The relative importance of texture, structure, organic matter and clay mineralogy to the nature of the soil moisture characteristic is examined for an extensive group of Australian soils using numerical classification and diagnostic methods.The presence of pedality, particle size composition and grade of structure were the soil properties most consistently associated with differences between the groups of soils with similar moisture characteristics. By association, field texture was shown to be a useful property. Although the presence of pedality and grade of structure were important, the shape and size of ped had only weak associations with differences in the soil moisture retention.Montmorillonite, iron oxide, vermiculite and quartz were the minerals in the clay size fraction which appeared to be important if they were present. In contrast, the presence of illite did not show any strong associations with a particular position or form of the moisture characteristic.The soil moisture characteristic was successfully modelled as a power function. It appears that being able to group and classify the soil moisture characteristic and then to provide a description of these groups both in terms of soil properties and model parameters is a valuable means of developing simple predictive models for field soils. The error of our predictions for 44 horizons based on this simple approach appears to be only marginally larger than that encountered in conventional laboratory methods, and in view of soil heterogeneity it is argued that following further development these predictions may be adequate in many hydrological and agricultural applications.

Errors of measurement of soil water content with the neutron probe can be made small enough for most water balance studies; and apart from the possible damage resulting from access tube implacement, they can be readily specified. A more important problem is estimating how much water is lost from soil by drainage.A simple, graphical method is described to distinguish between drainage and evaporation from soil, based upon identification of the maximum depth at which measurable quantities of water are extracted by roots. As a result of errors in soil water measurements and uncertainties in the distinction between drainage and evaporation, estimates of soil water deficit and hence evaporation for a single profile are likely to have a precision of about 2 5 mm. Larger variability i s commonly found in soil water storage changes measured in replicate profiles partly because of uneven wetting of the soil during rain and partly because of non-uniform losses by drainage and root extraction. This variability requires careful attention to replication and location of access tubes when asscssing the evaporation for a particular site.

No abstract

SUMMARYDeterminations were made of the available-water capacities and moisture release characteristics of six soils of different texture from Rothamsted, Woburn and Saxmundham. The amount of available water retained in the surface foot of soil ranged from 1·59 in in a sandy loam to 2·75 in in a silt loam under permanent grass. With one exception soils which had received additions of farmyard manure had significantly higher available-water capacities than unmanured soils.