Increasing the Homogeneity of Soil Map Units Using the Level of Landform Phase in the Geopedologic Approach

Delineating landscape into homogenous units is fundamental to managing resources and delivering sustainable development. The importance of this has long been recognized as a critical issue in various studies and it has been examined from different aspects. In soil mapping, the geopedologic approach is used for landscape classification, which was defined by Zinck (1989). This approach differentiates landscapes into landforms to increase the purity of soil map units. Therefore, the aim of this study was preparing geopedologic maps of the study area on the level of landform phases intending to make more homogeneous soil units. Honam sub-basin in Lorestan province is one of the most important agricultural areas in the Karkheh River watershed. Soil moisture and temperature regimes of the area were Xeric and Mesic, respectively. After a primary interpretation, a geopedology map of the study area at the landform level was prepared according to the geopedologic approach. After soil surveying, 31 profiles were excavated, described, and sampled in the largest delineation of this map. Ultimately, this landform unit was differentiated to the landform phase units using morphometric features and normalized difference vegetation index. Pedodiversity index was computed for each landform phase unit to investigate the credibility of the geopedological approach for this unit. The conditional probability of each soil family was also measured in each landform phase unit to compare statistical differences between landform phase units. Furthermore, statistical comparisons were made between the Shannon indices of each unit. The soils of the study area were classified into seven soil families according to Soil Survey Staff (2014). Based on the geopedology map, this area includes two landscapes of Piedmont plains and valleys. These two landscapes were differentiated to 6 and 3 relief/molding, respectively. Geologically, the study area has 5 lithologic units. Finally, 22 landform units were identified in this area. The area of the widest landform with the code of Pi461 was 1223.35 ha. With individual use of NDVI, TRI, and aspect map, this landform unit was differentiated into 3 phases, and with the use of these 3 maps collectively, 11 phases were differentiated. The results showed landform map unit of Pi641 with 31 soil profiles and 7 soil families has the highest value of diversity indices, such as 1.59 for the Shannon index. In addition, this map unit is a compound map unit consisting of several soils, where the highest probability of observing soils is related to soils A and B with 32.5% probability. By differentiating this landform unit into phases, the Pi461 map unit is separated into smaller units that are more homogeneous. For example, when it is separated according to the vegetation characteristics, the three phases Pi4611 (N), Pi4612 (N), and Pi4613 (N) were differentiated that have medium, low, and high vegetation, respectively. In this case, Pi4612 (N) map unit with 75% probability of soil C observation and Pi4613 (N) map unit with 87.50% probability of soil B observation are two homogeneous map units. The Shannon index of these two units is 0.56 and 0.37, respectively, which indicates the purity of these map units. The results also showed that diversity indices and conditional probabilities, respectively, were decreased and increased in most of the landform phase map unit compared to the landform map unit. The use of normalized difference vegetation index compared to other environmental features has been effective in separating the landform phase units and preparation of homogeneous map units. So, the most probability of observing the dominant soils of the study area increased from 32.25% in the landform unit to 52.63, 75.75, and 87.50% in the landform phase unit, and the Shannon index decreased from 1.59 in the landform unit to 1.36, 0.56, and 0.37 in the landform phase units. The use of other environmental features to increase the purity of the landform phase map unit is suggested in future studies. Results of using geopedological approach at landform level in the study area showed that this level is useful at highest levels of soil classification (from order to great group), but due to the complex nature of soils at lower levels of classification (family and soil series) does not have enough efficiency. Therefore, for improving the geopedology method accuracy and to present more uniform map units at lower levels of classification, landform phase maps were presented according to the environmental characteristics of the selected landform. The statistical comparisons between Shannon indices calculated for each map unit in the landform phase map showed a significant difference at the 90% probability level between most of the units, which indicates an increase in the purity of these units at the soil family level.