مدل‎سازی تغییرات پوشش سرزمین شهرستان تبریز با استفاده از شبکۀ عصبی مصنوعی و زنجیرۀ مارکف

نوع مقاله : مقاله کامل

نویسندگان

1 کارشناس ارشد محیط زیست، دانشکدۀ منابع طبیعی، دانشگاه تربیت مدرس، مازندران

2 استادیار گروه محیط ‌زیست، دانشکدۀ منابع ‌طبیعی، دانشگاه تربیت مدرس، مازندران

3 استادیار گروه محیط زیست، دانشکدۀ منابع طبیعی، دانشگاه تربیت مدرس، مازندران

4 کارشناس ارشد ارزیابی و آمایش سرزمین، دانشگاه پیام نور، تهران

چکیده

هدف از پژوهش پیش رو، مدل­سازی تغییرات کاربری اراضی شهرستان تبریز برای سال­های 1395 و 1400 با استفاده از مدل­ساز تغییر سرزمین (LCM) در محیط سامانۀ اطلاعات جغرافیایی است. برای این کار، تجزیه‎وتحلیل و بارزسازی تغییرات کاربری­ها، به‎کمک سه دوره از تصاویر ماهوارۀ لندست مربوط به سال­های 1367، 1380 و 1390 انجام گرفت و نقشه­های پوشش اراضی جداگانه‎ای برای هر سال تهیه شد. مدل­سازی پتانسیل انتقال، به‎کمک الگوریتم پرسپترون چندلایۀ شبکۀ عصبی مصنوعی با استفاده از شش متغیر مستقل صورت پذیرفت و میزان تخصیص تغییرات کاربری­ها به همدیگر، به‎روش زنجیرۀ مارکف مورد محاسبه قرار گرفت. نتایج حاصل نشان داد که در کل دورۀ مورد بررسی، یعنی بین سال­های 1367 تا 1390، حدود 5195 هکتار به وسعت مناطق شهری و مسکونی افزوده شده است که اراضی مرتعی به‎ویژه مراتع درجۀ یک، اراضی کشاورزی و درنهایت اراضی بایر و شوره­زار، به‎ترتیب با مساحت 3488، 1007 و 484 هکتار تبدیل اراضی، بیشترین سهم را در افزایش وسعت اراضی شهری و مسکونی داشته­اند. نتایج پیش­بینی پوشش اراضی نیز نشان داد که میزان توسعه و رشد شهری تبریز تا سال 1395 مساحتی برابر با 1037 هکتار و تا سال 1400 حدود 2033 هکتار خواهد بود.
 

کلیدواژه‌ها


عنوان مقاله [English]

Land Cover Changes Modeling of Tabriz Township Using Artificial Neural Network and Markov Chain

نویسندگان [English]

  • Mohammad Taheri 1
  • Mehdi Gholamalifard 2
  • Alireza Riahi Bakhtiari 3
  • Shahin Rahimoghli 4
1 M.S. in Environment, Department of Environment, Faculty of Natural Resources, Tarbiat Modares University, Noor, Mazandaran, Iran
2 Assistance Prof., Department of Environment, Faculty of Natural Resources, Tarbiat Modares University, Noor, Mazandaran, Iran
3 Assistance Prof., Department of Environment, Faculty of Natural Resources, Tarbiat Modares University, Noor, Mazandaran, Iran
4 Department of the Environment, Eastern Azerbaijan, Tabriz, Iran
چکیده [English]

