Plastic was originally engineered as an inexpensive, strong, and durable material with numerous benefits to human society, but it has become a symbol of the unintended consequences of human ingenuity. High rates of plastic accumulation in natural ecosystems have made plastic pollution a global environmental threat that proposes severe and potentially irreversible impacts to ecosystem functionality and its contribution to sustain human livelihood1,2,3.

The ecological functioning of the seafloor is extremely important for us humans4, but it experiences major plastic accumulation and especially coastal sediments suffer from direct and continuous inputs of new plastic from anthropogenic sources on land5,6,7. In coastal sediments, plastics can alter algal and microbial abundances, community structures and activity, as well as reduce benthic fauna energy reserves and alter fauna behavior and movement8,9,10,11,12,13. Such effects originate from the ingestion of the plastic itself14, the leaching of plastic additives15, and/or the colonization of the plastic as a new substrate16,17. These plastic-induced effects on marine ecosystems have been documented to occur within 0–75 days of plastic exposure. Disturbances from plastics and associated additives on the marine seafloor can have cascading and interactive physical, chemical, and biological effects, ultimately impacting coastal ecosystem functionality critical to planetary survival9,18,19. Sediment oxygen consumption (SOC) is a key process in coastal ecosystems as it integrates physical, chemical and biological processes influencing oxygen dynamics and it is indicative of ecosystem metabolism (i.e., the consumption and conversion of organic matter and energy). SOC plays a key role in multiple coastal ecosystem functions, including decomposition, carbon sequestration, nitrogen removal, nutrient recycling, and productivity20. Hence, if plastic pollution causes a response in sediment oxygen consumption, it is a sign of disruption of ecosystem functionality.

While recent research efforts have illuminated plastic pollution effects on ecosystem functionality, a crucial knowledge gap remains in testing real-world impacts19,21. Manipulative experiments isolate specific variables through the simplification of food webs and species composition in controlled laboratory environments, which serves as a double-edged sword. On one hand, it simplifies the complex interactions observed in natural ecosystems, allowing for a more precise examination of individual factors and a more mechanistic understanding of responses. On the other hand, this simplification obscures the complexity of natural ecosystems which is characterized by habitat heterogeneity, spatial and temporal variation, and interactions and feedback processes within natural communities22. For example, controlled studies collectively demonstrated that mesocosms with microplastics had higher sediment oxygen consumption rates than those without and attributed this to fauna type, fauna behavior or activity, and interactions between microplastics and fauna9,10,23. In a field survey, however, Ladewig et al. observed that the direction and magnitude of sediment oxygen consumption in Waitemata Harbour (New Zealand) is dependent upon the spatial variation of fauna abundance and type as well as the spatial variation of microplastic abundance and type19. Apart from spatial variation, another equally important driver of ecosystem functionality and stressor responses is temporal variation. Yet, assessments of disturbances to ecosystem functionality over time are lacking in plastic pollution research, limiting our capacity to inform decisions on plastic pollution management. To develop ecologically relevant strategies that reduce plastic pollution impacts and sustain ecosystem functionality and resilience, we need comprehensive ecological assessments of real-world ecosystem responses.

In this study, we investigate the real-world ecosystem response of coastal sediment metabolism to plastic pollution by conducting a manipulative field experiment. We inserted polyester netting pieces, representative of flexible fibers, into intertidal sediments and captured their effect on sediment oxygen consumption rates over time and relative to undisturbed sediments using sediment chamber incubations. To further link changes in the chemical characteristics of polyester pieces to metabolic disturbance, polyester pieces were chemically analyzed before and after experimental deployment. We first compare sediment oxygen consumption rates (SOC) of the control and plastic treatment (polyester netting) plots over the course of our experiment (7–72 days after plastic deployment) to identify effects of plastic pollution on sediment metabolism. Secondly, we identify additives present in the polyester, quantify the additives remaining in the polyester pieces over the time period of the experiment and test for correlations between additive concentrations and SOC to determine whether any additive lost from the deployed polyester is linked to changes in SOC.

