Seasonal dynamics of iron and phosphorus in reservoir sediments in Eucalyptus plantation region

Abstract

Background

Iron (Fe) and phosphorus (P) dynamics in sediments have direct and indirect impacts on water quality. However, the mobility of P and Fe in reservoir sediments in Eucalyptus plantation region remains unclear. This study examined P and Fe pollution in sediments in a Eucalyptus plantation region using the novel planar optode, the ZrO-Chelex DGT, and the DIFS model.

Results

Direct in situ investigations showed that the levels of labile P and Fe were smaller in the Eucalyptus species-dominated sediments (X2) compared to sediments without Eucalyptus species (X1). The mean concentration of labile P and Fe decreased by 25% and 42% from X1 to X2. The decrement was insignificant (p = 0.20) in the surface sediment concentration for labile P. The significant disparity for DGT-Fe (Fe²⁺) (p = 0.03) observed in the surface sediments could be attributed to the Eucalyptus species’ elevated organic matter (tannins) concentration at X2, which reacted and consumed labile Fe. For both regions, the maximum concentration of labile P and Fe occurred in November (autumn). The reductive decomposition of Fe/Mn oxides was recognized as the main driver for their high P efflux in July and November. Low concentration of labile P and Fe was observed in December (winter) due to the adsorption of Fe/Mn oxides. The concentration of labile Fe synchronizes uniformly with that of labile P in both sediments indicating the existence of a coupling relationship (r > 0.8, p < 0.01) in both regions. The positive diffusion fluxes in both regions suggested that the sediments release labile P and Fe. The fluxes of labile P and Fe in both regions were substantially higher (p < 0.05) in the summer (anoxic period) than winter (aerobic period), indicating that hypoxia and redox conditions influenced the seasonal efflux of labile P and Fe. From the DIFS model, the replenishment ability of reactive P was higher during the anoxic period (R = 0.7, k₁ = 79.4 day^− 1, k_-1 = 0.2 day^− 1) than the aerobic period (R = 0.4, k₁ = 14.2 day^− 1, k-₁ = 0.1 day^− 1), suggesting that oxygen inhibited the efflux of P in the sediments.

Conclusion

Our results indicated that hypoxia, Eucalyptus species (organic matter (tannins)), and redox conditions influenced the seasonal mobility of sediment labile P and Fe. Our findings provided an insight into the mobility of labile P and Fe in Eucalyptus-dominated sediments and, moreover, serves as a reference for developing future studies on Eucalyptus-dominated sediments.

Introduction

Sediments are a heterogeneous system that plays a significant role in water quality deterioration due to the dynamics of nutrients between the sediments and the benthic water. Recent research has shown that that the ecological degradation of water bodies owing to the mobility of phosphorus (P) and iron (Fe) at the sediment-water boundary (SWB) is a global challenge that has harmed the aquatic ecosystem (Krueger et al. 2020). Shallow water bodies are prone to sediment resuspension (via wave and wind disturbances), resulting in the efflux of nutrients from the sediments to the benthic waters. This causes the water quality of shallow water systems to be easily affected by sediment pollution compared to that of deep-water systems. Currently, the Tianbao (dominated by Eucalyptus trees) reservoir, which is a shallow reservoir, is experiencing water quality deterioration at an alarming pace due to the presence of Fe, P, and Eucalyptus species, which has resulted in changes in color and taste (Norgbey et al. 2020b). In our previous study on the Tianbao reservoir, we demonstrated that the dominant Eucalyptus sp. plantation was the main cause of the paroxysmal black water events in early winter (Luo et al. 2020). Eucalyptus species is popular in many countries because of its high economic benefit and fast growth rate. Studies by Yang et al. (2019) further revealed that black water formation in this region occurs because of the complexation reaction between Fe in sediments and tannins from Eucalyptus sp. trees. Other studies have attributed the black water formation in this region to the production of iron sulfide (FeS) at the SWB during hypoxia (Norgbey et al. 2020b). Due to these concerns, stakeholders and researchers are developing new techniques to improve the water quality in this region to protect the aquatic ecosystem. One way of achieving this objective is by investigating the seasonal interactions and patterns of P and Fe in the Eucalyptus-dominated reservoir.

Fe is a well-known element that plays a significant role in the transformation and mobility of soluble P at the SWB due to its coupling relationship with P at a micro-level (Zhang et al. 2020). Zhang et al. (2020) revealed that Fe bound P is primarily responsible for the efflux of P in sediments and the correlation between P and Fe is significant. In addition, the efflux of P has a direct relationship with the Fe/P ratio where the efflux of P decreases tremendously when the ratio is greater than 30 (Jensen et al. 1992). Other factors that primarily influence the transformation of P and Fe in freshwater systems are dissolved oxygen and oxidation-reduction potential. When oxygen is high (> 2 mg/L), Fe²⁺ is oxidized to Fe³⁺ and are present in their insoluble forms. On the other hand, when oxygen is low (< 2 mg/L), microbial reduction of Fe occurs, producing soluble forms of P and Fe into the benthic water. Although this phenomenon is common in thermally stratified reservoirs where oxygen content is low in summer and relatively high in winter (Azadi et al. 2020), studies about the efflux of P and Fe in SWB in a Eucalyptus species-dominated reservoir is limited. Therefore, investigating the synchronous efflux of P and Fe at the SWB at a micro-scale will help throw more light on the seasonal behavior of P and Fe in a Eucalyptus species-dominated reservoir.

