Dwivedi et al., 2018

Talk/Poster

An improved and practical approach for estimating catchment-scale response functions through power spectral analysis

Dwivedi, R., Meixner, T., Mcintosh, J.C., Ferre, T.P.A., Eastoe, C.J., Castro, C.L., Wright, W.E., Niu, G.-Y., Minor, R.L., Knowles, J.F., Barron-Gafford, G., Abramson, N., Mitra, B., Stanley, M., Chorover, J. (2018)
Abstract H13J-1876 presented at 2018 AGU Fall Meeting, Washington, D.C., 10-14 Dec  

Abstract

While catchment-scale response functions, such as transit time distribution (TTD) and evapotranspiration time distribution (ETTD) are considered fundamental descriptors of a catchment’s hydrologic and ecohydrologic behavior, their estimation is challenging, as existing methods require gapless time series data. Using Kirchner and Neal (2013)’s weighted wavelet spectral analysis method for a conservative tracer, we propose improved and practical methods for estimating TTDs and ETTDs, which are resistant to gaps in naturally observed tracer time series data. Using a multiple tracers (δ18O and 3H), weighted wavelet spectral analysis, and the convolution integral approach for 3H from Stewart et al. [2016], we estimate TTDs for shallow and deep subsurface water reservoirs that contribute to streamflow in a high-elevation, headwater catchment in Arizona, USA. Through testing a number of TTDs, we found a gamma type was most appropriate for streamflow using either δ18O or 3H tracers. For ETTD estimation, we first identified source waters for vegetation and streamflow, using stable water isotopes and found they were the same, which negates the ecohydrologic water source separation hypothesis. Whereas the ETTD could be approximated as piston flow (i.e. vegetation uses recent precipitation, if available) at lower periods (up to 0.15 years), a gamma ETTD type was closest to the ETTD observations at higher periods (at least up to 5 years). We also show that the fraction of young water, which is recommended in recent literature for better understanding catchment storage characteristics and their relationship to streamflow, curtails the TTD, and results in an incomplete description of catchment hydrologic behavior. As a result, we suggest deriving the TTDs for multiple tracers using methods such as weighted wavelet analysis and convolution integrals in a time-averaged sense, to allow for a more complete understanding of catchment hydrogeochemical behavior.

 

Citation

Dwivedi, R., Meixner, T., Mcintosh, J.C., Ferre, T.P.A., Eastoe, C.J., Castro, C.L., Wright, W.E., Niu, G.-Y., Minor, R.L., Knowles, J.F., Barron-Gafford, G., Abramson, N., Mitra, B., Stanley, M., Chorover, J. (2018): An improved and practical approach for estimating catchment-scale response functions through power spectral analysis. Abstract H13J-1876 presented at 2018 AGU Fall Meeting, Washington, D.C., 10-14 Dec.