The impact of urbanization on land use land cover change using geographic information system and remote sensing: a case of Mizan Aman City Southwest Ethiopia

Abstract

Land use land cover change due to urbanization is the prime driving forces to environmental problem and land surface temperature. The gap of the study is the lack of awareness of stakeholders regarding the protection of native forests, fruit trees, and BEBEKA coffee plantations. Deforestation for urban functions, including timber production, construction materials, and firewood, adversely affects the environment. The aim of this study was to analyze the effect of urbanization on Land Use Land Cover Change (LULCC) at Mizan Aman city, southwest Ethiopia from 1992 to 2022 using geographic information systemand remote sensing technique. The study employed systematic sampling household surveys and high-resolution remote sensing techniques to identify the impact of urbanization on land use land cover change and land surface temperature change. Sample household survey was focused on family size, education level, parcel, year of construction of the house, type of employment and monthly household income. The LULC classification were based on eight land cover class (settlement, dense forest, moderate forest, sparse forest, closed grassland, open grassland, open shrub land, annual crop land). Preprocessing, classification of the images and accuracy assessment were tested separately using the kappa coefficient. The analysis incorporates factor graph optimization for ambiguity resolution. The results indicated that cumulative accuracy were 81.52%, 82.96%, 85.41% and 84.46% and kappa coefficient 82.41%, 84.86%, 89.45% and 88.76%% for the year 1992, 2002, 2012 and 2022 respectively. This research showed that dense forest, moderate forest, sparse forest and open shrub land were significantly decreased by 68.96%, 24.60%, 31.36% and 8.28% respectively in the last 30 years. Urban settlement were increased at alarming rate due to land demand for housing, infrastructure and manufacturing. Therefore, urban planners must prioritize sustainable environmental management, integrated land use zoning, and active community involvement in order to protect against unsustainable changes in land use and land cover. For future research, incorporating methodologies such as multi-source remote sensing and high-resolution imaging will help differentiate land cover more effectively. Mizan Aman City experiences a nine-month rainy season with a hot climate, and cloud cover can affect image quality, making it challenging to map land covers clearly. Utilizing SENTINEL high-resolution data can enhance ambiguity resolution and improve spatio-temporal monitoring frameworks. Furthermore, integrating CO₂ estimation techniques could offer deeper insights into the environmental changes associated with urbanization.

Introduction

Change in land use is a longstanding human practice that modifies the availability of various natural resources such as plants, soil, and water¹. This modification directly impacts processes such as evapo-transpiration, groundwater infiltration, and rainwater runoff. Land use and land cover changes are significant when examining global dynamics in response to environmental and socioeconomic influence. Urbanization and its activities, anthropogenic factors (e.g., economic activity, construction, and demographic growth), and environmental factors are all drivers of LULC change².Factors such as urbanization, anthropogenic activities, and environmental conditions drive these changes, leading to both positive and negative impacts on agriculture, forests, the atmosphere, deforestation, and afforestation³.

Gathering accurate and timely geospatial information is the primary challenge when using remote sensing data to change land use and land cover⁴.Urbanization has conventionally been seen as a sign of regional economic vitality, but it now has negative consequences such as the loss of native trees, coffee plantations, bee wax, hot-climate weather change, loss of farming area, social fragmentation, and infrastructure costs in Mizan Aman wereda. Accurate land cover information is essential for natural resource management, planning, and monitoring programs in the region, as stated⁵. According to some studies⁶, changes in land use and land cover (LULC) are mainly influenced by population growth, economic growth, and physical factors such as land form, slope condition, soil type, and weather.

The rapid growth of cities due to urbanization has led to a significant increase in population and building density, resulting in a reduction in green areas and an increase in impervious surfaces⁷. This has caused the city center to become warmer than the surrounding areas, creating an urban heat island effect. Changes in land use and land cover have negative impacts on both global and local climate, land surface temperature, natural hazards, and socio-economic dynamics⁸. To meet the growing passion for fundamental people needs and wellbeing, it is crucial to have accurate information on land use and cover, as well as an understanding of the potential benefits and drawbacks of different land use choices. This information is essential for urban planning, sustainable urban development, efficient use of resource, and balancing hydrological cycle.

Many studies conducted on land use and land cover (LULC), such as agriculture, urbanization, and alterations in landscapes, find value in change detection in Ethiopia⁹. This highlights the gap in addressing implantation issues based on empirical evidence. Stakeholders should prioritize sustainable urban development that integrates nature into urban planning. It is crucial to comprehend patterns in landscapes, alterations, and the interplay between human actions and the environment for effective efficient use of resources without harming natural landscape¹⁰. Remote sensing data has become a valuable tool for studying changes in land use and cover due to its repetitive data collection, processing suitability, and precise geo-referencing. As a result, remote sensing data is widely utilized for identifying, measuring, and quantifying LULC patterns¹¹.

The growth of the population has been a significant factor leading to changes in land use and land cover in many developing nations, primarily Ethiopia¹². Numerous research studies¹³ have been carried out on the changes in land use and land cover in Ethiopia, with different factors driving these changes, such as demographic pressure and the resulting demand for environmental resources. This study¹⁴ found that the nature, direction, and extent of land use and land cover changes varied depending on the location. Additionally, research into land use and land cover changes is crucial for devising strategies for environmental protection and long-term environmental management. Ethiopia is experiencing substantial changes in land use and land cover, primarily due to alterations in natural resources, farming practices, and urban fabrics. Mizan Aman City is the largest city and the capital of the Bench-Sheko zone in the southwest region of Ethiopia. The city has been expanding into its periphery, leading to deforestation and a reduction in valuable forest areas. Mizan Aman is situated in the heart of a native dense forest, where fruit and coffee plantations thrive, supporting both domestic trade and export, particularly in coffee beans and tea leaves and spices. Due to rapid population growth and increasing utility demands, the timber production for household materials is also on the rise in the city. Mizan Aman city physical characteristics such as stream flow and land use land cover deteriorated over the last 20 years. The change has a direct impact on the natural forest, agriculture, and weather change. To monitor and evaluate land-use dynamics, well-timed management of LULC alteration dynamics and quantifying landform and built-up patterns for the natural forest and environment is required¹⁵. As a result, the purpose of this research was to investigate the impact of urbanization on Land Use/Land Cover Change (LULCC) in Mizan Aman city, Bench-Sheko zone, southwest Ethiopia from 1992 to 2022 using a geographic information system and remote sensing technique. It also aims to address the lack of awareness among stakeholders regarding the impact of deforestation of native forests for urban activities and construction. These actions contribute to climate change, increase the urban heat island effect, and elevate water runoff due to impervious surfaces. Therefore, this study aims to quantify the effects of urbanization on land use change and surface temperature change in order to inform sustainable planning and environmental policy. Furthermore, the study’s findings could be useful to urban governance, mayors, and urban planners for effective and resilient urban development and natural resource protection.

