Analysis of Factors Affecting Urban Land Use Changes: A Case Study of Urmia City, Iran

doi:10.21203/rs.3.rs-4528684/v1

Download PDF

Research Article

Analysis of Factors Affecting Urban Land Use Changes: A Case Study of Urmia City, Iran

https://doi.org/10.21203/rs.3.rs-4528684/v1

This work is licensed under a CC BY 4.0 License

You are reading this latest preprint version

The intricate relationship between human activities and the natural environment has a profound impact on land use patterns. These changes, whether direct or indirect, often lead to a myriad of environmental challenges that jeopardize the well-being of our planet. Therefore, understanding the nature of these transformations and the factors driving them is crucial for ensuring the sustainability of life and mitigating the adverse effects of land use changes. This research adopts an applied approach and follows an exploratory methodology. To identify land use changes, satellite imagery from the years 1993, 2003, 2013, and 2023 was analyzed using the ENVI, GIS, and Google Earth software. The study area encompasses the city of Urmia and its surrounding areas, covering a total of 26,768.7 hectares. The findings reveal a dynamic pattern of land use changes over the three decades from 1993 to 2023.

Land use planning

Change detection

ENVI

Urmia City

Iran

Urban growth and land-use change are just two of the many factors that will shape our cities in the future (Alwedyan, 2022). Land, however, is the primary factor in urban development and expansion. Land is a unique and finite economic resource with diverse uses for various land uses that cannot be separated from the urbanization process (Piotr Borawski et al., 2019). As urban settlements expand, there are significant changes in land use and ecosystems, particularly the uncontrolled consumption of urban land, low-density urban sprawl, which leads to problems such as loss of biodiversity, landscape fragmentation, and increased energy consumption and carbon emissions (Linlin Lu et al., 2019). In general, urban land use refers to land with specific uses and functions in an urban complex and system, and the process of urban land-use growth and expansion is closely related to sustainable urban development (Wang et al., 2020). The entire world is witnessing a rapid urbanization process due to the remarkable growth of population, socio-economic development, and rapid urban expansion (Yuan et al., 2018). The rapid process of urban development and expansion has been characterized not only by population growth but also by the expansion of industrialization, increased economic and social activities, and intensified use of land resources. Urbanization has overstretched the use of infrastructure and services and has led to shortages and inequities in access at the city level, resulting in a large proportion of the population living in slums and marginalized peri-urban areas. With land-use changes, forests and agricultural land are often converted into developed urban land-use features and human infrastructure. Urban land uses related to urban growth and expansion include buildings, roads, airports, and others. They should have the useful and necessary characteristics to preserve and support the growth of agricultural areas, water bodies, and forests. Land-use change in the world has been a significant factor in environmental degradation, occurring at a faster rate than population growth. On the other hand (M. Hersperger et al., 2018), land-use change during urban growth is increasingly complex and difficult to regulate.

Urban land-use planning examines how urban activities and functions are used, distributed, protected, organized, and spatially arranged based on the desires and needs of the urban community (Ziari, 2004). Therefore, urban land-use planning can always be used to prevent and manage urban sprawl issues. The process of changes and the identification of effective factors and their results can lead to logical understanding, analysis, and problems of dynamic and variable cities, and can be used as a tool for urban managers and planners to predict future trends and ultimately manage them effectively (Ghorbani et al., 2014). The process of urbanization has increased the complexity, heterogeneity, and fragmentation of the urban expansion pattern. Therefore, analyzing the transformations of urban growth and expansion and the factors affecting it in urban processes is necessary and can provide scientific support for principled and logical urban spatial planning and promote coordinated socio-economic, environmental, and other urban indicators (Deng et al., 2022). With the increase in population and economic and cultural growth, the city of Urmia has experienced significant expansion since the Islamic Revolution. Following urban expansion, there have always been many land-use changes at the city level and around the city limits.

Urmia, the tenth most populous city in Iran, has experienced substantial growth since the Islamic Revolution. Its population has surged from 357,399 in 1991 to 736,224 in 2016. This rapid expansion has been accompanied by significant land-use changes, driven by natural population growth, rural-urban migration, and land speculation. The uncontrolled physical growth, high-rise construction, and unplanned expansion of construction zones have resulted in a plethora of urban challenges, including environmental pollution, traffic congestion, air pollution, and a host of other issues.

Despite the lack of adequate amenities, essential services, and basic infrastructure in Urmia's peri-urban areas, the city continues to attract a steady influx of migrants. This relentless population growth, coupled with the rapid physical expansion of the city, has triggered a cascade of problems, including excessive migration to urban fringe areas, infrastructure deficiencies, urban sprawl, urban heat islands, environmental degradation, the uncontrolled expansion of slums and peri-urban areas, agricultural land destruction, and a host of other challenges.

The lack of comprehensive data management and analysis in Urmia hinders effective assessment of the current situation and impedes future planning efforts. To effectively address the city's myriad challenges, a comprehensive and holistic urban land-use plan is urgently needed. This plan should guide and control Urmia's expansion, identify the factors and drivers influencing land-use changes within and outside the city, and facilitate sustainable urban growth and development. By understanding the current and historical patterns of urban land use, the decision-making process can be optimized to meet the needs of Urmia's urban society and pave the way for a healthy and sustainable urban future.

Urbanization is a global phenomenon, with Asia being the fastest urbanizing region in the world (Bakker, 2022). Urbanization has led to changes in economic, social, and environmental aspects (Jonathan Dolley and at al, 2020). Land use change occurs not only within the city limits but also in the surrounding peri-urban or rural areas to support the urban area. This surrounding area of the city is defined as a peri-urban area (Dipanwita Dutta and et all, 2020). When a city is built, the physical form and land use patterns can remain constant for generations, leading to unsustainable expansion. Urban land consumption is outpacing population growth by up to 50%, and 1.2 million square kilometers of new urban area are expected to be added to the world by 2030 (Biello, 2018). Today, about 56% of the world's population, 4.4 billion people, live in cities; this trend is expected to continue, with the urban population doubling in size by 2050 and, at that time, nearly 7 out of 10 people will live in cities. With over 80% of the world's GDP generated in cities, urbanization can contribute to sustainable growth if well managed through increased productivity and innovation. However, the speed and scale of urbanization pose challenges such as meeting the rapid demand for affordable housing, sustainable infrastructure including transportation systems, basic services, and jobs, especially for the nearly one billion urban poor living in informal settlements. Growing conflicts are helping to put pressure on the over 50% of internally displaced persons living in urban areas (World Bank-Urban Development 2023).

