Land use land cover (LULC) dynamics have long been recognized as a significant driver of natural resource change. As a result, understanding the spatial and temporal variation of LULC in the watershed is essential for effective natural resource management and long-term development. This study attempts to analyze the dynamics and change drivers from 1990 to 2020 and predict the situation for 2035 and 2050 in the Ajora-Woybo watershed. ArcGIS 10.3 and ERDAS 2015 were used to analyze quantitative data from Landsat imagery. For supervised image classification, a Maximum-Likelihood classification algorithm was used. To identify driver variables, focus groups and key informants' interviews were done. TerrSet 18.31 software was used to predict LULC utilizing the Multi-Layer Perceptron Neural Network and Cellular Automata-Markov Chain models incorporated in Land Change Modeler. Six LULC classes were discovered: cultivated land, built-up, shrub land, forest land, bare land, and water body. Cultivated land, built-up area, and bare land have increased at the expense of shrub land and forest land over the last three decades. Trends in water bodies show both decreasing and increasing trends. According to the predicted outcomes, cultivated land, built-up and bare land has increased, while shrub land and forest land have declined. Finally, agricultural expansion, population growth, wood extraction, resettlement, urbanization, and lack of environmental consideration were identified as the major drivers of LULC change. The study demonstrated that there have been significant changes in the watershed LULC. As a result, reversing the predicted conditions is critical to ensuring the watershed long-term viability.

中文翻译：

---

埃塞俄比亚 Omo-Gibe 盆地 Ajora-Woybo 流域土地利用土地覆盖的历史和未来动态及其驱动因素

长期以来，土地利用土地覆盖 (LULC) 动态一直被认为是自然资源变化的重要驱动力。因此，了解流域 LULC 的空间和时间变化对于有效的自然资源管理和长期发展至关重要。本研究试图分析 1990 年至 2020 年的动态和变化驱动因素，并预测 Ajora-Woybo 流域 2035 年和 2050 年的情况。ArcGIS 10.3 和 ERDAS 2015 用于分析 Landsat 影像的定量数据。对于监督图像分类，使用最大似然分类算法。为了确定驱动变量，进行了焦点小组和关键信息提供者的访谈。TerrSet 18. 31 软件用于利用土地变化建模器中包含的多层感知器神经网络和元胞自动机-马尔可夫链模型来预测 LULC。发现了 6 个 LULC 类：耕地、建成区、灌木地、林地、裸地和水体。在过去的三十年里，耕地、建成区和裸地的增加是以灌木地和林地为代价的。水体的趋势既有减少的趋势，也有增加的趋势。根据预测结果，耕地、建设用地和裸地增加，灌木用地和林地减少。最后，农业扩张、人口增长、木材开采、重新安置、城市化和缺乏环境考虑被确定为 LULC 变化的主要驱动因素。研究表明，流域 LULC 发生了显着变化。因此，扭转预测条件对于确保流域的长期生存能力至关重要。