Literature review

Examining the historical context, urban growth trends, and concepts related to land use and land cover changes is essential for understanding the impact of urban growth on these dynamics. Population growth and increasing land consumption, particularly in urban areas, are driving factors in land use and land cover change patterns¹⁶.The utilization of remote sensing and geographic information system technologies for evaluating natural resources and monitoring environmental transformations is becoming increasingly prevalent¹⁷. By analyzing time series data captured through remote sensing and integrating it with socio-economic or biophysical information using geographic information system, the dynamics of land use and land cover changes can be effectively studied¹⁸. The advancements in GIS and remote sensing technologies have facilitated research endeavors such as detecting changes in land cover and predicting future scenarios by leveraging technological progress, reduced data costs, and the availability of historical spatiotemporal data and high-resolution satellite imagery. Many studies have been conducted to track patterns of land use and land cover changes through time and forecast the future expansion of urban regions¹⁹.

Remote sensing and GIS

Geographic information science (GIS) is made up of two major components: remote sensing and geographic information systems (GIS)²⁰, which also includes global positioning system technology, geodesy, and traditional cartography²¹.This sentence was made the last two decades²².The merging of remote sensing (RS) and geographic information system (GIS) technology has been acknowledged as a significant advancement in recent years²³. Additionally, global positioning system (GPS) technology, geodesy, and traditional cartography have also been highlighted as crucial elements in the field. Scholars have emphasized the importance of integrating remotely sensed data (RS) and GIS technology as a valuable approach for analyzing land use and land cover changes²⁴.The combination of remote sensing and GIS is widely regarded as effective in obtaining precise and timely information regarding the spatial distribution of LULC changes.

Urbanization and change of land use

Urbanization is widely and increasingly accepted as an unavoidable phenomenon. Economic activity and population growth have caused time to change²⁵. Rapid urbanization is primarily caused by pull and push factors from rural to urban, urban growth to peri-urban areas, and natural population increase in cities. Land use and land cover can be affected by urbanization. Land use is defined as provisions, activities, and involvements humans to produce change, or preserve a particular land cover type²⁶. This definition emphasizes the connection between land use and land cover change. The perceived biophysical features on the land surface is referred to as land cover. Land cover in non-urban areas is typically defined by the dominant vegetation type, such as forest, grassland, or cropland. Changing how land is used (for example, by building towns and cities on it) alters the land cover and has numerous direct and indirect consequences²⁷. The direct effect is a diminishing in agricultural land, an increase in the area of hard surfaces impervious area, which reduces water infiltration rate into ground and raises storm water or flooding, and extraction of building materials, which reduces natural forest and environmental degradation by mining cement stone and road construction material.

Implications of population increase in urban centers

The effects of urban agglomeration on resource productivity, including infrastructure and service, are both positive and negative. Furthermore, cities could provide housing more effectively through cost-effective, high density development and can be a sustainable goal for rural residents in terms of employment creation, market services and healthy living style²⁸. However, this can only occur if urbanization is well managed, and it has to be noted that every increase in population puts additional strain on the provision of basic urban services. It is also associated with increased land use change and environmental issues, particularly urban heat islands, deforestation, and agricultural. This is a need that is not easily met financially or technically in the country’s urban centers.

The land sat program

The Land sat Program was founded in 1969 by the United States Geological Survey(USGS) and National Aeronautics and Space Administration(NASA).NASA designs remote sensing tools and spacecraft, tests them, and puts them into orbit²⁹.USGS manages the satellites, ground-data reception, archiving, product creation, and dissemination. The program produces a comprehensive record of global environmental changes over time. It also tracks land use and its change effectively due to weather and urbanization.

Change detection technique

The technique describes the process of change detection using remote sensing (RS) technology. It explains that change detection involves observing an object or phenomenon at different times to identify differences in its state³⁰. Change detection highlights the importance of Earth surface features in understanding the relationships and interactions between human and natural features, which can lead to informed decision-making. The accessibility of historic remote sensing data, lower data costs, and higher resolution from satellite platforms have contributed to the increased impact of RS technology on monitoring land use and land cover change at multiple three-dimensional level. To effectively use RS for land use and land cover (LULC) change detection, an adequate perception of the study area, the satellite imaging system, and the several information extraction methods is crucial³¹.It further notes that LULC change detection has become a main application of remote sensing technology in recent years due to its monotonous coverage at tight intervals and reliable image quality. Therefore, remotely sensed data is recognized as a realistic source of LULC information that can be effectively extracted to assess and monitor these changes.

Accuracy assessment

Evaluating the accuracy or validating the results is critical and has become a standard component of any remotely sensed land use land cover map. This is essential for perceiving the improved outcomes and employing these results to inform decision makers. The error matrix is one of the most commonly adopted methods for assessing the accuracy of image classification. It can be used to generate a variety of descriptive and inferential statistics³². Typically, the error matrix compares the relationship between the reference field data (ground truth) and the classification results. To well construct an error matrix, several factors should be considered. Ground truth data collection, classification scheme, sampling techniques, spatial autocorrelation, sample size, sampling frame and unit. Generally, two types of data are expected to support accuracy assessment. These are ancillary data and ground based data. This data must be collected as close to the time of acquisition as possible, during the same vegetation season. Accuracy data should be collected consistently and independently of reference and ground truth data.

There are several approaches to assessing accuracy, including overall accuracy, errors of omission, errors of commission, user accuracy, producer accuracy, Kappa statistics, and fuzzy accuracy. Kappa statistics have become a regular measure of calculating classification accuracy over the last 30 years. The Kappa statistic is a distinct multivariate technique that measures the agreement between two images to assess the accuracy and precision of change detection and classification³³. The Kappa statistics used to calculate the variation among the actual agreement among the reference data and an automated classifier, and the chance agreement between the reference data and a random classifier.

$$K=\frac{{{\text{observed}}\;{\text{accuracy}} - {\text{random}}\;{\text{agreement}}}}{{1 - {\text{chance}}\;{\text{agreement}}}}$$

(1)

Equation (1) statistic indicates the degree to which the accurate percentage values of a confusion matrix can be attributed to “genuine” agreement versus “random” agreement. As the observed genuine agreement approaches 1 and the random agreement approaches 0, the kappa coefficient approaches 1. This represents the ideal situation. In reality, the kappa coefficient typically ranges between 0 and 1.