Urban land-use changes and urban development play a crucial role in addressing climate change, as exposure to climate hazards and disasters increases with urban growth and sprawl. Since 1985, around 76,400 square kilometers of newly urbanized land have been added in locations with flood depths exceeding 0.5 meters during extreme floods. This is equivalent to 50 times the area of Greater London. Worldwide, 1.81 billion people (or one in four) live in flood-prone areas. Exposure to risk is particularly high in densely populated river deltas and rapidly urbanizing coastal zones in developing countries, where 89 percent of the world's flood-exposed population lives. Cities are also on the frontlines of fighting epidemics. The COVID-19 pandemic has been a major challenge for cities and their citizens, both rich and poor. Its impact and the measures taken to control the spread of the virus have had a disproportionate impact on the poor, marginalized and vulnerable, exposing fault lines in the economic structure of cities and their preparedness for such a crisis, particularly their public health status and service delivery systems, which have led to increased construction in peri-urban areas around cities. National and local governments currently play a critical role in shaping the future of their own development and creating opportunities for all.

This study adopts a descriptive-analytical approach, employing an exploratory research design with both quantitative and qualitative methodologies. Data collection involves library research, document analysis, field studies, and the utilization of comprehensive urban plans, statistical centers, and municipal data from various sources such as the municipality of Urmia, the Environmental Protection Agency of West Azerbaijan, the Governorate of West Azerbaijan, the Department of Roads and Urban Planning, the Provincial Budget Planning Organization of West Azerbaijan, the Iranian Statistical Center, and satellite imagery. The research is structured into three main phases:

1) Examination and analysis of changes over the past three decades.

2) Analysis of spatial growth and urban development in both past and present contexts.

3) Analysis of factors influencing changes in land use.

In this research, Landsat and OLI satellite images from 1992 to 2023 are initially processed in ENVI software, incorporating various indices after geometric and atmospheric corrections. Land use patterns are then analyzed using Google Earth and GIS software, enabling the identification and assessment of land use changes. Final land use maps are generated and the observed changes in the years 1992, 2003, 2013, and 2023 are compared and evaluated. For the analysis of factors influencing land use, statistical data from the National Statistical Portal are utilized, along with Fuzzy Delphi method and factor analysis techniques implemented in SPSS software.

3.1. Study Area

Urmia is one of Iran's major cities, serving as the capital of West Azerbaijan Province and Urmia County in northwestern Iran, Situated in the Azerbaijan region. According to the 2016 census, the city had a population of 736,224, making it the tenth most populous city in Iran and the second most populous city in northwestern Iran. Urmia, located at an altitude of 1332 meters, lies to the west of Lake Urmia, nestled amidst the foothills of the Sīr Mountains and within the Urmia Plain (Fig. 1).

Urmia County is one of the twelve counties in West Azerbaijan Province, situated in its central region. It is bordered by Salmas County to the north, Naghadeh and Maku counties to the south, Lake Urmia to the east, and the Iran-Turkey border to the west. With an area exceeding 5251 square kilometers, the county encompasses approximately 14% of the province's total area.

From a national administrative perspective, Urmia County is divided into 5 districts, 20 rural districts, 5 cities, and 615 settlements. The city of Urmia itself is located approximately 18 kilometers away from Lake Urmia, positioned at 45 degrees and 4 minutes east longitude and 37 degrees and 33 minutes north latitude. It spans 70 kilometers in length and 30 kilometers in width within a plain.

Due to its geographical location and its position along the Caucasus-Armenia-Mesopotamia route, as well as being a northern branch of the Silk Road, Urmia County (as well as the city) has been subject to numerous invasions and has endured significant damages throughout history. However, this very position has also led to ethnic-cultural diversity and the presence of significant historical heritage in the region.

To quantify land use changes within the study area, satellite images (Landsat) with minimal cloud cover from Landsat 8/9, Landsat 7, and Landsat 5 satellites were utilized from the Landsat Collection 2 – Level 1 dataset. Radiometric corrections were performed to rectify existing errors and atmospheric distortions using ENVI, GIS, and Google Earth software. Initially, land uses were classified into 4 and 5 classes for the years 1993, 2003, 2013, and 2023. The accuracy of classification was calculated, and the area of each land use type for each year was determined.

Subsequently, in Google Earth software, samples were selected for all images separately, and classification methods such as Maximum Likelihood, Minimum Distance, Neural Network, and Support Vector Machine were employed to obtain the best Kappa coefficient. This process allowed for the identification of the most accurate method with the highest Kappa coefficient for determining changes that occurred over three decades, delineated in ten-year intervals. This analysis aimed to reveal which land use types converted to others and the corresponding areas during each decade.

Classification: In this paper, to determine land uses in the years 2023 and 2013, four land use classes were used: Urban (man-made), Gardens (forest), Agricultural lands (plowed) and Barren lands (mountains, etc.)

Additionally, for the years 2003 and 1993, five classes were utilized, including the aforementioned four classes plus a class for water bodies. For the classification of images in the four time periods, a supervised classification method was employed. For each class, 70 samples were used, totaling 280 samples for the 2023 and 2013 maps and 350 samples for the 2003 and 1993 images.

Class Separability

The accuracy of separating the selected classes depends on the accuracy, number, and size of the samples. To compare classes and selected samples, the Compute ROI separability command is used, which yields results in a table or matrix around the number 2. If the values are closer to 2, it indicates a higher possibility of separating classes from each other with greater ease. However, as the distance from the number 2 increases, the homogeneity of the area decreases, making it harder to discern phenomena or classes, and the likelihood of errors increases.

The fundamental point in producing land use maps lies in the precision of determining the training samples (first sample) and, importantly, the validation sample (essentially based on a thorough understanding of the region through either physical survey or software such as Google Earth). Furthermore, increasing the number of selected pixels for each category in each sample will enhance the accuracy of the generated raster map. This is reflected in the ROI Separability report, where in this study, the minimum separability value is 1.94150149 for the barren lands with agriculture in 1993, 1.995442302 for gardens with agriculture in 2003, 1.98556922 for agricultural lands with barren lands in 2013, and 1.99945853 for gardens with urban lands in 2023.

