NSC projects November 2017 - October 2018

Global and regional convection-permitting ensemble forecasts with NCAR’s MPAS model

Project lead: Craig Schwartz, MMM
Cheyenne allocation: 15.1 million core-hours

Convection-permitting ensembles (CPEs) have repeatedly been shown to benefit extreme weather prediction. However, CPEs are computationally expensive and have only been run over limited-area (regional) domains. Thus, lateral boundary conditions (LBCs) provided by coarser-resolution models are a necessary component of CPEs. Nearly all CPEs have insufficient spread, with “truth” regularly occurring outside CPE envelopes. Part of this spread deficiency may be related to lateral boundaries, as errors that grow on the high-resolution grid eventually leave the domain and are replaced by coarser-resolution, larger-scale errors at the boundaries, which infiltrate the domain and may lack characteristics needed to achieve appropriate CPE spread. This work will examine whether global CPEs, which are free of constraints imposed by lateral boundaries, have more appropriate spread than regional CPEs. The global CPE forecasts will be produced by NCAR’s Model for Prediction Across Scales (MPAS) and compared to limited-area CPE forecasts produced by the regional version of MPAS currently under active development. As both the global and regional CPEs will use identical dynamics and physics, differences between the forecast sets can be attributed to LBC impacts. Five-day forecasts will be produced and objectively evaluated to determine whether global CPEs yield improved performance compared to regional CPEs. The global CPE forecasts will be unprecedented and lead to new knowledge regarding CPE design. Moreover, the 5-day forecasts will provide insight regarding reliability and predictability of medium-range CPE forecasts of extreme weather. Finally, through collaboration with MPAS developers, this work will accelerate the development of MPAS’s regional capability.

Collaborative Research: Near-surface mixing driven by breaking internal tides: thermal refugia for coral reefs in the western tropical Pacific.

Project lead: Scott Bachman, CGD
Cheyenne allocation: 15 million core-hours

Internal waves are an important mechanism for vertical energy transport in the ocean. The earth's tides are a major source of these waves, particularly in areas of complex bathymetry and steep topographic features such as trenches, sills, and seamounts. Observations indicate that only 2-20% of the wave energy generated in these locations actually breaks near the topography, implicating the tides as a primary mechanism leading to mixing in the mid- and upper ocean. Recent work suggests this mixing may play a role in reducing surface ocean temperatures in reef environments, reducing heat stress and coral bleaching. Numerical study of these effects is a significant challenge, however: very high resolution is needed to resolve the internal wave spectrum, but within a sufficiently large domain to permit tide-bathymetry interactions. The study proposed here aims to address both of these issues using a 500 m resolution model of the Coral Triangle (CT) region. This study will build on previous research which highlighted the role of the internal tides in the CT, but the substantially higher (10x) resolution used here will resolve the wave spectrum in unprecedented detail. A recently developed technique for diagnosing the wave energy spectrum will clarify the tide's role in mixing in this region, while providing novel insight into the propagation of these waves throughout the water column. Lastly, the results of this study will help to identify potential thermal refugia sites within the CT, where the reef environment may be better conditioned to survive the effects of climate change.

Addressing model uncertainty through stochastic parameter perturbations within the Thompson microphysics scheme as part of a High Resolution Rapid Refresh ensemble

Project lead: Jamie Wolff, RAL
Cheyenne allocation: 13 million core-hours

In many regional ensemble NWP systems, model-related uncertainty is addressed by using multiple models, dynamic cores, physics suites, or a combination of these methods. While these approaches have demonstrated potential, maintenance of such systems is time consuming, particularly in an operational environment. In an effort to move toward a more sustainable and unified system (that also aligns with NCEP goals), the Developmental Testbed Center (DTC) proposes continued testing of stochastic parameter perturbations within physics parameterizations. Based on promising results from previous studies, the DTC Regional Ensemble Task is proposing continued testing of stochastic physics in 2017. Research will focus on implementing SPP within the Thompson microphysics scheme (Thompson and Eidhammer 2014). In particular, we will apply the SPP method to treat known parameter uncertainties within the hybrid graupel/hail category of the scheme. Two particular aspects will be investigated: (a) the relationship used to specify the Y intercept parameter of the assumed inverse exponential size distribution, and (b) the shape factor of the generalized gamma size distribution of graupel/hail particles.

