Journal of Spatial Analysis

Urmia Lake is one of the largest hyper saline lakes in the world and largest inland lake in Iran which located in the north west of Iran, between the provinces of East Azerbaijan and West Azerbaijan. The lake basin is one of the most influential and valuable aquatic ecosystems in the country and registered as UNESCO Biosphere Reserve. In addition, it is very important in terms of water resources, environmental and economic. Unfortunately, lake water level has dramatically decreased in recent years, due to various reasons. This issue has created some problems for Local people, especially people living in rural area in east of the Lake. The results of this research are of great importance for regional authorities and decision-makers in strategic planning for people of inhabits in east coast village.

The present paper is an attempt to integrate a semi-automated Object-Based Image Analysis (OBIA) classification framework and a CA-Markov model to show impacts of Urmia Lake Retrogression On eastern coastal villages. OBIA present novel methods for image processing by means of integration remote sensing and GIS. Process and outcome of this methodology can be divided in three step including: Segmentation, Classification and Accuracy assessment.in the process of segmentation aims to create of homogeneous objects by considering shape, texture and spectral information. A necessary prerequisite for object oriented image processing is successful image segmentation. In our research the segmentation step was performed by applying multi-resolution segmentation and considering 0.2 for shape and 0.4 for the compactness. The scale of segmentation is also an important option which leads to determine the relative size of each object. Having great values for scale leads to create large objects while smaller value would result small objects respectively. In this study the scale parameter of 100 has been selected based on the size of objects in Scale of study area as well as spatial resolution of the satellite images were used for segmentation. In doing so, we employed spectral and visual parameters contains: texture, shape, color tone and etc. for developing object based rule-sets.  To determine the characteristics of the spectral data and geometric features classes the fuzzy based classification was performed by employing fuzzy operators including: or (max) operator with the maximum value of the return of the fuzzy, the arithmetic mean value of fuzzy and the geometric mean value of fuzzy, and (min). After this step, the validation process was performed by using overall accuracy and Kappa coefficient. Then, using the CA-Markov Model The trend of changes was predicted in the future (For 2020). Another way to predict changes in land use and cover, used the CA-Markov model. Markov chain analysis is a useful tool for modeling land use changes. Markov chain model consists of three step: First step Calculating the probability conversion using Markov chain analysis, second step, Calculating the Cover and land use maps competently on the basis of multi-criteria evaluation, third step, assign locations cover and land use simulation based on the CA position operator.

Results of Satellite image processing indicate that the area of garden, Farmland, Zones of muddy-salty (Saline soils), moist salt and newly formed salt have increased while area of Urmia lake has rapidly dropped between 1984 and 2015. The area of Urmia lake declined from 4904.51 square kilometers in 1984 to 676.79 square kilometers in 2015. The farmland area increased from 177.72 square kilometers in 1984 to 542.37 square kilometers in 2015. The garden area increased from 83.71 square kilometers in 1984 to 227.28 square kilometers in 2015. The moist salt area increased from 111.89 square kilometers in 1984 to 945 square kilometers in 2015. Zones of muddy-salty (Saline soils) area increased from 859.01 square kilometers in 1984 to 2986.5 square kilometers in 2015. The newly formed salt increased from 171.27 square kilometers in 1984 to 921.99 square kilometers in 2015. Markov chain model results indicate in 2020 the garden area will be 638 square kilometers, the moist salt area will be 717 square kilometers, Zones of muddy-salty (Saline soils) area will be 4127 square kilometers, the farmland area will be 644 square kilometers, the newly formed salt area will be 363 square kilometers and the Urmia lake area will be 118 square kilometers.

One of the most important calamities that affect the arid and semi- arid regions and is taken into account as threatening factors for human- life and destroying the natural resources is desertification, so recognizing and forecasting this phenomenon is very important. Desertification is a complex phenomenon, which as environmental, socio-economical, and cultural impacts on natural resources. In recent years, the issues of desertification and desert growth have been stated as important debate on global, regional and national levels and extensive activities have been carried out to control and reduce the its consequences. Desertification is considered as the third important global challenge in the 21th century after two challenges of climate change and scarcity of fresh water. At present, desertification as a problem, involves many countries, especially developing countries and includes some processes that caused by natural factors as well as human incorrect activities. In the other word, Desertification is the ecological and biological reduction of land that maybe occur naturally or unnaturally. The desertification process influences the arid and semiarid regions essentially and decrease the lands efficiency with increment speeds. The study area is located in the east and south of Isfahan. This region has been faced to increasing rate of desertification, because of drought, vegetation removal, change of rangelands to dry farming lands, water and wind erosion and lack of proper land management over previous years. Hence, aim of this research is monitor and forecasting of desertification changes in the east and south of Isfahan during the period of (1986-2016). In this research, the Landsat satellite images used as studies base for studying region desertification. Therefore, at first, satellite images of the study area were extracted from United States geological survey(USGS)website during the period of (1986-2016) and data and satellite images of TM5, ETM+ 7 and LDCM8 sensors of Landsat satellite were used which include thermal and spectral bands. In this relation, for studying the desertification condition in the south and east region of Isfahan, the Landsat satellite images of 4, 7 and 8 during 5 periods of 1986, 1994, 2000, 2008 and 2016 have been utilized. After completing the information data base, first, the soil salinity(S1, S2 and S3) and vegetation NDVI indices exerted on the satellite images. According to Fuzzy ARTMAP method, the land use changes during the period of (1986-2016) recognized in the studied region. In the other word, the vegetation NDVI and soil salinity (S1, S2 and S3) indices have been utilized for identifying vegetation and the desert and salty regions. For preparing the region land use map, the Fuzzy ARTMAP supervised classification method have been utilized and 5 land uses(desert and salty lands, vegetation, city, arid and Gavkhouni) in the region were identified by TerrSet software. The changes calculation in region uses during 5 periods accomplished by LCM model. Also, the Markov chain and Cellular automata synthetic model have been utilized for changes forecasting. This research results indicated that the greatest changes during studied period belonged to vegetation. This volume of change had been during 1986- 1994 that indicate 1062 km2 desertification. In the other hand, the greatest intensity of increasing the salty and desert regions have been occurred during 1994-2000 which indicate 495 km2 increasing. The CA- Markov synthetic method have been utilized for forecasting the land uses changes trend, too. In this relation, for assessing the forecast accuracy, the Kappa coefficient have been utilized which indicate 78%. Finally, it has been specified that the greatest changes during 2016-2024 will be in vegetation which about 60% of region vegetation will disappear and arid lands will be replace them. The salty and desert lands will disappear about 1% of vegetation, 3.3% of arid land and less than 0.01% of city and Gavkhouni. During 2016-2024 about 32% of Gavkhouni lagoon area will disappear and arid lands will be replace them.