Introduction

Large river floodplain wetland systems have complex food webs that support significant ecosystem goods and services1. However, these large river floodplain wetlands are impacted by extreme climate change, and anthropogenic disturbances such as increased river regulation and water pollution2. The integration of palaeoecology, modern ecology and carbon energy and nutrient mass flow approaches to understand food web dynamics can resolve many key questions of the modern ecosystem structure and function of shallow floodplain wetland systems3,4. This combined approach provides an historic perspective that should facilitate best practices for the management of the hydrology and ecology of regulated river systems5.

The lower Murray River system in southeast Australia has witnessed rapid hydrological transformation over the past century. This has impacted biodiversity and ecosystem structure and the function of wetlands across temporal and spatial scales6. The majority of these wetlands have shown marked declines in submerged aquatic macrophyte communities following the arrival of Europeans in the mid-1800s7. River regulation and widespread catchment modifications for agriculture are implicated in driving these changes8 but waste from mining and settlements9,10,11 has also had considerable impact on most river tributaries. The variability in river flows, from both natural and anthropogenic, together with increased nutrient and sediment fluxes and a subsequent decline in water clarity significantly reduces underwater light and photosynthesis7,8,12.

The cascading effects of eutrophication and disruption in the connectivity with the river system have driven unprecedented changes in floodplain wetland ecosystems13. The changes include decline in the water level of some wetlands, and perennial inundation of others followed by the partial or complete loss of sub-merged vegetation as well as the proliferation of fringing emergent macrophytes (Typha orientalis, T. domingensis and Phragmites australis)14 and floating plants (e.g. Azolla sp., Lemna sp.) and the invasion of exotic carp (Cyprinus carpio)15. Alteration of carbon and nitrogen sources in the system have modified the food web structure and trophic pathways16. Particularly, the increased disturbance of terrestrial and pelagic or periphytic habitats has altered carbon energy and nitrogen mass flows from the base of the food web to higher trophic levels17,18.

Various monitoring approaches have been employed to better understand the changes in floodplain wetland ecosystems in southeast Australia to inform management practices designed for their restoration. For example, the water requirements of riparian forests has been quantified to mitigate the impact of water resource development and drought on keystone river red gum communities19. The macroinvertebrate community composition and densities have been monitored to assess low flow impacts as they are seen as critical in the successful recruitment of native fish and water birds20. Similarly, palaeolimnological approaches have been used to assess the changes in water quality due to excessive nutrients such as total phosphorous (TP) and total nitrogen (TN) loads and salinization and acidification (pH) in the region over the past century21. However, these approaches have not explored the changing nature of food web dynamics of the Murray River floodplain wetland system over the recent past.

Palaeoecology provides evidence of benchmarks for undisturbed ecosystems, and can show how the system has responded to change following a disturbance, while modern ecology defines current system behaviour, including linkages across trophic pathways22,23,24,25. The carbon energy and nutrient mass flows in floodplain wetlands provide information on consumer food sources and trophic changes over time26,27. The combination of subfossil cladocerans, such as chydorids, bosminids and daphnids, and the stable isotope analysis (SIA) of carbon and nitrogen in primary producers (e.g. aquatic macrophytes) and primary and secondary consumers (zooplankton and fish), is likely to prove a useful tool for understanding the contemporary condition of floodplain wetlands which can be used to underpin a management framework in the lower Murray River system5,28. The variability in the conditions can be achieved through the use of mixing models that provide critical information regarding changing food sources and predator–prey interactions25,26,29,30. However, the dynamics of stable isotopes of carbon and nitrogen across the trophic levels are complex. This is largely dependent on the type and pool source of the organic matter and the fractionation of elements during the movement from lower to higher trophic levels31.

In this study, we have integrated palaeoecology and palaeo-and-modern food web approaches to assess recent changes in Kings Billabong, a shallow regulated floodplain wetland, in the lower Murray River system of southeast Australia, by using carbon and nitrogen stable isotopes. We hypothesize that wetland impoundment due to river regulation by early European immigrants has led to the collapse of sub-merged vegetation leading to a change in the source of carbon. The integration of cladoceran palaeoecology, the SIA of carbon in subfossil Daphnia and bulk sediment, as well as SIA of carbon and nitrogen in contemporary submerged and emergent macrophytes, and modern Daphnia and fish, provides key evidence for the nature of wetland change including nutrient (carbon and nitrogen) movements across the trophic levels.

