2008 Progress Report: Exposure Assessment of Children and Metals in Mining Waste: Composition, Environmental Transport, and Exposure Patterns

EPA Grant Number: R831725C002
Subproject: this is subproject number 002 , established and managed by the Center Director under grant R831725
(EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).

Center: Harvard Center for Children’s Environmental Health and Disease Prevention Research
Center Director: Hu, Howard
Title: Exposure Assessment of Children and Metals in Mining Waste: Composition, Environmental Transport, and Exposure Patterns
Investigators: Shine, James P. , Spengler, John D.
Institution: Harvard T.H. Chan School of Public Health
EPA Project Officer: Callan, Richard
Project Period: June 1, 2004 through May 31, 2009 (Extended to May 31, 2011)
Project Period Covered by this Report: June 1, 2008 through May 31,2009
RFA: Centers for Children's Environmental Health and Disease Prevention Research (2003) RFA Text |  Recipients Lists
Research Category: Children's Health , Health Effects , Health


Following the severe flooding event that occurred in Miami in July 2007, we have conducted additional sample collection and analysis to characterize the role of this major disturbance in distributing metals from Tar Creek and Neosho River sediments into the floodplain, including household yard soils and public use areas. We have developed a geochemical fingerprinting approach to characterize the mixing of flood deposit material from metal input sources to the Tar Creek floodplains and residential yards. We have expanded upon our aims to assess the relative spatial and temporal distribution of metals from upstream Tar Creek at the mine waste piles to downstream Neosho River approaching the Grand Lake. We collected sediment cores in Tar Creek and the Neosho River at several locations between the confluence with Tar Creek and Grand Lake. We also compared the relative mobility of Pb and Zn downstream of the source to determine which metal has the greatest potential for impacting areas at large distances from the source. Furthermore, in collaboration with Project 3, we have collectively been assessing the oral in vivo bioavailability and in vitro bioaccessibility of pure-phase Zn minerals and fine grain chat samples.

Progress Summary:

i. Geochemical fingerprint using principal component analysis
Following a major flooding event in Miami in early July 2007, the worst flood in nearly 50 years, there were community concerns about metal exposure due to the deposition of sediments from Tar Creek and
the Neosho River throughout the town (Figure B2.1).

Figure B2.1. Left: Environmental Protection Agency (EPA) estimation of the extent of the July 2007 flood in the town of Miami, OK. Top Right: Example of Neosho River flood deposit on the floodplains. Bottom Right: Example of Tar Creek flood deposit.
The flood event also posed an opportunity for us to develop a geochemical fingerprint for different metal sources using an exploratory statistical method, principal component analysis, to examine the mixing that occurred during the flood and the resulting concentration patterns.
Due to the hydrodynamics of the flood, the Neosho River (with no known inputs of metal contamination) flowed backwards into Tar Creek, impeding normal flow. Tar Creek also flooded, transporting mine waste and mining impacted sediments downstream. By using PCA, we determined that there are three distinct sources of metal inputs on the floodplains: Neosho River sediment, mining waste (chat), and hydrous ferric oxides (HFOs) found in upstream Tar Creek (Figure B2.2).

Figure B2.2. Principal component analysis scores for mine waste, Neosho River floodplain samples and sediment cores, Tar Creek floodplains, hydrous ferric oxides (HFOs), and mine waste samples separated into coarse and fine grain size fractions.
Principal component 1 (PC1) separates mine waste from Neosho River sediment and soils. Principal component 2 (PC2) separates coarse grain mine waste (> 0.25 mm) from fine grain mine waste (< 0.25 mm). From a previous study, we have determined that fine grain mine waste has higher concentrations of Pb, Zn, and Cd than coarse mine waste. Fine grain mine waste has a slower settling velocity than coarse mine waste and therefore is more likely to travel farther from the source. Thus, it is important to distinguish between these two grain size categories. Principal component 2 also separates high trace metal concentration sources (HFOs and fine grain mine waste) from lower trace metal concentration sources (coarse mine waste and Neosho River sediment). The post-flood floodplain soils of Tar Creek are positioned between the three metal input sources demonstrating that the floodplain soils have components of all three sources. When the original vectors are displayed relative to PC1 and PC2, we can determine which samples are enriched in the important elements relative to the other samples (Figure B2.3).

