Abstract:
This study analyzes LULC (Land use and land cover) changes in the Urmia Plain, northwestern Iran, from 2000 to 2020, and projects future patterns for 2040. Landsat, ETM, and OLI images were classified into six categories—agriculture, orchards, urban areas, saline lands, barren lands, and water bodies—using the supervised maximum likelihood algorithm, achieving Kappa coefficients above 0.85. Between 2000 and 2020, agricultural lands declined by 17.4%, while barren and saline areas expanded by over 25% combined. Urban areas increased modestly (2.3%) , reflecting localized expansion near major settlements. Water bodies, mainly associated with Lake Urmia and its wetlands, experienced a sharp decline and in some years nearly disappeared, highlighting severe hydrological stress. Using a Markov chain model derived from 2000–2010 transitions, the 2040 scenario predicts an additional 12% agricultural loss, with barren lands encroaching on former orchards. Results indicate that anthropogenic factors, including unsustainable agriculture and groundwater depletion, are the dominant drivers of degradation. The integrated remote sensing–Markov framework offers transferable tools for historical and predictive LULC assessment, supporting targeted land restoration and sustainable management in dryland environments.
Keywords: Land use change, Remote sensing, Urmia Plain, Markov chain, Landsat

Introduction
Land use and land cover (LULC) changes represent significant contributors to environmental degradation, particularly in arid and semi-arid regions where limited water resources and fragile ecosystems exacerbate vulnerability. The Urmia Plain, located within the Lake Urmia Basin in northwestern Iran, serves as a crucial agricultural and ecological zone that is currently facing significant stress due to prolonged drought, climate change, and anthropogenic pressures. The overexploitation of groundwater, coupled with inefficient irrigation practices and uncontrolled land conversion, has resulted in soil salinization, loss of vegetation, and the expansion of barren lands, thereby jeopardizing regional food security and ecological stability (Daryanto et al. 2016; Kheyruri et al. 2024b).
Numerous investigations in Iran have utilized remote sensing techniques to observe land use and land cover (LULC) trends, employing statistical or simulation models for forecasting changes. Nonetheless, only a limited number have concentrated on the Urmia Plain or integrated historical trend assessments with future predictions. Given its vital importance within the Lake Urmia ecosystem, it is essential to assess both historical alterations and anticipated future conditions to guide effective management approaches (Fatema et al. 2023; Vohra et al. 2024).
Remote sensing provides long-term, consistent datasets for monitoring LULC changes, while Markov chain modeling offers a probabilistic approach for predicting future land use patterns based on observed transitions (Mortezaii et al. 2020; Feizizadeh et al. 2022). Integrating these techniques enables both accurate historical mapping and robust forecasting.
This study aims to: (1) classify LULC for 2000, 2010, and 2020 using Landsat imagery; (2) evaluate classification accuracy; (3) analyze dominant LULC transitions over two decades; and (4) project the 2040 LULC map using a Markov chain model. The findings provide essential information for sustainable land management and conservation planning in the Urmia Basin.
Methodology
The study focuses on the Urmia Plain, West Azerbaijan Province. Land use/land cover (LULC) maps were generated for 2000, 2010, and 2020 using a supervised maximum likelihood classification. Six categories were defined: agriculture, orchards, urban areas, barren lands, saline lands, and water bodies. Accuracy assessment using confusion matrices and Kappa statistics confirmed values above 0.85, ensuring reliable spatial classification.
Post-classification statistics were used to quantify the areal extent of each class, supported by corresponding maps (Figures 2–5). Transition probability matrices were calculated to analyze historical changes between the periods, allowing the identification of spatial patterns, trends, and areas with significant land use changes. Comparative percentage changes across the three periods are summarized in Table 5.
This approach provides a comprehensive assessment of past spatial dynamics, highlighting areas of degradation and land use change, which can inform future studies and support planning for sustainable land management in this vulnerable region. No predictions for future land use are made, and the focus remains on understanding historical patterns.
Results and Discussion
The analysis of the 2005 LULC map (Figure 2) and statistical data (Table 1) reveals that mixed orchard–crop lands dominated the Urmia Plain, accounting for nearly 40% of the area, followed by orchards (18.9%) and urban areas (12.1%). In contrast, water bodies represented only 0.16% and wetlands 0.42%, showing their marginal presence even at the beginning of the study period. These results highlight that while agricultural and orchard activities were still relatively strong, natural water resources were already critically reduced, and barren lands began expanding around the lake’s margins, suggesting early signs of ecological stress.
By 2010, significant shifts were evident (Figure 3, Table 2). Surface water completely disappeared, and wetlands declined to 0.67%, marking the onset of severe hydrological degradation. Mixed orchard–crop expanded to nearly 47%, but this was accompanied by rising poor vegetation (11%) and the continued encroachment of barren lands in the north and east. The spatial patterns indicate unsustainable agricultural expansion, particularly in water-stressed areas, which contributed to land fragmentation and ecological instability. This period also reflects accelerated land conversion driven more by human intervention than by climatic variability.
The 2015 classification (Figure 4, Table 3) demonstrated intensification of these trends. Mixed orchard–crop surged to nearly 60%, while independent orchards dropped to just over 6%. Poor vegetation areas increased to 12.7%, confirming the weakening of ecological resilience. Wetlands remained below 1%, and water bodies continued to be absent, reflecting the persistent water crisis. Urban areas grew to almost 13%, driven by the expansion of Urmia city and surrounding settlements. The combination of orchard expansion into unsuitable lands and the sharp decline of natural vegetation emphasizes the anthropogenic stress dominating land-use transitions during this period.
By 2020, urban growth reached its highest share at 15.3%, while independent orchards rebounded to 17.9% (Figure 5, Table 4). Mixed orchard–crop declined to 42.3%, reflecting a shift toward more fragmented agricultural practices. Water bodies remained absent, and wetlands persisted at only 0.82%. The Markov projection for 2040 (Figure 6, Table 5) predicts a further ~12% decline in agriculture, significant orchard losses compared to 2005 and 2020, and a +26% urban increase relative to 2000. Barren and poor vegetation classes are expected to dominate the northern and eastern plain, while water and wetlands remain negligible. These findings confirm that without immediate intervention, the Urmia Plain faces continued ecological degradation and heightened desertification risk.

Conclusion
This study demonstrates that the Urmia Plain has undergone substantial LULC changes over the past two decades, with agricultural and orchard areas declining sharply and barren and saline lands expanding significantly. The Markov chain model projects further degradation by 2040, with agricultural lands potentially decreasing by an additional 12% if current practices continue.
Findings highlight the predominance of anthropogenic factors—particularly unsustainable agriculture and groundwater overexploitation—over climatic drivers in shaping LULC patterns. The integration of multi-temporal Landsat imagery and Markov chain modeling offers a reliable and transferable framework for both retrospective analysis and future prediction in semi-arid regions.
Urgent adoption of conservation-oriented land management, improved irrigation efficiency, and groundwater regulation is essential to halt degradation. The study’s results can guide policymakers and stakeholders in designing targeted strategies for restoring land productivity and safeguarding the ecological integrity of the Urmia Basin.