Tissue Hydrogen Peroxide Concentration Can Explain the Invasiveness of Aquatic Macrophytes: A Modeling Perspective

In recent years, an invasive macrophyte, Egeria densa, has overwhelmingly colonized some midstream reaches of Japanese rivers. This study was designed to determine how E. densa has been able to colonize these areas and to assess the environmental conditions that limit or even prevent colonization. Invasive species (E. densa and Elodea nuttallii), and Japanese native species (Myriophyllum spicatum, Ceratophyllum demersum, and Potamogeton crispuss) were kept in experimental tanks and a flume with different environmental conditions. Tissue hydrogen peroxide (H2O2) concentrations were measured responding to either individual or multiple environmental factors of light intensity, water temperature, and water flow velocity. In addition, plants were sampled in rivers across Japan, and environmental conditions were measured. The H2O2 concentration increased in parallel to the increment of unpreferable levels of each abiotic factor, and the trend was independent of other factors. The total H2O2 concentration is provided by the sum of contribution of each factor. Under increased total H2O2 concentration, plants first started to decrease in chlorophyll concentration, then reduce their growth rate, and subsequently reduce their biomass. The H2O2 concentration threshold, beyond which degradation is initiated, was between 15 and 20 µmol/gFW regardless of the environmental factors. These results highlight the potential efficacy of total H2O2 concentration as a proxy for the overall environmental condition. In Japanese rivers, major environmental factors limiting macrophyte colonization were identified as water temperature, high solar radiation, and flow velocity. The relationship between the unpreferable levels of these factors and H2O2 concentration was empirically obtained for these species. Then a mathematical model was developed to predict the colonization area of these species with environmental conditions. The tissue H2O2 concentration decreases with increasing temperature for E. densa and increases for other species, including native species. Therefore, native species grow intensively in spring; however, they often deteriorate in summer. For E. densa, on the other hand, H2O2 concentration decreases with high water temperature in summer, allowing intensive growth. High solar radiation increases the H2O2 concentration, deteriorating the plant. Although the H2O2 concentration of E. densa increases with low water temperature in winter, it can survive in deep water with low H2O2 concentration due to diffused solar radiation. Currently, river rehabilitation has created a deep zone in the channel, which supports the growth and spreading of E. densa.


