Biomechanical effects, lithological variations, and local pedodiversity in some forest soils of Arkansas

Abstract

A high degree of soil variability over short distances and small areas is common, particularly in forest soils. This variability is sometimes, but not always, related to readily apparent variations in the environmental factors that control soil formation. This study examines the potential role of biomechanical effects of trees and of lithological variations within the parent material in explaining soil diversity in the Ouachita Mountains of Arkansas. The diversity of soils on Ouachita sideslopes is high, and the soil series vary primarily with respect to morphological properties such as soil thickness and rock fragment content. Soils vary considerably within small more-or-less homogeneous areas, and richness–area analysis shows that the overall pattern of pedodiversity is dominated by local, intrinsic (within-plot) variability as opposed to between-plot variability. This is consistent with variation controlled mainly by individual trees and local lithological variations. Given the criteria used to distinguish among soil types, biomechanical as opposed to chemical and hydrological effects of trees are indicated. Results also suggest divergent evolution whereby the pedologic effects of trees are large and long-lived relative to the magnitude of the initial effects and lifespan of the plants.

Introduction

Spatial variability of soils has long been recognized as a crucial issue in a variety of practical contexts and is emerging as a key concern in the geosciences. Because soils reflect the interacting influences of geology, climate, hydrology, geomorphic processes, and the biosphere, the understanding and interpretation of soil patterns and variability is of concern in the use of paleosols to reconstruct environmental change, and in comprehending contemporary earth surface systems. The value of spatial analysis of the soil cover in relation to environmental constraints on soil formation, soil processes and evolution, and the architecture of the environment has been amply demonstrated Fridland, 1976, Grzebyk and Dubrucq, 1994, Hole and Campbell, 1985, Ibañez, 1994, Ibañez et al., 1990, Ibañez et al., 1995, Ibañez et al., 1998, McBratney, 1992, McBratney, 1998.

A high degree of soil variability over short distances and small areas is common. This variability is sometimes, but not always, related to readily observed variations in soil-forming factors. Even when variability is related to (for instance) microtopography or tree throw, which are incorporated in the factors of soil formation conceptual framework, it may occur at a spatial scale which is too fine for typical applications of the soil-landscape model and the soil survey paradigm. While variation in individual soil chemical and physical properties is increasingly measured at very detailed spatial scales, variation in soils themselves (pedodiversity) is primarily treated at the scale of 1:10,000 or smaller soil maps.

Soils are influenced by multiple interrelated environmental factors. They may also include relic or inherited properties unrelated to contemporary environmental controls. Further, pedogenesis may sometimes be convergent, so that variations in environmental factors are reduced and obscured, and sometimes divergent, exaggerating the effects of minor initial variations or disturbances. Thus, even without consideration of the technical and practical problems of measurement and observation of environmental variability, linking soil variability to variations in vegetation, topography, hydrology, etc.—and vice versa—is no simple matter.

The purpose of this paper is to investigate the potential role of several factors, including biomechanical effects of trees, lithological variations in parent material, and microtopography in determining soil spatial variability in the Ouachita Mountains of Arkansas. This arose as a consequence of our observations of a great deal of soil variability at our study sites over small areas, which seemed to suggest a potential role for the effects of individual trees and microtopography. The criteria used to distinguish among the soils in the study area are morphological properties that would be influenced primarily by biomechanical (rather than biochemical or hydrological) effects of trees, thus we focus on this possibility. The importance of lithological variations emerged as the study progressed.

We were also interested in the relative importance of readily observable and measurable variations in soil-forming factors (for example, differences in topographic setting or parent material) in determining soil spatial variability, versus variations attributable to the unstable persistence and growth of minor variations in initial conditions or small disturbances. Several studies have suggested or demonstrated that dynamical instability and deterministic chaos can contribute to local-scale soil variability Culling, 1988, Ibañez, 1994, Liebens and Schaetzl, 1997, Minasny and McBratney, 1999, Phillips, 1998, Phillips, 1999, Phillips, 2001, Phillips et al., 1996, Webster, 2000. Instability and chaos leads to divergent soil development whereby small variations and perturbations persist and become exaggerated over time, rather than convergent development characterized by the muting of variations. The latter is an issue because, if effects of individual trees are invoked as a cause of the observed spatial pattern of soils, then the pedologic impacts of the trees must persist much longer than the trees themselves. Likewise, centimeter-scale microtopographic variation must involve some divergence from initial conditions to lead to morphologically distinct soil profiles at scales an order of magnitude or more broader.

The concern here is with soil morphology and soil stratigraphy rather than soil characteristics which tend to be fast-reacting and transient, such as nutrient status, organic matter, carbon, or pH. The latter are quite important, but this study is concerned with soil changes which would be relevant to soil mapping, the evolution of soils and regolith covers, and efforts to detect or reconstruct past vegetation boundaries based on soil properties. Thus the focus is on characteristics such as soil thickness, presence and thickness of master horizons, drainage status as indicated by redox features, the content and distribution of rock fragments, and the presence or absence of specific pedogenic features. Differences in these factors give rise to variation in soil taxa within the study area. Accordingly, the primary concern is pedodiversity—that is, the richness and variability of soils—as opposed to the variability of specific soil attributes.

While classification of soils, mapping of soil types, and analysis of the spatial patterns thereof is quite routine in practical applications of pedology and soil geography, and among many researchers, it must be acknowledged that not all soil scientists are comfortable with the basic idea that there exist qualitatively, categorically different types of soil that can be so identified and classified in a way analagous to biological taxonomy. In this view analysis of soil maps or of data on spatial distributions of soil types is of little value; rather the analysis should be based on specific soil features and characteristics such as nitrogen content, pH, or depth. This study is based on the premise that it is reasonable to identify qualitatively different types of soil, and that the analysis of the variability of these entities provides insight not obtainable from the analysis of separate soil properties. This is based on three assertions. First, soil classifications integrate the effects of a number of specific soil properties and are thus more comprehensive indicators of soil variability. Second, soil classification, while clearly imperfect and sometimes arbitrary, is a systematic, rule-based technique for grouping similar and distinguishing among dissimilar soils in a way that numerical values of soil properties cannot. Third, we believe the record and tradition of this type of analysis has produced scientifically useful and practically relevant results that clearly legitimize it (e.g., Beckett and Bie, 1978, Bregt et al., 1992, Fridland, 1976, Grzebyk and Dubrucq, 1994; Guo et al., 2003; Hole and Campbell, 1985, Ibañez et al., 1995, Ibañez et al., 1998). We also believe this reasoning applies more generally to factors such as lithologies and vegetation communities, which may be similarly imperfectly and sometimes arbitrarily classified. In short, this work is based on the premise that there is value in studying the spatial structure of the soil cover.