Doctoral Dissertation Research: Vegetation Processes as Mediated by Flooding and Fire in the Okavango Delta, Botswana

该项目重点关注博茨瓦纳北部的奥卡万戈三角洲，这是世界上最大的淡水湿地之一（约20,000平方公里）。奥卡万戈三角洲每年都会经历两次重大的动乱，洪水和火灾。本项目可获得1989 - 2003年期间的遥感时间序列数据库，包括96多幅中分辨率（30米）和4幅高分辨率（4米）图像。本研究的主要目的是通过以下方法提高对这些干扰对植被过程的影响的理解：(1)从遥感数据中提取由洪水和火灾干扰机制介导的时空植被模式；(2)利用生态模型在不同时空尺度上研究与上述干扰机制相关的生态系统过程。来自遥感图像的信息将用于绘制三角洲地区因干扰（洪水和火灾）而产生的植物反应。干扰的频率将使用一维傅立叶分析提取，并与已知的降水振荡相关联。将利用遥感资料跟踪1989-2003年期间观测到的扰动引起的生态变化。此外，叶面积指数（衡量植物活力的指标）将使用来自Landsat TM、ETM+和ALI卫星数据的植被指数来计算，并与叶面积指数的野外测量结果相关联。土地覆盖信息、叶面积指数和气候数据将被输入生态系统模型，通过15年的图像序列跟踪净生态系统生产力。地球上的生态系统过程发生在不同的尺度上，这里提出的工作将解决通过遥感数据将植被特征纳入生态系统模型的问题。这些结果不仅可以更好地描述跨尺度的关键景观过程，而且还可以提供关于观测和建模尺度如何影响对全球气候变化的更广泛理解的详细信息。将野外收集的生态系统信息与遥感数据联系起来的能力是一个正在进行的研究课题，对未来的科学方向有许多影响。遥感提供了在大范围内进行测量的手段，但从林分水平到观测水平的地表参数换算机制尚不清楚。大多数生物地球化学过程不是从一个尺度到另一个尺度的线性缩放，并且由于景观的空间异质性而变得复杂。该项目将解决生态规模化和持续成功利用遥感数据所面临的许多问题。

This project focuses on the Okavango Delta of northern Botswana, one of the world's largest freshwater wetlands (approximately 20,000 km2). The Okavango Delta annually experiences two significant disturbance regimes, flooding and fire. A remote sensing time series database spanning the period 1989 - 2003 period is available to this project and consists of over 96 medium resolution images (30 m) and 4 high resolution images (4 m). The primary objective of this research is to improve understanding the impacts of these disturbances on vegetation processes by 1) extracting the vegetation patterns over both space and time as mediated by flooding and fire disturbance regimes from remotely sensed data and 2) examining the ecosystem processes related to the previously mentioned disturbance regimes at various spatial and temporal scales using ecological modeling. Information derived from remotely sensed imagery will be utilized to map the vegetative response due to disturbances (flooding and fire) that occur in the Delta. Frequency of disturbances will be extracted using a 1-D Fourier analysis and correlated to known oscillations of precipitation. Ecotonal shifts due to observed disturbances will be tracked for the 1989-2003 period using remotely sensed data. Additionally, leaf area index (a measure of plant vigor) will be computed using vegetation indices derived from Landsat TM, ETM+, and ALI satellite data and correlated to field derived measurements of leaf area index. Land cover information, leaf area index, and climatic data will be fed into ecosystem models to track net ecosystem productivity through the 15-year sequence of imagery.Ecosystem processes on the Earth occur at various scales, and the work presented here will address the incorporation of vegetation characterization via remotely sensed data into ecosystem models. These results will provide not only better descriptions of critical landscape processes across scales but moreover refined information on how the scale of observation and modeling impacts broader understanding of global climate change. The ability to link ecosystem information collected at the field level with remotely sensed data is an ongoing research topic and has many implications for future directions of science. Remote sensing provides the means to make measurements over extended areas, yet the mechanisms of scaling surface parameters from the stand level to the level of observation are not known. Most biogeochemical processes do not scale linearly from one scale to another and are complicated by spatial heterogeneity across a landscape. This project will address many of the issues facing ecological scaling and the continued successful utilization of remotely sensed data.