Top-down rehydration of the mangrove Avicennia marina in response to wetting events caused by deliquescence of salt, accumulation of dew, and interception of rainfall under an arid climate DP150104437

Absorption of atmospheric water by branches of Avicennia marina was assessed by field-based studies of sap flow velocity and shoot water relations under dry season conditions along Giralia Bay in arid Western Australia. Two hypotheses were tested: 1) that deliquescence of salt secreted onto leaf surfaces can drive top-down rehydration, and 2) that absorption of atmospheric water from wetting events makes a functional contribution to shoot water balances, subsidising transpiration and reducing stem vulnerability to hydraulic failure. Study design: Three similar trees of the mangrove Avicennia marina (Forssk.) Vierh subsp. marina were studied under field conditions. Four branches of similar size and orientation were selected on each tree and randomly assigned to four treatments: Natural (N), Wet (W), Dry (D) and Ambient (A). Trees are identified by number and branches are identified by the letter denoting the treatment received. For example, an ID of 2D indicates Tree 2, branch allocated to Dry treatment. The Wet and Dry treatments were imposed only during a brief nocturnal manipulation experiment and were otherwise subject to the natural variation in weather. Ambient and Natural treatments were only subject to natural variation in weather but branches in the Natural treatment remained untouched throughout the whole monitoring period. The branches were used in two experiments in which sap flow characteristics were analysed in relation to either natural or manipulated conditions. Measurements of sap flow velocity: Detection of top-down rehydration events was based on reversal of sap flow (ie negative sap flow), indicating downward movement of water from branches to the main stem. A sap flow meter (SFM1, ICT International Pty Ltd, Armidale, Australia) was installed (sensor depth 12.2 mm) in the leafless region of each branch near its junction with the main stem, approximately 70-80 cm from the branch canopy. Branch level sap flow was monitored at 20 min intervals using the heat ratio method (Burgess et al., 2001) together with micro-meteorological measurements to enable analyses of directional water fluxes in relation to air humidity, air temperature and the occurrence of wetting events due to deliquescence, dew or precipitation. Environmental data were recorded at 10 minute intervals throughout the study by a portable weather station (Kestrel 3500 Delta T Meter, Neilsen-Kellerman Co, Boothwyn, PA). Air vapour pressure deficit (VPD) was calculated according to Murray (1967): VPD = Pv � ((RH/100)*Pv), where air temperature (Ta) and relative humidity (RH) were recorded and vapour pressure (Pv) was calculated as Pv = 0.611 exp [17.27 Ta/(Ta + 237.3)]. Branches were harvested when the study concluded, and zero flow baselines were determined for each meter/branch combination as recommended by the manufacturer. All leaves were harvested from each branch and imaged for determination of leaf area using J IMAGEJ software (Schneider et al., 2012). Total leaf dry mass was measured with a field balance (XP 205 Metter Toledo balance, Mettler � Toledo Ltd., Greifensee, Switzerland) after oven drying at 80�C for 72 h. Monitoring sap flow under natural weather conditions: Data sets 1 through 6 Environmental conditions and sap flow velocities wererecorded simultaneously under natural conditions for seven days. During this period, branches were subject to a natural sequence of dry and wet conditions. The whole data set is summarized in Data Set 2 in which reverse sap flow was detected under conditions conducive to leaf wetting by deliquescence of salt, dew formation and interception of intermittent rainfall. Observations of two natural events in which deliquescence of salt on leaf surfaces drove reverse sap flow are highlighted in two data sets. Data Set 3 is a subset of Data Set 2 and contains sap flow data from the three branches in which reverse sap flow was detected under deliquescent conditions in the absence of dew. Data Set 4 is also a subset of Data Set 2 and contains additional data collected when a sudden change in the weather provided an opportunity to study leaf rehydration during a natural foliar wetting event driven by salt deliquescence on leaf surfaces. Once water began to accumulate on leaf surfaces in late afternoon, leaves were harvested from each of the three replicate trees at each time point for measurement of water content and water potential (ΨLeaf) from branches adjacent to those monitoring sap flow. Leaf fresh mass was measured before determining leaf water potential ΨLeaf with a Scholander pressure chamber (1050D, PMS Instruments, Albany, USA), followed by measurements of leaf area and dry mass. Measurements continued until light became too low to work (18:20 pm). Then, one twig with two fully expanded leaves was collected from each of the three trees to measure water uptake by detached leaves. Each twig was incubated in an individual zip-lock plastic bag with moist paper towels for 60 min under darkness at ambient temperature. Then fresh leaf mass and ΨLeaf were determined for one leaf in each pair, followed by determinations of leaf area and dry mass. These data were used to calculate rates of rehydration and water uptake by leaves on detached twigs by assuming that their starting points were the same as those measured at the time of twig collection (18:20). All sap flow data collected under natural or ambient conditions were then pooled to analyse patterns in sap flow velocities in relation to the vapour pressure deficit of the air (Data Set 5) and relative humidity of the air (Data Set 6). Monitoring sap flow under manipulated conditions: data sets 1, 7 and 8. Conditions were manipulated during one night to assess effects of nocturnal Dry, Wet and Ambient treatments on branch sap flow velocity (Data Set 7), and water relations of leaves and twigs together with branch sap flow velocity under common ambient conditions during the following photoperiod (Data Set 8). In late afternoon, six twigs of similar size and orientation were selected on each of the three treatment branches on each tree and allocated to one of six sampling times for measurement of diurnal variation in ΨLeaf. The two youngest, fully expanded leaves on each twig were allocated to either transpiring or non-transpiring treatments. The latter treatment was tightly covered with plastic cling wrap and an outer layer of aluminium foil. Branch treatments were then imposed just before sunset. Wet branches were doused with freshwater to fully wet all leaf and branch surfaces. Six soaking wet sponges each containing approximately 100 g of water were distributed near the designated twigs. Then each branch was sealed inside a heavy gauge black garbage bag. Dry branches each received six envelopes made of 30% shade cloth containing 150 g of dried silica gel. The envelopes were secured near the twigs, and the branches were then covered with a heavy gauge black garbage bag which was then purged with several volumes of dry, compressed air from a scuba tank before the garbage bag was sealed. Ambient branches were left exposed to natural conditions. Each branch was fitted with an i-button (Hygrochron DS1923, Whitewater, Wisconsin) placed in a sheltered canopy position to monitor humidity and temperature during nocturnal treatments. Branches were liberated from Wet and Dry treatments at sunrise (7:30) by removing the garbage bags, sponges and envelopes containing silica gel. Water potentials of paired transpiring and non-transpiring leaves from the designated twigs were measured at six sampling times arrayed at two hour intervals from 30 minutes after sunrise, 8:00 to 18:00 to assess effects of nocturnal treatments on branch water status under common conditions in the subsequent photoperiod. At each sampling time, the designated leaves were collected and water potential measurements were made within minutes at a field lab established approximately 20 m from the study trees. Measurements of leaf fresh mass were followed sequentially by measurements of ΨLeaf, area and dry mass.

