Registration Dossier

Environmental fate & pathways

Additional information on environmental fate and behaviour

Currently viewing:

Administrative data

Endpoint:
additional information on environmental fate and behaviour
Type of information:
calculation (if not (Q)SAR)
Remarks:
Migrated phrase: estimated by calculation
Adequacy of study:
key study
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
other: The estimations are based on model that has been developed to assess the the long-term fate and effects of metals in the environment

Data source

Reference
Reference Type:
study report
Title:
Unnamed
Year:
2013

Materials and methods

Principles of method if other than guideline:
The two main objectives of this work are:
1. Simulate copper removal from the water column of a generalized lake environment using a transport and fate model, TICKET-UWM, to see if the rapid removal benchmark of 70% removal of dissolved copper in 28 days is met and assess the extent to which copper deposited to sediment is remobilized to the water column.

2. Assess the predictive capabilities of the TICKET-UWM by testing it against data from several laboratory and field datasets including a lake, a reservoir, a large enclosure in a lake, and laboratory microcosms.
GLP compliance:
no
Type of study / information:
Model calculatations and literature reviews.

Test material

Reference
Name:
Unnamed
Type:
Constituent

Results and discussion

Applicant's summary and conclusion

Conclusions:
Dissolved copper, added to a standard lake model (EUSES) is removed by >70% within 5 days. Sensitivity analysis and validations from 1 mesocosm study with continuous copper dosing at 5 concentrations and 2 lake systems with intermettent copper dosing also demonstrate that copper is rapidly removed, in excess of 70% within 28 days. For one study, with exterme combinations of low settling rate, low suspended solids and low distribution coefficients (much lower then defined in the EUSES model) the removal rats was <70% in 28 days. Such extreme conditions are not considered as relevant to a hazard classification.

It is therefore concluded that for typical conditions in Europe, copper-ions are rapidly removed from the water-column. This conclusions is critical to the chronic environmental classification.

It is therefore concluded that copper ions are rapidly removed from the water column.
Executive summary:

Introduction:

European Union (EU) regulations pertaining to Classification, Labelling, and Packaging (CLP) of chemical substances follow the United Nations Globally Harmonized System (UN GHS). The EU CLP guidance document (ECHA, 2011), being aligned with the GHS, recognizes the importance of environmental transformation of metals to potentially less bioavailable forms. This guidance document includes a provision for demonstrating removal from the water column to assess the “persistence” or lack of degradation of metal ions, responsible for the toxicity of metals and metal compounds. In analogy to organic chemicals, “rapid degradation” for metals requires greater than 70% removal within 28 days. However, unlike organic chemicals, where removal from the water column occurs via degradation, metal removal occurs through changes in speciation. Therefore, in line with the GHS guidance, “rapid degradation” for metals requires one to demonstrate not only rapid loss from the water column, but also limited remobilization potential from sediment. 

The two main objectives of this work are:

1. Simulate copper removal from the water column of a generalized lake environment using a transport and fate model, TICKET-UWM, to see if the rapid removal benchmark of 70% removal of dissolved copper in 28 days is met and assess the extent to which copper deposited to sediment is remobilized to the water column.  

2.     Assess the predictive capabilities of the TICKET-UWM by testing it against data fromseveral laboratory and field datasets includinga lake, a reservoir, a large enclosure in a lake, and laboratory microcosms. 

The Tableau Input Coupled Kinetics Equilibrium Transport Unit World Model for Metals in Lakes (hereafter referred to simply as the TICKET-UWM) (Farley et al., 2007; Farley et al., 2008; Farley et al., 2010) was developed to address the complexities of metal speciation and its influence on the fate and effects of metals in the environment. Processes considered by the model include complexation by aqueous inorganic and organic ligands such as dissolved organic carbon (DOC), adsorption to particulate phases such as particulate organic carbon (POC) and iron/manganese oxides, binding to biological receptors (biotic ligands), dissolution kinetics of metals powders, and cycling of organic matter and sulfide production in lakes.

