used to elucidate a complex series of questions regarding trophic interactions and their shifts within stygofauna species comparing two distinct environmental seasons. Bulk and compound specific stable isotope analysis coupled with environmental factors are linked together to explain a series of changes in stygofauna composition, feeding behavior and trophic shifts. The authors do a good job justifying their research and stating the importance of groundwater ecosystems and the stygofauna. They present a clear and concise background on the role that stygofauna play in the carbon cycle and how it incorporates into the trophic web, setting the stage for the methods the authors selected to answer the research questions. The objectives of this work are stated clearly.

It would make this paper a lot easier to read if the authors would include a broader description of precisely how d13C CSIA on EAA and NEAA allows the identification of source carbon signatures (lines 81-86), giving a hint to what the reader should expect to find and thus, would make the reader fully understand the differences pointed out in the results. As a suggestion, including what the authors expected to find on CSIA EAA and NEAA would also be nice.

Thank you for your suggestion. We included a section in the Introduction (Lines 83-87) explaining the difference between EEA and NEAA and its implication for the carbon fingerprinting. In the methodology section, we also specified why we focus in EAA and referred to their potential of elucidating shifts in stygofaunal OM incorporations under different rainfall conditions (Lines 267-277). In addition, and for clarity purposes, we differentiated between NEAA involved into Krebs cycle and glycolysis. We incorporated new results in Table 1 and referred to them in the results (Line 321-326) and discussion (Lines 471-481), stressing the need to widen the knowledge about isotopic ecology patterns in invertebrates.

Their data supports a change in feeding habits of several stygofaunal components from different levels of trophic web, thus demonstrating the importance of carbon input in these subterranean ecosystems.

Overall, the text is well structured, the results are very interesting and I believe this article would serve as a good baseline for future work in Australian calcretes. However, to my understanding, some of the assertions in the discussion are not fully supported by the data and some of the results are not fully discussed.

Thank you for your comment. We considerably improved the discussion and support the suggestions provided with bibliographic references (Lines 457-647).

Annotations and suggestions:

Line number: Comment.

Methods: Some specifications would help the reader to have a better understanding of exactly how the sampling took place and would be vital for replication: How deep below the surface is the water table in this region? Does it vary throughout the year? Are 11m enough to get deep into the groundwater? How far is the distance from the water table to the bottom? how much does the net go underwater?

Thank you for your suggestions. We incorporated information about the calcrete (Line 113-115). We also referred to the in press manuscript (Ecohydrology Journal) “Stygofaunal community trends along varied rainfall conditions: deciphering ecological niche dynamics of a shallow calcrete in Western Australia” (reference [38] in the manuscript) for further details about the sampling design, monitoring of water depth and hydrogeological background at Sturt Meadows (Lines 121-122).

78: It is not clear if the authors will use SIA to refer to stable isotope analysis or to bulk material Stable isotope analysis.

Thank you for your comment. We removed “SIA” in the Line 78 as suggested also by the reviewer 2 for clarity purposes.

127: The authors should state how the boreholes were selected (randomly or selected and the selecting criteria). Could there be any bias in the boreholes selected?

Thank you for your notes. We included information about the simple random criteria for the sampling (Line 137) and reference to the previously mentioned in press manuscript (reference [38] for further details). Overall, we included a sampling procedure that involves a third of the accessible and non-dry bores in situ. This extensive sampling effort provides a robust and reliable methodological approach that gives us confidence that the stygofaunal diversity and abundance ranges of the area are representatives and non-biased.

131: In an effort to describe a replicable method, please comment on the methods used to obtain such samples: were the sediments collected from the bottom of the borehole, how were they obtained and what instruments were used?

Thank you for your notes. We included further details about sediment samples collection, separation and pre-treatment (Lines 142-146).

133: The authors should mention why they consider that these 2 boreholes are representative of the rest and how they were chosen.

Thank you for your comment. We included further detailed explanation on how these two bores where selected and why we consider them representative of the hydroecological groundwater dynamics (Lines 150-153).

145: Figure caption requires modification of EEA to EAA as is referred to in the figure. Initials in figure need to be explained: LC-iRMS (liquid chromatography- Isotope Ratio Mass Spectrometry), GC-iRMS (Gas), EA-iRMS (Elemental analysis).

Thank you for your notes. We replaced EEA with EAA in the Figure 2 and explained the analytical acronyms in the figure caption (Lines 172-175).

161: Possibly the authors would use the above defined abbreviation?

Thank you for your suggestion. We replaced the sentence with the above defined abbreviation (Line 190).

