Characterising forest landscapes using a stationary and a roving tower eddy covariance system: linking point data to landscape-level characterisations of growth and yield

在生长期森林中采用的基于涡动协方差的净生态系统生产力（NEP）测量提供了测量林分木材产量的最佳方法。基于森林过渡规则的林分水平森林模型或单株树木之间模拟的相互作用（林隙模型）提供了林分生长和木材产量在不同精度水平上的概貌。这些模型大多依赖于历史树数据进行模型公式校准，其生长行为有很好的记录。在不断变化的气候条件下，树木的生长行为最有可能发生变化，这使得用森林过渡模型预测未来森林生长不合适。此外，景观应用间隙模型是令人望而却步的，因为需要大量的计算机资源来校准和运行可能有数以万计的树木的景观模型，其中许多具有不同的属性。由于模型的复杂性和对数据的要求，现有的基于生态系统的NEP模型在景观中的应用也不现实。由于这些限制，建议的工作是通过将模型的每月生长预测与当地生物物理条件联系起来，将现有的10参数NEP模型扩展到景观，从而开发森林G&Y。这些条件将通过将基于遥感的温度和叶面积图像与太阳辐射和土壤含水量的景观级模拟相结合而得出。森林生态系统的特征和模型参数的设定将主要基于固定基站和漫游塔系统收集的信息。37米高的固定塔将提供长期的基础天气和生态数据。这些信息以及在森林站点通过流动塔收集的新生态环境预测和天气数据将用于获得不同森林覆盖所需的森林特定生长参数，并延长3-4周的新生态环境预测时间序列，以便在景观水平上应用。提出的工作提供了一个重大的科学创新，在G&Y模型。

Eddy covariance-based net ecosystem productivity (NEP) measurements taken over growing forests provide the best possible measure of stand wood production. Stand-level forest models based on forest transition rules or modelled interactions between individual trees (gap models) provide generalisations of stand growth and wood production at various levels of accuracy. Most of these models rely on historical tree data for model formulation-calibration, for which growth behavior is well documented. In a changing climate regime, tree growth behavior is most likely to change, rendering future forest growth projections with forest transition models unsuitable. Also, landscape application of gap models is prohibitive because of the huge amount of computer resources needed to calibrate and run the model for a landscape of potentially tens of thousands stands, many with different attributes. Landscape application of existing ecosystem-based NEP models is also not practical, because of the models' complexity and data requirements. Because of these limitations, the proposed work is to develop a forest G&Y by extending an existing 10-parameter NEP model to the landscape by linking the model's monthly growth projections to local biophysical conditions. These conditions are to be derived by combining remote sensing-based images of temperature and leaf area with landscape-level simulations of solar radiation and soil water content. Forest ecosystem characterisation and model parameter setting are to be based mostly on information collected by a stationary base station and roving tower system. The 37-m stationary tower will provide long term base weather and ecological data. This information along with NEP and weather data collected at forest sites with a roving tower will be used to derive the required forest-specific growth parameters for different forest cover, and extend 3-4 week-long time series of NEP for application at the landscape level. The proposed work provides a major scientific innovation in G&Y modelling.