The application of expert knowledge in Bayesian networks to predict soil bulk density at the landscape scale

Abstract

This paper investigates the use of expert knowledge as a resource for digital soil mapping. To do this, three models of topsoil soil bulk density (D b ) were produced: (i) a random forest model formulated and cross-validated with the limited data available (which served as the benchmark), (ii) a naïve Bayesian network (BN) where the conditional probabilities that define the relations between D b and explanatory landscape variables were derived from expert knowledge rather than data and (iii) a 'hierarchical' BN where model structure was also defined by expert knowledge. These models were used to generate spatial predictions for mapping topsoil D b at a landscape scale. The results show that expert knowledge-based models can identify the same spatial trends in soil properties at a landscape scale as state-of-the-art mapping algorithms. This means that they are a viable option for soil mapping applications in areas that have limited empirical data.

Machine learning for digital soil mapping: Applications, challenges and suggested solutions

“…An example of model structure providing algorithmic explanations in DSM is found with the use of Bayesian belief networks (BBN, Cooper, 1990) in Mayr et al (2010) and later Taalab et al (2015). BBN is a probabilistic graphical model predicting the likely value of a soil property or class given conditional dependencies between covariates.…”

Robust spatial estimates of biomass carbon on farms

“…Probabilistic Bayesian Belief Networks (BBNs) provide a way of understanding social-environment interactions where complex sets of geospatially interdependent variables can affect environmental outcomes and consequently human decisions on landscape management (Grafius et al, 2019;Karimi et al, 2021). The capability of BBNs to describe complex systems (Borsuk et al, 2004;Heckerman, 1997;Pearl, 1988;Taalab et al, 2015aTaalab et al, , 2015b potentially provide an approach to quantifying biomass stocks and their certainty for a situation where we typically have incomplete knowledge and information (Hassall et al, 2019;Korb and Nicholson, 2010). Moreover, BBNs when combined with Geographical Information Systems (GIS) can be used to guide on-farm decisions by providing an idea of the probability of events, defining key variables and conditional dependencies, and identifying how these could affect the outcomes (Grafius et al, 2019).…”

“…Moreover, BBNs when combined with Geographical Information Systems (GIS) can be used to guide on-farm decisions by providing an idea of the probability of events, defining key variables and conditional dependencies, and identifying how these could affect the outcomes (Grafius et al, 2019). The C stored in a farm can be affected by many factors (Krauss et al, 2022;Branca et al, 2013), however, conditional dependencies can assist in narrowing these down to the ones affecting most the parameter that needs to be predicted (Taalab et al, 2015a(Taalab et al, , 2015b.…”

Distinct respiratory responses of soils to complex organic substrate are governed predominantly by soil architecture and its microbial community

“…Bayesian Belief Networks are a graphical representation of a probabilistic dependency model. They represent the variables that affect the response of interest in the form of a graph or network and describe the relationships between the drivers and responses as a set of conditional probabilities (Talaab et al., 2015). Soil property, microbiology, land management, and environmental context data were combined in the BBN and used to predict cluster identity for each site.…”