Effect of elevated magnesium sulfate on two riparian tree species potentially impacted by mine site contamination

Globally, mining activities have been responsible for the contamination of soils, surface water and groundwater. Following mine closure, a key issue is the management of leachate from waste rock accumulated during the lifetime of the mine. At Ranger Uranium Mine in northern Australia, magnesium sulfate (MgSO4) leaching from waste rock has been identified as a potentially significant surface and groundwater contaminant which may have adverse affects on catchment biota. The primary objective of this study was to determine the effect of elevated levels of MgSO4 on two riparian trees; Melaleuca viridiflora and Alphitonia excelsa. We found that tolerance to MgSO4 was species-specific. M. viridiflora was tolerant to high concentrations of MgSO4 (15,300 mg l-1), with foliar concentrations of ions suggesting plants regulate uptake. In contrast, A. excelsa was sensitive to elevated concentrations of MgSO4 (960 mg l-1), exhibiting reduced plant vigour and growth. This information improves our understanding of the toxicity of MgSO4 as a mine contaminant and highlights the need for rehabililitation planning to mitigate impacts on some tree species of this region.

Mining activities have significantly impacted terrestrial and aquatic ecosystems at both local and regional scales [1][2][3] . Critical to minimising these impacts is appropriate management of waste rock or spoil, that is, the vast quantities of extracted material remaining following segregation and removal of the relatively small amount of the desired substance 4 . Oxidation and weathering of exposed waste rock material can result in acidic and/or sodic leachates 5,6 . Many elements can increase in concentration in groundwater passing through the rock and potentially contaminate the receiving environment. For example, chlorite schists can result in leachate with elevated levels of sulfate (SO 4 ), bicarbonate (HCO 3 ), calcium (Ca) and magnesium (Mg) 5,7,8 . Long-term site rehabilitation requires knowledge of the tolerance of the receiving environment to run-off contaminants and appropriate management to minimize potential ecological impacts 9 .
Waterways have been impacted from off-mine pollution [9][10][11] . For example, the U.S. Environmental Protection Agency (USEPA) stated in 2000 that 40% of headwater streams in the western USA were polluted from mining 12 . These streams had significant impacts resulting from the influence of surface or groundwater that flowed through and chemically interacted with rock waste piles 13 , affecting both instream communities and adjacent riparian habitats [14][15][16][17] . There has been increasing focus on the prevention, management and mitigation of the potential impacts of mining on rivers and their riparian ecosystem 18 . A healthy riparian zone is critical at both site-and regional-scales, influencing the hydrology and morphology of fluvial systems, supporting terrestrial riparian biota and influencing instream biota 19 .
Mine site rehabilitation and mine abandonment have emerged as major issues in Australia 20,21 with many current and legacy mines raising questions about long-term site management, particularly of waste-rock and its pollutants 22 . In Australia, provision for mine site rehabilitation is now a requirement of all active mining operations. Ranger Uranium Mine (RUM) occurs in a 79 km 2 leasehold area surrounded by the World Heritage-listed Kakadu National Park in Australia's Northern Territory (NT). Uranium mining ceased at RUM in 2012, with all decommission works to be completed by 2026 23 . Rehabilitation is underway with waste rock used as capping for the final landform. Due to the composition of the waste rock, the landform will generate significant . The potential effect of MgSO 4 on riparian plants has been identified as a key knowledge need for rehabilitation planning at RUM 30 . Although Mg and S are important macronutrients for plant development 31,32 , elevated levels can have a detrimental impact on plant growth. For example Mg above ≥8.5 mM (207 mg l −1 ) in soil solution was found to impact development of Arabidopsis thaliana plants 33 and sulfate concentrations of 400 mg l −1 had a negative impact on an aquatic moss in soft water 34 . The concentration at which plants are impacted differs between species 35 and varies with site-specific factors, such as the ratio of Ca to Mg in the soil [36][37][38] . There is significant literature describing physiological effects of Mg deficiencies on photosynthesis and plant growth, but far fewer studies on effects of elevated Mg. Sulfate is generally found to be non-toxic to plants, although at very high concentrations the increased salinity can induce plant osmotic stress [39][40][41] . There are no studies examining the impacts of MgSO 4 on native Australian tree species. This paucity of research means there is limited information to guide long-term management of the riparian vegetation of the Magela Creek catchment post RUM closure, or other areas potentially impacted by elevated MgSO 4 concentrations.
This study assessed the effect of elevated concentrations of MgSO 4 on two riparian tree species; Melaleuca viridiflora Sol. Ex Gaertn. and Alphitonia excelsa (Fenzl) Benth. These species occupy different riparian zone habitats within the Magela Creek catchment and both are common downstream from the RUM. The aim was to determine the range of MgSO 4 concentrations in soil solution where changes to plant physiology and growth could be detected, and to see if responses differed between the two species. To address this aim we undertook three glasshouse trials.

