Mohammad Ali Mahmoodi; Sohaila Momeni; Masoud Davari
Abstract
Introduction: Land use and Land cover (LULC) information has been identified as one of the crucial data components for a range of applications including global change studies, urban planning, agricultural crop characterization, and forest ecosystem classification. The derivation of such information increasingly ...
Read More
Introduction: Land use and Land cover (LULC) information has been identified as one of the crucial data components for a range of applications including global change studies, urban planning, agricultural crop characterization, and forest ecosystem classification. The derivation of such information increasingly relies on remote sensing technology due to its ability to acquire valuable spatiotemporal information on LULC. One of the major approaches to deriving LULC information from remotely sensed images is classification. Numerous image classification algorithms exist. Among the most popular are the maximum likelihood classifier (MLC), artificial neural network (ANN) classifiers and decision tree (DT) classifiers. Conventional parametric method like MLC is based on statistical theory and assumes a multivariate normal distribution for each class. In case of data that has non-normal distribution (which is common with LULC data), the parametric classifiers may fail since the inability to resolve interclass confusion. This inability is the major limitation of parametric classifiers. Nonparametric classifiers like ANNs and DTs, which do not rely on any assumptions for the class distributions of data, could overcome the aforementioned limitations of parametric classifiers. The support vector machines (SVMs), a nonparametric classifier, that has recently been used in numerous applications in image processing, represents a group of theoretically superior machine learning algorithms. The SVM employs optimization algorithms to locate the optimal boundaries between classes. It was found competitive with the best available classification methods, including ANN and DT classifiers. The classification accuracy of SVMs is based upon the choice of the classification strategy and kernel function. The objective of this study was to investigate the sensitivity of SVM architecture including classification strategy and kernel types to identify LULC information from Landsat Enhanced Thematic Mapper (ETM) remote sensing data in Gavshan dam watershed in west of Iran.
Materials and Methods: SVMs were used to classify orthocorrected Landsat ETM images of May, 2016. Image pre-processing such as atmospheric correction were conducted before utilization. Three classification strategies (One versus one, one versus all and ordinal) and three types of kernels (linear, polynomial and radial basis function) were used for the SVM classification. A total of 18 different models were developed and implemented for sensitivity analysis of SVM architecture. A two-layer feed-forward Perceptron network classifier with sigmoid hidden and softmax output neurons was also used for comparison. The network was trained using scaled conjugate gradient backpropagation algorithm. A total of 1320 ground control points were collected to train, validate and test the SVM and ANN models. Ground truth locations on each image were identified using the GPS coordinates for extracting spectral reflectance data of seven bands (Bands 1-7) of Landsat ETM images. The LULC class of each point was identified using land survey or Google earth images. The identified LULC classes were agriculture, buffer forests, orchard, ranges brush, range grasses, urban areas, roads and water.
Results and Discussion: The results suggest that the choice of classification strategy and kernel types play an important role on SVMs classification accuracy. Statistical evaluation of the SVM models against the ground control points showed that the one versus one classification strategy had the highest accuracy than the two other ones for any kernel function type and the polynomial kernel function had the highest accuracy than the two other kernels for any classification strategy. The SVM model with polynomial (n=3) kernel and one versus one classification strategy outperformed all SVMs models and gave the highest overall classification accuracy of 78.5 and Kappa coefficient of 68.5. The McNemar’s test clearly showed significant improvement of the best SVM model in comparison to the ANN model (P<0.001). Also, the user accuracy and producer accuracy achieved by best SVM model were higher than ANN model for all LULC classes. In both approaches water and agriculture categories have high accuracy while roads have low accuracy. The resulting LULC map indicated that most parts of the studied area (52.8%) have been assigned to the agriculture. The ranges brush and range grasses categories cover 12.5% and 26.8% of the watershed, respectively. Only about 2.7% of the watershed have been covered with trees.
Conclusions: This study suggests that the SVMs approach based on Landsat ETM bands may provide reliable and accurate LULC information even better that best ANN approaches. However, choice of classification strategy and kernel types play an important role on SVMs classification accuracy. Best model of polynomial kernel and one versus one classification strategy outperformed all SVMs and ANN models and gave the highest classification accuracy.
hojjat ghorbani vaghei; M. Davari
Abstract
Introduction: Soil organic carbon (SOC) has great impacts on soil properties, soil productivity, food security, land degradation and global warming. Similar to other soil properties, SOC has a strong spatial heterogeneity as a result of dynamic interactions between parent material, climate and geological ...
Read More
Introduction: Soil organic carbon (SOC) has great impacts on soil properties, soil productivity, food security, land degradation and global warming. Similar to other soil properties, SOC has a strong spatial heterogeneity as a result of dynamic interactions between parent material, climate and geological history, at both regional and continental scales. However, landscape attributes including slope, aspect, altitude, and land use types are dominant factors influencing on SOC in areas with the same parent materials and climate regime. Understanding and identifying the spatial and temporal distribution of SOC is essential to evaluate soil quality, agricultural management, watershed modeling and soil carbon sequestration budgets. Therefore, the objectives of this study was to estimate soil organic carbon content in the Aligodarz watershed, and to investigate the effects of altitude, slope, and land use type on SOC.