Validation using Kappa Coefficient

The Kappa coefficient measures accuracy inversely to the overall accuracy based on all pixels that are correctly and incorrectly classified. In the current paper, Kappa coefficients above 79% along with appropriate accuracy of classified images will yield acceptable results. Following the classification, classes are supervised and necessary sampling is done using Google Earth software, which requires precise knowledge of the study area. Sampling for each image is done separately; for each class in ENVI, 70 samples are taken, and for images of 2023 and 2013 in Google Earth Pro, 50 samples per class are taken, totaling 200 samples for each image. For images of 2003 and 1993, considering they have five classes, 250 samples are selected, totaling 900 samples for all four images and all classes. Specifically, 400 samples are for images of 2013 and 2023, and 500 samples are for images of 1993 and 2003 in Google Earth. In total, approximately 1260 selected samples are drawn. The results, including the concordance of these samples in terms of error percentage and Kappa coefficient for each year using the introduced methods, are presented in the following tables

Table 1

Results of error matrix and Kappa coefficient in 1993
	Overall accuracy	Number of errors	kappa
Maximum likelihood	85.42%	164/192	0.8114
Minimum-distance	71.3542	137/192	0.6303
Neural network	80.2083	154/192	0.7402
Support vector machine	80.2083	154/192	0.7429

Based on the Kappa coefficients obtained, the best classification method for the year 1993 (Table 1) is the Maximum Likelihood method, with a Kappa value of 0.81 and an accuracy of 85%. Out of 192 selected samples, 164 samples were accurately classified while 28 samples were misclassified. Among these, 56 samples were urban areas, 47 samples were agricultural land use, 31 samples were barren land, 45 samples were gardens, and 13 samples were water bodies. The highest error occurs between the barren land and agricultural land use classes, while the lowest error is observed in the urban land use class, with an accuracy of 93%.

Table 2

Results of error matrix and Kappa coefficient in 2003
	Overall accuracy	Number of errors	kappa
Maximum likelihood	75.3676	205/272	0.6782
Minimum-distance	59.5588	162/727	0.4837
Neural network	69.8529	190/272	0.6075
Support vector machine	67.6471	184/272	0.5855

For the land use classification of the year 2003, 272 samples were selected. The Maximum Likelihood method exhibits the least error, with a Kappa coefficient of 0.67 and an accuracy of 75% (Table 2). Out of these samples, 205 samples were consistent with the Google Earth imagery. Specifically, 64 samples were urban areas, 86 samples were agricultural land, 40 samples were barren land, 73 samples were gardens, and 9 samples were water bodies. Gardens achieved the highest accuracy at 99%, followed by agricultural land at 79%, urban areas at 78%, barren land and water bodies at 48% and 42% accuracy respectively. However, in terms of error, it's notable that barren land and water bodies exhibit relatively higher inaccuracies, which could potentially result in significant inaccuracies in area estimation.

Table 3

Results of error matrix and Kappa coefficient in 2013
	Overall accuracy	Number of errors	kappa
Maximum likelihood	79.4425	228/287	0.7247
Minimum-distance	76.3066	219/287	0.685
Neural network	90.2439	259/287	0.8699
Support vector machine	89.547	257/287	0.86.6

Considering the classifications conducted using the four methods, the neural network emerges as the best approach with an accuracy of 90.24% for the image of the year 2013 (Table 3). The error rate for the four classes is 10%, which is acceptable, resulting in a Kappa coefficient of 0.8625. This coefficient is deemed acceptable for validation and matching the selected classes with reality. According to the Kappa coefficient table, out of the 28 errors observed, 7 errors occurred between urban areas and agricultural land, barren land, and gardens; 8 errors occurred from agricultural land to urban areas, barren land, and gardens; 9 errors occurred from barren land to agricultural land and gardens, and 4 errors occurred from gardens to both barren land and agricultural land classes. Urban areas had a 91% accuracy rate, agricultural land had an 89% accuracy rate, barren land had an 88% accuracy rate, and gardens had a 95% accuracy rate.

Table 4

Results of error matrix and Kappa coefficient in 2023
	Overall accuracy	Number of errors	kappa
Maximum likelihood	79.4425	228/287	0.7247
Minimum-distance	84.4203	233/276	0.792
Neural network	92.029	254/276	0.8932
Support vector machine	92.3913	255/276	0.8981

Among the employed methods, the neural network appears to be the most suitable approach for the image of the year 2023, achieving the highest possible accuracy (Table 4). Out of 276 selected samples, 254 were correctly classified, while 22 samples contributed to errors across all classes. The overall measurement accuracy for this image was 92%, with a corresponding Kappa coefficient of 0.89.

In this image, urban areas had a 90% correct selection rate, agricultural land had a 64% correct selection rate, barren land had an 88% correct selection rate, and gardens had a 97% completely correct selection rate. The distribution of errors in terms of percentages is as follows: the highest error rate, at 2.63%, occurred between agricultural land and urban areas and gardens. Barren land had a 12% error rate with agricultural land, gardens had a 3% error rate with agricultural land, and urban areas had the highest error rate of 8% with barren land.

4.1. Image Classification

Classification of the 1993 Image

For the image of the year 1993 with 5 land use classes, considering the changes and classifications conducted, the Maximum Likelihood method, with the highest Kappa coefficient of 0.81, is the most suitable method for classifying this image (Fig. 2).

Table 5

classified map data in 1993
Name	Count	Area	Percent
Urban	4427	4630.63	17.30
Agriculture	4062	6448.03	24.1
Water	805	227.26	0.85
Barren	6050	5210.61	19.47
Gardens	1296	10252.15	38.3
Total	16440	26768.7	100.0

Based on the data provided in the table, the areas of the five land use classes - urban, agricultural, water, barren, and gardens - with pixels measuring 30*30 square centimeters, are respectively 4630.63, 6448.03, 227.26, 5210.61, and 10252.15 hectares out of a total area of 26768.7 hectares. These areas cover the specified region, with gardens covering the largest area at 38.3% and water covering the smallest area at 0.85%, which is linear and only represents the Silvana River. Agricultural land, barren land, and urban areas represent 17.30%, 19.47%, and 24.1% of the total land, respectively (Table 5).

