Spatial and Temporal Variability of Corn Grain Yield on a Hillslope

the relationships between topography and crop yield. Graveel et al. (1989) observed that corn yields were The purpose of this study was to relate the temporal and spatial more variable on steep slopes but found no significant variability of corn (Zea mays L.) grain yield on a Typic Fragiochrept soil on a hillslope to soil properties and topographic features. Corn differences for degree of slope in a study of 24 yr of grain yields were sampled from a field that measured 280 by 150 m corn yields in Tennessee. Miller et al. (1988) found no using a grid and five transects. One-hundred forty yield measurements correlations between slope percentage and wheat (Tritiwere taken on the grid (1983–1985) and 190 measurements on the cum aestivum L.) yields. Moulin et al. (1993) found that transects (1984 and 1985) from plots 5.3 m long and two corn rows wheat yields were lowest on more highly elevated knolls wide. Measurements of soil surface elevation, soil organic matter where soil erosion losses were greatest. The effect, how(OM), P, and K contents were also taken at the grid plot locations. ever, was modified by surface shape, whether convex These data were analyzed using the methods of spectral analysis. or concave. Yield spatial and temporal variability was strongly related to surface Many soil properties are spatially correlated (Gajem undulations and the value of surface curvature was found to be a et al., 1981; Burrough, 1993), i.e., close locations are useful parameter to quantify variations in topography. The intraannual differences in weather had the largest effect on grain yield at more likely to have similar soil parameters than remote locations where the magnitude of curvature was large. Where the locations. Spatial correlation of crop yields has also been magnitude of curvature was small, the correlations of yield for the reported. Early work demonstrated this correlation unwet (1984) and dry (1985) years were highly significant. Yields correder irrigated conditions (Bressler et al., 1981, 1982; Morlated with soil P, K, and OM only in the dry year, 1985. Yields in the koc et al., 1985). In later work, Miller et al. (1988) relatively dry years (1983 and 1985) correlated with depth to fragipan. reported a range of spatial correlation for wheat yield Elevation data helped us interpret the spatial and temporal variability of about 80 m in California. This value was similar for of grain yield by separation of areas with convex curvature from areas several soil properties, which included surface soil thickwith concave curvature. ness and sand content. Boyer et al. (1991) observed a strong spatial correlation between biomass production and soil pH on a hillslope. Bhatti et al. (1991) found I on the spatial variability and distribution that the spatial correlation of wheat yields was related to of crop yield can be used to tailor management pracorganic matter distributions that resulted from erosion. tices for specific locations. Yields can now be mapped Because soil properties and associated crop yields are in detail, and techniques are needed to interpret yield often spatially correlated, spatial analysis should be used maps in terms of soil variability and to develop sitewhen quantifying their variability. Miller et al. (1988) specific management practices based on that variability found that classical statistical techniques that did not (Mulla, 1991). account for spatial correlations were inadequate for The variability of soil properties is large in complex showing the effects of soil erosion on dryland wheat hills (Miller et al., 1988). Soil physical properties such yields. Methods adapted from time series analysis, minas clay content and distribution with depth, sand coning geostatistics, and studies of turbulence have been tent, and pH have been shown to be highly correlated applied as statistical tools to describe the interdepenwith landscape position (Ovalles and Collins, 1986). Sidence and structural arrangement of soil properties nai et al. (1981) showed that, in an arid region, moisture (Burgess and Webster, 1980; Kachanoski et al., 1985a,b). contents were highly correlated with the curvature of These methods have also been applied to a description the soil surface. Organic matter (Miller et al., 1988; of the dependence of crop yields on soil properties Bhatti et al., 1991) as well as water-holding capacity (Bresler et al., 1981, 1982; Boyer, et al., 1991). Bhatti (Hanna et al., 1982) have also been shown to vary by et al. (1991) suggested that easily measured covariates slope position. The depth of the Ap horizon, which is could be used to divide fields into management units. an important soil fertility factor, is often greater in Soil properties such as A horizon thickness or toposwales than on knolls (Miller et al., 1988). graphic properties such as curvature have been shown to Crop yields on hillslopes are affected by topography have a periodic component in their spatial distribution and the attendant differences in soil properties. Stone (Kachanoski et al., 1985a,b; Folorunso and Rolston, et al. (1985) reported correlations of crop yield with 1985). Spectral analysis was used in these studies to landscape position to be higher than with erosion class. quantify the periodic distribution of variance. It is difficult, however, to make generalizations about Weather interacts in a complex way with topography to affect crop yields because of the relationships among D.J. Timlin, USDA-ARS, Remote Sensing and Modeling Lab., Bldg. soil relief, root growth, and hydrologic regime. Ciha 007, Rm. 008, BARC-West, Beltsville, MD 20705; Ya. Pachepsky, (1984) demonstrated that winter wheat yields were most Dep. of Botany, Duke Univ., Durham, NC 27708; V.A. Snyder, Colclosely related to slope position. The relationship was lege of Agricultural Sciences, P.O. Box 21360, Rio Piedras, PR 00928; and R.B. Bryant, Dep. of Soil, Crop, and Atmospheric Sciences, Bradespecially noticeable during dry years and appeared to field Hall, Cornell Univ., Ithaca, NY 14853. Received 25 Oct. 1996. be related to the depth of the surface layer. In this *Corresponding author ([email protected]). context, surface curvature can be an informative characteristic of topographic relief. Kachanoski et al. (1985a) Published in Soil Sci. Soc. Am. J. 62:764–773 (1998).