Materials and Methods: The research was carried out in the Aligodraz watershed in Lorestan province of Iran. The study area is located between latitudes N 33° 10' 51.72"to N 33° 34' 28.22" and longitudes E 49° 27' 17.99"to E 49° 58' 40.84" 14 that covers an area of 1078.9 km2. It has an altitude between 1866.3 and 3200 m above sea-level. The primary land uses within the watershed include pasture, dryland and irrigated farming. In this study, soil samples were randomly collected from 206 sites at depth of 0– 15 cm during June and August 2003. The mean distance between samples was about 5 km. Soil samples were air-dried in the shade for about 7 days and then passed through a 0.25 mm prior to determination of SOC. Soil organic carbon content was determined in triplicate for each sample using the Walkey-Black method. Basic statistical analyses for frequency distribution, normality tests, Pearson's correlation and analysis of variance were conducted using SPSS (version 18.0). Calculation of experimental variograms and modeling of spatial distribution of SOC were carried out with the geostatistical software GS+ (version 5. 1). Maps were generated by using ILWIS (version 3.3) GIS software.
Results and Discussion: The results revealed that the raw SOC data have a long tail towards higher concentrations, whereas that squareroot transformed data can be satisfactorily modelled by a normal distribution. The probability distribution of SOC appeared to be positively skewed and have a positive kurtosis. The square root transformed data showed small skewness and kurtosis, and passed the K–S normality test at a significance level of higher than 0.05. Therefore, the square root transformed data of SOC was used for analyses. The SOC concentration varied from 0.08 to 2.39%, with an arithmetic mean of 0.81% and geometric mean of 0.73%. The coefficient of variation (CV), as an index of overall variability of SOC, was 44.49%. According to the classification system presented by Nielson and Bouma (1985), a variable is moderately varying if the CV is between 10% and 100%. Therefore, the content of SOC in the Aligodarz watershed can be considered to be in moderate variability. The experimental variogram of SOC was fitted by an exponential model. The values of the range, nugget, sill, and nugget/sill ratio of the best-fitted model were 6.80 km, 0.058, 0.133, and 43.6%, respectively. The positive nugget value can be explained by sampling error, short range variability, and unexplained and inherent variability. The nugget/sill ratio of 43.6% showed a moderate spatial dependence of SOC in the study area. The parameters of the exponential smivariogram model were used for kriging method to produce a spatial distribution map of SOC in the study area. The interpolated values ranged between 0.30 and 1.40%. Southern and central parts of this study area have the highest SOC concentrations, while the northern parts have the lowest concentrations of SOC. Kriging results also showed that the major parts of the Aligodarz watershed (about 87%) have statistically SOC content less than 1%. Lower SOC concentrations were associated with high altitude (r = −0.265**). The results of Pearson correlation analysis showed that soil organic carbon content has a significantly negative correlatiton with slope gradient (r = −0.217**). The results also indicated that the SOC content was variable for the different land use types. The irrigated lands had the highest SOC concentrations, while the pasture lands had the lowest SOC values.
Conclusion: The square-root transformed data of SOC in Aligodarz watershed of Lorestan province, Iran, followed a normal distribution, with an arithmetic mean of 0.81%, and geometric mean of 0.73%. The coefficient of variation and nugget/sill ratio revealed a moderate spatial dependence of SOC in the study area. The results indicated that the major parts of the Aligodarz watershed have SOC content less than 1%. The land use type had a significant effect on the spatial variability of SOC and that lower SOC concentrations were associated with higher altitude and slope gradients. The irrigated and pasture lands had the highest and lowest SOC concentrations, respectively.
M. Davari; M. Homaee
Abstract
Abstract
Soil Contamination by heavy metals is yet one of the most important environmental concerns. Among heavy metals, Nickel and Cadmium have dangerous influences on human, animals and plants. The objective of this study was to derive a new model for simultaneous phytoextraction of Ni and Cd from ...
Read More
Abstract
Soil Contamination by heavy metals is yet one of the most important environmental concerns. Among heavy metals, Nickel and Cadmium have dangerous influences on human, animals and plants. The objective of this study was to derive a new model for simultaneous phytoextraction of Ni and Cd from contaminated soils. Consequently, a macroscopic model was derived by combining yield reduction functions and relative concentrations of Ni and Cd in plant tissues. To verify the derived model, a clay loam soil was simultaneously contaminated with different concentrations of Ni and Cd. The Ornamental Kale seeds were then seeded in these packed contaminated soils in three replicates. Plants were harvested after full development. The Ni and Cd contents of soil samples and plant materials were extracted by 4M HNO3 oxidation and wet oxidation methods, respectively. The Ni and Cd concentrations were measured by Atomic Absorption Spectrometer (Shimadzu, AA 670-G) and Inductively Coupled Plasma Optical Emission Spectrometry (Varian Vista-PRO). The results indicated that relative yield of Ornamental Kale in the contaminated soils with both Ni and Cd was reduced more than the soil polluted with separate Cd or Ni. The results also indicated that at any given soil Cd concentration, the Ni content of Ornamental Kale increases with increasing soil Ni concentration. Meanwhile, with increasing soil Cd, the Ni content in Ornamental Kale was decreased. Further, at any given Cd content, the amount of Cd in Ornamental Kale was increased by increasing Ni concentration in soil. The results further indicated that the proposed model can well predict Ni phytoextraction from soils contaminated with both Ni and Cd. However, this model could only provide an overall estimate for Cd phytoextraction. It was further concluded that Ornamental Kale due to its high biomass production and high tolerance to Ni and Cd concentrations can be used to remediate low to moderate combined Ni -Cd contaminated soils.
Keywords: Cadmium, Multiplicative theory, Nickel, Ornamental kale, Phytoextraction
