Pacific Coast Environmental Metrics
"Specializing in measuring intangibles."

Zero Waste Planet


Our Mission

Social Games

Common Pool Resources

Collective Action

Publications/Working Papers


Frequently Asked Questions

About Us


Top Level

Box 63
Malahat, B.C.
V0R 2L0
(250) 730-3738
metrics [at]

Problems with Probabilities: When protecting 95% of the species 50% of the time means protecting only 80% of the species 12% of the time.


This study compares a small set of guidelines between Canada and the Netherlands that were derived from socially similar approaches. This small set is reflective of the elimination of non-similar approaches and differing regulated products from the larger national guidelines of both countries. The conclusion is that in both countries the guidelines favored economic interests over ecological protection, although no statement is made about the quality of the guidelines themselves.

Netherlands approach to soil guidelines

The Netherlands methodological approach to developing soil guidelines for heavy metals is based on three methods. For toxicants that few data exist, either the “modified EPA-style” or “equilibrium partitioning” methods are used. These approaches are not discussed in this study. For toxicants for which data exist for “four or more taxonomic species,” the approach uses a “statistical extrapolation” approach. This approach generally considers a protection level of 95% of the species present, and the guidelines often cite “Aldenberg and Slob” as the origination of the “statistical extrapolation” approach (Crommentuijn, et al 1997):

The aim of environmental quality objectives is that the (Maximum Permissible Concen-
tration) is set at a level that protects all species in an ecosystem. However, in order to be
able to use extrapolation methods like the one of Aldenberg and Slob (1993), a 95% pro-
tection level is chosen as a sort of cut-off value...From the estimated distribution a
concentration can be derived which is assumed to be ‘safe’ for ecosystems. (p 43).

A review of Aldenberg and Slob (1993) raises some questions about the application of their paper in this context. Aldenberg and Slob routinely refer to two preceding papers (Kooijman, and Van Straalen and Denneman) as deriving the approach as well as the protection level, and instead focuses on analyzing the statistical probability that the extrapolated protection levels are correct. Their paper argues that it is insufficient to extrapolate a concentration level which will be protective of a percent of species without calculating the probability that the extrapolation is incorrect. This adds a second parameter to the “95% protection” statement, that of the confidence level of that statement. Aldenberg and Slob are very specific in suggesting that a protection statement should be calculated for both 50% and 95% confidence levels, i.e. determine the range for which an individual can be 95% certain that 95% of the species will be protected. A protection statement with a 50% confidence level will mean that statistically 50% of the time less than 95% of the species will be protected. Since this failure is open-ended, the degree of unprotected species can be substantial. One calculation in the paper indicated that for the given 50% confidence scenario a 12% chance existed that 20% of the species would actually be harmed.

Aldenberg and Slob (1993) are quite clear regarding their beliefs of use of the 50% and 95% confidence levels:

The great virtue of regarding the 95% confidence value as the safe value is that it tends to
outweigh ecological and economical interests. If this safe value, based on available data,
appears to be low enough to have important economic drawbacks, one would not hesitate
to investigate more species, since the associated reduction of uncertainty might quite well
result in higher values for the safe concentration. On the other hand, using the 50%
confidence value as an indication of the safe value results in a strong bias toward economic
interests. One would test a minimum number of species, hoping that the coin falls on the
right side; if not, one could always extend the number of test organisms afterward.
Obviously this situation would be quite harmful from an ecological viewpoint. (p

In the Netherlands’ guidelines (Crommentuijn et al 1997), the choice is for 50% confidence:

When PAF (Potentially Affected Fraction of species) is equal to 0.05, the corresponding
concentration can be calculated with 50% and 95% confidence. In the Netherlands the MPC
is set equal to the former value. (p. 45)

For the 1997 guidelines for heavy metals, the Netherlands modified the statistical extrapolation approach based on work by Struijs, van de Meent, Peijnenburg, van den Hoop, and Crommentuijn (1997). This was labeled “Added Risk Approach” and resulted in a Maximum Permissible Concentration that was the sum of the background concentration and the Maximum Permissible Addition (MPA). The MPA was derived from protecting 95% of the species not already at risk from the background levels in conjunction with bioavailability of the additional metals, with a 50% confidence level. The background concentration (Cb) was set by a single value per heavy metal for all of the Netherlands.

