N-15 IN SOIL RESEARCH - APPROPRIATE APPLICATION OF RATE ESTIMATION PROCEDURES

In soil research, N-15 has most often been used to quantify the fate and efficiency of applied fertilizers. Useful information has been gained from these studies, but the greater power of N-15 experiments, the ability to estimate gross rates, has been exploited less. Analytical solutions for differential equations describing mineralization and immobilization processes are an efficient means to obtain rate estimates, provided that certain important assumptions are met. Through the use of data from simulation models, we identified some of the conditions under which analytical solutions do, or do not, yield good estimates. These studies indicated that serious rate estimation errors are more likely to arise from an incorrect assumption about system structure than either (1) the use of incorrect kinetic order or (2) perturbations caused by the introduction of tracer atoms. Non-linear regression, coupled to numerical simulation, was explored as an alternative to analytical rate equations. Simulation is recommended as an aid to system identification, whereas non-linear regression is used to estimate rates and pool sizes, while providing an indication of statistical reliability. Non-linear regression is particularly useful in estimating rates of slower processes and difficult to measure pool sizes.