Changes in the physical, chemical, and biological state of sediments affect the dissolved oxygen (DO) content, pH, and oxidation redox potential of sediments. Maintaining this natural state of the sediment samples during sampling will help understand the cycling of P and Fe at the SWB correctly. Currently, the use of conventional techniques in the measurement of P and Fe in sediments is accompanied by many disadvantages such as the disturbance of the physical, chemical, and biological state of sediment samples. During the transportation of the samples from the field to the laboratory, the natural state of the samples is distorted. Furthermore, the process involved in measuring samples using conventional methods is slow and cumbersome. In this study, a diffusive gradient in thin films (DGT) (based on Fick’s law) was used to measure the kinetics and changes in diffusion fluxes of P and Fe at the SWB. Compared to conventional sampling techniques, the DGT technique is easy to operate and has a wide range of applicability. The DGT device can easily analyze the kinetics in P and Fe at SWB without giving ambiguous results (Luo et al. 2020). Unlike conventional techniques, the DGT technique does not contaminate or disturb the sediments during testing (Wang et al. 2016). Also, the planar optode (PO) machine examined changes in DO concentration across the SWB in 2-dimensions while the DGT-induce flux in soil (DIFS) program provided information on the replenishment ability of sediments. These high resolutions and frequency measurements provide good insight into the mechanism of P and Fe mobility at the SWB in the Eucalyptus species-dominated reservoir.

Currently, there is extensive literature on P and Fe mobility at the SWB for many shallow lakes. However, research on seasonal interactions of P and Fe in reservoir sediments in a Eucalyptus plantation region is limited. Additionally, eutrophication of reservoirs in southern China is an emerging problem (Chen et al. 2019d; Norgbey et al. 2020b), and studies on phosphorus dynamics in these reservoirs are rare. From this perspective, more dedication and studies are needed in this field to tackle the issue.

In this research, the novel PO device and the ZrO-Chelex DGT were incorporated to compare the simultaneous efflux of labile P and Fe in Eucalyptus-dominated sediments to sediments without Eucalyptus species. This millimeter-scale technique (ZrO-Chelex DGT) has been successfully used in past studies and has proved to be a reliable way to simultaneously measure P and Fe in situ without destroying the samples (Rougerie et al. 2021). From this context, we selected the Tianbao reservoir (a shallow reservoir, dominated by Eucalyptus trees) as our area of study. The DGT was incorporated to synchronously calculate the interaction and pattern of P and Fe at the SWB during the different seasons in a year. Also, the DIFS model was used to investigate the replenishment ability of the sediment during hypoxia and aerobic periods in a year. The information garnered from this research can provide management knowledge (designs) on how to find long-lasting solutions to the water blackening issue in this region due to the presence of P, Fe, and Eucalyptus species.

Materials and methods

Study area

Tianbao reservoir-China, located within the latitude 22° 52′ 15.84′′–22° 53′ 16.51′′ N and longitude 108° 13′ 31.33′–108° 14′ 9.37′′ E, experiences black water event in early winter due to the complexation reaction between iron and tannins (Luo et al. 2020). The shallow reservoir has a volume of about 13.5 million m³ and the watershed area of 51 km². Tianbao reservoir has a water surface area of 733,000 m² and a maximum depth of 14 m (Norgbey et al. 2020b). Tianbao reservoir is thermally stratified for a greater part of the year with no stratification occurring for a short time in winter (December to February). The reservoir is located in a subtropical monsoon climatic zone. The reservoir has an annual mean temperature of 21.9 °C and annual mean precipitation of 1304 mm. The annual mean relative humidity in this region is about 78% (Luo et al. 2020). Some parts of the reservoir are surrounded mainly by Eucalyptus sp. plantation. There are four main seasons in this study area: spring (March–May), summer (June–August), autumn (September–November), and winter (December–February). Two sampling points (X1, depth = 14 m and X2, depth = 10 m) were chosen in the reservoir. Sediments located at X2 are dominated by the Eucalyptus species, whiles X1 sediments are not close to the Eucalyptus species (Fig. 1). The sampling stations X1 and X2 with depths of 14 m and 10 m, respectively, were not close to activities that occur in the drainage area and the water inlet. Thus, rainfall and runoff effects were insignificant on the water quality parameters at X1 and X2.

Fig. 1

Location of the study area (Tianbao reservoir)

Buoy description

A buoy was stationed at the deepest point in the reservoir with a depth of about 14 m (Fig. 1). The yellow spring instrument (YSI) multi-parameter water quality monitor (YSI EX02) was strapped to the buoy to measure the water column temperature (T), dissolved oxygen (DO), conductivity (EC), and oxidation-reduction potential (ORP). These parameters were taken on-site at an interval of 0.5 m from the surface (0.2 m) to the maximum (14 m) depth (at every 3 h interval during the day). The YSI multi-parameter and the buoy were both purchased from YSI Corporation, Yellow Springs, OH, USA. In this study, the data recorded at 12 am was used for the analysis since thermal stratification is stable and intense at this time (Liu et al. 2019). Temperature and dissolved oxygen data presented in the research work covers a period of 18 months. The trend recorded in 2018 was consistent with the trend in 2019. Thus, the analysis done in the study with regard to reservoir oxygen content and temperature can be depended upon to provide an unambiguous reflection of the study region.

Data collection routine

The data collection routine was done in three steps. Firstly, the water samples were collected to measure their physico-chemical properties. Secondly, the collected sediments were analyzed with DGT, PO device, and the trace metals within the sediments were measured. Thirdly, the DIFS software evaluated the resupply kinetics of P on a seasonal scale. These processes have been described in detail in the following sections.

Seasonal water quality analysis

Water samples were collected in April 2019, July 2019, November 2018, and January 2019 representing the spring season, summer season, autumn season, and winter season, respectively. Results of the received samples (water) in 2019 were consistent with water sampling done in 2018. Please see the supplementary sheet, demonstrating that the 2-year data is accurate and can be depended upon to provide an unambiguous reflection of the study region. The main reason we choose these months (April (2019), July (2019), November (2018), and January (2019)) because these periods truly reflect the peak periods of the major seasons in the study area. Specifically, April (2019), July (2019), November (2018), and January (2019) represent the peak period of spring, summer, autumn, and winter, respectively. This assertion is based on the weather temperature observed in the study area from 1960 to 2018. The weather station data has been presented in the supplementary material.

