Along with advancements in the observing system, computer power, and high-resolution NWP model simulations and ensemble forecasting, NWP becomes a big data problem. Our research is expanding to explore the applications of artificial intelligence, especially machine learning, in NWP and related areas. We currently have projects collaborated with engineering interdisciplinary.
- Big Data and Machine Learning in NWP
Accurate forecasts of the intensity and structure of a hurricane at landfall can save lives and mitigate social impacts. However, among recent efforts in hurricane forecast improvements, few studies have focused on landfal
ling hurricanes. This reflects the complexity of predicting hurricane landfalls and the uncertainties in repres
enting the atmospheric boundary layer conditions in numerical weather prediction (NWP) models. The aim of our research is to study the interaction between landfalling hurricanes and the atmospheric boundary layer using ensemble-based data assimilation.
We incorporate the airborne Doppler radar and ground-based NEXRAD observations with other available in-situ and satellite data into an ensemble data assimilation system with the community mesoscale weather research and forecasting (WRF) model to address the related science questions.
- The interaction between landfalling hurricanes and the atmospheric boundary layer
Taking advantage of Utah's location in Intermountain West, our study focuses the predictability of flows over mountainous terrain at mesoscale, in particular, the error growth (i.e., the sensitivity to initial conditions at various lead times), and develop meaningful measures of skill relative to appropriate conditional climatologies (i.e., the skill of capturing specific phenomena when they are supposed to appear; e.g., turbulence generation when a Richardson number criterion is satisfied). Specifically, our research emphasizes on investigating the sensitivities of model forecasts to input properties (initial conditions and model parameters) and boundary conditions, data assimilation, and comparison of different techniques (e.g., 4DVAR, ensemble Kalman filtering, 3DVAR) for their abilities in analyses and forecasts over the regions of complex terrain.
- Mountain Terrain Atmospheric Modeling and Observations
In addition, we explore better understanding, observing, and prediction of cold fog over complex terrain.
The principal objective of our research in this topic is to create realistic estimates of high-resolution (1 km by 1 km horizontal grids) atmospheric boundary layer structure and the characteristics of precipitating convection, including updraft and downdraft cumulus mass fluxes and cold pool properties over a region the size of a GCM grid column from analyses that assimilate the surface mesonet observations of wind, temperature, and water vapor mixing ratio and available profiling data from single or multiple surface stations using advanced data assimilation methods.
- Cloud life cycles - cloud permitting scale and large-eddy simulations
In addition, we also use high-resolution numerical simulations with data assimilation to study the interaction between mesoscale convective system and MJO.
The objective of our study is to investigate large-scale environmental conditions, mesoscale phenomena and small scale convective bursts as well as their interactions that are responsible for TC formation and intensity changes. Specific areas include 1) Characterize the intensity of convection over the western Pacific and Atlantic oceans from radar, aircraft and satellite data; 2) Derive an accurate mesoscale environment of convective systems through the assimilation of satellite, radar, lidar and in-situ data; 3) Evaluate the quality of the global forecast system (e.g., NCEP and NOGAPS ensembles) for accurate TC analyses and forecasts; 4) Understand the environmental factors that determine tropical cyclone formation and rapid intensification.
Past Ph. D. and M. S. Students and Research Topics (years denote the graduation years)
- Tropical cyclone formation and rapid intensification
- L. Xu, 2007: Seasonal variations of snow cover over the Tibetan Plateau and ultimate effects on the East Asian summer monsoon. M. S. Thesis.
- X. Li, 2008: High-resolution numerical simulations of tropical cyclone intensity change with assimilation of satellite, radar and in-situ data. Ph.D. Dissertation.
- A. Snyder, 2009: Tacking and verification of tropical cyclone development in global ensemble prediction systems: Evaluations during recent field programs. M. S. Thesis.
- L. Thatcher, 2010: How vertical wind shear affects the rapid intensification of Typhoon Jangmi (2008). M.S. Thesis.
- M. Ma, 2011: Numerical simulations of the extreme atmospheric boundary layer heights over Northwest China and their impacts on chemical tracer transportations. Ph.D. Dissertation.