164-166: This procedure is missing or unclear in Fig. 2

Thank you for your comment. We included CF (continuous flow) in the acronym – and caption – of the Figure 2 (Line 174).

228: It would help the reader know the number of specimens (n) the authors used in the different sets of analysis.

Thank you for your notes. As reflected in the section “Samples preparation and study design”, all the individuals collected were pooled together and separated into LR and HR. For this reason, instead of considering the number of specimens, for clarity purposes we specified the range of mg per samples considered in each analysis in the text (Line 191,199, 208 and 228) and referred to a newly created supplementary table (Table S2) for further details. This approach is conventionally reported in many isotopic studies (Takano et al., 2010 Rapid Communications in Mass Spectrometry 24.16 (2010): 2317-2323; Chikaraishi et al., 2009 Limnology and Oceanography: methods, 7(11), 740-750) and allows comparison of analytical effort between different laboratories procedures and machine setups.

234: Tukey

Thank you for your comment. We amended the spelling mistake (Line 274).

272-277: I would suggest the authors to be consistent with the abbreviations, probably using NEAA and EAA would eliminate the need for a new abbreviation in this table.

Thank you for your suggestion. We used NEEA and EAA for consistency and clarity purposes along the manuscript.

Check for misspellings in Table 1.

Thank you for your note. We amended the misspellings in the Table 1.

Table 1. Annotation 2: using only Ser, Gly and Ala is not explicitly explained or justified in the text by the authors and is suggested.

Thank you for your comment. We agree with the review that further information is required, and we considerably improved this section according to the suggestion. In the section titled ‘Data treatment and statistical analyses’ we provided details about the two final biochemical pathways (Krebs cycle: intermediates serine (Ser), glycine (Gly) and alanine (Ala); Glycolysis: intermediates) considered for the NEAA analyses and comparison to EAA (Lines 267-272). Results from the new comparison (EAA vs NEAA(Glycolisis)) were integrated in the Table 1, detailed in the results (Line 321-326) and discussed in the section ‘Shifts in OM basal assimilation’ (Lines 471-481).

279: The results shown here should be further explained and clarified to suggest patterns of predators or prey so the readers understand exactly what the authors are pointing out.

Thank you for your suggestion. These results refer to the average values of EAA and NEEA, and their comparison across the stygofaunal community studied. As detailed in the text, δ13C EAA average values were significantly different (and more depleted) compared to NEAA. Aside the mentioned exceptions, same patterns were observed for both preys and predators, so we did not include further explanation/categorisation as suggested. It is worth underlining that EAA are the main focus of the present study for the analysis of potential shifts in OM incorporations. For this reason, the suggested clarification and analysis were provided for EAA as reported in the section spanning from the line 327 to 334. We consider that this section, including the Figure 3 and the improved Table 2, provides clear displaying of the results linked to the EAA patterns and consequent OM shifts (further discussed in the Discussion section).

286-293: This section of results was very nicely explained.

Thank you for your note.

303: Table 2: The readers would appreciate the taxon names along with the ID´s

Thank you for your suggestion. We included the names and ID’s as per Table 1.

320: the authors should identify which is LR and HR within a and b in the caption of Figure 4.

Thank you for your note. We included (a) and (b) in the caption of the Figure 4 (Lines 360-361).

325: I would suggest the authors be consistent with the description style and include the taxon name before the abbreviation ID, as they did in 232-235.

Thank you for your suggestion. For consistency purposes, we ‘formatted’ the taxa descriptions along the MS as ‘species name’ followed by ‘(Sample ID)’.

326: A sample size number (n) would help the readers have an idea of what is a low abundance in these sites, it could help the authors to remember it may be very different in other subterranean environments.

Thank you for your comment. We included the number of total individuals per each one of the two taxa (Lines 365-368), and referred to Table S1 (former Table S2) for the abundance data.

336: table 3 caption: It would help the reader to know the n size in each of the samples, the authors should also specify what the range after each value is referring to (SD, error, variance, etc). The code description under the TP pairwise comparison column is missing in the caption.

Thank you for your comment. We included details about the group lettering and the SD in the Figure caption for clarity purposes (Lines 376-378). As specified in the response to the comment about the Line 228, we refer to the mg of sample employed in each analysis instead of the total number of specimens for clarity purposes. The inclusion of the total number of individuals sampled in this table would expose the reader to the misleading assumption on the totality of the sampled individuals employed for nitrogen CSIA. For this reason, we are inclined not to incorporate the suggested data. However, for clarity purposes we increased the number of citations of the abundance Table (Table S1) along the text in order to avoid misunderstanding between abundance data and isotope samples.

