Soil map update: Procedure and problems encountered for the island of Réunion

Many soil maps were drawn up after World War II with different soil classifications that have significantly evolved since. Updating such old maps with a new version or a new classification system is always complex: (i) we do not always possess all the original information; (ii) the criteria for determining references are often different, and (iii) on the most accurate scales, correlations come up against the complexity and specificities of each classification system. On Reunion, a volcanic tropical island in the Indian Ocean, we undertook a comprehensive overview of the old existing soil studies. This article describes (i) the procedure used to update the soil maps and the toposequence acquired with the old French Commission de Pédologie et de Cartographie des Sols (CPCS) classification system, without any new information, using the World Reference Base for soil resources (WRB); (ii) the construction of a new soil map drawn up with completely new information, and (iii) a comparison of these two approaches. At elevations below 350m asl (above sea level), without any new pedological information, we updated Brown ferruginous soils, Reddish-brown ferrallitic soils, and Fersialitic soils into Haplic Nitisols (Humic, Eutric). The acquisition of new data showed that this update was incorrect because not all the diagnostic criteria of the Nitic horizons were met. The correct diagnostic horizons were a Mollic horizon when the thickness was 25cm or more, or a Cambic horizon. Leptic Phaeozems and Leptic Cambisols were then the correct Reference Soil Group (RSG). At elevations from 350 to 900m asl, without any new information, Brown and Reddish-brown ferrallitic soils, Andic ferrallitic soils, and Brown and Andic brown soils were updated into Haplic Nitisols (Humic, Dystric) and Andic Umbrisols (Humic). The acquisition of new data showed that this update was incorrect because Andic properties and the diagnostic criteria of the Nitic horizons were not met. Over 900m asl, Pozols were correctly updated, as were the Andosols except from 900 to 1050m asl where not all the Andic properties were met. Without any new information, incorrect updates were observed for both the determination of RSG and the qualifiers. Despite the field descriptions, the lack of any analytical determinations on the old soil studies was a source of updating errors for the more developed soils formerly qualified as ferrallitic. In order to update limits for Andic properties and Andosols, the systematic use of analytical determinations has to be considered for updating old soil maps, as the diagnostic criteria are more restrictive than in the past.