Influence of Spatial Variability of Soil Potassium and Nonuniformity of Fertilizer Application on Crop Response

Spatial variability of soil properties important in plant growth is common in soils of the tropics and subtropics. The description of this variation is the first step towards understanding and correcting the associated detrimental effects. Geostatistical methods have been used to characterize inherent soil variation, yet the variation resulting from fertilizer or amendment addition can be equally important. Geostatistical techniques have been developed to estimate proportions of a region above or below selected concentrations of elements. These methods have seldom been applied to the description of or the modification of nutrient concentrations in soils but hold promise in dealing with spatial variability of soil properties with inherent variation or that added by fertilizer application. Conventional statistical analysis and geostatistical analysis were used to investigate the field spatial variability of selected soil properties. Comparison of coefficients of variation (CV) indicated that exchangeable K was the least variable (CV= 12.6) and exchangeable Na the most variable (CV= 34.4). Semi-variograms of exchangeable Ca, Mg, K, and Na revealed strong spatial dependence. Large nugget variances accounted for 41-52% of the total variances. Ranges of spatial dependence were about 17 m for Kand Na and more than 20 m for Ca and Mg. Kriged estimates were used to draw contour maps of selected soil properties. The structure of spatial dependence of soil K was used to estimate the proportion of a given experimental plot below a specified threshold of exchangeable K. Different patterns of fertilizer distribution were simulated in a field study, and a measure of uniformity (UCF = 1 - CV) of fertilizer distribution was calculated from the CV of K application. The results indicated that fertilizer rates necessary for 90-95% relative yield could be predicted from the response of uniform application and the CV of the actual application. Variable rates of fertilization could then be applied if the deficient zones could be identified in the field. The response of Chinese cabbage to potassium application in the field experiment was strongly affected by the pattern of fertilizer distribution. There was a 9.52% decrease in maximum yield with increased unevenness of fertilizer distribution. The amount of K associated with 95% maximum yield was 97, 113, and 444 kg K/ha for UCF values of 1.0, 0.42, and -0.02, respectively. The corresponding tissue K concentrations associated with maximum yield were 4.00, 4.15, and 4.30%. The decrease in yield due to nonuniform fertilizer application was quantitatively expressed through the fluctuation response index (FRI) i.e., the product of the variance of fertilizer application and the second derivative of the yield-fertilizer function obtained under uniform fertilizer application. The fluctuation response index holds promise as a way to estimate crop loss due to excessive variability, and therefore financial loss to the farmer. The farmer may then have the opportunity to control the application so that losses are kept within acceptable limits. Nonuniform fertilizer application significantly affected the critical level of exchangeable K for Chinese cabbage. Increased variation in exchangeable K from CV= 44.1% to CV= 96.6% resulted in critical levels of soil K of 0.28 and 0.73 cmol(+)/kg, respectively. The response of maize to residual K was closely related to the degree of spatial variability of residual K. Increased variation of residual K from CV= 37.0% to CV= 48.5% resulted in yield losses and in increased level of exchangeable K associated with 90% maximum yield from 0.16 to 0.24 cmol(+)/kg, respectively. There was a gradual decrease in the correlation coefficient between corn yield and ear-leaf K with increasing spatial variability of residual K. The potassium retention capacity of the Puaulu soil was investigated. The results indicated that the Puaulu soil has a low affinity for K as evidenced by the very low Gapon selectivity coefficient (KG= 0.15 (mol-1L) 1/2). The topsoil had significantly lower retention capacity than the subsoil