350: the authors should use formal English.

Thank you for your comment. We replaced ‘couldn’t’ with ‘could not’ (Line 391).

356: VPBD and AIR are not explained in caption of figure. I also suggest the authors to include 13C and 15N values of sources in their figures, it can give a better understanding to the reader of how these sources relate to each of the stygofaunal elements and how these may also shift during LR and HR.

Thank for your notes. We detailed VPBD and AIR references in the methodology section (Lines 196-199). We refer to Table S3 in the figure caption for the specific values 13C and 15N values since we think that displaying the data on the graph would duplicate results.

367: I would suggest the authors include the number of specimens obtained.

Thank you for your suggestion. We included in brackets the average value for LR (between LR1 and LR2) and HR (Line 409).

368: LR has been defined as dry season previously.

Thank you for your note. We eliminated ‘dry season’ and phrased ‘during low rainfall conditions (LR)’ (Lines 410-411)

368-377: It appears as this paragraph has been separated or is unclear and the authors need to revise it.

Thank you for your notes. We included the expression ‘Regarding dietary preferences of the amphipod S. axfordi (AM1)’ (Line 410) to link this paragraph to the previous section.

370: AM1 is mentioned in the text and figure caption but AM2 is shown in the figure. Are the differences shown significantly different?

Thank you for your comments. There was an error in the figure. We corrected the name of the species (AM1, S. axfordi) in the Figure 6 and improved the description of the figure caption by reporting that AM2 illustrated the same trends as AM1 (Lines 419-420). Bayesian models do not employ P values but work with posterior probability distributions. As detailed in the methodology section (Lines 303-305), MCMC convergence was tested via Gelman-Rubin diagnostic (with the conventional threshold of 1.1) to investigate the fit of the posterior predictive model. Please refer to Parnell et al., 2013 Environmetrics, 24(6), 387-399 for further details.

377: It would help the reader if this was mentioned along with the reference to fig 6.

Thank you for the suggestion. We included the sentence ‘AM2 illustrated the same dietary preferences as AM1 under both rainfall conditions’ in the caption of Figure 6 (Lines 419-420).

379-381: it would help the reader if the authors would include the taxon name again.

Thank you for your suggestion. We included the taxon names for consistency purposes See response to the comment on the Line 325).

381: The authors are suggested to include the n values in table S1.

Thank you for your suggestion. Please see the previous responses to the comment about line 228 and 336.

382: table S2 has not been mentioned and the authors are invited to consider including it into their results, considering it will be discussed further ahead and the abundance data (and figure) provide interesting information on the community assemblage.

Thank you for your comment. We referred to the abundance data in the Table S1 (former Table S2) in the methodological section (Line 138) so it does not come as a surprise later on the manuscript (as suggested also by the Reviewer 2). However, we are inclined to leave the Figure 10 for the discussion since it mixes both abundance and newly generated isotopic data (TP). Given our study has the main focus to illustrate ecological patterns via isotopic approaches, we consider that Figure 10 provide useful insights that have to be discussed in the discussion section of the manuscript for consistency and clarity purposes.

387: misspelling

Thank you for your note. We amended the misspelling.

388: C and H are not used in this figure.

Thank you for your comment. We homogeneised the coding for copepods and harpacticoids along the manuscript in the figure and figures (Lines 398-399,417 and 434-435). For clarity purposes, we employed the codes (C and H) in the figures and refereed to their taxa Cyclopoida sp. (C) and Harpacticoida sp. (H) in the figure caption. Supplementary Tables included either both the ID code and taxonomic name (Table S1) or just the code ID (Table S2 and Table S3) for copepods (cyclopoids and harpacticoids).

396: There is no sustaining evidence from the authors to claim a change in amphipod behavior in HR; the authors should explain what the amphipod ethology is in LR and HR (and if it differs among species), and further explain how these changes -coupled with increased OM inputs- trigger the species specific predatory response from the P. mesosturtensis larvae (Blv).

Thank you for your comment. We restructured the sentence and eliminated the ethological suggestion (Lines 447-449). Beyond the evident shift from roots to POC, no further information about changes in amphipods’ ethology are available to date. As a result, in this section we limited our dissertation to the outcomes of the Bayesian modelling, and left further suggestion/potential explanations for the final discussion.

398: There appear to be several things to note in figures 7 & 8: S is not present in figure 7, which could mean P. microsturtensis has no cannibalistic behavior (although mentioned earlier) and that other beetles are not feeding on S either (if "B and M illustrated same trends" as noted in fig 7.