The samples (water) taken were tested for total phosphorus (TP), total nitrogen (TN), total iron (TFe), manganese (Mn), dissolved organic carbon (DOC), sulfide (S), and ammonia nitrogen (NH₃-N) at the sampling point X1 and X2 (Fig. 1). Water samples were collected using the plexiglass hydrophore water sample collector at different depths of the reservoir (0.5, 3, 6, 9, 12, 13, and 14 m). The device was obtained from the Guangxi Institute of Water Resources Research located in Nanning, China. The collected water samples were stored in a polyethylene water storage container with a volume of 500 ml. For temporary storage of the water samples, 1.0 ml of nitric acid was added to the water sample for the testing of TFe and Mn. For S, 1 ml of sodium hydroxide was added to the water sample while 1 ml of sulfuric acid was added to the water sample for the testing TP. No reagent was added to water samples for the testing of DOC. Before being sent to the research facility for further study, the water samples were held at 4 °C. Assessment of TFe and Mn was performed using the Perkin Elmer Optima 8300 ICP-OES (Inductively Coupled Plasma-Optical Emission Spectrometry), Winston-Salem, USA (Krueger et al. 2020). TP was tested in the laboratory using the ammonium molybdate spectrophotometry technique using the UV-1800 UV-Vis Spectrophotometer, Shimadzu Scientific Instrument, Columbia, USA (Shyla & Mahadevaiah 2011) while DOC was analyzed using the non-dispersive (combustion oxidation) infrared method (Bisutti et al. 2004). Sulfide in the water samples was tested following studies by Norgbey et al. (2020a, 2020b) using the methylene blue spectrophotometry. Tannins concentration in the water column was tested with the ultraviolet spectrophotometry, Shimadzu UV-1900i, Japan (Huang et al. 2018). TN was tested using the combined persulfate digestion, and NH₃-N was tested using the indophenol blue method following works done by Nwankwegu et al. (2020).

Seasonal sediment sampling

The undisturbed sediment core samples were taken from the Tianbao (Eucalyptus dominated) reservoir using the UWITEC sediment core sampler (obtained from EasySensor Institute-Nanjing, China). The samples were collected in autumn (November 2018), winter (January 2019), spring (April 2019), and summer (July 2019). The results of the sediment samples collected in 2019 in this study were consistent with water sampling done in 2018 (See supplementary sheet), demonstrating that the 2-year data is accurate and can be depended upon to provide an unambiguous reflection of the study region. The sediment cores, with a diameter of 11 cm and length of 30 cm, were taken at X1 and X2 (Fig. 1). X1 had a depth of 14 m while X2 had a depth of 10 m. The presence of trace metals (Mn, Ni, Cu, Zn, As, Cd, Pb), total iron (TFe), total phosphorus (TP), and organic matter were measured in the surface sediments (0–25 cm). The trace metals in the surface sediments in the Tianbao reservoir were analyzed and tested using the inductively coupled plasma mass spectrometry (ICP-MS) (ThermoFisher Scientific, USA) following studies by Zhang et al. (2017). The sediment specimen was dried and passed through a plastic sieve with a size of 0.16 mm. A high-efficiency microwave digester (MLS-1200 MEG) was used to microwave the sample at a temperature of 150 °C and a digestion time of 30 min. The microwaved sieved sample was mixed with concentrated nitric acid, hydrogen peroxide, and hydrofluoric acid (ratio of 3:1:1). The samples were then tested in three replicates. The TFe in sediment samples was determined using the atomic absorption spectrophotometry (ICE 3000 Thermo Fisher Scientific – USA) while TP in the sediments was tested with the X-ray fluorescence-XRF-Rigaku Supermini200 (Fang et al. 2018). The organic matter was analyzed with the potassium dichromate-external heating method (Zhang et al. 2020).

Diffusive gradient in thin films (DGT)

DGT and PO sample analysis

The sediments at sampling sites (X1 and X2) were collected during the four seasons with the UWITEC sediment core sampler. The distribution of DO in the sediment cores was measured using the PO machine (PO2100) purchased from EasySensor Institute, Nanjing, China. Detailed information about the preparation, calibration, and usage of the PO has been provided in past research works (Yuan et al. 2020). The ZrO-Chelex DGT devices were used to simultaneously determine the variation in concentrations and the diffusive fluxes of labile P and Fe at the SWB. Before deploying the DGT device, the device (DGT) was kept in deionized water for a time of about 16 h. The deployment time for the DGT device was 24 h. After removing the DGT device from the sediment samples. The device was rinsed with deionized water, kept moist in a plastic bag and stored at a low temperature. During the deployment of the DGT devices, the PO was used to monitor the DO distribution at the SWB. To obtain information on the labile P and Fe, the ZrO-Chelex DGT was cut using the ceramic knife at an interval of 2 mm. 400 μl 1 mol/L HNO₃ was added to the binding membrane in a centrifuge tube. After 16 h of incubation, the HNO₃ was collected and tested for labile Fe. Similarly, 400 μl 1 mol/L NaOH was added to the sample and incubated for about 16 h and later tested for labile P following works by Xu et al. (2013). Detailed information about the DGT device and the simultaneous measurement of labile P and Fe have been extensively discussed and can be found in previous research works (Pedersen et al. 2015; Menezes-Blackburn et al. 2019).

DGT data analysis

The variation in concentration and the fluxes P and Fe with depth at the SWB were obtained according to Eqs. 1, 2, and 3 (Xu et al. 2013).

$$ {C}_{\mathrm{DGT}}=\frac{M\triangle g}{DAt} $$

(1)

where M - amount (mass) in the device’s sticking gel (ng), g - diffusive layer thickness (0.09 cm), D - coefficient of metal diffusion (cm²/s), t - deployment time of the device (s), and A - exposure window area of the DGT device (cm²). C_DGT is the concentration (effective) (mg/L)

$$ F= Fw+ Fs=-{D}_W{\left(\frac{\partial {C}_{DGT}}{\partial {X}_W}\right)}_{\left(x=0\right)}\kern2.25em -\upvarphi {D}_S{\left(\frac{\partial {C}_{DGT}}{\partial {X}_S}\right)}_{\left(x=0\right)} $$

(2)

where F (g cm^-2 s^-1) refers to the summation of benthic water and sediment fluxes, Fw (g cm^-2 s^-1) refers to the labile fluxes of Fe/P from benthic water to SWB, Fs (g cm^-2 s^-1) refers to the labile fluxes of Fe/P from sediment to SWB$, {\left(\frac{\partial {C}_{DGT}}{\partial {X}_W}\right)}_{\left(x=0\right)}$ refers to slope of concentration in the benthic water is, ${\left(\frac{\partial {C}_{DGT}}{\partial {X}_S}\right)}_{\left(x=0\right)}$ refers to slope of concentration in the sediment, Ds (cm²/s) refers to the coefficients of diffusion of reactive P in the sediment, Dw (cm²/s) refers to the coefficients of diffusion of reactive P in the benthic water, and Φ refers to the surface sediment’s porosity. The porosity was measured following Eq. 3.

$$ \Phi =\frac{W{d}_s}{\left(1-W\right){d}_w+{Wd}_s} $$

(3)

where W refers to mass water content, d_s refers to sediment’s mean density, and d_w refers to benthic water’s mean density.