INTRODUCTION
Macrophyte responses to environmental conditions are species specific, and invasive plants tend to exhibit more tolerance than native species (Zerebecki and Sorte, 2011;Bates et al., 2013). Therefore, invasive species are able to dominate or distribute in areas where native species fail to survive. Among different invasive aquatic macrophytes, Egeria densa is a well-known worldwide species that causes significant ecological issues in freshwater ecosystems. In Japan, E. densa was used as an ornamental aquarium plant in the early 19th century. However, it has escaped into natural freshwater bodies and became naturalized in the 1940s. Although E. densa mainly invaded lakes during the initial spreading stage in the 1970s (Kadono, 2004), this species has been recorded increasingly in many western Japanese rivers over the last two decades (MLIT, 2019). These rivers were originally nearly free of macrophytes and consisted of gravel beds and hyporheic flow (Tanida, 1984;Hauer et al., 2016). Though native species (e.g., Myriophyllum spicatum, Potamogeton crispuss, and Ceratophyllum demersum) were colonized patchily, no large colonies were found in major rivers (Kunii, 1982;Kadono, 2004). Another alien species, Elodea nuttallii, also invaded at nearly the same time in 1961. However, it did not produce large colonies except for lakes and small streams. In contrast, E. densa spread to cover the entire river channel of major rivers. The widespread colonization of E. densa has led to extreme changes in these river ecosystems. After establishment, E. densa behaved as ecological engineers, changing the environment to their benefit (Schoelynck et al., 2012;Schoelynck et al., 2014). They reduced water flow velocity and attenuated wave energy, leading to particle settlement and, consequently, hyporheic flow capacity reduction (Madsen et al., 2001;Boano et al., 2014). It has also caused economic losses. For example, the presence of macrophytes substantially decreases the yield of Ayu (Plecoglossus altivelis altivelis), a grazer of benthic algae (Kawanabe, 1970). Casual monitoring between present-day abiotic conditions and plant traits, such as growth rate and biomass, is the method commonly used to evaluate the preferable habitat for macrophyte species (Barko et al., 1991;Riis et al., 2012;O'Hare et al., 2018). However, environmental conditions frequently change, and there are various types of effective factors in the natural rives. Thus, it is difficult to apply the monitoring system in the field, particularly to derive the most influential factor.
Aquatic macrophytes growing in their natural environment often face an array of unpreferable environmental conditions, for example, too low or too high water temperatures, high flow velocity , pollution, or substrate alteration . They can survive and propagate if the conditions remain within the plants' tolerance levels. When the environmental conditions exceed the tolerance thresholds for a considerable period of time, macrophytes become stressed, lose their colonization capacity, and ultimately decay. However, following a short-term exposure to such conditions, they can recover, depending on the extent of the damage caused and the characteristics of the species (Weerakoon et al., 2018). Thus, the presence of a specific macrophyte species in an area depends on whether environmental factors are within their tolerance levels as well as on the duration of the exposure. When plants are subjected to unpreferable environmental conditions, reactive oxygen species (ROS) are generated in different organelles (Zaman and Asaeda, 2013;Das and Roychoudhury, 2014;Choudhury et al., 2017;Helaly et al., 2017;Parveen et al., 2017;Asaeda et al., 2018;Elsheery et al., 2020a;Elsheery et al., 2020b), which damages the plant body by the oxidative stress. Some ROS are scavenged relatively quickly by antioxidants (Omar et al., 2012), and the homogeneity of ROS in tissues is maintained by balancing the ROS and antioxidants. The balance flips over when oxidative stress surpasses the scavenging capacity of the antioxidants (Naser et al., 2016;Dumont and Rivoal, 2019). Among ROS, hydrogen peroxide (H 2 O 2 ) is widely generated (Asada, 2006;Sharma et al., 2012), relatively stable, and can be easily measured (Satterfield and Bonnell, 1955;Zhou et al., 2006;Asaeda et al., 2020). The concentration of H 2 O 2 in plant tissues does not depend on a particular stress but is subjected to sum magnitude of unpreferable environmental conditions (Suzuki et al., 2014;Asaeda et al., 2020). Thus, H 2 O 2 concentration in the plant tissue can be used as an indicator of the physiological status of a particular macrophyte species (Smirnoff and Arnaud, 2019). The system has been used for E. densa, which has successfully identified the channel slope that it can colonize (Asaeda et al., 2020).
The trend of H 2 O 2 concentration is likely as a result of a long history of acclimatization to the natural condition of a particular area; thus, it may vary widely between native and invasive species. To apply tissue H 2 O 2 concentration as an indicator to elucidate the intensive growth of invasive species, it is necessary to determine the relationship between H 2 O 2 concentration and environmental factors both for native and invasive species. The main objective of the present study is to 1) empirically determine the H 2 O 2 concentration generated by unpreferable conditions of abiotic environmental factors for both native and invasive species, 2) develop the model to predict the environment where these species can colonize, and 3) elucidate the reason for the overwhelming growth of E. densa in rivers of particular areas.

Experimental Methodology
In the experiment, invasive macrophyte species (E. densa and E. nuttallii) and major Japanese species (C. demersum, P. crispuss, and M. spicatum) were tested (MLIT, 2019). They were exposed to different types of physical conditions, temperature, irradiance, and water flow velocity, following the range of the rivers where these species were colonized from ∼8°C in winter to 30°C in summer for water temperature, 0-1,200 µmol/m 2 /s for the irradiance in water, and 0-50 cm/s for flow velocity (MLIT, 2019). For the laboratory experiments, healthy macrophyte stocks were collected from the Saba River (E. densa) and the Moto-Arakawa River near Tokyo (E. nuttallii, C. demersum, P. crispuss, and M. spicatum). Collected plants were cleaned with water to remove debris, and any attached macro-algae were carefully separated with tweezers. The plants were then cultured in a glass tank at 25 ± 2°C under a 12/12 h photoperiod with photosynthetically active radiation (PAR) (∼125 µmol/m 2 /s using fluorescent lamps) for over 2 months. Commercial sand (D 50 < 0.1 mm) was used as a substrate, and 5% Hoagland solution was provided as the nutrient medium . Algae were removed weekly, and algae-free plants were used in the experiments. Three types of experiments (triplicate) were conducted in total, each focusing on different combinations of environmental factors.