Results
There were marked differences in the response to elevated MgSO 4 concentrations between the two study species. There was no relationship between MgSO 4 concentration and plant dry mass for M. viridiflora in both trial 1 (ANOVA, F 2,15 = 0.04, P = 0.96;) and trial 2 (ANOVA, F 2,15 = 0.50; Table 1). By contrast, there was a significant decrease in plant mass with increased MgSO 4 concentration for A. excelsa (ANOVA with Tukey HSD post hoc test, F 3,16 = 9.54, P < 0.001; Table 1). At the end of the experiment, mean plant mass of A. excelsa individuals in the lowest treatment (5 mg l −1 ) was more than double those in the highest treatment (9,100 mg l −1 ) (56.0 g c.f. 22.3 g, respectively). Plant biomass values were supported by visual assessments of plants throughout the experiment. At the highest treatment concentration (9,100 mg l −1 ), A. excelsa had dropped or desiccated leaves by week 10 ( Supplementary Fig. 1h), with some leaf loss and desiccation evident in the next highest treatment (3,900 mg l −1 ). (Supplementary Fig. 1g).
Differences in mean plant mass at week 10 were reflected in chlorophyll fluorescence and pre-dawn water potentials. For A. excelsa, stomatal conductance decreased with increasing MgSO 4 concentration, declining from 144.6 m −2 s −1 in the 5 mg l −1 treatment to 42.9 m −2 s −1 in the 3,900 mg l −1 treatment (ANOVA with Tukey HSD post hoc test, F 2,11 = 16.46, P < 0.001). Only one A. excelsa individual in the 9,100 mg l −1 treatment had leaves  Fig. 1a) and there were no significant differences between treatments for stomatal conductance (Fig. 1b). There were no significant differences in chlorophyll content between MgSO 4 treatments for either species (A. excelsa F 2,11 = 0.08, P = 0.923; M. viridflora ANOVA F 2,15 = 2.98, P = 0.08; Fig. 1d). Overall, mean leaf chlorophyll content across treatments was higher in A. excelsa, with an average of 9.6 mg g −1 compared with 2.4 mg g −1 for M. viridiflora in both trial 1 and 2. For A. excelsa predawn water potential was significantly lower at a treatment concentration of 3,900 mg l −1 (ANOVA with Tukey HSD post hoc test, F 2,11 = 29.04, P < 0.001). At lower concentrations of 5 mg l −1 and 960 mg MgSO 4 l −1 , A. excelsa seedlings did not indicate water stress, however, at 3,900 mg l −1 the majority of replicate plants had predawn shoot water potentials lower than wilting point (−1.5 MPa). There was only one replicate in the 9,100 mg l −1 treatment due to leaf-loss by the majority of the plants, and again this value was below wilting point (excluded from analysis). For M. viridiflora plant water potential was lowest at the highest MgSO 4 treatment concentration of 15,300 mg l −1 (ANOVA with Tukey HSD post hoc test, F 2,15 = 19.97; P < 0.001), although values remained above −0.8 MPa, indicating that plants were not water stressed (Fig. 1c).
In each trial there was a general trend of higher foliar concentrations of Mg and S in plants receiving higher concentrations of MgSO 4 ( Fig. 2 and Table 2); however, there were differences in uptake between the two species.   Table 2).
There was a significant positive relationship between foliar Mg and Ca concentrations in M. viridiflora (except for upper leaves in trial 1; Fig. 3a), and this relationship was strongest in trial 2. There was a weak positive relationship between Ca and Mg in the upper leaves of A. excelsa, however there was no relationship for the lower leaves (Fig. 3b).

