What artificial urine composition is adequate for simulating soil N2O fluxes and mineral N dynamics?

Artificial urine, an aqueous solution of various nitrogenous compounds and salts, is routinely used in soil incubation studies on nitrous oxide (N2O) emissions and related nitrogen (N) and pH dynamics. There is, however, no consensus on artificial urine composition, and a wide variety of compositions are used. The aim of this study was to test which artificial urine composition is adequate for simulating N2O fluxes, respiration, soil mineral N and pH dynamics of real cattle urine in both short- and long-term incubation studies. Urine solutions of differing compositions were applied to a sandy soil and incubated for 65 days, and results of measurements on N2O fluxes and soil mineral N were analyzed over the first 5 days as well as over the whole incubation period. Results from two real cattle urines with known nitrogenous composition (R1 and R2) were compared with three artificial urine types: (i) urea + glycine (AG), (ii) urea + hippuric acid (AH) and (iii) an artificial urine identical to the nitrogenous composition of real urine R1 (AR). During the first 5 days, only cumulative N2O emissions for AG deviated significantly (P = 0.02) from the N2O emissions for real urines, with 0.4% of applied N emitted compared with 0.0% and 0.1% for R1 and R2, respectively. Respiration from R1 was significantly (P < 0.001) higher than from R-2 and all artificial urines. Over the whole incubation period, no significant differences could be detected for N2O emissions or respiration with urine type. From all artificial urine types, AH yielded N2O emissions closest to those from real urine. AG deviated most from real urine, both in short- and long-term incubations. Over the whole period, soil NH4+ was higher for all artificial urines (P < 0.001) and pH-KCl was lower for AG and AR (P = 0.004) than for the real urines. AH was not significantly different from real urine R2 with respect to all measured properties except soil NH4+. We conclude that only AG did not adequately simulate N2O emissions, and that glycine is therefore not an appropriate substitution for hippuric acid in artificial urine. For future studies using artificial urine we recommend therefore a mixture containing at least urea and hippuric acid as sources of N. As no artificial urine composition resembled real urine with respect to all measured variables, even when nitrogenous composition was identical (AR), we recommend the use of real urines whenever possible. (c) 2006 Elsevier Ltd. All rights reserved.