From 5624e499a0634c1787e018025c4ee1f358cd52c2 Mon Sep 17 00:00:00 2001 From: Xia Sun <58949533+XiaSun-Atmos@users.noreply.github.com> Date: Thu, 30 Jun 2022 09:53:36 -0600 Subject: [PATCH] Add Hierarchical Testing Framework section (#28) * add HTF section * update indexes * Update Intro to reflect HTF work` * Update docs/source/Intro.rst Co-authored-by: ligiabernardet * Update docs/source/HTF.rst Co-authored-by: ligiabernardet * Update docs/source/HTF.rst Co-authored-by: ligiabernardet * Update docs/source/HTF.rst Co-authored-by: ligiabernardet * Update docs/source/HTF.rst Co-authored-by: ligiabernardet * Update docs/source/HTF.rst Co-authored-by: ligiabernardet * Update docs/source/HTF.rst Co-authored-by: ligiabernardet * Add links to AMS abstracts * Add link to ccpp scm * Add ref for CCPP SCM * Update CAD description based on Ligia's comment Co-authored-by: ligiabernardet --- docs/source/HTF.rst | 27 +++++++++++++++------------ docs/source/Intro.rst | 4 +++- 2 files changed, 18 insertions(+), 13 deletions(-) diff --git a/docs/source/HTF.rst b/docs/source/HTF.rst index 3ee30f3..8458dc2 100644 --- a/docs/source/HTF.rst +++ b/docs/source/HTF.rst @@ -8,44 +8,47 @@ Hierarchical Testing Framework ===================================== -Hierarchical Testing Framework (HTF) refers to test physics innovations using tools ranging in complexity from simple parameterization simulators to complex three-dimensional global models. Unified Forecast System (UFS) Case Studies Platform (Sun et al., 2021) provides a set of cases that reveal the forecast challenges of NOAA’s operational numerical global weather prediction model, the Global Forecast System (GFS). Our catalog of test cases currently contains 10 cases that cover known biases of the GFS operational implementations GFS v15.2 and GFS v16, such as hurricanes, mid-latitude storms, heat waves and cold blast events, and a cold air damming case. We used the 2020 July Convective Available Potential Energy (CAPE) case and the 2020 Cold Air Damming (CAD) cases to perform hierarchical testing of physics innovations. The results were presented to the community at the American Meteorological Society annual meetings (Sun et al., 2022; Pan et al., 2022). +In the context of development of physical parameterizations and suites, Hierarchical Testing Framework (HTF) refers to testing physics innovations using tools ranging in complexity from simple parameterization simulators to complex three-dimensional global models. The Unified Forecast System (UFS) Case Studies Platform (Sun et al., 2021) provides a set of cases that reveal the forecast challenges of NOAA’s operational numerical global medium-range weather prediction model, the Global Forecast System (GFS). Our catalog of cases contains cases that illustrate known biases of the GFS operational implementations GFS v15.2 and GFS v16, such as hurricanes, mid-latitude storms, heat waves and cold blast events, as well as a cold air damming case. We used the 2020 July Convective Available Potential Energy (CAPE) and the 2020 Cold Air Damming (CAD) cases to perform hierarchical testing of physics innovations. The results were presented to the community at the 2022 American Meteorological Society annual meeting (Sun et al., 2022; Pan et al., 2022). .............................. 2020 July CAPE Case .............................. -For the 2020 CAPE case, we exercised the HTF to investigate the factors that lead to the low CAPE bias. This case was initialized on July 23, 2020 using GFS v15.2 and GFS v16. A large CAPE reservoir appeared over the Great Plains with a warm front passage over the northern Great Plains and no significant synoptic pattern impacting the Southern Great Plains (SGP). A comprehensive evaluation of GFS forecasts was made, focusing on the SGP and verifying against the observational dataset from the Atmospheric Radiation Measurement network. We investigated the large-scale advection fields over a 120 km × 120 km region centered at the SGP central facility site using European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5). Experiments were conducted using the Common Community Physics Package (CCPP) Single Column Model (SCM) to probe the impacts of initial conditions (IC) and surface forcing. +For the 2020 CAPE case, we exercised the HTF to investigate the factors