Country-wide mapping of harvest areas and post-harvest forest recovery using Landsat time series data in Japan
Country-wide mapping of harvest areas and post-harvest forest recovery using Landsat time series data in Japan
复制标题
DOI：
10.1016/j.jag.2021.102555
复制
发表时间：
2021-12
期刊：
影响因子：
--
通讯作者：
Katsuto Shimizu;H. Saito
中科院分区：
文献类型：
--
作者：
Katsuto Shimizu;H. Saito
森林采伐和随后的植被恢复的特征为有效的森林管理提供了宝贵的见解。虽然陆地卫星时间序列数据提供了关于大面积森林干扰和恢复的空间上的明确信息，但对全国范围的收获和收获后恢复的详细描述是不够的。尽管在收获地区种植很重要，但通常不考虑在收获后由森林管理部门重新种植的地区的地图。这项研究调查了一种方法，利用1984至2020年的陆地卫星时间序列数据，在全国范围内探测收获和其他森林干扰地区，并表征收获后的恢复情况。为此，使用随机森林算法，利用从5个光谱指数的LandTrendr时间分段以及地形和气候数据获得的预测变量，对干扰剂和稳定的土地覆盖类别进行分类。收获后恢复的特点是森林物种组成(即针叶林/阔叶林)，然后用于连接收割区的补种地区。对于森林收获类，干扰因子/稳定土地覆盖分类的准确率分别为80.1%(±4.8%)和93.8%(±3.8%)。收获后恢复的总体准确率很高(83.9%)，与全日本补种面积的统计数据比较，趋势和估计面积的一致性很好，均方根误差为2687.8公顷(15.2%)。总体而言，过去35年间，日本森林采伐面积占国土总面积的4.6%。结果表明，约60.0%的针叶林被恢复为针叶林。针叶林采伐后的时空分布可能代表了采伐区森林经营的补种活动。本研究提出的方法有可能利用陆地卫星时间序列数据提供关于全国范围内收获区再植活动的有价值的信息。
The characterization of forest harvesting and subsequent vegetation recovery provides valuable insights for effective forest management. Although Landsat time series data offers spatially explicit information regarding large-area forest disturbance and recovery, the detailed characterization of country-wide harvest and post-harvest recovery is insufficient. Despite the importance of planting in harvest areas, the mapping of areas replanted by forest management after harvest are not usually considered. This study investigated an approach to detect harvest and other forest disturbance areas country-wide and to characterize post-harvest recovery using annual Landsat time series data from 1984 to 2020. To do so, a random forest algorithm was used to classify disturbance agents and stable land cover classes using predictor variables derived from the LandTrendr temporal segmentation of five spectral indices and topographic and climate data. Post-harvest recovery was characterized as forest species composition (i.e., coniferous/broadleaved forests) and then used to link replanted areas in harvest areas. The disturbance agents/stable land cover classification achieved producer’s and user’s accuracies of 80.1% (±4.8%) and 93.8% (±3.8%), respectively, for the forest harvest class. The overall accuracy of post-harvest recovery was high (83.9%) and a comparison with statistical data of replanted areas for entire Japan showed good agreement in trends and estimated area with a root mean square error of 2687.8 ha (15.2%). Overall, harvested forest area accounted for 4.6% of the total land area in Japan during the past 35 years. The results indicated that approximately 60.0% of coniferous plantation forests were recovered as coniferous forests. The spatial and temporal distribution of coniferous forests after harvest presumably represented the replanting activity of forest management in harvest areas. The approach presented in this study has the potential to provide valuable information on country-wide replantation activity in harvest areas using Landsat time series data.