Results

Rapid effects of plastic on SOC

Time since the addition of the netting had a significant effect on SOC rates (two-way ANOVA: df = 4, F = 8.214, p < 0.001), reflecting natural temporal variability typical of an in situ experiment. Plastic addition (treatment) also had a significant effect on SOC rates (two-way ANOVA: df = 1, F = 5.059, p = 0.032). Post-hoc pairwise comparisons revealed that there were significantly lower SOC rates in the polyester treatment 7 days after plastic deployment (two-way ANOVA, Holm-Sidak pairwise comparison, p = 0.021) with the average SOC rate of the polyester treatment being 16% lower with 1.57 mmol O2 m−2 h−1 compared to 1.87 mmol O2 m−2 h−1 (Fig. 1A). At 13 and 22 days after plastic deployment, the average SOC rates were 12% and 13% lower in polyester treatment plots compared to control plots, but significant differences (p < 0.05) in these rates were not significantly different (two-way ANOVA, Holm–Sidak pairwise comparison, day 13: p = 0.138, day 22: p = 0.054) Interestingly, alongside the difference in mean SOC rates, we also observed a 77% lower standard deviation in the plastic treatment compared to the control at day 7 (Fig. 1A). This difference in variation of SOC rates was also present but weaker at day 13 and 22, with 33% and 38% lower standard deviation in the plastic treatment compared to the control, respectively. 51 and 72 days after plastic deployment, SOC rates were similar in both treatments (two-way ANOVA, Holm–Sidak pairwise comparison, day 51: p = 0.510, day 72: p = 0.608) and standard deviations of plastic treatment and control differed by 24% and 12% (Fig. 1A). There was no interaction effect of time and treatment in SOC rates (two-way ANOVA: df = 4, F = 2.249, p = 0.090). The deviation of SOC rates of polyester treatment plots from respective control plots (absolute value of control SOC minus absolute value of plastic SOC) significantly and negatively correlated with time (Pearson Correlation, r =  − 0.92, p = 0.03, Fig. 1B). These results suggest that polyester pollution rapidly reduces SOC rates and its natural variability within 7 days of plastic exposure and recovery to control SOC rates and natural variability is possible a few weeks after the polyester pollution impact.

Figure 1
figure 1

(a) SOC (mean ± std) of plastic (black) and control (white) plots, (b) absolute difference of mean SOC in plastic treatment plots to control plots over time. Day 0 = day of plastic deployment.

Full size image

Loss of organic additive DEHP

Of the inorganic and organic additives targeted from pristine and recovered plastic pieces, only the organic additive DEHP was observed above the method detection limit in all extracts with a clear decrease in concentration over time (Fig. 2A). The amount of DEHP extracted from the plastics decreased by 89% over the first week (from day 0 to day 7), followed by another 5% drop in DEHP concentration from day 7 to day 13. After day 13, the average DEHP concentrations did not differ as time progressed, reaching an average concentration of ~ 1.5 ± 0.3 µg DEHP g−1 polyester. DEHP concentrations extracted from pieces incubated in running seawater under dark conditions (day 120) were comparable to those extracted from field-deployed pieces (2.46 ± 1.73 µg DEHP g−1 polyester), indicating that the leaching of additives from plastic can occur in the absence of photodegradation. The sharp decrease in extracted DEHP concentrations from the plastic pieces at early time points indicates that DEHP was rapidly lost from pristine polyester pieces during the first 13 days of the experiment. Correlations between the extracted DEHP concentrations and the change of SOC in polyester treatment compared to control plots further suggest that the plastic impact on SOC could be linked to the loss of DEHP (Pearson’s correlation analysis, r = 0.87, p = 0.05, Fig. 2B). Fourier-transform infrared (FTIR) spectroscopy confirmed the loss of chemical functional groups (Fig. S1), which may be associated with any part of the plastic, including additives.

Figure 2
figure 2

Concentrations of the organic additive di(2-ethylhexyl) phthalate (DEHP) from polyester extracts at the time points of the experiment (A) and correlation between DEHP concentrations and the change of SOC on the polyester treatment relative to the control (B). The DEHP concentration at timepoint − 2 represents the DEHP mass concentration of pristine polyester before deployment. Data in panel A are presented as averages and standard deviation of three replicate measurements using liquid injection GC/MS.

Full size image

Discussion

A rapid ecosystem response to plastic pollution

The rapid change to SOC provides real-world evidence that polyester contamination affects seafloor ecosystem metabolism. This ecosystem response was observed from day 7 to day 22 followed by recovery by 72 days, which suggests it was caused by a quickly acting stressor, such as leachate of the toxic additive DEHP.