Drivers of land cover change

Several interrelating factors originating from diverse level of institutions of the connected human-environment and activities always cause land use change. These driving forces change dramatically over time and space, depending on the specific human environment situations. Land use change driver has been broadly recognized and studied at the global, regional, and local levels³⁴. All biophysical and human-induced drivers that have been identified can be classified as instantaneous and ultimate causes, as well as systematic and random drivers.

Methods and materials

Study area

Mizan Aman is a city in Ethiopia’s southwest people region’s Bench-Sheko zone as shown in Fig. 1. It is the biggest city and administrative center in the southwest people region, about 185 km from Jimma and 561 km from Ethiopia’s capital, Addis Ababa. Mizan Aman has an elevation of 1407 m and is located at 6°59′00′′ N and 35°35′ 00′′E UTM zone 36³⁵. The city of Mizan Aman gets its name from Mizan Teferi, which, along with the nearby town of Aman, creates a discrete sub-city known as Mizan and Aman. This was surrounded by Bench Wereda on the south side. According to the CSA’s 2013 Census, the projected population of Mizan Aman in July 2019 is 112,807 people, with 56,757 males and 56,050 females. Mizan-Aman is one of the reform towns in the southwest region, with a third-grade city administration, five kebeles, ten sub-kebeles, and 45 localities.

Mizan Teferi (Mizan Aman after 2003) was born in 1944 as the current capital of the Bench-Sheko zone. Captain Alemayehu Filate, who coined the name and took the first significant step in the establishment of the city, was the driving force behind this event. Following his victory over the Bench fighters, Captain Alemayehu took the initiative to establish a garrison in the area to keep the people under control.In this regard, he chose modern-day Mizan- Teferi as his focal point. According to elders, Fitawrari Alemayehu Filate named the town “Mizan -Teferi” to mean “Teferi is fair” in his governance. It was named on the occasion of Emperor Hailesilassie’s birthday on July 24, 1944, solely to honor and commemorate him. Mizan Aman was granted municipal status in 1952.

Mizan Aman city amenities such as digital telephone access, postal service, bank, and hospital were established in the last three decades, according to the southwest region bureau of finance and economic development. Bebeka coffee plantation is under construction near the town. It is also home to two higher education institutions, Aman college of health sciences and Mizan - Tepi University³⁶.

Fig. 1

Location map of the study (prepared by using ArcMap 10.8and ERDAS IMAGINE software’s).

Method of data collection

This study used a sample household survey, key stakeholder interviews, site observation, a review of historical urban plans, and urbanization trends, land use land cover change to collect both qualitative and quantitative data. Mapping and describing urbanization trends and changes in land use and land cover. The following major components comprised the research methodology: (i) data purchase and gathering; (ii) data preprocessing; (iii) classification scheme; (iv) accuracy assessment; and (v) land use land cover change investigation. Change detection was gathered utilizing satellite images in this LULC study. The Arc GIS 10.8³⁷ and ERDAS IMAGINE software’s³⁸ were used to edit map preparation and categorization. Landsat satellite images of the Mizan Aman city were collected in 1992, 2002, 2012, and 2022. The study was included by socio-economic questionnaires of the sample households. According to³⁹, graph optimization significantly enhance the analysis of spatiotemporal data by improving ambiguity resolution. These factor graph optimization offers a structured framework for modeling complex relationships between variables and enabling the integration of diverse data sources. This approach facilitates the identification and correction of data inconsistencies, resulting in more accurate estimates of spatial and temporal dynamics.

Sampling frame, sample size and sampling procedure

The sampling frame is a collection of elements from which the sample is selected. Thus, Mizan Aman city has five kebele administrations: Shesheka, Kometa, Ediget, Hibret, and Addis ketema. How the urban environment has changed over the study period (1992–2022), including horizontal expansion, land conversion to urban land use, and other activities. Systematic sampling techniques was used to collect socio-economic data. There were 112,807 population from the five strata (kebele). The average family size is 8.57, resulting in 13,065HHs. As a result, the sampling frame was 13,065. From each strata five kebeles there were 112,807 population. Average family size was 8.57 and hence, 13,065HHs. Therefore, sampling frame was 13,065.

Using household survey formula sample size (n) was calculated by using Eq. (2).

$${\text{no}}=\frac{{{{\left( {{\text{Za}}/2} \right)}^2}*{\text{P}}\left( {1 - {\text{P}}} \right)}}{{{{\text{d}}^2}}}$$

(2)

no = initial sample size ( un adjusted sample size).

Z = the desired level of confidence 95% (Zα/2 = 1.96).

d = Margin of error, let say 5% (0.05).

P = the proportion of population which is identified from the previous study is 0.5 or 50%.

${\text{no}}=\frac{{{{\left( {{\text{Za}}/2} \right)}^2}*{\text{P}}\left( {1 - {\text{P}}} \right)}}{{{{\text{d}}^2}}}={(1.96)^2}*0.5(1 - 0.5)=384$

Meanwhile, the population correlation between sample and population (no/N) is greater than 5% (0.05), the researchers have to be adjusted using the formula: n = no/(1 + no/N) where N is sampling frame. no/N = 384/13,065 = 0.294. Thus 0.294 > 0.05. Thus, n = no/(1 + no/N) = 384/(1 + 384/13,065) = 373.04 = 373. As shown in Fig. 2, the sample size were373 in proportional stratified sampling from each stratum (2.85%) from the total population. The sampling frame was13,065 households is (N), and (n) chosen sample size of 373 was expected by using systematic sampling techniques to gather household socioeconomic data. The researchers have found the fraction of these two numbers, N/n = 13,065/373 = 35 to obtain the sampling interval. Then, the samples interval were 35. Hence, the beginning sample was between 1 and 35. The data were collected by starting at 10th household and continued 10, 45, 80, 115, 150, 185, 220,……13,065 to the end.

Fig. 2

Sampling procedures.

Research design

Data source

Both primary and secondary data were utilized for land use and land cover change in the study area. The time series satellite images, the ground control points and Google satellite image data were employed as main input data for this study in general as given in Fig. 3. The primary data source and materials listed in Table 1.

Fig. 3

Research design.

Table 1 Datasets and software used for LULC preparation.

