Status Reports
Metal exposure profiles at metal-contaminated sites in rivers across Japan
2025 Volume 5 Pages 35-39
Details
2025 Volume 5 Pages 35-39
Understanding realistic exposure profiles of mixtures is crucial for effectively predicting the ecological risks and effects of metal mixtures in natural environments on a large scale (e.g., at a country level). In this study, we aimed to compile available measurement data for metals relevant to ecological risk assessment in rivers across Japan, identify metals of particular concern based on their relative ecological risks, and derive realistic exposure profiles of these metals based on the compiled data. We focused on six metals of concern (Ni, Cu, Zn, Pb, Cd, and Al) selected by comparing available measurement data and 10% inhibition concentrations for the daphnid Ceriodaphnia dubia. We then compiled measurement data on these metal concentrations, ensuring sufficiently low detection/quantification limits relevant to ecological risk assessments, from a total of 531 riverine sites. At 194 metal-contaminated sites, concentrations generally increased in the following order: Cd (median: 0.013 μg/L), Pb (0.072 μg/L)<Ni (0.45 μg/L)<Cu (1.2 μg/L)<Zn (9.1 μg/L), Al (22 μg/L). Using hierarchical cluster analysis, we classified the metal-contaminated sites into three groups (Group 1: 56 sites; Group 2: 104 sites; and Group 3: 34 sites). Groups 1 and 3 were characterized by higher concentrations of Cd and Ni, respectively, compared to Group 2. Further compilation and accumulation of measurement data, particularly in small rivers (e.g., tributaries of major rivers), are required to more accurately assess contamination levels and ecological risks from metals in rivers nationwide.
In aquatic environments such as streams and rivers, contamination by trace metals such as copper and zinc is a long-standing environmental issue (Luoma and Rainbow, 2008; Namba et al., 2020). The risk assessment approaches for many individual metals, such as copper, zinc, and nickel, have seen considerable advancements, particularly with the development of biotic ligand models (Adams et al., 2020). However, predicting ecological impacts of metal mixtures remains a challenging task (Farley et al., 2015), although a general risk assessment framework has been proposed to address this issue (Nys et al., 2018). To effectively understand and predict ecological risks and effects of metal mixtures in natural environments on a broad scale (e.g., a country level), it is crucial to understand the mixture exposure profiles that are likely to occur (Mebane et al., 2017).
In Japan, there has been no comprehensive attempt to compile exposure data or evaluate exposure profiles for multiple elements, including metals, on a nationwide scale. There are three national-level databases that compile results of water quality monitoring in aquatic environments throughout Japan: the Comprehensive Information Website for Water Environment managed by the Ministry of the Environment (https://waterpub.env.go.jp/water-pub/mizu-site/), the Water Information System managed by the Ministry of Land, Infrastructure, Transport and Tourism (http://www1.river.go.jp/), and the Database of Water Quality of Aqueduct managed by the Japan Water Works Association (http://www.jwwa.or.jp/mizu/). However, the reported quantification limits for metals in these databases are not sufficiently low to be used for ecological risk assessments, and measurements at a majority of monitoring sites have either not been performed or fall below the reported limits. For instance, the reported quantification limits for Cu, Cd, and Pb are generally higher than 10, 0.3, and 1 μg/L, respectively, which are close to or approximately an order of magnitude higher than the U.S. Environmental Protection Agency (EPA) hardness-adjusted water quality criteria for aquatic life (e.g., 2.3, 0.21, and 0.42 μg/L at water hardness of 30 mg/L; U. S. EPA (2016); U.S. EPA (2002)).
The objectives of this study were thus threefold: (1) to compile available measurement data of metals relevant to ecological risk assessment in rivers across Japan, (2) to identify metals of particular concern based on their relative ecological risks, and (3) to elucidate and categorize the realistic exposure profiles (i.e., compositions and concentration levels) of these metals based on the compiled data. The outcomes of this study provide valuable insights into realistic concentration levels and ratios, which are critical for assessing the ecological risks and effects of exposure to metal mixtures in Japanese rivers.
We first selected metals of concern by comparing measurement data for as many metals as possible in rivers across Japan with toxicity test results performed under the same conditions. As the measurement dataset, we used results of dissolved metal concentrations surveyed at 450 sampling sites across 45 major rivers (i.e., 10 sites per river) in Japan during 2002–2006 (Uchida et al., 2007). For this comparison, using water quality benchmarks, such as water quality criteria and standards, as reference values for toxicity might be desirable; however, the number of metals for which these benchmarks are available is limited. Therefore, for the toxicity test data, we used the 10% inhibition concentration (IC10) values for 50 metals, determined by chronic toxicity tests with Ceriodaphnia dubia (Okamoto et al., 2021). The toxicity tests were consistently conducted using filtered and sterilized tap water with a water hardness of approximately 70–80 mg/L. Although water quality parameters important for considering metal bioavailability (e.g., pH, water hardness, and dissolved organic matter; Adams et al., 2020) can vary among sampling sites and differ between the toxicity tests and sampling sites, the results of Okamoto et al. (2021) are considered ideal for the preliminary screening of metals of particular concern.