DIFS model

The DIFS model, developed and first used by Harper et al. (2000), can be used to measure the remobilization of P in sediments at a micro-level. The DIFS software program parameters were obtained from Eqs. 4, 5, 6, 7, 8, 9, 10, 11.

$$ Pc=\frac{W_1}{\left(\frac{W_0-{W}_1}{\uprho}\right)} $$

(4)

$$ {\varphi}_s=\frac{d_p}{\left({d}_p+{P}_c\right)} $$

(5)

$$ Ds=\frac{Do}{1-\ln {\left({\varphi}_s\right)}^2} $$

(6)

$$ R=\frac{C_{DGT}\ }{C_{Peeper}\ } $$

(7)

$$ Kd=\frac{C_s}{C_{Peeper}} = \frac{H_1}{C_{Peeper}} = \frac{1}{P_c\ }.\frac{k_1}{k_{-1}\ } $$

(8)

$$ {T}_c=\frac{1}{k_1+{k}_{-1}} $$

(9)

$$ {k}_1=\frac{1}{T_c} $$

(10)

$$ {k}_{-1}=\frac{k_1}{k_d.\kern0.5em {P}_C} $$

(11)

where C_DGT refers to DGT estimated concentration (mg/L), C_s refers to solid phase concentration (mol/g), C_peeper refers to dissolved concentration (mg/L), R refers to DGT concentration divided by dissolved concentration, K_d refers to distribution equilibrium coefficient (cm³/g), P_c refers to density of particle (g/cm³), φ_d refers to porosity of diffusion layer, φ_s refers to porosity of sediment, D_o refers to coefficients of diffusion layer (cm²/s), D_s refers to sediment’s diffusion coefficients (cm²/s), T refers to time for deployment (h), Δg refers to thickness of diffusion layer (cm), T_c refers to time of sorption or desorption (s), k₁ refers to sorption rate constant (s^− 1), k₋₁ refers to desorption’s constant (s^− 1), C_o refers to concentration in the gap water (mol/cm³), H₁ refers to loosely sorbed P (mg/g), and K_d and T_c investigate the adsorption/desorption kinetics in the SWB (Harper et al. 1998, 2000).

Statistical analysis

Statistical analysis and data processing were performed. One-way ANOVA from the origin software was used to show the significant difference between different groups with a p value less than 0.05 indicating a significant difference (Huang et al. 2017). Pearson’s correlation was processed using Origin software. Diagrams were drawn with Origin 9 software. Data analyses were analyzed (one-way ANOVA) using Origin 9. The sample relationship (R²) was statistically evaluated by the linear regression analysis of SPSS Statistics 17.0 (SPSS Inc., USA). All chemical analysis determinations for sediment samples were performed in a triplicate form (Huang et al. 2019; Jiang et al. 2020; Norgbey et al. 2020a).

Results and discussion

Changes observed in the physical and chemical state of reservoir water and sediments

Variations in the physical property of reservoir water during the seasons

The seasonal changes in water column temperature (T) and dissolved oxygen (DO) content with depth at the two sampling regions (Eucalyptus dominated sediments and sediments without Eucalyptus species) have been presented in Fig. 2 and S1. The high sensor YSI machine attached to the buoy station measured the T and DO profiles at sites X1 and X2. The pH and ORP of the reservoir water over the sampling period are presented in Table S1, S2, and Fig. S1. The pH of the benthic water at X1 and X2 ranged from 6.7 to 7.2 and 6.8 to 7.3, respectively, during the seasons while the ORP ranged from 111 to 188 mV and 92 to 201 mV, respectively. The T, DO, pH, and ORP values were similar at the two sampling sites (X1 and X2) with no significant difference (p > 0.05).

Fig. 2

Variation in depth of a water temperature and b dissolved oxygen profiles of Tianbao reservoir with time from March 2018 to December 15, 2019. Data was collected at 12:00 each day at sampling site X. Dashed vertical lines show dates of sampling (spring-April 2019, summer-July 2019, autumn-November 2018 and winter-January 2019)

Generally, the reservoir was thermally stratified from March to November where the water temperature decreased substantially from the surface to the bottom of the reservoir (Fig. 2 and S1). During the thermally stratified period, the mean surface (0–2 m) and bottom (11–14 m) temperatures were 28.1 °C and 15.8 °C, respectively, suggesting that the difference between the surface and bottom water temperature was significant (p < 0.05). The difference in surface temperature and the bottom temperature reached its lowest value in winter (December to February) where there was complete mixing of the reservoir water. In winter (no thermal stratification), the surface water temperature was 16.9 °C whiles the bottom water temperature was 15.6 °C. This trend observed in the study (2019 data) with regard to water column temperature (T) and dissolved oxygen (DO) were consistent with results measured in 2018. Please see supplementary material for the seasonal changes of water column temperature (T) and dissolved oxygen (DO) in 2018. From this context, the Tianbao reservoir can be classified as a subtropical monomictic since the reservoir experiences a single overturn period in a year (Liu et al. 2019).