Figure B2.3. PCA of three distinct sources, post-flood floodplain samples, and original vectors including Pb, S, Zn, Cd, Fe, Al, and Si.
The elements Pb, Zn, Cd, and S are all highly correlated suggesting that they co-vary in the floodplain samples. The fine grain mine waste and HFO samples tend to be enriched in these metals and sulfur as expected given our previous findings on the elevated concentrations in fine grain mine waste and HFOs. The ore minerals mined at this site contain sulfide as their main anion suggesting that there is a primary ore mineral fingerprint in the watershed. The hydrous ferric oxides are enriched in iron (Fe) relative to the other samples. While coarse mine waste is enriched in silica (Si). This corresponds with our assumptions that chert (a silica rich mineral) is found in large quantities in coarse mine waste. Through our principal component analysis, we determined that the flood deposited fine chat, coarse chat, Neosho River and Tar Creek sediment and hydrous ferric oxides on the floodplains. Using this analysis, we will determine quantitatively which floodplain locations were primarily impacted by Neosho River or Tar Creek inputs.
ii. Temporal metal loading to Tar Creek
Mine waste pile runoff and seepage from underground mines have transported metals into the Tar Creek watershed since the mines became active. Mine seepage became more of a problem when the mines were abandoned in the 1970s. A temporal record of mining activity can be preserved in the creek sediment by examining metal concentration levels. The temporal record is also useful to determine whether or not the system has reached a steady state. We find that the concentrations of Pb, Zn and Fe all peak at a similar depth in a sediment core taken in Tar Creek upstream of the confluence with the Neosho River (Figure B2.4).

Figure B2.4. Concentration of Zn, Cd, and Pb in sediment as functions of depth from the sediment surface. One way of examining the source of metal contamination is to determine how the metals co- vary with depth. Both Pb and Fe were found to co-vary significantly with Zn along the whole depth of the core (p=0.001, p=0.02). The co-variation of Zn and Fe suggest that acid mine drainage is a significant source of metal loading at this location.
iii. Spatial metal loading downstream of source
Zinc is known to be a more labile metal than Pb and this was also demonstrated in mine waste samples at this site in a previous study we conducted using sequential extractions. Thus, we hypothesized that we would see greater metal loading of Zn than Pb downstream from the source. Comparing the ratio of Pb to Zn concentrations in a variety of media (sediment, soil, HFOs, mine waste) at increasing distances from the source, we see that the Pb to Zn ratio has an inverse relationship with the distance from source (Figure
Figure B2.5. Relationship of Pb to Zn concentration ratios in different media at different distances from source. Type of sample is indicated by shape. Location in the Tar Creek watershed is indicated by color.

Mine waste has the greatest Pb to Zn ratio (0.15 – 0.8) while the downstream Neosho River sediment core samples are among the samples with the lowest Pb to Zn ratios (0.06 – 0.02). These data agree with our previous findings that Zn is more labile than Pb.
iv. Comparisons of in vivo bioavailability and in vitro bioaccessibility of Zn in pure phase Zn minerals and mine waste
As part of our ongoing collaboration with Project 3, we are assessing the bioaccessibility and
bioavailability of Zn present in various mineralogical forms. We obtained samples of 5 zinc minerals from the Mineralogical Collection at Harvard’s Museum of Natural History. Two of these minerals, sphalerite (ZnS) and hemimorphite (Zn silicate), were primary minerals present at the Tar Creek site. The mineralogical samples were pulverized and sieved to <37 µm. These fine-grained samples, in addition to a <37 µm chat sample, were neutron activated (NA) at MIT and aliquots were used in both in vivo bioavailability studies in rats and in vitro bioaccessibility tests using the Simple Bioaccessibility
Extraction Test (SBET), a simulated gastric fluid extraction.
Bioaccessibility was measured in two sets of extractions. In one set, non-NA mineral and chat samples were run at a 1:100 (g:mL) ratio and analyzed by ICP-MS. In a second set, extractions of 65Zn from NA samples were run at 1:1000 and 1:5000 (g:mL) ratios, in the presence and absence of rat food. As predicted, our results showed that sphalerite was the least bioaccessible form of Zn (Table B2.1), which was consistent with previous studies showing that metal sulfides have low bioavailability. Similarly, in vivo Zn absorption studies in rats showed that sphalerite had the lowest bioavailability of all the minerals tested. We also expected that smithsonite (ZnCO3) would have high bioaccessibility, since metal carbonates are thought to be highly bioavailable. While the bioaccessibility of smithsonite was substantially higher than that of sphalerite, the other three minerals all had even higher bioaccessibility than smithsonite. There were no differences between the 1:1000 and 1:5000 extractions, suggesting that the results are not dependent on the solid:liquid ratio used, and the presence of food did not decrease the Zn bioaccessibility in most samples, although for smithsonite and sphalerite, bioaccessibility in the presence of food decreased 30-36%. For all samples, the bioaccessibility as measured using NA samples was higher than those measured using ICP-MS. We will determine whether these differences are due to the different solid:liquid ratio used, to the NA process and resulting pool of potentially leachable metals, or to the methods of determining the total amount of Zn present. For instance, the “total” Zn in the NA samples was determined directly by measuring 65Zn in the samples prior to extraction, whereas the “total” Zn in the non-NA samples was based on theoretical values, which could underestimate bioaccessibility if there are impurities in the minerals.
We will compare our results with Project 3’s in vivo bioavailability studies in order to determine if the SBET provides a reasonable prediction of Zn bioavailability in rats, which would be useful information given the expense and time required to perform animal studies. In the future, we also plan to run sequential extractions on the same samples in order to compare their results with SBET results, as well as to verify their specificity for certain mineral phases. Furthermore, we would like to extend this work to other metals for which there also have not been extensive comparisons of in vivo bioavailability and in vitro bioaccessibility.

