Large-Eddy Simulations: WORK IN PROGRESS!
IMPORTANT: These results are NOT final. They are
simply made available on our website for internal organization
and to facilitate discussion among ourselves and with other
interested investigators. Comments are welcome.
Please e-mail them to
shane@lidar.ssec.wisc.edu.
THE OBJECTIVE is to use the University of Wisconsin's scalable nonhydrostatic
modeling system (UW-NMS, see Greg Tripoli, 1992, Monthy Weather Review) to
simulate the leading edge of the lake-induced internal convective boundary
layer and COMPARE the simulations with volume imaging lidar data from Lake-ICE.
The TIBL simulations are interesting because they attempt to capture a non-homogeneous
boundary layer growing in a non-moving grid. The list of simulations
below also include examples of homogeneous convective (#40, 42 and 47),
an Ekman BL (#91), and homogeneous shear-driven ML (#190 for example).
Each simulation is identified by a different number.
Please click on the simulation number you wish to see.
All simulations were run in parallel on an
IBM J50
SMP computer with 8 processors and 2 GB of memory. (We would love to have
an
IBM s80!)
Many thanks to IBM Corporation for making this research possible.
LES #266: Upwind (of 13 Jan 98)
(created March 2001) Same as #263 except HDKR=VDKR=0.15.
200x120x69, albedo=0.66, dx=dy=dz=15 m, All over land.
Non-enstrophy conserving. Periodic lateral BCs.
LES #263: Upwind (of 13 Jan 98)
(created Feb 2001) Same as #252 except NON enstrophy conserving.
200x120x69, albedo=0.66, dx=dy=dz=15 m, HDKR=VDKR=0.05. All over land.
Periodic lateral BCs.
LES #260: IBL (of 13 Jan 98)
(created Jan-Feb 2001) Best IBL simulation so far!
Initialized by replicating the 30-minute solution of
#252. 800x120x69, left-half over snow-covered land with albedo=0.66;
right-half over 279 K water. dx=dy=dz=15 m. HDKR=VDKR=0.05.
Flow recycler turned on to maintain turbulent inlet.
LES #256: IBL with tracer inflow problem (of 13 Jan 98)
(created Jan 2001) Initialized by replicating the 30-minute solution of
#252. 800x120x69. left-half over snow-covered land with albedo=0.66;
right-half over 279 K water. dx=dy=dz=15 m. HDKR=VDKR=0.05.
Flow recycler turned on to maintain turbulent inlet.
LES #252: Upwind (of 13 Jan 98)
(created Dec 2000) used to initialize #260. 200x120x69. Albedo=0.66 for
dT/dt=0.8 degrees per hour at 7.5 m AGL.
dx=dy=dz=15 m. HDKR=VDKR=0.05. All over land.
Periodic lateral BCs.
LES #251: Upwind (of 13 Jan 98)
(created Dec 2000) 200x120x69 with albedo=0.78 for
no temperature change at surface with time.
dx=dy=dz=15 m. HDKR=VDKR=0.05. All over land.
Periodic lateral BCs.
LES #234: Upwind (of 13 Jan 98)
(created Dec 2000) First run with radiation and soil-model turned off.
Completely shear driven.
200x120x69, dx=dy=dz=15 m. HDKR=VDKR=0.05. All over land.
Periodic lateral BCs.
LES #228 and #229: Upwind (of 13 Jan 98)
(created Dec 2000)
#228 has Arakawa & Lamb and #229 has momentum conserving.
Everything else same as #225 except smaller domain (40 x 40 x 69) for testing
faster.
LES #225: Upwind (of 13 Jan 98)
(created Dec 2000) Shear & heat driven. 200x120x69, dx=dy=dz=15 m.
HDKR=VDKR=0.05. All over land. Periodic lateral BCs.
This version has a 3-piece wind profile for initialization.
LES #223: TIBL (of 13 Jan 98)
(created Dec 2000) Continuation of LES#212 except with
VDKR=HDKR=0.00. Everything else same as LES#215, #216 and 222.
LES #215: TIBL (of 13 Jan 98)
(created Oct 2000) Continuation of LES#212 except with
VDKR=HDKR=0.10 instead of 0.20. TR1P and TKE fixed on the inflow wall.
Building is still located at the shoreline.
LES #214: TIBL (of 13 Jan 98)
(created Oct 2000) Duplicate of LES#213 except with
VDKR=HDKR=0.1 instead of 0.2 in LES#213. LES#214 is also a
continuation of LES #212 with the
Sadourney advection scheme (instead of the Arakawa & Lamb scheme in #212).
Building is still located at the shoreline.
LES #213: TIBL (of 13 Jan 98)
(created Oct 2000) Continuation of LES #212 with the
Sadourney advection scheme (instead of the Arakawa & Lamb scheme in #212).
