Dispersal, heterogeneity and resistance: challenging soil quality assessment

Abstract

One of the most common, fundamental and intuitively attractive methods to assess environmental impact of pollutants is a survey of species density at a number of sites in a gradient around a known or expected source of pollution. The numbers derived are gross estimates of the site-specific response in mortality, reproduction, and immigration/emigration taken together and would normally represent the cumulated population performance, usually integrated over more than one generation. For certain groups that are more susceptible or exposed than others to the pollutants, a pattern may appear that relates the density variation in the gradient to the exposure data. One such example is given by Bengtsson et al (1983) from their survey of earthworms in forests around a brass mill in south-east Sweden. It seems justifiable to suggest from their Figure 2 that the density of earthworms be inversely related to the concentration of metals in the soil. The predictive power of these data is, however, very weak because of a great variability in numbers between replicate samples. This is especially true if individual species are considered (coefficient of variation, CV, ranging from 100 to 500% for n = 15, Table 1 in Bengtsson et al, 1983). The confidence interval for the correlation between the soil metal concentration and the earthworm density is such that a huge number of replicates would be required from a randomly selected site to tell whether earthworm density was influenced by the soil metal concentration or not (Figure 9.1).