B is not present in fig8.b, possibly meaning that Blvs do not feed on adults of their same species during HR. I would suggest the authors to mention these differences in the results section to avoid misinterpretation from the readers.

Thank you for your comment. Figure 7 refers to the dietary proportions of P. microsturtensis (S) and, among the other Paroster species, included only sister species B and M for two main reasons. First of all, no evidence of adult cannibalism (i.e. S adults actively predating/scavenging on other S adults) has been reported in previous studies [40]. With this in mind, we included ‘(and potential active predatory pressures) on sister species’ in the Line 183 to clarify that molecular evidence indicates interspecific - and not intraspecific - interactions. Second, the investigation of cannibalistic behaviours on adult beetles is out of the scope of our manuscript for methodological/analytical reasons. While several studies focuses on isotopic techniques to investigate cannibalistic behaviours, the vast majority of them employ feeding/ethological experiments in the lab (e.g. Venturelli et al., 2006. Transactions of the American Fisheries Society, 135(6), 1512-1522). In these mesocosm experiments, individuals from the same species are usually put in strict contact for a certain amount of time and anomalous nitrogen enrichments are used as an indicator of cannibalistic behaviour. Our study focuses on the functioning of trophic interactions at community level and did not incorporate mesocosm experiments because of both the bias that artificial environments can provide and the lack of genetic data supporting the cannibalistic hypothesis at Sturt Meadows. Indeed, we agree that additional information is key to understand the whole extent of the stygofaunal trophic interactions at Sturt Meadows, and we advocate for further analysis able to bring insights into these complex dynamics (i.e. Lines 636-641 in the discussion).

We also restructured the sentence in the caption of the Figure 7 (Lines 433-441) by specifying that B and M illustrated the same trends of dietary proportion with the exclusion of cannibalism (we considered Paroster prey items M and S for B’s diets, and Paroster prey items B and S for M’s diets).

Stygofaunal contributions to the diet of P. macrosturtensis larvae (Blv) did not consider cannibalism (Blv feeding on B) during HR because the Gelman-Rubin diagnostic reported a higher number than the conventional 1.1 threshold, indicating a not reliable posterior predictive model for P. macrosturtensis adults (B). For clarity purposes, we included this explanation in the caption of the Figure 8 (Lines 452-454).

420-421: POC and DOC would hint the readers to think of carbon instead of matter (and they were also defined earlier in the text).

Thank you for the We removed the expression in brackets for clarity purposes (Line 484).

416-427: Apparently reference [50] says the epilithic biofilms are consumed preferentially from the OM that enters from the surface, and refers to the “high quality” carbon from outside as an assumption, but I could not find their quality measurements in this reference. I would suggest the authors to consider looking at J. Pohlman´s work (1997, 2011) and Brankovits et al 2017 [16] for additional information on "quality" carbon and other production sources in flooded caves. An alternative, would be to consider in situ production, such as it has been observed in other subterranean ecosystems and would probably correspond to a “groundwater microbe functional and metabolic richness” increment after OM recharge event, as demonstrated by [27].

Thank you for your comment. We included the Pohlman, 2011 ([57]) reference and removed the fomer reference [50] from the sentence for consistency purposes. We referred to the in situ production later on along the discussion (Lines 526-537 and 538-545) as we agree that might play a key role in these systems.

435-437: Although it is not this papers scope, is it possible there was an abundance difference within these species among your monitored sites? If so, could these differences be suggesting intraspecific competition?

Thank you for your notes. Interesting point. However, as reported in the MS in press on Ecohydrology (Saccò et al., 2019b) [38]: ‘The community was not distributed differently across the five geological areas and the number of individuals, taxa, Shannon and Evenness indexes did not change significantly according to the geological zones across the different rainfall periods’. Other abundance-based preliminary analyses (unpublished data) also discarded conclusive information about inter and intraspecific competition patterns. Given the opportunism that these species manifest, we argue that other techniques such as eDNA or CSIA could provide useful insights into ecological patterns otherwise very difficult to detect in groundwater systems (Saccò et al., 2019a)[22].

449-451: Including the sources 13C and 15N in the figures shown above, could help the reader follow this description.

Thank you for your suggestion. We consider that reporting the isotopic values of the sources would break the flow of the discussion. d13C and d15N values are reported in the Table S3 and its reference is mentioned in the caption of the figures mentioned by the reviewer (Figure 5,6,7,8).

452: This statement could confuse the readers. Would the authors please expand their explanation on how 13C depleted POC (sources with a greater abundance in 12C) is fresh carbon? Did the authors mean 14C, the replenishment of C sources or something else?

