An Overview of the Noah-Distributed Land Surface Model

1
An Overview of the Noah-Distributed Land Surface
Model

• David J. Gochis, Wei Yu, Fei Chen, Kevin Manning
• WRF Land Surface Modeling Workshop
• Sep. 13, 2005

2
Brief Rationale for Noah-Distributed

• Standard land surface parameterizations
  characterize exchanges of radiation, heat, mass
  and momentum between the land and atmosphere
• Historically, treatment of terrestrial hydrology
  has been simplified 1-d formulations
• With high resolution implementations/applications
  there is now a need to explicitly account for
  enhanced hydrological processes
• Violating early assumptions
• 1. Surface runoff can not assumed to be
  captured by a stream channel
• 2. Lateral transfers from one cell may form
  significant input to adjacent cell

3
Brief Rationale for Noah-Distributed

• Standard land surface parameterizations
  characterize exchanges of radiation, heat, mass
  and momentum between the land and atmosphere
• Historically, treatment of terrestrial hydrology
  has been simplified 1-d formulations
• With high resolution implementations/applications
  there is now a need to explicitly account for
  enhanced hydrological processes
• Higher resolution capabilities ? land surface
  heterogeneity
• Earth systems-biogeochemcial cycling
• Mitigation of high-impact weather events (e.g.
  floods)

4
Outline

• Brief overview of the Noah LSM
• Noah-distributed core features
• Implementation of Noah-distributed into the
  NCAR/HRLDAS framework
• Ongoing and planned upgrades to Noah-distributed

5
(No Transcript)
6
Community Noah Land Surface Model Recent
Enhancements

• Recent Enhancement of the Community Noah LSM
  (released in WRF V2.0, May 2004) Noah-Unified
• Nearly-identical implementations of Noah LSM
  development effort NCAR, NCEP, U.S. Air Force
  Weather Agency, NASA, university community
• Fully modularized, F90 code conventions
• Seasonal surface emissivity
• surface emissivity is introduced as function of
  landuse
• Added surface emissivity in surface energy
  balance equation for both snow and non-snow
  surfaces
• Urban model improvements (the simple approach)
  such as
• Large roughness length
• Low surface albedo
• Large thermal capacity and thermal conductivity

7
Overland Flow Processes in Noah-Router
(NCAR Tech Note Gochis and Chen, 2003)

• New Parameters retention depth, surface
  roughness
• Ponded water in excess of retention depth subject
  to overland flow
• Overland flow fully-unsteady, explicit,
  finite-difference, 2-dimensional diffusive wave
  (generally applicable to length scales lt 1km)

8
Dynamic modeling of land-surface hydrology with
Noah-Router Ponded Water Processes
(NCAR Tech Note Gochis and Chen)

• New Parameters None
• Currently no formulation for partial area
  coverage
• Ponded water consists of residual of
  infiltration excess from previous time step and
  routed surface water
• Direct evaporation of ponded water reduces
  potential evaporation (no adj. for temp/albedo)
• Ponded water not evaporated is subject to
  infiltration

Issue May need to revise infiltration
formulation when using routed runoff to
calibrate Surface runoff can not assumed to be
captured by a stream channel
9
Subsurface Flow Routing Noah-Router
(NCAR Tech Note Gochis and Chen)

• New Parameters Lateral Ksat, n exponential
  decay coefficient
• Critical initialization value water table depth
• 8-layer soil model (2m depth, sealed bottom
  boundary)
• Quasi steady-state saturated flow model, 2-d
  (x-,y-configuration)
• Exfiltration from fully-saturated soil columns

Surface Exfiltration from Saturated Soil Columns
Lateral Flow from Saturated Soil Layers
Saturated Subsurface Routing Wigmosta et. al, 1994
10
Noah-Distributed Core Features

• Present issues in treatment of subsurface
  routing
• Frozen soil adjustment to soil water
• Remove soil ice from total soil moisture and
  route only liquid component
• Update of conductivity as a function of soil
  temp/fraction of frozen soil
• Inclusion of variable depth soils

11
Subgrid Routing

• Noah LSM is run at a variety of grid spacings
• Subsurface and overland flow routing need to be
  performed on a terrain grid (lt 1 km)
• Required fields are aggregated/disaggregated
  using a simple averaging scheme
• Soil water, infiltration excess, routing
  parameters
• Can offer significant computational savings
  compared to full resolution implementations of
  Noah LSM
• Sacrifice detail in current formulation

12
Noah-Distributed Core Features

• Subgrid disaggregation proposed new method
  carrying over weighting factors between LSM model
  executions
• Eliminates the loss of distributed information
  between routing time-steps

Noah land surface model grid
Routing Subgrids
13
Noah-Distributed Software Features

• F90, up to date with recent version in HRLDAS
• Routing routines (1-d and 2-d) are contained
  within a single module (all agg./disagg. Routines
  will be included into routing module)
• Routing and sub-grid options are switch-activated
  though a namelist file
• Options to output sub-grid state and flux fields
  to WRF consistent netcdf files
• Basic Flow
• LSM gt Disagg. gt Subsfc gt Overland gt Agg. gt LSM
  gt

14
Outline

• Brief overview of the Noah LSM
• Noah-distributed core features
• Implementation of Noah-distributed into the
  NCAR/HRLDAS framework
• Ongoing and planned upgrades to Noah-distributed

15
NCAR-HRLDAS (High Res. Land Data Assim. System)

• Rationale and basics of HRLDAS
• Create globally-deployable variable resolution
  equilibrated land surface conditions for NWP
  initializations
• Current static forcing data

Time
Variable Interval Dataset Grid Resolution Reference
Precipitation Hourly NEXRAD Stage IV 4 km Fulton et al, 1998
Surface Meteorology Hourly EDAS 40 km Rogers et al., 1995
Land Use Static USGS-24 Category 1 km Loveland et al., 1995
Greenness Fraction Monthly n/a 0.15 degree Gutman and Ignatov, 1998
Soil Classification Static STASGO-16 Category 1 km Miller and White, 1998
Overland Flow Roughness Coefficient Static n/a Mapped to Land Use Adapted from Vieux, 2001

16
HRLDAS

• Recent tests over International H2O Project
  (IHOP) domain
• 18 month execution 1 Jan, 2001 30 Jun, 2002

17
Top Layer Soil Moisture (fraction)
18
Total Column Soil Moisture (mm)
19
Total Surface Evapotranspiration (mm)
20
Ponded Water Evaporation (mm)
21
(No Transcript)
22
(No Transcript)
23
(No Transcript)
24
Outline

• Brief overview of the Noah LSM
• Noah-distributed core features
• Implementation of Noah-distributed into the
  NCAR/HRLDAS framework
• Ongoing and planned upgrades to Noah-distributed

25
Future Upgrades to Noah Distributed

• Improve runtime performance
• 2-d vs 1-d formulations
• DEM-based steepest descent method is much faster
• Strictly DEM based routing (kinematic)
  problematic in flat areas where change in sfc
  water influences flow direction (e.g. backwater)
• Working on a compromise algorithm
• default to DEM based routing
• check for backwater
• Perform search
• 1-d 129 model days/wall clock d vs. 2-d 97
  model days/wall clock day (1/3 faster)

26
Future Upgrades to Noah Distributed

• Complete parallelization
• Couple to stream channel model (via DESWAT
  project)
• Develop better method to nudge/assimilate
  groundwater and river/stream stage into modeling
  system
• Develop enhanced method to characterize
  stream-aquifer exchange

27
(No Transcript)