Muthanna Journal of Engineering and Technology
Volume (12), Issue (2), Year (30 December 2024), Pages (39-49)
DOI:10.52113/3/eng/mjet/2024-12-02/39-49
Research Article By:
Basheer S. Jasim
Corresponding author E-mail: basheer.jasim@atu.edu.iq
ABSTRACT
Rapid population growth is one of the serious and common issues in Iraq, and parts of this country have been facing an increase in the number and density of the population in recent decades, putting much pressure on Iraq’s natural resources and sometimes economic activities. The growth of residential constructions and industrial zones has caused direct or indirect destruction of ecosystems and their natural lands. Landsat satellite imagery, TM sensors for 2000, and the OLI-TIRS sensor 2024 were used to detect land cover change. A supervised classification technique by Random Forest (RF) method was used for image classification, and the land cover map was obtained in two different years (2000 and 2024), with overall accuracy of 88.33% and 90.83%, respectively. The analysis results have shown that during 24 years, there has been an increase in urban areas: Urbanization increased significantly from 45.24% to 67.98%, indicating significant population and economic growth. Whereas there was a decrease in green spaces, the percentage of vegetated land decreased from 32.56% to 9.09%, which indicates the diminishment of agricultural and green spaces due to urban expansion.
Regarding the relative stability of water bodies, the percentage of water bodies decreased slightly from 3.08% to 2.53%. Finally, there was a slight increase in arid lands; the percentage of barren land increased from 19.12% to 20.40%, which may reflect land degradation and increasing desertification. Comparing land use and land cover changes over a long period shows the impact of human activities and climate change on the environment, allowing for a deeper assessment of environmental degradation and identifying the most affected areas.
Keywords: Machine Learning, Land Use Land Cover, Classification, Random Forest Algorithm.