Thanks for your notes. Since we refer to the replenishment of C sources, we used the expression ‘replenished carbon’ in the text (Line 519).

481: Authors should clarify whether these results are from another study or include their environmental data in the supporting information.

Thank you for the comment. We referred to the in press MS (Saccò et al., 2019b)[38] reporting the information referred in the text.

483-484: I would suggest the authors to mention that these amazing beetles actually breathe through their exoskeleton. Some readers are unfamiliar with the Australian stygofauna.

Thank you for your suggestion. We included the suggested information in the Line 569.

479-489: I suggest the authors are consistent throughout the document when referring to the taxon or ID´s of the species.

Thank you for your suggestion. For consistency purposes, we ‘formatted’ the taxa descriptions along the MS as ‘species name’ followed by ‘(Sample ID)’.

500: which species are the authors referring to?

Thank you for your comment. We refer to P. macrosturtensis larvae (Blv) and we improved the sentence for clarity purposes (Line 587).

501: This statement is confusing to the reader, it appears to suggest that feeding on another predator is less nutritious than feeding on a primary consumer, is there supporting evidence?

Thank for your comment. We agree with the reviewer. We removed the part ‘(and less nutritious)’ (Line 588)

502: please include the ID code in the taxon box within figure 8, since the text refers to the code instead of the name.

Thank you for your comment. We improved the Figure 8 by referring to Cyclopoida sp. and Harpacticoida sp as ‘C’ and ‘H’, same as per Figure 7.

502-504: this sentence is a little confusing, it would help the reader if the authors would be more specific in their description of when the increased predatory focus occurred (Hr or LR); Consistency of abbreviations would be clearer: "under HR, Blv...compared to the LR"

Thank you for your comment. We improved the sentence according to the reviewer’s suggestions (Lines 587-592).

505-510: In this paragraph it is not clear when the authors are referring to data from [66] and when they are addressing their own findings.

Thank you for your comment. We added ‘According to their results’ for clarity purposes (Line 594).

510-513: The authors have not made a statement of any estimation of prey density in HR or LR and there is no reference to this data. This is unsupported. -Figure 10 and table S2 need to be mentioned earlier; probably in methods or results. Statistical analysis should be provided for this claim.

Thank you for your notes. Table S1 (former Table S2) is mentioned in the methods section for clarity purposes. However, we are inclined to leave the Figure 10 for the discussion section as explained in the response to the comment on the Line 382. Our suggestion in the (former) lines 510-513 does not involve any consideration of the prey/predator abundances across the rainfall periods, but it rather focuses on biochemically-based observations. In fact, this section of the manuscript (former 510-513) focuses on the interpretation of trophic shifts by considering the isotopical data, presented in the result section, composed by Blv’s dietary contributions, TP estimations and amphipods’ (AM1 and AM2) shifts in OM incorporations. For this reason, we consider that this suggestion is supported and pertinent to the discussion of the results. Taxa abundances were not significantly different between LR (mean between LR1 and LR2) and HR and we included this information in the figure caption. We also referred to Saccò et al., (2019b) [38] for further details about the analysis of abundance patterns along LR1, LR2 and HR (Lines 633-635).

519-527: As a reader who potentially does not know the particular conditions of this site, is it possible that the authors could consider an autochtonous (in situ) primary production that could fuel the trophic web throughout the LR as observed in other studies (16)? –probably a good place in the manuscript to discuss?-

Thank you for your suggestion. For clarity purposes, we included information about potential in situ production in two different sections within the paragraph ‘Shifts in basal OM assimilation’. First, microbial carbonates incorporations/production are inferred from the d13C fingerprints of the sediment samples (Lines 526-537). Second, we suggest alternative nitrogen sources linked to different microbial baselines potentially involving ammonia oxidation metabolisms (Line 538-545).

528: Do the authors refer to natural predators?

Thank you for your note. We replaced ‘enemies’ with ‘predators’ (Line 616).

529: References 77 and 78, though doctoral dissertations, have not been through a peer reviewed process and thus should be used carefully. -78 reference year is 2018, not 2010.

Thank you for your comment. We agree with the reviewer that conclusions from not peer reviewed works should be treated carefully. However, we consider that the two cited works provide interesting insights that improve the understanding of the ecological processes at the studied aquifer. Moreover, being groundwater processes notoriously site-specific, the cited doctoral dissertation provides vital genetically-based information that contribute to comprehension of the trophic dynamics at Sturt Meadows.

547: Being this the only plausible explanation exposed, I would suggest the authors expand their explanation and present or point to supporting information that leads to their hypothesis.

