Premium
Temporal dynamics of catchment transit times from stable isotope data
Author(s) -
Klaus Julian,
Chun Kwok P.,
McGuire Kevin J.,
McDonnell Jeffrey J.
Publication year - 2015
Publication title -
water resources research
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.863
H-Index - 217
eISSN - 1944-7973
pISSN - 0043-1397
DOI - 10.1002/2014wr016247
Subject(s) - tracer , surface runoff , watershed , flow (mathematics) , mixing (physics) , hydrology (agriculture) , environmental science , transit time , series (stratigraphy) , boundary (topology) , drainage basin , geology , mathematics , computer science , geography , mechanics , physics , engineering , ecology , paleontology , geotechnical engineering , quantum mechanics , machine learning , nuclear physics , transport engineering , biology , mathematical analysis , cartography
Time variant catchment transit time distributions are fundamental descriptors of catchment function but yet not fully understood, characterized, and modeled. Here we present a new approach for use with standard runoff and tracer data sets that is based on tracking of tracer and age information and time variant catchment mixing. Our new approach is able to deal with nonstationarity of flow paths and catchment mixing, and an irregular shape of the transit time distribution. The approach extracts information on catchment mixing from the stable isotope time series instead of prior assumptions of mixing or the shape of transit time distribution. We first demonstrate proof of concept of the approach with artificial data; the Nash‐Sutcliffe efficiencies in tracer and instantaneous transit times were >0.9. The model provides very accurate estimates of time variant transit times when the boundary conditions and fluxes are fully known. We then tested the model with real rainfall‐runoff flow and isotope tracer time series from the H.J. Andrews Watershed 10 (WS10) in Oregon. Model efficiencies were 0.37 for the 18 O modeling for a 2 year time series; the efficiencies increased to 0.86 for the second year underlying the need of long time tracer time series with a long overlap of tracer input and output. The approach was able to determine time variant transit time of WS10 with field data and showed how it follows the storage dynamics and related changes in flow paths where wet periods with high flows resulted in clearly shorter transit times compared to dry low flow periods.