Discussion
Elevated concentrations of Mg and MgSO 4 are emerging issues in land and water management 42 , with data urgently required to support informed management of contaminated water from RUM lease which occurs within Kakadu National Park. Our trials on M. viridiflora indicated that extremely high MgSO 4 concentrations (~15,300 mg l −1 ) did not significantly affect leaf-scale physiological processes (stomatal conductance, chlorophyll fluorescence and predawn water potential), nor plant biomass of M. viridiflora. In contrast, we show that A. excelsa is a more susceptible species, with plant water status and plant biomass reduced by elevated concentrations of MgSO 4 (~960 mg l −1 ), a significant outcome given the paucity of data previously available. Management of MgSO 4 from mine waste rock and capping will need to consider species-specific responses to elevated MgSO 4 , with further research required on more species across a similar range of treatment concentrations. , it was evident that M. viridiflora was unable to fully exclude excess ions, as indicated by increasing foliar concentrations of Mg and S (Fig. 2a,b). However, this was limited to lower leaves, indicating translocation of ions to older leaves in order to maintain growth and function 49 . Thus, M. viridiflora exhibits mechanisms of root exclusion and translocation of excess ions, resulting in minimal negative response to elevated concentrations of MgSO 4 .
Root exclusion and translocation of ions, as inferred for M. viridiflora, are well described mechanisms for halophytic plants to manage salt balance 50 . There is evidence that M. viridiflora is tolerant of brackish water, with the species distribution within the Magela Creek catchment including reaches immediately upstream from mangrove stands (P. Christophersen, pers. comms.). Other common Melaleuca species, namely M. cajuputi and M. leucadendra may also have a similar tolerance to MgSO 4 given the salt tolerance of M. viridiflora 51 . Such tolerant species would be suitable for riparian rehabilitation if dieback was observed due to elevated concentrations of MgSO 4 in contaminated mine water from RUM. In contrast, A. excelsa does not extend into estuarine environments 52 and its distribution is more representative of common tree species in the area, with the majority constrained to fresh water environments. Thus, testing additional species across a treatment regime informed by potential contamination concentrations is required for a comprehensive assessment of post-rehabilitation MgSO 4 risks.
Our study showed that two common riparian trees from northern Australia have different tolerances to elevated concentrations of MgSO 4 , a mine water contaminant. It is likely that these differences are related to the relative salt tolerance of the two species, with the distribution of M. viridiflora indicating greater salt tolerance than A. excelsa. We infer that M. viridflora excludes uptake of Mg and SO 4 , and redistributes ions to older leaves. In contrast, A. excelsa demonstrated a lower tolerance to MgSO 4 , and is more likely to be impacted by increased MgSO 4 levels in the environment. The outcomes of this work provide important information that will assist with mine site rehabilitation in an area surrounded by a World Heritage-listed national park, as well contribute to our understanding of plant response to elevated MgSO 4 more broadly.