Questions that strike me are tied to their biological reproductive cycle (for both prey and predators), such as discussed below.

Thank you for your notes. We restructured the sentence by stressing that this suggestion is in light of the dynamics described along the manuscript (Line 636). In addition, we have used caution in inferring ethological changes in amphipods, and referred to the ecological changes rather than purely behavioural/biological patterns. We also removed any reference to changes in biological fitnesse for clarity and consistency purposes (Lines 636-641).

569: authors should revise the sentence.

Thank you for your comment. We included ‘methods’ (Line 658) as suggested by the reviewer 2.

Reviewer #2: Sacco et al. use stable isotopes of carbon and nitrogen (both bulk and amino acids) to unravel trophic structure of a stygofaunal community. They find clustering of organisms around trophic level 3 and report a change in source use between a low and high rainfall period. The paper is generally well written and the topic is interesting. I have a few comments I hope will improve the manuscript.

1. I think the intro or methods could use a bit more description of the conditions in the boreholes. How far does light penetrate? Should we expect some primary production in the surface waters that could sink to the bottom, or is it entirely runoff from the surface? Are the copepods that were sampled from near the surface, or are they also in permanent darkness? What about POC samples (surface or dark)? How were the roots and sediment samples obtained, and what does the latter represent? Sediment is a mixture rather than a source. Adding detail on the source pathways will help contextualize the isotope data that come later.

Thank you for your comments. We included further description of the boreholes (Lines 117-118). Light does not penetrate since the bores are capped. In addition, the natural habitat in the calcretes is in permanent darkness and there is nowhere access to light. Spatial information about primary production on site is not available to date and it is limited to results from functional genomics involving proteobacteria (Hyde et al., 2018). Copepods were sampled via haul nets along the water columns, same as per the rest of stygofauna, and therefore no further explanation is provided. Further explanation of the POC sampling methodology is provided (Lines 153-156) for clarity purposes. Roots and sediment were collected through the haul netting procedure (Lines 142-144), the latter representing the particulate sediment from the bottom of the aquifer as stated in the Line 142. We also referred to the in press manuscript titled “Stygofaunal community trends along varied rainfall conditions: deciphering ecological niche dynamics of a shallow calcrete in Western Australia” (Saccò et al., (2019b)) [38] for further details about sampling design, monitoring of water depth and hydrogeological background at Sturt Meadows (Lines 121-122).

2. The number of organisms captured should be expressed as per sampling unit (e.g. # of organisms per tow), since five tows were used at each site at each time. Comparing abundances at different times relies on standardized effort across sites and times. Also, the abundance data should probably be reported in the results section so it doesn’t come as a surprise later (i.e. before line 526 and Figure 10).

Thank you for your comments. Abundance data (Table S1, former Table S2) is referred in the methods section, earlier in the manuscript for clarity purposes (Line 138). Standardised procedures, by one operator, were carried out to maximise the representativeness of the sampling methodology. Five hauls, the most reliable sampling approach at Sturt meadows as reported by Allford et al. (2005), were repeated for every sampled bore to retain specimens along the water column (and not just in the upper layer in close proximity to the borehole case). As a result, the sampling unit considered is composed by 5 hauls and it is not possible to separate the single tows as suggested. While we concord on the necessity to dig into representativeness of sampling procedures in groundwater environments, we also consider that the methodology applied - with the technology and knowledge available to date - provides the rigour and robustness required. After sampling the aquifers over more than 15 years now we have found a general consistency with which bore holes are associated with different taxa and how abundant these taxa are (see Hyde et al. 2018; when taking account of seasonal effects), which provides us with some confidence that our sampling methods are reliable and allow a comparison of relative abundance patterns over time.

3. The sediment has unusually high d13C values, suggestive of carbonate contamination. It is noted that POC samples were acid washed. What about the sediments? Were they also acid washed? Without acid washing, I don’t trust the reported value and it could skew the mixing model very strongly (see Phillips et al. 2014 Can. J. Zool. 92: 823-835).

