فصلنامه سنجش از دور و سامانه اطلاعات جغرافیایی در منابع طبیعی
سال ششم شماره 3 (پاییز 1394)

Preparing the maps of land use/cover for spatial planning and management is essential. Nowadays, satellite images and remote sensing techniques have widespread applications according to their capabilities to produce the updated data and analyze the images in all disciplines such as agriculture and natural resources. In the present study, Artificial Neural Network, Support Vector Machines and Object-Based techniques wereutilized for drawing the land use and vegetation maps in Ardabil, Namin, and Nir counties. The images of LandSat-8 Operational Land Imager (OLI) (2013) were usedafter geometric correction and topographic normalization and classified into 9 land use/cover classes including water bodies, irrigated farming, rainfed farming, meadows, outcrops, forests, rangelands, residential and airport areas. After the accuracy assessment, overall accuracy for the produced maps of ANN, Support Vector Machine (SVM) and Object-based (OB) techniques was estimated as 89.91, 85.68 and 94.37%, respectively and Kappa's coefficients were 0.88, 0.82 and 0.93, respectivelyindicating that the object-based method in comparison with two other methods has more advantages;on the other hand, all three methods could provide the desirable accuracy for the land use/covermaps. Overally, three advanced classification methods were examined in the heterogeneous area with elevation changes up to 3600m using the images of new lunched Landsat 8 and the most appropriate land use/cover mapping method was introduced.

Using time series of satellite images is a cheap and efficient way to study trend changes of natural and human phenomena. The aim of the studyis to analyze the development of Urban Heat Iisland (UHI) in Rasht using time series satellite images. For this study, time series of LandSat images during 1990 and 2015 were used. Thresholding Normalized Difference Vegetation Index (NDVI) and Fraction Vegetation Cover (FVC) method has been applied to obtain the land surface emissivity; in addition, Planck's law for TM and ETM+ images and Split Window (SW) algorithm for OLI/TIRS images were utilized in order to retrieve land surface temperature. UHI and FVC trends were analyzed by statistical and Mann-Kendall methods. Statistical analysis showed that the average of FVC has decreased during the study periodand data skewness has changed to low FVC. The reduction trendhas increased FVC caused an average normalized temperature during the study period and also enhanced the data skewness of land surface temperature. The Mann-Kendall spatial analysis showed that in most of the study area, the land surface temperature and vegetation fraction haveincreased and decreasing trends, respectively; these mentioned trends have been intensified in the places where gardens and agricultural land uses were changed into built-up ones.

The aim of this study was to evaluate the efficiency of three support vector machine algorithms, fuzzy decision trees and neural networks for mapping land vegetation map of Arakvaz watershed using OLI sensor of Landsat images (2014). Geometric correction and image pre-processing were utilized to determine the training samples of land vegetation classes for the classification operations. Sample resolution in the vegetation classes has been evaluated using a statistical divergence index. On the next stage, to evaluate the accuracy of algorithm's classification results, ground truth map with the dimensions of 550 m was designed using systematic approach and land vegetation types in the sampling plots were determined. Finally, the efficiency of each classification methodwas investigated bysuch criteria as overall accuracy, kappa coefficient, producer accuracy and user accuracy.Comparing the accuracy and kappa coefficient obtained for three categories with a proper band set in comparison with the ground truth map indicates that the Support Vector Machine (SVM) classifier with overall accuracy of 91.26% and kappa coefficient of 0.8731 has had more appropriate results than other algorithms. The results showed that the separation and classification of forest landswith high accuracy have beenperformedas compared to the other land use classes.

