Yousef Hasheminejhad; Mehdi Homaee; Ali Akbar Noroozi
Abstract
Introduction: Soil salinization is increasing across developing world countries and agricultural production is decreasing as a result of this stress. Climate change could adversely affect soil salinization trend through the decrease in rainfall and increased evapotranspiration in arid regions. Policy ...
Read More
Introduction: Soil salinization is increasing across developing world countries and agricultural production is decreasing as a result of this stress. Climate change could adversely affect soil salinization trend through the decrease in rainfall and increased evapotranspiration in arid regions. Policy and decision makers require continuous and quantitative monitoring of soil salinity to adapt with the adverse effects of climate change and increasing need for food. Indices derived from near surface or satellite based sensors are increasingly applied for monitoring of soil salinity so a considerable number of these indices are introduced already for soil salinity monitoring. Different regression methods have been already used for modeling and verification of developed models amongst them multiple linear regression (including stepwise, forward selection and backward elimination) and partial least square regression are the most important methods.
Materials and Methods: To evaluate different approaches for modeling soil salinity against remotely sensed data, an area of about 50000 ha was selected in Sabzevar- Davarzan plain during 2013 and 2014 years. The locations of sampling points were determined using Latin Hypercube Sampling (LHS) strategy. Sampling density was 97 points for 2013 and 25 points for 2014. All points were sampled down to 90 cm depth in 30 cm increments. Totally 366 soil samples were analyzed in the laboratory for electrical conductivity of saturated extract. Electromagnetic induction device (EM38) was also used to measure bulk soil electrical conductivity for the sampling points at the first year and sampling points and 8 points around it at the second year. Totally 97 and 225 EM measurements were also recorded for first and second years respectively. Mean measured soil EC data were calibrated against the EM measurements. Finding the fair correlations, the EM and EC data could be converted to each other. 23 spectral indices derived from Landsat 8 images in the sampling dates along with DEM were used as independent variables. Multiple Linear Regression (MLR) and Partial Least Square Regression (PLSR) methods were evaluated for their fitness in predicting soil salinity from independent variables in different calibration and verification datasets.
Results and Discussion: Different multiple linear regression approaches using the first year data for training and second year data for testing the models and vice versa were evaluated which produced determination coefficients of about 22 to 88 percent in the training dataset but this regression did not reach to 29 percent in the test dataset. Due to the multiple co-linearity amongst the independent variables the multiple linear regression methods were not applicable to all variables. Excluding the co-linear variables, log- transforming and randomizing them into train and test datasets improved the determination coefficient of model and its validation at an acceptable level. Application of partial least square regression using the original and log- transformed data of first and second years as train and test datasets and vice versa introduced determination coefficients of about 39 to 85 percent in the training dataset but were not able to predict in the test dataset. Random dividing of all data into train and test datasets considerably increased the determination coefficient in the verification dataset. Repeating the randomization showed that the approach has the required consistency for predicting the coefficients of variables.
Conclusions: Wide range of independent variable could be used for predicting soil salinity from remotely sensed data and indices. On the other hand the independent variables generally show multi-colinearity amongst themselves. Correlation matrix, variance inflation factor and tolerance indices could be used to identify multi-colinearity. Removing or scaling the variable with high colinearity could improve the regression. Different data transformation methods including log- transformation could also significantly improve the strength of regression. In this research EM data showed more significant correlations with spectral indices in comparison with laboratorial measured EC data. As the EM38 device measures the reflectance in special range of spectrum this higher correlation could be expected. Such models should be calibrated and verified against ground truth data. Generally a part of data set is used for calibrating (making the model) and the remained for verifying (testing the model). Random dividing of the total data of 2 years into calibration (2/3 of data) and verification (1/3 of data) could significantly improve the regression in the verification data set. This procedure increases the range of variability for data used for calibration and verification and prevents outlier predictions.
Yousef Hasheminejhad; Mahdi Homaee; Ali Akbar Noroozi
Abstract
Introduction: Monitoring and management of saline soils depends on exact and updatable measurements of soil electrical conductivity. Large scale direct measurements are not only expensive but also time consuming. Therefore application of near ground surface sensors could be considered as acceptable time- ...
Read More
Introduction: Monitoring and management of saline soils depends on exact and updatable measurements of soil electrical conductivity. Large scale direct measurements are not only expensive but also time consuming. Therefore application of near ground surface sensors could be considered as acceptable time- and cost-saving methods with high accuracy in soil salinity detection. . One of these relatively innovative methods is electromagnetic induction technique. Apparent soil electrical conductivity measurement by electromagnetic induction technique is affected by several key properties of soils including soil moisture and clay content.