Data analysis method

In remote sensing studies, having a representative ground truth dataset is essential for linking reflectance properties to objects, training classifiers, and enabling accurate automatic classification. The study⁴⁰ involved analyzing trend changes in different land cover types by classifying the entire image using a supervised classification method. Land sat images from specific years were downloaded and purchased from Ethiopian geospatial information institute for the study area, and random control points were used along with ERDAS IMAGINE and GIS 10.8 software for further land use and land cover classification. Image data files were obtained in zipped format from the US Geological Survey website and converted to Tiff format for analyzing land surface temperature. During the pre-processing procedure, the downloaded images were changed to image format using ERDAS IMAGINE software. Following the collection of demographic, socio-economic, environmental, and infrastructure data via questionnaires in Mizan Aman. The data were analyzed and interpreted using SPSS in the form of descriptive manner (tables, graph, and percentages)⁴¹. Emphasizes spatio-temporal monitoring frameworks offer a novel approach to understanding and analyzing dynamic phenomena across both spatial and temporal dimensions. These frameworks integrate advanced analytical techniques with state-of-the-art data collection methods, such as remote sensing and geographic information systems (GIS), to track changes in various environmental, social, and economic factors over time. The capacity to visualize and interpret complex interactions within a spatiotemporal context is crucial for informed decision-making in areas like urban planning, environmental management, and disaster response.

Land use and land cover classification

During the initial processing phase, standard steps typically involve identifying and repairing damaged lines, aligning or registering images, adjusting radiometric values, correcting for atmospheric effects, and accounting for topographic variations. Established land cover classification standards are globally recognized through the Food and Agricultural Organization (FAO) and the European Environmental Agency (EEA), with specific regional criteria like the CORINE classification system for Europe. A recent land cover change analysis conducted as part of the FAO/Africover initiative yielded comparable categories of land cover modifications. The Land sat image was distinguished by the contents of the land cover class, generalized patterns, representative (typical of class photographs), and class particularities. Eight land cover classes participated in this study: settlement, dense forest, moderate forest, sparse forest, closed grassland, open grassland, open shrub land, and annual crop land. Factor graph optimization is introduced to tackle classification ambiguities in land cover classification tasks⁴². This method effectively integrates multiple sources of information, enabling a more robust inference process by modeling the relationships between various features and their corresponding classes as a factor graph.

Supervised image classification

In this phase, we chose pixels that represented recognizable patterns or could be recognized with external sources like Google Earth. Prior to choosing samples, it was crucial to have a thorough understanding of the data, class specifications, desire of the class, and algorithms to be employed. The researchers oversaw the classification of pixels by assigning class values based on prioritizing these classes.

Normalized difference vegetation index (NDVI)

The NDVI has been a prevalent tool in remote sensing applications for an extended period, particularly for vegetation analysis. By examining the visible and near-infrared bands of the electromagnetic spectrum, this technique can differentiate between various types of vegetation and other elements⁴³. It utilizes standardized measures to assess the presence of live green vegetation in satellite imagery over different timeframes, offering valuable insights into the health and distribution of plant life. NDVI is a standardized method for comparing the greenness of vegetation across seasons and years in satellite images. It is a different way of measuring the quantity and situation of vegetation.

It presents an alternative method for determining vegetation quantity potentially linked to canopy features like biomass, leaf area index, and vegetation cover percentage⁴⁴. NDVI stands out as a commonly utilized index, with its effectiveness in satellite analysis and vegetation monitoring established over the past twenty years⁴⁵. The NDVI values of pixels range from − 1 to + 1, with higher values close to + 1 indicating full-bodied, thriving vegetation^46,47, while values near 0 and − 1 signify water bodies or non-vegetated areas (bare lands). In this study, NDVI is computed to distinguish vegetated regions from non-vegetated ones in the specified area by applying a Eq. (3) that involves the red and near-infrared bands.

$$\:NDVI=\frac{NIR-R}{NIR+R}$$

(3)

Where, NDVI = Normalized Difference Vegetation Index, NIR = Near Infra-Red and R = Red.

Accuracy assessment

Field observations provided the main reference (ground control point) data in this study, with Google Earth points created for accuracy evaluation). A total of 300 reference points from the field survey were utilized to assess accuracy. The classification accuracy was verified by overlaying classified maps and test samples. The assessment of image classification accuracy included calculating the Kappa coefficient ‘k’, which measures the agreement between image data and ground truth. The coefficient ranges from 0 to 1, with 1 indicating perfect agreement, often expressed as a percentage for classification accuracy evaluation given by Eq. (4).

$$\:\text{K}=\frac{\text{N}{\sum\:}_{\text{i}=1}^{\text{r}}\text{x}\text{i}\text{i}-{\sum\:}_{\text{i}=1}^{\text{r}}(\text{x}\text{i}+\text{*}\text{x}+\text{i})}{{\text{N}}^{2}-{\sum\:}_{\text{i}=1}^{\text{r}}(\text{x}\text{i}+\text{*}\text{x}+\text{i})}$$

(4)

Where, N is totals samples, xii is the number in row i and column i, K is kappa statistics, x + i is column total, xi + is row total.

A Kappa coefficient greater than 0.85 indicates excellent agreement, while a value between 0.70 and 0.85 signifies very good agreement. A Kappa coefficient between 0.55 and 0.70 signifies good agreement, while one between 0.40 and 0.55 indicates fair agreement. A Kappa coefficient less than 0.40 signifies poor agreement⁴⁸. The accuracy assessment is a crucial end step in the classification process, as it measures how many ground truth pixels were correctly classified. It’s important to remember that no map is a perfect representation of reality and that all maps contain errors. However, accuracy is generally higher when viewing geographic patterns over larger areas. The overall accuracy as per Eq. (5) of the map is determined through an accuracy assessment, which calculates the mean accuracy for each class in the map .

$$\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:\:OA\left(\%\right)=\frac{SCC}{TS}*100$$

(5)

Where OA stands for overall accuracy, SCC stands for the sum of correctly classified items, and TS stands for total sample. In order to conduct an accuracy assessment, it is necessary to have a set of reference pixels that represent specific geographic locations on the classified image. The error matrix is commonly used to assess accuracy by comparing the classification of each pixel to the actual land cover conditions determined from ground truth data. Producer accuracy assesses how accurately different land cover types are classified based on omission errors, while user accuracy measures the probability of a classified pixel corresponding the land cover type of its real-world location⁴⁹. Moreover, the assessment of accuracy evaluation involves the use of an interpreted map or classified image obtained from remote sensing data, in conjunction with a reference map with high-resolution or ground truth data. The error matrix is used to compare the information from the classified image and the reference map information source, while the Kappa coefficient provides a comprehensive measure of agreement among the different elements of the error matrix.

