Geography and Environmental Studies, Arba Minch University (AMU), P.O. Box: 21, Arba Minch, Ethiopia.
Center for Development Research (ZEF), University of Bonn, Bonn, Germany.
Environ Monit Assess. 2018 Apr 25;190(5):309. doi: 10.1007/s10661-018-6688-1.
Earlier studies on land change (LC) have focused on size and magnitude, gains and losses, or land transfers between categories. Therefore, these studies have failed to simultaneously show the complete LC processes. This paper examines LCs in the Legedadie-Dire catchments in Oromia State, Ethiopia, using land-category maps with intensity analysis (IA) at three points in time. We comprehensively analyze LC to jointly encompass the rate, intensity, transition, and process. Thirty-meter US Geological Survey (USGS) Landsat imagery from 1986, 2000, and 2015 (< 10% cloud) is processed using TerrSet-LCM and ArcGIS. Six categories are identified using a maximum likelihood classification technique: settlement, cultivation, forest, water, grassland, and bare land. Then, classified maps are superimposed on the images to statistically examine changes with an IA. Considerable changes are observed among categories, except for water, between 1986-2000 and 2000-2015. Overall land change occurred quickly at first and then slowly in the second time interval. The total land area that exhibited change (1st ≈ 54% and 2nd ≈ 51%) exceeded the total area of persistence (1st ≈ 46% and 2nd ≈ 49%) across the landscape. Cultivation and human settlements were the most intensively increased categories, at the expense of grassland and bare ground. Hence, when grassland was lost, it tended to be displaced by cultivation more than other categories, which was also true with bare land. Annual intensity gains were active for forest but minimal for cultivation, implying that the gains of forest were associated with in situ reforestation practices and that the gains in cultivation were caused by its relatively large initial area under a uniform intensity concept. This study demonstrates that IA is valuable for investigating LC across time intervals and can help distinguish dormant vs. active and targeted vs. avoided land categories.
早期的土地变化（LC）研究主要集中在规模和幅度、增益和损失，或类别之间的土地转移。因此，这些研究未能同时展示完整的 LC 过程。本文以埃塞俄比亚奥罗米亚州 Legedadie-Dire 集水区的土地类别图为基础，利用强度分析（IA），在三个时间点上对土地变化进行了研究。我们全面分析 LC，共同涵盖速率、强度、转移和过程。使用 TerrSet-LCM 和 ArcGIS 处理了 1986 年、2000 年和 2015 年的 30 米美国地质调查局（USGS）陆地卫星图像（<10%的云）。使用最大似然分类技术确定了六个类别：定居点、种植、森林、水、草原和裸地。然后，将分类图叠加在图像上，通过 IA 进行统计检查变化。除了水之外，1986-2000 年和 2000-2015 年期间，各类别之间观察到相当大的变化。整个土地变化起初迅速，然后在第二个时间间隔缓慢。在景观中，表现出变化的土地总面积（第一阶段≈54%，第二阶段≈51%）超过了保持不变的土地总面积（第一阶段≈46%，第二阶段≈49%）。种植和人类住区是增长最强烈的类别，以牺牲草原和裸地为代价。因此，当草原消失时，它往往被种植所取代，而不是其他类别，裸地也是如此。森林的年强度增益是活跃的，但对种植的增益是最小的，这意味着森林的增益与原地重新造林实践有关，而种植的增益是由于其在统一强度概念下相对较大的初始面积引起的。本研究表明，IA 对于研究跨时间间隔的 LC 非常有价值，并且可以帮助区分休眠与活跃以及有针对性与避免的土地类别。