DEHP is a common plasticizer which is used in the production of synthetic textiles24. It is not chemically bound to plastic polymers and can be easily released to the environment25,26. Indeed, the correlations between the mass concentration of DEHP in our polyester pieces and SOC variables, and the remaining presence of the polyester pieces throughout the study, suggests that the rapid effects to SOC were not from the polyester pieces itself but instead associated to the 89% reduction in DEHP concentration in plastic pieces over the first 7 days of plastic deployment. Similarly, Paluselli et al. measured the highest release of phthalates from PVC and PE to seawater within the first two weeks, with an increase of phthalate concentration in seawater by ~ 75–90%27. There is a growing body of evidence that DEHP and other additives leaching from plastics accumulate in marine sediments and the marine food web25,28,29, raising concerns about its potential toxicity30,31. DEHP has been found to affect growth, activity, and energy metabolism of marine fauna32,33,34, as well as microbial activity and community structure (i.e., carbohydrate metabolism, amino acid metabolism, membrane transport, and signal transduction)35. Additionally, interactive effects such as the reduction of microbial aerobic mineralization due to reduced fauna bioturbation activity through either eco-toxicological effects on fauna or migration of larger macrofauna away from plastic contaminated sediments may have influenced sediment oxygen consumption in our plastic treatment9,23. Since the activity, structure and interaction of both benthic fauna and microbial communities are a major part of sediment ecosystem metabolism, the presence of DEHP likely played a role in the drop of magnitude and variability of the measured SOC rates in our plastic treatment. Photodegradation of DEHP could be another process that led to the decline in DEHP mass concentration in our polyester pieces36. However, our polyester nettings were deployed vertically within the sediment and only a rim of the netting reached out of the sediment surface (~ 0.2 cm). With the majority of the nettings being positioned inside the sediment where light does not penetrate, photodegradation likely played only a minor role in the decline of DEHP concentrations in plastic-deployed pieces.

Through the extended time period of our in situ experiment (72 days) we could further observe that our treatment plots recovered from the plastic impact roughly 2 months after plastic deployment. Previous mesocosm studies detected effects of microplastic pollution on sediment oxygen consumption in around the same time frame as our study but only measured one sampling time point (i.e., after 7 days—Coppock et al. 2021, after 20 days—Urban-Malinga et al. 2021, and after 24 days—You et al. 2023)9,10,23. Therein, microplastic effects to SOC were explained to be driven by species23, fauna behavior or activity9,10,23, and interactions between microplastics and fauna9. While these authors illustrate the complexity of microplastic effects on sediment oxygen consumption in controlled mesocosm experiments, temporal effects of plastic additives were not considered to be the cause of any of these experimental outcomes. Temporal aspects of plastic pollution are often integrated in studies on the distribution and accumulation of plastic and their additives in marine ecosystems, but it is currently unknown how ecosystems respond to plastic pollution changes over time. The recovery of SOC in our in situ experiment clearly shows that there is a temporal aspect of polyester impacts on ecosystem metabolism, likely driven by the physical dissipation of DEHP, re-distribution of leached DEHP by macrofauna37,38, or microbial degradation of leached DEHP39. FTIR spectra confirmed losses of functional groups from the plastic over time (Fig. S1), which may be an indication of plastic degradation (to form microplastics) and/or the loss of plastic additives. Further testing is required to confirm if effects to SOC were associated with the potential degradation of polyester pieces to microplastics. However, since effects were only present early on in the experiment, it is unlikely effects were due to the generation of microplastics which is expected to increase with time. While 6 time points were achieved in this in situ study amongst constraints around tide timetables, daylight availability, and health and safety measures related to intertidal field work, more frequent and consistent sampling would contribute further to the understanding of plastic pollution impacts on SOC, specifically in tracking inorganic and organic plastic-derived leachates.