Building is still located at the shoreline.
LES #208: TIBL (of 13 Jan 98)
(created Aug-Sept 2000) Continuation of LES #195 with
east-west periodic condition turned off and restoring zone
turned off. First TIBL run with integrated mass-flux on
east & west walls being saved.
LES #198: TIBL (of 10 Jan 98)
(created July 2000)
Second simulation of 10 January 1998. 6.5 km tall domain.
800x120x69 grid-points, dx=dy=30 m.
dz=30 m up to 1200 m. Expands to 250 m above.
DKR and ISCALE variables changed.
LES #197: TIBL (of 10 Jan 98)
(created July 2000)
First simulation of 10 January 1998.
800x120x69 grid-points, dx=dy=dz=30 m.
Features 1-km deep ML with surface flow from the WSW.
Domain is 24-km long, 3.6 km wide, and 2-km tall.
LES #195: TIBL (of 13 Jan 98)
(created May 2000) 800x120x69 grid-points, dx=dy=15 m, dz=15 m
up to 675-m then stretches gradually to 30-m. Purpose of this
run was to test the effects of a slightly taller (300-m more)
domain. Also, passive tracer is being run through the
restoring zones on the ends.
LES #193: TIBL
(created May 2000) 800x120x69 grid-points, dx=dy=dz=15 m.
Folded 1-hr solution to #190 four times as initial condition.
Periodic in lateral directions with no restoring zones.
LES #190: Upwind PBL
(created Apr 2000) 200x120x69 grid-points, dx=dy=dz=15 m.
All over land. Stationary grid, periodic in lateral directions.
Domain is 25% of size of LES #181.
Using 6 x 1-2-1 filter on initial perturbations.
New style of MPEGs created using Matlab.
LES #188 & 189: Upwind PBLs
(created Mar 2000) Both 200x60x69 grid-points, dx=dy=dz=15 m.
All over land. Only difference is #188 is stationary grid and
#189 is moving grid.
Using 6 x 1-2-1 filter on initial perturbations.
LES #186 & 187: Upwind PBLs
(created Mar 2000) Both 200x60x69 grid-points, dx=dy=dz=15 m.
All over land. Only difference is #186 is stationary grid and
#187 is moving grid.
Using 1 x 1-2-1 filter on initial perturbations.
LES #174 & 183: Upwind PBLs
(created Mar 2000) Both 200x60x69 grid-points, dx=dy=dz=15 m.
All over land. #174 uses the old enstrophy conserving scheme
and #183 uses the new enstrophy conserving scheme.
LES #181: TIBL with restoring zones
(created Mar 2000) 800x120x69 grid-points, dx=dy=dz=15 m. (12 km long, 1.8
km wide by 1 km tall.) 6-km land, 6-km water. Periodic in both
horizontal directions. 900-m restoring zones in east and west ends of
domain. This page includes animations created using Matlab.
Click on the thumbnail to the left to see a comparison of results from
simulation numbers
120, 121, 163, and 149. All are from 3-D grids with 15-m resolution.
#120 and #121 show a TIBL developing in
laminar flow. #163 shows a shear-generated ML developing over 12-km of
land. #149 (bottom) shows a TIBL developing within a
shear-generated ML. Note the color bar is log-scale of the variance of
water vapor and the images are from one instant in time. In the future,
simulated lidar backscatter will be used and several snapshots
averaged together to obtain the mean TIBL shape. Click on the
individual simulation numbers below to learn more about each run.
LES #163: Land boundary layer
(created Dec 99) 800x110x69 grid-points, dx=dy=dz=15 m. (12 km long, 1.65
km wide by 1 km tall.) Same as LES #149 except entire domain is
over land.
LES #149: TIBL
(created Nov 99) 800x110x69 grid-points, dx=dy=dz=15 m. (12 km long, 1.65
km wide by 1 km tall.) Same as LES #147 except FIXED western bounary
condition instead of floating.
LES #147: TIBL
(created Oct 99) 800x110x69 grid-points, dx=dy=dz=15 m. (12 km long, 1.65
km wide by 1 km tall.) 6.07-million grid-points! Reduced the magnitude of
perturbations on the west-wall to 1 cm/s and location of grid-points to be
disturbed to the bottom 3. Using "floating" western boundary condition.
LES #143: TIBL
(created Oct 99) 800x50x69 grid-points, dx=dy=dz=15 m. (12 km long, 0.75
km wide by 1 km tall.) Our first attempt at simulating a transitional
flow that goes from turbulent to turbulent MLs. (Previous TIBL simulations
(i.e. LES120 and 121) were laminar to turbulent transitions.) Here, in
LES143, we randomly wiggled all the v-components of flow below 400 m on
the west wall by + or - 0.5 m/s on each time step.