Introduction
     Cities with their physical development are the major causes of land use and land cover changes which lead to many problems such as loss of agricultural land, loss of green space, water pollution, soil erosion and declining of environment quality. The main reason for physical development of cities is population growth in urban areas which enables the physical expansion of the cities in geographical space. Tabriz, as the center of East Azerbaijan Province, has experienced considerable changes in recent decades due to increasing population in the surrounding lands and rural communities located in regions of influence sphere. The present study was conducted for modeling of land use changes in Tabriz City for the years 2016 and 2021 by using the Land Change Modeler (LCM) in GIS environment.
Methodology
     The study area in this research is the Tabriz, as the center of East Azerbaijan Province in the north west of Iran. Extent of the area is approximately 2179 square kilometers and is situated in the geographical position of 45˚ 52' to 46˚ 34' of eastern longitude and 37˚ 46' to 38˚ 28' of northern latitude and the average altitude is about 1340 m above sea level. In order to produce land cover maps of the study area, Landsat satellite images of TM sensor for years 1989 and 2011 and ETM+ sensor for 2000 were used. In the following steps of production of land cover maps, the maximum likelihood supervised classification method was used as one of the most accurate methods of cell based classification. The first step in performing the supervised classification is introduction of training samples for each land cover class. By using the 1:50.000 topographic maps and visual interpretation of false-color images, the training samples for each land cover class were introduced. At the end of this phase, 6 classes of land cover were specified. These land cover classes were: 1. Irrigated agriculture and gardening, 2. Grassland (grade 1) and dry farming, 3. Grassland (grade 2), 4. Grassland (grade 3), 5. Salt marsh, and 6. Urban and residential areas. In order to assess the accuracy of produced land cover maps, 389 accuracy assessment points systematically in the networks in 2.5×2.5 km by using GPS were marked on the images. Finally, modeling of land cover changes by using LCM was done in four major steps:

Change Analysis
Modeling of Transition Potentials
Change Prediction and Modeling
Accuracy Assessment

Results and Discussion
     Analysis and detection of land cover changes during the study period between the years 1989 to 2011 showed that during this period, 5195 hectares of the area have been added to the urban and residential areas. The rangeland especially grassland (grade 1), agricultural lands and finally salt marsh lands have by 3488, 1007 and 484 hectares of land conversion, respectively. The land uses have the highest roles in the increasing the extent of urban and residential areas. It is also important to note that the rate of changes in the urban and residential areas during the period under study is not identical and the growth rate for urban and residential areas has increased in recent years. For example, from 1989 to 2000, 2271 hectares and between the years 2000 to 2011, 3062 hectares have been added to the extent of urban and residential areas. During this period, grasslands (grade 1) with 36958 hectares decrease in area and agricultural and gardening lands with 26617 hectares increase in area have the most of changes in the study area, respectively. Throughout the period studied, grasslands (grade 1) have been formed in most brigades within the city of Tabriz that from 1989 to 2000, 22944 hectares and from 2000 to 2011, 26803 hectares have shown reduction in area. The highest annual rate of land cover changes between the years from 1989 to 2011 were for urban and residential areas, so that this rate for urban and residential lands, agricultural and gardening lands, grasslands (grade 2) and salt marsh and bayer lands were calculated 3.16%, 2.33%, 1.46% and 1.18%, respectively . The grasslands (grade 1) and (grade 3) had a reduction with annual rate of -1.01% and -0.34%, respectively.
Conclusion
     The results obtained in this study showed that over the past 23 years, 5195 hectares have been added to  urban and residential areas that the shares of desirable agricultural and pasture lands have been very higher in comparison with salt marshes and bayer lands. In particular, growth and change rates of urban and residential areas in recent years have been faster. The results of this study indicate that the explosion and unbridled expansion of Tabriz based on the demands and needs of the urban population will cause widespread changes and developments in diverse areas. The most important of the changes can be observed in Tabriz integration with some of the surrounding villages. This ultimately will lead to a vast expanse of space with unit performance. Certainly, this vast expanse of space will lead to in addition of absence the certain physical boundary between urban and rural areas that the economic, cultural and other functions in each of them are not realized. Therefore, we can largely avoid future problems by compliance and enforcement of laws related to land use and participation in organizations, public institutions, trading firms and land speculators.
     The results of this study based on the nature of land cover changes in the city of Tabriz in the past, now and future can provide useful insights and perspectives for managers and urban planners to manage properly the land uses in Tabriz city. The results of this research can also be an introduction and complementary for environmental impact assessment projects, feasibility and site selection projects and other studies that seek to make identification of sensitive areas and vulnerable zones.
 

کلیدواژه‌ها [English]

  • Artificial Neural Network
  • Land Change Modeler (LCM)
  • Markov Chain
  • Tabriz