Extended Abstract 
Introduction
Over time, land co Extendedver and land use patterns of cities and villages undergo fundamental changes and transformations, and human and natural factors cause these changes and transformations. Thus, the present study aims to identify critical points of land use changes in Bandar Abbas city using methods to detect urban land changes such as image subtraction, image segmentation (NDVI index), principal component analysis, and maximum similarity classification (MLC) on TM images of 1990 and OLI images of 2020 of Bandar Abbas city. Land use maps have been produced in five categories: cliffs, barren lands, residential areas, urban vegetation, and wet lands. The results show that the use of barren lands in 2020 has decreased compared to 1990, and the largest change in this type of use is to residential areas in the northern and northeastern parts of Bandar Abbas city. Also, parts of barren lands have been converted into green spaces, so that green space has changed from 2.9 square kilometers in 1990 to 22.11 square kilometers in 2020. Therefore, the change in use related to barren lands has been negative and related to green space has been positive. In the wet lands section, there have been changes in the western part of the pier in Shahid Bahonar Port.
 
Methodology
In this study, in order to achieve the objectives of the first stage of the work, the aim is to identify unstable and critical areas in relation to land use change within the city of Bandar Abbas. To achieve this goal, in the first step, land use maps were extracted in the years 1990 and 2020. In the second step, land use for the year 2030 was predicted using the Markov chain model. One of the most important models for predicting the physical expansion of cities is the Cellular Automata-Marcov model, which is actually a combination of the cellular automata model and Markov chains in predicting cities, which is most widely used in predicting cities with greater accuracy. This model is tested in several stages using regression methods to have the best outputs in predicting the physical expansion of the city. The satellite images processed in this stage are the most important input data of the model. In the third step, by determining the weight of different land uses in the form of FANP method and using regional data, unstable and critical areas were identified in relation to land use change. Statistical and station data including meteorological, hydrological and demographic statistical data as well as social data collected from related data sources and classified in the form of a data bank. Topographic map of 1/50,000 of the region to extract slope and other information, geological map of 1/100,000 of the region to extract information on faults and rock types, use of remote sensing tools such as satellite images and aerial photographs with a distance from the ground to compare historical changes in urban development. Findings: In the present study, remote sensing techniques and geographic information systems were used to extract land use and vegetation cover of Bandar Abbas city. For this purpose, Landsat TM, ETM, and OLI satellite images of this region were downloaded in three frames in 1990 and 2020, then geographic, radiometric, and atmospheric corrections were performed using the ENVI 5.3 software package. Next, the NDVI index was calculated to detect and identify vegetation cover for the region. Then, using sampling and the maximum likelihood (ML) method and with the help of the NDVI index, remote sensing of satellite images was performed. The classification accuracy of the images was evaluated using systematic random sampling. For this purpose, 300 samples of classified images were collected and then their accuracy was evaluated using aerial photographs, land use and topographic maps, and satellite images for a specific land use type and an error matrix related to each class was formed, and the kappa index and overall accuracy were calculated. The overall accuracy index was calculated by dividing the number of pixels that were correctly classified for each of the aforementioned land use types by the total number of pixels in the sample examined. However, the kappa index calculates the accuracy rate, in contrast to the overall accuracy method, based on all pixels that were correctly and incorrectly classified. Future land use change prediction was performed using the Markov chain model. This model analyzes land use types at different times and produces data, transition matrices, and conditional probability values. The transition matrix represents the number of Pixels that change from one land use type to another in a given time unit. Therefore, the probability of changing from one land use type to another forms the data of the transition matrix. In the Markov chain model, land use change has random processes and different types of land use represent a state (condition) of the chain. Prioritization of
 
unstable and critical areas was carried out according to the five extracted land use types using the Fuzzy ANP method. For this purpose, questionnaires were initially completed to extract the pairwise comparison matrix of land use types from 30 environmental experts with sufficient knowledge of Bandar Abbas city.
Results and Discussion
The present study aimed to investigate and predict critical points of land use change under urban development on the outskirts of Bandar Abbas using Markov algorithms and remote sensing data. In this regard, the results showed that in all periods considered in the study, man-made land use has an increasing trend and green space use has a decreasing trend. The results also showed that a significant increase in the percentage of changes in water use was observed, which is indicative of intense construction around the sea in the years 1990-2020. Figure 1 shows the map of land use changes between 1990-2020. During this period, changes were made in rocky lands and then clay lands, and the water layer faced the least changes. This map shows the expansion of the city into peri-urban areas throughout the north of the city (from the northeast to the northwest); this could be due to the cheapness of land in the northern areas of the city compared to land near the coast. A notable point in these changes is the significant increase in the percentage of changes in water use, which indicates that construction on wetlands has increased during this period from 1990 to 2020; this could be due to the government's attention to beaches and ports for tourism and coastal trade. The table below shows the changes in each of the land cover classes to built-up land during the period 1990 to 2020. On the other hand, in the northern areas of Bandar Abbas, due to the possibility of city development, these peri-urban areas have been continuously seized during the study period. A point that should be particularly noted is that speculative activities in the land market have caused the growth of the city in the north of the city to progress rapidly.
Conclusion
In a final conclusion, it can be stated that further development of Bandar Abbas city is scattered and It has been spontaneous and has caused extensive land use changes. Constructions carried out around the sea and in areas with good climate have caused the emergence of heterogeneous and inconsistent urban expansion. In most cases, this incoherent process has been so great that it has caused significant changes in relation to changes in agricultural lands, and in general, the urban constructions of Bandar Abbas and the development process of this city have caused urban creep and changes in agricultural land use and the conversion of these lands to consumer and service use. Finally, the results of the research are consistent with the studies of Abbas and his colleagues (2020), Abedi (2012), Feizollah Pour (1403), based on the use of Markov models to simulate urban growth. It is also consistent with the studies of Mostafazadeh et al. (2022), Talebizadeh et al. (1401), based on the reduction of agricultural land and vegetation cover due to land use changes.