seasonal modeling (noaa) jian-wen bao sara michelson jim wilczak curtis fleming emily piencziak

Post on 13-Jan-2016

219 Views

Category:

Documents

1 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Seasonal Modeling (NOAA)

Jian-Wen BaoSara Michelson

Jim WilczakCurtis FlemingEmily Piencziak

Accomplishments

• Preparation of gridded data sets for the entire summer of 2000 to initialize MM5

• Refinement of the FDAA data preparation:

1. Elevation of each sigma level varies with observation sites.

2. Observations are interpolated from two adjacent data levels to a given model level.

3. An anisotropic spatial influence function is being implemented and tested.

• Observational data preparation for model evaluation:

1. Hand editted (level 1C) winds and RASS from 25 wind profilers for 60 days (3 August – 2 October, 2000)

2. Data from 1 June – 2 August remain to be editted

3. PBL depths for entire period, all profilers, have been calculated

PBL Depth Estimation

36km grid 95x91

12km grid 91x91

4km grid 190x190

All have 50 layers, with 22 in lowest 1km

Subregions for Model Evaluation

Wind profilersites

A Major Issue to Settle

What is the best model configuration for the seasonal simulations?

Through:

• Evaluation of chemical model simulations

• Evaluation of major transport processes

• Comparison of the simulations using V6 and V7

• Comparison of WRF and MM5

Chemical Model Evaluation

• MM5 runs of two 5 day periods ( Jul 24- Jul 29 and Aug 3- Aug 8 2000) were completed.

• Analysis and comparison of the above two runs with observations have been started.

Comparison of the Simulated and Observed Surface Winds for the Jul-Aug Case

ABL Height Comparisons

(Colored contours are TKE, and dots indicate the observed ABL height)

Observed clouds vs Model Simulation

Solar Radiation Fluxes from Various Versions of Models for the Jul-Aug Case

Physics Configuration in MM5:• the MYJ ABL and surface layer schemes• the NOAH land surface model (LSM)• the Dudhia short-wave, RRTM long-wave radiation schemes• the Reisner microphysics parameterization• the Grell convective scheme (only on the 36 and 12 km grids) Physics Configuration in WRF:• the MYJ ABL and surface layer schemes• the NOAH land surface model (LSM)• the Dudhia short-wave, RRTM long-wave radiation schemes• the Lin et al. microphysics parameterization• the Kain-Fritsch convective scheme (only on the 36 km and 12 km grids)

NCEP’s ETA 40-km isobaric analysis is used to initialize both WRF and MM5 at 1200 UTC 29 July 2000.

Domain and Physics Configurations of MM5 and WRF

MM5WRF

36 kmTopography

4 kmTopography

4 kmTopography

36 kmTopography

4 kmVeg-Frac

4 kmVeg-Frac

4 kmLanduse

4 kmLanduse

MM5WRF

Differences of MM5 and WRF

Differences in the LSM Initialization

MM5WRF

Temperature at Soil Layer 1

Differences in the LSM Initialization

MM5WRF

Temperature at Soil Layer 2

Differences in the LSM Initialization

MM5WRF

Temperature at Soil Layer 3

Differences in the LSM Initialization

MM5WRF

Temperature at Soil Layer 4

Differences in the LSM Initialization

MM5WRF

Moisture at Soil Layer 1

Differences in the LSM Initialization

MM5WRF

Moisture at Soil Layer 2

Differences in the LSM Initialization

MM5WRF

Moisture at Soil Layer 3

Differences in the LSM Initialization

MM5WRF

Moisture at Soil Layer 4

Comparison of the simulated and observed areal averaged 2m temperatures and 10m winds

Comparison of the simulated and observed areal averaged 2m temperatures and 10m winds

Comparison of the simulated and observed areal averaged 2m temperatures and 10m winds

Comparison of the simulated and observed areal averaged 2m temperatures and 10m winds

WRF MM5WRF-mm5ics

from 1200 UTC 29 July to 1200 UTC 2 August

from 1200 UTC 29 July to 1200 UTC 2 August

from 1200 UTC 29 July to 1200 UTC 2 August

from 0000 UTC 30 July to 1200 UTC 2 August

from 0000 UTC 30 July to 1200 UTC 2 August

from 0000 UTC 30 July to 1200 UTC 2 August

from 1200 UTC 30 July to 1200 UTC 2 August

from 1200 UTC 30 July to 1200 UTC 2 August

from 1200 UTC 30 July to 1200 UTC 2 August

comparison of simulated forward trajectories

Conclusions

• Averaging meteorological input in time

• Optimizing “tunable” parameters through sensitivity experiments

• Improving cloud physics and cloud-radiation interaction

• Using MM5V3-6 for the seasonal modeling

• Undesirable noise in the FDDA run

• Uncertainties in the LSM

• Some differences in the simulated and observed clouds on cloudy days

Recommendations

ETL CCOS Web Sitewww.etl.noaa.gov/programs/

modeling/CCOS/data

top related