Materials and methods

Study site

Kings Billabong is a shallow (max. depth = 2 m) wetland of the lower Murray River system in northwest Victoria (34° 11′ 0″ S, 142° 09′ 0″ E) (Fig. 1). The ecosystem of Kings Billabong was likely to have been modified after it was used as a water storage basin from the early 1890s32,33. Construction of the pumping station on the River at Psyche Bend, in 1891, enabled the filling of Kings Billabong for the emerging irrigation industry. Prior to the pumping station the connectivity of Kings Billabong with the Murray River was natural, influenced largely by regular flooding and drying patterns. After 1891, the Billabong became permanently inundated for the first time with pumping from the river maintaining water levels during dry times. Flow regimes in the Murray River channel itself were interrupted severely following the phase of river regulation after the construction of locks along the river during the 1920–1930s34. The regulation triggered changes in the condition of aquatic ecosystems and the floodplains through both reduced overbank flooding and the permanent filling of wetlands linked to weir pools. For example, the river red gum (Eucalyptus camaldulensis) forests around the Billabong began to die through inundation19. There was also a gradual decline in the cover of the submerged aquatic plant species such as, Vallisneria americanus and charophyte species such as Nitella sonderi, N. hyalina, N. stuarti, and Chara gymnopytis (unpublished subfossil records) as well as the development of dense stands of Typha domingensis and Phragmites australis around the wetland margins35. Over the decades after river regulation, the relatively shallow Kings Billabong did not respond abruptly, but underwent a gradual and sustained phase of ecosystem degradation33. The slow and sustained response in ecological state to the gradual change of the wetland environment led to the collapse of the submerged vegetation stands during the late twentieth century33. The first sign of change was visible with the growth of Typha sp. during the 1950s. The aerial photos show the areas of emergent macrophytes began to expand rapidly around the wetland margins during the 1970s.

Figure 1
figure 1

Map of Kings Billabong wetland showing the deepest point (2.3 m) where the sediment core was collected. The shaded areas are the areas of water, green areas are catchment emergent aquatic macrophytes and terrestrial catchment vegetation.

Analysis of subfossil cladocerans from the sediment core

A 94 cm long sediment core was retrieved from the centre of Kings Billabong wetland in 2011. A central core was taken to represent conditions across the wetland and the sediments proved to preserve microfossils of both littoral and planktonic cladoceran zooplankton36. The core was then subsampled at high resolution (1 cm interval) for subfossil cladoceran analysis. Sub-samples were treated with 10% KOH solution, and heated at 60 °C on a hotplate for at least 45 min. The subsample mixture was then sieved through a 38 µm mesh. Safranin was used to stain the remains before the slides were prepared42. Cladoceran remains were counted under 400 × magnification and identified and described following37,38. Cladoceran species were classified as ‘littoral’ and ‘planktonic’ according to their preference to water quality and habitats and the grouping was tested using constrained cluster analysis (CONISS)39.

Collection and analysis of subfossil ephippia of cladocerans (Daphnia)

Subfossil Daphnia ephippia were used for the analysis of δ13C and δ15N. The ephippia reflect the stable isotopic composition of the parent Daphnia, their diet and the environmental water in which they have lived40. The δ13C of Daphnia ephippia in different seasons has also been found to match the δ13C of their body parts41. Based on the assumption of the close relationship of the δ13C between Daphnia ephippia and their body parts, altogether six subfossil Daphnia sub-samples were analysed. These were taken at the depth intervals of 0–10 cm, 10–20 cm, 20–30 cm, 30–40 cm, 40–50 cm, 50–60 cm, respectively from the same core used for subfossil cladoceran analysis32. Prior to picking Daphnia ephippia, the sediment was treated with 10% KOH as described in Kattel et al.42. Daphnia ephippia were identified under a binocular microscope following43.