The short-term reduction in average SOC rate by 12–16% and variability by 30–80% serve as an indicator of ecosystem health, and—here—indicates a temporarily compromised intertidal ecosystem. Intertidal flats are among the most productive areas of the coastal zone with high turnover rates of organic matter and nutrients40. In Mahurangi Harbour, specifically, intertidal flats, like the one used in our experiment, are functional hotspots of the estuary, i.e. a habitat that has high value for ecosystem functioning41. Since SOC is strongly linked to various ecosystem functions, such as organic matter decomposition, carbon sequestration, nitrogen removal, nutrient recycling, even a small change in average SOC can be indicative of a disruption in the functioning of the intertidal flat. Moreover, the 80% reduction in SOC variability compared to control SOC variability alongside the lower average SOC rates indicates homogenization of this ecosystem process in response to plastic contamination. Homogenization of the seafloor community through stressors like physical disturbance have been linked to reduced ecosystem resilience42,43. While monitoring data have shown declines in macrofauna community composition linked to land-use, water and sediment quality, Mahurangi Harbour remains an oligotrophic and biodiverse estuary44. In this context, the short-term response in SOC we detected, while not indicative of severe issues, reveals the ecosystem’s vulnerability to change, which likely exacerbates under multiple co-occurring stressors. To our knowledge, no research currently exists in the published peer-reviewed literature regarding the cumulative effects of plastic pollution combined with other stressors on sediment oxygen consumption. Exploring combined impacts of plastic pollution with additional stressors, such as contaminant pollution (metals, PAHs, herbicides/pesticides), terrestrial sediment loading, nutrient enrichment or climate change could shed light on cumulative effects and real-world ecosystem responses.

A directed approach to guide change

We designed this in situ experiment to detect ecosystem-level effects from polyester contaminated sediments over a time-scale of days to weeks. While it is essential to acknowledge that our experiment was not designed to assess longer-term effects spanning a year or more, this aspect merits consideration in future research endeavors, as plastics will persist in seafloor sediments for generations. Long-term investigations are better positioned to investigate the accumulation of stressors, including the sorption of environmental contaminants by plastic45, which may exert more potent influence on organisms than the immediate impact and dispersal of additives. We can rule out that effects to ecosystem functioning were associated with physical changes to the sediment structure, as our choice of flexible plastic netting and deployment strategy ensured minimal disturbance to sediment structure, as the plastic netting pieces were vertically positioned, accommodating the vertical movements of the resident organisms. The netting’s strategically sized apertures of 2–3 mm also ensured unimpeded horizontal migration of subsurface organisms. Lastly, and importantly, we retrieved 98% of the deployed polyester pieces, even in the face of severe weather events. With the in situ ecosystem response we detected, we are confident that the concept of our study opens the door to further field research obtaining information beyond this study’s capabilities, such as long-term effects of polyester remaining in coastal sediments, in situ leachate effects on benthic carbon or nitrogen cycling processes, and in situ effects of other or combined plastic types in coastal marine sediments.

Conclusion

Polyester and DEHP, amongst other plastics and additives, have already raised environmental and human health concerns, but this is the first assessment of their influence on marine ecosystem functionality in situ. One of the key questions posed by this research is whether the observed effects on ecosystem functioning are primarily attributable to the plastic itself or to bound chemicals with such findings applicable to judicious choices regarding material usage. In our experiment, it appears the effects were caused by the rapid release of DEHP from the polyester material. Although the polyester lost the majority of DEHP content within just seven days, the polyester managed to retain some DEHP—and thus preserve its physical properties, i.e. flexibility and durability—throughout the 72-day exposure to the environmental conditions of coastal intertidal sediments. Making polyester materials more flexible and durable is the function of plasticizers; DEHP makes polyester materials soft (e.g., clothing and bed sheets) and easier to process during manufacturing. This raises the compelling question of whether plastic manufacturers could either reduce plasticizer addition or shift manufacture to less toxic plasticizers. The DEHP mass concentration in our pristine polyester netting extracts was 23.1 ± 4 µg/g, which is at the higher end of reported concentrations of DEHP in polyester fibers (range: 1.1–26 µg/g)46. With DEHP concentration at the lower end of those reported by Tang et al.46, polyester fragments may not cause an ecosystem-level response in coastal sediments thus reducing at least one of the multiple impacts that plastic pollution induces. Moreover, explorations on ‘greener’ plasticizers and chemical recycling of polyesters and their additives pose possible alternatives supporting ecosystem conservation from plastic pollution47,48. Ultimately, the findings of our experiment underscore the urgency of addressing plastic pollution realistically whilst encouraging that less DEHP-bound plastic material should be released from society into marine environments. Future plastic pollution research needs comprehensive assessments of ecosystem responses to lay a solid groundwork for informed conservation initiatives and policy formulations aimed at safeguarding the vitality and adaptability of marine ecosystems.