Generalized Lake Simulations:

The model was used in time-variable mode to simulate copper removal following a single short-term addition to the water column of a generalized lake based upon the EUSES model lake (RIVM, 2004; ECHA, 2010). The initial copper concentrations used for the simulations were set at the pH-specific chronic ERVs (Table 2-1), 0.1 mg/L, and 1 mg/L. The water chemistries for three pH values (6, 7, and 8) were based upon directives in Annex IV of the Guidance on the Application of the CLP Criteria (ECHA, 2011) and Annex 10 of theGlobally Harmonized System of Classification and Labelling of Chemicals(GHS) (United Nations, 2011). 

For the generalized lake removal calculations, copper adsorption onto suspended solids was described by means of two different approaches: 1) using empirical distribution coefficient values (logKD) from the copper risk assessment (Cu RA) document (Heijerick et al., 2005) for the water column and sediment; and 2) using the speciation models within TICKET-UWM to calculate “instantaneous” distribution coefficients based upon water chemistry and the concentration of particulate sorbents (e.g., particulate organic carbon, POC). Based on the description of the rapid removal definition in Annex IV, removal was evaluated by comparing the concentration of dissolved copper at a particular time to the initial concentration.

Using the empirical distribution coefficient, copper removal was rapid: 31% of the copper initially added to the system was removed immediately via equilibrium partitioning onto particles, and the remaining 39% left the water column within 3.3 days. In an alternate, more conservative approach in which adsorbed copper was considered equally bioavailable to dissolved copper,totalcopper was compared to the initial concentration and the rapid removal benchmark was met 4.7 days after copper addition.

Using the speciation calculation approach, model-estimated distribution coefficients at the three pH values were higher than the empirical value from the Cu RA document. As a result, 70% removal of dissolved copper occurred instantaneously via initial partitioning for most test cases. The time required for 70% removal of total copper varied between 1.5 and 3.2 days. Therefore, for a generalized lake environment, copper removal from the water column satisfies the definition for rapid removal of 70% dissolved copper removal in 28 days. 

Various sensitivity analysis, with varying loadings (0.01, 0.1 and 1 mg/L), increased DOC concentration (15 mg/L) and settling velocity were  carried out and further confirmed that copper-ions are rapidly removed (70% within 28 days).

To examine the potential for remobilization of copper from sediments, a series of 1-year simulations were made. These focused on resuspension, diffusion, and burial to/from the sediment layer and their net effect on copper concentrations in the water column. For the base-study, , the baseline EUSES parameters at pH 7 water chemistry at a loading of 35µg Cu/L was used. Sediment bulk and porewater chemistry was specified based on data from the Besser et al. (2010), Flemish waterways (Vangheluwe et al., 2000), and sediment monitoring data from 1995 (Personal communication with M. Vangheluwe, 2010). Rates ofresuspension, diffusion, and burial were set to EUSES model lake values. Remobilization from the sediment was evaluated by examining the water column copper concentration response with and without feedback from the sediment. Simulations were made with an oxic sediment layer as well as with an anoxic sediment layer (with varying concentrations of AVS) and varying resuspension rates (up to 10 times the default EUSES model lake value). For the oxic case, sulfide production and metal sulfide precipitation were not included. Metals can sorb to POC, HFO, and HMO in the sediment and precipitate as carbonates and/or hydroxides. For the anoxic case, metal binding to HFO and HMO was not considered. Metals can sorb to POC and precipitate as sulfides, carbonates, and/or hydroxides. A model run was also made with empirical distribution coefficients from the Cu RA document.    

In the simulations without sediment feedback (i.e., no resuspension or diffusion), water column total and dissolved copper concentrations decreased rapidly and within 20 days of copper additional were more than 4 orders of magnitude below the concentration corresponding to 70% removal. With feedback, the water column copper concentrations leveled off within 50 days of addition as the resuspension and settling fluxes set up a pseudo steady-state in the water column. For the remainder of the simulation time, copper was slowly depleted out of the water column / active sediment layer domain via the effect of burial. In simulations with AVS present, copper in the sediment was precipitated as insoluble copper sulfide solid (CuS or Cu2S). In simulated sediments with AVS present in excess of copper, essentially all sediment copper was present as copper sulfide precipitate. As a result of this strong binding, the sediment logKDgreatly exceeded the water column logKDand the net diffusive flux of copper was directed into the sediment. For all cases considered, the pseudo steady-state total and dissolved copper concentrations were lower than the concentration corresponding to 70% removal. Research (Simpson et al., 1998; Sundelin and Eriksson, 2001) suggests that the potential for copper release from sulfides and other sediment binding phases is limited. This supports the idea that additional metal immobilization capacity afforded by sulfides in sediment will be long-lived. 