Collection/preparation of modern aquatic emergent macrophytes, fish and Cladocera (Daphnia)

For the SIA of C and N of fish: Murray cod (Maccullochella peelii), common carp (Cyprinus carpio) and Australian smelt (Retropinna semoni) were collected using gill and seine nets in March 2014. Living Daphnia were collected from Kings Billabong wetland using hand-nets (100 µm mesh size) from both open and littoral areas on the day the fish were collected.

For vegetation samples, both upper and lower sections of the stem and leaves of the two dominant stands of emergent aquatic macrophytes: Typha domingensis and Phragmites australis, and one sub-merged charophyte species, Nitella sonderi, were collected from the littoral zone at the same time. The upper and lower sections of the plantstand and leaves were selected to ensure that the isotopic compositions would be representative of the whole plant. Each sample was washed and transported to the laboratory.

Murray cod and common carp, once caught by gill nets, were placed live in a submerged cage overnight, while the Australian smelt was collected using seine nets, and transported to the laboratory. Live Daphnia were placed in an aquarium for 24 h to void their gut contents. Cladoceran zooplankton were acid washed with 10% HCl and rinsed in distilled water to avoid contamination of the organic carbon from the carbonate fraction44. All biological materials were oven dried at 30 °C for 72 h before they were pulverised and submitted for SIA32.

Dating of sediment core

The biological remains, including subfossil cladocerans, were thought to be mixed in the water column before being deposited in the sediment at 2–3 m depth in Kings Billabong, where the 94 cm long 80 mm piston core was collected in 2011. Rather than from the lake centre, the core collected from the deepest point, as the deepest area of the wetland is usually the best for studying palaeoecology due to the integration of both littoral and pelagic cladoceran zooplankton microfossils36.

The 210Pb-based age model was established from a total of nine subsamples45 down to 51 cm in the sediment core based on both constant initial concentration (CIC) and constant rate of supply (CRS) models. The age-depth model for the rest of the core was established using extrapolation methods46. This analysis was undertaken at the Australian Nuclear Science and Technology Organization (ANSTO), Lucas Heights, Sydney. More details of dating method is described in a previous study on Kings Billabong33.

Analysis of δ13C and δ15N of emergent and sub-merged macrophytes, fish, and modern and subfossil cladoceran zooplankton

Each dry sample (both modern macrophytes, Daphnia, fish and subfossil Daphnia), weighing 100–200 µg, was packed into a small tin cup for isotope analysis. The subfossil Daphnia ephippia samples were oxidised and the resultant CO2 and N2 analysed with a Finnigan Delta Plus mass spectrometer interfaced via a Conflo II to a NC2500 Elemental Analyzer. The δ13C, δ15N isotopic values, were determined using a Thermo Finnigan Flash EA 112 interfaced via a Thermo Conflo III with a Thermo Delta V Plus IRMS47,48. Samples of acetanilide of known isotope composition were analyzed within each run to verify isotope values. Reproducibility for both δ13C and δ15N was 0.2‰ while that for both %C and %N was 1%47,48. Ratios of 13C/12C and 15N/14N were expressed as the relative per mil (‰) difference between the sample and conventional standards (PDB carbonate and air N2, respectively). These ratios were expressed as: δX = [R (sample)/R (standard) − 1] × 1,000 (‰), where X = 13C or l5N and R = 13C/l2C or l5N/14N32.

The modern Daphnia isotopic values of nitrogen were derived from the samples collected in 2014, while the subfossil Daphnia isotopic values of nitrogen were from the top sections (0–10 cm) of the core that dated back to c. 2006–2011.

Analysis of δ13C and δ15N values of bulk sediment

Analysis of δ13C and δ15N values was also carried out to compare with the stable isotope values of C and N of subfossil Daphnia and chironomids. The method for the analysis of δ13C values for 45 bulk sediment samples of the core can be found in detail in Kattel et al.32. Details of the analysis of δ15N for the same number of bulk sediment samples can be found in Kattel et al.33. Samples, once weighed (c. 100–200 µg), were treated with 1 mL HCl to remove carbonates and then analyzed on an Elementar VarioMICRO Elemental Analyser and a Continuous Flow Isotope Ratio Mass Spectrometer (GV Instruments IsoPrime). The acid-treated fraction and results of δ13C were normalized to the reference standard IAEA C8 (consensus value: δ13CV-PDB = − 18.31‰)49. All δ15N analyses were performed on the untreated fraction and a two point normalization was performed on the data using international reference standards IAEA N-2 (consensus value: δ15NAIR =  + 20.3‰ and USGS-25 (consensus value: δ15NAIR = − 30.4‰)50.

