西北欧全新世中期的Elm Decline研究新进展

在５０００ａ ＢＰ西北欧发生的Ｅｌｍ Ｄｅｃｌｉｎｅ，是全新世突发性环境事件或灾变过程的表现。它在６０年来一直是欧洲环境变迁，考古和＾１４Ｃ年代学界关心的热点论题，对其研究进展作了系统的评述，以了解其研究现状，促进我国全新世环境变迁和人地关系演变科学研究。     

基于面向对象与深度学习的榆树疏林识别方法研究

榆树疏林是浑善达克沙地中一种特殊的植被类型,它对于维持区域生态系统稳定具有重要意义,在防风固沙、涵养水源、调节气候等方面发挥着重要的作用。本文利用无人机影像与GF-2影像,对高分辨率数据源中榆树疏林的两种自动识别方法进行了研究。在面向对象方法中,首先通过计算影像对象的局部方差变化率得到了最佳分割尺度;其次采用随机森林算法对初选特征的重要性进行排序,并删除无关特征;最后分别对支持向量机（SVM）、随机森林（RF）、深度神经网络（DNN）3种分类器进行参数寻优与榆树疏林提取。此外,在ENVI5.5中基于TensorFlow框架,利用U-Net构建深度学习模型对榆树疏林进行了提取,并与面向对象方法进行对比。结果显示：① 通过面向对象方法过程的优化,最终的识别精度较以往研究有所提升,GF-2影像中SVM总体精度为90.14%,RF总体精度为 90.57%,DNN总体精度为91.14%;无人机影像中SVM总体精度为97.70%, RF与DNN总体精度为97.42%。② 深度学习方法中,GF-2影像的总体精度为91.00%,无人机影像的总体精度达到了98.43%。研究结果说明在榆树疏林提取中,无人机影像具有更高的空间分辨率,更丰富的纹理、形状等信息,能达到比GF-2影像更高的精度。面向对象方法对于2种影像都有较高的适用性;深度学习的方法在本文中更适用于无人机影像,它可以有效地减少无人机影像中的错分现象。     

Vegetation and Community Changes of Elm （Ulmuspumila） Woodlands in Northeastern China in 1983-2011

Elm (Ulmus pumila), widely distributed in the north temperate zone, contributes to a special savanna-like woodland in typical grassland region in the northeastern China. This woodland performs a variety of ecological functions and environmental significance, such as decreasing soil erosion, stabilizing sand dunes, preserving species diversity. However, in the last approximate 30 years, the species composition, productivity and distribution area of elm woodland has decreased severely. A series of studies have been carried out to find out whether the climate changes or human disturbances caused the degradation of elm woodland and how these factors affected elm woodland. In this study, undisturbed, plowing and grazing elm woodland were investigated in 1983 and 2011 by using Point-Centered Quarter method. The relationship between vegetation changes and environmental factors was analyzed by Bray-Curtis ordination. The results show that in 2011, species diversity and understory productivity of undisturbed elm woodland decrease slightly compared to those of undisturbed elm woodland in 1983. However, nearly 60% of the species is lost in the plowing and grazing elm woodland relative to the species undisturbed elm woodland in 1983. Interestingly, plowing stimulates the growth of elm and certain understory species through furrowing soil and accelerating soil nutrient turnover rate. Grazing disturbance not only leads to species loss and productivity decrease, but also induces changes in elm growth (small, short and twisted). The mean age of the elm was 29 ± 2 yr in undisturbed and plowing elm woodland, while only 15 yr in the grazing elm woodland. The results of Bray-Curtis ordination analysis show that all sample stands clustered to three groups: Group I including the undisturbed sample stands of 83UE (undisturbed elm woodland in 1983) and 11UE (undisturbed elm woodland in 2011); Group II including sample stands of PE (elm woodland disturbed by plowing); Group III including samples stands of GE (elm woodland disturbed by grazing). The results indicate that the long time disturbance of the plowing and grazing have converted elm woodland to different community types. Climate change is not the primary reason causing the degradation of elm woodland, but plowing and grazing disturbance. Both plowing and grazing decrease the vegetation composition and species diversity. Grazing further decreases vegetation productivity and inhibits the growth of elm tree. Therefore, we suggest that reasonable plowing and exclusive grazing would be favorable for future regeneration of degraded elm woodland.