Bioavailability Correction Factors

Correction for Leaching and Ageing Effects

The Ni soil toxicity data used for PNEC derivation (see Fact Sheet 2) are generally based on soils spiked in the laboratory with a soluble Ni salt. Comparative studies show that toxicity tests in freshly spiked soil generally overestimate toxicity of Ni to soil organisms/microbial processes compared to field-contaminated soils or aged soils. It is observed that Ni solubility in soils appreciably decreases with increasing equilibration time after an initial rapid sorption phase. Hence it can be expected that testing soils immediately after spiking with a soluble Ni salt will also result in an overestimation of Ni toxicity compared to long-term equilibrated soils.

Moreover, spiking a soil with a soluble metal salt also changes the physico-chemical characteristics of the soil by increasing the ionic strength and decreasing the pH of the soil.  These spiking artifacts can either directly affect the response of the endpoint tested or alter the Ni bioavailability and the toxicity of the soil. These issues are corrected by leaching and ageing the soil after spiking.

Figure 2To correct for the ageing and leaching effects when using Ni soil toxicity data for PNEC derivation, a correction factor (i.e., leaching-ageing factor) was developed. The L/A (leaching-ageing) factor2 is the ratio of toxicity values in leached and aged soils to toxicity values in corresponding freshly spiked soils. For Ni, there is a clear effect of pH on the change in toxicity after leaching and ageing (Figure 2). Toxicity is ameliorated least in acidic soils (median ageing factor of 1.2) and most in calcareous soils (median ageing factor of 8.4). This is consistent with the differences in the amount of added Ni that is isotopically exchangeable in freshly spiked and corresponding aged soils for 16 soils sampled across Europe with contrasting soil properties and land use.

The L/A factor is estimated from an empirical model fitted to the ratio of added Ni that is (isotopically) exchangeable in freshly amended soils (1–21 days after amendment) to that after at least one year of ageing, i.e., L/A = 1 + exp (1.4(pH-7.0)). This equa-tion is calibrated on soil aged maximally 1.5 year and soil pH ranged between pH 3.6 and 7.7. This factor is a conservative estimate for the changes in toxicity.

The L/A factors vary between 1 to 4 and exhibit a positive rela-tionship with pH, but only become significant beyond pH values of about 6.0.

Normalization for Soil Types

Bioavailability and chronic toxicity of Ni to soil organ-isms/microbial communities vary as a result of the characteristics of the soil media.  In order to make comparisons between laboratory toxicity data, results must be normalized to a standardized set of conditions using bioavailability models.  Bioavailability models can be used to derive site-specific HC5 and PNEC values for sites in which appropriate soil properties have been quantified.

Chronic regression bioavailability models for Ni have been de-veloped using laboratory experiments for three different trophic levels, i.e., for microbial function [using nitrification (PNR), sub-strate induced respiration (SIR), and maize respiration (MR)], for higher plants (using the tomato Lycopersicon esculentum and barley Hordeum vulgare), and for both hard-bodied (using the collembole Folsomia candida) and soft-bodied (using the worm Eisenia fetida) invertebrates.

The chronic regression models for Ni were developed/calibrated based on soils that represent the full range in physico-chemical parameters (pH, clay, OM, eCEC) that represent soil conditions in the EU. Accounting for differences in soil properties significantly explained variation in Ni toxicity to all endpoints tested, and it was observed that chronic Ni toxicity was best correlated with the eCEC of the soils.

The same trends were observed for all of the species tested:

          as eCEC ↑, toxicity ↓

Linear regression models (log EC50 (mg/kg) = a + b log eCEC) were developed to allow for normalization based on differences in Ni toxicity between soils with different properties. An overview of all significant regression models is presented in Table 1.

Table 1: Overview of All Significant Regression Models Relating the Toxicity of Nickel ([Ni] in mg/kgdw, after correction for ageing) to eCEC

Organism or
Microbial Function

Regression Model3

R3

Invertebrates
Eisenia fetida

log[Ni] = 0.95 log (eCEC) + 1.76)

0.72

Folsomia candida

log[Ni] = 1.17 log (eCEC) + 1.70)

0.71

Higher Plants
Hordeum vulgare

log[Ni] = 1.12 log (eCEC) + 1.57)

0.83

Lycopersicon esculentum

log[Ni] = 1.27 log (eCEC) + 1.06)

0.67

Microbial Community
Nitrification

log[Ni] = 1.00 log (eCEC) + 1.42)

0.60

Substrate induced respiration

log[Ni] = 1.34 log (eCEC) + 1.38)

0.92

Maize induced respiration

log[Ni] = 1.22 log (eCEC) + 1.37)

0.72


2.  L/A factor: the term leaching-ageing factor (L/A factor) refers to the combined effect of leaching (due to changing ionic strength) and ageing (due to long-term reactions) on Ni bioavailability and toxicity in soil.

3.  The slopes of the bioavailability models were based on regressions of EC50 with soil properties, because EC50 values are more robust and less sensitive to experimental error compared to the NOEC or EC10 values.