Modification of GFS Land Surface Model Parameters to Mitigate ...

Modification of GFS Land Surface Model Parameters to Mitigate ...

Modification of GFS Land Surface Model Parameters to Mitigate the Near-Surface Cold and Wet Bias in the Midwest CONUS: Analysis of Parallel Test Results 04 September 2012 Fanglin Yang, Shrinivas Moorthi, Helin Wei, Glenn White, Geoff Manikin,, Mike Ek, John Derber and Bill Lapenta NOAA/NWS/NCEP Environmental Modeling Center 1 Problem: Customers Noted the GFS PBL was too Cold and Wet Observed and simulated soundings valid 00 UTC 23 July 2012 Observed: Solid; GFS 24h Fcst: Dashed Lincoln, Illinois (ILX) Td=+9C

T=-7C Omaha, Nebraska (OAX) Td=+10C T=-8C The GFS 2m temperature and dew point forecasts valid in the late afternoon were too low and high, respectively 2 Spatial Distribution of Biases in 2-Meter Temperature and Dew Point 2 Meter Temperature GFS approximately 10-15 F too cool in the late afternoon for the mid-west GFS 24h Fcst

OBS 2 Meter Dew Point GFS dewpoint approximately 15-20 F too high in the late afternoon for the mid-west GFS 24h Fcst OBS 3 Sequence of Events, Actions and Results 24 May: EMC Model Evaluation Group (MEG) identified potential cold/wet bias in GFS over the Central

US 14 June: MDL and EMC received complaints from field that the GFS 2m temperature and dew point forecasts valid in the late afternoon were too low and high, respectively 19 June: EMC verified field observations and disassociated problem from the hybrid GDAS upgrade implemented 22 May 21 June: EMC isolated the problem to land surface model: Look up table for vegetation canopy resistance and root zone depth used in GFS was the culprit Implemented erroneously in the May 2011 upgrade Change allowed too much evaporation through vegetation types in central CONUS

24 June: EMC conducted limited tests using original LSM look up table and found positive impact in poor performing areas 29 June: EMC notified users of suspect cause and plans to correct through MDL June-August: EMC Conducted a retrospective parallels: June-August 2012 (CONUS warm season) Jan-Feb 2012 (CONUS cold season) CCS Availability limited due to SREF testing and service interruptions in August Change of LSM table has expected impacts on improving warm season PBL structure and precipitation No impact on CONUS PBL structure and precipitation during cold season

August 25: NCEP Director approves fix to be implemented 5 Sept. 4 Land Surface Model Look-up Table Minimum Canopy Resistance (RSMIN) and Root Depth Number (NROOT) were changed in May-2011 Implementation leading to increased evapotranspiration over some regions depending on vegetation type Operational GFS RSMTBL =(/300.0, 175.0, 175.0, 300.0, 70.0, 70.0, & 20.0, 70.0, 70.0, 70.0, 70.0, 20.0, & 70.0, 0.0, 0.0, 0.0, 0.0, 0.0, & 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, & 0.0, 0.0, 0.0, 0.0, 0.0, 0.0/) NROOT_DATA =(/4,4,4,4,4,4,4,4,4,4,4,4,4,0,0,

& 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0/) Prior to May 2011 implementation RSMTBL =(/300.0, 175.0, 175.0, 300.0, 300.0, 70.0, & 45.0, 225.0, 225.0, 225.0, 400.0, 45.0, & 150.0, 0.0, 0.0, 0.0, 0.0, 0.0, & 0.0, 0.0, 0.0, 0.0, 0.0, 0.0, & 0.0, 0.0, 0.0, 0.0, 0.0, 0.0/) NROOT_DATA =(/4,4,4,4,4,4,3,3,3,2,3,3,2,0,0, & 0,0,0,0,0,0,0,0,0,0,0,0,0,0,0/)

Parameter differences in table highlighted in RED Lower canopy resistance (i.e. lower values in table) and deeper root zone depth (i.e., larger values in table) will allow more evapotranspiration to occur for those vegetation types Modified table is used in CFS 5 Changes in Canopy Resistance Mixed forest: Reduced from 300 to 70 units Canopy Resistance (Operational Corrected) Cropland: Resistance Reduced from 45 to 20 units Bare soil: Reduced from 400 to 70 units

6 Changes in Root Zone Depth Root Zone Depth Layer (Operational Corrected) Positive values indicate shallower rooting depth 7 Numerical Experiments Use the LSM table employed in the operational GFS before May 2011 implementation Warm season experiment prt2mexp: Started from the operational GFS 20120531 18Z initial conditions Results presented for the period of Jun 14 Aug 22, 2012 Cool season experiment prt2mwin:

Started from the prd12q3s hybrid-enkf parallel 20120101 06Z initial conditions Results presented for the period of Jan 15 March 01, 2012 8 Summary of Results Warm Season Experiment (prt2mexp) Reduced late-afternoon surface cold bias and moisture bias over the Northern and Southern Great Plains found in the operational GFS Significantly improved CONUS precipitation skill scores especially for light amounts No significant change to: Hurricane track and intensity Standard forecast skill scores such as 500-hPa height AC and tropical wind RMSE

Slightly worsened the mid-day warm bias over the SE US Cool Season Experiment (prt2mexp) No significant change in: CONUS meteorological fields Standard global forecast skill metrics Detailed verification at Fanglins websites http://www.emc.ncep.noaa.gov/gmb/wx24fy/para/t2mbias/exp2012 / http://www.emc.ncep.noaa.gov/gmb/wx24fy/para/t2mbias/exp2012w/ 9 Time Averaged 2 Meter Temperature (C) 14 June to 22 August 2012 (00Z Cycles Only) CONUS Northern and Southern Great Plains regions Northern Great Plains CONUS East

CONUS West Southern Great Plains 10 10 Example of Correction Obtained in GFS Parallel 21h GFS 2-m Temperature Forecasts (F) Valid 21UTC 22 July 2012 Operational GFS LSM GFS Parallel GFS approximately 10-15 F too cool in the late afternoon for the mid-west (see slide #3 for OBS) Correction to LSM alleviates cold bias (see slide #3 for

OBS) Corrected LSM in GFS found to mitigate 4 to 16 F cold bias in operations and compares favorably with observations Difference Ops Para 11 Example of Correction Obtained in GFS Parallel 24h GFS 2-m Dewpoint (F) Forecasts Valid 00 UTC 23 July 2012 Operational GFS LSM GFS Parallel GFS dewpoint approximately 15 F too high in the late

afternoon for the midwest (see slide #3 for OBS) Correction to LSM alleviates moist bias (see slide #3 for OBS) Difference Ops Para Corrected LSM in the GFS found to reduce dewpoint by 8 to 16 F and compares more favorably with observations 12 Time Averaged 2 Meter Temperature (C) 14 June to 22 August 2012 (00Z Cycles Only)

CONUS East and West regions CONUS East CONUS West 13 13 Quantitative Precipitation Forecasts (CONUS) 14 June to 22 August 2012 (00Z Cycles Only) Day 1: 24h Period ending at 36h forecast 14 14 Quantitative Precipitation Forecasts (CONUS) 14 June to 22 August 2012 (00Z Cycles Only)

Day-2: 24h Period ending at 60h forecast 15 15 Quantitative Precipitation Forecasts (CONUS) 14 June to 22 August 2012 (00Z Cycles Only) Day-3: 24h Period ending at 84h forecast 16 16 500 hPa Anomaly Correlation 14 June to 22 August 2012 (00Z Cycles Only) Northern Hemisphere Southern Hemisphere

Operational LSM Para 17 17 Large ScaleTemperature and Moisture Verification 14 June to 22 August 2012 (00Z Cycles Only) 24 and 48h Forecasts for the Northern Hemisphere and North America Bias RMSE North America Temperature (C) Pressure (hPa)

Pressure (hPa) Northern Hemisphere Temperature (C) Bias RMSE RMSE North America Mixing Ratio(g/kg) Pressure (hPa) Pressure (hPa) Northern Hemisphere Mixing Ratio (g/kg) Bias

Operational GFS= Solid LSM Parallel = Dotted Bias 24h Fcst = Black 48h Fcst = Red RMSE 18 Large ScaleTemperature and Moisture Verification 14 June to 22 August 2012 (00Z Cycles Only) 24 and 48h Forecasts for the Southern Hemisphere and Tropics Bias RMSE

TropicsTemperature (C) Pressure (hPa) Pressure (hPa) Southern Hemisphere Temperature (C) Bias RMSE RMSE Tropics Mixing Ratio(g/kg) Pressure (hPa) Pressure (hPa)

Southern Hemisphere Mixing Ratio (g/kg) Bias Operational GFS= Solid LSM Parallel = Dotted Bias 24h Fcst = Black 48h Fcst = Red RMSE 19 Atlantic Hurricane Track Errors 01 June to 23 August 2012 All 4 GFS cycles

20 Atlantic Hurricane Intensity Errors 01 June to 23 August 2012 All 4 GFS cycles 21 East-Pacific Hurricane Track Errors 01 June to 19 August 2012 All 4 GFS cycles 22 East-Pacific Hurricane Intensity Errors 01 June to 23 August 2012 All 4 GFS cycles 23

Time Averaged 2 Meter Temperature (C) 15 January to 01 March 2012 (00Z Cycles Only) CONUS West and East regions CONUS West CONUS East 24 24 Quantitative Precipitation Forecasts (CONUS) 15 January to 01 March 2012 (00Z Cycles Only) Day 1: 24h Period ending at 36h forecast 25 25 Quantitative Precipitation Forecasts (CONUS)