Experiment 1: Water Temperature and Irradiance
A number of studies have reported that water temperature can significantly affect the abundance of different aquatic plant species (Pip, 1989;Barko et al., 1991;Lougheed et al., 2001;Pandit, 2002). An experiment was conducted to identify the increment of H 2 O 2 concentration of the plant tissue under different water temperatures and irradiance levels, and, thereby, to make empirical relations between these factors. Several light levels (0-1,300 µmol/m 2 /s of PAR) were tested in small aquaria (dimensions: 50.0 cm × 35.0 cm × 35.0 cm). Temperature level was maintained at 10 ± 2 (E. densa), 15 ± 2 (E. densa), 20 ± 2, 25 ± 2 (E. densa), and 30 ± 2, 35 ± 2°C using a temperature controlling system (Aquarium cooler ZC-100α, Zensui Corporation, Tokyo, Japan). PAR intensity was irradiated under natural solar radiation or using LED lights (Model LT-NLD85L-HN, OHM Electric Inc., Japan) with a 12 h light:12 h dark photoperiod for 3 weeks.

Experiment 2 and 3: Flow Velocity and Irradiance
This experiment was designed to test the effect of water flow velocity on the H 2 O 2 concentration of the plant tissues and the interaction with irradiance . Two sets of experiments were conducted. In the first experiment, experimental plants (E. nuttallii, P. crispus, C. demersum) were exposed to two water flow levels (16 and 25 cm/s) using custom-made recirculating flumes (dimensions: 240 cm long × 25 cm width × 28 cm depth) exposed to artificial light intensity by the LED lights, or dark conditions. Pre-aerated tap water was circulated by centrifugal electric motor pumps. Pre-acclimatized potted plants were allocated to a section in the flume where water was introduced though a gradually shrinking entrance section to reduce turbulence. Plants were continuously exposed to low or high mean flow velocities for up to 4 days. During the experiment, mean water flow velocity was detected using an ultrasonic velocimeter (Tokyo Keisoku Co. Ltd., Japan) directly above the plant leaf surface and recorded daily to minimize flow variation. Temperature level was maintained at 15 ± 2°C using an aquarium water temperature controlling system (Aquarium cooler ZC-100α, Zensui Corporation, Tokyo, Japan). Stress assays by means of H 2 O 2 measurements were performed every 3 h from 6:00 to 18:00 after 4 days' exposure, and each treatment contained three replicate flumes. For another experiment, a flume channel 2.4 m long, 25 cm wide and 22 cm depth was constructed outdoors. Eighteen flat pots with more than three E. densa plants were carefully and randomly installed. Water temperature was kept at 25 ± 2°C throughout the experiment. Flow velocities from stagnant to 40 cm/s were employed under different solar radiation, and after 3 h, three plants were sampled at each time and a stress assay was conducted immediately. PAR intensity in the water was measured with a portable quantum flux meter (Apogee, MQ-200, United States).

Field Observations
Several rivers that are highly colonized by E. densa were selected from the species distribution database in Japan (MLIT, 2019). The selected rivers were assessed to obtain detailed location information pertaining to the colonization of E. densa. respectively. E. nuttallii was sampled in July and September 2018 from the same river. Sampling was conducted approximately every 3 h in the light-exposed and dark-adapted conditions to remove the effect of solar radiation. The dark treatment involved placing a black plastic sheet (3 m × 3 m) floating over part of the plant colony for 30 min. The 30 min pre-dark period was determined by laboratory experiments, which were specifically conducted to determine the optimum pre-darkness duration (data not shown). In August 2017, a sampling of M. spicatum was conducted in the Sakuradabori of the Imperial Palace Moat, at the center of Tokyo, where M. spicatum made a mono-specific stand. The depth of the sampling site was 0.3 m-2.5 m. Both solarexposed and dark-adapted samples were taken. Plant biomass was sampled from 50 cm × 50 cm quadrats in all sampling sites. The plant samples were placed in plastic bags and immediately stored in a cooling box containing dry ice for transfer to the laboratory where it was stored at −80°C until an H 2 O 2 assay and chlorophyll estimation were conducted.