For the 15 metals (Ni, Cu, Zn, Pb, Cd, Al, Mn, Fe, Co, Rb, Sr, Ba, Ti, V, and Cr) for which data were available from both sources, we calculated the ratio of the dissolved concentration of each metal to the corresponding IC10 (hereafter, toxic unit [TU]) as well as the sum of the TUs at each sampling site. By calculating the contributions of each metal to the sum of the TUs at each sampling site and ranking the metals based on the 99th percentiles of these contributions, the metals in descending order were Ni (0.94), Zn (0.64), Cd (0.61), Al (0.31), Cu (0.14), Fe (0.13), and Co (0.09) at sites where the sum of the TUs ≥1 (i.e., metal-contaminated sites) (see Fig. 1 for more details). Although the 99th percentiles of the contributions for Cu, Fe, and Co were similar, Cu was included in the following analysis because its IC10 of 12 μg/L (Okamoto et al., 2021) was somewhat higher than the U.S. EPA hardness-adjusted water quality criterion (e.g., 6.6 μg/L at a water hardness of 70 mg/L) and because ecological risks associated with Cu have been a concern in Japan (Han et al., 2016). Similarly, although the 99th percentile of the contribution for Pb was low (0.002), Pb was included in the following analysis because the IC10 for Pb reported by Okamoto et al. (2021) was 67 μg/L, which is more than an order of magnitude higher than the U.S. EPA hardness-adjusted water quality criterion of 1.7 μg/L at a water hardness of 70 mg/L. It should be noted that the IC10 of 0.37 μg/L for Ni is two orders of magnitude lower than the U.S. EPA water quality criterion of 38 μg/L at a water hardness of 70 mg/L. Although the direct use of IC10 values (i.e., without bioavailability correction) requires some caution, IC10 values are likely acceptable for screening metals of ecological concern. The following data collection and analysis were conducted for a total of six metals (Ni, Cu, Zn, Pb, Cd, and Al).
In order to compile as much available monitoring data as possible for these metals in rivers across Japan, we reviewed the Comprehensive Information Website for Water Environment and the Water Information System as well as existing literature covering over 1,000 sampling sites in total (excluding the two databases). The literature included peer reviewed papers (e.g., Han et al., 2013) and Japanese project reports. Based on examination of whether the measured concentrations (preferably, dissolved concentrations) of all six metals were relatively low, we selected three datasets: the large-scale measurement data of Uchida et al. (2007); measurements from metal-contaminated rivers affected by legacy mines as well as nearby reference rivers (Iwasaki et al., 2023: 26 sites; Iwasaki et al., unpublished data: 12 sites); and measurements from rivers to assess the ecological impacts of Ni (Takeshita et al., 2019: 50 sites). All these studies reported dissolved metal concentrations by filtering water samples through filters with a pore size of 0.45 μm and analyzing them using inductively coupled plasma mass spectrometers (ICP-MS) or inductively coupled plasma optical emission spectrometers (ICP-OES). We analyzed data from a total of 531 sampling sites (excluding seven estuarine sites from Uchida et al., 2007; see Fig. 2a). All the complied data are available in the supporting information (Table S1).
To obtain realistic metal exposure profiles at the metal-contaminated sites where ecological risks due to the five metals were of concern, we first calculated the sum of TUs by dividing the measured dissolved metal concentrations by the U.S. EPA hardness-adjusted water quality criteria for aquatic life (U.S. EPA, 2002; U.S. EPA, 2016). The water quality criteria for a water hardness of 20 mg/L were used as conservative “safe” concentrations (Cu: 2.3 μg/L, Zn: 30.2 μg/L, Pb: 0.42 μg/L, Cd: 0.21 μg/L, Ni: 13.3 μg/L, and Al: 220 μg/L) because water hardness at many of the sampling sites was as low as 20 mg/L (Table S1). We used U.S. EPA water quality criteria because water quality standards for aquatic life are available in Japan for only total Zn (30 μg/L). Any concentration below the quantification or detection limit was operationally replaced with half the corresponding limit for operational purposes (see Table S1 for specific values). We then selected 194 sites where the sum of TUs ≥1 as metal-contaminated sites (Fig. 2b) to elucidate the exposure profiles at sites where the ecological risks could be of concern. A sum of TUs >1 was used as the conservative criterion for selection, not to accurately assess risk. Together with empirical evidence that the exceedance of the sum of TUs beyond 1 does not necessarily lead to population- or community-level effects on aquatic organisms, such as fish and macroinvertebrates (Namba et al., 2021; Iwasaki et al., 2023), it is also important to note that such impacts may not always be observed at the selected metal-contaminated sites.
Based on concentrations of the six metals, the 194 metal-contaminated sites were classified using hierarchical cluster analysis with Ward’s method based on Euclidean distances (Ward, 1963). To ensure classification based on metal concentration compositions rather than absolute concentration values at individual sites, metal concentrations were first log10-transformed and later standardized for each site before performing the hierarchical cluster analysis.
