Soil system modelling and generation of field hypotheses

Abstract

Listen

Translate article

The relationship between soil modeling and field pedology is recursive. Models should be firmly grounded in field observations. Field data are then used to test, calibrate, and refine models. One often-overlooked relationship between models and field studies is the use of models to generate field-testable hypotheses unrelated to the model itself—propositions derived from model outputs or implications, the testing of which provides pedologic insight independent of the model and its underlying assumptions. This paper provides an example using a model of soil thickness. The case study illustrates a stepwise, recursive relationship between field evidence and models. The relationship between bedrock weathering, soil thickness, and surface erosion used in most numerical models of soil and hillslope evolution was generalized into a qualitative nonlinear dynamical systems model, the interaction matrix of which is dynamically stable. This supports the notion of a steady-state equilibrium soil thickness where weathering is balanced by surface removal. However, empirical data in the Ouachita Mountains, Arkansas, USA, shows nonequilibrium soil thickness. This in turn indicates either recent and/or large disturbances or changes in boundary conditions, or that factors other than weathering and erosion play a significant role in determining thickness. The inconsistency between model results and field evidence led to further field investigations, suggesting that biomechanical effects of trees on soil depth are highly significant. A model devised to explore this notion showed that the interrelationships between individual trees and soil depth are dynamically unstable, leading to generation of a specific field-testable hypothesis, that soil should be systematically thicker under trees than in adjacent sites. This relationship was confirmed by augering to bedrock at 108 pairs of tree stumps and immediately adjacent sites. The recursive relations between soil modeling and field pedology thus led to specific findings regarding the influence of trees on soil thickness and nonequilibrium soil thickness that would not have arisen otherwise. Another key lesson is that field observations contrary to a model may be of greater importance in advancing pedology than those consistent with the model.