Determination of Shoot Growth Rate, H 2 O 2 and Chl-a Concentrations
The length of the plants grown in the experimental units was measured using a millimeter scale at 5-7 day intervals. The shoot growth rate (SGR) was calculated as the difference in shoot length between two observations divided by the duration, and it was expressed in cm/day. At the end of each experiment, fresh plant shoots were extracted (∼500 mg) in an ice-cold phosphate buffer (50 mM, pH 6.0) that contained polyvinylpyrrolidone (PVP), and the extractions were centrifuged at 5,000 × g for 20 min at 4°C. This extraction was used to analyze the H 2 O 2 content spectrophotometrically following the TiSO 4 method (Satterfield and Bonnell, 1955) with modifications. The reaction mixture contained 750 µl of enzyme extract and 2.5 ml of 1% TiSO 4 in 20% H 2 SO 4 (v/v), which was centrifuged at 5,000 × g for 15 min at 20°C. The optical absorption of the developed yellow color was measured spectrophotometrically at a wavelength of 410 nm. The H 2 O 2 concentration in samples was determined using the prepared standard curve for known concentration series and was expressed in µmol per gram fresh weight (µmol/gFW). Chlorophyll a (Chl-a) concentrations of experimental plants were determined spectrophotometrically (UV Mini 1210, Shimadzu, Japan) by extracting pigments with N,N-dimethylformamide after keeping them in darkness for 24 h, and they were expressed in terms of fresh weight (FW) (Wellburn, 1994).

Statistical Analysis
Data were tested for normality with the Shapiro-Wilk test before statistical analyses. All results were presented as the mean ± SD of more than three replicates. Data were subjected to a one-way analysis of variance (ANOVA) with Tukey's post-hoc test for mean separation. The t-test was performed where necessary. Bivariate analysis was used and followed by Pearson's correlation to evaluate the relationship among parameters. Statistical analyses were performed in IBM SPSS V25. Asaeda et al. (2020) proposed the total H 2 O 2 concentration formed in plant tissues for a particular temperature (Temp) by the sum of H 2 O 2 generated by metabolism (H 2 O 2met ), flow velocity (H 2 O 2vel ), and solar radiation (H 2 O 2rad ). If the value is between 15 and 20 µmol/gFW, then E. densa growth deteriorates.

Development of the Species-Specific Model to Identify the Colonization Zones
For other species, empirical formulas obtained by experiments and field observation were introduced to H 2 O 2 concentrations generated by each environmental component, solar radiation, H 2 O 2rad (Temp), temperature inclement, H 2 O 2rad (Temp), the basal level of the metabolism, the H 2 O 2met (Temp) (Apel and Hirt, 2004), and the threshold level to deteriorate, H 2 O 2cr .
In rivers flowing with moderate velocity, water is fully mixed. Therefore, the light attenuation coefficient is nearly uniform at all depths, and the light intensity is given by I 0 exp(−kz), where I 0 is the light intensity just below the water surface, k( 0.083 cm −1 ) is the attenuation constant of light in water, and z is the canopy depth. The intensity of solar radiation, I 0 (µmol/m 2 /s), and water temperature (°C) at the Eno and Saba rivers are empirically given as a function of month, month: Flow velocity in a river channel "Vel" (cm/s) is estimated by the Manning's equation, assuming the channel is sufficiently wide compared to the depth and is longitudinally uniform, such that: where R is the hydraulic radius, approximately given by the depth H (cm), S is the channel bed slope, and n is the Manning's roughness coefficient, where n is ∼0.090 in the river zones considered in the present study (personal information).

Empirical Relationships of H 2 O 2 Concentration With Abiotic Factors
Combined effects of temperature and light intensity on H 2 O 2 formation in macrophyte tissues showed a species-specific response ( Figure 1). The basal H 2 O 2 concentrations were 4.6 µmol/gFW at 20°C for E. densa and E. nuttallii, and 3.0 µmol/gFW at 20°C for other species, respectively, after being exposed to dark conditions. Water flow velocity and light intensity had significant impacts on the H 2 O 2 metabolism in macrophytes. The tissue H 2 O 2 concentration linearly increased responding to increasing water flow velocity for all these species (Figure 2) Figure 3 indicates that a similar relationship is available for E. nuttallii and P. crispuss without a large error (r 0.97, p < 0.01 for E. nuttallii and r 0.99, p < 0.01, respectively). For M. spicatum and C. demersum, the increasing rate of H 2 O 2 with respect to solar radiation was slightly lower, decreasing the effect of the solar radiation. Therefore, a different equation was derived for M. spicatum, and C. demersum (r 0.98, p < 0.01 for M. spicatum, and r 0.89, p < 0.01 for C. demersum, respectively).

