ETS in Neighborhood context PQPF Verification Localized Probability Match (LPM) LPM vs PM West Virginia Flash Floods In 2016 and 2017, CAPS continued to play major role in the NOAA Hazardous Weather Testbed (HWT) Spring Forecast Experiment, and expanded into contributing to the HMT Flash Flood and Intensive Rainfall (FFaIR) Experiment by producing convection-allowing storm-scale ensemble forecasts (SSEF) over the entire CONUS domain at 3-km grid spacing. A FV3 forecast was added in 2017. The CAPS SSEF is a large part of a Community Leveraged Unified Ensemble (CLUE) “grant ensemble” for the HWT SFE2016 and SFE2017, coordinated among various groups including NSSL, SPC, CAPS, NCAR, UND, EMC, GSD, and DTC, in an effort to provide guidance to the design of near-future operational SSEF systems for EMC. The NSSL Multi-Radar/Multi-Sensor (MRMS) QPE (Zhang et al. 2011) data were used as verification dataset for the SSEF QPF. This work is primarily funded by the NOAA CSTAR program, HWT and HMT grants. 2017 CAPS SSEF Highlight • 3-km horizontal grid spacing (1620×1120) • WRF version 3.8.1 (coupled with ARPS v5.4) • HWT: A 3DVAR-based SSEF with 23 ARW members, initiated with 3DVAR radar analysis at 0000 UTC, running on Lonestar5 at TACC; 1 FV3 member • HWT: A GSI+EnKF-based ensemble forecast with 10-member ensemble forecast starting at 0000 UTC. Running on Bridges at PSC • HMT: A 3DVAR-based SSEF with 10 ARW members, running on Stampede2; plus a FV3 member, running on OSCER’s Schooner Storm-Scale Ensemble Forecast for the NOAA HWT Spring Forecasting Experiment and HMT FFaIR Experiment Fanyou Kong 1 , Ming Xue 1,2 , Youngsun Jung 1 , Keith Brewster 1 , Nathan Snook 1 1 Center for Analysis and Prediction of Storms and 2 School of Meteorology, the University of Oklahoma, Norman, OK 73072 Forecast domain for the 2017 HWT SFE and HMT FFaIR (thick inner box is the verification sub-domain) . CONUS domain (1620x1120, at 3-km) For HMT, the following settings are used based on preliminary experiments: – Local patch size: 5 × 5 grid points – Calculation area size: 65 x 65 grid points – Gaussian smoother standard dev.: 1 grid point ETS of 3-h accumulated precipitation >= 0.01 inch (top) and 0.5 inch (bottom) averaged over all 2017 3DVAR SSEF forecasts initiated at 0000 UTC. ETS of 3-h accumulated precipitation >= 0.5 inch, averaged over all 2017 3DVAR SSEF forecasts initiated at 0000 UTC, with R=8 referring to a neighborhood scale in grid points. AUROCs for the ensemble probability of 3-h accumulated precipitation≥0.25, 0.5, and 1.0 inch with r=8 (~25km). AUROCs for the ensemble probability of 3-h accumulated precipitation≥0.25, 0.5, and 1.0 inch with r=0. Contact Email: [email protected] ETS of 3-h accumulated precipitation ≥0.01 (left), and ≥0.5 inch (right), with different neighborhood scale values.
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
ETS in Neighborhood context
PQPF Verification
Localized Probability Match (LPM)
LPM vs PM
West Virginia Flash Floods
In 2016 and 2017, CAPS continued to play
major role in the NOAA Hazardous Weather
Testbed (HWT) Spring Forecast Experiment,
and expanded into contributing to the HMT
Flash Flood and Intensive Rainfall (FFaIR)
Experiment by producing convection-allowing
storm-scale ensemble forecasts (SSEF) over
the entire CONUS domain at 3-km grid
spacing. A FV3 forecast was added in 2017.
The CAPS SSEF is a large part of a
Community Leveraged Unified Ensemble
(CLUE) “grant ensemble” for the HWT
SFE2016 and SFE2017, coordinated among
various groups including NSSL, SPC, CAPS,
NCAR, UND, EMC, GSD, and DTC, in an
effort to provide guidance to the design of
near-future operational SSEF systems for
EMC.
The NSSL Multi-Radar/Multi-Sensor (MRMS)
QPE (Zhang et al. 2011) data were used as
verification dataset for the SSEF QPF.
This work is primarily funded by the NOAA CSTAR
program, HWT and HMT grants.
2017 CAPS SSEF Highlight
• 3-km horizontal grid spacing
(1620×1120)
• WRF version 3.8.1 (coupled with ARPS v5.4)
• HWT: A 3DVAR-based SSEF with 23 ARW members, initiated with 3DVAR radar analysis at 0000 UTC, running on Lonestar5 at TACC; 1 FV3 member
• HWT: A GSI+EnKF-based ensemble forecast with 10-member ensemble forecast starting at 0000 UTC. Running on Bridges at PSC
• HMT: A 3DVAR-based SSEF with 10 ARW members, running on Stampede2; plus a FV3 member, running on OSCER’s Schooner
Storm-Scale Ensemble Forecast for the NOAA HWT Spring Forecasting Experiment and HMT FFaIR Experiment
Fanyou Kong1, Ming Xue1,2, Youngsun Jung1, Keith Brewster1, Nathan Snook1 1Center for Analysis and Prediction of Storms and 2School of Meteorology, the University of Oklahoma, Norman, OK 73072
Forecast domain for the 2017 HWT
SFE and HMT FFaIR (thick inner box
is the verification sub-domain) .
CONUS domain (1620x1120, at 3-km)
For HMT, the following settings are used based on
preliminary experiments:
– Local patch size: 5 × 5 grid points
– Calculation area size: 65 x 65 grid points
– Gaussian smoother standard dev.: 1 grid point
ETS of 3-h accumulated precipitation >=
0.01 inch (top) and 0.5 inch (bottom)
averaged over all 2017 3DVAR SSEF
forecasts initiated at 0000 UTC.
ETS of 3-h accumulated precipitation >=
0.5 inch, averaged over all 2017 3DVAR
SSEF forecasts initiated at 0000 UTC, with
R=8 referring to a neighborhood scale in
grid points.
AUROCs for the
ensemble probability
of 3-h accumulated
precipitation≥0.25,
0.5, and 1.0 inch with
r=8 (~25km).
AUROCs for the
ensemble probability
of 3-h accumulated
precipitation≥0.25,
0.5, and 1.0 inch with
r=0.
Contact Email: [email protected]
ETS of 3-h accumulated precipitation ≥0.01
(left), and ≥0.5 inch (right), with different
neighborhood scale values.
Related Documents
