Spatial patterns of throughfall isotopic composition at the event and seasonal timescales
Spatial patterns of throughfall isotopic composition at the event and seasonal timescales
复制标题
DOI：
10.1016/j.jhydrol.2014.12.029
复制
发表时间：
2015-03
影响因子：
6.4
通讯作者：
S. Allen;R. Keim;J. McDonnell
中科院分区：
文献类型：
--
作者：
S. Allen;R. Keim;J. McDonnell
森林中穿透雨同位素组成的空间变异性表明冠层中发生的复杂过程，并且仍然没有得到充分的理解，以正确描述降水对集水区水平衡的影响。在这里，我们研究穿透雨同位素组成的变异性，目的是：（1）量化事件规模样本的空间变异性，（2）确定是否存在对变异性的持续控制以及这些变化如何影响季节性累积穿透雨的变异性，以及（3）分析与不同采样方案相关的测量穿透雨同位素组成的分布。我们在美国西部俄勒冈州的花旗松树冠下测量了两三个月的穿透降水。到2009年秋季（11次事件）和2010年春季（7次事件），每次事件的δ 18 O平均空间范围分别为1.6‰和1.2‰。然而，同位素组成的空间格局在时间上并不稳定，导致季节总穿透降水的变化小于事件穿透降水（1.0‰; 2009年秋季累积δ 18 O的范围）。同位素组成没有空间自相关，不能解释的位置相对于树干。对代表不同抽样方案的野外测量数据和蒙特-卡罗模拟数据的抽样误差分析表明，与真实平均值的差异的标准差高达0.45‰（δ 18 O）和1.29‰（d过量）。这种同位素变化的幅度表明，小样本量是大量实验误差的来源。
Spatial variability of throughfall isotopic composition in forests is indicative of complex processes occurring in the canopy and remains insufficiently understood to properly characterize precipitation inputs to the catchment water balance. Here we investigate variability of throughfall isotopic composition with the objectives: (1) to quantify the spatial variability in event-scale samples, (2) to determine if there are persistent controls over the variability and how these affect variability of seasonally accumulated throughfall, and (3) to analyze the distribution of measured throughfall isotopic composition associated with varying sampling regimes. We measured throughfall over two, three-month periods in western Oregon, USA under a Douglas-fir canopy. The mean spatial range of δ18O for each event was 1.6‰ and 1.2‰ through Fall 2009 (11 events) and Spring 2010 (7 events), respectively. However, the spatial pattern of isotopic composition was not temporally stable causing season-total throughfall to be less variable than event throughfall (1.0‰; range of cumulative δ18O for Fall 2009). Isotopic composition was not spatially autocorrelated and not explained by location relative to tree stems. Sampling error analysis for both field measurements and Monte-Carlo simulated datasets representing different sampling schemes revealed the standard deviation of differences from the true mean as high as 0.45‰ (δ18O) and 1.29‰ (d-excess). The magnitude of this isotopic variation suggests that small sample sizes are a source of substantial experimental error.