LES #141: Upwind Wall Experiment
(created Oct 99) 50x50x69 grid-points, dx=dy=dz=15 m. (0.75 km square by
1 km tall.) Purpose was to experiment with randomly perturbing the west
inflow wall in an attempt to generate large-eddies downsteam of it. To
do this we randomly wiggled all the v-components of flow below 400 m on
the west wall by + or - 0.5 m/s on each time step.
LES #125: Upwind PBL
(created Sept 99) 200x50x69 grid-points, dx=dy=dz=15 m.
This domain is 3 km (east-west) by 750 m (north-south) wide by 1025 m tall.
Purpose was to simulate the shear driven mixed layer west of the Lake.
Laterial boundary conditions are periodic in both directions.
LES #121: TIBL
(created Aug 99) 400x50x80 grid-points, 15 m x 15 m
(horizontal) x 1-->15 m (vertical) resolution.
This domain is 6 km (east-west) by 750 m (north-south) wide by 1025 m tall.
Purpose was to simulate the lake-induced thermal internal boundary layer
using a sounding from LES #107 as the horizontally homogeneous initial condition.
LES #120: TIBL
(created Aug 99) 400x50x69 grid-points, 15 m x 15 m
(horizontal) x 15 m (vertical) resolution.
This domain is 6 km (east-west) by 750 m (north-south) wide by 1025 m tall.
Purpose was to simulate the lake-induced thermal internal boundary layer
using a sounding from LES #107 as the horizontally homogeneous initial condition.
LES #104: Upwind PBL
(created June 99) 40x40x40 grid-points, 100 m x 100 m
(horizontal) x 50 m (vertical) resolution.
This domain is 4 km (east-west) by 4 km (north-south) wide by 2.0 km tall.
Purpose was to simulate the 400 m deep ML upwind of the IBL domain
and obtain a "steady-state" or "equilibrium" solution. Large
gravity waves develop in the inversion.
LES #91: Neutrally-stratified, Ekman PBL
(created late May 99) 40x40x40 grid-points, 50 m (cross-flow) x 100 m
(downstream) x 37.5 m (vertical) resolution. Grid is moving with
10 m/s westerly geostrophic wind.
This domain is 4 km (east-west) by 2 km (north-south) wide by 1.5 km tall.
Purpose is to compare results with Andren et al. (1994).
LES #47: Homogeneous CBL
(created March 99) 40x40x60 grid-points, 160x160 m horizontal resolution.
This domain is 6.4 km (east-west) by 6.4 km (north-south) wide.
Same as LES#40 except for new executable and 10 more grid-points
in the vertical. Will start putting line plots in here.
LES #42: Homogeneous CBL
(created February 99) 80x80x50 grid-points, 160x160 m horizontal resolution.
This domain is 12.8 km (east-west) by 12.8 km (north-south) wide.
Purpose of this simulation is see if doubling the domain size in LES#40
has any effect on the structures produced.
LES #40: Homogeneous CBL
(created February 99) 40x40x50 grid-points, 160x160 m horizontal resolution.
This domain is 6.4 km (east-west) by 6.4 km (north-south) wide.
Purpose of this simulation is to compare with horizontally homogeneous
CBL turbulence statistics with those from Nieuwstadt et al. (1993).
LES #37 & 39: Lake-ICE TIBL
(created Dec 98 - Jan 99) 400x200x80 (6.4 million!) grid-points, 15x15 m horizontal resolution.
This domain is 6.0 km (east-west) by 3 km (north-south) wide.
LES #10: Lake-ICE TIBL
(created fall 98) 140x70x80 grid-points, 50x50 m horizontal resolution.
This domain is 7 km (east-west) by 3.75 km (north-south) wide.
LES #16: Lake-ICE TIBL
(created fall 98) 140x70x80 grid-points, 25x25 m horizontal resolution.
This domain is 3.5 km (east-west) by 1.75 km (north-south) wide.
LES #17: Lake-ICE TIBL
(created fall 98) 140x70x80 grid-points, 10x10 m horizontal resolution.
This domain is 1.4 km (east-west) by 750 m (north-south) wide.
UW Lidar // July 25, 1999 // root@lidar.ssec.wisc.edu
Return to the
Index
Click on the thumbnail to the left to see a comparison of results from
simulation numbers
120, 121, 163, and 149. All are from 3-D grids with 15-m resolution.
#120 and #121 show a TIBL developing in
laminar flow. #163 shows a shear-generated ML developing over 12-km of
land. #149 (bottom) shows a TIBL developing within a
shear-generated ML. Note the color bar is log-scale of the variance of
water vapor and the images are from one instant in time. In the future,
simulated lidar backscatter will be used and several snapshots
averaged together to obtain the mean TIBL shape. Click on the
individual simulation numbers below to learn more about each run.