8th NCEP Ensemble User Workshop
8th NCEP Ensemble User Workshop
NCWCP, College Park, 20740 MD
August 27-29 2019
Day one (August 27):
8:00 - 8:30am Refreshment/Registration
Session 1: Introduction/logistics; NWS roadmap; NCEP ensemble models review
Chair: Vijay Tallapragada
08:30 - 08:45 1.1 Opening remarks (Brian Gross; NCEP/EMC director)
08:45 - 09:00 1.2 Introduction and logistics - Walter Kolczynski (EMC/IMSG)
09:00 - 09:20 1.3 NWS roadmap - Yan Xue (OSTI; Invited)
09:20 - 09:50 1.4 NCEP Global Ensemble Forecast System - Xiaqiong Zhou (EMC/IMSG)
09:50 - 10:00 1.5 NCEP Global Wave Ensemble Forecast System - Henrique Alves (EMC)
10:00 - 10:30 Break
Session 2: Global ensemble forecast systems: ECMWF and NOAA partners
Chair: Jack Kain
10:30 - 11:10 2.1 ECMWF Global Ensemble Forecast System - Simon Lang (Invited; ECMWF)
11:10 - 11:30 2.2 CMC Global Ensemble Forecast System and Application - Stéphane Gagnon (CMC)
11:30 - 11:50 2.3 Navy S2S coupled ensemble system - Justin McLay (NRL)
11:50 - 12:00 2.4 Air Force Global Ensemble Forecast System - Evan Kuchera (557th Weather Wing)
12:00 - 01:30 Lunch break and posters
Session 3: Ensemble products and application for weather - Part I
Chair: Jason Levit
01:30 - 02:00 3.1 Overall Center’s Review - Dave Novak (Invited; WPC) 02:00 - 03:30 NCEP Center + JTWC Overviews
02:00 - 02:15 3.2 WPC - Bruce Veenhuis
02:15 - 02:30 3.3 OPC - Joseph Sienkiewicz
02:30 - 02:45 3.4 AWC - Brian Pettegrew
02:45 - 03:00 3.5 SPC - Israel Jirak
03:00 - 03:15 3.6 NHC - Eric Blake
03:15 - 03:30 3.7 JTWC - Matt Kucas
03:30 - 04:00 Break
Session 4: Panel discussion on model development
Discussion topics: See below
04:00 - 05:30 Panelists: Vijay Tallapragada (EMC) - Lead, Jacob Carley (EMC), Matt Rosencrans (CPC), Eric Blake (NHC), Ernie Wells (OWP), Israel Jirak (SPC), Partha Mukhopadhyay (IITM
Day two (August 28):
Session 5: Ensemble products and application for weather - Part II
Chair: James Nelson
08:30 - 09:00 5.1 Overall NWS fields products review - Jeff Craven (MDL; Invited)
09:00 - 10:00 NWS Region Overviews
09:00 - 09:10 5.2 Eastern Region - Josh Watson
09:10 - 09:20 5.3 Southern Region - Greg Patrick
09:20 - 09:30 5.4 Central Region - Greg Mann
09:30 - 09:40 5.5 Western Region - Matthew Jeglum
09:40 - 09:50 5.6 Alaska Region - Gene Petrescu
09:50 - 10:00 5.7 Pacific Region - Bill Ward
10:00 - 10:30 Break
Session 6: Ensemble applications from community - Part I
Chair: Kathryn Gilbert
10:30 - 11:00 6.1 Tools to Utilize Ensembles During the Forecast Process: A Stony Brook CSTAR Perspective - Brian Colle (Invited; Stony Brook U.)