Image Classification 2003

For the image of the year 2003 with 5 land use classes, the Maximum Likelihood method has been utilized. According to the map and its legend, urban areas are represented in red, gardens or forest-like areas in green, barren land in light blue, river water areas in light blue, and plowed agricultural land, with wheat and barley, in yellow (Fig. 3).

Table 6

classified map data in 2003
Name	Count	area	Percnt
Urban	4778	5569.77	20.81
Water	509	89.21	0.33
Agriculture	10861	7854.64	29.34
Barren	7420	4254.98	15.90
Gardens	8125	9000.08	33062
Total	31693	26768.7	100.00

The satellite image of the year 2003 comprises 5 land use classes with a total area of 26768.7 hectares, with pixels measuring 15*15. Urban areas cover 5569.77 hectares, water areas cover 89.1 hectares, agricultural land covers 7854.64 hectares, barren land covers 4254.98 hectares, and gardens cover 9000.08 hectares. Barren land constitutes the largest area at 33.62%, while water areas constitute the smallest area at 0.33%. Gardens, agricultural land, and urban areas account for 20.81%, 29.34%, and 15.90% respectively. Compared to 1993, the changes in each class are as follows: a decrease of 0.57% in barren land, an increase of 5.25% in agricultural land, a decrease of 4.68% in gardens, a decrease of 0.52% in water areas, and an increase of 3.51% in urban or human-made areas (Table 6).

Image Classification 2013: For the image of the year 2013 with 4 land use classes, the best method is the neural network. The classes are well separated from each other, with the coloring as follows: urban areas are depicted in yellow, garden or forest-like areas in light green, barren land in gray, and plowed agricultural land, with wheat and barley, in light green (Fig. 4).

Table 7

classified map data in 2013
Name	Count	area	Percent
Urban	8999	5501.27	20.55
Agriculture	19762	8585.91	32.07
Barren	5270	6104.7	22.81
Gardens	10453	6576.7	24.57
Total	44484	26768.7	100

The changes in the year 2013 are as follows: initially, the reduction in the number of classes from 5 to 4 is due to the fact that after the completion of the Silvana Dam project, the Silvana River, which flows from Urmia city to Lake Urmia, partially opens its reservoirs solely for agricultural purposes. Consequently, the riverbed remains dry for most seasons and days of the year. Therefore, starting from 2013, we have removed the water class. Subsequent changes are as follows: urban areas increased by 20.55%, from 5569.7 hectares to 5501.57 hectares. Agricultural land, in terms of cultivation, experienced a 2.73% change, increasing from 7854 hectares to 8585 hectares. Barren land also increased by 6.9%, from 4254 hectares to 6104 hectares, and gardens decreased by 9.05%, from 9000 hectares in 2003 to 6576 hectares in 2013.

The satellite image of 2013, with pixels measuring 15*15 square centimeters, covers a total area of 26768.7 hectares. Agricultural land constitutes 32.07% of this area, gardens cover 24.57%, urban areas cover 20.55%, and barren land covers 22.81% (Table 7).

Image Classification 2023: Similar to the image of 2013, the image of the year 2023 also consists of 4 land use classes, and the best method employed is the neural network. The classes are well distinguished from each other, with the following color scheme based on the map and its legend: urban areas (urban 2003) are depicted in yellow, gardens or forest-like areas (gardens 2003) in light green, barren land in gray, and plowed agricultural land, with wheat and barley (agriculture), in light green (Fig. 5).

Table 8

classified map data in 2023
Name	Count	Area	percent
Urban	9832	7697.96	28.76
Agriculture	19527	8652.81	32.32
Barren	10533	5930.24	22.15
Gardens	5118	4487.68	16.75
Total	45010	26768.7	100

In 2023, relative to the year 2013, there have been notable changes in the four existing land classes over a span of 10 years. Urban lands have experienced an 8.21% increase in area, expanding from 5501 hectares to 7697 hectares. Agricultural lands have seen a negligible increase of 0.25%, maintaining a relatively stable figure over the decade. Barren lands have decreased by 0.65%, shrinking from 6104 hectares to 5930 hectares. Additionally, gardens have witnessed a significant decline of 7.80%, decreasing from 6567 hectares in 2013 to 4487 hectares in 2023. In 2023, out of the total area of 26768.7 hectares, agricultural lands cover 8652.81 hectares, gardens cover 4487.68 hectares, urban and human-made lands encompass 7697.96 hectares, and barren lands contribute 5930.24 hectares to this total. This distribution is observed within a grid of dimensions 15*15 square centimeters (Table 8).

The results obtained from the four periods of 1993, 2003, 2013, and 2023 are as follows:

Table 9

Land use area in 4 time periods (1993-2003-2013-2023)
Name	Urban	Agriculture	Barren	Gardens	Water
1993	4630.63	6448.03	5210.61	10252.15	227.26
2003	5569.77	7854.64	4254.98	9000.08	89.21
2013	5501.27	8585.91	6104.7	6576.7	-
2023	7697.96	8652.81	5930.24	4487.68	-
Min 2023 − 1993	3067.33	2204.78	719.63	5764.47-	-

According to the Fig. 6, over three decades from 1993 to 2023, the changes in land use within the studied area of 26768.7 hectares are as follows: riparian lands for the Silvana River, which were only flowing in 1993 and 2003 due to the completion of the Silvana dam construction, are now only used for agricultural purposes. Therefore, since 2003, water reservoirs are only opened in spring and summer months, and the riverbed also passes linearly through the city center, its width has decreased, and urban constructions and vegetation cover have been established around it. This classification was omitted in the 2013 and 2023 images.

Urban or human-made lands, which are among the most significant objectives of this study, have consistently shown an upward trend from 1993 to 2023, as indicated in the chart. There has been significant growth in such a way that during these three decades, an area of 3067 hectares has been added to it, approximately doubling its area compared to 1993, representing a 66% increase in area over these three decades. This expansion consistently leads to environmental degradation, destruction of natural lands, steppes, and gardens.