Methods

Study site and experimental design

Experiments were conducted in Mahurangi Harbour, North Island, New Zealand (36° 28ʹ 8.9256ʺ S, 174° 44ʹ 1.5108ʺ E). Mahurangi Harbour is a shallow (< 15 m), tidal (range ~ 3 m), well-mixed estuary covering an area of 25 km2, of which ~ 60% are intertidal flats49,50. Our study site was established in a shallow bay on an intertidal sandflat (mid tide water depth = 2 m, mean grain size = 133 µm, 11.8% mud content). The experimental study was employed from January 2022 to April 2022 with six sampling timepoints. Experimental areas were 2 × 2 m, each containing an control and plastic treatment of 0.5 × 0.5 m. Each treatment had 24 replicates with 4 replicates randomly allocated to sampling on each of the 6 time points, totaling 48 plots.

Plastic pieces were cut from pristine polyester netting material (www.spotlightstores.com, 100% polyester net ivory, EAN 9349336114663, thickness: 10 µm) to the size of 6 × 2 cm (Fig. 3A). Polyester is used in various textiles and materials, has a density (1.38 g cm−3) greater than seawater and thus would be expected to accumulate on the seafloor. Flexible plastic netting was selected for this study to reduce disturbance to biophysical processes, such as sediment stabilization, porewater, and particle transport. We deployed polyester (i.e., polyethylene terephthalate, PET) pieces vertically into sediments to allow for vertical migration of flora and fauna around the inserted plastic pieces (Fig. 3B, C, D). To insert the net consistently, 1 cm of the end of each piece was wrapped around the end of a thin spatula and carefully pushed vertically into sediments so that the top of the pieces was flush with the sediment surface and the bottom of the pieces reached 5 cm sediment depth (Fig. 3B, D). A subsurface clay layer met the bottom 1 cm of plastic pieces which helped to hold them in during deployment. Sediments in control plots were disturbed by the spatula similarly to plastic plots but without the addition of plastic pieces. In each plot, 81 pieces of plastic were deployed, and in total took up 0.06% of surface area (1.5 cm2 of 2500 cm2) of each plastic plot. The abundance of plastic pieces applied in our experimental plots is in the same order of magnitude as abundances found for microplastic in the same region, but at the higher end of abundances reported for a similar size class from other regions in the world (Table S1)8,19,52,55,56. Plastic pieces were polyester material, confirmed via FTIR (Fig. S1)51. Fiber shape polyester pieces were selected because they are one of the most common types and shapes of plastics found in sediments of this region of New Zealand19,52,53. Additionally, polyester has hydrolysable chemical bonds in its hydrocarbon backbone that are more susceptible to degradation than some other types of plastic54. Plots were left to stabilize for one week after plastic deployment before in situ chamber incubations were carried out. Plastic pieces were also incubated in a bucket with overflowing, filtered (1 µm) seawater under dark conditions with no sediments for 120 days to assist understanding of polyester degradation in seawater versus marine sediments.

Figure 3
figure 3

Equipment and plot design. (a) Flexible polyester netting pieces used in the experiment. (b) Method for vertical deployment of plastic pieces by spatula into the seafloor. (c) Illustration of the top view of the plastic plot after plastic deployment. (d) Image of the top view of plastic plot after plastic deployment. (e) Experimental plots outlined by the base of benthic chambers holding a pump and O2 optode.

Full size image

Sediment oxygen consumption

In situ chamber incubations were carried out 7, 13, 22, 37, 51, and 72 days after plastic deployment based on previously published literature noting plastic pollution effects to SOC9,10,23. Chamber bases (50 × 50 × 15 cm height, stainless steel) were carefully pressed 5 cm into the sediment just after the incoming tide (10–20 cm water depth) inundated the site (Fig. 3E). Once the incoming tide reached ~ 50 cm depth, acrylic domes were used to seal 40 L of ambient seawater over the sediments. Black PVC-sheets were fixed atop the chambers to create dark conditions, negating photosynthetic activity. Hobo light pendants (Onset HOBO Pendant Temperature/Light Data Logger, Bourne, Massachusetts, USA) monitored light and temperature conditions inside and outside of chambers and were compared after incubations to assure dark conditions were achieved and temperature variation was minimal. Oxygen optode loggers (PME miniDOT Logger) were mounted inside all chambers of control and plastic treatment plots for continuous oxygen concentration measurement at an interval of 2 min during the incubation (Fig. 3E). Water inside the chambers was mixed with pulsed, non-directional pumps (SBE5M-1, Seabird electronics, flow rate of 25 mL s−1 for 5 s every 45 s, Fig. 3E). Chambers were incubated for 3.5–5 h over midday high tide. SOC was determined as flux of oxygen into the sediment using.