11:00 - 11:15 6.2 A Near-Median Density Index: An Order Statistics Approach to Assessing Ensemble Uncertainty; Keith Brill (WPC/IMSG)
11:15 - 11:30 6.3 The Spread of Tropical Storm Tracks in NCEP’s Global Ensemble model -
Frank Colby (UMass-Lowell)
11:30 - 11:45 6.4 The Testing of an Ensemble Based Canonical Tool For Prediction in the Day 8-10 Forecast Period at the Weather Prediction Center - Mike Bodner (WPC)
11:45 - 12:00 6.5 Tracking and Usage of Ensemble Model Heavy Precipitation Objects at the Weather Prediction Center; Michael Erickson (WPC)
12:00 - 1:30 Lunch break and posters
Session 7: Ensemble application from community - Part II
Chair: Yuejian Zhu
01:30 - 02:00 7.1 The MOGREPS Ensemble Systems - Ken Mylne (UK Met Office; Invited)
02:00 - 02:15 7.2 Impact Based Forecast for the Canadian Armed Forces: from a deterministic
to a probabilistic approach - David Dégardin (MSC)
02:15 - 02:30 7.3 Enabling Better Decisions via The Weather Company’s Probability Forecast Platform - James Belange (Weather Company - IBM)
02:30 - 02:45 7.4 Probabilistic Precipitation Calibration Using Two-parameter Ensemble Model Output Statistics; Xiang Su, (NCAR visitor)
02:45 - 03:00 7.5 The Use of Multi-Model Ensemble Clustering in The Weather Prediction Center’s Extended Range Forecast Experiment - Bill Lamberson (WPC/IMSG)
3:00 - 3:15 Break
Session 8: Panel discussion on product development and user feedback
Discussion topics: see below
03:15 - 04:15 Panelists: Jason Levit (EMC) - Lead, Kathy Gilbert (WPC), Israel Jirak (SPC), Danny Sims (FAA), Bill Myers (Global Weather Corporation)
Adjourn for day
Day three (August 29):
Session 9: S2S prediction and applications
Chair: Matthew Rosencrans
08:30 - 09:00 9.1 S2S overview - Dave DeWitt (CPC; invited)
09:00 - 09:20 9.2 CPC Review - Emerson LaJoie (CPC)
09:20 - 09:35 9.3 GEFS review for subseason - Wei Li (EMC)
09:35 - 09:50 9.4 Seasonal Forecasts: A Shift Towards Probabilistic at the Weather Company, an IBM Business - Michael Ventrice (Weather Company - IBM)
09:50 - 10:05 9.5 Update of GEFS Reanalysis/Reforecast - Hong Guan (EMC)
10:05 - 10:30 Break
Session 10: S2S prediction and applications
Chair: Mark Fresch
10:30 - 11:00 10.1 Review of National River Forecast Services - Ernie Wells (NWS Hydrologic Services Branch)
11:00 - 11:15 10.2 Forecast Informed Flood Management: Folsom Dam Reservoir Operations - Brett Whitin (California/Nevada RFC)
11:15 - 11:30 10.3 The importance of an ensemble forecast for the New York City’s Water Supply – a shift in the operational decision making approach - Adao Matonse (NYC Dept Environmental Protection - Bureau of Water Supply)
11:30 - 11:45 10.4 Ensemble Streamflow Predictions for Risk Based Reservoir Operations of Lake Mendocino in Mendocino County, California - Chris Delaney (Sonoma Water)
11:45 - 12:00 10.5 Enhancing Ensemble Streamflow Forecasting Through an Integrated Multimodel System - Sanjib Sharma (Penn State)
12:00 - 01:30 Lunch and Posters
Session 11: National/international collaboration - Part I
Chair: Yuejian Zhu/Zoltan Toth
01:30 - 02:00 11.1 National ESPC overview - Dave McCarren (Office of the Oceanographer of the Navy; Invited)
02:00 - 02:20 11.2 High-resolution GEFS T1534 ensemble prediction system for better prediction of extremes - Parthasarathi Mukhopadhyay (IITM)
02:20 - 02:40 11.3 The development and operational implementation of GRAPES global ensemble prediction system at CMA - Xiaoli Li (CMA)
02:40 - 03:00 11.4 Ensembles: Dynamically or Statistically Generated? - Zoltan Toth (GSD)
03:00 - 03:30 Break
Session 12: National/international collaboration - Part II
Chair: Huiling Yuan
3:30 - 3:50 12.1 Wave ensemble verification; Benoit Pouliot (ECCC)
3:50 - 4:10 12.2 Predictability of the Meiyu front rainbelt position for the 30 June to 4 July 2016 extreme rainfall period - Jie Ma, (CMA/NMC)
4:10 - 4:30 12.3 Overview of data access to ensemble products from Canadian Meteorological Center; Benoit Archambault, (CMC)
4:30 - 4:50 12.4 The Impact of Major Changes of CMA Global Ensemble on GRAPES; Jing Chen, (CMA/NWP Center)
Adjourn the workshop
Posters
P1. WPC The Application of Ensembles and Fuzzy Clustering to Enhance Winter Weather Predictability - Mike Bodner (WPC)