Agricultural lands, typically cultivated with wheat, barley, and chickpeas within the study area, have also undergone incremental changes. Due to the proximity to Urmia city and expansion beyond its boundaries, coupled with advancements in agricultural tools, more steps are being cultivated, and more natural lands are being destroyed. Specifically, these lands increased from 6448 hectares in 1993 to 8652 hectares in 2023, representing a 34% increase over the past three decades. Steppes increased from 5210 hectares in 1993 to 5930 hectares in 2023, partly due to methodological errors.

The most significant degradation and transformation occurred in the classification of gardens, which consistently showed a declining trend. Its area decreased from 10252 hectares in 1993 to 4487 hectares in 2023, marking a 56% reduction compared to the total area. It can be said that this area halved in 2023 compared to 1993. Most of the changes observed are related to the conversion of gardens to urban or human-made lands and agricultural lands (Table 9).

4.2. Investigating Changes

Using the Thematic Change Workflow, we derive land use changes from 1993 to 2023 by comparing images from different periods. Initially, we compare the images of 1993 with 2003, then the changes obtained from the 2003 images are compared with those from 2013, and finally, we compare the images of 2013 with 2023. Ultimately, to ascertain the land use changes over the past three decades within the study area, we utilize a comparison between the images of 1993 and 2023.

Changes from 1993 to 2003

In the examination of land use changes within the study area, according to the land use change map of 2003 compared to 1993, alterations have occurred in 35 different scenarios among the defined 5 land use change classes. However, in 7 instances of these classifications, where no changes have occurred, the least change is attributed to barren lands remaining as barren lands, amounting to 18 pixels with an area of 3.05 hectares. The most substantial land use transformation over the span of 10 years for barren lands amounts to 621.91 hectares, converted into agricultural lands. This transformation is primarily due to human activity, as barren lands surrounding and extending beyond the urban limits have been cultivated, leading to environmental degradation.

Agricultural lands have experienced minimal changes, with 9.17 hectares transitioning from agriculture to water bodies, and the most significant change being 775.8 hectares of agricultural lands remaining uncultivated and barren. It should be noted that due to seasonal fluctuations when crops are not cultivated, most agricultural lands are identified as barren during certain months of the year.

The least changes observed pertain to water bodies, particularly the Silvana River basin, where the least alteration along its course amounts to 11.19 hectares, converted into barren lands. This segment could be attributed to the depletion of water resources in this area, leading to the emergence of barren lands. The most significant transformation involves the conversion of these lands into urban areas, amounting to 92.64 hectares.

Subsequently, changes in gardens exhibit the least alteration, with 17.67 hectares transitioning into water bodies, while the most substantial change involves their conversion into gardens in the vicinity of Urmia city, totaling 832.36 hectares.

The most crucial changes pertinent to this study are those associated with urban and human-made lands. It is imperative to understand the predominant types of transformations occurring within urban areas and identify the predominant land use changes. The least urban land use change amounts to 9.2 hectares transitioning into water bodies, converted into riverbeds. Conversely, the most significant change amounts to 490 hectares transitioning into gardens. It can be inferred that urban areas have experienced more extensive development than expansion into surrounding areas over the ten-year period from 1993 to 2003 (Table 10, Fig. 7).

Furthermore, the most substantial urban degradation has occurred on the outskirts of the city, with changes from suburban to agricultural and gardens, indicating a population influx towards the city center, resulting in high population density. The largest area of land transformed into urban areas is attributed to the conversion of gardens in the northern and northeastern parts of Urmia, totaling 781.13 hectares. Conversely, the least transformation into urban areas amounts to 92.64 hectares, transitioning from water bodies to urban and human-made lands.

Table 10

Land use change from 1993 to 2003
Name	Urban	Agriculture	Barren	Gardens	Water
Urban	3635	353.87	140.93	490.91	9.2
Agriculture	666.85	4401	775.8	557.2	9.17
Barren	454.33	1671.07	2384.19	621.91	3.05
Gardens	781.13	1268.47	832.36	7380.77	17.67
Water	92.64	23.4	11.19	46.7	40.06

Changes from 2003 to 2013

The changes that occurred from 2003 to 2013, according to the table and land use change map, indicate that the least alterations in the water bodies class amount to 1.93 hectares, which have been converted into agricultural lands, while the most significant changes in this class amount to 46.09 hectares, transformed into barren lands in 2013.

In the agricultural land’s class of 2003, the least changes amount to 1125 hectares, converted into barren lands, while the most substantial alterations in this class amount to 522 hectares, transformed into gardens. This implies that more than 522 hectares of agricultural lands were destroyed and converted into gardens from 2003 to 2013 (Table 11).

The least changes in the barren lands class amount to 202.75 hectares, converted into urban areas, while the most significant alterations in this class over the ten-year period amount to 1621.58 hectares, transformed into agricultural lands. These changes predominantly occur in the northern and southwestern outskirts of Urmia city. The most minor changes in the gardens class amount to 523 hectares, converted into urban and human-made lands, while the most significant alterations in this class over the ten-year period occur in barren lands. However, this change is not in the form of gardens disappearing entirely; rather, it is due to the preservation of garden land use itself, but changes arise due to alterations in irrigation methods and the age of gardens.

The most significant changes occur in urban lands, with the least area of change being 8.946 hectares, converted from human-made lands into agricultural lands, while the most substantial alterations amount to 110.997 hectares, transformed into barren lands. Most of these changes manifest as access roads between urban areas and their surroundings. Furthermore, during this period from 2003 to 2013, the most significant changes in transitioning to urban lands amount to 1901745 hectares, converted into urban areas (Fig. 8).

Table 11

Land use change from 2003 to 2013
Name	Urban	Agriculture	Barren	Gardens	Water
Urban	3550.87	234.3	1695.28	185.43	-
Agriculture	1376.82	4752	1125.53	522.56	-
Barren	202.75	1621.58	1842.84	490.15	-
Gardens	523.14	1693.95	1851.55	5017.21	-
Water	1.93	14.46	46.09	19.52	-

Changes from 2013 to 2023

During the years 2013 to 2023 (Table 12), in the agricultural lands class, the least changes occurred, amounting to 313.16 hectares, which were converted into barren lands, while the most significant changes amounted to 1020.55 hectares, transformed into urban and human-made lands. In the gardens class, the least area of change was 80.49 hectares, converted into urban and human-made lands, while the most substantial change was 598 hectares, converted into agricultural lands.