本研究针对澳大利亚西部干旱地区吉拉利亚湾（Giralia Bay）旱季条件下的白骨壤（Avicennia marina）枝条液流速率（sap flow velocity）与枝条水分关系（shoot water relations）开展野外研究，以评估其吸收大气水分的能力。本研究检验了两项假说：1）叶片表面分泌的盐分潮解可驱动自上而下补水（top-down rehydration）过程；2）湿润事件中吸收的大气水分可对枝条水分平衡起到功能性贡献，补偿蒸腾作用并降低茎部水力失效风险。

研究设计：选取3株生长状况一致的红树林白骨壤（Avicennia marina (Forssk.) Vierh subsp. marina）野外个体。每株树选取4个规格与朝向一致的枝条，随机分为4组处理：自然组（Natural, N）、湿润组（Wet, W）、干燥组（Dry, D）与对照组（Ambient, A）。植株以编号标识，枝条以其所属处理的字母标识，例如编号2D代表2号植株接受干燥处理的枝条。

湿润组与干燥组仅在夜间短期操控实验中施加处理，其余时间受自然天气波动影响；对照组与自然组仅受自然天气波动影响，但自然组枝条在整个监测周期内均不做任何干预。本研究通过两组实验分析枝条液流特征与自然或操控环境的关联。

液流速率测量：自上而下补水事件的检测基于液流逆转（即负液流），该信号指示水分从枝条向主茎的向下移动。采用液流计（SFM1，ICT国际私人有限公司，阿米代尔，澳大利亚），将传感器安装于枝条与主茎连接处的无叶区域，距离枝条冠层约70~80 cm，传感器插入深度为12.2 mm。采用热比率法（heat ratio method，Burgess等，2001）以20分钟间隔监测枝条水平的液流，同时结合微气象测量，以分析定向水分通量与空气湿度、气温以及由盐分潮解、露水或降水引发的湿润事件的关联。