Linking full chemistry to CAM-SE and the effects of resolution and improved chemistry on observed CESM2 model bias

Project lead: Forrest Lacey, ACOM
Cheyenne allocation: 11.2 million core-hours

This project focuses on using the Community Earth System Model to explore the effects of varying the model resolution and the increasingly complex chemical mechanisms on model bias when compared to a suite of observations; particularly focused on secondary organic aerosols and ozone. By running different configurations of the model including horizontal resolutions ranging from 1.9 degrees all the way down to 0.25 degrees and increasing the number of chemical species tracked in the model to approximately 360, we are able to develop an understanding of how each specific change impacts model bias, comparing to flight campaigns in the years 2013 and 2016. These model outputs will then be used to identify regions for which the resolution and chemical complexity is critical, which will then be tested using a regionally refined version of the Community Atmosphere Model, Spectral Element to confirm the reduction in model bias in a more computationally efficient manner.

Modeling impacts of winds on ocean submesoscales and mixed layer dynamics

Project lead: Daniel Whitt, CGD
Cheyenne allocation: 10.6 million core-hours

An accurate representation of small-scale processes in the upper ocean, which must be parameterized in regional and global models, is crucial to the grand challenge of achieving an accurate prediction of the consequences of natural and anthropogenic climate variability and change at regional and global scales. Here, we propose ocean large eddy simulations that resolve the interactions between boundary layer turbulence (1-100 m) and submesoscales (.1-10 km) to obtain the model data necessary to update the Fox-Kemper et al. 2008 parameterization to include the dependence of submesoscale upper-ocean vertical fluxes on atmospheric forcing in regional to global ocean models (like MOM). The proposed simulations are motivated by preliminary results Whitt and Taylor 2017, which show that winds can be the dominant energy source for submesoscales, in contrast to Fox-Kemper, in which submesoscales obtain energy exclusively from larger scale oceanic flows. In this scenario, the submesoscale buoyancy flux <w'b'> ~10x stronger than predicted by Fox-Kemper.

Climate and air quality impacts including an interactive fire model for future climate scenarios with and without geoengineering

Project lead: Simone Tilmes, ACOM/CGD
Cheyenne allocation: 9.5 million core-hours

Changes in climate are expected to impact soil moisture and precipitation, and therefore the occurrence and strength of fires. However, interactive fire emissions are currently not included in the standard CMIP6 versions of CESM2 and CESM2(WACCM). For past and present-day simulations fire emissions are currently prescribed based on inventories derived from observed fire counts or satellite observations. For future scenarios, biomass burning emissions are uncertain, especially for projected geoengineering scenarios using stratospheric aerosol modifications. The slowing in the hydrological cycle under solar dimming has potentially large impacts on fire activity and consequently on aerosol burden and chemistry. This may significantly impact air quality, which will in turn affect crops and therefore impact agriculture. Using CESM2(WACCM) with interactive fire injections and CLM5 with an interactive crop model provides the first opportunity to explore the interactions between changes in fire emissions, air-quality and agriculture in future scenarios with and without proposed solar geoengineering approaches. The proposed computer time would allow us to test the interactive fire model within CESM2 for a transient simulation between 1850-2014. Furthermore, we plan to use CESM2(WACCM) to assess the ability of the interactive fire model to reproduce observed chemistry and aerosol distributions for present day, using satellite and in-situ observations. Finally, we plan to perform future climate simulations with CESM2(WACCM) that are part of a new Geoengineering Model Intercomparion Project (GeoMIP) Tier 1 experiment to analyze the importance of an interactive fire model for impacts on air-quality and agriculture.

Turbulent flow over two- and three-dimensional forested hills

Project lead: Edward Patton, MMM
Cheyenne allocation: million core-hours

This project proposes to perform a number of turbulence-resolving simulations over canopy-covered two- and three-dimensional canopy-covered low hills. Analysis of these simulations aims to extend current theory of flow over two-dimensional canopy-covered hills to three-dimensional canopy-covered hills. The new theory targets parameterization of orographic drag imparted by low forested hills from an understanding of the physical processes creating the drag toward improved weather and climate prediction.