Food web mixing model

An integrated palaeoecological, and palaeo-and-modern food web analysis approach was used to test the changes of wetland trophy, and the trophic pathways of the food web in Kings Billabong. The subfossil assemblages of cladocerans were used to infer trophic changes of the wetland before and after river regulation. However, key questions remained and these are the focus of this study, namely how the carbon and nitrogen sources of cladocerans in the past compare with those at present, , and how the palaeo- and modern cladoceran diets differ with respect to basal food sources and how are the isotopic signatures of carbon and nitrogen transferred to upper trophic levels of the ecosystems. Hence, the contemporary and past responses of wetland ecosystem state and food web dynamics were evaluated by using a mixing model of stable isotopes of carbon and nitrogen from both modern macrophytes, cladoceran zooplankton (Daphnia), fish (large vs small) and subfossil Daphnia51. The linkages among cladocerans and other biota, including fish and macrophytes, were then established by applying stable isotope values of carbon and nitrogen of each organism to a mixing model.

Mixing models are mathematical equations (Eqs. 1 and 2), often applied in analyses of carbon and nitrogen stable isotopes. The models describe the observed isotopic composition as a mixture of the assimilated isotopic compositions based on the isotopic mass balance29. Individual isotopic values of δ13C and δ15N are measured in primary producers and consumers using mixing models. For example, the mixing (combined) values of the carbon isotopes are given in the equations below:

$$\begin{aligned} & \delta 13Cmix = f1\delta 13C1 + f2\delta 13C2, \\ & f1 + f2 = 1 \\ \end{aligned}$$
(1)

where, δ13Cmix is the isotope signature of the consumer, usually a combination of the δ13C of individuals (e.g. subscripts of Carbon i.e. 1 and 2 are combination of two individuals usually prey 1 and prey 2), weighted by their unknown diet fractions (f1 and f2, respectively). Assuming that the combination of diet fractions is 1, then f1 and f2 are usually estimated algebraically by rearranging the equations to:

$$\begin{aligned} & f1 = \frac{{\left( {\delta 13Cmix - \delta 13C2} \right)}}{{\left( {\delta 13C1 - \delta 13C2} \right)}}, \\ & f2 = 1 - f1 \\ \end{aligned}$$
(2)

The isotopic fractionation across trophic levels is much smaller for carbon isotopes (range = 0–l‰ per trophic level)52. The measurement of mean δ13C values primarily indicate the basal source of carbon generated by primary producers in the food web that is important for the energy of higher-level consumers53,54. Hence, the δ13C values are used for the information about consumer food preferences. Unlike carbon, the isotopic fractionation for nitrogen is much larger (range = 1.3–5‰ per trophic level), hence the measurement of mean δ15N is used as evidence for the number of trophic levels present in a food web55.

Results

Age modelling

There was a very small variation between the supported 210Pb (sediment containing a background level of 210Pb by the decay of 226Ra eroded from rocks) and unsupported 210Pb (sediment already with excess of 210Pb, occurring through the decay of 238U followed by 222Rn in the atmosphere then to 210Pb, washed out of the atmosphere) (Table 1). The constant rate of sediment accumulation model (CRS) suggested that, at 51 cm sediment depth, the rate of accumulation was 0.6 g cm−2 year−1 yielding a corresponding age of c. 1970 (Fig. 2a,b). A linear extrapolation approach46 of SAR showed that the sediments at 75 cm were deposited in c. 1930.

Table 1 The 210Pb dating of the sediment core collected from Kings Billabong.
Figure 2
figure 2

(a) King Billabong core unsupported 210Pb activities plotted against depth, (b) King Billabong core CIC and CRS model ages plotted against depth.

Palaeoecology of cladocerans

Change in the ecosystem of Kings Billabong wetland was revealed by shifts in the composition of assemblages of subfossil cladocerans over time (Fig. 3). As the wetland was used as a water storage basin from c. 1891 the record of change commenced from this time. This early phase of inundation supported a diverse cladoceran fauna that reflects a healthy community. Subfossil cladoceran abundances suggest relatively clear water existed in Kings Billabong wetland between c. 1891 and c. 1930 (Fig. 3). There appears to have been limited influence of the pumping of water into Kings Billabong on the wetland condition during this time.

