Premium
Spatial Analysis of Soil Fertility for Site‐Specific Crop Management
Author(s) -
Cahn M. D.,
Hummel J. W.,
Brouer B. H.
Publication year - 1994
Publication title -
soil science society of america journal
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.836
H-Index - 168
eISSN - 1435-0661
pISSN - 0361-5995
DOI - 10.2136/sssaj1994.03615995005800040035x
Subject(s) - semivariance , spatial variability , precision agriculture , spatial ecology , environmental science , soil fertility , soil water , spatial analysis , geostatistics , soil test , sampling (signal processing) , soil science , agronomy , mathematics , statistics , geography , agriculture , ecology , archaeology , filter (signal processing) , computer science , computer vision , biology
Spatial patterns of soil properties and nutrient concentrations need to be characterized to develop site‐specific farming practices that match agricultural inputs with regional crop needs. The spatial variation of soil organic C (SOC), soil water content (SWC), NO 3 ‐N, PO 4 ‐P, and K were evaluated in the 0‐ to 15‐cm layer of a 3.3‐ha field (Typic Haplaquoll and Argiaquic Argialboll) cropped with maize ( Zea mays L.) and soybean [ Glycine max (L.) Merr.]. The range of spatial correlation was determined from semivariance analyses of the data and was found to vary among and within fertility parameters. Nitrate had the shortest correlation range (<5 m) and SOC had the longest (>180 m), whereas SWC, PO 4 ‐P, and K had intermediate spatial correlation ranges. In addition, SOC was found to have small‐scale spatial variation nested within large‐scale spatial variation. The spatial pattern of NO 3 ‐N changed with time. Frequency distributions of SOC and SWC were close to normal, whereas the distributions of NO 3 ‐N, K, and PO 4 ‐P data were skewed. Median polishing detrending and trimming of outlying data were useful methods to remove the effects of nonstationarity and non‐normality from the semivariance analysis. The results suggest that reducing sampling intervals from 50 to 1 m would reduce the variance of SWC, SOC, NO 3 ‐N, PO 4 ‐P, and K estimates by 74, 95, 25, 64, and 58%, respectively. A useful sampling pattern for characterizing the spatial variation of several soil properties‐nutrients and scales should be random with sample spacing as close as 1 m and as far apart as the longest dimension of the field.