Land surface temperature

The MODIS Land Surface Temperature/Emissivity (LST/E) information, which has a spatial resolution of 1 m and a data type of 16-bit unsigned integer, was adjusted by multiplying it with a scale factor of 0.02 to convert the digital numbers (DN) to temperature data in degrees Celsius (°C)⁵⁰. The following Eq. (6) was used to convert the digital numbers of land surface temperature to degrees Celsius.

$$Temerature\;(^\circ C)=(DN * 0.02) - 273.15$$

(6)

To calculate the mean monthly LST for each year during the study period (1992–2022), four 8-day composite images of each year were summed and then divided by 4 using a raster calculator. Subsequently, the mean LST of the dry season for each year was computed by adding the mean monthly LST for each month of the year and dividing the total by 7, where 7 denotes the number of months in the dry season from October to April. Furthermore, integrating CO₂ estimation techniques⁵¹ could provide further insights into environmental changes associated with urbanization.

Results

Population and administrative size of the City

Mizan Aman city has got the reform status in 1995 E.C. Under recent structuring the city is separated in to two sub cities and five kebeles. The name of the sub-cities are said to be Mizan and Aman. The city covers an area of 9,164 hectares of land through an average elevation of 1250 m above mean sea level. According to⁵², the most recent population and housing census result (projected population as of July 1, 2019), the city’s total population is 112,807, with males represent for 56,757 and females represent for 56,050. The estimated rate of population growth is 4.8%. In terms of age structure, 37% of the city’s entire population is between the ages of 0 and 14. The majority of the population (62% of the total) was between the ages of 15 and 64, with those 65 and older constituting only 1% of the total. The dependency ratio is calculated using a population size of less than 15 people who were over the age of 64. The 15–64 age group is the most economically active. The number of elderly, working and children age discussed in Table 2. The dependency ratio is a method of calculating the economic burden that the population must bear. The computation was based on many years of experience.

Table 2 Age group of the population and dependency ratio.

Demographic characteristics of sample households

From the total (373) respondents, 342 (91.7%) were males household heads while, 31 (8.3%) were females household head. Respondents’ age were categorized into four age categories. 18 to 29 years old young age, 30–49 years old adult age (19 years difference), 50–60 years old adult age (10 years difference) and the last one greater than 60 years old older people. The percentage of respondents were organized from maximum number of participants to lowermost age groups. These were age between 30 and 49 (72.7%), 50–60 (21.7%), > 60(4.6%) and 18–29 (1.1%), respectively.

Socio-Economic characteristics of sample households

The total number of sample household have been represented by gender given in Table 3 with their percentage and frequency.

Table 3 Sample household heads by gender.

Among 373 sample household heads 342 are males and 31 are females. This indicated that almost males are household heads.

Table 4 Sample households by Kebele name.

Mizan Aman city has a total of five kebeles. Each sample size was taken proportional from each kebeles. These were 81, 90, 61, 62, 79 from Shesheka, Kometa, Ediget, Hibret and Addisketam respectively shown in Table 4.

Causes and impact of urbanization on land use change

According to CSA data, Mizan Aman’s projected population size in 2020 is 112, 807 people. When projected until 2030 (for a ten-year planning period), the population would be 182,747 at the end of the planning year. This indicates that by removing forests, there will be a rapid population growth and demand of land for housing, service, and infrastructure construction at the periphery shown in Fig. 4. Mizan Aman has compact residence of minor farms that are helped by very profitable fruits, with coffee being the central money produce. Soils are productive, and the sub-region has a long and plentiful rainy time of year, at list 8 months in a year. The overall population density was 1230 people per square kilometer, or 12.3 people per hectare. The city was bounded to the north by Garkin, to the northeast and east by Kosokol, to the south by Shonga, Zemika, and to the southwest by Mashambi.

Migration, natural growth, industrial economic activities (coffee, honey, and mineral trade) are among the major factors driving urbanization in Mizan Aman. The city’s built-up area is rapidly expanding, particularly between 1992 and 2022. This rapid urbanization has resulted in a essential effect on the surroundings and native forests, in addition an imbalance in the region’s ecosystem in the Bench-Sheko zone.

Fig. 4

Mizan-Aman city administration landscape.(Prepared by using ArcMap 10.8³⁷).

Drivers of land use land cover change for Mizan Aman City

Land use change

The structural plan of the city was prepared for the next 10 years (i.e. 2012–2022 E.C).During the planning period; the plan is expected to address all current land demand backlog and land needed for all urban development activities in the planning period time frame. To fulfill all the demand of land required for all development activities new expansion area of land must be known. Current population of Mizan Aman City at the beginning of planning period 2020 was 112,807.

Average plot size for single housing unit is about 10 m by 20 m = 200m². 110m² is insignificant to take weight that is low cost housing cooperatives. For replacement of residence on expansion area 500m2 is taken. Here 9294 buildings are identified at expansion are and from which 121 buildings are not residential but 9197 are residential. Taking the trend 1/3(9197) = 3058 of the buildings are the owners of the farm land and expected to be provided by 500m2 plot of land. Hence out of 21,186 housing unit 18,128 are weighted by 200m² and 3,058 are weighted by 500m². So land required for residential use in next planning period (18128 by 200m²) plus (3058 by 500m²) = 3,625,600m² plus 1,529,000m² = 515.5hectars. Total land for residential expansion use in the future require 695.9 ha shown in Fig. 5. To classify the share of residential land use a simple arithmetic was applied, if previous 70% of total built up is lowered to 40% total built up in new planning approach, the new residential land share 23–29% of the previous residential share 40–50%. Thus, Mizan Aman city has been expanding from 1992 to 2022 in higher consumption of land. The current total land area is 9,164 ha.

The city serves as the primary commercial hub for farming and manufacturing products. Fruits, spices, incense, coffee and timber production, in particular, are transported to the country Ethiopia main market places⁵³. It also has a number of educational institutions; however, there are no private colleges in the city. The city appears to be growing at a rapid pace, with new constructions and business activities⁵⁴ sprouting up all along the city’s highways. According to the data, rural-urban migration is more likely than natural population growth. It has a stronger and more immediate political, social, and economic impact than any other demographic dynamic. According to this, the city’s the settlement or urbanization has been rapidly growing outward.

The elevation from contour⁵⁵ and slope analysis were mapped as given in Fig. 6. The maximum elevation was 862 m and the minimum was 739 m. The northern part of the city is sloppy area and difficult to provide infrastructure and unsuitable for settlement.

Fig. 5

Elevation and slope analysis of Mizan-Aman city.(Prepared by using ArcMap 10.8³⁷).