The seasonal changes in the DO profiles for the reservoir (X1 and X2) are presented in Fig. 2 and S1. The entire water column in the reservoir during the mixing period (December to February) had a high DO concentration, which exceeded 6 mg/L. From March to November 2018/2019, there was a significant decline in the DO concentration of the bottom layer (11–14 m) with a value of less than 2 mg/L indicating an anoxic environment (Chen et al. 2019a). Although the bottom layer lacked oxygen during this period, the surface water layer had a high DO concentration with an average value of 8.4 mg/L. The DO concentration in the bottom water reached its lowest value (mean: 0.8 mg/L) in summer (June–August).

Variations of the chemical property of reservoir water during the seasons

From Table S1 and S2, the seasonal variation in concentrations at X1 and X2 for TP, NH₃-N, TN, S, and DOC in the water column were similar (p > 0.05), suggesting that the presence of Eucalyptus species did not influence these parameters. For both sampling regions, the concentration of TP, NH₃-N, and TN in the bottom water decrease from the anoxic period (spring, summer, and autumn) to the aerobic period (winter). The concentration of S was low (< 0.05 mg/L) in the bottom water with the difference being insignificant (p > 0.05) throughout the four seasons for X1 and X2. The concentration of TP (X1) in the bottom for spring, summer autumn, and winter were 0.02 mg/L, 0.03 mg/L, 0.05 mg/L, and 0.01 mgL, respectively (Table S1 and S2). The TP concentration at X1 reduced by 80% from autumn to winter. The trend was the same for X2. Similarly, the concentration of NH₃-N and TN also reduced by 55% and 52% from the anoxic season to aerobic season at X2. This indicates that the anoxic condition in the bottom water accelerated the release rate of TP, NH₃-N, and TN from the reservoir sediments to the benthic water; thus, their high concentration during the anoxic period (Nürnberg 2005). This trend observed in the study (2019 data) with regard to TP, NH₃-N, and TN were consistent with results measured in 2018. Please see supplementary material for the seasonal changes of TP, NH₃-N, and TN in 2018.

Variations in the physico-chemical property of sediments during the seasons

The seasonal physical (ORP and pH) and chemical properties (Mn, Ni, Cu, Zn, As, Cd, Pb, TP) in the sediments (both sampling regions) have been presented in Fig. S1, Table S3 and S4. The difference in concentration (Mn, Ni, Cu, Zn, As, Cd, Pb, TP) between the two sampling regions were insignificant (p > 0.05) suggesting that the presence of Eucalyptus species had no primary influence on the concentration of the trace metals (Table S3). For both sampling regions, the sediment ORP showed significant (p < 0.05) variations across the seasons. The concentration of Mn, Ni, Cu, Zn, As, Cd, and Pb in the sediments (X1) were 769.59 mg/kg, 25.24 mg/kg, 39.26 mg/kg, 85.77 mg/kg, 13.35 mg/kg, 0.41 mg/kg, and 51.77 mg/kg, respectively, in winter. In summer, their concentration decreased by 511.88 mg/kg, 13.4 mg/kg, 22.21 mg/kg, 20.2 mg/kg, 9.82 mg/kg, 0.31 mg/kg, and 31.69 mg/kg, respectively. A similar trend was observed at sampling point X2 (Fig. S1, Table S3 and S4), with the concentration of the trace metals decreasing from winter to summer. Previous research works by (Chen et al. 2019a, 2019c) suggest that the adsorption by Fe/Mn oxides inhibits the efflux of trace metals in the sediments, thus supporting the high concentration recorded in our study in winter. From Table S3, there was a significant (p < 0.05) reduction in sediment’s TP concentration in summer (hypoxia) compared to winter (no hypoxia) at X1 and X2. Coincidentally, the TP concentration in the bottom water was high in summer compared to that of winter (Table S1 and S2). The results suggest that significant changes in the redox conditions (Table S4) at the SWB during the different seasons influenced the concentration of trace metals and TP (Chen et al. 2019a, 2019c). This trend observed in the study (2019 data) with regard to sediment trace metals (Mn, Ni, Cu, Zn, As, Cd, Pb, TFe, TP) were consistent with results measured in 2018. Please see supplementary material for the seasonal changes of trace metals in 2018.

Studies by Jensen et al. (1992) proposed the Fe/P ratio as an indicator to describe if the sediment absorbs or releases P. This ratio helps in controlling the internal loading of P in a water system leading to eutrophication. When the Fe/P ratio is less than 15, the sediments are unable to keep soluble P (PO₄^3-) and release the P to the benthic water. On the other hand, when the Fe/P ratio is greater than 15, the adsorption of P occurs due to the formation of insoluble Fe/Mn oxide (in the presence of oxygen). In this study, the TFe/TP ratio at X1 during summer and winter was 13.5 and 29.3, respectively. At X2, the TFe/TP ratio was 12.3 in summer and 31.5 in winter. The low Fe/P ratio (X1 = 13.5, X2 = 12.3) for summer implied the discharge of P from the sediments to the benthic water while the high Fe/P ratio (X1 = 29.3, X2 = 31.5) in winter indicated the adsorption of soluble P by Fe/Mn oxide. The findings suggest that hypoxia caused the efflux of P from sediments in summer due to the reductive decomposition (dissolution) of Fe/Mn oxide in both the Eucalyptus species-dominated sediments (X2) and sediments without Eucalyptus species (X1) (Mortimer 1941).

Seasonal variation and coupling relationship between phosphorus and iron in the SWB

The seasonal changes of labile P and Fe²⁺ at X1 and X2 were compared (Figs. 3 and 4). For both sampling sites, the content of labile P and Fe showed variations suggesting that P mobility within the sediments is sensitive to redox conditions. The mean content of labile P and Fe in the surface sediment (0–30 mm) for spring (April 2019), summer (July 2019), autumn (November 2018), and winter (January 2019) in the reservoir was 9.2 mg/L, 6.7 mg/L, 11 mg/L, and 8.2 mg/L and 0.04 mg/L, 0.05 mg/L, 0.05 mg./L, and 0.02 mg/L, respectively (Table S5). For the benthic water, the mean concentration (X1 and X2) of labile Fe in spring (April 2019), summer (July 2019), autumn (November 2018), and winter (January 2019) were 0.85 mg/L, 0.93 mg/L, 1.02 mg/L, and 0.18 mg/L, respectively. While for P during this same period were 0.005 mg/L, 0.004 mg/L, 0.004 mg/L, and 0.005 mg/L, respectively (Table S5). The planar optode measured the seasonal variation in DO profiles at the SWB for X1 and X2 pictured in Figs. 3 and 4. The DO level was typically low for spring, summer, and autumn in the samples due to the intensity of thermal stratification in the reservoir. For winter, thermal stratification ceased due to the mixing of the reservoir and the DO content was significantly higher in winter compared to the other seasons (Figs. 3 and 4).