that lead to the low CAPE bias. This case was initialized on July 23, 2020 using GFS v15.2 and GFS v16. A large CAPE reservoir formed over the Great Plains with a warm front passage over the northern Great Plains and no significant synoptic pattern impacting the Southern Great Plains (SGP). A comprehensive evaluation of GFS forecasts was made, focusing on the SGP and verifying against the observational dataset from the Atmospheric Radiation Measurement (ARM) network. We investigated the large-scale advection fields over a 120 km × 120 km region centered at the ARM SGP central facility site using European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5). Experiments were conducted using the Common Community Physics Package (CCPP) Single Column Model (SCM) to probe the impacts of initial conditions (IC) and surface forcing. Key findings for the CAPE case are listed below: - The lower CAPE in GFS is attributed to air at low levels being drier than observed. - The larger-than-observed Bowen ratio in GFS suggests an incorrect surface energy partitioning associated with soil moisture and soil temperature. - GFS v16 exhibits more vertical mixing leading to a deeper boundary layer than v15.2 and observations. -- The drier air in GFSv16 is mostly related to local physics not large-scale advection in this case. +- The drier air in GFSv16 is mostly related to local physics and not large-scale advection in this case. - With realistic IC of vertical profiles and surface fluxes, the GFS v16 physics suite is able to produce the large CAPE. .............................. 2020 Cold Air Damming .............................. -We adopted the latest version of Short-Range Weather (SRWeather) App (version 1.0) to investigate the 2020 CAD case1 developed as part of the UFS Case Studies project to investigate the impact of code updates, horizontal and vertical resolution, and physics suites. The SRWeather App is the foundation for building NOAA’s future convection-allowing Rapid Refresh Forecast System (RRFS). The limited area configuration needs fewer computer resources and can run models in high resolution and cycle frequently. Thus the case studies conducted with the SRWeather App are a useful supplement for the global configuration case studies. The physics suites used in this study, which focuses on the CONUS domain, include GFS v15.2, GFS v16, GFS v17α, and RRFS v1α. The results from UFS SRWeather App were compared with the results from the UFS global configuration. The SRWeather App was also tested with different configurations, such as different versions of the code base, model grid spacing (13 km vs 3 km, 64 levels vs 127 levels), and physics suite employed. The model forecasts were verified against station observations and analysis data. +We adopted the Short-Range Weather (SRWeather) App v 1.0 to investigate the 2020 CAD case to investigate the impact of code updates, horizontal and vertical resolution, and physics suites. The SRWeather App is the foundation for building NOAA’s future convection-allowing Rapid Refresh Forecast System (RRFS). The limited area configuration requires fewer computer resources so that the model can be run in high resolution and be cycled frequently. Thus the case studies conducted with the SRWeather App are a useful supplement for the global configuration case studies. The physics suites used in this study, which focuses on the CONUS domain, include GFS v15.2, GFS v16, GFS v17α, and RRFS v1α. The results from UFS SRWeather App were compared with the results from the UFS global configuration. The SRWeather App was also tested with different configurations, such as different versions of the code base, model grid spacing (13 km vs 3 km, 64 levels vs 127 levels), and physics suite employed. The model forecasts were verified against station observations and analysis data. The key findings for the CAD case are: - The spatial pattern of the 2-meter surface temperature field obtained with the SRWeather App is slightly improved over the one obtained with the global configuration. -- Using the code version of April 10 th, 2022 and more vertical levels leads to -- Improvements for this case when using the GFS v16 suite -- Degradation when using the (uncoupled) GFS v17α suite -- When the 3-km horizontal grid spacing configuration is used with the GFS