Soil nitrogen mineralization during laboratory incubation: Dynamics and model fitting

Soil nitrogen mineralization kinetics were studied for eight treatments of two soils in an aerobic long-term (30 wk) incubation experiment. Soil mineral-N (NH4+ and NO3−) in the leachates was measured every week during the first 9 wk and every 2 or 3 wk thereafter. The NH4+ percentage of the mineral-N ranged between 85 and 99% for all treatments in the first week of incubation and remained high (> 80%) in several treatments until the end of wk 4. Starting at wk 7, NH4+ concentrations were negligible in all treatments. The net N mineralization rate was 15–24 mg N kg−1 wk−1 during the first 4–6 wk and 2–5 mg N kg−1 wk−1 from wk 8 until the end of the incubation. Four models, (i) a one-component, first-order exponential model (the single model), (ii) a two-component, first-order exponential model (the double model), (iii) a one-component, first-order exponential model including a constant term (the special model), and (iv) a hyperbolic model, were fit to the cumulative mineral-N vs time data using a non-linear regression procedure. The goodness of fit of the four models depended on the duration of incubation. With 30 wk data, the double and special models were significantly better than the other two models; with the first 15 wk data, the four models had essentially the same goodness of fit for seven out of eight treatments. The values of the regression parameters derived from each model also depended on the incubation duration. Results from this study show that the pool size and mineralization rate parameters in the different models are merely mathematically-defined quantities obtained from the kinetic analysis of the net N mineralization and do not represent any rigorously-defined pool sizes of potentially-mineralizable N and their mineralization rate constants in the soils.