Aldenberg and Slob (1993) reevaluated data on cadmium in soil from one of the referenced articles (Van Straalen and Denneman) and calculated the levels for protecting 95% of the species both 95% and 50% of the time (n=7). The 95/95 level was 0.03 mg/kg, while the 95/50 level was 0.53 mg/kg. The 1997 MPA level for cadmium in soil was 0.76 mg/kg, and when the MPA was added to the background concentration of 0.8 mg/kg the MPC was 1.6 mg/kg.

Canadian approach to soil guidelines

The approaches to deriving Canadian soil guidelines centered on the use of EC25 data for a distribution of species. Canada’s guidelines break down usage into four categories (agricultural, residential/parkland, commercial, and industrial). For commercial and industrial usage, the 50th percentile of the EC25 data is selected when possible, and this number is labeled the “Effects Concentration Low (ECL)” level. For agricultural and residential/parkland usage, the 25th percentile of the EC25 data is selected, but additionally divided by an uncertainty factor to arrive at a “Threshold Effects Concentration (TEC).” When EC25 is not available, data meeting guideline requirements from both effects and no-effects studies are combined and used as EC25. If this is still not possible, the TEC is derived from the Lowest Observable Effects Concentration (LOEC) through the use of dividing by an uncertainty factor (CCME 2006).

The ECL used for commercial and industrial usage attempts to protect 50% of the species that showed 25% effects rates for the ECL chosen. This level is likely to show higher effects rates for the 50% of the species whose EC25 is exceeded. The TEC chosen is likely to protect more than the 75% of the species that showed 25% effects rates, since the TEC is divided by an uncertainty factor. Both the ECL and TEC base their determinations on socially acceptable levels of effects. The ECL levels of effects are substantially higher than for the Netherlands. The TEC levels of effects are dependent on the chosen uncertainty factor and could protect a higher number of species than in the Netherlands.


Of the seven regulated heavy metals in common between Canada and Netherlands, five guidelines in the Netherlands used the statistical extrapolation method. Table 1 lists the comparison of Netherlands’ soil guidelines and Canada’s soil guidelines (CCME 2007 and Crommentuijn, et al 1997).

















































Table 1. All units are mg/kg.

Canadian guidelines:
Agri. - agricultural
Res. - residential/parkland
Com. - commercial
Ind. - industrial
Netherlands guidelines:
MPA - Maximum Permissible Addition
Cb - background concentration
MPC - Maximum Permissible Concentration


As can be seen from the table, all five concentration guidelines from the Netherlands are equal to or less than Canadian residential guidelines, and three (copper, mercury, and zinc) are lower than Canadian agricultural guidelines. Given that Aldenberg and Slob (1993) already indicated the Netherlands’ 50% confidence level reflected economic interests over ecological protections, the choice of higher concentration levels in Canadian guidelines reflect the economic interests even greater.

Aldenberg, T,, and Slob, W. (1993). Confidence Limits for Hazardous Concentrations Based
on Logistically Distributed NOEC Toxicity Data. Ecotoxicology and Environmental Safety 25(1), 48-63. Retrieved Nov. 14, 2008 from ScienceDirect database

Canadian Council of Ministers of the Environment (CCME) (2006). Summary of A Protocol for
the Derivation of Environmental and Human Health Soil Quality Guidelines. Retrieved
Nov 12, 2008 from CCME database

Canadian Council of Ministers of the Environment (CCME) (2007). Canadian Soil Quality
Guidelines for the Protection of Environmental and Human Health, Summary Tables.
Retrieved Nov. 14, 2008 from CCME database

Crommentuijn, T., Polder M.D., & van de Plassche, E.J. (1997). Maximum Permissible
Concentrations and Negligible Concentrations for metals, taking background
concentrations into account. Retrieved Nov. 14, 2008 from RIVM database

Struijs, J., van de Meent, D., Peijnenburg, W. J. G. M., van den Hoop, M. A. G. T., &
Crommentuijn, T. (1997). Added Risk Approach to Derive Maximum Permissible
Concentrations for Heavy Metals: How to Take Natural Background Levels into
Account. Ecotoxicology and Environmental Safety, 37 (2), 112-118. Retrieved Nov. 15,
2008 from ScienceDirect database.