Soil texture (% sand, silt and clay sized particles) is one of the most important of soil characteristics affecting the function of soils. To better understand the behaviour of soils, reliable spatial estimates of soil texture need to be available. Digital soil mapping has been an enabler in delivering this sort of information. Delivered as two connected pieces, we present new efforts to update the soil texture maps for Australia (Version 1 was delivered in 2015). The main distinguishing enhancement is the merging of field descriptions of soil texture with the traditional laboratory analysed data. This greatly increases the number of available data, yet also calls for an elaboration of methods of how to convert texture class data into continuous variables, how to deal with the associated uncertainties of these conversions, and how these can be propagated in any sort of spatial modelling. Here we report on research to re-calibrate the soil texture centroids that were first determined by Minasny et al. (2007). Then we describe our approach on how the centroids and their uncertainty can be used to generate acceptable soil texture fractions for all qualitive soil profile texture descriptions in the Australian soil database.

土壤质地（％的沙子，粉砂和黏土大小的颗粒）是影响土壤功能的最重要的土壤特征之一。为了更好地了解土壤的行为，需要可靠的土壤质地空间估计。数字土壤制图已成为提供此类信息的促成因素。作为两个连接的部分交付，我们提出了新的工作来更新澳大利亚的土壤质地图（版本1已于2015年交付）。主要的显着增强是将土壤质地的野外描述与传统的实验室分析数据相结合。这大大增加了可用数据的数量，而且还要求详细说明如何将纹理类数据转换为连续变量，如何处理这些转换的相关不确定性的方法，以及如何在任何类型的空间建模中传播它们。在这里，我们报告了有关重新校准Minasny最初确定的土壤质地质心的研究等。（2007）。然后，我们描述如何针对澳大利亚土壤数据库中所有定性土壤剖面质地描述如何使用质心及其不确定性生成可接受的土壤质地分数的方法。