Fig. 3

Seasonal variations at sampling site X1. a The DGT-labile P and labile Fe concentrations with depth at the sediment-water interface (SWI) and b oxygen concentration at the SWI (2D imaging) of Tianbao Reservoir. The position of the SWI is represented by zero

Fig. 4

Seasonal variations at sampling site X1. a The DGT-labile P and labile Fe concentrations with depth at the sediment-water interface (SWI) and b oxygen concentration at the SWI (2D imaging) of Tianbao Reservoir. The position of the SWI is represented by zero

Many studies have proposed the mechanism responsible for the discharge of reactive P and Fe²⁺ from sediments to bottom water (Krueger et al. 2020). These include the reductive decomposition (dissolution) of Fe/Mn oxides by anoxia, organic matter (OM) decomposition, and changes in water environment behavior such as pH and redox potential. Other research works have shown that efflux P is sometimes dependent on Mn content (Chen et al. 2019b). For the two sampling points in this study, the sediment concentration of Mn was substantially smaller than that of Fe (Table 2). Thus, the influence of Mn on the efflux of P in the different ecotype sediments was less compared to Fe (Wang et al. 2018). In this study, the seasonal variations suggest that the hypoxia caused the Fe³⁺ ions in the sediments to be reduced to soluble Fe²⁺ ions causing the efflux of labile P within the sediment (Zhang et al. 2020), thus high labile P content from the time of hypoxia to the time of aeration. Specifically, the mean content of labile P (0–30 cm) at X1 and X2 were 0.05 and 0.04 mg/L in summer and 0.03 and 0.01 in winter, respectively. The content of labile P and Fe (for both sampling sites) was significantly smaller in the benthic water than in the sediments. The lack of oxygen within the sediments intensified the decomposition of Fe oxide and the efflux of P at the sampling sites. Thus, displaying an elevated amount of P in sediments compared to benthic water.

For a better insight on the pattern and synchronization of labile P and labile Fe, a linear correlation analysis was done between these two parameters (Fig. S2). The relationship (correlation coefficient-r² value) between DGT P and Fe were significantly positive during the sampling seasons at X1 and X2 (Fig. S2). The r² value between DGT P and Fe at the sampling sites (X1 and X2) in spring (April 2019), summer (July 2019), autumn (November 2018), and winter (January 2019) were 0.8, 0.7, 0.8, 0.7, and 0.7, 0.9, 0.8, and 0.8, respectively. The highly significant correlation (positive) recorded (p < 0.01) suggested that the efflux of labile P and Fe at the SWB was synchronous and a coupling relationship existed. The gradient of the linear (correlation) equation (Fig. 4) describes the mechanism for the mobility of Fe²⁺ and reactive P at the SWB. The gradient of labile P and Fe was lower in the anoxic period compared to the aerobic period. The slopes (X1) in summer and autumn were 251.97 and 208.46, respectively, compared to 264.95 in winter. The high gradient recorded in winter revealed that sediments have a high tendency to fix P due to the presence of oxygen. Similarly, the low gradient recorded during the anoxic period shows that the sediments tend to release reactive P from the sediments (Zhang et al. 2020). A research work by Norgbey et al. (2020b) revealed that under hypoxia (reductive conditions), the complexes of Fe³⁺ are reduced to Fe²⁺. During this process, Fe bound P decomposes and releases reactive P from the sediments to bottom water. Similarly, during the aerobic condition, Fe²⁺ is oxidized to Fe³⁺ causing the adsorption of P by Fe/Mn oxide in the sediments. The trend was similar at X2. The study showed that hypoxia enhances the efflux of labile P and Fe at both X1 and X2. The elevated concentration of labile P and Fe in summer compared to winter in both sampling sites is thus explained. Similar studies (Wang et al. 2016) were conducted in the Hongfeng reservoir (thermally stratified reservoir) in southern China where the labile P and Fe concentrations were high in summer compared to winter indicating that hypoxia influenced the process. Also, Rong et al. (2020) used the DGT (ZrO-Chelex) to measure the dynamics of P and Fe at the SWB. These authors revealed the efflux of Fe²⁺ and P was synchronous and significantly correlated, supporting the results from our study.

Role of Eucalyptus species in the seasonal dynamics of labile phosphorus and iron in the SWB

The seasonal (spring, summer, autumn, and winter) disparity of labile P and Fe assessed by DGT at X1 and X2 are shown in Figs. 3 and 4. The surface sediment (0–30 mm) content of labile P and Fe was lower in the Eucalyptus species-dominated sediments (X2) compared to sediments without Eucalyptus species (X1) (Table S5). The mean concentration of labile P at X1 and X2 were 0.04 ± 0.01 mg/L and 0.03 ± 0.01 mg/L, respectively. We observed an insignificant disparity (P: p = 0.20) in the concentration of labile P at X2 compared to X1. Similarly, the mean concentration of labile Fe at X1 and X2 were 11.1 ± 1.65 mg/L and 6.4 ± 2.8 mg/L, respectively. The significant disparity for (Fe: p = 0.03) DGT-Fe (Fe²⁺) observed in the surface sediments may be attributed to the elevated content of Eucalyptus species at X2 (Table S4). Many research works have observed that Eucalyptus species have had a significant impact on soil properties. For example, the trees are known as “green deserts” because they absorb a significant amount of water in the soil (Cook et al. 2017; de Barros Ferraz et al. 2019). Yang et al. (2019) also connoted that Eucalyptus species affect water quality by chemically interacting with Fe to produce black products, implying that Eucalyptus species clearly influence the natural state of soil and water. In this study, the concentration difference of DGT-Fe at X1 and X2 (surface sediments) was statistically significant (Table S5) (Fe: p value = 0.03). Research findings (Hladyz et al. 2011; Shen et al. 2014; Li et al. 2020) have connoted that the mineralization (degradation) of organic matter process in the SWB depletes oxygen causing hypoxia (or even anoxia). Our study recorded an abundant amount of organic matter in the surface sediments of X2 compared to X1 (Table S4). In fact, the content of organic matter increased by 56% from X1 to X2. This could be attributed to the high litter-fall, root biomass in the Eucalyptus species-dominated sediments (X2) compared to X1. The mineralization (degradation) of organic matter process during hypoxia at the SWB facilitated the productions of Fe²⁺/FeS from the reduction dissolution of Fe(III) oxide. This process occurs via microbial and chemical iron reduction presented in Eqs. 12 and 13, respectively (Li et al. 2020).

