Estimation of an active fraction of soil nitrogen

Publisher Website
Google Scholar
Abstract
Recent simulation models of soil organic matter and N transformations postulate several pools of soil N with differing stabilities. A nonlinear regression procedure was used with published data (Stanford and Smith for 34 soil samples to estimate the size of a pool (N1) which mineralized within four weeks, and the pool size (N1) and first order mineralization rate constant (k2) of a pool which mineralizes more slowly and which is similar to the “active”; N fraction described by Parton et al. Regression equations including all 34 samples based on total soil nitrogen, organic carbon, pH, and dummy variables for soil order, temperature and moisture regimes, texture, and presence of free calcium carbonate accounted for 55%, 86%, and 53% of the variation in N1, N2, and k2, respectively. Several simulation models of soil organic matter transformations have been developed recently (Parton et al.4, Jenkinson and Rayner5, Paul and Van Veen6. These models assume that soil organic matter exists in pools differing in their rates of decomposition. Jenkinson and Rayner5 and Paul and Van Veen6 use first order equations to simulate the rate of soil organic matter decomposition. The rate constants used in these equations are not sensitive to soil moisture or temperature, and simulated organic matter decomposition corresponds to average temperature and moisture conditions where the models were developed. The model developed by Parton et al4. simulates both soil C and N transformations, and actual decomposition rates are functions of soil temperature and moisture. All three models postulate one or more pools of non‐residue organic matter with decomposition rate constants ranging from 0.001 d‐1 to 0.02 d‐1. For example, Parton et al4. describe an “active fraction”; of soil C and N with a first order rate constant for decomposition under ideal temperature and normal moisture conditions of 2.0 y‐1 (0.0055 d‐1 ). The size of the active fraction probably depends on prior cultivation history and management. For a virgin grassland near Sidney, Montana, the initial size of the fraction was estimated to be about 17% of the total N in the topsoil (W. J. Parton, personal communication). Each of the models discussed above was developed for a rather narrow range of soils, and none describes an independent method for estimating the initial sizes of the various pools for a wide range of soi1s. Stanford and Smith3 used 39 widely different but agriculturally important soils from the continental U.S. to test an incubation technique for estimating “potentially mineralizable”; soil N (No). That study contains data which can be used to study differences among soils in mineralization of soil N under near‐ideal conditions (Jones et al.7). In the study, thirty‐week incubations were conducted at near‐optimal moisture, temperature, and aeration; and net mineralization was measured by leaching nitrate from the samples at intervals during the incubation period. Stanford and Smith3 described the mineralization of N during the period of incubation with the exponential equation where Nt is the cumulative amount of inorganic N released in time t and k is a time‐invariant rate constant. Both Molina et al.8 and Talpaz et al.9 have suggested that nonlinear regression analysis be used to estimate the parameters of the exponential equations describing the release of inorganic N during the incubation. Molina et al.8 also pointed out that this exponential equation does not account for the mineralization of large amounts of N during the first four weeks of incubation, and they suggest that a more appropriate model is one with two pools of mineralizable N, one which decomposes within the first four weeks of incubation (Birch 10,11) and another which decomposes more slowly. This paper describes a method by which the results of Stanford and Smith3 were reanalyzed to obtain independent estimates of the pool sizes of a rapidly mineralizing N pool responsible for the initial flush of mineralization (N1), a more slowly mineralizing N pool similar to the active fraction described by Parton et al.4 (N2), and the first order rate constant describing the mineralization of N2 under near‐ideal temperature and moisture conditions (kp). Multiple regression analyses were then used to estimate N1, N2, and k2 from soil chemical and taxonomic characteristics.