Table B2.1. Bioaccessibility of Zn in 5 pure phase mineralogical samples, as measured by the Simple
Bioaccessibility Extraction Test.
v. Analysis of hydrous ferric oxides preserved on tree bark
In-situ FP-XRF analysis of hydrous ferric oxides deposited on steam-side tree bark during the flood
events further contributes to our understanding of HFO’s role in metal fate and transport (Figure B2 6a).
Field based measurements were made at three sites along Tar Creek (Douthat Bridge, 69th Street Bridge, and 22nd Street Bridge). At each site, two or three trees of the same species were selected for analysis along a transect away from the streambed. 90-second measurements were obtained at 20 cm intervals from 20 cm to 160 cm above the ground. Bark samples from 20 cm and 80 cm were also collected for additional lab based characterization. Data from the in-situ analysis confirm a strong correlation between Pb, Zn, and Fe consistent with HFO deposition. Metal correlations with Fe are most apparent at the 69th Street Bridge though the correlations are partly defined by the high concentrations recorded at 20 cm (Figure B2 6b). While Pb and Zn correlate closely at all sites, [Pb] decreases relative to [Zn] farther downstream (Figure B2 6c). These results suggest that Pb is selectively removed from the creek (and adsorbed to HFO’s) more efficiently than Zn.

Figure B2.6 a) HFO coatings on trees near 22nd Street Bridge; b) Correlation between [Pb], [Zn], and [Fe] from 20 to 180 cm at 69th Street Bridge; c) Pb]/[Zn] at 80cm height at all locations Arrow points in downstream direction.

Journal Articles:

No journal articles submitted with this report: View all 6 publications for this subproject

Supplemental Keywords:

children, Native American, tribal, mixtures, lead, PBPK, community, Superfund, intervention, environmental management, environmental management, international cooperation, Scientific Discipline, Waste, Health, RFA, Risk Assessment, Health Risk Assessment, Children's Health, Hazardous Waste, Biochemistry, Hazardous, epidemiology, neurodevelopmental toxicity, developmental toxicity, fate and transport , children's environmental health, mining wastes, human health risk, mining waste, community-based intervention, metal contamination, metal wastes, biological markers, metals
, RFA, Health, Scientific Discipline, INTERNATIONAL COOPERATION, ENVIRONMENTAL MANAGEMENT, Waste, Health Risk Assessment, Hazardous Waste, Children's Health, Hazardous, Risk Assessment, community-based intervention, epidemiology, fate and transport , developmental toxicity, Human Health Risk Assessment, neurodevelopmental toxicity, children's environmental health, biological markers, mining waste, metal wastes, metals, human health risk, metal contamination

Progress and Final Reports:

Original Abstract
  • 2004 Progress Report
  • 2005 Progress Report
  • 2006
  • 2007
  • 2009
  • Final

  • Main Center Abstract and Reports:

    R831725    Harvard Center for Children’s Environmental Health and Disease Prevention Research

    Subprojects under this Center: (EPA does not fund or establish subprojects; EPA awards and manages the overall grant for this center).
    R831725C001 Metals, Nutrition, and Stress in Child Development
    R831725C002 Exposure Assessment of Children and Metals in Mining Waste: Composition, Environmental Transport, and Exposure Patterns
    R831725C003 Manganese, Iron, Cadmium, and Lead Transport from the Environment to Critical Organs During Gestation and Early Development in a Rat Model
    R831725C004 Metals Neurotoxicity Research Project