A comparison between three legacy soil maps of Zambia at national scale: The spatial patterns of legend units and their relation to soil properties

Abstract

We examined three soil maps of Zambia, two published at scales of 1:1 million – the Exploratory Soil Map of Zambia (ESMZ) and the Vegetation–Soil Map produced by Trapnell and colleagues in 1947 – and one at 1:3 million, the Soil Atlas of Africa (SAA). We estimated components of variance for measurements of clay, sand and organic carbon content and bulk density of the soil across the country using models which included different mean values for soil map units as random effects. For all but organic carbon content there was significant variation accounted for by differences between legend units for two of the maps, ESMZ with legend units based on the FAO-Unesco and SAA with legend units based on the World Reference Base respectively. This was despite their small cartographic scale. For the Vegetation–Soil Map, we examined differences between broad soil physiographic units. These did not account for significant variation in the soil properties. There were clear similarities between the soil physiographic units of the Soil–Vegetation Map and broader physiographic units into which the legend units of the ESMZ are grouped. The spatial pattern of soil units of the SAA was the most spatially heterogeneous, as measured by the sum of indicator variograms, despite being at the smallest published scale. It was apparent that some of the soil variation within the largest physiographic unit of the Soil–Vegetation Map, the Plateau Soils, as expressed by the map units of the SAA was significantly associated with the different vegetation units mapped in 1947. These studies show how quantitative assessment of legacy soil information may help us understand its potential and limitations.