Thank you for your comment. As per POC samples, sediment samples were hydrolysed to remove the inorganic component. These samples indicated d13C values in the range of dissolved inorganic carbon values in water as reported by several studies in groundwater environments (Lipar et al., 2017. Palaeogeography, palaeoclimatology, palaeoecology, 470, 11-29; Cartwright et al., 2007. Australian Journal of Earth Sciences, 54(8), 1103-1122; Rowe et al., 2000. Palaeogeography, Palaeoclimatology, Palaeoecology, 157(1-2), 109-125.), suggesting that the organic component of the sediment (lipids, sugars, microbes and biofilm) is strictly related to carbonate metabolisms. This is not unusual, as widely detailed by Portillo et al., 2009. Naturwissenschaften, 96(9), 1035-1042. Indeed, we agree with the Reviewer 2 that further CSIA carbon fingerprinting analyses are necessary to unravel the carbon assimilatory pathways in the system. However, provided the whole-system focus of our work, we consider that the improved methodology (Line 117-160), results (i.e. Table 1, Lines 321-326) and discussion (Line 471-481; Lines 526-545) sections on OM incorporations provided in the MS allow untangling of the main biochemical processes shaping Sturt Meadows calcrete aquifer.

4. I find it interesting that none of the organisms analysed for CSIA came back as primary consumers. All were near TL = 3. So who are the primary consumers? This points to microbial conditioning of detrital sources as the likely pathway leading to metazoans. This should be highlighted in the discussion.

Thank you for your comment. Very interesting point. Our isotopically-based data suggest that copepods act as pure primary consumers at Sturt Meadows. We agree with the reviewer that this outcome indicates linkages between primary consumers and microbes. For this reason, in the paragraph ‘Shifts in basal OM assimilation’ we included two sections about potential DIC microbial metabolisms (as suggested by the low d13C values of the decarbonated sediments, Lines 526-537) and the linkage copepods-microbes as suggested by the high d15N values of copepods and harpacticoids under both rainfall regimes (Lines 538-545). Unfortunately, methodological constraints prevented the analysis of CSIA on copepods. Once combined with functional genetic studies, these biochemically-based results will bring further light to the crucial linkage copepods-microbes (Line 546-556).

5. The mixing model did not resolve the diets of adult beetles particularly well (lots of overlap in 95% CIs), so this shouldn’t be stated as a significant difference between seasons.

Thank you for your note. We corrected the wrong reference to the 95% CIs by replacing it with the correct description “Medians and quartiles of each prey category are represented in the boxplot” in the captions of the figures 6 (Line 418), (Lines 435-436) and 8 (Lines 454-455). As a result, the box plots presented in this study do not refer to potential overlapping in dietary proportion. While we agree with the reviewer that the overall diet of P. microsturtensis (S) is quite similar under the two rainfall periods (LR and HR), mixing models revealed considerably different contributions from amphipods (AM1 and AM2) and sister species of beetles (B and M) under LR and HR. We are therefore inclined to report the observed differences due to the fact that they provide interesting insights into trophic shifts (amphipods-based diet vs sister species predation/scavenging) under contrasting rainfall periods.

Specific comments

Line 25 and elsewhere – when using the term “depleted”, always refer to the heavier isotope i.e. “13C-depleted”.

Thank you for your comment. We added “13C-“ to “depleted” along the manuscript.

Line 78 – delete “SIA”

Thank you for your note. We deleted “SIA” (Line 74).

Line 84 – change “signatures” to “isotopes” since signature refers to large well-buffered reservoirs

Thank you for your suggestion. We changed “signatures” to “isotopes”.

Line 140 – delete “the”

Thank you for your comment. We deleted “the” (Line 164).

Line 157 – I was surprised to hear the biofilm referred to as “shredding”. Is that an error? Do microbes shred?

Thank you for your comment. Indeed, it is an error. We restructured the sentence for clarity purposes. (Lines 186-187).

Line 162 – are the copepods not considered to be stygofauna?

Thank you for your comment. We deleted “copepods” for clarity purposes (Line 190). Meiofauna was considered as stygofauna as suggested by the reviewer.

Line 320 – add (a) and (b) to the figure caption

Thank you for the comment. We added (a) and (b) to the figure caption (Lines 360-361).

Line 348 – it is worth noting here that copepods had very high d15N relative to the rest of the food web (not just in HR), despite being assumed to be primary consumers. This implies a distinct N source for these organisms (i.e. different baseline), and it is unfortunate that they weren’t analysed for CSIA. If they truly are primary consumers, then the CSIA would have revealed it.

Thank you for your comments. We agree with the reviewer that copepods revealed anomalously high d15N values compared to rest of stygofauna. As detailed in the response to the main point 4, we discussed about the different baseline and potential metabolisms implied in the discussion section (Lines 538-545). Indeed, CSIA on groundwater copepods would provide key information due to their crucial role in linking microbes and upper stygofaunal taxa. Hopefully, future technological advances will allow analysis of even smaller amount of sample, allowing better understanding of the energy flows in groundwaters.