Fig. 6

Proposed land use of Mizan-Aman city.(Prepared by using ArcMap 10.8³⁷).

Land cover change of Mizan Aman City from 1992 to 2022

Then land use land cover change of 1992 to 2022 in terms of area coverage of each land cover were discussed in Tables 5, 6 and 7, and Table 8. The red value is one particularly at settlement land cover.

Table 5 Land use land cover of 1992.

Table 6 Land use land cover of 2002.

Table 7 Land use land cover of 2012.

Table 8 Land use land cover of 2022.

Fig. 7

Mizan Aman city built-up growth from 1992 to 2022.

Built up growth changed rapidly from 1992 to 2022 given in Fig. 7. This change result land use land cover change and land surface temperature increment.

Change detection

The classification technique⁵⁶ was applied using Land sat data obtained from the geospatial information institute. The amount of change over the enquiry periods was determined using supervised classification based on the maximum likelihood classifier and the addition of knowledge from field experience.

Table 9 Land use land cover change in hectares from 1992 to 2022.

LULC maps were created for 1992–2002, 2002–2012, and 2012–2022 (Figs. 1 and 2, and 3). Table 9 compares the area of every land cover class for the entire area, using four classification period from 1992, 2002, 2012, and 2022.The result showed that sparse forest was the foremost land cover class of 2250.96Ha,2494.53Ha, 2360.09Ha and 1997.07Ha next to sparse forest, annual crop land was the second largest land cover2460.95Ha, 2363.20Ha, 2303.56Ha, and 2256.95Ha in 1992, 2002, 2012 and 2022, respectively. Moderate forest was also dominant in the year 1992 and 2002. It was covered 1937.56Ha and 1112.71Ha respectively. The land cover gradually decreased and changed to urban settlement and crop land. Urban settlement highly increased from 578.31Ha in 1992 to 2541.50Ha in 2022. During this timeframe, agriculture emerged as the primary revenue generator, attributed to the rise in population and the increasing demand for housing, social service, infrastructure and manufacturing construction. Generally, the result indicates that dense forest, sparse forest, and moderate forest and shrub land had decreased while urban settlement, closed grassland and open grassland were enlarged. This trend can be attributed to the growth in population and land demand for urban expansion and industrial development. Analysis of land use and land cover changes in the area showed that agricultural land and residential areas expanded between 1992 and 2022 over the past three decades(Tables 6, 7, 8 and 9).

Land use and land cover from 1992 to 2022

The spatial analysis was conducted to delineate the patterns of land cover changes and overall transformations⁵⁷ in land use over time as illustrated in Figs. 8 and 9. The study region was classified into categories such as dense forest, moderate forest, sparse forest, open shrub land, open grassland, closed grassland, crop land, and urban settlement area to analyze land use and land cover changes⁵⁸.

Fig. 8

Land use land cover map of 1992 and 2002. (Prepared by using ArcMap 10.8³⁷).

Fig. 9

land use land cover map from 2012 and 2022. (Prepared by using ArcMap 10.8³⁷)

Table 10 Kappa coefficient agreement.

Table 11 Summary of accuracy assessment from 1992–2022 (%).

Keppa coefficient were categorized into five groups given in Table 10. Each categories were excellent, very good, good, fair and poor with the value of the coefficient greater than 0.85, 0.70 < kc < 0.85, 0.55 < kc < 0.70, 0.40 < kc < 0.55 and less than 0.40 respectively.

The overall accuracy and Kappa statistics in land use and land cover analysis refer to the assessment of the accuracy of classification models given in Table 11. The overall accuracy is determined by comparing the total number of correctly classified pixels to the total number of pixels in the confusion matrix. A minimum accuracy of 85% is generally considered reliable for land cover classification, but users’ acceptance of accuracy levels⁵⁹ may vary depending on the specific task. The Kappa statistic is used to measure the agreement between predictive models and field sample points, as well as to assess the agreement between two sets of classifications of a dataset.

The impact of LULC change on climate change

Actually, climate change is a universal concern that is exacerbated in part by LULC change. Therefore, we have included the impact of forest, farmland, and cropland land use and land cover change on the city and expansion area by extraction of building material and deforestation particularly and climate change in general in a concise manner.

Land use and land cover changes (LULCCs) have a significant impact on climate change at universal, regional, and local levels. They contribute to the accumulation of greenhouse gases in the atmosphere, lead to increased temperatures⁶⁰, create urban heat islands, and consequently drive global warming. Additionally, the degradation of natural ecosystems in the area leads to higher carbon dioxide levels and disrupts the native ecosystems of forests, coffee plantations, and their surroundings. In this context, deforestation is the main factor driving changes in land use, followed by urban expansion, which triggers the release of soil carbon due to disturbances in native forests from urban development.

The relief of Mizan Aman city is diverse and mostly characterized by ups and downs. The existing settlement area is established on sloppy area. The city has sloppy gorges, hills rugged mountains for instance Barges and Borsa areas are dominated by gorge. The relief of the south and southwestern part of the city are dominated by plain topography and forest. Thus, these part of the city is the main future expansion direction of the city. The central part (the gap between Mizan and Aman) and the eastern part of the city are also suitable for future expansion. This change will increase the impact of urban heat island and environmental problem (deforestation) and double increase in surface temperature change and drought in the region⁶¹.

The correlation of urbanization and LST from 1992 to 2022

Traditionally, the climate of Mizan Aman city is classified in to Kolla type. Because of its location southwestern highlands of Ethiopia. Mizan Aman city and surrounding get plenty of rainfall throughout the year and number of rivers cross the city. So Mizan Aman city and surrounding rural areas are favorable for settlement. The prevailing wind over Mizan Aman city blows from southwest to northeast. The land surface temperature (LST) increased through the year 2002, 2012 and 2022 from 25.95, 27.41 and 28.39 °C respectively illustrated in Fig. 10. This has indicated that urbanization has a negative impact to increase surface temperature given in Figs. 11 and 12, and 13.

Fig. 10

The correlation of urbanization and land surface temperature(LST).

Fig. 11

Land surface temperature of Mizan-Aman city in 2002 by degree Celsius (Prepared by Geographic Information Systems (GIS) ArcMap 10.8 version⁶²).

Fig. 12

Land surface temperature of Mizan-Aman city in 2012 by degree Celsius (Prepared by Geographic Information Systems (GIS) ArcMap 10.8 version⁶²).

Fig. 13

Land surface temperature of Mizan-Aman city in 2022 by degree Celsius(Prepared by Geographic Information Systems (GIS) ArcMap 10.8 version⁶²).