This indicates that the potential for copper remobilization from sediment is limited. 

Various sensitivity analysis with respect to loadings (0.01, 0.1 and 1 mg/L), AVS concentrations (0, 0.77 and 9.1 µmol/g), water pH (6-8) and corresponding sediment pH (7-7.5),  hardness (doubled) and sediment solid concentrations (15 and 500 g/l) were carried out and further confirmed that the potential for copper remobilization is limited.

TICKET-UWM Testing with Laboratory and Field Datasets:

The ability of the TICKET-UWM to described copper removal in laboratory and field systems was evaluated using data from

1. Two shallow lakes in the Limousin region of France: Lake Courtille and the Saint Germain les Belles Reservoir.   

2. A mesocosm study using large enclosures in Lake Baldegg (Lucerne, Switzerland).  

3. A microcosm study conducted at theFraunhofer Institute for Molecular Biology and Applied Ecology (IME).

Lake Courtille and the Saint Germain les Belles Reservoir were dosed with copper sulfate (CuSO4.5H2O) to control the algae population and the copper concentrations in the water column were monitored (Van Hullebusch et al., 2002a, 2002b, 2003a, 2003b, 2003c). 

Observed dissolved and total copper removal from the two waterbodies was rapid. For Lake Courtille, 70% removal of dissolved and total copper occurred 15 and 17 days after copper addition, respectively. For the Saint Germain, 70% removal of dissolved and total copper occurred 1.5 and 7 days after copper addition, respectively.

For the model testing, physical and chemical parameters serving as input for the TICKET-UWM were specified based on measurements in Van Hullebusch et al. (2002a, b; 2003a, b, c). TICKET-UWM input parameters not directly measured in the studies, such as settling velocity and burial rate, were set to regional values from the EUSES model lake. Copper partitioning to suspended solids was described using the two approaches discussed above as well as using an observed logKDbased upon data from the actual sites. While the settling velocity was initially set at the EUSES model lake value of 2.5 m/d, it was adjusted as necessary to optimize the model fit to the measured data.

Key findings from model testing with theLake Courtille and the Saint Germain les Belles Reservoir datasets include the following: · The Cu RA logKDwas more consistent with observed values than logKDvalues resulting from TICKET-UWM speciation calculations. These tended to overestimate the extent to which copper binds to particles;

· Predicted copper removal rates with the EUSES settling velocity value of 2.5 m/d were notably higher than observed; and · Reasonable model fits to the data were achieved with settling velocities between 0.68 and 1.02 m/d. These values are within thesettling velocity ranges for organic particles indicated by Burns and Rosa (1980) and O’Connor (1988).

The MELIMEX (MEtal LIMnological EXperiment) study was undertaken to study the effects of increased metal loading (relative to natural levels) on lacustrine biota and investigate the speciation, distribution and fate of added metals (Gächter, 1979). The experiment was conducted in Lake Baldegg (Lucerne, Switzerland) using large enclosures called limno-corrals (12 meters in diameter and 10 meters deep) to isolate portions of the lake water column and sediment for study. Copper was added continuously the limno-corrals and periodically the water column was sampled at several depths. The samples were analyzed for several water quality parameters including total and dissolved copper.

This study involved continuous copper addition to enclosures and the associated response was an increase in copper in the water column. The performance of the TICKET-UWM was evaluated based upon its ability to reproduce the copper increase in the water column. To address the rapid removal benchmark, additional TICKET-UWM simulations (referred to as post-loading simulations) were made in the absence of copper loading. The initial copper concentration for these simulations was the final model-predicted concentrations from the continuous load runs.

