Area of Ecology, Faculty of Biological and Environmental Sciences, University of León, 24071,León, Spain.
Area of Ecology, Faculty of Biological and Environmental Sciences, University of León, 24071,León, Spain.
J Environ Manage. 2022 Mar 1;305:114373. doi: 10.1016/j.jenvman.2021.114373. Epub 2021 Dec 24.
In burned landscapes, the recruitment success of the tree dominant species mainly depends on plant competition mechanisms operating at fine spatial scale, that may hinder resource availability during the former years after the disturbance. Data acquisition at very high spatial resolution from unmanned aerial vehicles (UAV) have promoted new opportunities for understanding context-dependent competition processes in post-fire environments. Here, we explored the potentiality of UAV-borne data for assessing inter-specific competition effects of understory woody vegetation on pine saplings, as well as intra-specific interactions of neighboring saplings, across three burned landscapes located along a climatic/productivity gradient in the Iberian Peninsula. Geographic object-based image analysis (GEOBIA), including multiresolution segmentation and support vector machine (SVM) classification, was used to map pine saplings and understory shrubs at species level. Input data were, on the one hand, multispectral (11.31 cm·pixel) and Structure-from-Motion (SfM) canopy height model (CHM) data fusion, hereafter MS-CHM, and, on the other, RGB (3.29 cm·pixel) and CHM data fusion, hereafter RGB-CHM. A Random Forest (RF) regression algorithm was used to evaluate the effects of neighborhood competition on the relative growth in height of 50 pine saplings randomly sampled across the MS-CHM classified map. Circular plots of 3 m radius were set from the centroid of each target pine sapling to measure percentage cover, mean height of all individuals in the plot and mean height of individuals contacting the target sapling. Competing shrub species were differentiated according to their fire-adaptive traits (i.e. seeders vs resprouters). Object-based image classification applied on MS-CHM yielded higher overall accuracy for the three sites (83.67% ± 3.06%) than RGB-CHM (74.33% ± 3.21%). Intra-specific competitive effects were not detected, whereas increasing cover and height of shrub neighbors had a significant non-linear impact on the growth on pine saplings across the study sites. The strongest competitive effects of seeder shrubs occurred in open areas with low vegetation cover and fuel continuity, following a gap-dependent model. The non-linear relationships evidenced in this study between the structure of neighboring shrubs and the growth of pine seedlings/saplings have profound implications for considering possible competing thresholds in post-fire decision-making processes. These results endorse the use of UAV multispectral and SfM photogrammetry as a valuable post-fire management tool for measuring accurately the effect of competition in heterogeneous burned landscapes.
在火烧迹地中，树木优势种的繁殖成功主要取决于在精细空间尺度上运作的植物竞争机制，这些机制可能会在干扰后的前几年阻碍资源的可用性。来自无人机 (UAV) 的超高空间分辨率数据为理解火烧迹地后环境中依赖情境的竞争过程提供了新的机会。在这里，我们探索了使用 UAV 数据评估在伊比利亚半岛气候/生产力梯度上的三个火烧迹地中，林下木本植被对松树苗的种间竞争影响以及相邻树苗的种内相互作用的潜力。地理对象图像分析 (GEOBIA)，包括多分辨率分割和支持向量机 (SVM) 分类，用于在种水平上对松树苗和林下灌木进行映射。输入数据一方面是多光谱（11.31cm·pixel）和运动结构（SfM）冠层高度模型（CHM）数据融合，简称 MS-CHM，另一方面是 RGB（3.29cm·pixel）和 CHM 数据融合，简称 RGB-CHM。随机森林 (RF) 回归算法用于评估邻域竞争对 50 株随机采样的松树苗相对生长高度的影响，这些树苗是在 MS-CHM 分类图上进行的。从每个目标松树苗的质心设置半径为 3m 的圆形样方，以测量样方内所有个体的百分比覆盖度、样方内所有个体的平均高度和与目标树苗接触的个体的平均高度。根据其火适应特征（即种子传播者与再生者）来区分竞争灌木物种。在 MS-CHM 上应用基于对象的图像分类，三个地点的总体精度均较高（83.67%±3.06%），而 RGB-CHM 的总体精度较低（74.33%±3.21%）。未检测到种内竞争效应，而灌木邻居的覆盖度和高度增加对研究地点内的松树生长有显著的非线性影响。在植被覆盖度低、可燃物连续性差的开阔区域，种子传播者灌木的竞争效应最强，遵循依赖空隙的模型。在这项研究中，证明了相邻灌木的结构与松树苗/松苗生长之间的非线性关系，这对考虑火灾后决策过程中的可能竞争阈值具有深远意义。这些结果支持使用 UAV 多光谱和 SfM 摄影测量作为一种有价值的火灾后管理工具，用于准确测量异质火烧迹地中竞争的影响。