Discussion

This research demonstrated how GIS and Remote Sensing can effectively capture spatio-temporal information for Mizan Aman city and its nearby areas, including the native forest that acts as the city’s lungs, as well as environmental problem and temperature change resulting from urban population increment impacts on land use and land cover. A time-series analysis approach was employed to compare remotely sensed images⁶³ from Land sat missions to investigate urban growth and the associated environmental changes over time.

The changes in land use and land cover dynamic occur from intricate interactions among various biophysical and socioeconomic factors⁶⁴. While people activities tend to have immediate and often significant impacts, natural influences typically unfold over a longer period. Between 1992 and 2022, the forest and open spaces in southwest Ethiopia experienced a troubling decline⁶⁵. During the period of urbanization, the southwestern Ethiopian highlands were still predominantly covered by 9 months annual rain with natural forests. Historically, shifting agriculture were practiced in the region for centuries without significantly threatening forest resources. However, with the rise of urbanization, population growth, and urban development, the landscape has changed dramatically. There is now a high demand of land for housing, services, infrastructure, manufacturing, and urban utilities. Additionally, a new urban farming system that was not well-suited to the local environmental situation has been introduced⁶⁶.

The findings indicated that the overall accuracy for 1992 was 83.76% (Table 11), with producer’s and user’s accuracy⁶⁷ varying between 76.41 and 90.78% and 71.34–90.54%, respectively. The Kappa statistic derived from the error matrix showed a classification coefficient of 79.02% for that year, placing it in a very good category. In general, the results (Tables 9 and 11) indicated overall accuracies of 84.76%, 85.91%, and 84.46% for the years 2002, 2012, and 2022, respectively. In this analysis, the producer’s and user’s accuracy ranged from 75.51 to 89.46% and 73.45–90.24% for 2002; from 78.94 to 93.24% and 80.43–93.24% for 2012; and from 79.15 to 91.45% and 82.15–91.34% for 2022. The Kappa coefficients for classification were recorded at 84.86%, 89.45%, and 88.76% for the years 2002, 2012, and 2022, respectively. Thus, the classifications⁶⁸ were consistently within a very good range and an excellent scale achieving in the years 2002, 2012, and 2022. Settlement, open grassland and closed grassland were significantly increased by 192.51%, 4.49% and 2.72% respectively in the last 30 years period (Figs. 7 and 8). This indicated that the major change occurred in annual crop land, dense forest, sparse forest and moderate forest were reduced gradually overall accuracy were 81.52%, 82.96%, 85.41% and 84.46% and kappa coefficient 82.41%, 84.86%, 89.45% and 88.76%% for the years 1992, 2002, 2012 and 2022 respectively.

Studies show that forest deforestation in Ethiopiais strongly associated with the continuous rise in population. A larger population typically results in a heightened demand of land for both housing and agricultural activities. Additionally, the usual restoration of forest is hindered by the significant presence of grazing and browsing livestock within these forested areas⁶⁹. The conditions have got further problem in the land use land cover and temperature change when migration rate of people increases and rapid urban expansion occur without proper urban planning and environmental protection⁷⁰. Consequently, the pressure on the native forest, fruit plantation and natural coffee farm resources⁷¹. Moreover, the demand on fuel wood and construction timber has been increasing through time. Lastly, unmanaged urban expansion and the cutting forest particularly “WARKA” for timber production. Most timber produced for urban development within the city and zonal administration and other city beyond the region.

Ethiopia is the least urbanized country in the world. Studies revealed that only 22% population live in urban area that deficit infrastructure, social service and electricity⁷². Therefore, rapid urbanization is occurring in Ethiopia without considering sustainable urban development. Many cities and towns are developing spontaneously due to poor economic development⁷³, lack of good urban governance and absence of strong rural urban linkage and autonomy of regional capital cities. According to the interview the city administration, the city of Mizan Aman has been expanding particularly after 1992 to the recent. Investor economic activates, establishment of Mizan-Tepi university, natural population increment, increasing housing demand, internal migration, evicted from other region due to conflict and drought were the main factors that expand the city. That is why the city built up has increased from 578.31hectares in 1992 to 2541.50hectares in 2022.

The major impact of uncontrolled urban growth and development without taking into account (the impact of environmental problems, social impact, economic development, and climate change) would pose a challenge for many Ethiopian cities. According to the study findings, the city has faced the challenge of deforestation of native dense forest, which serves as a lung for the people of southwest Ethiopia. The area has a hot climate, but the southwest region is covered in dense forest, which reduces hot weather conditions for nearly eight months of the year. Without managing environmental resources, urban development practices, population growth, and timber production have all had a negative impact⁷⁴. The region is well-known for its natural coffee, fruit, and honey production. This has gradually reduced its output, affecting the economy of city residents and export markets.

The main land cover in this study were dense forest, moderate forest, sparse forest, closed grassland, open grassland, open shrub land, annual crop land and settlement in the city boundary. Among forest land cover type has rapidly changed to urban settlement without considering high density development and transit oriented development⁷⁵. The city has started construction of airport at the edge of the city in 2022. The next two years at the beginning of 2025Mizan Aman city will be expected to start flight to the capital city of Addis Ababa. This would increase commerce, integrated land use and sustainable growth of southwest region. The city structure plan has mainly focused on urban function economic development, social service and manufacturing rather than environmental protection. Of course, urban development is uncontrolled phenomena due to population growth, infrastructure and housing demand. However, we can implement alternative choice for sustainable urban development such as compact city development, family planning, smart growth principle and insure the essence of good urban governance. Since 1992, the city has experienced extensive land use change due to urban expansion, agricultural expansion and intensification, deforestation, growing human populations, infrastructure and manufacturing expansion, and frequent natural temperature change⁷⁶. This study’s findings were consistent with spatial mapping land use and cover change in the Mekong basin (spruce et al., 2020). Investigating global change necessitates examining alterations in land use and land cover (LULC) and classification accuracy. Grasping the dynamics of LULC changes and the underlying factors is essential for predicting future transformations, which can inform sustainable and effective policy decisions and strategies.

