Estimation of reference bulk density from soil particle size distribution and soil organic matter content

Abstract

Crop growth and yield are affected by soil compactness. Rather than using bulk density, it has been suggested that soil compactness be described by relative bulk density, e.g. the degree of compactness, previously defined as the ratio of bulk density (ρ) to reference bulk density (ρ_ref). This study investigated relationships between ρ_ref as defined by Håkansson (1990) and soil particle size distribution (PSD) by analysing soil data from 171 experimental sites in Sweden, two in Poland and three in Finland. PSD was characterised either by the common size fractions (i.e. clay, silt and sand content) or by fitting a (continuous) mathematical function to the experimental PSD data. The Rosin–Rammler equation was used for the latter, and the PSD was characterised by the Rosin–Rammler parameters α and β. We present equations for estimation of ρ_ref from either soil textural classes or from α and β in combination with soil organic matter content (OM). It was shown that ρ_ref is largely controlled by OM. The best model (i.e. the model with the smallest value of Akaike Information Criterion) was found to be one that estimates ρ_ref from α, β and OM. The regression models for calculation of ρ_ref presented here could be incorporated into models for calculation of crop yield losses due to soil compaction. Furthermore, we found good agreement between ρ_ref and values for critical bulk density for root growth reported in the literature, indicating that ρ_ref is a critical bulk density for root growth.

Introduction

Crop growth is affected by soil structure and compactness. With respect to crop growth, there is an optimum soil compactness above or below which crop growth and yield are reduced (e.g. Håkansson, 1990). While yield loss due to over-compaction is largely a result of high penetration resistance for roots and poor soil aeration, yield losses due to too loose soil are less well understood but are generally associated with reduced transport of water and nutrients to the roots under unsaturated conditions. The optimum soil compactness depends on internal soil attributes such as soil texture and soil organic matter content, but may be different for different crops and climatic conditions (Håkansson and Lipiec, 2000, Håkansson, 2005, Suzuki et al., 2006, Reichert et al., 2009).

Bulk density (or porosity) is the parameter usually used to describe soil compactness. If soil is compacted, bulk density increases and porosity decreases correspondingly. However, absolute values of bulk density are unsuitable for characterising soil compactness with respect to crop yield when comparing different soils, as optimum and critical limits of bulk density for crop growth strongly depend upon soil type, i.e. different values of bulk density are optimum for different soils, as demonstrated by Reichert et al. (2009). A bulk density that indicates a compact state in one soil may imply a loose state in another (Håkansson, 1990).

Therefore, relative bulk density has been suggested to describe the compactness of soil. The relative bulk density is generally obtained as the ratio of actual field bulk density (i.e. bulk density measured either directly in the field or on undisturbed soil samples collected in the field) to a reference bulk density. As bulk density changes with soil moisture for swelling–shrinking soils, field bulk density should be measured at a defined soil moisture content (e.g. at field capacity) in order to obtain reliable and comparable relative bulk density values. Different authors have used different ways of obtaining the reference bulk density. Pidgeon and Soane, 1977, Carter, 1990, da Silva et al., 1994 obtained reference bulk density from the Proctor test at a given amount of impact energy. Other researchers define the reference bulk density as the bulk density obtained by uniaxial compression at 200 kPa (Håkansson, 1990, Lipiec et al., 1991, da Silva et al., 1997) and 1600 kPa (Suzuki et al., 2006). These tests all result in relatively high bulk densities. However, stresses higher than those used for obtaining the reference bulk density may occur due to agricultural field traffic and consequently field bulk densities higher than the reference bulk density can be observed. Recently, Reichert et al. (2009) compared different methods for obtaining reference bulk density and evaluated the usefulness of the relative bulk density approach.

Håkansson (1990) related crop yield to the degree of compactness, DC (%), which is given as:

$$DC=100\frac{\mathit{\rho }}{{\mathit{\rho }}_{\text{ref}}}$$

where ρ is the field bulk density, and ρ_ref is the reference bulk density determined by a drained, uniaxial compression test using wet soil samples loosely filled in an oedometer and loaded with a vertical stress of 200 kPa until drainage ceases. In a series of 100 field experiments with spring barley (Hordeum vulgare L.), Håkansson (1990) found that on all types of mineral soils, the highest crop yield was obtained at the same DC (Eq. (1)), viz. at a DC value of 87. This is further supported by other studies relating crop yield to soil compactness (Lipiec et al., 1991, Riley, 1998, Braunack et al., 2006). The main reason, as shown by Håkansson and Lipiec (2000), is that critical limits of soil aeration as well as of soil mechanical resistance to root growth are similarly related to DC in all soils. However, the optimum value for DC may slightly differ between crops as shown by Håkansson, 2005, Reichert et al., 2009.

Analytical soil compaction models used in agricultural research (e.g. O'Sullivan et al., 1999, van den Akker, 2004, Keller et al., 2007) incorporate calculation of (changes in) bulk density due to mechanical stresses (i.e. due to loading with agricultural machinery), but do not include effects on crop yield. Models for estimation of the effects of soil compaction on crop yield are purely empirical (e.g. Arvidsson and Håkansson, 1991). One reason for not including crop response in analytical models is that, as mentioned above, absolute values of bulk density are not useful for characterisation of crop growth. A relative bulk density parameter such as DC (Eq. (1)) would facilitate the inclusion of crop response in analytical models.

In order to calculate DC, ρ and ρ_ref have to be known (Eq. (1)). While ρ is readily measured or easily calculated by means of soil compaction models, determination of ρ_ref is labour-intensive. Therefore, it is relevant to investigate relationships between ρ_ref and readily-quantifiable soil characteristics for potential development of prediction equations.

The objective of this paper was to investigate relationships between the reference bulk density as defined by Håkansson (1990) and the particle size distribution and organic matter content based on extensive data from field experiments conducted in Sweden.