Scaling Analysis of Irregularly Sampled Soil Properties Using Second Generation Wavelets.

Soil spatial variability is a result of the combined action of soil physical, chemical and biological processes operating in different intensities at different scales. An adequate understanding of soil spatial variability is necessary for developing logical, empirical and physical models of soil landscape processes, environmental management, precision agriculture, and soil quality assessment. Two aspects of this spatial variability restrict the application of general spatial analysis methods: 1) localization of different scale processes, and 2) nonstationarity. Wavelet analysis can deal with these by partitioning the sample variation into positions (or locations) and frequencies (or scales). The traditional wavelet analysis including the continuous and discrete wavelet transforms, requires regular sampling intervals in space or time. However, soil scientists are often restricted to short and irregularly spaced samples, or regularly sampled data with missing values. A new and adaptive ‘second generation’ wavelet can be used in this situation to delineate the scale and location dependent spatial variability. In this analysis, constructed wavelets adapt their shape to the data, such as sampling gaps and edges. The objective of this study is to analyze spatial variability of irregularly-sampled soil water storage using a second-generation wavelet. Soil water was measured along a transect over hummocky landscape at St Denis National Wildlife Area, Saskatchewan, Canada, using Neutron Moisture Meter. Though the transect was established for 128 regularly sampled points, water stagnation in the depression prevented soil water storage measurement making almost an irregularly sampled transect. Second generation wavelets will be used to elucidate the scale-location specific spatial variability of soil water storage and compared with the traditional wavelets.