Mendeley Data
Description
This dataset contains soil spectral reflectance and associated physico-chemical and hydraulic properties of 130 soil samples collected from five provinces in Iran. The data were used to estimate key soil water retention curve (SWRC) parameters, including residual and saturated volumetric water content (θr, θS), and fitting parameters α and n. Measurements include soil physical and chemical properties, bulk density, and organic matter, alongside pre-processed spectral reflectance using methods such as multiplicative scatter correction (MSC), Savitzky-Golay derivatives, and standard normal variate (SNV). The dataset supports the development of pedotransfer functions (PTFs) and evaluation of machine learning models (Random Forest and Multiple Linear Regression) for predicting SWRC parameters. It can be used for soil hydraulic modeling, machine learning research, and sustainable soil and water resource management.
Files
Steps to reproduce
Software: All analyses were performed using Statistica. Models: Two models were applied for predicting SWRC parameters: Random Forest (RF) and Multiple Linear Regression (MLR). Data Preprocessing: Spectral and soil data were processed using the following techniques: No preprocessing Multiplicative Scatter Correction (MSC) Savitzky-Golay first derivative (SG1) Savitzky-Golay second derivative (SG2) Standard Normal Variate (SNV) Other preprocessing methods (F, NI, FI, N) Data Splitting: The dataset was divided into training (90%) and testing (40%) subsets. Performance Metrics: Model performance was evaluated using Mean Error (ME), Root Mean Square Error (RMSE), Coefficient of Determination (R²), and Akaike Information Criterion (AIC).
Institutions
- Bu Ali Sina University