Estimating mountain basin-mean precipitation from streamflow using Bayesian inference
Estimating mountain basin-mean precipitation from streamflow using Bayesian inference
复制标题
DOI：
10.1002/2014wr016736
复制
发表时间：
2015-10-01
影响因子：
5.4
通讯作者：
Lundquist, Jessica D.
中科院分区：
文献类型：
--
作者：
Henn, Brian;Clark, Martyn P.;Lundquist, Jessica D.
由于相对于流域的降水量计的地形代表性的不确定性，在复杂地形中估计流域平均降水量是困难的。为了解决这个问题，我们使用贝叶斯方法加上一个多模型框架来推断流域平均降水量的径流观测，我们将这种方法应用于雪为主的流域在内华达州的加州。使用径流观测，迫使低海拔站的数据，贝叶斯总误差分析（BATEA）的方法和理解结构误差（FUSE）的框架，我们推断流域平均降水量，并将其与流域平均降水量估计使用地形知情插值仪（PRISM，独立斜坡模型的参数-海拔回归）。BATEA推断的降水的空间格局与PRISM协议的排名从湿到干的流域，但不同的绝对值。在一些流域，这些差异可能反映了PRISM的偏差，因为一些隐含的PRISM径流率可能与区域气候不一致。我们还推断流域降水量的年时间序列使用两步校准方法。的精度和鲁棒性的BATEA方法的评估表明，在BATEA推断的降水的不确定性主要与水文模型结构的不确定性。尽管有这些局限性，在不同的模型和参数假设下推断的年降水量的时间序列是强烈相关的彼此，这表明这种方法是能够解决流域平均降水量的年际变化。
Estimating basin-mean precipitation in complex terrain is difficult due to uncertainty in the topographical representativeness of precipitation gauges relative to the basin. To address this issue, we use Bayesian methodology coupled with a multimodel framework to infer basin-mean precipitation from streamflow observations, and we apply this approach to snow-dominated basins in the Sierra Nevada of California. Using streamflow observations, forcing data from lower-elevation stations, the Bayesian Total Error Analysis (BATEA) methodology and the Framework for Understanding Structural Errors (FUSE), we infer basin-mean precipitation, and compare it to basin-mean precipitation estimated using topographically informed interpolation from gauges (PRISM, the Parameter-elevation Regression on Independent Slopes Model). The BATEA-inferred spatial patterns of precipitation show agreement with PRISM in terms of the rank of basins from wet to dry but differ in absolute values. In some of the basins, these differences may reflect biases in PRISM, because some implied PRISM runoff ratios may be inconsistent with the regional climate. We also infer annual time series of basin precipitation using a two-step calibration approach. Assessment of the precision and robustness of the BATEA approach suggests that uncertainty in the BATEA-inferred precipitation is primarily related to uncertainties in hydrologic model structure. Despite these limitations, time series of inferred annual precipitation under different model and parameter assumptions are strongly correlated with one another, suggesting that this approach is capable of resolving year-to-year variability in basin-mean precipitation.