15 January to 01 March 2012 (00Z Cycles Only) Day 3: 24h Period ending at 84h forecast 26 26 500 hPa Anomaly Correlation 15 January to 01 March 2012 (00Z Cycles Only) Northern Hemisphere Southern Hemisphere 27 27 Large ScaleTemperature and Moisture Verification 15 January to 01 March 2012 (00Z Cycles Only)

24 and 48h Forecasts for the Northern Hemisphere and North America Bias RMSE North America Temperature (C) Pressure (hPa) Pressure (hPa) Northern Hemisphere Temperature (C) Pressure (hPa) Pressure (hPa) RMSE

RMSE Northern America Mixing Ratio (g/kg) Northern Hemisphere Mixing Ratio (g/kg) Bias Bias Hybrid Parallel= Solid LSM Parallel = Dotted Bias 24h Fcst = Black 48h Fcst = Red RMSE

28 Large ScaleTemperature and Moisture Verification 15 January to 01 March 2012 (00Z Cycles Only) 24 and 48h Forecasts for the Southern Hemisphere and Tropics Bias RMSE TropicsTemperature (C) Pressure (hPa) Pressure (hPa) Southern Hemisphere Temperature (C) Bias

RMSE Hybrid Parallel= Solid LSM Parallel = Dotted RMSE Tropics Mixing Ratio (g/kg) Pressure (hPa) Pressure (hPa) Southern Hemisphere Mixing Ratio (g/kg) Bias Bias 24h Fcst = Black

48h Fcst = Red RMSE 29 Summary of Results Warm Season Experiment (prt2mexp) Reduced late-afternoon surface cold bias and moisture bias over the Northern and Southern Great Plains found in the operational GFS Significantly improved CONUS precipitation skill scores especially for light amounts No significant change to: Hurricane track and intensity Standard forecast skill scores such as 500-hPa height AC and tropical wind RMSE Slightly worsened the mid-day warm bias over the SE US

Cool Season Experiment (prt2mexp) No significant change in: CONUS meteorological fields Standard global forecast skill metrics Detailed verification at Fanglins websites http://www.emc.ncep.noaa.gov/gmb/wx24fy/para/t2mbias/exp2012 / http://www.emc.ncep.noaa.gov/gmb/wx24fy/para/t2mbias/exp2012w/ 30

Recently Viewed Presentations

  • Senior E.D.

    Senior E.D.

    Intrinsic factors Cognitive impairment Polypharmacy - four or more medications Sedatives, antihypertensive and psychotropic medications Alcohol Impaired mobility Fall history Sensory defects (hearing and vision) Frailty Postural hypotension ESI Severity Index 1, 2, or 3? 5 Level Triage System (2003...
  • participation and community engagement PACE lindie clark, academic

    participation and community engagement PACE lindie clark, academic

    staged implementation of PACE over the next 4 years. from 2012: participation unit a core requirement in some 31 degree programs (about half of the total number of degrees), and 8 majors/streams. a range of other participation activities open to...
  • Learning Target: I can use data to compare

    Learning Target: I can use data to compare

    Learning Target: I can use data to compare mass, volume and density based on graphical information. Calculate Density . Practice reading Mass vs. Volume Graphs. Interpreting Mass vs. Volume Graphs . Practice citing data (Density Unit test preparation)
  • GOING VIRAL! - Socorro Independent School District

    GOING VIRAL! - Socorro Independent School District

    Virus Tag 2. Viral comic strip Title Panel (book cover): Project title (including the infection name), team members Cause Panel: What type of virus causes the infection? Pathogen Characteristics Panel: Describe the- virus (normal or retrovirus), shape (envelope?), DNA or...
  • Using 21st Century Tools and Ideas To Make YOU Indispensable

    Using 21st Century Tools and Ideas To Make YOU Indispensable

    Using 21st Century Tools and Ideas To Make YOU Indispensable (Winter Ed.) ... GLOGSTER.EDU can post assignments, monitor students & have free peer sharing . ... Using 21st Century Tools and Ideas To Make YOU Indispensable Last modified by:
  • Hw: All Chapter 5 problems and exercises Average

    Hw: All Chapter 5 problems and exercises Average

    Hw: All Chapter 5 problems and exercises Test 1 results Average 75 Median 78 >90 >80 >70 >60 >50 <50 521 - 79 522 - 76 523 - 73.3 524 - 76.5 525 - 76 526 - 69.7 Outline Applications...
  • Confucius &amp; Confucianism

    Confucius & Confucianism

    Confucianism Confucius believed that teaching and learning was the way to improve self, family, and society as whole Concerned primarily with restoring social stability and order What is the basis of a stable, unified, and enduring social order? a system...
  • Chapter 1

    Chapter 1

    Complete T-chart Advantages and Disadvantages or venn diagram * * Know characteristics of Hybrid organizations. ... markets Chevrolet's Nova Advanced manufacturing and communications eBusiness Hybrid organizational structures S-Corp, LLP, LLC, LP (pg. 16) In-closing Using your index card ...