Variability within a soil mapping unit mapped at the soil type level in the Wanganui district

Publisher Website
Google Scholar
Abstract
The variability in a range of soil chemical properties is described from 32 profiles of a soil mapping unit mapped at the soil type level (scale 1:31, 608). The variation in the measured chemical properties is discussed in terms of 3 broad groups of properties showing increasing variability. Mean profile CV [coefficient of variation] values were 17% (range 12-19%) for the least variable Group I-type properties such as total Ca and total K; 25% (21-31%) for intermediate Group II-type properties such as total N, CEC [cation exchange capacity], TEB [total exchangeable bases], percentage base saturation and P retention; and 64% (36%-111%) for the most variable Group III-type properties such as exchangeable Ca, Mg, K, Na, and 0.5 M H2SO4-P. This observed level of variation is broadly similar to that found in similar overseas studies. It results mainly from the inherent natural variation within the soil mapping unit together with the effects of management. A relationship was established between mean profile CV and the number of profiles having depth functions for individual chemical properties within given constraints. Using this relationship it was shown that for Group II-type properties, which are perhaps the most important for characterization of soil survey samples, only about 5% of the profiles within the mapping unit have depth functions falling within .+-. 10% of the mean profile depth function, but about two-thirds will have depth functions falling within .+-. 25% of the mean. The maximum number of samples required to obtain a sample mean for chemical properties of the Group II-type within .+-. 15% of the total population mean is approximately 20 at the 95% confidence level and 10 at the 85% confidence level. For Group III-type properties, the corresponding numbers of samples are approximately 250 and 140. The concept of presenting results according to a rating system rather than as individual analyses was considered. In a 5-class rating scale the ratings given to most soils would probably cover 2 classes.