Main authors: Else K. Bünemann, Giulia Bongiorno, Zhanguo Bai, Rachel E. Creamer, Gerlinde De Deyn, Ron de Goede, Luuk Fleskens, Violette Geissen, Thom W. Kuyper, Paul Mäder, Mirjam Pulleman, Wijnand Sukkel, Jan Willem van Groenigen and Lijbert Brussaard
Editor: Jane Brandt
Source document: Bünemann, E. K. et al. (2018) Soil quality - A critical review. Soil Biology and Biochemistry, Volume 120, May 2018, pp 105-125


An indicator is only useful if its value can be unequivocally interpreted and reference values are available. Reference values for a given indicator could be either those of a native soil, which may however not be suitable for agricultural production, or of a soil with maximum production and/or environmental performance (Doran and Parkin, 1994). In the Netherlands, for example, ten reference soils for good soil biological quality were selected out of 285 sites that had been monitored for over ten years (Rutgers et al., 2008). These reference soils represent specific combinations of soil type and land-use (e.g. arable land on clay soil). Soil quality indicators at a given site could thus be compared to those at the reference site as well as to the mean value, and 5% and 95% percentiles of all sites under a given land-use, with the percentiles given as a means to express the frequency distribution. An important drawback of this approach is that the reference may not be at an optimum in all parameters (Rutgers et al., 2012).

Acceptable values for an indicator can also be defined as those at which there is no loss or significant impairment of functioning (Loveland and Thompson, 2002). In the context of pollution, thresholds of contamination are often used (Chen, 1999). Likewise, Arshad and Martin (2002) list threshold levels for soil quality indicators, but this is rarely found in other publications on soil quality assessment. For plant nutrients, most agricultural advisory services use thresholds of available reserves below which plant production may become nutrient-limited, while maximum values are related to the risk of losses (Allen et al., 2006; Schoumans et al., 2014). Indicator thresholds for other soil functions are absent from most soil quality assessment approaches.

A more advanced way to evaluate soil quality indicators is the establishment of standard non-linear scoring functions, which typically have the shapes i) more is better, ii) optimum range, iii) less is better, or iv) undesirable range, with i-iii being most common in soil science. The shape of such curves is established based on a combination of literature values and expert judgement (Andrews et al., 2004). When scoring curves are based on regional data, such as in the Cornell Soil Health Assessment (Moebius-Clune et al., 2016), then scores are relative to measured values in the respective region. Each indicator measurement is transformed to a value between 0 and 1 (or 0 and 100) using a scoring algorithm (Karlen and Stott, 1994), with a score of 0 being the poorest (lower threshold) and a score of 1 (or 100) the best (upper threshold). The baseline value equals the midpoint between threshold values. Validation of scoring curves is possible if datasets with measurements of the given soil quality indicator and a related soil process are available.

Obviously, acceptable target ranges of soil quality indicators need to be soil- and land use-specific, and they depend not only on targeted soil functions, but also on both spatial and temporal scale of soil quality assessments, with regional target ranges typically being narrower than national ones (Lilburne et al., 2004; Wienhold et al., 2009). In addition, acceptable ranges of a soil quality indicator for one property or process are often highly dependent on the value of another soil property or process, e.g. dependence of microbial biomass or soil organic carbon on soil texture (Candinas et al., 2002; Johannes et al., 2017).

It has been claimed that the interpretation of soil quality indicators, i.e. the establishment of target or workable ranges, will always remain contentious, which is partly due to a lack of data, partly due to the curvilinear pattern that many indicators follow and partly because the use of expert judgement is contentious itself (Merrington, 2006). A comparative approach in which indicator values or scores of a given sampling point are put in relation to other sampling points may be the most intuitive and flexible basis for interpretation, since it gives a relative assessment (e.g. top 25%) and allows continuing evolution of the system. This approach is being implemented in the iSQAPER project, where the variation in soil quality indicator values within pedo-climatic zones is determined. Ranges are defined for specific land uses (e.g. arable land, grassland), and benchmark scores based on relative frequency are given. This approach may also introduce modular extensions of indicators that are only relevant in specific contexts, where stakeholders can relate to them. Decision trees based on environmental conditions, management systems and relevance of ecosystem services can guide the selection of specific indicators.


Note: For full references to papers quoted in this article see

» References

Go To Top