The Threshold H 2 O 2 Concentration for Growth Deterioration
Chl-a concentrations and SGR as functions of H 2 O 2 concentrations at different flow velocities, water temperatures, and light intensities of E. densa are shown in Figure 4A.

Simulated Results
The Comparison With the Observed Data Figure 5 shows the comparison between the observed H 2 O 2 concentration and simulated H 2 O 2 concentration for experimental and observed results. Satisfactory agreement between the simulated and observed values were found in the simulation (r 0.798, p < 0.01 for E. densa, r 0.700, r < 0.05 for E. nuttallii, r 0.919, p < 0.01 for M. spicatum, r 0.976, p < 0.01 for C. demersum, and r 0.974, p < 0.05 P. spicatum).

The Depth-Wise Distribution of H 2 O 2 Concentration of Different Species
The H 2 O 2 concentration of E. densa was simulated for channel slopes of 1/300 at 10, 20, and 30°C, which were close to the condition of the observed reaches of the Eno and the Saba rivers in March, May/June, and October as well as August to September, respectively. Figure 6 shows the simulated results with respect to the depth, observed H 2 O 2, and macrophyte biomass. The threshold H 2 O 2 concentration was assumed as 16 µmol/gFW. The H 2 O 2 concentration was high at the water surface and gradually decreased. With deeper depth, increasing velocity increases the H 2 O 2 concentration. The decreasing or increasing trend with respect to depth depends on the combination of these two factors. The H 2 O 2 concentration of the stagnant water is lower than the sloped channel, as H 2 O 2 generated by the velocity is zero. In the case of E. densa, the H 2 O 2 concentration is higher with lower temperature, and FIGURE 3 | Additional H 2 O 2 concentration of light-exposed tissues with respect to dark-adapted ones of different rivers and experiments as a function of the light intensity for different species.  Takahashi and Asaeda (2014), and Kunii (1984), respectively. The fractions in the parentheses are channel slopes. Frontiers in Environmental Science | www.frontiersin.org January 2021 | Volume 8 | Article 516301 (Asaeda et al., 2020;Elsheery et al., 2020a;Elsheery et al., 2020b), and it is highly correlated with the intensity of a single environmental factor Parveen et al., 2017;Chalanika De Silva and Asaeda, 2018). The present study elucidates that under a combination of different environmental factors, the total H 2 O 2 concentration is provided as the sum of H 2 O 2 generated by individual factors and the amount generated by metabolism (Apel and Hirt, 2004). In addition, the relationship between H 2 O 2 concentration and the intensity of each environmental factor does not vary much between seasons and phenological stages of the plant (Asaeda et al., 2020). Therefore, the H 2 O 2 concentration is considered an indicator of the degree of the unpreferable condition. The Chl-a concentration and the growth parameter decreased with increasing intensity of the total H 2 O 2 concentration (Coleman et al., 1989;French and Moore, 2003;Boustany et al., 2010). Interestingly, when the tissue H 2 O 2 concentration exceeded 16-20 µmol/gFW, the plants became brownish and deteriorated. Therefore, the environmental conditions reflected by H 2 O 2 concentrations below this threshold allows macrophytes to form a large and healthy colony. This system can be applied to elucidate the growth area of macrophyte species, by formulating the H 2 O 2 concentrations and abiotic conditions in the environment.