In the urban lands class during this period, the least transformation occurred, with an area of 19.51 hectares converted into gardens, while the most significant change, with an area of 346 hectares, was converted into barren lands. Most of these alterations occurred in large areas outside the city, such as access roads and abandoned workshops, which were transformed into barren lands. Barren lands experienced the least conversion to urban areas, with an area of 905 hectares, while the most significant changes, amounting to 2959 hectares, were converted into agricultural lands (Fig. 9).

Over the course of these ten years, the least area converted to urban lands was 80 hectares, transformed from gardens into urban areas. The most significant changes to urban lands occurred in the outskirts and urban expansion areas, with an area of over 1020 hectares converted from agricultural lands into urban areas. In total, more than 200 hectares of garden, barren, and agricultural lands were transformed into urban areas over the ten-year period, indicating urban expansion within Urmia city.

Table 12

Land use change from 2013 to 2023
Name	Urban	Agriculture	Barren	Gardens	Water
Urban	5851.44	199.78	346.02	19.51	-
Agriculture	1020.55	4067.72	313.16	377.76	-
Barren	905.1	2959.96	4671.63	924.11	-
Gardens	80.49	1127.11	598.32	3305.96	-
Water	-	-	-	-	-

Changes from 1993 to 2023

Over the three decades from 1993 to 2023 (Table 13), the summary of changes for five land use classes between 1993 and 2003, and four land use classes between 2013 and 2023 is as follows

For urban lands, the least changes occurred, amounting to 233.33 hectares converted into gardens, while the most significant area transformed, totaling 1384 hectares, was converted into barren lands. In agricultural lands, the least changes observed were 165 hectares converted into gardens, whereas the most significant alteration amounted to 2518 hectares transformed into urban areas.

In the water class, measured only over two decades, the smallest conversion area was 20.65 hectares, transformed into urban lands, while the most substantial changes were 121.39 hectares converted, respectively, into urban and barren lands.

Throughout these 30 years, barren lands experienced the least changes, with 295.31 hectares converted into gardens, and the most significant alterations, with 2183 hectares converted into agricultural lands. Gardens had the least changes, with 1546 hectares converted into urban lands, and the most substantial changes, with 2887 hectares converted into gardens (Fig. 10).

The most considerable changes occurred in the urban and human-made land use class, with the least transition observed in water class, converting 20.65 hectares into urban lands, and the most significant change, with 2518 hectares converted from agricultural to urban lands. Moreover, over these three decades, the least change occurred from water to urban lands, with 20.65 hectares, and the most significant transformation was from gardens to agricultural lands, with 2887.41 hectares.

Table 13

Land use change from 1993 to 2023
Name	Urban	Agriculture	Barren	Gardens	Water
Urban	2691.8	316.34	1384.08	233.33	-
Agriculture	2518.85	2973.33	763.82	165.79	-
Barren	1059.81	2183	1622.62	295.31	-
Gardens	1546.21	2887.41	1972.77	3834.4	-
Water	20.65	34.93	121.39	43.67	-

4.3. Urban Areas

According to the results obtained from land use changes, urban areas do not deteriorate over time and are converted to other uses. Therefore, urban areas should increase compared to previous periods. To rectify this error, we utilized the ERASE tool in GIS software to minimize these errors on urban land parcels according to the maps (Fig. 11).

In 1993, the urban land area was 3953.95 hectares, which increased to 5179.74 hectares in 2003. By 2013, the urban or human-made land area had reached 5411.37 hectares, and finally, in 2023, it reached 7697.96 hectares. In the comprehensive plan of Urmia city in 2016, the urban area boundary was declared as 10,000 hectares. However, the results obtained indicate discrepancies due to processing errors, unclear delineation of the urban boundary, and the absence of a road network matching the declared boundary in the Urmia comprehensive plan.

4.4. Factors

To identify the factors influencing land use changes, various dimensions and reasons for urban growth and expansion at a global level were identified after research in various sources. In this regard, the Fuzzy Delphi method was used to achieve this goal. By collecting opinions from urban experts in Urmia, several steps were taken to identify the factors affecting land use changes in Urmia through defining the research problem, dimensions, and various components using factor analysis. The criteria introduced by the experts categorized these factors into 6 dimensions as follows (Table 14). In these 6 categories, the first three dimensions are particularly important and are the main drivers of changes, while the other three dimensions are supportive and complementary to the first three dimensions, hence not significantly impactful on land use changes.

Table 14

Defined Dimensions
Criteria	Explanation
Social Dimension (Population)	Rate, growth, and density of the population
Living Standards	Economy, income levels, and livelihood of individuals
Technological Advancement	Transportation, construction of power plants, industries, etc.
Political Economy	Production and distribution of income and public wealth in various regions of Urmia
Political Structure	Political system, role of the government, plans and policies, management
Attitudes and Values	Culture, customs, and traditions of the city's ethnic groups and religions (Urmia)

By selecting these 6 dimensions according to the experts' opinions, the main criteria for each dimension were determined. Then, through reviewing the literature of various research studies and investigations and considering the opinions and suggestions collected from 60 experts, 60 indices were introduced for each of these dimensions. To identify the effective factors, 25 components out of the mentioned 60 sub-criteria entered the analysis. The oblique rotation method was used for Factor analysis, determining the factors based on the assumptions of the test, with eigenvalues of 0.5 or 1. The aim was to identify the best model and eliminate inefficient variables with lower factor loads that had insignificant correlations with other variables. Meaningful factor loads with higher significance coefficients in each factor analysis round, exceeding 0.5, were selected for a more precise and better analysis. As a result, 6 factors were extracted, encompassing 19 variables. Variables related to land use changes were categorized into these 6 factors based on their eigenvalues.