Figure 3
figure 3

Trophic changes of Kings Billabong revealed by the percentage abundance of the most common planktonic and littoral cladocerans. The assemblage of planktonic species prior to river regulation (1930) is low. River regulation caused hydrological shift of the Murray River causing (low littoral:planktonic ratios) as well as the composition of B. meridionalis has increased. Among littoral cladocerans, the assemblage of turbid water preferring species mostly small Alona, A. guttata, and other species, A. quadrangularis, Biapertura longispina and C. sphaericus all have increased, but the composition of the pioneer clear water macrophyte preferring species, D. crassa, declined from c. 1970. The post 1990 shows a sign of reduced water quality due to increased turbidity revealed by increased cladoceran ephippia. Depth information in the y-axis (in cm) is from the sediment core.

The principal evidence of clear water during this period is the high relative abundance of the littoral cladoceran, Dunhevedia crassa, a species preferring clear water conditions. Despite this prevailing state, relative to D. crassa, species tolerant of turbid water condition, such as the planktonic species Bosmina meridionalis, and the littoral forms Alona quadrangularis and Biapertura longispina, showed an upward trend of relative abundance for a short period indicating the response to the occasional flooding of the Billabong and subsequent inundation leading to a nutrient-rich environment, albeit temporarily (Fig. 3).

The most obvious change detected in the subfossil cladoceran assemblages coincides with river regulation from c. 1930. The aquatic ecosystem of Kings Billabong from this time began to change steadily. This phase of degradation was sustained until 1970. While this wetland was artificially inundated from 1891, the onset of this change in the ecology of the wetland occurred particularly after the modification of the natural hydrology of the Murray River itself from ~ 1930, when the effects of river regulation were established. From this time, the abundance of D. crassa began to decline, while the abundance of small Alona, such as Alona guttata and the planktonic B. meridionalis, increased. This trend was also matched by sustained declines in the littoral to planktonic (L:P) ratio and inferred reduction in water clarity (Fig. 3).

By c. 1970, the abundance of the littoral, plant-preferring species D. crassa markedly declined reflecting a significant loss, or possibly a ‘collapse’, of sub-merged aquatic vegetation in Kings Billabong. A rapid growth and development of an emergent and fringing macrophyte community, such as Typha sp. and Phragmites sp., as well as increased invasion of common carp, was recorded at or around this time56,57. The increasing ecological stress, such as the disappearance of submerged macrophytes, which began in 1930, followed by prolonged eutrophication was reflected by the highest concentrations of subfossil cladoceran ephippia in the sediment from this time until the late 1990s (Fig. 3).

The δ13C values in subfossil Daphnia, subfossil chironomids and bulk sediments

The δ13C values of subfossil Daphnia, subfossil chironomids and bulk sediment in the Kings Billabong wetland sediment core began to decline after c. 1910s (Fig. 4). All three proxies showed a relative depletion of δ13C values along with the gradual loss of the sub-merged macrophyte community over time (1910–2006). The δ13C values in bulk sediments were the highest followed by those from subfossil chironomids and subfossil Daphnia. After river regulation, the δ13C values in both consumers depleted reaching as low as − 29.5‰ in the 2000s. However, unlike the subfossil Daphnia, the subfossil chironomid δ13C values showed a relatively close relationship with the bulk sediment δ13C values (Fig. 4).

Figure 4
figure 4

The δ13C values extracted from subfossil Daphnia ephippia, chironomid headshields and bulk sediment samples from Kings Billabong. A marked shift in carbon source is revealed over time in all three proxies. From 1910 to 1930, the source of carbon was relatively high (δ13C > − 23‰). From 1930 to 1970 and 1980 the carbon source continued to decline, and was variable among the proxies; and after 1980 the carbon source further depleted in all proxies. The depleted δ13C values also corresponded to the change in the dynamics of subfossil littoral cladoceran (e.g. D. crassa) assemblage as shown in Fig. 3 reflecting the possible reduction in the sub-merged macrophyte-derived carbon source for primary and secondary consumers.