$$ {\mathrm{CH}}_2\mathrm{O}+8{\mathrm{H}}^{+}+4\mathrm{FeOOH}={\mathrm{CO}}_2+4{\mathrm{Fe}}^{2^{+}}+7{\mathrm{H}}_2\mathrm{O} $$

(12)

$$ 3{\mathrm{H}}_2\mathrm{S}+2\mathrm{FeOOH}=2\mathrm{FeS}+{\mathrm{S}}^0+4{\mathrm{H}}_2\mathrm{O} $$

(13)

Consequently, FeS and Fe²⁺ are produced (Eqs. 12 and 13) more at X2 compared to X1 due to high organic material at X2 compared to X1. The latter interacts chemically with sulfide in the SWB (Table S3). Consequently, Fe²⁺ will dwindle at X2 compared to X1. This supports the significant reduction of DGT-Fe at X2 compared to X1.

Another reason for the dwindling of DGT-Fe at X2 compared to X1 could be the high Eucalyptus trees’ tannins reacting with Fe^2+/3+, causing the DGT-Fe content to dwindle significantly at X2. Yang et al. (2017, 2019) deduced that the tannins from Eucalyptus trees are substantially greater than that of different plants. Water bodies that contain Eucalyptus leaves change color due to the existence of tannins (Yang et al. 2017, 2019). In this study, Eucalyptus trees, which are mainly concentrated at X2 compared to X1, influenced the nutrients (Fe) at the SWB. From Table S1 and S2, the mean concentration of tannins in the water column at X2 (mean 0.4 ± 0.40 mg/L) was higher than that of X1 (mean 0.2 ± 0.21 mg/L). In fact, the content of tannins increased by 59.5% from X1 to X2. The high Eucalyptus trees’ tannins reacted with Fe^2+/3+, thus causing the DGT-Fe content to be used up. Yang et al. (2019) and Luo et al. (2020) have supported this assertion about Eucalyptus trees’ tannins reacting with Fe²⁺/Fe³⁺. These authors further proved that Eucalyptus trees’ tannins and Fe^2+/3+ react to produce a black substance, thus explaining the significantly low DGT-Fe concentration at surface sediment at X2 compared to X1. In short, the findings from the study showed that the Eucalyptus species had no significant influence on P mobility at the SWB. However, the influence of Eucalyptus species on the mobility of DGT-Fe was significant.

Seasonal variation in the diffusion fluxes of labile P and Fe across the SWB

The seasonal diffusive fluxes of labile P and Fe were calculated based on Fick’s law (Fig. 5). The fluxes for soluble Fe²⁺ for spring, summer, autumn, and winter at X1 (X2) were 7.9 (10.1) mg m⁻² d⁻¹, 12.9 (6) mg m⁻² d⁻¹, 14.6 (8.9) mg m⁻² d⁻¹, and 5.5 (4.3) mg m⁻² d⁻¹, respectively, with no significant difference (p > 0.05) detected between the sediments in X1 and X2. However, a seasonal reduction (Fe²⁺) of about 54% (28%) was observed at X1 (X2) from summer (hypoxia period) to winter (no hypoxia). Similarly, the flux for reactive P in X1 (X2) for spring, summer, autumn, and winter was 0.05 (0.04) mg m⁻² d⁻¹, 0.08 (0.05) mg m⁻² d⁻¹, 0.08 (0.12) mg m⁻² d⁻¹, and 0.04 (0.002) mg m⁻² d⁻¹, respectively (p > 0.05). The flux of reactive P reduced by 46% (98%) at X1 (X2) from the anoxic period to the aerobic period. The flux of reactive P at X1 (X2) was two (twenty-five) folds higher in summer than winter, suggesting that hypoxia was the primary force responsible for the efflux of reactive P from the sediments (Wang et al. 2016). During the four seasons, the diffusive fluxes of Fe²⁺ and reactive P varied and were positive (X1 and X2) indicating that both sediments acted as the source of Fe²⁺ and reactive P to the benthic water (Zhang et al. 2020). High fluxes were observed in the anoxic period whiles low fluxes were recorded during the period of no hypoxia. The high fluxes in summer in both ecotype sediments (X1 and X2) revealed that Fe³⁺ reduction promoted the release of Fe²⁺ and reactive P from the sediments (Rong et al. 2020). The mean flux of Fe²⁺ (8.8 mg m⁻² d^− 1) recorded in our study was substantially higher than that of the Hongfeng reservoir (average: 0.89 mg m⁻² d^− 1) and Lake Taihu (average: 0.85 mg m⁻² d^− 1) in China. Similarly, the average flux of soluble P (0.07 mg m⁻² d^− 1) was lower than that of the Hongfeng reservoir (average: 0.22 mg m⁻² d^− 1) and Lake Taihu (average: 0.15 mg m⁻² d^− 1), both experiencing serious eutrophication due to the presence of phosphorus (Chen et al. 2019d).

