The Amazon hydrological cycle: past, present and future

南美洲热带地区作为主要的对流中心之一，是热带(Walker环流)的重要组成部分，也是全球大气环流(Hadley Cell)和水文循环的重要组成部分。例如，亚马逊河就说明了这一点，它向海洋排放了大约17%的淡水。热带南美洲还拥有世界上最大的森林区域之一，这是一个巨大的碳库，可以通过森林破坏或干旱反馈等方式快速释放到大气中。现有的水文循环记录、奥比多斯亚马逊河流量(整合了亚马逊集水区77%的面积)以及降水气候学记录显示，亚马逊盆地的水文循环在过去大约20年里呈现出显著增强的趋势。这一增加主要发生在雨季，导致河流流量的季节性幅度增加。降水量日最大值有更强的趋势，日最小降水量有减少的趋势，表明降雨和干旱事件也在加剧。热带南美洲在2005年和2010年确实经历了严重的干旱，最近的一次是在2009年和2012年。由于干旱和洪水对生计造成了重大破坏，从全球气候变化的角度来看，了解亚马逊水文循环的变化非常重要。然而，我们对亚马逊水文循环正在发生的变化的了解很差。主要原因是控件较多，受数据约束较差。它们包括外部因素，如通过热带大西洋主要气流输入的水蒸气或热带辐合带位置的变化，以及内部因素，如热带雨林功能的变化，特别是通过森林将水重新循环回大气。这项建议的目的是将新的和现有的数据与补充的建模和归类技术相结合，以了解亚马逊水文循环的当前和过去趋势，以帮助预测未来会发生什么。我们提出的工作建立在我们研究的两个最新结果的基础上。首先，我们发现香柏树种表现出非常清晰的年轮，树轮纤维素中的氧同位素18O与亚马逊的大规模水文循环密切相关。具体地说，在玻利维亚亚马逊网站的八棵树上记录的18O与奥比多斯的亚马逊河流流量之间存在着强烈的相关性。其次，我们最近成功地利用大气气团轨迹和植被类型遥感数据来估计植被对空气中水汽的贡献，从而估计降水的再循环。因此，我们建议对玻利维亚的18O降水记录进行补充，使之在整个盆地的另外五个地点得到良好的时空复盖，涵盖过去两个世纪的降水18O和间接降水。其次，我们提出了一种模式分析，既使用了带有同位素的气候植被模式来检查一系列过程及其对降水和降水中18O及其时间趋势的影响，又同时使用了一种反向轨迹方法来将观测到的同位素特征与气团轨迹联系起来，以估计盆地水循环的变化。通过我们的方法，我们希望能够查明过去20年亚马逊水文循环加剧的原因，以及这些原因在多大程度上是由于陆地植被功能的变化，从而能够预测未来几十年的情况。我们还希望能够找出树木年轮降水180度长达一个世纪的趋势的原因，以及它们告诉我们什么会导致系统的长期变化。

As one of the major centers of convection, tropical South America is an important component of the tropical (Walker Circulation) and also the global atmospheric circulation (Hadley Cell), and hydrological cycle. This is for example illustrated by the Amazon river who discharges approximately 17 % of all freshwater to the oceans. Tropical South America also hosts one of the largest forested areas in the world which is a huge carbon store amenable to fast release to the atmosphere, e.g. by forest destruction or drought-induced feed-backs.Existing records of the hydrological cycle, the Amazon river discharge at Obidos, integrating 77% of the Amazon catchment area, and precipitation climatologies, reveal that the Amazon basin hydrological cycle exhibits a substantial intensifying trend over approximately the last two decades. The increase occurs mainly during the rainy season leading to an increase in the seasonal amplitude of river discharge. There is an even stronger trend in the daily maxima in precipitation and a decreasing trend of minimum daily precipitation pointing to an intensification of rain and drought events as well. Tropical South America has indeed witnessed severe droughts in 2005 and 2010 as well as strong flooding, most recently in 2009 and 2012.Both because of substantial damage to livelihood by droughts and floods as well as from the perspective of global climate change, understanding changes in the Amazon hydrological cycle is important. However our understanding of ongoing changes of the Amazon's hydrological cycle is poor. The main reason is that there are many controls, which are poorly constrained by data. They include external factors, like water vapor input via the main air stream from the tropical Atlantic or changes of the location of the inter-tropical convergence zone, as well as internal factors like changes in rainforest functioning specifically the recirculation of water back to the atmosphere via forests. The purpose of this proposal is to combine novel and existing data, with complimentary modelling and attribution techniques to understand ongoing and past trends of the Amazon hydrological cycle in order to help predict what to expect in the future. Our proposed work builds on two recent results from our research. First we have discovered that the tree species Cedrela odorata exhibits very clear annual rings and that the oxygen isotope 18O in tree ring cellulose is closely linked to the large-scale hydrological cycle of the Amazon. Specifically there is a strong correlation between 18O recorded in eight trees at a Bolivian Amazon site and Amazon river discharge at Obidos. Secondly we have recently succeeded to use atmospheric air parcel trajectory and remote sensing data of vegetation type to estimate the contribution of vegetation to water vapor in the air and thus recirculation of precipitation. We therefore propose to complement the Bolivian 18O precipitation record to further five sites across the basin to produce a good spatio-temporal coverage of precipitation 18O and indirectly precipitation over the last two centuries. Secondly we propose a modelling analysis employing both a climate vegetation model with isotopes to examine a range of processes and their effect on precipitation and 18O in precipitation and their time trends, and in parallel a back-trajectory approach to link observed isotope signatures along air parcel trajectories to estimate changes in water recycling in the basin. With our approach we expect to be able to pinpoint the causes of the intensification of the Amazon hydrological cycle over the last two decades and to what extent they are due to changes in functioning of the land vegetation and therefore to permit predictions of what to expect over the next decades. We also expect to be able to pinpoint the causes of the century long trends in tree ring based precipitation 18O and what they tell us about what causes longer term changes of the system.