Department of Forestry and Forest Products, Ayancık Vocational School, Sinop University, 57402, Sinop, Turkey.
Department of Forest Management and Planning, Faculty of Forestry, Bartın University, 74100, Bartın, Turkey.
Environ Monit Assess. 2021 Jul 23;193(8):507. doi: 10.1007/s10661-021-09281-x.
The purpose of this study, covering the northern Ulus district of Turkey, was to analyze the forest and land use/land cover (LULC) changes in the past period from 2000 to 2020, and to predict the possible changes in 2030 and 2040, using remote sensing (RS) and geographic information systems (GIS) together with the CA-Markov model. The maximum likelihood classified (MLC) technique was used to produce LULC maps, using 2000 and 2010 Landsat (ETM +) and 2020 Landsat (OLI) images based on existing stand-type maps as reference. Using the historical data from the generated LULC maps, the LULC changes for 2030-2040 were predicted via the CA-Markov hybrid model. The reliability of the model was verified by overlapping the 2020 LULC map with the 2020 LULC model (predicted) map. The overall accuracy was found to be 80.90%, with a Kappa coefficient of 0.74. The total forest area (coniferous + broad-leaved + mixed forest) grew by 10,656.4 ha (15.4%) in the 2000-2020 period. Examination of the types within the Forest Class revealed that the coniferous forest area had grown by 5.9% in the period 2000-2010, whereas it had decreased by 4.7% in the period 2010-2020. The broad-leaved forest area had grown by 1.2% and 3.1%, respectively, between 2000 and 2010 and 2010 and 2020. The mixed forest area had been reduced by 7.1% in the period 2000-2010 but had grown by 1.7% in the 2010-2020 period. In the Non-Forest Class, although the water area had increased in the 2000-2020 period, agricultural land and settlement areas had decreased by 11,553.9 ha (32.3%) and 34.6 ha (0.5%), respectively. According to the 2020-2040 LULC simulation results, it was predicted that there would be 3.8% and 26.4% growth in the total forest and water surface areas and 13.9% and 5.3% reduction in the agricultural and settlement areas, respectively. Using the LULC simulation to separate the Forest Class into coniferous, broad-leaved, and mixed forest categories and subsequently examining the individual changes can be of great help to forest planners and managers in decision-making and strategy development.
本研究以土耳其北部乌鲁斯地区为研究区域，利用遥感（RS）和地理信息系统（GIS）结合 CA-Markov 模型，分析了 2000 年至 2020 年期间的森林和土地利用/土地覆盖（LULC）变化，并预测了 2030 年和 2040 年可能的变化。采用最大似然分类（MLC）技术，使用 2000 年和 2010 年 Landsat（ETM+）和 2020 年 Landsat（OLI）图像，根据现有林型图作为参考，生成 LULC 图。利用生成的 LULC 图的历史数据，通过 CA-Markov 混合模型预测 2030-2040 年的 LULC 变化。通过将 2020 年 LULC 图与 2020 年 LULC 模型（预测）图进行叠加，验证了模型的可靠性。模型的总体精度为 80.90%，kappa 系数为 0.74。在 2000-2020 年期间，森林总面积（针叶林+阔叶林+混交林）增加了 10656.4 公顷（15.4%）。对森林类内的类型进行检查发现，2000-2010 年期间针叶林面积增加了 5.9%，而 2010-2020 年期间减少了 4.7%。阔叶林面积分别增加了 1.2%和 3.1%，2000-2010 年和 2010-2020 年。混交林面积在 2000-2010 年期间减少了 7.1%，但在 2010-2020 年期间增加了 1.7%。在非森林类中，尽管 2000-2020 年期间水域面积有所增加，但农业用地和定居点面积分别减少了 11553.9 公顷（32.3%）和 34.6 公顷（0.5%）。根据 2020-2040 年 LULC 模拟结果，预计森林和水面总面积将分别增加 3.8%和 26.4%，农业和定居点面积将分别减少 13.9%和 5.3%。通过利用 LULC 模拟将森林类分为针叶林、阔叶林和混交林类别，并随后检查各个类别中的变化，这对森林规划者和管理者在决策和战略制定方面非常有帮助。
