Department of Remote Sensing, Space Science and Geospatial Institute, Entoto Observatory and Research Center (EORC), PO Box, 33679, Addis Ababa, Ethiopia; Geo-Information Science Program, School of Geography and Environmental Studies, Haramaya University, PO Box 138, 3220, Dire Dawa, Ethiopia.
J Environ Manage. 2024 Dec;371:123336. doi: 10.1016/j.jenvman.2024.123336. Epub 2024 Nov 15.
The Middle Awash Basin (MAB) faces severe ecological degradation due to the rapid spread of the invasive Prosopis juliflora (P. juliflora), which threatens native vegetation. The study characterizes and predicts the spatiotemporal dynamics of rangelands affected by P. juliflora in the MAB. Using three Landsat images from ETM+ (2003) and OLI (2013 and 2023), we applied a supervised random forest (RF) classification technique processed on the Google Earth Engine (GEE) platform. This classification was integrated into an intensity analysis to examine temporal transitions between land use and land cover (LULC) classes. The predictive modeling included 12 variables, including climatic, topographic, edaphic, phenological, hydrological, and anthropogenic factors, using Terrset 2020. Using multitemporal satellite remote sensing, machine learning (ML), and cellular automata markov chain (CA-MC) methods, LULC was mapped from 2003 to 2023, and future scenarios were predicted up to 2060. The P. juliflora coverage quadrupled from 2.16% in 2003 to 8.61% in 2023, while rangelands were decreased by more than 25%. Models predict that P. juliflora could occupy 22% of the land by 2060 and over 40% of rangeland areas as of 2003, expanding two to three times faster than the intensities of the LULC baseline changes, primarily targeting rangelands. Our analysis is based on a single business-as-usual scenario; however, it highlights the worrying invasion patterns. The study's limitations include the absence of multiple scenarios and climate model integration, which could offer further insights into future invasion dynamics. Nonetheless, our findings indicate that the MAB faces imminent widespread ecosystem transformation without prompt action, which will severely affect pastoral livelihoods and biodiversity conservation. Therefore, we advocate for a management strategy involving prevention, eradication, and restoration measures, underpinned by policy reforms and stakeholder cooperation.
马萨瓦洼地（MAB）面临着严重的生态退化，原因是入侵的普氏相思树（P. juliflora）迅速蔓延，这威胁到了本地植被。本研究对 MAB 中受 P. juliflora 影响的牧场的时空动态进行了特征描述和预测。我们使用了 ETM+（2003 年）和 OLI（2013 年和 2023 年）上的三张 Landsat 图像，应用了在 Google Earth Engine（GEE）平台上处理的监督随机森林（RF）分类技术。该分类被整合到强度分析中，以检查土地利用和土地覆盖（LULC）类之间的时间转换。预测模型包括 12 个变量，包括气候、地形、土壤、物候、水文和人为因素，使用 Terrset 2020 进行了预测。使用多时相卫星遥感、机器学习（ML）和元胞自动机马尔可夫链（CA-MC）方法，从 2003 年到 2023 年绘制了 LULC 地图，并预测了到 2060 年的未来情景。2003 年，普氏相思树的覆盖率从 2.16%增加到 2023 年的 8.61%，而牧场面积减少了 25%以上。模型预测，到 2060 年，普氏相思树可能会占据 22%的土地，到 2023 年，将占据 2003 年牧场面积的 40%以上，扩张速度将比 LULC 基准变化的强度快两到三倍，主要针对牧场。我们的分析基于单一的常规情景；然而，它突出了令人担忧的入侵模式。研究的局限性包括缺乏多种情景和气候模型的集成，这可能会进一步深入了解未来的入侵动态。尽管如此，我们的研究结果表明，如果不采取紧急行动，MAB 将面临迫在眉睫的广泛生态系统转型，这将严重影响畜牧业生计和生物多样性保护。因此，我们倡导采取一种管理策略，包括预防、根除和恢复措施，同时进行政策改革和利益相关者合作。
