Global spatial sensitivity of runoff to subsurface permeability using the active subspace method

Abstract Hillslope scale runoff is generated as a result of interacting factors that include water influx rate, surface and subsurface properties, and antecedent saturation. Heterogeneity of these factors affects the existence and characteristics of runoff. This heterogeneity becomes an increasingly relevant consideration as hydrologic models are extended and employed to capture greater detail in runoff generating processes. We investigate the impact of one type of heterogeneity – subsurface permeability – on runoff using the integrated hydrologic model ParFlow. Specifically, we examine the sensitivity of runoff to variation in three-dimensional subsurface permeability fields for scenarios dominated by either Hortonian or Dunnian runoff mechanisms. Ten thousand statistically consistent subsurface permeability fields are parameterized using a truncated Karhunen–Loeve (KL) series and used as inputs to 48-h simulations of integrated surface-subsurface flow in an idealized ‘tilted-v’ domain. Coefficients of the spatial modes of the KL permeability fields provide the parameter space for analysis using the active subspace method. The analysis shows that for Dunnian-dominated runoff conditions the cumulative runoff volume is sensitive primarily to the first spatial mode, corresponding to permeability values in the center of the three-dimensional model domain. In the Hortonian case, runoff volume is sensitive to multiple smaller-scale spatial modes and the locus of that sensitivity is in the near-surface zone upslope from the domain outlet. Variation in runoff volume resulting from random heterogeneity configurations can be expressed as an approximately univariate function of the active variable, a weighted combination of spatial parameterization coefficients computed through the active subspace method. However, this relationship between the active variable and runoff volume is more well-defined for Dunnian runoff than for the Hortonian scenario.

[1]  A. Mantoglou,et al.  The Turning Bands Method for simulation of random fields using line generation by a spectral method , 1982 .

[2]  Reed M. Maxwell,et al.  Quantifying the effects of three-dimensional subsurface heterogeneity on Hortonian runoff processes using a coupled numerical, stochastic approach , 2008 .

[3]  Reed M. Maxwell,et al.  Quantifying the effects of subsurface heterogeneity on hillslope runoff using a stochastic approach , 2011 .

[4]  R. Maxwell,et al.  A comparison of two physics-based numerical models for simulating surface water–groundwater interactions , 2010 .

[5]  Narendra N. Das,et al.  Characterization of effective saturated hydraulic conductivity in an agricultural field using Karhunen‐Loève expansion with the Markov chain Monte Carlo technique , 2010 .

[6]  R. H. Moore,et al.  Regression Graphics: Ideas for Studying Regressions Through Graphics , 1998, Technometrics.

[7]  Wendy D. Graham,et al.  Improved hydrograph prediction through subsurface characterization: conditional stochastic hillslope simulations , 2014, Hydrogeology Journal.

[8]  J. Jaimegomezhernandez,et al.  ISIM3D: An ANSI-C three-dimensional multiple indicator conditional simulation program , 1990 .

[9]  D. A. Woolhiser,et al.  Effects of Spatial Variability of Saturated Hydraulic Conductivity on Hortonian Overland Flow , 1996 .

[10]  R. D. Black,et al.  An Experimental Investigation of Runoff Production in Permeable Soils , 1970 .

[11]  R. Horton.  The Rôle of infiltration in the hydrologic cycle , 1933 .

[12]  Paul G. Constantine,et al.  Active Subspaces - Emerging Ideas for Dimension Reduction in Parameter Studies , 2015, SIAM spotlights.

[13]  I. Rodríguez‐Iturbe,et al.  On the synthesis of random field sampling from the spectrum: An application to the generation of hydrologic spatial processes , 1974 .

[14]  Keith Loague.  Impact of rainfall and soil hydraulic property information on runoff predictions at the hillslope scale , 1988 .

[15]  Matthew D. Wilson,et al.  Uncertainty in predicted runoff due to patterns of spatially variable infiltration , 2007 .

[16]  J. P. Delhomme,et al.  Spatial variability and uncertainty in groundwater flow parameters: A geostatistical approach , 1979 .

[17]  R. Maxwell.  A terrain-following grid transform and preconditioner for parallel, large-scale, integrated hydrologic modeling , 2013 .

[18]  R. Allan Freeze,et al.  A stochastic‐conceptual analysis of rainfall‐runoff processes on a hillslope , 1980 .

[19]  J. Jaime Gómez-Hernández,et al.  ISIM3D: and ANSI-C three-dimensional multiple indicator conditional simulation program , 1990 .

[20]  R. Ababou,et al.  Implementation of the three‐dimensional turning bands random field generator , 1989 .

[21]  Olaf Kolditz,et al.  Surface‐subsurface model intercomparison: A first set of benchmark results to diagnose integrated hydrology and feedbacks , 2014 .

[22]  S. P. Neuman,et al.  Effects of kriging and inverse modeling on conditional simulation of the Avra Valley Aquifer in southern Arizona , 1982 .

[23]  D. Gleich,et al.  Computing active subspaces with Monte Carlo , 2014, 1408.0545.

[24]  Paul G. Constantine,et al.  Active subspaces for sensitivity analysis and dimension reduction of an integrated hydrologic model , 2015, Comput. Geosci..

[25]  Roger E. Smith,et al.  A Monte Carlo Analysis of the hydrologic effects of spatial variability of infiltration , 1979 .

[26]  Baxter E. Vieux,et al.  Impact on Sahelian runoff of stochastic and elevation-induced spatial distributions of soil parameters , 2002 .

[27]  Vijay P. Singh,et al.  Effect of spatial and temporal variability in rainfall and watershed characteristics on stream flow hydrograph , 1997 .

[28]  S. Ashby,et al.  A parallel multigrid preconditioned conjugate gradient algorithm for groundwater flow simulations , 1996 .

[29]  R. Maxwell,et al.  Integrated surface-groundwater flow modeling: A free-surface overland flow boundary condition in a parallel groundwater flow model , 2006 .

[30]  Keith Loague,et al.  The quixotic search for a comprehensive understanding of hydrologic response at the surface: Horton, Dunne, Dunton, and the role of concept‐development simulation , 2010 .