社区水平森林景观格局的动态变化与未来情景模拟

论文以重庆市石柱县三星乡石星村为样区,基于2004年森林景观格局的相关数据,运用Logistic逐步回归方法筛选出了不同时期对石星村森林景观格局具有决定作用的影响因子,并采用CLUE-S模型对石星村2014年森林景观分布格局进行了模拟与验证。此基础上,分别基于历史发展趋势、“农二代”返乡潮以及工商资本介入3种情景模拟石星村2024年的森林景观格局,利用景观格局指数对2004-2024年间3种情景下的景观格局动态及破碎度进行研究。结果表明：1） 利用石星村2004年的森林景观格局分布图对2014年森林景观格局进行模拟,总体精度达到了85%,Kappa指数均值达到0.816,这表明CLUE-S模型对石星村森林景观的格局变化具有良好的模拟能力;2） 在预设的3种情景下,样区森林景观一直占据20 a间景观基质的主导地位,林区分析结果显示出林地总面积没有明显减少,且相较于2004年的林地总面积表现出增加的趋势。3种情景下退化林地面积均呈减少的趋势,其中,相较于历史发展趋势,“农二代”返乡潮情景下退化林地的减少均伴随着人工林和农用地面积的增加,而工商资本介入情景下退化林地的减少伴随着人工林面积的增加和农用地面积的减少;3） 在林地的空间分布上,各林地类型的分布存在一定的规律性,其中退化原始林主要集中分布在中东部低山、深丘区,次生林、退化林地和人工林呈镶嵌格局,分散于主要基质性景观中;4） 样区在3种情景模拟下,森林景观破碎化程度存在差异。总体来说,“农二代”返乡潮和工商资本介入两种情景下森林景观恢复较好。研究结论可为石星村未来森林景观管理、规划及政策的制定等提供参考依据和决策支持。

Forest Landscape Pattern Dynamic Change and Scenarios Simulation at Community Level

This paper took Shixing village in Sanxing town, Shizhu County as the sample area. Based on the forest landscape pattern in 2004, the Logistic stepwise regression was adopted to select factors playing essential roles in the forest landscape pattern in different periods, which were terrain, altitude, drainage course, transportation and distribution of the settlement. CLUE-S was applied to simulate the distribution of the forest landscape pattern in Shixing village in 2014, and the result was validated by comparing with the actual pattern. Then the forest landscape pattern ten years later was simulated based on three scenarios: the historical development trend, the boom of returning home of the second generation of farmers, and the intervention of industrial and commercial fund. Besides, this paper analyzed the landscape dynamics from 2004 to 2024 by using the landscape pattern index. There are four aspects of the result. Firstly, based on the distribution pattern of the forest landscape in 2014, the precision of forest landsacpe pattern in 2014 was 85%, and the average Kappa coefficient was over 0.816, which illustrated the applicability of CLUE-S. Secondly, in the three scenarios, the forest landscape always occupied the main position in the landscape matrix during the 20 years from 2004, and the result illustrates that the total area of the forest increased compared with that in 2004. In addition, the degraded forest lands decreased in all three scenarios. In the second scenario, the reductions of the degraded forest land accompanied with the increase of the artificial forest and the agricultural land. In the third scenario, the reductions of the degraded forest land accompanied with the increase of the artificial forest and the decrease of agricultural land. Thirdly, the spatial distributions of the forest lands are regular. The degraded primary forests are mainly seen in the hills and deep hillock areas, while the secondary forests, degraded forest land and artificial forests are in mosaic structure, being scattered in some matrix landscape. Fourthly, the degree of forest landscape fragmentation is different in three simulated scenarios. Generally, the recoveries of forest landscape are better in the second and the third scenarios. The result of this study will provide the reference and support to the administration, planning, and policy making of the forest landscape in following years.