Food web mixing models

Food web mixing models of fish, modern and subfossil Daphnia and submerged and emergent aquatic macrophytes revealed different relationships among consumers and primary producers in Kings Billabong (Fig. 5a,b). According to the isotopic fractionation of carbon (~ 1‰) and nitrogen (~ 3.4‰) at each trophic level58, the δ13C values between modern Daphnia and a submerged macrophyte (N. sonderi), and the δ15N values between modern Daphnia and Murray cod, showed a relatively close trophic relationship (Fig. 5a). The difference in the mean δ15N values between modern (8.9‰) and subfossil (4.6‰) Daphnia was apparent as obviously the modern Daphnia utilized more enriched nitrogen sources in Kings Billabong over recent time than the subfossil Daphnia 5 years prior.

Figure 5
figure 5

Isotopic representations of the mean (± SE) of δ13C and δ15N in Kings Billabong derived in mixed models. (A) comparison of individual species of fish (Australian smelt, Murray cod and common carp), Daphnia (modern and subfossil), sub-merged macrophyte (N. sonderi) and emergent macrophytes (Typha and Phragmites); (B) comparison after all fish species, all Daphnia and all emergent macrophytes are combined.

The mean δ13C values of all fish species, Daphnia, emergent and submerged macrophytes showed that the transfer of carbon from the base of the food web to Daphnia may have occurred mainly from the submerged macrophytes community, while the mean δ15N values in most biota revealed that submerged and emergent macrophytes and Daphnia, all may have contributed nitrogen masses to the fish community in Kings Billabong (Fig. 5b).

Discussion

Changes in the ecology of Kings Billabong

The integrated approach of analysing modern and subfossil biotic assemblages and stable isotopes of carbon and nitrogen in different biota and bulk sediment reveals a sustained change in the ecology of Kings Billabong after the artificial inundation of the wetland in 1891. The timing of the onset of sediment accumulation coincides with the pumping of water from the nearby Psyche Bend into the Billabong for irrigation purposes. The onset of pumping converted the wetland from a naturally intermittent to a permanently inundated basin allowing for the preservation of a continuous sediment sequence. Most intermittent sites in the region today are dominated by semi-terrestrial plants and aquatic communities that grow rapidly once inundated. The water quality effect on those sites is usually localized and dependent on the degree of disturbance, both anthropogenic and climatic.

1891–1930 AD

The onset of permanent inundation, together with rainfall events during the mid-1890s, was a significant change enabling the protection of soft sediments in Kings Billabong59,60. A relatively stable, clear water state persisted, even during the Federation drought (1895–1903), between 1891 and 1930, but the episodic floods in 1917 kept the Billabong wet supporting a relatively resilient hydrology61. However, climate, and the deliberate pumping of water into the basin allowed gradual sedimentation and the accumulation of biological and chemical remains in the bottom sediments as revealed by both the dated sediment core and preserved cladoceran subfossils along with their carbon and nitrogen isotopes. The relative abundance of littoral taxa such as D. crassa, and the increased littoral to planktonic (L:P) ratios of cladocerans during this time (Fig. 3), suggest that both the Murray River and Kings Billabong were dominated by a submerged macrophyte community with high water clarity33. The pumping of river water via the Psyche Bend during this period played an important role in maintaining the water level as well as supporting diverse floral and faunal communities, and retaining subfossil cladocerans, and suspended sediments including carbon and nitrogen in Kings Billabong.

1930–1970 AD

After 1930, the influence of anthropogenic disturbance in the Kings Billabong wetland ecosystem increased. A high abundance of B. meridionalis, with reduced density of littoral cladocerans, are indicative of increased disturbance with the onset of changes in water quality (Fig. 3)62 which coincided with river regulation and land use intensifications across northwest Victoria63. Onset of riverbank erosion caused by forest clearances and increased household and animal wastes led to the variability in the flux of sediments including carbon and nitrogen into the wetland. However, efforts were also made in waste management practices from the local government to avoid water pollution further. The rate of ecosystem change in the wetland during this time was relatively slow indicating ecological resistance and the integrity of the source water from the Murray River33. Natural ecosystems often show resilience until an abrupt change happens64. However, gradual change in the ecosystem condition with dissolved organic carbon (DOC) concentration and a reduced light regime can lead to the loss of submerged macrophytes and diversity in the invertebrate faunas. In Kings Billabong, the cumulative impacts of drivers were not strongly visible at this period. However, the 1930s, the time of the major phase of weir construction for river regulation, was the point at which the system began to change. The water quality, ecosystem and food web of the wetland, all responded to the impact of river regulation. The reduction in the δ13C values in Daphnia after 1930 (Fig. 4) is an indication of the simplification of the food web that resulted from a gradual switch of Daphnia food sources derived from different primary producers.

