An improved empirical model of soil carbon dynamics in temperate ecosystems

土壤中储存的碳是植被或大气中的两倍多，而全球碳收支和大气中二氧化碳的浓度对土壤碳的任何变化都很敏感。以前，人们认为大多数土壤碳被安全地锁起来了。但最近完成的一项由拟建项目成员对英格兰和威尔士土壤进行的调查显示，在过去25年里，碳的损失规模巨大：足以抵消英国减少的化石燃料碳排放。这项调查测量了近6000个地点的土壤样本中的碳含量，这些地点包括所有类型的土地利用/农田、草地、林地、高地荒原和沼泽。该研究于1978年至1983年首次进行，1994年至2003年在40%的地点重复进行。大多数地点都出现了损失，而且损失发生在所有类型的土地利用中。虽然土地管理的各种变化将导致损失/例如，20世纪50年代在草原上耕作作物/所有形式的土地利用都发生了损失，这一事实表明与气候变化有关。其机制是众所周知的。目前，土壤和植被被认为吸收了燃烧化石燃料排放到空气中的大约四分之一的二氧化碳。由于植物的生长速度更快，二氧化碳含量增加，所以这个比率正在上升。但随着气温的升高和土壤含水量的变化，土壤中死亡植被和其他有机质的周转速度以及相应的CO2释放也在增加。在某一时刻，二氧化碳释放的速度将超过吸收的速度，全球陆地表面将从二氧化碳的汇变成二氧化碳的源。因此，稳定二氧化碳水平需要更大幅度的减排。基于计算机建模研究，人们认为这需要几十年的时间。但新的结果表明，这种情况已经在英格兰和威尔士发生了，这意味着减排的紧迫性更加迫切。分离出造成英格兰和威尔士各地损失的各种可能原因是很复杂的。这可能涉及各种各样的土地管理因素，它们的影响与气候变化相互作用。因此，对所有可能的因素组合进行测量是不切实际的。在任何情况下，从这些测量中得出的统计关系都不能揭示因果关系，而且它们在揭示各种非线性关系和正在运行的反馈过程方面也很差。基于对底层过程的理解，需要某种形式的数学建模。但目前的模型还不足以解决这个问题。在这里，我们建议通过改进处理土壤碳转化的组件，以及它们对土壤水分、温度和其他变量的依赖，以及土壤水分、温度和其他驱动变量的计算，来改进现有的最佳土壤碳模型/世纪模型。我们将汇集(1)上面讨论的关于英格兰和威尔士大范围土壤、气候和土地利用的土壤碳随时间变化的新的国际独特数据集；(2)现有土壤碳转化和相关土壤条件动态的最佳模型，以及模拟这些模型的新方法；(3)新的空间统计技术用于分析待建模数据集和模型预测的可靠性；(4) 4位土壤模型专家和2位土壤空间统计专家，他们在这些领域的专业知识都得到了国际上的认可。

More than twice as much carbon is stored in soils as in vegetation or the atmosphere, and the global carbon budget and the concentration of carbon dioxide (CO2) in the atmosphere are sensitive to any changes in soil carbon. Previously it was thought that most soil carbon was safely locked away. But a recently-completed survey of the soils of England and Wales by members of the proposed project has shown losses of carbon over the past 25 years on an enormous scale: enough to cancel the UK's reductions in carbon emissions from fossil fuels. The survey measured carbon in soil samples from almost 6,000 sites across all types of land use / cropland, grassland, woodland, upland heath, bogs. It was first carried out in 1978-83 and then repeated at 40% of the sites between 1994 and 2003. Most sites showed losses, and the losses occurred across all types of land use. Whilst various changes in land management will have contributed to the losses / for example, the ploughing of grassland for crops in the 1950s / the fact that the losses occurred across all forms of land use suggests a link to climate change. The mechanisms for this are well known. At the moment, soils and vegetation are thought to absorb about a quarter of the CO2 emitted into the air from burning fossil fuels. The rate is increasing because plants are growing faster, fertilized by higher CO2 levels. But as temperatures rise, and the moisture content of the soil changes, the rate of turnover of dead vegetation and other organic matter in the soil and the corresponding release of CO2 are also increasing. At some point, the rate of release will overtake the rate of absorption, and the global land surface will switch from being a sink for CO2 to a source. Stabilising CO2 levels will then require much greater cuts in emissions. It was thought, based on computer modelling studies, this would take several decades. But the new results suggest it is happening already in England and Wales, implying an even greater urgency to cut emissions. Separating out the various possible causes of the losses across England and Wales is complicated. A huge variety of land management factors may be involved, and their effects interact with climate change. It is therefore impractical to make measurements of all the possible combinations of factors. In any case statistical relations derived from such measurements do not reveal cause-and-effect, and they are poor at revealing the sorts of non-linear relations and feedback processes that are operating. Some form of mathematical modelling is required, based on understanding of the underlying processes. But current models are inadequate for this. Here we propose to improve the best-existing soil carbon model for our purposes / the Century model / by improving its components dealing with soil carbon transformations, their dependences on soil moisture, temperature and other variables, and the calculations of soil moisture, temperature and other driving variables. We will bring together (1) the new, internationally-unique dataset discussed above on soil carbon changes over time across the large range of soils, climates and land uses in England and Wales; (2) the best existing models of soil carbon transformations and of the dynamics of associated soil conditions, and new approaches to modelling these; (3) new spatial statistical techniques for analyzing the datasets to be modelled and the soundness of the model's predictions; (4) four specialists in soil modelling with different perspectives on these topics, and two specialists in soil spatial statistics, all internationally-regarded for their expertise in these areas.