Fig. 5

The diffusive fluxes of Fe and P at the sediment-water interface in X1 of Tianbao Reservoir

Studies have revealed that factors such as DO, pH, ORP, and organic matter strongly influence the efflux of Fe²⁺ and reactive P moving from the sediments to the bottom water (Zhang et al. 2020). Research work by Wang et al. (2016) demonstrated that sediment pH and DO content were the primary drivers responsible for the seasonal release of soluble Fe²⁺ and P in the Hongfeng reservoir, China. Similar research by Rong et al. (2020) suggested that organic matter mineralization and redox conditions contribute substantially to the high labile P fluxes in sediments. To confirm the main factors responsible for the efflux of labile P and Fe in both Eucalyptus species sediments and sediments without Eucalyptus species, the Pearson’s correlation was used to explore the correlation (relationship) between the flux of labile Fe/P with their physico-chemical parameters within the sediments. The study showed that the release flux of labile P and Fe showed a strong correlation with pH_sediment, ORP_sediment, organic matter_sediment, and DO_water (Table 1) during the seasons in the different ecotype sediments, while the release flux of labile P and Fe showed no significant correlation with pH_water, ORP _water, and DOC_water. The results suggest the changes in the DO content (hypoxia/no hypoxia) at the bottom water influenced the efflux of reactive P. The results further indicated that the redox conditions, pH, and the presence of organic matter in the sediments influenced the seasonal changes in the efflux of labile P and Fe at X1 and X2.

Table 1 Pearson’s correlation of Fe and P efflux with physico-chemical properties of Tianbao reservoir

Resupply kinetics (replenishment) of P using the DIFS model

In this study, the replenishment characteristics of P in sediments was done at X2 (Eucalyptus dominated sediments). The output results of the DIFS software program are pictured in Table 2. Figure 6 shows the resupply kinetics of P in the surface sediments (0–40 mm) for summer (anoxic period) and winter (aerobic period) which was simulated by the DIFS model. From Fig. 6, the resupply kinetics of P in sediments in summer can be attributed to the steady condition of “partial case” while the resupply kinetics of P in sediments in winter can be termed as the non-steady condition of “partial case” (Harper et al. 2000).

Table 2 The input and output parameters of DIFS model for the surface (0–40 mm) sediment for summer and winter

Fig. 6

Time-dependence (DIFS model) of the measured R values for phosphorus in Tianbao reservoir for summer and winter

The R value provided a good analysis about the supplemental ability of reactive P (Harper et al. 1998). The mean R value for the surface sediments was higher in summer (0.7) than in winter (0.4) suggesting that the resupply ability of soluble P of the sediments is very high during hypoxia. Consequently, the T_c value during the anoxic period was lower (0.01 day) than that of the aerobic period (0.07 day) indicating that the equilibrium time of the sediment system during the anoxic period was shorter. This means that the response time of the sediment to depletion (release reactive P) is shorter in summer compared to winter (Harper et al. 2000; Yan et al. 2019). Oppositely, the k₁ and k_-1 values were shorter during the aerobic period (14.2 day^− 1 for k₁, 0.101 day^− 1 for k_-1) than the anoxic period (79.4 day^− 1 for k₁, 0.2 day^− 1 for k_-1) (Table 2). This revealed that a relatively high adsorption and desorption process occurred in anoxic sediment compared to the aerobic sediment (Harper et al. 2000). These results reveal that the sediment particles during hypoxia (summer) had a higher replenishment ability of phosphorus than during the aerobic period. Moreover, the findings show that the presence of oxygen in winter caused the adsorption process of Fe/Mn oxide to take place. This inhibited the efflux of P from the sediments.

Conclusion

The study investigated the seasonal dynamics of iron and phosphorus in the reservoir sediments in a Eucalyptus plantation region. This was achieved by using the novel planar optode, ZrO-Chelex DGT, and the DIFS model.

The study showed that there was no significant (p > 0.05) difference in the sediment concentration and fluxes of labile P at X1 (sediments without Eucalyptus species) and X2 (Eucalyptus species-dominated sediments), suggesting that the presence of Eucalyptus species did not significantly influence the efflux of labile P. However, the levels of labile Fe were significantly smaller (Fe: p value = 0.03) in the surface sediment at X2 compared to X1. This could be attributed to the elevated presence of Eucalyptus species’ organic matter (tannic acid) content at X2, which reacted and consumed labile Fe.

The concentration and the efflux of P and Fe at the SWB were synchronous at X1 and X2. This assertion was supported by the high correlations (positive) recorded in the study at both sampling sites (r > 0.8, p < 0.05). The findings further demonstrated that hypoxia influenced the seasonal efflux of labile P and Fe at the SWB through the decomposition (reduction) of Fe(III) in both Eucalyptus species-dominated sediments and sediments without Eucalyptus species. The results also suggest the changes in the DO content (hypoxia/no hypoxia) at the bottom water influenced the seasonal efflux of reactive P, providing valuable information to stakeholders/engineers on how to control Fe/P deterioration in a reservoir by controlling the oxygen content in the benthic region since DO is the driving factor influencing the efflux of P. The measured diffusion fluxes, based on Fick’s law, were positive for all the seasons and at all sampling sites (X1 and X2) suggesting that the sediments discharged labile P and Fe to the benthic water. A strong correlational relationship was pictured between labile P/Fe flux and pH_sediment, ORP_sediment, organic matter_sediment, and DO_water for X1 and X2, suggesting these parameters influenced the seasonal efflux of P in the Eucalyptus dominated reservoir. This information about the parameters (pH, DO, ORP, OM) which influence the efflux of P, serves as an indicator to stakeholders and engineers, that controlling pH_sediment, ORP_sediment, organic matter_sediment, and DO_water can serve as counteracting measure to control Fe/P mobility in a reservoir. The DIFS model further supported the above findings and showed that the replenishment rates of reactive P in the sediments were higher during the anoxic period (no oxygen) than the aerobic period, indicating that the adsorption process by Fe/Mn oxide (in the presence of oxygen in sediments) in winter inhibited the efflux of P from sediments. Our results gathered from this study (variations in dissolved oxygen, redox conditions, pH, and organic matter in the benthic region influence P and Fe mobility) provides valuable information to stakeholders, engineers, and researchers on understanding the geochemical transformation of P and Fe in the Eucalyptus dominated reservoir.