Choosing the right pedological support for digital soil assessment modeling

Recently a spline tool was developed within the GLobalSoilMap.net Project to build a continuous representation of the vertical variability of soil properties. This approach is meaningful in addressing the continuous representation of a soil profile and the gaps infilling of missing samples at any depth. However soils are natural bodies that often behave in a way related to its vertical structuring. That is how any soil function at hand (i.e. the storage or infiltration of water in soils) acts over a period of time depends on the type of considered soils as an emergent behaviour, which in turn depends on the behaviour of the genetic horizons components. This consideration recalls two important things: the need to survey soils using a classical approach in which the fundamental survey unit is the genetic horizon (as opposed to the fixed depths surveying), and the need to consider soil properties/behaviours as attached to the horizons themselves. In our opinion the response at the DSA frame should account the profile properties in terms of a step instead of a spline function. Different models of spatial inference are compared to crystallize this thinking and to highlight the importance of choosing the right pedological support for any definite DSA issue. Preliminary results show that integrated 3D spatial modelling, nonlinear statistical approaches and a step function accounting of the vertical variability of soil attributes are promising in bridging the actual gap between the corroborated DSM procedures and the needed DSA products.