For model testing, physical and chemical parameters serving as input for the TICKET-UWM were specified based on measurements from the study itself. TICKET-UWM input parameters not directly measured in the studies were set to regional values from the EUSES model lake. Baccini et al (1979a) estimated a logKDvalue and settling velocity of 4.12 and 0.2 m/d, respectively, for the limno-corrals.         

Key findings from model testing with theMELIMEX study dataset include the following:

· Using the above logKDand settling velocity, a reasonable fit to the observed copper data was achieved;

· As was the case for Lake Courtille and the Saint Germain les Belles Reservoir, the experimental data were well-described using settling velocities markedly lower than the EUSES default value of 2.5 m/d;

· TICKET-UWM speciation calculations tended to overestimate the logKD; and

· For many of the post-loading simulations, the rapid removal benchmark was not met.

However, because of the low settling velocity, low distribution coefficient, and low suspended solids concentration (relative to the EUSES value), this field test case is more representative of a “worst-case” scenario for copper removal from the water column and therefore not necessarily an appropriate field test case to compare to the rapid removal definition.

A microcosm study was undertaken at the Fraunhofer Institute for Molecular Biology and Applied Ecology (IME) to study the effects of continuous copper exposure on aquatic organisms (Schäfers, 2003). Microcosms were filled with water and sediment collected from a manmade pond near Schmallenberg-Oberkirchen, Germany and dosed to achieve six different nominal concentrations: 5, 10, 20, 40, 80, and 160µg/L. Model testing was performed using dissolved copper data sampled 1, 24, and 48 hours after copper addition and total copper data sampled 24 hours after addition.  

Based on half-lives calculated from the measured dissolved copper data,the time required for 70% copper removal (relative to the initial nominal copper concentration), is between 2.4 and 7.6 days. This is consistent with the definition for rapid removal.

For the model testing, physical and chemical parameters serving as input for the TICKET-UWM were specified based on measurements from the study itself. TICKET-UWM input parameters not directly measured in the studies were set to regional values from the EUSES model lake. Initial copper was specified in the model by 1) setting the initial total copper concentration to the nominal copper concentration for each microcosm, and 2) setting the initial total copper to produce the initial dissolved concentration extrapolated from measured values at1, 24, and 48 hours after copper addition.

Key findings from model testing with theIME microcosm study dataset include the following:

· By optimizing settling velocities in the simulations for the 80 and 160 µg/L microcosms, general agreement between model-predicted and experimental dissolved copper removal rates was observed in each of the examined partitioning scenarios and initial copper specification approaches. 

· For the 5, 10, 20, and 40 µg/L microcosms, optimization of the model fit to the data when the initial total copper was set at the nominal values was complicated by measured total copper concentrations above the nominal value. 

· With initial dissolved copper,CD(0), specified in TICKET-UWM by extrapolation from measured data, agreement between model-predicted and experimental dissolved copper removal rates was observed for all microcosms once the settling velocity was optimized. 

· Optimized settling velocities for simulations using the logKDcalculated from experiment data and the logKDobtained from TICKET-UWM speciation calculations were 0.67 and 0.89 m/d, respectively. These values are consistent with the range observed in the Lake Courtille and Saint Germain les Belles (0.68 – 1.02 m/d) and the range associated with POC (Burns and Rosa, 1980);

· Unlike model applications to Lake Courtille, Saint Germain les Belles Reservoir, and the MELIMEX mesocosms, TICKET-UWM speciation calculations for the IME microcosms tended to underestimate the logKD;

· Both model simulations (i.e., withCd(0) specified) and measured dataindicate rapid removal of dissolved copper; and

· The relatively shallow depth of the microcosms (75 cm) favors rapid removal.

The TICKET-UWM testing results described above confirm that a relatively simplistic model (e.g. one water column cell and one sediment layer) can be used to simulate copper fate in the water column of lakes in a reasonably accurate manner. The critical issue is accurate parameterization of the characteristics and processes associated with the water body, particularly metal partitioning and particle settling velocity.