Study species.
A glasshouse-based pot trial was undertaken at the University of Western Australia to determine the effect of elevated MgSO 4 on two riparian tree species; Melaleuca viridiflora Sol. Ex Gaertn. and Alphitonia excelsa (Fenzl) Benth. Both species are widespread in the monsoonal wet-dry tropics of northern Australia, and occur in the riparian zone at Magela creek downstream of RUM in the Northern Territory (12.66°S, 132.89°E). M. viridiflora grows to 16 m and occurs in riparian habitats and seasonally inundated wetlands, and across a range of different soil types 53 . A. excelsa grows to 10 m and occurs across a broader range of habitats including riparian corridors, monsoon vine forests associated with permanent freshwater streams and savanna www.nature.com/scientificreports www.nature.com/scientificreports/ woodlands 52 . Temperatures at RUM range between 18 and 38 °C and the long-term average rainfall is 1,565 mm per year (Jabiru Airport 014198, Bureau of Meteorology, 2019). It is likely that riparian tree species are reliant on shallow groundwater (1 to 3.5 m below ground) during the dry season 54,55 . Experimental design. Three pot trials were undertaken (Table 3); trial 1 and 2 focussed on M. viridiflora and trial 3 focussed on A. excelsa. Each trial ran for 10 weeks, a period deemed long enough to detect the usually rapid response of plants to salinity and toxicity [56][57][58] . Treatments were applied daily as a liquid solution to each pot for 10 weeks. The liquid solution included a diluted Hoagland's nutrient mixture (Supplementary Table 1) and each plant received 300 ml of solution per day. There is evidence that Ca ameliorates the effect of Mg on biota 36 . Previous ecotoxicology studies of aquatic biota in Magela Creek identified that a Ca:Mg of 1:9 has an ameliorating effect on the toxicity of Mg for biota from this location 43 . In this current study we maintained Ca concentration at 1 mg l −1 , the background level at Magela Creek 43 , exceeding the 1:9 ratio for the majority of the treatments. This represents a worst case scenario where high levels of MgSO 4 are released into the low Ca environment.
M. viridiflora seedlings were sourced from a commercial nursery and A. excelsa plants were grown from seed in a glasshouse. Seedlings were removed from pots and all soil carefully washed from the roots. Seedlings of each species were transplanted into experimental pots of 9 cm diameter and 100 cm tall, filled with washed and steam-sterilised river sand, then acclimated for a minimum of two months in glasshouse conditions ( (Table 3). Following trial 1, trial 2 commenced when plants were 12 months old and assessed the effect of three substantially higher concentrations 6,000, 9,100 and 15,300 MgSO 4 (n = 6) due to the lack of detectable impact on M. viridiflora plants during trial 1. Trial 3 commenced when plants were 12 months old and tested the effect of 5, 960, 3,900 and 9,100 mg l −1 MgSO 4 (Table 3) on A. excelsa (n = 5). The electrical conductivity (EC) of the applied solutions was measured using an Aqua-CP/A with Conductivity Sensor and a Vernier LabQuest 2 with Salinity Sensor for higher treatment concentrations (e.g. trial 2). The osmotic potential of treatment solutions was calculated based on the concentration of MgSO 4 following Colmer et al. 49 .
Leaf physiology. Plant vigour was assessed by measurements of leaf chlorophyll content, total plant dry weight, root:shoot ratio, concentration of key elements in leaf tissue (all trials), and measurements of stomatal conductance (g s ), leaf chlorophyll fluorescence (Fv/Fm) and predawn plant water potential (ψ pd ) (trials 2 and 3). All measurements were made at the end of the trial in week 10.
Leaf chlorophyll content was assessed using a colorimeter (SPAD502Plus, Konica Minolta Pty, (SPAD)). We quantified the chlorophyll content in leaves across the full range of measured SPAD values (n = 18 and 20 for A. excelsa and M. viridiflora respectively) following the methods of Hendry and Grime 59 and the relationship between SPAD values and chlorophyll content (r 2 = 0.74, P < 0.001 and r 2 = 0.63 and P < 0.001 for M. viridiflora and A. excelsa respectively) was used to determine leaf chlorophyll content ( Supplementary Fig. 2). Fv/Fm was measured on dark-adapted leaves using a Pocket PEA (Hansatech Instruments) and g s was measured with a leaf porometer (SC-1 Decagon). Fv/Fm and g s were measured on four leaves from each replicate plant between 08:30 and 11:30AM local time. Predawn leaf water potential was measured using a Scholander-type pressure chamber (Model 600, PMS Instrument Company) on a small twig for M. viridiflora and one leaf for A. excelsa, sampled from the upper (younger) portion of each replicate plant. For predawn water potential, Fv/Fm, g s and chlorophyll content there were 6 replicate plants for trial 2 and 5 replicates for trial 3, except at the higher treatment levels (3,900 and 9,100 MgSO 4 mg l −1 ) because most leaves had abscised or desiccated, therefore measurements were limited to a subset of replicates (n = 4 and 1 respectively). The treatment with only one replicate (3,900 MgSO 4 mg l −1 ) was not included in the analysis.

Species
MgSO 4 (mg l -1 ) www.nature.com/scientificreports www.nature.com/scientificreports/ Nutrient content was determined for upper and lower leaves in week 10 for M. viridiflora, and week 7 for A. excelsa when it was evident that leaves were abscising from the higher treatment plants. Dried samples were ground, acid digested and the concentrations of major ions were analysed using ICP-OES. MgSO 4 in solution dissociates into Mg and SO 4 , thus foliar S concentrations are considered indicative of SO 4 concentration, with SO 4 the only applied source of S. All plants were destructively sampled at the end of the trials, and sand was carefully washed from the root material. Leaf, stem and root material was dried at 60 °C until mass stabilised and dry mass of each component was determined.
For leaf physiological variables (Fv/Fm, stomatal conductance, predawn water potential and chlorophyll content) differences between treatments within each trial was tested using one-way analysis of variance (ANOVA) with Tukey honestly significant difference (HSD) post hoc test. For foliar concentrations of Mg and S, 2-way ANOVAs were used to test for differences between MgSO 4 treatments and between upper and lower leaves. Homogeneity of variance was tested using Levene's test and normality of data distribution was determined through Shapiro-Wilk test and a visual assessment of the residuals. ANOVAs were on untransformed data, except for water potential for A. excelsa and foliar Mg content for M. viridiflora in trial 2, with analyses instead performed on log-transformed data. The relationships between foliar concentrations of Ca and Mg were determined using linear models. All analyses were completed in R 3.5.2 60 .

Data availability
Data is available through the University of Western Australia's research repository (https://research-repository. uwa.edu.au/en/datasets/).