Multi-scale assessment of river-floodplains

由于自然和人为两种力量的作用，河流生态系统是高度动态和异质的生态系统。沿海河岸森林的异质性使它们成为最具生物多样性的温带森林之一，但在广阔地区绘制地图也是一项挑战。滨海洪泛区森林内各种重要的生态结构导致了不同尺度的空间格局。其中许多图案包括纹理、空间和光谱特性，这有助于利用高空间分辨率卫星图像区分它们。细比例尺图像(分辨率低于米)用于河岸测绘的潜力尤其令人信服，因为使用较粗糙的图像识别这类区域一直是一项挑战。这项拟议的研究计划的目标是使用一套新的遥感工具和地面真实信息，严格量化河流泛滥平原自然和人为变异性的等级模式。我们强调在我们的试点开发的技术在Clayoquot Sound内新的更大的泛滥平原上的使用、推广和测试，重点是稀有的受保护关注的栖息地以及自然和人为干扰的模式。我们将使用最先进的基于对象的分类软件对不列颠哥伦比亚省沿海三条河流沿线的高空间分辨率图像进行分类，并使用更粗尺度的图像和景观格局指数评估大范围的可变性。使用这一工具的组合来了解河流泛滥平原在精细和广泛空间尺度上的可变性，将有助于理清导致河流泛滥平原生态系统变化的人为和自然因素。作为我们长期目标的一部分，我们还审查了海达瓜伊州和斯基纳河沿岸的更多地点，其中包括原始地区以及那些经过历史森林管理改造的地点。这些相同的河流系统也将在历史框架内进行审查，方法是利用历史重建和航拍照片量化过去的变化。这些不同的方法将增加我们对公元前一些最具生物多样性和管理最严格的河流泛滥平原系统的理解的关键维度。

Riverine ecosystems are highly dynamic and heterogeneous ecosystems, due to both natural and anthropogenic forces. The heterogeneity of coastal riparian forests makes them among the most biodiverse of temperate forests, but also challenging to map over broad areas. A variety of ecologically important structures within coastal floodplain forests result in spatial patterns at different scales. Many of these patterns include textural, spatial and spectral properties which aid in their discrimination using high spatial resolution satellite imagery. The potential of fine-scale imagery (with sub-metre resolution) for use in riparian mapping is particularly compelling, as such areas have been challenging to discern using coarser imagery. The goal of this proposed research program is to rigorously quantify hierarchical patterns of natural and anthropogenic variability in river-floodplains using a suite of new remote sensing tools and ground-truth information. We emphasize the use, extension, and testing of techniques developed in our pilot site to new larger floodplains within Clayoquot Sound, with an emphasis on rare habitats of conservation concern and patterns of natural and anthropogenic disturbance. We will classify high spatial resolution imagery along three rivers in coastal British Columbia using state-of-the-art object-based classification software, as well as assess broad-scale variability using coarser-scale imagery and landscape pattern indices. Using this combination of tools to understand the variability within river-floodplains, at both fine and broad spatial scales, will help disentangle the anthropogenic and natural agents of change in river-floodplain ecosystems. As part of our long-term objectives, we also examine additional sites in Haida Gwaii and along the Skeena River, which include pristine areas as well as those modified by historic forest management. These same river systems will also be examined within an historical framework, by quantifying past variability using historic reconstruction and aerial photos. These different approaches will add a critical dimension to our understanding of some of the most biodiverse and heavily managed river-floodplain systems in BC.