The dynamics of carbon at the base of the food web in shallow lakes is influenced largely by the submerged macrophyte density and the periphyton attached to it, and phytoplankton. The primary producers utilize available dissolved inorganic carbon (CO2 or HCO3) with varying isotopic signatures in the wetland pool during photosynthesis18. In the presence of submerged macrophytes, periphyton has a wider boundary layer and corresponding reduced CO2 uptake than free floating phytoplankton, and as such, have enriched 13C isotopes25. This is due to increased fixation of HCO3 rather than CO227 a phenomenon reflected in Kings Billabong prior to the 1930s, when there were abundant submerged macrophytes. However, when the wetland went through a change to lower to moderate primary productivity, algae were able to utilize the lighter 12C isotope preferentially from the available CO2 pool resulting in the depleted δ13C31. Most suspended organic matter, composed of phytoplankton, plant detritus, and riverine organic matter, was likely to have been the diet of Daphnia in Kings Billabong after the 1970s as indicated by depleted 13C27. Changes of the dynamics of the riverine organic matter derived from C3 and C4 plants from the catchments can also alter δ13C values in sediment over time65. Further, in simplified food webs, trophic links between primary producers and primary consumers are reduced66, and the 13C enriched macrophyte may be replaced by 13C depleted phytoplankton under hydrological regulation67. As a result, consumers, such as Daphnia, are forced to rely on single and poorly resourced phytoplankton food27,66. This is also consistent with previous studies that show Daphnia to switch their diet when food sources in the system change due to natural and anthropogenic disturbances17,68,69.

1970–2011 AD

This period experienced a further gradual decline in ecosystem condition with the further shift to 13C carbon depleted-food sources to Daphnia. By the 1980s the food web was represented by highly depleted 13C in Daphnia and by 2011, the Kings Billabong ecosystem had shifted to one driven by low quality diet such as phytoplankton and fringing macrophytes.

Food web dynamics: causes and consequences

The submerged macrophytes at the base of the food web of the early Kings Billabong ecosystem played an important role in supporting littoral cladocerans18 and carbon and nitrogen dynamics across the trophic levels. In the presence of sunlight, the fish, cladocerans and vegetation interactions and external nutrient inputs maintained ecosystem processes. Sun light increases underwater photosynthesis and carbon and nutrient recycling70 while Daphnia utilizes both littoral and planktonic habitats for food and to avoid predators32,71,72. In Kings Billabong, the macrophyte-derived DOC transferred to higher trophic levels via Daphnia and other zooplankton in the post 1930s, was different to that in the pre-1930s25. During the post 1930s, and particularly after the 1970s, the decline of the submerged littoral vegetation community triggered a shift in the food web structure. Low abundance of D. crassa, a littoral cladoceran taxon, indicates a profound disruption in the base of the food web and fish diet after the 1930s. Studies suggest that exotic common carp in the Murray river system intensify submerged macrophyte losses, increase competition for food among native fish community15 and alter predator–prey interactions73. The reduced δ13C values in subfossil Daphnia, subfossil chironomid and bulk sediments over time indicate the rapidly changing nature of food sources at the base of the food web (Fig. 4). Relatively higher values of δ13C in bulk sediment, compared to that in subfossil Daphnia and chironomids (Fig. 4), suggest that the sediment carbon in Kings Billabong is composed of a combination of allochthonous (catchment C3 and C4 plants) and autochthonous organic matter that accumulated in the catchment over a longer period65 and is weakly associated with Daphnia. Primary producers and other allochthonous organic matter pools in the seston and sediments often confound δ13C and δ15N measurements of the consumer food source74. Climate change also intensifies food webs and top-down and bottom-up processes by affecting consumer physiology and their feeding behavior with effects on resource nutrient contents including the stable isotopes of carbon and nitrogen69.

