Soil surveys and soil classification, are based on a static view of soil prope1ties (subsoil properties) that tend not to change significantly on the human time scale; however, most growers and land managers identify soils as dynamic systems (topsoil properties). The current conventional soil classifications therefore, fall sho1t of describing the dynamic/functional
behaviour of soils or the soil quality indicators, which is of most interest to land managers. The scope of this thesis is to investigate whether broad soil types defined by traditional soil classification, can be used to predict the soil quality indicators (SQis) and whether the SQis can be used to classify soils which can predict soil function and bacterial biodiversity. In addition,
we investigated whether other factors ( e.g. vegetation classes) regulate the SQ Is and whether there are critical limits in the SQI in different soil types or vegetation types (A YCs). To achieve this, we monitored (I) Carbon turnover rates monitored over a (i) 90 day and (ii) 1.5 y period using 14C-labelled artificial root exudates or 14C-labelled plant leaves, (2) soil quality factors and
the dominant attributes in the factors, (3) Soil respiration, mineralisation and biodiversity, in the different soil types and AVCs. Results from several statistical methods employed on these SQis revealed significant differences between soil types or A VCs, however, the differences were small. In most cases only the Peat or Peloso! soils were distinctly different from the rest of the soils or the Heath and Bogs, Moorland and Grass Mosaic, and Upland Wooded from the rest of
the rest of habitats. The definition of the class limits remained ambiguous, as exclusive reference values for each soil type or A VCs could not be established due to overlaps in SQ! ranges.
Statistical soil classification by cluster analysis based on selected soil physico-chemical prope,ties did not improve its predictability of the soil function and diversity. We conclude that conventional soil classification provides a poor predictor of most SQis. We fu1ther conclude that long-term laboratory mineralisations in soils at constant temperature failed to reveal major
differences between soil types and that laboratory mineralization studies may provide a poor proxy for predicting soil C sequestration potential. We asc,ibe this to the inability of sho1t tenn biological assays to represent pedogenic processes which have taken ca. I 0,000 y to become manifest.