25 ◦ F MOS Chinook Errors Billings 25 ◦ F MOS Temperature Errors in the US Background Chinook Belt Types of Errors Distribution Recent Cases Closing Thoughts Brian J. Billings Millersville University, Millersville, PA June 26, 2018
25 ◦ F MOS ESCI 390 - Mountain Meteorology Chinook Errors Billings Background Types of Errors Distribution • One-time elective offered Fall 2017 Recent Cases • A range of class projects/activities: Closing Thoughts • 2 NSF Educational Deployments ( P11 ) - analyzed data from DOW visit - participated in King Air visit • In MOS project students obtained data for a complex terrain station • mesowest.utah.edu • mesonet.agron.iastate.edu • Using Excel macro, searched for events with extreme model errors
25 ◦ F MOS Methodology Chinook Errors Billings Background • Areas east of Rockies affected Types of Errors by chinook warming Distribution • No formal chinook belt definition Recent Cases Closing Thoughts • c.f. Alberta: Longley (1967) • Selected the 6 NWS WFO’s nearest to mountains • 72 hour forecasts back to 2002 • 25 ◦ F (14 ◦ C) errors ≤ 24 hours apart grouped as event • Event defined by largest error and shortest forecast time Available at www.metviz.com/MOS.xlsx
25 ◦ F MOS Overchinooking Chinook Errors Billings Background • MOS chinook wind timing exceed observed Types of Errors • Break-in Broadening • Onset Broadening Distribution Recent Cases Closing Thoughts • Missed Pause • Phantom Chinook
25 ◦ F MOS Cold Surges Chinook Errors Billings • Region also affected by cold surges Background • Colle and Mass (1995); Schultz et al (1997) Types of Errors Distribution Recent Cases Closing Thoughts • Late Cold Surge • No Cold Surge
25 ◦ F MOS Other Types of Error Chinook Errors Billings • 24% of total number of events Background Types of Errors Distribution • Late Chinook/Early Cold Surge (10/5) Recent Cases • 6 of former in 2007 Closing Thoughts • Thunderstorm Outflow (27) • 17 in KRAP and KGTF-NAM • Cloud/Precipitation (33) • Southerly WAA (10) • 9 in KRAP and KCOS
25 ◦ F MOS Monthly and Hourly Distribution Chinook Errors Billings Background Types of Errors Distribution Temporal Spatial Recent Cases Closing Thoughts • January maximum; • Mid-late daytime • Maximum for slightly more after maximum longest forecasts • more NAM events • 2 extra 6-hourly • more GFS events forecasts issued for shorter leads outside cold season
25 ◦ F MOS Annual Distribution Chinook Errors Billings Background • 357 total events Types of Errors • 119 (33.3%) in dataset’s first 22-25 months (13%) Distribution Temporal • 19 of 40 No Cold Surge events occur in this timeframe Spatial Recent Cases Closing Thoughts • (Lott and Miller 1997, QJRMS; Alpert 2004, WAF/NWP) • Low years have 8-14 events • High years have 20-29 events • 2005-06, 2010, 2012-13, • 2007-09, 2011, 2014, 2017 2015-16
25 ◦ F MOS Total Events by Model Chinook Errors Billings Background Types of Errors Distribution Temporal Spatial Recent Cases Closing Thoughts • Similar NAM totals along most of Rockies • GFS minimum around Columbia-Snake channel • Higher totals in KRAP; Highest totals in KGTF
25 ◦ F MOS Total Events by Error Type Chinook Errors Billings CHK = Overchinooking; CDS = Late or No Cold Surge Background Types of Errors Distribution Temporal Spatial Recent Cases Closing Thoughts • Cold surges are largest total in 3 cities • Slightly more other events in KRAP and KCOS • Overchinooking significantly more important at KGTF
25 ◦ F MOS KTFX/KBIL Phantom Chinooks Chinook Errors Billings Background • 7 of 10 phantom chinooks occur in winter 2016/17 Types of Errors Distribution Recent Cases KTFX Chinooks KDEN Cold Surges Closing Thoughts KTFX Observed Skew-T • Some chinooks observed at upstream MT DOT stations; others entirely absent from plains
25 ◦ F MOS KDEN Cold Surges Chinook Errors Billings Background 2014-2017 GFS and NAM MOS Types of Errors Distribution Recent Cases KTFX Chinooks KDEN Cold Surges Closing Thoughts • Extreme/large errors more often affect two/six forecast periods • Forecast excess electricity sold to market/utilities needs replacing
25 ◦ F MOS Closing Thoughts Chinook Errors Billings Background Types of Errors • A wider range of classroom activities leads to a better chance of Distribution students getting more deeply involved Recent Cases Closing Thoughts • Knowledge of historically important changes to model configurations is critical for comparing errors from past years • Four dominant causes of extreme MOS errors in this region are: 1 Late Cold Surges 2 Overchinooking 3 Clear or Overcast 4 Thunderstorms • Field campaigns for some phenomenon could be anywhere along the lee of Rockies (1 and 3 above); others should be in MT/AB • Benefit for users in improving short-range model details, as well as extending long-range
Recommend
More recommend