Urbanization stands out as one of the greatest apparent global transformations. Nowadays, researchers have taken an interest in the LULC change processes occurring at local and regional levels, positing that analyzing spatial patterns can illuminate the connection between human activities and land use changes. The area under study experiences a tropical climate characterized by distinct wet seasons and hot weather, with average land surface temperature ranging from 25.95 to 28.39 °C.Land surface temperature has increased from the year 1992 to 2022 by 3°C.Land Surface Temperature depends on the nature of land surface. Water bodies, dense and moderate forest, and built-up areas remain cooler than open grassland and urban areas. Land Surface Temperature (LST) refers to the surface’s skin temperature. LST can be significantly reduced by vegetation, particularly tree canopy. The change in LULC has a significant impact on LST. The LST has been influenced by urbanization, which has negative environmental effects such as rising surface temperatures with increased built-up due to urban expansion. Temperatures in cities are higher than in rural areas. Urban heat islands (UHIs) are the occurrence in which atmospheric and surface temperatures in urban region is greater than in surrounding rural region. Changes in land surface temperature contribute to the expansion of urban heat islands. The primary contributor to urban heat island is impervious surface area. It needs to introduce novel satellite-based techniques for calculating CO₂ partial pressure, which hold great promise for assessing the carbon footprint of Mizan Aman City’s urban growth. Researchers can track and measure CO₂ emissions in various urban environments by utilizing high-resolution satellite imagery and advanced remote sensing methods.

This research thus offers consistent LULC information that provide the percentage of change of land use land cover that happened in Mizan Aman city in southwest Ethiopia between 1992 and 2022. The major findings identified that the expansion of urban settlement by deforesting native forest and rise of land surface temperature as a result of urbanization. This indicated that less awareness and implantationof land management policy and sustainable planning system by decision-makers. Additionally, conservation and integrated development with nature through management strategy and technology. This study was also recommended to begin the apparent socio-economic, infrastructure, environmental impact assessment, and spatial drivers of land cover change between 1992 and 2022 in Mizan Aman city, southwest Ethiopia.

The study confirms an increase in LST from 25.95 °C in 1992 to 28.39 °C in 2022, correlating with urban expansion. Utilizing multi-angle Sentinel-2 data for estimating fractional biocrust coverage (Sun et al., 2024) can enhance LST analysis.It also raised urban heat island effect. Findings align with global trends, where increased built-up areas exacerbate localized warming. The study could benefit from integrating synthetic aperture radar (SAR)-based ocean eddy detection techniques (Zi et al., 2024) for a refined atmospheric analysis.

Conclusions and policy formulation

The LULC change and expansion of Mizan Aman city were examined in this study over the last thirty years. The findings revealed that significant variations in land use land cover happened throughout the period 1992 to 2022. Sparse forest, closed grassland, agriculture, and urban settlement areas increased by 24.83%, 471.79%, 0.11%, and 107.88%, respectively, whereas dense forest, moderate forest, open shrub land, and open grassland decreased by 22.99%, 15.25%, 43.98%, and 50.24%, respectively. The effects of land cover land use change and urbanization have a significant impact on rainfall, the increase of impervious surface, water runoff, the generation glazing building and the change in surface temperature. The findings portrayed that the conversion of forest, open grassland, and shrubs to crop land and urban development created problems such as changes in land use land cover and issues with native forest, natural coffee plantation, and climate change. These changes may have an impact on native forest management and the living condition of the city residents. The scale of land use change differs significantly across time and space, reproducing alterations in urban fabric characteristics, urban development across locations, population growth, and future development plans. Land use land cover change in the last thirty years had implication on land surface temperature. Land use and land cover significantly impact climate variability, particularly regarding the carbon sequestration capabilities of different ecosystems. The study highlights that changes in land use such as deforestation, urbanization, and agricultural expansion can greatly affect an ecosystem’s ability to absorb and store carbon, ultimately leading to increased greenhouse gas emissions.

The formulation of key polices that give a high priority on protecting the area’s natural resources such as coffee plantation and dense forests. Controlling urban growth through zoning regulations that restrict the development of impervious surfaces should be one of the main policies. Encouraging green spaces integration and environmental protection policy to resilient development. Green infrastructure policy can also diminish the effects of rising land surface temperatures. It is important to institutionalize continuous monitoring and evaluation of LULC changes in order to guide future policy and guarantee adaptive management techniques to handle new issues arising from urbanization and climate change.

Recommendations and future research.

• The landscape (landform), featuring sparse and moderate forests, creates considerable shadowing, which poses a notable challenge in digital image processing. To enhance digital classification, it is recommended to use techniques that incorporate additional data, like a digital elevation model (DEM). By integrating DEM into image classification, the accuracy of digital data classification can be improved, as it provides insights into the dissemination of terrain features that influence spectral responses and helps in identifying locations for fieldwork sites.

• Utilizing a high-resolution remote sensing imagery can enhance mapping and detect changes. Further using SENTINEL along with GIS and Land sat satellite imagery help to advance quality mapping and monitor changes in urban area, as well as in Ethiopia as a whole.

• Policy and urban planning Land use changes caused by unplanned urbanization can have an impact on the environment and cause climate change at all level local, provincial, and international. The impact of urbanization on land use change and temperature change is expected to be a major issue in the future. The demand for housing, infrastructure, social services, and industry has risen over time. Urban planners can minimize urban sprawl, preserve natural landscapes, and reduce dependency on cars by emphasizing compact, mixed-use development and improving public transit systems. As a result, planners and stakeholders should improve urban policy and planning approaches such as biodiversity protection strategy, smart growth, transit-oriented development for sustainable and resilient future.

• Urban planning and smart growth strategies are essential for sustainable urban development. They must address the need for adaptive policies that promote land degradation neutrality and combat climate change resulting from urbanization. Additionally, the conservation of natural resources is crucial, especially as rural depopulation reshapes plant diversity, a phenomenon that can be contrasted with the impacts of urbanization in Ethiopia. Moreover, integrating socioeconomic indicators derived from structural urban imagery predictions can help validate trends in urban expansion.

• Limitation and further research The land cover data of 2002, particularly settlements was very condensed due to poor image quality and cloud cover. As a result, it was challenging to clearly identify the land cover types in the southwestern region of Ethiopia throughout the year. The use of a confusion matrix for accuracy assessment has become a standard practice in evaluating the quality of remote sensing products. However, this method has its limitations. It assumes that the reference data accurately represent the entire classification, which is unlikely. It also requires that the reference data and the classified image be perfectly co-registered, an impossible condition, and it presumes that there are no errors in the reference data, which is again highly unlikely. Consequently, the actual accuracy of our classification remains unknown, as it is impossible to accurately assess the true class of every pixel.

• Furthermore, the study offers a thorough analysis of land use and land cover change (LULCC) in Mizan Aman City. However, it could be greatly enhanced by integrating advanced remote sensing techniques, environmental monitoring methodologies, and policy-oriented frameworks. Future research should explore the use of multi-source remote sensing data fusion to improve the accuracy of land cover classification and predictive modeling.