Quantification of the compaction problem of selected vineyard soils and a critical assessment of methods to predict soil bulk density from soil texture
Thesis (Ph. D. Agric.) -- University of Stellenbosch, 1989.
ENGLISH ABSTRACT: Besides this overall abstract, each chapter has a separate.abstract under its own heading. Soil compaction is a problem common to many South African vineyard soils, and it has substantial adverse effects on root growth. Although literature reveals that much is known about the inherent soil properties affecting compactibility, extrapolation beyond the sample population remains a problem. The permanence of expensive soil loosening actions is uncertain on some soils, because a measure bf the bulk density to which the soil will recompact is not available. This study was conducted tq_ docu-me-nt- th-e compaction problem in vineyard s_oils in a total perspective. The final obj~_ftive was to ~ ~ ·-··----- pr~9ict both maximum compactibility (MBD) a_~9 equilibrium field bul_~ Q~£l~i!Y_ (FBD) fr9_r:ri _soil_ textu_r~I data, and to use other physical soil properties, e.g. soil structure, modulus of rupture, and air-to-water ·---- - --------~·~ ---------- permeability ratio as ~ackground information to explain _observed and predi<?!~~ soil -~ulk densities (BO). As a starting point, the effect of artificial subsoil compaction on grapevine shoot and root growth was studied in 85 dm3 pots. In a follow-up experiment, three deep tillage methods were evaluated in terms of the size and looseness of the rooting. volume. The soil in this experiment had a natural dense subsoil. Subsequently, 71 soil samples comprising of a wide textural range, and representing different degrees of compaction in the field, were collected from the most important viticultural areas of the Republic of South Africa. Soils were tested for various soil physical and chemical properties. Regression _techniques were used to relate soil compactibility to soil texture or some measure of soil structure . . Finally, a packing model of Gupta and Larsonwas tested for its applicability to predict soil BD. No critical BD or penetrometer soil strength (PSS) for the impedance· bf grapevine roots was found in the pot experiment for any of the five different soils investigated. However, aboveground grapevine performance decreased linearly with increasing subsoil BD's. Very low optimum compaction levels have been suggested, but this needs further investigation in field trials. In the field experiment, potential rooting volume was defined by BD and PSS values. Great care should be taken to define the position of PSS and BD measurements. During sample collection, the many types of compaction in the field, as it varied with soil type, were documented. A multiple regression model was fitted to the textural data and its distribution measures. It was possible to predict Proctor MBD with fairly good accuracy for most of the soils (R2 = 61 %) by using the strict norm of ±0,05 Mg m-3 variation. Soil structural properties (e.g. air-to-water p.ermeability ratio, modulus of rupture), were not necessarily related to compactibility, but were useful in explaining the differences between observed and predicted BD's. The Gupta-Larson model complemented the regression model in that it was able to predict FBD for the majority of the soils with a typical ·standard error of estimate of 0,08 Mg m-3. Soils whose FBD was under- or overpredicted could be sorted into logic groups.- based on origin, morphological characteristics such as clay illuviation, a specific textural composition, e.g. total sand content >60%, or tillage history. Some of the anomalies are outside the scope of the packing model. Some of the general conslusions of the study are listed below: (i) The 2 to 6 mm particle size class should be included for compaction studies and soils should be separated into at least 1 o size fractions. (ii) Increasing coarse sand contents led to increasing BD's, while increasing clay contents and higher coefficient of kurtosis values led to lower BD's. _ (iii) None of the relationships between predicted BD and Observed BD or any of the independent variables was nonlinear. (iv) None of the measured structural characteristics could statistically be used in regression models to describe variations in BD, and thus it was concluded that soil structural properties are not necessarily related to compactibility. (v) An air-to-water permeability ratio of 40 is suggested as a threshold value for soil structural stability. (vi) Modulus of rupture, determined after 12 hours soaking, successfully identified soils that get very hard upon drying. (vii) The claims that have been made in the literature for the suitability of soil texture as a characteristic to predict BD were justified for selected South African vineyard soils. (viii) Using the simple relationships, developed in this thesis, one can routinely predict MBD and FBD from textural data. (ix) The predicted BD's should, however, be interpretated in light of the current knowledge about the relationship between BD and root growth. (x) For screening purppses, e.g. to map a field in compactibility classes, only relative figures for BD are suffice. (xi) Although several areas for future research in soil compaction are suggested, this thesis has made a promising start t6 the understanding and alleviation of this problem.