Estimating Forest Forest Structure to inform Predictive Ecosystem Mapping (PEM) Approaches across British Columbia

不列颠哥伦比亚省以森林生态系统为主。绘制和了解这些森林生态系统的一个关键方法是通过预测生态系统制图（PEM）系统。PEM的一个关键优点是，作为一个自动化预测系统，它可以开发和应用于大的空间区域。最近可用的光探测和测距（LIDAR）数据提供了前所未有的地形和森林结构信息，可以与PEM预测相结合，为该省关键地区提供非常精细的空间细节和准确性。该项目的重点是(a)开发新的模型和方法，根据激光雷达数据预测森林结构的组成部分，供PEM从业者使用；(b)评估森林结构的哪些变化可以与树种、地点、年龄、干扰历史和森林健康的变化联系起来；(c)研究开发的方法在PEM景观单元内部和之间的可转移性，包括创新地使用深度学习；（iv）评估数据模型，以便将基于栅格的森林结构估计有效地总结为多边形单元。该项目将通过专门的博士HQP提供额外的智力支持，并利用已分配给该项目的个人、数据和实地观察方面的重要资源，帮助提出和开发创新方案，以绘制BC省景观中的生态单元。

British Columbia is dominated by forested ecosystems. One key approach to map and understand these forested ecosystems is through the Predictive Ecosystem Mapping (PEM) system. A key benefit of PEM is that, as it is an automated predictive system, it can be developed and applied over large spatial areas. The recent availability of light detection and ranging (LIDAR) data which provides unprecedented information on both terrain and forest structure can be used in association with PEM predictions to provide a very fine level of spatial detail and accuracy acros key areas of the Province. The project focuses on (a) development of new models and methods to predict components of forest structure from LiDAR data, for use by PEM practitioners, (b) an assessment of what variations in forest structure can be linked to changes in tree species, site, age, disturbance history, and forest health, (c) examine how transferable the developed methods are within and across PEM landscape units including the innovative use of deep learning and (iv) an assessment of data models to allow raster based estimates of forest structure to be effectively summarized into polygon units. This project will support the BC Ministry PEM Methods team by providing additional intellectual support through a dedicated PhD HQP and, using the significant resources already assigned to this project in terms of personal, data and field observations, help propose and develop innovations to map ecological units across the BC landscape.