An effective hyper-resolution pseudo-3D implementation of small scale hydrological features to improve regional and global climate studies P. Hazenberg1 , P. Broxton1 , M. Brunke1 , D. Gochis2 , G.-Y. Niu1 , J. Pelletier1 , P.A. Troch1 , X. Zeng1 1 University of Arizona, Tucson, AZ. 2 NCAR, Boulder, CO. 25 February 2014 Introduction: Current hydrology in CLM Lateral subsurface flow Subsurface-river network flow No lateral subsurface flow taken into account! Model resolution increases due to computational developments Lateral subsurface flow Increase in resolution Decrease importance of river flow Subsurface-river network flow Importance of: I river network decreases, I lateral subsurface flow increases, I elevation differences increase. Question: How to improve the CLM’s hydrology while still making use of vertical column information and focus? Develop a new hybrid 3-D hydrological model Separating the vertical and lateral hydrological response (pseudo 3-D): I Vertical response: I I I Use CLM soil column, Extend depth of column to bedrock, Lateral response: I I I I Design a new model (no linear reservoir), Defining a generic approach resolution independent, Differentiating different hydrological response units, Here: 1 km pixel resolution (hyperresolution), I Using different resolutions for vertical and lateral hydrological response, I Making use of high resolution datasets. Make use of high resolution datasets Barren Snow/Ice Cropland/Natural Urban Croplands Permanent Wetlands Grasslands Savannas Woody Savannas Open Shrublands Closed Shrublands Mixed Forests Deciduous Broadleaf Forests Deciduous Needleaf Forests Evergreen Broadleaf Forests Evergreen Needleleaf Forests Water Broxton et al., 2014 I Vegetation data I Digital elevation model (DEM) I Soil databases Make use of high resolution datasets Pelletier & Rasmussen, 2009 I Vegetation data I Digital elevation model (DEM) I Soil databases Make use of high resolution datasets Drainable porosity 30 m DEM I Vegetation data I Digital elevation model (DEM) I Soil databases Differentiation between different hydrological units 1st order, head water hillslopes Riparian zone D1 2nd order side slopes QCH Channel D2 α1 Dch st 1 order hillslope QHS,GW α2 Drpz Riparian zone or Wetland QGW QHS,CH Hillslope Flat region Channel Lake nd 2 order side slope QGW QCH QGW Make use of high resolution DEM to differentiate between hydrological units 30 m DEM QCH QHS,GW QGW QHS,CH Hillslope Flat region Channel Lake QGW QCH QGW Fraction of the different lateral hydrological components for each 1 km pixel. A hybrid 3-D approach for the hillslope type w P Eact D1 Channel dx D2 θ α1 Dch 1st order hillslope Tact D QOF,IE Roots D α2 Drpz Riparian zone or Wetland QOF,SE 2nd order side slope α QGW + QOF,SE QGW The 1-D Richards equation: ∂θ ∂ ∂Ψ = Kv (Ψ) + 1 − G(h) ∂t ∂z ∂z The 1-D hillslope storage Boussinesq equation (Troch et al., 2003) ∂h ∂ 1 ∂ (f (h) h) = w Kl h sin α + cos α + cos α Rgw ∂t w ∂x ∂x hgwl Comparison of hsB and CLM at the hillslope scale UNIFORM CONVERGENT DIVERGENT wtop = 50 m wtop = 90 m wtop = 10 m Distance to bottom (m) 100 80 L = 50 m 60 40 20 dy = 5 m 0 dx = 2 m Distance (m) 0 20 40 60 80 100 0 Distance (m) 20 40 60 80 100 Distance (m) Uniform Convergent Divergent CLM 0 −50 −100 −150 −200 −250 −300 −350 Theta (−) Sandy Loam 20 40 60 80 100 80 60 40 20 0 0.435 0.400 0.350 0.300 0.250 0.200 0 −50 −100 Clay loam 0 wbottom = 90 m whs (m) whs (m) wbottom = 10 m 80 60 40 20 0 Vertical depth, z (cm) whs (m) wbottom = 50 m 80 60 40 20 0 −150 −200 −250 −300 −350 0 50 100 150 0 50 100 150 0 50 100 150 0 50 100 150 Time (days) Theta (−) 0.476 0.400 0.350 0.300 0.250 0.200 A hybrid 3-D approach for lateral interaction between pixels P Tact Eact QOF QCH Roots QHS,GW QGW QHS,CH Hillslope Flat region Channel Lake QGW QCH D QCH QCH θ QGW hgwl QGW QGW,CH QGW The 1-D Richards equation: ∂θ ∂ ∂Ψ = Kv (Ψ) + 1 − G(h) ∂t ∂z ∂z The 2-D subsurface Boussinesq equations (Darcy) A ∂(h f ) ∂h ∂E ∂h ∂E = −∂ wx h kL + −∂ wy h kL + +Rgw