环境数据由便携式气象站（Kestrel 3500 Delta T Meter，尼尔森-凯勒曼公司，布斯敦，宾夕法尼亚州）以10分钟间隔全程记录。空气水汽压亏缺（vapour pressure deficit, VPD）根据Murray（1967）的方法计算：VPD = Pv – ((RH/100)*Pv)，其中记录气温（Ta）与相对湿度（RH），水汽压（Pv）计算公式为Pv = 0.611 exp [17.27 Ta/(Ta + 237.3)]。

研究结束后采收枝条，按照制造商建议为每个液流计-枝条组合确定零流量基线。采收每个枝条的全部叶片，采用ImageJ软件（Schneider等，2012）成像以测定叶面积。将叶片在80℃烘箱中烘干72小时后，使用梅特勒-托利多XP 205电子分析天平（Mettler Toledo Ltd., 格赖夫斯瓦尔德，瑞士）测定总叶干重。

自然天气条件下的液流监测：数据集1至6。在自然条件下同步记录环境数据与液流速率，为期7天。此期间枝条经历自然的干湿交替序列。全部数据集汇总于数据集2，其中在盐分潮解、露水形成与间歇性降雨截留导致叶片湿润的条件下，检测到了液流逆转现象。有两次自然事件的观测结果被重点标注，即叶片表面盐分潮解驱动液流逆转，分别对应两个数据集。数据集3为数据集2的子集，包含3株枝条的液流数据，这些枝条在无露水的盐分潮解条件下检测到了液流逆转。数据集4同样为数据集2的子集，包含当天气突然变化时收集的额外数据，此时可研究由叶片表面盐分潮解引发的自然叶片湿润事件期间的叶片补水过程。

当日晚些时候叶片表面开始积水时，在每个时间点从3株重复植株的邻近监测枝条上采收叶片，以测定其含水量与叶水势（ΨLeaf）。先测定叶片鲜重，随后使用Scholander压力室（Scholander pressure chamber，1050D，PMS仪器公司，奥尔巴尼，美国）测定叶水势ΨLeaf，再依次测定叶面积与干重。监测持续至光线过暗无法开展操作为止（18:20）。随后从每株树采集1根带有2片完全展开叶片的小枝，以测定离体叶片的吸水速率。将每根小枝置于带湿纸巾的密封保鲜袋中，在室温黑暗条件下孵育60分钟。随后测定每对叶片中1片的鲜重与ΨLeaf，再依次测定叶面积与干重。基于小枝采集时（18:20）测定的初始参数，计算离体小枝上叶片的补水速率与吸水速率。

随后将所有自然或对照组条件下收集的液流数据合并，分析液流速率与空气水汽压亏缺（数据集5）以及空气相对湿度（数据集6）的关联模式。

操控条件下的液流监测：数据集1、7与8。在1个夜间开展环境操控，以评估夜间干燥、湿润与对照组处理对枝条液流速率的影响（数据集7），以及后续光照周期内共同环境条件下叶片与小枝的水分关系及枝条液流速率（数据集8）。

当日晚些时候，在每株树的3个处理枝条上各选取6个规格与朝向一致的小枝，分配至6个采样时间点，以测定ΨLeaf的日变化。每个小枝上的2片最幼嫩的完全展开叶片分别分配至蒸腾组与非蒸腾组：非蒸腾组用塑料保鲜膜紧密包裹，再覆盖一层铝箔。随后在日落前施加枝条处理。湿润组枝条用淡水喷淋，以完全湿润所有叶片与枝条表面。在指定小枝附近放置6块浸透水的海绵，每块海绵约含100 g水，随后将每个枝条密封于重型黑色垃圾袋中。干燥组枝条各放置6个由30%遮光布制成的信封，每个信封含150 g干燥硅胶，将信封固定于小枝附近，随后用重型黑色垃圾袋覆盖枝条，在密封垃圾袋前用潜水气瓶的干燥压缩空气吹扫数次。对照组枝条保持自然暴露状态。为每个枝条安装i-button（Hygrochron DS1923，怀特沃特，威斯康星州），置于冠层遮蔽位置，以监测夜间处理期间的湿度与温度。

日出（7:30）时移除垃圾袋、海绵与硅胶信封，解除湿润组与干燥组枝条的处理。从日出后30分钟（8:00）至18:00，以每2小时1次的间隔设置6个采样时间点，测定指定小枝上成对的蒸腾组与非蒸腾组叶片的水势，以评估夜间处理对后续光照周期内共同环境条件下枝条水分状态的影响。每个采样时间点采集指定叶片，在距离研究植株约20 m的野外实验室中于数分钟内完成水势测定。依次测定叶片鲜重、ΨLeaf、叶面积与干重。