Modelling loess landscapes for the South Island, New Zealand, based on expert knowledge

Abstract

In New Zealand, occurrence of loess often determines the spatial pattern of soil depth, and influences droughtiness, leaching potential, organic matter accumulation, nutrient retention, and natural plant-species distribution. Understanding loess distribution is therefore a major prerequisite for soil and land resource management. Although New Zealand's soil scientists have accumulated a good empirical knowledge of loess distribution through several decades of field investigation, only some of this knowledge is recorded in papers and reports. This study estimates loess thickness and percent cover, and provides loess landscape models for the internal loess distribution of all land units in the South Island based on expert knowledge. We derived loess depth classes and percent cover classes and assembled land units with similar loess distribution patterns. The soil sets underpinning the map units of the New Zealand Land Resource Inventory (NZLRI) were classified according to loess depth, loess cover, and loess pattern. New loess maps of the South Island were produced from those classifications, displaying loess coverage, thickness, loess pattern, and loess landscapes. These maps present our current knowledge of the coarse-scale loess distribution and provide a framework for fine-scale loess landscape modelling.

Automatic landform stratification and environmental correlation for modelling loess landscapes in North Otago, South Island, New Zealand

“…Loess textural variation is consistent with downwind fining from the source area (Young, 1964;Ruhe, 1969;Mason et al, 1999), with resultant influences on soil morphology: well-drained fine sandy Laminar Pallic Soils (Haplustalfs) are found adjacent to the Waitaki River while Perch-Gley Pallic Soils (Inceptisols) further south are comprised mainly of silts with fragic and reductimorphic features. Soil maps and conceptual soil-landscape models have been used to infer loess distribution (Mason et al, 1999;Schmidt et al, 2005a), and the 1:50,000-scale soil map of the study area and associated soillandscape models of Wilson (1970) may be used in this way. Wilson (1970) divided North Otago into distinct physiographic regions defined according to their lithology and pattern of landform components (e.g.…”

“…Loess maps derived from these models were more accurate and had greater resolution than earlier maps, and they provided a more detailed description, including uncertainties, of extent, depth, and pattern of loess distribution within loess-landscapes. Schmidt et al (2005a) concluded, however, that these models needed to be refined and tested by higher-resolution studies in selected areas, with quantitative models complemented by intensive field data.…”

“…However, due to variability in loess sources, transport paths and local potentials for deposition and erosion, the spatial distribution of loess is highly complex (Ives, 1972) and models of the spatial distribution of loess tend to be conceptual (Ives, 1972;Mason et al, 1999). Schmidt et al (2005a) recently attempted to map South Island loess by defining quantitatively coarse-scale loess-landscape models, the internal loess pattern of which was explained by a conceptual loess-landscape model developed from expert knowledge. Loess maps derived from these models were more accurate and had greater resolution than earlier maps, and they provided a more detailed description, including uncertainties, of extent, depth, and pattern of loess distribution within loess-landscapes.…”

“…The spatial distribution and thickness of Quaternary loess deposits strongly influences soil properties and other environmental subsystems on the plains and lowlands of New Zealand's South Island, and therefore modelling loess distribution is a prerequisite for soil and land resource management (Schmidt et al, 2005a). Local and regional models of loess generation and distribution have been developed (see Eden and Hammond, 2003), and spatial information on loess distribution has been captured in detailed regolith studies, as part of soil-landscape studies and as part of investigations into large-scale patterns (Raeside, 1964;Ives, 1972;Bruce, 1973;McCraw, 1975).…”

“…The objectives of this study are to 1) test the validity of the model of Schmidt et al (2005a) to identify those parts of the North Otago, South Island, landscape where the model performs well and where it performs poorly, 2) develop loess distribution models using environmental correlation based on field data, and using a landform stratification approach based on a 1:50,000 soil map and associated conceptual models, 3) validate and test the performance of the models against at dataset of field observations of loess distribution.…”