The widespread deforestation of eucalypt trees over the past century, increased erosion and sedimentation, variation in the carbon production75,76, interruption in the river-wetland connectivity after 1930 and subsequent resource movements within the wetland, destabilized the food web processes77. In the mixing model (Fig. 5), the δ13C values of Daphnia and zooplanktivorous fish (e.g. Australian smelt) show a very weak relationship indicating a switch of diet by the fish. If Australian smelt were predated by exotic common carp, as revealed by the δ13C values of the carp, the consequences in the trophic interactions between zooplanktivorous fish and Daphnia, and subsequent change in the food web structure, would not be easily verified. Differential feeding and predator avoidance behavior of Daphnia, and various environmental factors including river regulation and climate change, likely led to a switch of food web in Kings Billabong over time by replacing 13C enriched submerged macrophytes with 13C depleted phytoplankton (Fig. 5). The increased density of emergent macrophytes T. orientalis, T. domingensis and their δ13C values do not show a relationship with Daphnia and fish (Fig. 5a) suggesting that the mechanism behind the transfer of carbon from these plants is yet to be established.

The δ15N values of primary producers and consumers suggest that modern Daphnia may be receiving nitrogen from emergent macrophytes and this contributes to higher trophic levels including those of Australian smelt and Murray cod (Fig. 5a). The increase, over 5 years, to higher mean δ15N values in modern Daphnia, compared to that in subfossil Daphnia indicates a rapid nutrient enrichment shift and changes in the wetland ecosystem and its trophic interactions. Lakes with rapidly changing ecosystems, and increased concentrations of TP, elsewhere also show higher values of δ15N78. However, the degree of trophic interactions, as a result of nutrient enrichment in Kings Billabong, is not known. For example, the δ15N values of emergent macrophytes are the lowest suggesting the source of nitrogen to Daphnia could also be from plants other than the emergent macrophytes. The Daphnia population also appears to be increasingly vulnerable due to changes in the nutrient status of the wetland in recent times. There is a possibility that the replacement of Daphnia by other less vulnerable prey may lead to further predation intensification of common carp resulting in more stabilized trophic levels79. Such changes in the condition of the Kings Billabong ecosystem require further investigation.

Implications for management

A sustained transformation of water quality and ecosystem of Kings Billabong over the past century is indicative of the need of strong measures for wetland management in southeast Australia. Although the response of cladocerans, as revealed by their subfossil records in sediments, is attributed to a localized condition change, the changes are also likely to be associated with catchment level or basin scale modifications of land and water including the introduction of invasive species of common carp causing increased sediments and nutrient fluxes into the Murray River system, sediment resuspension and subsequent reduction in water clarity of the connected wetlands.

Like the proxies used to indicate wetland condition elsewhere21,80,81, our records suggest that the Kings Billabong ecosystem was once rich in biological diversity. The degradation of the ecosystem, the reduction in the diversity of its flora and fauna and the simplification of its food web is likely not attributable to proximal causes. Locally managed ecosystem practices may mitigate the drivers of change to a certain degree but the quality of the river itself remains the main cause of the Billabong’s decline. The trophic dynamics between producers and various consumers suggest that the gradual decline in the ecosystem, began in 1930, and owing to the recent changes, appears not to have reached to an endpoint yet. Being once intermittent, but now permanently inundated, the mere provision of water will not lead to the recovery of the diversity this wetland once supported. Management outcomes that may recover the rich community of the past will include a strong focus on recovering stands of submerged plants by improving the quality of the Billabong’s water, it being nourished by the Murray River catchment, by mitigating the nutrient and sediment flux.

The approach used in this study provides evidence of ecosystem processes and food web dynamics that can have implications for the management of wetland ecosystems. The δ13C and δ15N mixing models derived from modern and subfossil biota reveal much about the contemporary and past dynamics of the regulated river system in southeast Australia. More broadly, the approach also reveals in detail the nature of the switching of trophic positions, of primary and secondary consumers, when exposed to external drivers including nutrients, hydrological modification, sedimentation and invasion.