Nitrogen mineralization: a review and meta-analysis of the predictive value of soil tests

Accurate estimation of mineralizable nitrogen (N) from soil organic matter is essential to improve fertilizer management in agricultural systems. Mineralizable N refers to the amount of N in soil that is released during a certain period (ranging from 1 week to the length of a growing season). It has been estimated from increases in inorganic N during incubation or from N uptake by plants grown in a greenhouse or field. Many chemical soil tests measuring extractable organic N (EON) fractions have been proposed to predict mineralizable N. We evaluated the predictive value of these soil tests, using 2068 observations from 218 papers. Meta-analysis was used to find the best soil test, to analyse differences between field and laboratory experiments, and to determine whether their predictive value is affected by extraction intensity (% of total soil N that is extracted). The concentration of EON was positively related to mineralizable N, explaining on average 47% of the variation. It did not, however, explain more of the variation than total N. Best predictions (57% < R2 < 74%) were obtained when EON was extracted with hot CaCl(2), acid KMnO(4), acid K(2)Cr(2)O(7), hot water or hot KCl. Extraction intensity was not related to the strength of the above-mentioned relationship. Predictions of mineralizable N were significantly worse when mineralization was measured in the field compared with measurements under controlled conditions. We found no evidence of a causal and direct relationship between EON and mineralizable N. Accuracy of soil testing may improve when the current 'single soil test approach' changes to a more complex approach, which includes different soil tests, soil properties and environmental conditions.