Epsilon Open Archive

Abstract

In this thesis eleven microbial variables were evaluated as indicators of soil fertility. The variables were derived from short-incubation tests based on respiration, denitrification, ammonium oxidation, and alkaline phosphatase activity. During the incubation, the formation rate of metabolic end-products or intermediates was recorded and the variables were derived by use of linear or non-linear regression. Soils from three sites of a long-term field experiment and from a short-term laboratory incubation experiment were used for the evaluation. The microbial variables were evaluated in relation to standard soil chemical variables and yields from a whole crop rotation cycle. Substrate-induced respiration (SIR) and respiration rate under phosphorus limited conditions (Max-P) were indicators of labile organic material and available P, respectively. They were also the best yield predictors in unfertilised and nitrogen fertilised systems, respectively. SIR was shown to quantify both dormant and fully active microorganisms. Substrate amendment could transform the dormant biomass into an active stage. The division of SIR into active and dormant biomass fractions holds promise that the SIR assay can be of value as an indicator of the mineralisation dynamics at two different time scales. Moreover, the concept of active and dormant microbial biomass has implications on how to model the carbon fluxes in soil. Basal respiration was influenced by temporary substrate sources and rhizodeposits and was therefore not considered as a stable soil fertility indicator. Potential denitrification (PDA) and ammonium oxidation (PAO) was governed both by pH and substrate supply. The prediction models for PDA varied markedly between the sites and exhibited a relatively low predictive capacity, which renders PDA less sutable as a soil fertility indicator. However, PDA was satisfactorily predicted by a combination of PAO and SIR. When interpreting PAO, it should be regressed on pH to check for diverging soils that may be N limited. A combination of pH and organic matter characteristics was efficient for prediction of alkaline phosphatase activity (Alk-P), which implies that Alk-P is a pH-sensitive indicator of the microbial biomass.