being generated. The field-specific characteristics of the soil test levels (i.e., the range of soil test levels and the Grid soil sampling is typically used for establishing management nature of the spatial variability) ultimately determine zones for site-specific application of nutrients. The geostatistical procedures used to estimate values between sample locations require sam-Autocorrelation can be used as a method of assessing lrs.uoguelph.ca).how well data points are related with separation distance. Published in Agron. J. 97:524-532 (2005).

High resolu on (hourly) soil water content and temperature data have been collected for nearly 10 yr within the 0-100 cm depth profi les under no-ll (NT) and conven onal fall llage (CT) prac ces. The data indicate signifi cant diff erences between the two llage pracces, especially during winter and early spring freeze\thaw cycles. Results indicate that shallow minimum soil temperatures were lower under CT than NT; however, average winter shallow soil temperatures were very similar between the two treatments. Soil freezing characteris c curves (SFC) measured in situ with soil water content and temperature data were analyzed for diff erences between treatments as the NT system matured. The SFC shapes for NT evolved over a 7-yr period as the age of the NT system increased. An intensive freeze-thaw episode showed strong hysteresis in SFC, a phenomenon not analyzed in detail before this study based on data collected in the fi eld.

Transport properties vary considerably over small distances in most soils. The stochastic streamtube model offers one approach to incorporating heterogeneity into transport predictions. This study tested the ability of the streamtube concept to predict transport in heterogeneous fields using measurements from undisturbed columns. Fifty undisturbed columns (0.15‐m diam. by 1.5 m deep) were taken every 0.4 m from a 20‐m‐long transect in a loamy sand soil with variable horizon thickness. Each core was instrumented at 0.1‐m intervals with time domain reflectometry probes to measure resident fluid concentrations of a conservative (Cl−) tracer under steady flow conditions. Large‐scale concentration curves of Cl− from solution samplers and coring were obtained from field experiments conducted on the same soil under similar boundary conditions. Differences were observed in the solute spread and mass recovery, but not in the centers of mass. Horizontal scale dependence of transport was observed in the field but not in the columns. This suggests that a higher dimensionality of transport, probably along the horizon interfaces, may be responsible for the observed scale dependence in the field. Although the stochastic streamtube model gave good predictions of the center of mass, it does not appear to be a realistic physical analogue for describing solute dispersion in soils with spatially variable layer thickness.

O'Halloran, I. P., von Bertoldi, A. P. and Peterson, S. 2004. Spatial variability of barley (Hordeum vulgare) and corn (Zea mays L.) yields, yield response to fertilizer N and soil N test levels. Can. J. Plant Sci. 84: 307-316. Identification of management units for the variable application of fertilizer N is a critical component for the implementation of a site-specific N management program. Field studies were conducted to examine the spatial variability of soil nitrate levels, spring barley (Hordeum vulgare) and corn (Zea mays L.) yields and yield responses to fertilizer N applications on two sites in southwestern Ontario, Canada. Soil sampling on a 3 × 10 m grid indicated that soil NO 3 -N test values had a log-normal distribution and varied considerably at both sites with CVs exceeding 57% on the untransformed data. Ranges of spatial correlation varied from 20 to 95 m with 30 to 80% of the total variance of the ln-transformed data existing as either random or unsampled variance, and these parameters were not temporally stable. Although NO 3 -N tended to increase at lower slope positions in two of the 3 site-years, considerable within-slope variability of soil NO 3 -N levels was also observed. Spatial variations in soil N test levels, crop yields and crop yield responses to applied fertilizer N were not strongly related to one another indicating that it would be unlikely that either soil N test level or yield would adequately delineate management zones for the variable application of N fertilizer at these sites. 307-316. La mise en oeuvre d'un programme de gestion du N adapté au site exige l'identification des unités dont on se servira pour modifier la quantité d'engrais azoté appliquée. Les auteurs ont effectué des essais sur le terrain pour vérifier la variabilité spatiale de la concentration de nitrate dans le sol, le rendement de l'orge (Hordeum vulgare L.) et du maïs (Zea mays L.) ainsi que de la réaction du rendement aux applications d'engrais N à deux endroits, dans le sud-ouest de l'Ontario (Canada). Après prélève-ment d'échantillons de sol dans un quadrillage de 3 × 10 m, on constate que la concentration de N-NO 3 relevée lors des essais suit une distribution logarithmique normale et varie considérablement aux deux endroits, le coefficient de variabilité dépassant 57 % pour les données non transformées. La corrélation spatiale varie de 20 à 95 m et 30 à 80 % de la variance globale des données après transformation logarithmique se retrouve sous forme de variance aléatoire ou non échantillonnée. En outre, les paramètres ne sont pas stables dans le temps. Bien que la concentration de N-NO 3 ait tendance à augmenter aux emplacements plus bas dans la pente deux années-site sur trois, on remarque aussi une variabilité considérable de ce facteur dans la pente même. La variation spatiale de la concentration de N dans le sol, le rendement des cultures et la réaction du rendement aux applications d'engrais azoté ne sont pas très corrélés, signe que la concentration de N dans le sol ou le rendement s'avéreraien...