Spatiotemporal patterns of precipitation inferred from streamflow observations across the Sierra Nevada mountain range

JOURNAL OF HYDROLOGY

卷:556

Article

Henn, Brian^1,8  Clark, Martyn P.²  Kavetski, Dmitri³  Newman, Andrew J.²  Hughes, Mimi^4,5  McGurk, Bruce⁶  Lundquist, Jessica D.⁷ 

[1] Univ Washington, Dept Civil & Environm Engn, 201 More Hall,Box 352700, Seattle, WA 98195 USA

[2] Natl Ctr Atmospher Res, Res Applicat Lab, POB 3000, Boulder, CO 80307 USA

[3] Univ Adelaide, Sch Civil Environm & Min Engn, Adelaide, SA, Australia

[4] NOAA, Earth Syst Res Lab, Phys Sci Div, Boulder, CO USA

[5] Univ Colorado, Cooperat Inst Res Environm Sci, Boulder, CO 80309 USA

[6] McGurk Hydrol, Orinda, CA USA

[7] Univ Washington, Dept Civil & Environm Engn, Seattle, WA 98195 USA

[8] Univ Calif San Diego, Scripps Inst Oceanog, Ctr Western Weather & Water Extremes, 9500 Gilman Dr, La Jolla, CA 92093 USA

关键词: Precipitation;    Orographic enhancement;    Streamflow;    Mountain hydrology;    Sierra Nevada;    Bayesian inference;   

DOI  :  10.1016/j.jhydrol.2016.08.009

来源: Elsevier

【 摘 要 】

Given uncertainty in precipitation gauge-based gridded datasets over complex terrain, we use multiple streamflow observations as an additional source of information about precipitation, in order to identify spatial and temporal differences between a gridded precipitation dataset and precipitation inferred from streamflow. We test whether gridded datasets capture across-crest and regional spatial patterns of variability, as well as year-to-year variability and trends in precipitation, in comparison to precipitation inferred from streamflow. We use a Bayesian model calibration routine with multiple lumped hydrologic model structures to infer the most likely basin-mean, water-year total precipitation for 56 basins with long-term (>30 year) streamflow records in the Sierra Nevada mountain range of California. We compare basin-mean precipitation derived from this approach with basin-mean precipitation from a precipitation gauge-based,1 /16 degrees gridded dataset that has been used to simulate and evaluate trends in Western United States streamflow and snowpack over the 20th century. We find that the long-term average spatial patterns differ: in particular, there is less precipitation in the gridded dataset in higher-elevation basins whose aspect faces prevailing cool-season winds, as compared to precipitation inferred from streamflow. In a few years and basins, there is less gridded precipitation than there is observed streamflow. Lower elevation, southern, and east-of-crest basins show better agreement between gridded and inferred precipitation. Implied actual evapotranspiration (calculated as precipitation minus streamflow) then also varies between the streamflow-based estimates and the gridded dataset. Absolute uncertainty in precipitation inferred from streamflow is substantial, but the signal of basin-to-basin and year-to-year differences are likely more robust. The findings suggest that considering streamflow when spatially distributing precipitation in complex terrain may improve its representation, particularly for basins whose orientations (e.g., windward-facing) are favored for orographic precipitation enhancement. (C) 2016 Elsevier B.V. All rights reserved.

【 授权许可】

Free