Showing 4 results for Digital Soil Mapping
M. Bagheri Bodaghabadi, M. H. Saleh, I. Esfandiarpoor Borujeni, J. Mohammadi, A. Karimi Karouyeh, N. Toomanian,
Volume 16, Issue 61 (10-2012)
Abstract
Discrete Models of Spatial Variability (DMSV) have limitations for soil identification in traditional soil maps. New approaches, generally called digital soil mapping (DSM), using continuous methods (CMSV), try to predict soil classes or soil properties based on easily-available environmental variables. The objective of this study was to map the soil classes of the Borujen area, Chaharmahal-va-Bakhtiari province, using digital elevation model (DEM) and its attributes and Soil-Land Inference Model (SoLIM). To do this, eighteen terrain attributes were derived from the DEM of the area. The primary analysis showed seven attributes are the most important derivatives. These derivatives as well as three dominant soil subgroups and seven soil families of the region (41 profiles from 125 profiles) were used to construct the input data matrix of the model. Then, output fuzzy soil maps of SoLIM were converted to polygonal soil map, using ArcGIS. Results showed that different combinations of DEM attributes have different accuracy rates for soil prediction. The accuracy of the interpolation was twice that of the extrapolation. Although SoLIM had an acceptable accuracy for soil nomination, and identification of soil map units’ types, it did not have enough accuracy for the location of soil classes. It seems that using other data like parent material and geomorphic surface maps will increase the accuracy of the model prediction.
S. Ayoubi, R. Taghizadeh, Z. Namazi, A. Zolfaghari, F. Roustaee Sadrabadi,
Volume 20, Issue 76 (8-2016)
Abstract
Digital soil mapping techniques which incorporate the digital auxiliary environmental data to field observation data using software are more reliable and efficient compared to conventional surveys. Therefore, this study has been conducted to use k- Nearest Neighbors (k-NN) and artificial neural network (ANN) to predict spatial variability of soil salinity in Ardekan district in an area of 700 km2, in Yazd province. In this study, 180 soil samples were collected in a grid sampling manner and then soil chemical and physical properties were measured in laboratory. Environmental auxiliary variables were included topographic attributes, remote sensing data (ETM+) and apparent electrical conductivity (ECa). The result of the study showed that the K-mean nearest neighborhood had higher accuracy than ANN models for predicting soil electrical conductivity (ECe). Overall, k-NN models could provide significant relationships between soil salinity data and environmental auxiliary variables. The k-NN model had the root mean square and coefficient of determination of 12.10 and 0.92, respectively, between predicted and observed ECe data. Also, apparent EC, and remotely sensed indices and wetness index were identified as the most important factors for predicating the soil salinity in the studied area.
A. Fariabi, H. Matinfar,
Volume 22, Issue 3 (11-2018)
Abstract
One of the problems with the traditional mapping of soils is the expert’s opinion, it time-consuming and timely preparation, and the updating of the maps. While digital soil mapping, using different soil-earth models leads to the simplification of the complexity of the soil system. The purpose of this study was to investigate Soil-Environment Inference (SIE) in soil mapping with an emphasis on using the expert knowledge and fuzzy logic. For this purpose, the digital layer of geology and peripheral layers were derived from a digital elevation model including elevation, slope, and curvature of the ground surface, and auxiliary index, which comprised the input data of the SIE model. Then, the fuzzy maps prepared for the five soil types and the final map of soil prediction were created by hardening. The results showed that the SIE model, which used environmental variables, had a high ability to isolate soil types with more detailed compositions of soils with different maternal materials. The comparison of the error matrix showed that the overall accuracy of the derived map of the SIE model was equal to 75%, and the matching of the digital mapping results with conventional mapping accounted for 74.71% of the results. The difference in the compliance rate could be attributed to the difference in the nature of the two methods.
H. R. Matinfar, Z. Mghsodi, S. R. Mossavi, M. Jalali,
Volume 24, Issue 4 (2-2021)
Abstract
Knowledge about the spatial distribution of soil organic carbon (SOC) is one of the practical tools in determining sustainable land management strategies. During the last two decades, the utilization of data mining approaches in spatial modeling of SOC using machine learning algorithms have been widely taken into consideration. The essential step in applying these methods is to determine the environmental predictors of SOC optimally. This research was carried out for modeling and digital mapping of surface SOC aided by soil properties ie., silt, clay, sand, calcium carbonate equivalent percentage, mean weight diameter (MWD) of aggregate, and pH by machine learning methods. In order to evaluate the accuracy of random forest (RF), cubist, partial least squares regression, multivariate linear regression, and ordinary kriging models for predicting surface SOC in 141 selected samples from 0-30 cm in 680 hectares of agricultural land in Khorramabad plain. The sensitivity analysis showed that silt (%), calcium carbonate equivalent, and MWD are the most important driving factors on spatial variability of SOC, respectively. Also, the comparison of different SOC prediction models, demonstrated that the RF model with a coefficient of determination (R2) and root mean square error (RMSE) of 0.75 and 0.25%, respectively, had the best performance rather than other models in the study area. Generally, nonlinear models rather than linear ones showed higher accuracy in modeling the spatial variability of SOC.