P2. Performance analysis on Extended-Range Ensemble Prediction over middle and lower reaches of Changjiang River in Meiyu period;- Yong Li (CMA/NMC)
P3. Bias correction of ensemble precipitation forecasts in the improvement of summer streamflow prediction skill - Huiling Yuan (Nanjing University)
P4. The Generation of Site-Specific Guidance Based on the Ensemble Forecast - Chaoping Chen (Sichuang Observation Station, CMA)
P5. An Introduction to Week-2 Temperature Extremes Forecast Guidance of Central Weather Bureau - Hui-Ling Chang (CWB)
Panel Discussion Topics - Day 1
Model development
NOAA is moving towards unifying the NCEP Production Suite with Unified Forecast System (UFS), which is conceived to be fully coupled earth system model with coupled data assimilation for all spatial and temporal scales. The forecasts products will be based on ensembles with associated reanalysis and reforecasts. What are the fundamental issues associated with this approach?
In a unified modeling approach an ensemble ideally is based on a single-core, stochastic physics
/ forcing approach, while creating multi-model ensembles by combining products of different institutes (see Strategic Roadmap). How would users perceive single-model vs. multi-model ensembles?
Subseasonal to seasonal predictions are essentially probabilistic in nature, and hence the need for well calibrated ensembles. What are the primary scientific priorities we should focus on for improved S2S predictions?
Physical processes at the interface of coupled system components play a critical role in the model predictions. How do we improve the representation of uncertainties in model physics?
How do we validate and identify sources of errors with reference to various modes of variability (MJO, PNA, ENSO, NAO, AMO etc.)?
How to better represent uncertainty in the coupled system? What stochastic processes are important to consider in various earth system components?
Reanalyses and Reforecasts play a major role in developing products and also in various downstream applications (e.g., HEFS, NBM etc.). The need for more frequent model upgrades may not allow for producing reanalysis and reforecasts at the same frequency. What strategies can be adopted to address these issues?
Global ensemble systems are approaching resolutions that can match regional ensemble systems. How do we effectively cater to the needs of the users of regional ensembles with global ensembles? In other words, what needs to be done to retire SREF and move users to GEFS based products while we develop Rapid Refresh Forecast System (RRFS)?
Panel Discussion Topics - Day 2
Product development and user feedback
Who is/should be responsible for product development? EMC, MDL, NCEP Centers (for specialized guidance relevant to their mission), others? How much should users, especially non-meteorological ones, be responsible for product development?
Statistical post-processing, certainly the need for high quality and representative observations for the fields we need to forecast
EMC will be consolidating the production suite into eight separate ensemble models using the community-based Unified Forecast System. How should the community develop
post-processed products for users? What are recommendations on the display and distribution of the uncertainty information we can derive from ensemble data?
What type of statistically-post processed data should the community develop for users? What types of investments should be made in observing systems to aid in statistical
post-processing of high-impact fields, such as ocean wind and waves, and snowfall?
GRIB2 is the standard data format for the production suite. What other formats, such as netCDF, should be examined for data distribution?
File sizes are already very large and ensemble output will multiply the sizes even more (considering high resolution output from multiple members) likely overloading current dissemination paths. How can the output from ensembles be reduced while still retaining the necessary and relevant information?
Many users currently rely on deterministic output for decision making; will there be an option for deterministic output? If so, how will it be determined (no pun intended) or will users be “forced” to adapt to ensemble output?