Table 15

Eigenvalues, Percentage of Variance, and Cumulative Percentage
	Initial Eigenvalues			Sum of Squared Loadings			Sum of Squared Loadings After Rotation
	Total	Percentage of Variance	Cumulative Percentage	Total	Percentage of Variance	Cumulative Percentage	Total
1	4.625	18.499	18.499	4.625	18.499	18.499	3.908
2	3.627	14.509	33.008	3.627	14.509	33.008	3.521
3	2.941	11.765	44.773	2.941	11.765	44.773	3.185
4	2.402	9.61	54.383	2.402	9.61	54.383	2.582
5	2.294	9.175	63.558	2.294	9.175	63.558	2.572
6	1.929	7.716	71.274	1.929	7.716	71.274	2.728

Table 15 represents the number of factors extracted from the data (initial variables). Factors in this section are considered influential if their characteristic value (sum column) is greater than one. The last column of this table indicates the percentage of variance explained by all factors (from the first factor to the current factor) together. According to the table, 6 factors have been extracted, which collectively account for 71.274% of the variability of the main variables. The total variance for each test is equal to 100%. The closer this value is to 100, the better the interpretation of the number of factors. The eigenvalue for the first factor is 4.625. Other eigenvalues for subsequent factors are also listed in the Total column.

The scree plot (Fig. 12) graphically displays the eigenvalues of each of the extracted components, starting from the largest eigenvalue. It consistently exhibits a descending slope. The scree plot indicates which factor had a noticeable change. It's quite evident that the initial 3 factors have a steeper slope compared to the second set of 3 factors. In fact, the results from the Total Variance Explained table are observed graphically in the scree plot, revealing which factor contributes to a higher percentage of variability in the variables. It's apparent in the plot that the initial 3 factors encompass more variability, covering 44% of the total, while the subsequent 3 factors also cover nearly 27% of the variability, as visible in the plot. Considering the scree plot, factors 7 and 8 could also be taken into account.

Based on the changes from 1993 to 2003, the smallest change was from barren land to water land, covering an area of 11.19 hectares. The largest change during this period was from barren land to agricultural land, encompassing 1671 hectares. From 2003 to 2013, the smallest change was from water land to urban land, with an area of 1.93 hectares, while the largest changes involved barren land transforming into both barren and agricultural lands, totaling 1851 hectares, and agricultural land converting to urban land, covering 1376 hectares. In the period from 2013 to 2023, the smallest changes were from urban land to garden land, and the largest change was from barren land to agricultural land, spanning 19.51 hectares. Additionally, 2959.96 hectares of agricultural land were converted to urban land.

According to the land use changes from 1993 to 2023, the smallest change was from water land to urban land, with an area of 20.65 hectares. The largest changes over the 30 years were from agricultural land to urban land, covering 2518 hectares, barren land to agricultural land, spanning 2183 hectares, and garden land transforming, totaling 2887 hectares. This indicates that the most significant land use changes over the three decades pertain to urban and man-made lands, with their area approximately tripling.

Water Bodies: The water bodies in the specified area encompass the Silvana River, also known as the Berdeh Soo River, which flows through the city of Urmia and empties into Lake Urmia. In 1993, this stretch of land had an approximate area of 227.26 hectares, which decreased to 89.21 hectares by 2003. According to analytical tables, this reduction in area can be attributed to various reasons. Firstly, the widening of the riverbed due to urban development and beautification projects by the Urmia Municipality. Additionally, the construction of the Silvana Dam both upstream and downstream has led to a significant portion of the riverbed being utilized for agriculture and gardens due to reduced river volume. Furthermore, in the past decade, due to drought and limited rainfall across the country, water from the dam has only been released for agricultural purposes on specific days and months of the year, resulting in periods of water scarcity in the riverbed. Therefore, in 2013 and 2023, the classification of water classes in the images was omitted due to these reasons.

Urban Areas (Human-made): In 1993, the total identified urban area in Urmia was 4630.63 hectares, which significantly increased over three decades to 7697.96 hectares in 2023. Over 3000 hectares have been added during this period, representing a 65% increase compared to 1993. All land conversions to urban areas were calculated in season 4. The major reasons for the tripling of Urmia's urban area over these three decades include population growth and subsequent construction activities, particularly unauthorized constructions, mass housing developments in the northern, northwestern, southern, and southwestern parts of the city, along with infrastructural developments, highways, factories, industrial workshops outside the city limits. Residential and commercial township constructions outside the city limits and obtaining top positions in industrial township constructions nationwide, along with achieving the third position nationally in the executive performance of the construction movement, have contributed to numerous economic activities and constructions across the city. Gardens constructed outside the city limits have, inevitably, led to changes in culture and the construction of small orchards or garden villas, subsequently necessitating the construction of access roads for transportation due to the high cost of land and the consequent influx of people to the outskirts and suburban areas for housing construction. Additionally, the construction of large administrative centers in the city's periphery has been another major factor driving urban expansion and growth in the specified area.

Agricultural Lands: Agricultural lands covering 6448.03 hectares in 1993 increased to 8652.81 hectares in 2023, representing an addition of 2204 hectares, approximately a 35% increase compared to 1993. While the percentage change in agricultural lands may not seem significant at first glance, a closer examination reveals substantial changes in land use over the three decades. It's notable that for every unit of agricultural land converted to urban areas or gardens due to various reasons, such as construction activities, an equivalent area of agricultural land has been added annually. Barren lands have been increasingly cultivated, leading to the degradation of natural and pristine lands.

Barren Lands: Barren lands are among the most critical lands from an environmental perspective, upon which human life depends. In 1993, these lands covered an area of 5210.61 hectares within the specified area, but by 2023, this area decreased to 5930.24 hectares, nearly halving compared to 1993. More than 44% of the land area in 1993 has been lost, which is approximately 462 hectares converted from barren lands to other uses by 2023. Barren lands have undergone significant changes, being converted to agricultural lands, gardens, and most significantly, urban areas.

Land Use Changes: Urmia had 10,252 hectares of gardens in 1993, primarily consisting of vineyards and apple gardens. However, by 2023, due to significant urban development and human interventions, the area of these lands has decreased to 4,487.68 hectares. More than 22% of these gardens, totaling 129 hectares, underwent transformation into other land uses. This transformation mainly occurred through the construction of various urban roads, resulting in a change of land use along these roads. Additionally, all gardens in recent years have been equipped with recreational facilities such as garden pavilions or at least one residential structure within the gardens for leisure activities and other purposes.

