Soil classification, map legends and GIS logic: experiences from converting the legend of the soil map of Belgium into WRB | Classification de sols, légende de cartes et logique SIG : expérience suite aux travaux de conversion de la légende de la Carte des Sols de la Belgique dans le système international WRB

A systematic soil survey of Belgium was conducted from 1948 to 1991. The purpose wasto have a soil information base adequate for boosting agricultural production after theSecond World War. Field surveys were done at the detailed scale of 1:5000, while themaps were published at a 1:20,000 scale. To enable soil surveyors to identify soils in thefield, an original soil classification was developed based on readily observable physicaland morphogenetic characteristics. Within the European Union, the World Reference Base(WRB) for Soil Resources (IUSS Working Group WRB, 2007), has been adopted as thecommon classification system. As soil surveys in most European countries wereconducted independently, the challenge now is to convert the national legends into acommon WRB legend. In Belgium, such a conversion is being implemented and wepresent some lessons learned in relation to classification and map legends.The legend of the soil map of Belgium is based primarily on soil texture, drainage statusand profile development. “Soil series” are defined in an open and non-hierarchicalstructure by combining these three categorical variables, and to which modifiers can be added such as parent material, stoniness or depth to a substratum. The WRB-2007classification is based on diagnostic features defined by morphological, physical andchemical properties. The conversion of the legend of the soil map of Belgium to WRB isbased on insights gained from classifying, so far, more than 360 legacy soil profiles incombination with field observations. From these insights, heuristic rules have beendeduced regrouping soil series into Reference Soil Groups (RSG) and for which somequalifiers could also be identified. Other qualifiers were determined by relying ondatabases of the legacy soil profiles. To take regional variability into account, theconversion is done by physiographic zone.Converting the legend of the soil map of Belgium to WRB actually leads to a regroupingof soil series into broader WRB categories and which can neatly be represented on1:50,000 scale maps. Hence, it does not imply substituting one classification with anotherone. Users, who would need the detailed information, can still refer to the detailed symbolsof the soil series. The regrouping of soil series has been made possible thanks to theflexibility of WRB for combining various qualifiers. However, as the WRB-2007classification leads to a varying number of qualifiers it is less practical for constructingmap legends. Therefore, the WRB-2010 guidelines (IUSS Working GroupWRB, 2010) propose to organise qualifiers in main and optional ones with priority rulesfor the main qualifiers. As illustrated in Table 1, this approach may highlight, or hide,some qualifiers inconsistently. When for example only two qualifiers are retained in amap legend, the qualifier Endogleyic of stony, poorly drained Albic Podzols will not beindicated, while it will be indicated when such soils are not stony (Table 1). Moreover, theproposed hierarchy is sometimes also inconsistent when compared across Reference SoilGroups (RSG) as illustrated for the Arenosols and Regosols (Table 1). These drawbackscould be avoided if qualifiers would be organised into thematic groups ― such as profiledevelopment, texture, drainage, depth of substratum and fertility ― rather than by rankingthem. Grouping qualifiers thematically would have the advantage to give more flexibilityto the map user working with GIS, and indeed, it would render WRB a more “GIS logic”classification system.