نوع مقاله : علمی- پژوهشی
نویسندگان
1 دانش‌آموخته کارشناسی ارشد رشته سنجش از دور و GIS، دانشکده جغرافیا، دانشگاه تهران، تهران، ایران
2 دانشیار گروه سنجش از دور و GIS، دانشگاه تهران، تهران، ایران.
3 دانشجوی دکتری اقلیم‌شناسی، دانشکده جغرافیا، دانشگاه تهران، تهران، ایران
4 دانش‌آموخته کارشناسی ارشد رشته سنجش از دور و GIS، دانشکده جغرافیا، دانشگاه یزد، یزد، ایران
موضوعات
عنوان مقاله [English]
نویسندگان [English]
Background and objectives: Riparian forest ecosystems play a crucial role in maintaining ecological resilience and biodiversity in hot and arid regions. Additionally, these forests serve as vital safeguards against agricultural land erosion and alterations in riverbeds. Long-term monitoring is imperative to preserve the ecological capacity and biodiversity of riparian forests, which are continually impacted by climate change and land use/cover changes (LUCC). These changes have far-reaching implications for the environment, biodiversity, food security, and human health.
Methodology: This study utilized the Landsat satellite series archive for Gotvand County in Iran. The analysis commenced with an image from the MSS sensor (Landsat 1) in 1972, followed by images from the TM, ETM+, and OLI sensors placed in orbit on 7/16/1982, 4/15/1999, and 11/2/2013, respectively. Raw calibrated pixel values underwent conversion to surface reflectance through atmospheric correction. Classification input layers included spectral bands and two indices (NDVI, SAVI for vegetation cover, and MNWDI, NDWI2 for water class), derived from available spectral bands for each year. Object-based classification, employing the SVM algorithm, was implemented to extract forest areas, water bodies, agricultural lands, and other phenomena. Various values for scale, shape, and compression were applied in the object-based classification method to enhance separation. Evaluation metrics such as Overall Accuracy, Producer’s Accuracy, User’s Accuracy, and Kappa Coefficient were employed to assess classification results.
Results: The study observed the lowest difference between Red and NIR bands for the MSS sensor and the highest difference for the ETM+ sensor. Classification accuracy was lower for years when ground sample conditions were validated through satellite images compared to other years. In 2022, with improved spatial and spectral accuracy, the overall accuracy reached 98.9%, the Kappa coefficient was 0.89, and user and producer accuracies for the forest class were 97% and 99%, respectively. Agricultural land changes witnessed a staggering growth of over 4393% from 1972 to 2022. Riparian forest ecosystems, dominant in the area between 1972 and 2000 (ranging from 3670.6 to 2379.2 hectares), experienced a 35% loss. From 2000 to 2022 (covering 1569.6 hectares), an additional 34% of this plant ecosystem's area was lost.
Conclusion: The research findings highlight a 57.23% decrease in riparian forests over the past fifty years, reaching its lowest point of 1279.2 hectares in 2010. Concurrently, agricultural land area expanded by 45 times from 1972 to 2022, indicating a significant shift in land cover from forests to agriculture. The observed changes align with shifting precipitation (-1.5 mm/yr) and temperature (0.04 °C/yr) trends, impacting the studied ecosystems. This study serves as a crucial benchmark for the sustainable management of riparian forests along the Karun River in Gotvand County.
کلیدواژه‌ها [English]
)