Based on the results of the hierarchical cluster analysis (see Fig. S1 for the dendrogram), we classified the metal-contaminated sites into three groups (Group 1, 56 sites; Group 2, 104 sites; and Group 3, 34 sites). Although there is no absolute criterion for determining the number of clusters in hierarchical cluster analysis, using more than four groups resulted in clusters with fewer than 20 sites. We therefore used a three-group classification.
Despite the observed large variations in metal concentrations, the concentration levels at the 194 metal-contaminated sites generally increased in the following order: Cd (median: 0.013 μg/L)≤Pb (0.072 μg/L)<Ni (0.45 μg/L)<Cu (1.2 μg/L)<Zn (9.1 μg/L)≤Al (22 μg/L) (Fig. 3). Compared with Group 2, where the overall metal concentration levels were similar to those at all metal-contaminated sites, the Cd concentrations in Group 1 were notably higher (median: 0.12 μg/L in Group 1 vs. 0.007 μg/L in Group 2). Similarly, the Ni concentrations in Group 3 were relatively high (median: 17 μg/L in Group 3 vs. 0.44 μg/L in Group 2). Summary information on metal concentrations (i.e., 10th to 90th percentile values) for Groups 1–3 is available in Table S2. In Groups 1–3, the median (maximum) values of the sum of TUs calculated based on the U.S. EPA water quality criteria were 1.9 (34), 1.3 (9.0), and 3.6 (23), respectively.
Each of Groups 1–3 included sampling sites from at least two of the three different original datasets (Fig. 2, Table S1), and it was difficult to further interpret the geographic distribution shown in Fig. 2b. However, sampling sites from legacy mine surveys (Iwasaki et al., 2023) were more frequently included in Group 1, whereas those from the nickel survey (Takeshita et al., 2019) were more frequently included in Group 3. Importantly, similar three groups were derived even when sampling sites from the legacy mine and nickel surveys (Takeshita et al., 2019; Iwasaki et al., 2023) were excluded from the hierarchical cluster analysis (Fig. S2), although the Ni and Zn concentration levels in Group 3 of Fig. S2 were markedly reduced.
By compiling measurement data on metal concentrations relevant to ecological risk assessments, we were able to derive realistic compositions and concentration levels of six metals (Ni, Cu, Zn, Pb, Cd, and Al) at metal-contaminated sites in rivers across Japan (see Table S2 for the percentiles of metal concentrations in individual groups). This information would be useful in choosing metal concentration levels relevant to those found in Japanese rivers for metal mixture toxicity tests, enabling a better assessment of their actual ecological risks. However, it should be noted that the two field studies included in our dataset focused on the impacts of specific metals such as Ni (Takeshita et al., 2019) or Cu, Zn, Cd, and Pb (Iwasaki et al., 2023). In addition, Uchida et al. (2007) measured metal concentrations in the mainstreams of major rivers across Japan but did not make measurements in tributary streams. Because levels of contamination in small streams and rivers are more likely to be high owing to their limited capacity for dilution (Büttner et al., 2022; Iwasaki et al., 2022), further compilation and accumulation of measurement data are therefore required to more accurately categorize and assess contamination levels and ecological risks from metals in rivers across Japan. To this end, it is essential to encourage the open accessibility of measurement data, along with their geographical coordinates, in easy-to-use formats in publications such as peer-reviewed papers, particularly those written in Japanese. Furthermore, the bioavailability of metals was not considered when selecting the metal-contaminated sites in this study. While important water quality variables such as Ca, Mg, and dissolved organic carbon (DOC) were collected as complementary data (Table S1), DOC was not measured by Uchida et al. (2007). Because the influence of DOC on bioavailability can be significant for some metals (OECD, 2017), compiling measurement data on such water quality variables in addition to metals, as well as developing predictive models when measurements are unavailable, will be essential for more accurate metal risk assessments.
This paper does not necessarily reflect the policies or views of any government agency. This study was funded by the Environment Research and Technology Development Fund (JPMEERF20225005) of the Environmental Restoration and Conservation Agency of Japan provided by the Ministry of the Environment of Japan. We are grateful to Dr. Keiko Tagami for help with data collection.
During the preparation of this paper, the authors used ChatGPT to improve readability and language. After using this tool, the authors reviewed and edited the content as needed and take full responsibility for the content of the publication.
The authors declare no conflicts of interest.
Fig. S1, Dendrogram resulting from the hierarchical cluster analysis of 194 metal-contaminated sites; Fig. S2, Distributions of concentration of six metals (Ni, Cu, Zn, Pb, Cd, and Al) across all metal-contaminated sites in Uchida et al. (2007), as well as in Groups 1, 2, and 3; Table S1, Compiled data including metal concentrations at river sites in Japan; Table S2, 10th to 90th percentile values of metal concentrations in Groups 1–3.
This material is available on the Website at https://doi.org/10.5985/emcr.20240039.