$${\text{SOC}} = {\text{ s}} * {\text{V}}/{\text{A}}$$

where s is the slope of regression from linear decrease of O2 during incubation (mol L−1 h−1), V is the incubated water volume (L), and A is the incubated sediment area (m2). SOC was only calculated from incubations with linear oxygen concentration decreases over the incubation time with an r2 > 0.97. Timepoint 4 at 37 days after plastic deployment was one day after a cyclone and chamber incubations were not successful. Of the other 40 chamber incubations, 37 were successful with a linear oxygen decrease.

Characterization of polyester additives

After chamber incubations, plastic pieces were retrieved from the plots at a 98% recovery rate. A subset (n = 3) of plastic pieces was taken for chemical analysis from each time point, pristine pieces, and plastic pieces incubated in running seawater while under dark conditions. Plastic samples were screened for both inorganic and organic additives. Different pieces of plastic were used for each inorganic and organic extraction. Controls were performed in triplicate for each set of extractions where an extraction vessel was filled only with the solvent of interest before undergoing the same extraction and analysis procedures as the samples.

To allow for chemical analysis, the plastic pieces were solvent extracted in an incubator shaker (NB-205, N-Biotek) at 50 °C for 72 h. The incubator shaker was set to 100 rpm to ensure that the plastics were completely wetted for the duration of the extraction. Type I H2O (Sartorius, 18.2 MΩ) was used as an extraction solvent for the inorganic analyses while acetonitrile (Macron, ChromAR grade) was used as an extraction solvent for the organic analyses57. A harsher solvent was not selected in order to best preserve the integrity of the plastic pieces. Extracts intended for inorganic analysis were prepared as 1% HNO3 + 1% HCl solutions prior to injection onto the inductively coupled plasma – mass spectrometry (ICP-MS; Agilent 7700) system. Water extracts were screened for Ti, Fe, Co, Ni, Zn, As, Mo, Cd, Sb and Pb (Limit of detection (LOD) = 2 ppb). The acetonitrile extracts were injected directly onto a gas chromatograph—mass spectrometer (GC/MS) for organic analysis, based on previous work by Rindelaub et al.57. The GC/MS system employed a 5% phenyl/95% dimethylpolysiloxane capillary column (Phenomenex Zebron 5-MSi; 30 m, 0.25 mm i.d., 0.25 µm) that was kept at 40 °C for 5 min before being heated at a rate of 10 °C min−1 to 250 °C with a 3 min final hold. The column flow rate was 1 mL min−1, and the injection was kept at 250 °C. Ionization was accomplished using electron impact at 70 eV, along with an MS scanning range of 40–1000 m/z. Both untargeted and targeted analyses were performed. For targeted analysis, concentrations of benzyl butyl phthalate, bis(2-ethylhexyl) adipate, dibutyl phthalate, diethyl phthalate, dimethyl phthalate, di-n-octyl phthalate, and di(2-ethylhexyl) phthalate (DEHP) were quantified via comparison to external standards (EPA 506 Phthalate Mix, Sigma Aldrich). The limit of detection (LOD) and limit of quantification for DEHP were LOD = 0.074 µg mL−1 and LOQ = 0.24 µg mL−1. The LOD and LOQ were calculated as follows: LOD = 3σ/S and LOQ = 10σ/S, where σ is the standard deviation of the signal and S is the slope of an external calibration curve.

Data analysis

Two-way analysis of variance (ANOVA) was used to test the significance of differences (p < 0.05) between treatments (control vs. plastic treatment plots), time (chamber incubation time points), and the interaction (treatment * time). Normality of data was analyzed using the Shapiro–Wilk test and homogeneity of variances was analyzed using the Brown–Forsythe test. Post-hoc multiple comparison was conducted using the Holm–Sidak adjustment to identify differences between groups when ANOVA test results were significant (p < 0.05). Relationships between variables were tested using Pearson’s correlation coefficient (r). All statistical tests were run in SigmaPlot (version 14).