Alternative stable states between subtropical forest and grassland: A palaeoecological case study in southern Brazil

森林和草原是截然不同的陆地生态系统，它们提供不同的生态系统服务。频繁的火灾维持着草原，森林扑灭了火灾。因此，这两个系统都是自我稳定的，可能发生在相同的气候条件下。热带和亚热带的大片区域在气候上都适合这两个生态系统，州与州之间的变化可能会对区域乃至全球气候产生影响。了解这种双稳定性的决定因素最近受到了很大的关注，但主要是基于现代观测和模型模拟。古生态技术为研究更长时期内状态变化的动力学提供了机会，从而有助于正在进行的讨论。巴西南部的自然植被代表了潮湿的森林和草原的马赛克，在过去的4000年里，森林一直延伸到草原。这种情况为案例研究提供了一个理想的环境，该案例研究揭示了引爆点，确定了系统弹性的因素，以及这种状态转变的动态。有许多花粉图描述了这种变化，但它们来自早、中期全新世沉积速率较低的小泥炭沼泽，无法对这一过程进行详细分析。该方案旨在对一个小湖泊10m全新世沉积岩芯进行高分辨率的火灾和植被动力学多代理调查，以此为基础与动态植被模型的模拟进行比较。该区域内不同地点森林扩展的时间为了解这种植被变化的驱动因素和触发因素提供了更多信息，该项目的目的是改善已经调查的地点的这种变化的年代，并提供更多的地方调查。草原和森林物种在花粉产生和传播方面存在差异，为了减少这种偏见并有助于解释，将进行一项调查，以解决花粉与植被的关系。这将使人们能够量化估计草原的面积及其被森林取代的程度。综合信息将被用来检验三个相互矛盾的假设：a)全新世降水量随着南美季风的加强而逐渐增加，而森林扩张受到频繁火灾的限制，只有在超过阈值时才会发生；b)降水制度的逐步变化引发了森林的迅速扩张；人类协助了阿拉卡里亚森林的传播，因为种子具有很高的营养价值。一个机械的植被模型将被用来检验假设，通过强迫它在不同的气候情景下进行，并将模拟与定量的植被重建进行比较。将设计进一步的数据模型实验，以检验目前关于草原和森林之间替代稳定状态的理论。

Forest and grassland are vastly contrasting terrestrial ecosystems that provide different ecosystem services. While frequent fires maintain grasslands, fires are suppressed by forests. Thus both systems are self-stabilizing and may occur under the same climate conditions. Large regions of the tropics and sub-tropics are climatically suitable for both ecosystems, and shifts between states may have consequences for regional and possibly global climate. Understanding the determinants of this bi-stability has recently received much attention, but has been mainly based on modern observations and model simulations. Palaeoecological techniques provide the opportunity to study the dynamics of state changes over longer time periods and thus contribute to the ongoing discussion. The natural vegetation in southern Brazil represents a mosaic of humid forests and grasslands, with forests expanding into the grasslands over the last 4000 years. This situation provides an ideal setting for a case study uncovering tipping points, determining factors of system resilience, as well as the dynamics of such a state shift. A number of pollen diagrams exist that describe this change, but are from small peat bogs with low sedimentation rate during the early and mid-Holocene, precluding detailed analysis of the process. The proposal aims to carry out a high resolution multi-proxy investigation on fire and vegetation dynamics from a 10 m Holocene sediment core of a small lake as a base for comparison with simulations from dynamic vegetation models. The timing of the forest expansion at different localities within the region provides additional information to understand the drivers and triggers of this vegetation shift, and the project aims to improve the dating of this change for already investigated sites and to provide additional local investigations. Grassland and forest species differ in pollen production and dispersal, and to reduce this bias and help in interpretation, an investigation will address the pollen-vegetation relationship. This will allow quantified estimates of the extent of grassland and its replacement by forest. The combined information will be used to test three competing hypotheses: a) Holocene precipitation increased gradually with a strengthened South American Monsoon, while forest expansion was limited by frequent fires and only occurred once a threshold was passed; b) a stepwise change in the precipitation regime triggered the rapid forest expansion; humans assisted in the spread of the Araucaria forest as the seeds have high nutritious value. A mechanistic vegetation model will be used for hypotheses testing, by forcing it with different climate scenarios and comparing the simulations against the quantitative vegetation reconstructions. Further data model experiments will be designed to test the current theory on alternative stable states between grasslands and forests.