A ∂t ∂x ∂x ∂y ∂y x y Computational scheme for hybrid 3-D hydrological model • Atmospheric input fluxes • Neighboring grid cell states • Ocean information Setup and parameters Hybrid 3D grid cell Vertical response Lateral recharge/ drainage flux For each time step • • • • • • Dimensions Number of columns Soil parameters Bedrock, lake depth Width functions Etc. Vertical Vertical Vertical Vertical Vertical soil soil soil soil soil column column column column column Lateral response River River River River River network network network network network column column column column column • Water content • Pressure head • Drainage/recharge rate Vertical recharge/ drainage flux Hillslope Hillslope Hillslope Hillslope Hillslope Hillslope Order of implementation Temporal input Lake Lake Lake Lake Lake column column column column column • Water level • Drainage/recharge rate • Saturated subsurface states • Saturated outflow flux Flatland River network • River levels • Discharge rates Lake • Lake levels • Discharge rates Testing the hybrid 3-D hillslope response in LEO I Landscape Evolution Observatory (LEO) in Biosphere 2 I Convergent hillslope (11x30 m) Recharge and drainage experiment I I Compare to 3D Richards model CATHY (Niu et al., 2013) Atm BC (mm h−1) Testing the hybrid 3-D hillslope response in LEO 12 10 8 6 4 2 0 −2 a) Storage (m3) 350 1D−hsB: 0.996 CATHY : 0.993 300 b) 250 200 Observed 1D−B CATHY 150 100 I I Convergent hillslope (11x30 m) Recharge and drainage experiment Compare to 3D Richards model CATHY (Niu et al., 2013) 0.8 Qgw(m3 h−1) I Landscape Evolution Observatory (LEO) in Biosphere 2 1D−B: 0.984 CATHY : 0.977 0.6 c) 0.4 Opt. param. range: min−max 0.10−0.90 0.25−0.75 0.2 0.0 5 1D−B: 0.106 CATHY : −1.078 4 Qof(m3 h−1) I d) 3 2 1 0 0 10 20 30 Time (h) 40 50 10 Atm BC (mm h−1) 10 3 2 12 10 8 6 4 2 0 −2 a) 350 Storage (m3) 10 4 1.00 NS, range 0.99 0.98 0.97 LEO obs. LEO soil 1D−B opt. min−max 0.10−0.90 0.25−0.75 1D−hsB: 0.996 CATHY : 0.993 300 b) 250 200 Observed 1D−B CATHY 150 100 0.8 Qgw(m3 h−1) 10 5 101 1D−B: 0.984 CATHY : 0.977 0.6 c) 0.4 Opt. param. range: min−max 0.10−0.90 0.25−0.75 0.2 0.0 0.0 0.1 0.2 0.3 0.4 Soil moisture, θ [−] Good correspondence with observations! 5 1D−B: 0.106 CATHY : −1.078 4 Qof(m3 h−1) Matric potential, Ψ [−cm] Testing the hybrid 3-D hillslope response in LEO d) 3 2 1 0 0 10 20 30 Time (h) 40 50 Atm BC (mm h−1) Testing the hybrid 3-D hillslope response in LEO 12 10 8 6 4 2 0 −2 a) 0.33 0.36 0.39 0.42 0.45 0.45 0.42 0.39 0.99 0.33 40 60 80 100 120 140 160 −1 θs [−] NS 0.36 Storage (m3) −10 −20 −30 −40 −50 −60 −70 −80 θs [−] Ψae [cm] Ψae [cm] 350 −10 −20 −30 −40 −50 −60 −70 −80 0.98 40 60 80 100 120 140 160 0.95 Kv,s [cm h ] 0.90 −1 Kv,s [cm h ] 0.5 1.0 0.5 0.33 0.36 0.39 0.42 0.45 θs [−] −1 Kv,s [cm h ] 200 Observed 1D−B CATHY 150 1D−B: 0.984 CATHY : 0.977 0.00 −1.0 0.5 40 60 80 100 120 250 0.8 0.50 1.0 −80 −60 −40 Ψae [cm] −20 Qgw(m3 h−1) 1.0 2.0 b [−] b [−] 0.6 c) 0.4 Opt. param. range: min−max 0.10−0.90 0.25−0.75 0.2 0.0 Improved understanding on parameter sensitivity! 5 1D−B: 0.106 CATHY : −1.078 4 Qof(m3 h−1) b [−] 2.0 b) 100 0.80 2.0 1D−hsB: 0.996 CATHY : 0.993 300 d) 3 2 1 0 0 10 20 30 Time (h) 40 50 Conclusions We are developing a new hybrid 3-D hydrological scheme for CLM: I Keep the current vertical column framework of CLM, I Differentiate between lateral and vertical response, I Extending the depth of the vertical column to the bedrock, I Identifying the response of different hydrological units (e.g. hillslope, flatland, river network and lake). For next period: I Finish implementation of shallow water equations for river network and lake, I Couple to vertical column structure of CLM, I Test the possibilities of this model at different scales (catchment, continental, global). I For longer run: Add possibility of using multiple vertical columns. Thanks for your attention!!! Questions? Email: [email protected] Acknowledgements: Department of Energy research grant (DE-SC0006773)
© Copyright 2024 ExpyDoc
