Measuring the immeasurable: A structural equation modeling approach to assessing soil health. Science of The Total Environment 870: 161900.
McClellan Maaz, T., Heck, R.H., Tallamy Glazer, C., Loo, M.K., Rivera Zayas, J., Krenz, A., Beckstrom, T., Crow, S.E., Deenik, J.L. 2023.
Summary by Dobermann, A.
Terms such as ‘soil quality’ or ‘soil fertility’ were used in the past to describe the general functions and properties of soils used in agriculture or nature, but ‘soil health’ has gained a lot of popularity in recent years. Scientists, other professionals and farmers have struggled, however, to define soil health more precisely, particularly also in terms of measurable soil properties (‘indicators’) that matter most for a specific crop or cropping system, or for specific functions a soil is expected to provide. Many studies present and assess soil health indicators but very few provide a quantitative soil health score. In this paper, the authors use structural equation modelling – a form of regression – and data from 50 sites across six islands of Hawaii to derive a soil health scoring function. The analyses grouped the mostly biological and physical soil parameters into three factors. The resulting soil health score was signiﬁcantly lower in conventional cropping systems relative to other land management, which was likely related to differences in soil organic carbon and its various fractions. The main advantages of the structural equation modeling approach to soil health are the simultaneous assessment and weighting of soil health indicators, accounting of measurement error, and inclusion of a variety of indices to evaluate individual parameters. Of course, the results will depend on the soil properties entered into the analysis, which may make standardization more difficult. Different structural equation models will probably be required for different main uses (roles) of a soil, but this is certainly an area worth exploring more.