v15.2 and GFS v16 suites (but not with the GFS v17α suite), the cold air is not maintained and the CAD strength is reduced, which is a degradation when compared to experiments conducted at lower resolutions. for -- Sensitivity tests with suite GFS v17α show that changing the land surface model from NoahMP to Noah can cause a northward shift of the cold air. +- Using the April 10th, 2022 code version and more vertical levels leads to + + - Improvements for this case when using the GFS v16 suite + - Degradation when using the (uncoupled) GFS v17α suite + +- When the 3-km horizontal grid spacing configuration is used with the GFS v15.2 and GFS v16 suites (but not with the GFS v17α suite), the cold air is not maintained and the CAD strength is reduced, which is a degradation when compared to experiments conducted at lower resolutions. + +- Sensitivity tests with suite GFS v17α show that changing the land surface model from NoahMP to Noah can cause a northward shift of the cold air, which leads to a warmer 2-m temperature bias and a weaker CAD signal. **References** -Pan, L., Heinzeller, D., Bernardet, L., Sun, X., & Brown, J. M. (2022, January). Surface biases in forecasting extreme events using the Unified Forecast System with different physics suites. 31st Conference on Weather Analysis and Forecasting. +Pan, L., Heinzeller, D., Bernardet, L., Sun, X., & Brown, J. M. (2022, January). Surface biases in forecasting extreme events using the Unified Forecast System with different physics suites. 31st Conference on Weather Analysis and Forecasting. [`Link `_] -Sun, X., Heinzeller, D., Bernardet, L., Pan, L., & Brown, J. M. (2021, January). Case studies that exemplify known biases of the Unified Forecast System (UFS) Weather Model. 11th Conference on the Transition of Research to Operations. +Sun, X., Heinzeller, D., Bernardet, L., Pan, L., & Brown, J. M. (2021, January). Case studies that exemplify known biases of the Unified Forecast System (UFS) Weather Model. 11th Conference on the Transition of Research to Operations. [`Link `_] -Sun, X., Heinzeller, D., Bernardet, L., Pan, L., & Brown, J. M. (2022, January). Confronting the low summer CAPE behavior in GFSv16. 12th Conference on the Transition of Research to Operations. +Sun, X., Heinzeller, D., Bernardet, L., Pan, L., & Brown, J. M. (2022, January). Confronting the low summer CAPE behavior in GFSv16. 12th Conference on the Transition of Research to Operations. [`Link `_] Sun, X., Heinzeller, D., Bernardet, L., Pan, L., Li, W., Turner, D., & Brown, J. M. On the factors leading to the low surface-based CAPE behavior in the Global Forecast System. In preparation. diff --git a/docs/source/Intro.rst b/docs/source/Intro.rst index 0b665be..1bb6ef1 100644 --- a/docs/source/Intro.rst +++ b/docs/source/Intro.rst @@ -64,11 +64,13 @@ The goal of this ongoing effort is to provide the community, as well as the phys The case catalog in this documentation was created based on the known biases of GFS model. This is a list that we are updating diligently. Please come back to check updates anytime. -In addition, we will use the Hierarchical Testing Framework (HTF) tools, including the Single Column Model and case studies provided on this website, to document strengths and weakness of current UFS physics suites, and to inform future model development. +In addition, we will use Hierarchical Testing Framework (HTF) tools, including the Common Community Physics Package (CCPP) Single Column Model (Firl et al., 2022) and case studies provided on this website, to document strengths and weakness of current UFS physics suites, and to inform future model development. **References** +Firl, G., D. Swales, L. Carson, L. Bernardet, D. Heinzeller, and M. Harrold, 2021. Common Community Physics Package Single Column Model v6.0.0 User and Technical Guide.Firl, G., D. Swales, L. Carson, L. Bernardet, D. Heinzeller, and M. Harrold, 2021. Common Community Physics Package Single Column Model v6.0.0 User and Technical Guide. [`Link `_] + Stan C., Yang F., and Harris L. (2019). UFS Development Goals and Priorities for Medium-Range and S2S Applications. *Unified Forecast System Community*. [`Link `_] Yang F. (2020). Development and evaluation of NCEP's Global Forecast System Version 16. *Unified Forecast System Community Webinar*, Oct 22, 2020. [`Link `_]