One of the forest management tools which has a direct impact on the qualitative and quantitative characteristics is timber marking. In order to evaluate the effects of physiographic factors of land and road forms on timber marking for the logging, the multiple linear regression model was used in the Forestry Plan Series 2 in the seventh zone of forestry plans of Nekachoub Company, Mazandaran province. For this purpose, the location of marked trees was determined using GPS in a plot with an area of 215 hectares. Then, maps of marked trees, altitude, aspect and slope percent, soil and distance from the road were prepared in GIS environment using these points and digital elevation model of study area. To understand the relationship between factors and marked trees, the multiple linear model was used. So, maps of mentioned factors and marked tree's location were entered into the regression model as independent and dependent variables, respectively. The model results showed that the timber marking activity was done more in the areas with low altitude, gentle slope and deep soils on northern and eastern directions near the forest road as compared to the other areas. Timber marking is affected by some physiographic factors, and such elements as slope or adjacency to the road may make the timber marker to interfere in these areas and avoid working in the areas with higher slope or away from the road; it leads to an uneven interventions throughout the forest as well as an unbalanced reclamation of forest structure.

This study was carried out in the western catchment of the Behzisty Township in Gorgan, Golestan province in order to develop the regression models of soil moisture. In this study, the desired measurements were made in 18 locations using TDR during 6 running days after a relatively intense rainfall event at different depths of soil. The digital elevation model was prepared from UltraCam images taken in 2014 with the pixel size of 1×1 m in order to estimate the terrain properties. Finally, the regression model was developed with respect to these characteristics and soil moisture. The results showed that important terrain characteristics which are greatly associated with soil moisture are horizontal curvature, slope, aspect and wetness index in the regression model. These results suggest that the diversion of material flow has controlled the amount of water in the soil, slope aspect and soil moisture. The results showed that the developed regression models can forecast at least 44 to 60% of the total variations of soil moisture in the watershed scale.

In recent decades, remote sensing methods because of their economical aspects are frequently used in soil survey studies. The objective of this study was to compare the remote sensing and geostatistical methods for estimating and mapping the calcium carbonate equivalent and clay content of surface soil in the Chitgar Park with the area of 665 hectares. Therefore, 116 samples (0-20 cm) were taken by the intervals of 250 meters based on regular grid patterns. Two geostatistic methods of kriging and inverse distance weighting (IDW), and two remote sensing techniques of near-infrared analysis (NIRA) and Continuum removal (CR) were used for modeling the desired variations. The results of the cross validation showed that the accuracy of kriging was appropriate for modeling clayand calcium carbonate equivalent contents. The NIRA method due to the low correlation between TM bands and calcium carbonate equivalent was not validated for modeling the calcium carbonate equivalent; but this method was appropriate for estimating clay content (R2=0.52). CR method was not accurate for both variables of clay and calcium carbonate equivalent. Based on the overall accuracy and Kappa coefficient of producing maps, it is concluded that the kriging method has higher accuracy than remote sensing methods. Although high accuracy of geostatistic methods is expected due to the use of real data, the acceptable accuracy of the NIRA technique for modeling the clay variable should be considered with respect to lower costs of remote sensing methods.

This paper has been conducted to estimate the detection capability of LandSat satellite data for the detection and qualitative assessment of forest area and vegetation changes, land uses and vegetation percent in Rustam city. In this regard, using Landsat satellite images (1987 and 2010), forest land use map, Normalized Difference Vegetation Index (NDVI) and Ratio Vegetation Index (RVI) were obtained by Maximum Likelihood and Supervised Classification algorithms. The results showed that the area of extracted layers of forests with high, moderate and low density as well as barren regions has been estimated as 48.78, 348.67 and 281.42 and 81.68 km2, respectively. Assessing the classification results indicated that the overall accuracy, producer accuracy and user accuracy were given as 94.3, 91 and 95%, respectively and also, Commission error and Ommission error have been computed as 0.09 and 0.05. The highest and lowest producer accuracy estimated as 99 and 80% was related to low-density and high density forests and the highest and lowest percent of user accuracy given as 100 and 87% was attributed to the barren and moderate density forest. Also, comparing maps of vegetation percent and Ratio Vegetation index during 1987 and 2010 has shown no significant changes.