- L. Wei, 2012: Numerical simulation of wintertime inversion and data assimilating of late spring convection. Ph.D. Dissertation.
- X. Zhang, 2013: Practical implementation of an ensemble transform Kalman filter.
M. S. Project. Computational Engineering Science Program.
- L. Thatcher, 2013: High-resolution ensemble error growth and dimensionality in tropical cyclone genesis environments. Ph.D. Dissertation.
-  Z. Li, 2013: Studying the genesis of Typhoon Nuri (2008) with high-resolution numerical simulations and data assimilation. Ph.D. Dissertation.
- H. Zhang, 2014: Ensemble Kalman filter data assimilation in regions of complex terrain. Ph.D. Dissertation.
- C. Lin, 2014: Evaluation of double-moment representation of warm-rain and Ice
hydrometeors in bulk microphysical parameterization. M. S. Thesis.
- C. Chachere, 2016: Cold season inversion in Salt Lake City: Connections to valley
variables and numerical simulations. M. S. Thesis.
- Y. Yu, 2016: Evaluation of tropical cyclone forecasts from a global model and
comparison with regional mesoscale numerical simulations of Hurricane Joaquin. M. S.
- N. Hock, 2017: Numerical simulations and sensitivity studies of a Florida sea breeze and
its associated convection in the gray-zone spacing. M. S. Thesis.
- C. Yu, 2017: The impact of assimilation of GPM clear-sky radiance on hurricane
forecasts with HWRF model. M. S. Thesis.
-  F. Zhang, 2017: Influences of boundary layer vertical mixing and land surface
parameterizations on numerical simulations of landfalling hurricanes. Ph.D. Dissertation.
-  S. Zhang, 2017: Improving hurricane vortex initialization and prediction through
inner-core data assimilation with ensemble-variational hybrid methods. Ph.D. Dissertation.
-  P. Saunders, 2019: Studying the sudden onset and evolution of outer
precipitation of Hurricane Harvey (2017) using numerical simulations
with data assimilation and cloud initiation. Ph.D. Dissertation.
-  D. Hodges, 2020: Interactions between low-level jets and precipitation extremes with
climatology, mesoscale numerical simulations, and ensemble-based non-Gaussian data
assimilation. Ph.D. Dissertation.
-  X. Li, 2020: Characteristics and effects of turbulent structure in the atmospheric boundary
layer driven by weak and strong forcings. Ph.D. Dissertation.
The group uses the community Weather Research and Forecasting (WRF) model and its data assimilation systems (e.g., 3DVAR, 4DVAR, EnKF, DART, GSI) as basic research tools. Major research efforts in most recent activities involve the use of various types of satellite and airborne observations from NASA, NOAA and research communities. The group also uses the research version of NCEP operational models such as HWRF, GFS, FV3GFS and 3DEnVAR and 4DEnVAR data assimilation system as research tools.
The research group maintains an excellent track in the areas of atmospheric data assimilation and mesoscale numerical simulations. Recent developments are highlighted by assimilation of satellite and radar data, high-resolution numerical simulations of the hurricane and frontal systems, model validation and satellite data analysis. Most recent research efforts also included couple land-atmospheric data assimilation, atmospheric boundary layer, numerical simulation of high-impact weather systems (e.g., landfalling hurricanes, mesoscale convective systems, fog and winter storms), as well as interactions between tropical convection and MJO. We have the broad spectrum of research areas with continuous growth.
Our research laboratory is well-equipped by personal computers, four workstations (each has > 16 CPUs and > 64GB memory); and >500 TB hard disk spaces, as well as our own Linux cluster nodes that are available all the time for our group members. As of fall 2018, our group has 24 cluster nodes with a total of 512 CPUs to support our research needs. With the availability of these computer nodes, we have recently built up a real-time forecast capability. The group members also have access to the large Linux cluster (with more than 3000 CPUs) from the University of Utah's Center for High-Performance Computing(CHPC). Also, the group gets external computing recourses through NOAA, NASA, NSF, and other government agencies.