Hydrology and Earth System Sciences (Oct 2022)

A method for predicting hydrogen and oxygen isotope distributions across a region's river network using reach-scale environmental attributes

  • B. D. Dudley,
  • J. Yang,
  • U. Shankar,
  • S. L. Graham

DOI
https://doi.org/10.5194/hess-26-4933-2022
Journal volume & issue
Vol. 26
pp. 4933 – 4951

Abstract

Read online

Stable isotope ratios (isotope values) of surface water reflect hydrological pathways, mixing processes, and atmospheric exchange within catchments. Development of maps of surface water isotope values (isoscapes) is limited by methods to interpolate point measures across river networks. Catchment attributes that alter surface water isotope values affect downstream river reaches via flow, but some attributes such as artificial dams are no more likely to affect nearby unconnected catchments than distant ones. Hence, simple distance-based geospatial and statistical interpolation methods used to develop isoscapes for precipitation and terrestrial systems are less appropriate for river networks. We used a water-balance-based method to map long-term average δ2H and δ18O for New Zealand rivers, incorporating corrections using catchment environmental predictors. Inputs to the model are national rainfall precipitation isoscapes, a digital elevation layer, a national river water isotope monitoring dataset (3 years of monthly sampling at 58 sites), and river environmental databases covering around 600 000 reaches and over 400 000 km of rivers. Much of the spatial variability in δ2H and δ18O of New Zealand river water was explained using the initial combination of precipitation isoscapes and a simple water balance model. δ2H and δ18O isoscapes produced by subsequently applying residuals from the water balance model as a correction factor across the river network using regression kriging showed improved fits to the validation data compared to the correction using ordinary kriging. Predictors of high importance in the regression included upstream lake and wetland area, which was not strongly spatially autocorrelated nationally. Hence, additional hydrological process information such as evaporation effects can be incorporated into river isoscapes using regression kriging of residuals. The resulting isoscapes have potential applications in ecological, hydrological, and provenance studies that consider differences between surface water isotope values and those of other components of the hydrological cycle (e.g. subsurface runoff or local precipitation).