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

该项目的重点是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，面积约为20,000平方公里。奥卡万戈三角洲每年经历两个重大的干扰制度，洪水和火灾。该项目拥有一个跨越1989-2003年期间的遥感时间序列数据库，其中包括96幅中分辨率图像(30米)和4幅高分辨率图像(4米)。本研究的主要目的是通过1)从遥感数据中提取洪水和火灾干扰模式在空间和时间上的植被格局，以及2)利用生态学模型在不同的空间和时间尺度上检验与上述干扰模式相关的生态系统过程，从而提高对这些干扰对植被过程的影响的理解。来自遥感图像的信息将被用来绘制由于三角洲地区发生的干扰(洪水和火灾)而产生的植被反应。扰动的频率将使用一维傅立叶分析来提取，并与已知的降水振荡相关联。将利用遥感数据跟踪-2003年期间观测到的扰动引起的气候变化。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，并与野外叶面积指数的测量结果相关联。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，这里介绍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。这些结果不仅将更好地描述不同尺度上的关键景观过程，而且还将提供关于观测和建模规模如何影响对全球气候变化的更广泛理解的精炼信息。将实地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主题，对未来的科学方向有许多影响。遥感提供了在大片地区进行测量的手段，但从林分水平到观测水平衡量地表参数的机制尚不清楚。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尺度，而且由于景观的空间异质性而变得复杂。该项目将解决生态尺度和继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.