IRCEB: Biological Control of Terrestrial Carbon Fluxes

9977066 ojima从全球角度来看，了解陆地生态系统在碳循环中的作用对于量化碳储量至关重要。近年来，在测量网络（LTER和AmeriFLUX）、库存数据再分析和大空间尺度碳预算建模方面取得了很大进展。此外，对大气数据的“反向”分析开始提供大陆尺度的信息。在这个IRCEB研究项目中，我们将开发一个综合数据模型系统来分析关于生物和土地管理的不同假设对大气中二氧化碳模式的影响。这是针对特定地点数据的传统模型测试的有力补充，但测量网络和梯度研究的发展大大提高了原位数据模型比较的能力。对陆地生态系统和碳循环进行雄心勃勃的综合分析的时机已经成熟。这个项目的重点是美国，一个数据资源和验证信息无与伦比的地区，这个地区的多样性足以严重挑战我们的理解，同时保持可处理性。本项目有三个组成部分：(1)基于生态系统模型、过程研究和分析技术的最新进展，实施一个新的框架来分析控制净碳交换的过程。该项目将开发一个模块化框架，将干扰和土地利用与生命形式分布和生物地球化学联系起来，以便提供一种全面的方法来检查和估计陆地净碳交换。这个框架将由我们的团队开发，但会定期听取同事和潜在用户的意见。该模型将是一个“社区”模型，设计用于多个平台和远程操作。模型代码和来自重要实验的大型输出数据集，以及文档和元数据都将可用。该模型将利用参与者吸取的经验教训，并将利用现有模型中的科学知识。然而，我们将整合关键组成部分和过程，以评估陆地碳通量的变化，而不是简单地“连接”现有模型。(2)模型运行所需数据集的开发。这些课程包括土壤、气象学和土地利用史。数据活动的主要重点将是开发空间土地利用历史，其中包含操作模型所需的足够信息。这项活动将利用现有的努力，作为一项巨大的任务，将随着时间的推移逐步得到改进。(3)美国碳预算分析。我们将首先根据NPP、土壤碳/生物量数据等“传统的”现场观测来评估我们的模型系统。然后，我们将整合大陆模式，并将模拟的大气二氧化碳模式与观测结果进行比较。这将需要将陆地模式和估计的空间/季节化石燃料通量与大气模式相结合。我们将使用一个经过良好测试的模型（RAMS），该模型以“数据同化模式”运行，这是一种数学方法，该模型不断调整，以适应预测研究中广泛使用的关键变量的观测结果。数据同化是大气科学中的一项成熟技术，它将使模式产生接近观测条件的输送风、湍流通量和天气。这将使作为生态系统模型输入的天气和用于比较模拟和观测的二氧化碳的输送量保持一致并接近现实。

9977066OjimaFrom a global perspective, understanding the role of terrestrial ecosystems in the carbon cycle is crucial to quantifying carbon storage. Great progress has been made in recent years in measurement networks (LTER and AmeriFLUX), in the re-analysis of inventory data, and in modeling relevant to carbon budgets at large spatial scales. In addition, 'inverse' analyses of atmospheric data are beginning to provide information on continental scales. In this IRCEB research project, we will develop an integrated data-model system to analyze consequences of different assumptions about biology and land management on patterns of CO2 in the atmosphere. This is a powerful complement to traditional model testing against site-specific data, but the development of measurement networks and gradient studies greatly improves the power of in situ data-model comparisons. The time is ripe for an ambitious, integrated, analysis of terrestrial ecosystems and the carbon cycle. This project focuses on the conterminous US, an area where data resources and validation information are unparalleled, and an area diverse enough to seriously challenge our understanding while remaining tractable. This project has three integrated components: (1) The implementation of a new framework to analyze processes controlling net carbon exchange based on recent advances of ecosystem models, process studies, and analytical technologies. The project will develop a modular framework linking disturbance and land use to life form distribution and biogeochemistry, in order to provide a comprehensive way to examine and estimate terrestrial net carbon exchanges. This framework will be developed by our team, but with input from colleagues and potential users on a regular basis. The model will be a 'community' model designed to be operated on multiple platforms and remotely. Both the model code and large output data sets from important experiments will be made available, with documentation and metadata. This model will take advantage of lessons learned by the participants and will use science from extant models. However, rather than simply 'linking' extant models, we will integrate key components and processes to evaluate changes in terrestrial carbon fluxes. (2) The development of the data sets needed to operate the model. These will include soils, meteorology and land use histories. The main emphasis in the data activity will be the development of spatial land use histories containing sufficient information to operate the model. This activity will draw on existing efforts and, as a huge task, will be progressively improved over time. (3) The analysis of the US carbon budget. We will begin by evaluating our model system against 'traditional' in situ observations such as NPP, soil carbon/biomass data. We will then integrate the continental model and compare the simulated patterns of atmospheric CO2 against observations. This will require coupling the terrestrial model and estimated spatial/seasonal fossil fuel fluxes to an atmospheric model. We will used a well-tested model (RAMS) operated in a 'data assimilation mode', a mathematical approach where the model is continuously adjusted to observations of key variables widely used in forecast studies. Data assimilation is a mature technology in the atmospheric sciences and will permit the model to produce transport winds, turbulent fluxes, and weather close to observed conditions. This will make both the weather used as input to the ecosystem model and the transport used to compare simulated to observed CO2 consistent and close to reality.