Materials and Methods: Soil salinity and apparent soil electrical conductivity data of two years of 50000 ha area in Sabzevar- Davarzan plain were used to evaluate the sensitivity of electromagnetic induction to soil moisture and clay content. Locations of the sampling points were determined by the Latin Hypercube Sampling strategy, based on 100 sampling points were selected for the first year and 25 sampling points for the second year. Regarding to difficulties in finding and sampling the points 97 sampling points were found in the area for the first year out of which 82 points were sampled down to 90 cm depth in 30 cm intervals and all of them were measured with electromagnetic induction device at horizontal orientation. The first year data were used for training the model which included 82 points measurement of bulk conductivity and laboratory determination of electrical conductivity of saturated extract, soil texture and moisture content in soil samples. On the other hand, the second year data which were used for testing the model integrated by 25 sampling points and 9 bulk conductivity measurements around each point. Electrical conductivity of saturated extract was just measured as the only parameter in the laboratory for the second year samples.
Results and Discussion: Results of the first year showed a significant correlation between electrical conductivity and apparent conductivity with a regression coefficient of 0.78. Although multiple linear regression by inclusion of soil moisture and clay content as independent variables improved the regression coefficient to 0.80 but the effect of clay content was not significant in this multiple model. Sensitivity analysis by maintaining one variable at its average value and changing the second variable also showed greater sensitivity of the model to soil moisture in comparison with soil clay content. Generally under estimation of soil moisture and over estimation of soil clay content produced about 63 to 65 percent Mean Bias Error (MBE) while over estimation of soil moisture and under estimation of soil clay content produced about 35- 37 percent of MBE. So the model is quite sensitive to both parameters and they cannot be estimated in the field by feeling and the other field methods. Simple linear regression model between ECe and EMh was tested on the second year because the errors in estimating soil moisture could be imposed a significant error on estimating soil salinity. Once the model was tested for estimation of soil salinity in the central point based on EMh reading at the center and then it was tested for estimation of soil salinity based on the average EMh of 9 points in each location. Results showed that the correlation between soil salinity and apparent soil electrical conductivity could be improved to 0.98 using the average of 9 measurements instead of 1 measurement.
Conclusion: Based on the results the electromagnetic induction device is sensitive to soil moisture. Although its sensitivity to clay content is less than the sensitivity to moisture content, but the total model error as a result of over estimating soil moisture is about equal to its error resulted from under estimating clay content and vice versa. So the field and feeling methods could not be implemented as inputs for the multiple regression models but these methods have enough accuracy to divide soil samples into two groups of dry and wet soils or sandy or clayey soils, on the other hand measurements of these parameters imposes more cost and time to soil salinity surveys. Results also showed that the repeated EM measurements around each sampling point could improve the strength of the regression. Therefore regarding to the sensitivity of the technique to soil moisture three methods are suggested to improve accuracy of calibration: a)- measurement and calibration under the same moisture conditions; b)- field approximation of soil moisture and dividing soil samples into two groups of dry and moist soils and deriving two different groups of calibration equations.
Y. Hasheminejhad; M. Gholami touran poshti
Abstract
Abstract
The initial assumption of many laboratory, greenhouse and lysimetric experiments is the longtitudal and latitudal homogeneity of repacked soil columns as well as homogeneity between different replications. Despite the great importance of these assumptions, a few reports, especially among domestic ...
Read More
Abstract
The initial assumption of many laboratory, greenhouse and lysimetric experiments is the longtitudal and latitudal homogeneity of repacked soil columns as well as homogeneity between different replications. Despite the great importance of these assumptions, a few reports, especially among domestic studies, have discussed the details of packing procedure. On the other hand, in most cases no criterion is introduced as verification standards, while the methods introduced are dependent on soil sampling which itself is destructive. This paper has discussed the details of packing procedure used for filling 36 cylindrical lysimeters having 39.5 cm internal diameter and 150 cm height. In addition, 3 nondestructive criteria have been introduced to verify the packing procedure and homogeneity in repacked soil columns. These criteria include the soil surface depression, variation in field capacity (FC) point, and the slope of breakthrough curves (BTCs). Based on all introduced criteria, the introduced packing procedure produced longtitudal and latitudal homogeneity in repacked soil columns and similarity between 36 replications.
Key words: Soil packing, lysimeter, Breakthrough curve (BTC), Homogeneity