Environmental Conditions Influencing Macrophyte Colonization in Japanese Rivers
In Japanese rivers, the water quality is relatively good and there is no salinity in the midstream (Luo et al., 2011). Organic matter accumulates on the bottom in stagnant zones, which creates an anoxic zone in the sediment layer. There are such areas in the lowland zones; however, anoxia of the bottom sediment contributed only ∼3 µmol/gFW of H 2 O 2 (Parveen et al., 2017). Chalanika De Silva and Asaeda (2018) showed that the H 2 O 2 concentration differs between mono-and mixed-cultures of species in stagnant water, indicating the effect of species competition. However, the difference was only ∼2 µmol/gFW. In field sampling, Japanese native species, P. crispuss and C. demersum, were often found in thick E. densa colonies due to the reduction of flow velocity inside the colony (20 cm/s inside compared to 50 cm/s outside, according to our own observation). This indicates that the increment of H 2 O 2 concentration due to competition is less than the reduction of velocity-induced H 2 O 2 (∼3 µmol/gFW). In this study, the H 2 O 2 concentrations attributed to water temperature, high solar radiation, and high flow velocities of natural conditions are ∼10, ∼10, and ∼5 µmol/gFW, respectively. Therefore, water temperature, solar radiation and flow velocity are the major dominant environmental factors determining the colonization patterns of macrophytes in the midstream of Japanese rivers.  (Asada, 2006 In rivers where E. densa colonized overwhelmingly, water temperature decreases below 8°C in winter. However, even at that time, H 2 O 2 concentration remains below the threshold value at around 1 m deep in stagnant water. Small patchy colonies were found in the upstream of weirs. With increasing temperatures in spring, they started to grow and form a large summer colony, expanding to a shallow zone in the downstream. The channels were originally covered with gravel bed; the bed morphology easily changed under high flow and pools disappeared. However, river rehabilitation for the flood control has been intensively conducted in the last five decades. The shallow zone of the channels was excavated to deepen the channel, and E. densa can now colonize with low H 2 O 2 concentration. Weirs were constructed frequently, which created deep stagnant water in the upstream. Thus, the H 2 O 2 concentration of E. densa due to velocity and solar radiation may decrease. Gravel mining was conducted, substantially reducing the amount of gravel in the river channel, and there is no longer any sediment transport even at flood time . Thus, the modified river morphology does not change even during floods. The artificially created deep zone became a trigger for the overproduction of E. densa in the river channel. In the last three decades, river water temperature has significantly increased due to global warming at approximately 0.1°C/ year, often reaching 30°C, particularly in western Japanese rivers (Ministry of Environment, 2013). It seems difficult for Japanese native species and E. nuttallii to grow in these rivers. Particularly, E. densa and E. nuttallii are closely related species and came to Japan at nearly the same time. However, the overwhelming colonization of E. densa and the limitations of E. nuttallii seem to be attributed to the different temperature traits of these species. Local people emphasize the recent reduction of flow rate (personal communication). During the day, in addition to high solar radiation, water temperature is approximately 2°C higher than at night, particularly under a low summer flow rate. Thus, during this time, both solar radiation and temperature increases the H 2 O 2 concentration for Japanese native species. However, their effects are reciprocal and do not affect E. densa very much. It is likely another reason for the overwhelming presence of E. densa.

CONCLUSION
Under unpreferable environmental conditions, H 2 O 2 concentrations increase in plant tissues and reflect the macrophyte condition fairly accurately. Potentially, this could be a good indicator of submerged macrophyte colonization. This approach will save time by not requiring casual observations and biomass monitoring of macrophytes in ecosystem monitoring. The experimental and field observations indicated a clear positive relationship between the level of unpreferable conditions and H 2 O 2 concentrations, regardless of abiotic factors. The total H 2 O 2 concentration is provided by the sum of H 2 O 2 generated by each environmental factor, and <16-20 µmol/gFW is required for colonization. The relationships of H 2 O 2 concentrations and the contribution of each abiotic factor were obtained for invasive species (E. densa and E. nuttallii) and three major Japanese native species (M. spicatum, C. demersum, and P. crispuss). The system was applied to develop a mathematical model to simulate the colonization area of these species. The tissue H 2 O 2 concentration decreases with increasing temperature for E. densa and increases for other species, including native species. Therefore, native species grow intensively in spring; however, they often deteriorate in summer. For E. densa, on the other hand, H 2 O 2 concentration decreases with high water temperatures in summer, allowing intensive growth. High solar radiation increases the H 2 O 2 concentration, deteriorating the plant. Although the H 2 O 2 concentration of E. densa increases with low water temperatures in winter, it can survive in deep water with low H 2 O 2 concentration due to solar radiation. Currently, river rehabilitation has created a deep zone in the channel, which has supported the growth and spread of E. densa.

DATA AVAILABILITY STATEMENT
The datasets generated for this study are available on request to the corresponding author.

AUTHOR CONTRIBUTIONS
TA: contributed the conceptualization and field work, and wrote the manuscript together with other members; MR: contributed to field sampling, laboratory and data analyses, and helped write the manuscript; JS: reviewed and commented on the manuscript.

FUNDING
This work was financially supported by the Grant-in-Aid for Scientific Research (B) (19H02245) and Fund for the Promotion of Joint International Research (18KK0116) of Japan Society for the Promotion of Science (JSPS).