مقالات رزومه فاطمه رحمتی

The present research performed to estimate soil texture using visible near-infrared spectrometry in Semirom, Isfahan. A total number of 200 soil samples (0-10 cm) were collected from the Semirom area (51º 17' - 52º 3' E; 30º 42' - 31º 51' N), Isfahan. The samples were air dried and passed through a 2 mm sieve, and soil particles percentage was determined in the laboratory using hydrometry method. Reflectance spectra of all samples were measured using an ASD field spectrometer. Different pre-processing methods i.e., First Derivatives and Savitzky-Golay Filter, Multiplicative Scatter Correction and Standard Normal Variable were applied and performed on spectral data. The Partial Least Squares Regression, Support Vector Machine Regression and Artificial Neural Network models were used to estimate soil texture. The best result was obtained for Silt estimation, with excellent values of RPD >2, R2 =0.98 and RMSE=1.08 using Artificial Neural Network model with MSC pre-processing technique. The results indicated the desirable capability of Artificial Neural Network model with MSC and SNV pre-processing techniques in estimating the Clay (RPD >2, R2=0.94 and RMSE=1.21) and Sand (RPD >2, R2=0.84 and RMSE=6.24) contents of the soils, respectively. In general, based on the results of this study, VNIR spectroscopy was successful in estimating soil particles percentage and showed its potential for substituting laboratory analyses.

 Estimating soil properties on large scales using experimental methods requires specialized equipments and can be extremely time-consuming and expensive, especially when dealing with a high spatial sampling density. Soil Visible and Near-InfraRed (V-NIR) reflectance spectroscopy has proven to be a fast, cost-effective, non-destructive, environmental-friendly, repeatable, and reproducible analytical technique. V-NIR reflectance spectroscopy has been used for more than 30 years to predict an extensive variety of soil properties like organic and inorganic carbon, nitrogen, organic carbon, moisture, texture and salinity. The objectives of this study were to estimate soil properties (carbonate calcium equivalent (CCE), electrical conductivity (EC), pH, and organic carbon (OC)) using visible near-infrared and short-wave Infrared (SWIR) reflectance spectroscopy (350-2500 nm). In this study, the best predictions of all the soil properties, model and pre-processing technique were also determined. The Partial Least Squares Regression (PLSR), Artificial Neural Network, Support Vector Machine Regression and Principal Component Regression (PCR) models were also compared to estimate soil properties.

 A total number of 200 surface soil samples (0-10 cm) were collected from the Semirom region (51º 17' - 52º 3' E; 30º 42' - 31º 51' N), Isfahan, Iran. The samples were air dried and passed through a 2 mm sieve, and using standard procedures soil properties were determined in the laboratory. Accordingly, soil pH and the EC contents of soil samples were determined in saturated pastes and extracts, respectively. The CCE content of the soils were measured using back titration, and the OC contents of the samples were measured using Walkley-Black method. The Reflectance spectra of all samples were measured using an ASD field spectrometer. The selection of the best model was done according to the value of the Ratio of Performance to Deviation (RPD), the coefficient of determination (R2), and the Root Mean Square Eerror (RMSE).

 Once the models were constructed using PLSR, ANN, SVMR and PCR approaches, descriptive analysis was carried out for each property, for the data measured in the laboratory. The parameters calculated for the properties were mean, coefficient of variation (CV), minimum and maximum, standard deviation and range. Coefficient of variation for the organic carbon, CCE, pH, and EC values were 21.7, 12.4, 1.34, and 28.74, respectively. Wilding (1985) proposed low, medium, and high variability for the CV values less than 15%, 15-35%, and greater than 35%, respectively. Accordingly, the organic carbon and EC of soils could be classified in the group with moderate variability. However, the calcium carbonate equivalent and pH are in the group with low variability. Since spectral data preprocessing has an effective role on improving the calibration, in order to perform spectral preprocessing, two first nodes at the first (350-400 nm) and the end (2450-2500 nm) of each spectrum were removed. In addition, two interruptions were eliminated, due to the change in the detector in the range of 900 to 1700 nm. Different preprocessing methods i.e., Standard Normal Variable (SNV) and First (FD) and Second Derivatives (SD) and Savitzky-Golay preprocessing techniques were performed on spectral data. Then, using PLSR, the cross‐validation method was used to evaluate soil properties calibration and validation. According to Stenberg (2002), for agricultural applications, The values of RPD greater than 2 indicate that the models provide precise predictions, the values of RPD between 1.5 and 2 are considered to be reasonably representative, and the values of RPD less than 1.5 indicate poor predictive performance. The results indicated the desirable capability of the PLSR method in estimating the EC (RPD > 2, R2 = 0.94), CCE (RPD > 2, R2 = 0.88), and OC (RPD > 2, R2 = 0.89). The best results of the pH (RPD > 2, R2 = 0.79) were estimated by the SVMR method. In this study the best methods of preprocessing techniques were First (FD) and Second Derivatives (SD) and Savitzky-Golay filter.

 In general, based on the results of this study, VNIR spectroscopy was successful in estimating soil properties and showed its potential for substituting laboratory analyses. Moreover, spectroscopy could be considered as a simple, fast, and low-cost method in predicting soil properties. The PLSR model with First and Second derivatives and Savitzky-Golay pre-processing techniques seems to be more robust algorithm for estimating EC, OC, and CCE. The best results of the pH were estimated by the SVMR method with First and Second derivatives and Savitzky-Golay pre-processing techniques.