Line 370 – add “mean and 95% credible interval” to the figure caption

Thank you for your suggestion. We included “Medians and quartiles of each prey category are represented in the boxplot” for the captions of Fig.6, 7 and 8, since the intervals represented refer to the quartiles instead of the 95% credible intervals (and medians instead of means). See response to the main point 5 for further details.

Line 378 – are the depletions here referring to 13C or 15N or both?

Thank you for your note. We added “δ13C and δ15N” for clarity purposes (Line 425).

Line 397 – mean proportions were the same in both seasons so this statement is rather speculative

Thank you for your comment. We included the amphipods’ proportions (AM1 and AM2) under LR (52% in the Blv diet) to the underline their increase under HR (80%) (Line 445). We also restructured the following sentence without considering ethological changes in amphipods AM1 and AM2 for clarity purposes (Lines 447-449).

Line 407 – suggest adding “in that season” after “sources”

Thank you for the suggestion. We added “in that season” for clarity purposes (Line 463).

Line 479 – this paragraph is very speculative and could be deleted without affecting the paper’s interpretations.

Thank you for your comment. Further reference to the in press manuscript “Stygofaunal community trends along varied rainfall conditions: deciphering ecological niche dynamics of a shallow calcrete in Western Australia” (Saccò et al., (2019b))[38] has been added to support the increase in oxygen levels under HR (Line 567) and also sustain the S’s shifts in ecological niche occupations (Line 573). Moreover, we included a new reference to the study carried out by Bradford et al., 2013 [40] when S’s group feeding is mentioned (Line 574). Given all these updates, we consider that this paragraph is much less speculative and can provide interesting insights into S’s ecological patterns.

Line 496 – it might be worth noting that Blv were not at the top of the trophic web when bulk data were considered, perhaps owing to the different baseline d15N value of the copepods

Thank you for your comment. We are inclined to think that biochemical routing could have played a role in shaping this SIA result. Moreover, given CSIA provides a much more accurate pinpointing of the trophic positions along the food chain, we used it as reference and avoided mentioning any SIA-based trophic level characterisation for consistency purposes. Copepods played a marginal role in Blv’s diets (Figure 7) and provided the complexity of the biochemical flows along the stygofaunal community. We consider that further analyses are required to attribute the suggested linkage to different baseline assimilations.

Line 515 – field “of” trophic ecology

Thank you for your note. We included “of” (Line 603).

Line 569 – quantitative isotopic “methods”

Thank you for your note. We included “methods” (Line 658).

Submitted filename: Response to reviewers Sacco et al..docx

Click here for additional data file.

3 Oct 2019

Elucidating stygofaunal trophic web interactions via isotopic ecology

PONE-D-19-21542R1

Dear Dr. Sacco,

We are pleased to inform you that your manuscript has been judged scientifically suitable for publication and will be formally accepted for publication once it complies with all outstanding technical requirements.

Within one week, you will receive an e-mail containing information on the amendments required prior to publication. When all required modifications have been addressed, you will receive a formal acceptance letter and your manuscript will proceed to our production department and be scheduled for publication.

Shortly after the formal acceptance letter is sent, an invoice for payment will follow. To ensure an efficient production and billing process, please log into Editorial Manager at https://www.editorialmanager.com/pone/, click the "Update My Information" link at the top of the page, and update your user information. If you have any billing related questions, please contact our Author Billing department directly at authorbilling@plos.org.

If your institution or institutions have a press office, please notify them about your upcoming paper to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, you must inform our press team as soon as possible and no later than 48 hours after receiving the formal acceptance. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information, please contact onepress@plos.org.

With kind regards,

Jose M. Riascos, Ph.D.

Academic Editor

PLOS ONE

Additional Editor Comments (optional):

Reviewers' comments:

7 Oct 2019

PONE-D-19-21542R1

Elucidating stygofaunal trophic web interactions via isotopic ecology

Dear Dr. Sacco:

I am pleased to inform you that your manuscript has been deemed suitable for publication in PLOS ONE. Congratulations! Your manuscript is now with our production department.

If your institution or institutions have a press office, please notify them about your upcoming paper at this point, to enable them to help maximize its impact. If they will be preparing press materials for this manuscript, please inform our press team within the next 48 hours. Your manuscript will remain under strict press embargo until 2 pm Eastern Time on the date of publication. For more information please contact onepress@plos.org.

For any other questions or concerns, please email plosone@plos.org.

Thank you for submitting your work to PLOS ONE.

With kind regards,

PLOS ONE Editorial Office Staff

on behalf of

Professor Jose M. Riascos

Academic Editor

PLOS ONE