An Improved Model of Nitrogen Release for Surface‐applied Controlled‐release Fertilizer

Abstract

Empirical temperature models are commonly used to estimate N release from polymer‐coated controlled‐release fertilizers (CRFs) under field conditions where the influence of soil moisture is negligible. For surface‐applied CRF in bare‐root nurseries, the effect of soil moisture may be important. To quantify the effect of soil moisture on N release from surface‐applied CRF, two experiments were conducted in a sandy‐textured soil: (i) a laboratory experiment to develop two semiempirical models that characterized the effect of soil temperature (T model) and moisture (MT model) on N release, and (ii) a field experiment to evaluate the precision of these models. In the laboratory experiment, the N release from CRF (urea, 41–0–0, ∼2‐mo release) was measured under 16 combinations of four soil temperatures (5, 10, 23, and 35°C) and four soil moisture contents (0.002, 0.04, 0.11, 0.22 kg kg−1). The calculated time constants for each abiotic combination showed that both soil temperature and soil moisture content controlled the cumulative N release (CNR). The difference between model estimates and field measurements of CNR indicated that the T model significantly underestimated CNR by as much as 18% (P< 0.01). In contrast, this difference was reduced to 8% using the MT model; the slope of the least‐square fit for the MT model was not significantly different from unity at the 1% level and just significant at 5% (P= 0.052). Therefore, we argue that a model parameterized by both soil temperature and soil moisture will improve estimates of CNR from surface‐applied CRF in irrigated tree nurseries.