Through fuzzy logic analyses and factor analysis results, the driving forces behind land use changes and land cover in Urmia can be broadly categorized into six socio-demographic, economic or lifestyle strata (wealth), progress and technology, political economy, political-administrative structure, attitudes, and values (culture). The first three categories are associated with environmental changes, where environmental impact is a function of population, wealth, and technology.

Among these three categories of driving forces, the population generates the most significant changes in land use. Additionally, three other groups are added to these driving forces: political economy, which includes exchange systems, private ownership, government controls, and planning; political-administrative structure, which encompasses governmental institutions and organizations; and attitudes and values of individuals and groups. The categorized driving forces in this classification are much fewer than the previous three groups, and particularly, population growth is confirmed to have caused the land use changes in Urmia.

Ethical Approval

This study did not involve human or animal subjects. Therefore, ethical approval was not required.

Funding

This research received no specific grant from any funding agency in the public, commercial, or not-for-profit sectors.

Availability of Data and Materials

The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Conflict of Interest

The authors have no conflicts of interest to declare that are relevant to the content of this article

Talaghani, A., & Yousefi, A. (2021). "Explaining the Factors Influencing Land Use Changes in Peri-Urban Villages of Rasht." Journal of Peri-Urban Space Development. Retrieved from www.jpusd.ir.
Ziari, K. (2002). "Urban Land Use Planning (Case Study: Minab)." Journal of Geographic Research, 17(2-3).
Ziari, K., & Irji, H. (2022). "Investigating the Physical-Spatial Expansion of Shiraz Metropolis Using Remote Sensing." Geography and Planning Journal, 26(79), 171-186.
Alwedyan, S.,. (2022). "Monitoring Urban Growth and Land Use Change Detection with Gis Techniques In Irbid City, Jordan." International review for spatial planning and sustainable development D: Planning Assessment, Vol 11.
Bakker, K.,. (2022). "Water Security: Research Challenges and Opportunities." Springer, https://www.sciencedirect.com/science/article/pii/S1110982323000042#b0105.
Biello, D. (2018). "Gigalopolises: Urban Land Area May Triple by 2030." scientificamerican https://www.scientificamerican.com/article/cities-may-triple-in-size-by-2030/
Deng, L.; Zhang, Q.; Cheng, Y.; Cao, Q.; Wang, Z.; Wu, Q.; Qiao, J. (2022). Underlying the influencing factors behind the heterogeneous change of urban landscape patterns since 1990: A multiple dimension analysis. Ecol. Indic., 140, 108967.
Dutta, D., Rahman, A., Paul, S.K. et al. Estimating urban growth in peri-urban areas and its interrelationships with built-up density using earth observation datasets. Ann Reg Sci 65, 67–82 (2020). https://doi.org/10.1007/s00168-020-00974-8
Ziari, K., Safdari, A., Alavi, M., Qasemi, V. & Naorozi. H. (2014). Rehabilitation and renovation in the Islamic era urban planning style historical heritage." advances in environmental biology 8, no. 12: 708
Lu, L., Guo, H., Corbane, C., & Li, Q. (2019). Urban sprawl in provincial capital cities in China: evidence from multi-temporal urban land products using Landsat data. Science bulletin, 64 14, 955-957 .
Hersperger, A.M., Gennaio, M.-P., Verburg, P.H., Bürgi, M., 2010. Linking land change with driving forces and actors: four conceptual models. Ecol. Soc. 15 (4), art1 https://doi.org/10.5751/ES-03562-150401.
Molinero-Parejo, R., Aguilera-Benavente, F., Gómez-Delgado, M., & Shurupov, N. (2023). Combining a land parcel cellular automata (LP-CA) model with participatory approaches in the simulation of disruptive future scenarios of urban land use change. Computers, Environment and Urban Systems, 99. https://doi.org/10.1016/J.COMPENVURBSYS.2022.101895
Dolley, J., Marshall, F., Butcher, B. et al. Analysing trade-offs and synergies between SDGs for urban development, food security and poverty alleviation in rapidly changing peri-urban areas: a tool to support inclusive urban planning. Sustain Sci 15, 1601–1619 (2020). https://doi.org/10.1007/s11625-020-00802-0
Bórawski, P., Bełdycka-Bórawska, A., Szymańska, Elżbieta J., Jankowski, K.J., & Dunn, J. W., (2019). Price volatility of agricultural land in Poland in the context of the European Union," Land Use Policy, Elsevier, vol. 82(C), pages 486-496.
Estoque, R.C., Ooba, M., Togawa, T., Hijioka, Y., & Murayama, Y. (2021). Monitoring global land-use efficiency in the context of the UN 2030 Agenda for Sustainable Development. Habitat International, 115, 102403.
Wang L, Jia Y, Li X, Gong P. (2020). Analysing the Driving Forces and Environmental Effects of Urban Expansion by Mapping the Speed and Acceleration of Built-Up Areas in China between 1978 and 2017. Remote Sensing, 12(23):3929. https://doi.org/10.3390/rs12233929
Worldbank-Urban Development. 2023. "Urban Development." worldbank https://www.worldbank.org/en/topic/urbandevelopment/overview.
Yuan M, Song Y, Guo L. (2018). Exploring Determinants of Urban Form in China through an Empirical Study among 115 Cities. Sustainability, 10(10):3648. https://doi.org/10.3390/su10103648

No competing interests reported.

Download PDF

Editorial decision: Revision requested
01 Aug, 2024
Editor assigned by journal
26 Jul, 2024
Submission checks completed at journal
26 Jul, 2024
First submitted to journal
04 Jun, 2024

You are reading this latest preprint version

Analysis of Factors Affecting Urban Land Use Changes: A Case Study of Urmia City, Iran

Status:

Version 1

Abstract

Figures

1. Introduction

2. Literature review

3. Methodology

1) Examination and analysis of changes over the past three decades.

2) Analysis of spatial growth and urban development in both past and present contexts.

3) Analysis of factors influencing changes in land use.

3.1. Study Area

4. Results and Discussion

4.1. Image Classification

4.2. Investigating Changes

4.3. Urban Areas

4.4. Factors

5. Conclusion

Declarations

Ethical Approval

Funding

Availability of Data and Materials

Conflict of Interest

References

Additional Declarations

Status:

Version 1