Welcome to Computational Atmosphere Group!We Use Most Advanced Observations and Models to Explore our Planet Earth!
What Science Are We Doing?
Our planet Earth is an extremely complex system comprised of numemous interacting elements that influence one another through nonlinear and dynamical pathways. Earth science has entered an era that scientists need to combine diverse data sources and modeling tools with a system perspective of geospace. The research focuses of Dr. Lu's group are to explore two of the fundamental physical processes in the Sun-Earth system that shape the Earth's atmosphere and geospace, i.e., atmospheric wave coupling from the lower to the upper atmosphere and magnetosphere-ionosphere-thermosphere (MIT) coupling from the topside down. Both processes are universal involving multiscale coupled activities. The group is applying big data, numerical modeling, and cross-disciplinary techniques such as data assimilation and machine learning for the scientific studies.
Atmospheric Wave Coupling
The majority of the Earth's energy originates from the Sun through solar radiation and one of the important ways that Earth reacts is to excite a broad spectrum of atmospheric waves. Some waves are directly generated by differential solar heating, while others are indirectly generated by weather and wind systems forced by solar radiation. The atmospheric waves control the mean state and variability of the middle and upper atmosphere (100 miles above), and transport energy and momentum for thousands of miles. They play important roles in the climate change and space environment. We aim to draw a big picture of atmospheric waves and their roles in shaping the present and future of the Earth's atmosphere.
Underlying the breath-taking scene of aurora are entangled ions and neutrals, collisions of particles, and dynamic interaction across different fields and boundaries. During space storms, tons of energy carried by solar wind are channeled down towards the upper atmosphere through Joule heating and energetic particle precipitation. Dramatic changes occur in space weather. New observations challenge our understanding of the neutral-ion coupling and how magnetospheric energy is deposited and distributed in the ionosphere and thermosphere locally. We aim to understand the high-latitude convection pattern, electrodynamics, and impacts of the structured magnetospheric forcing on the Earth's upper atmosphere. We are interested in exploring the dynamical, chemical, and electrodynamical processes that drive the space weather.
What Tools Are We Using?
Numerical Modeling and Supercomputing
We are running global general circulation models (GCMs) including CESM/WACCM , CESM/WACCM-X , TIEGCM and CTIPe, and wave-dynamics mechanistic models on Clemson's Palmetto cluster. The numerical models excel at studying physical mechanisms and interpreting observations. They can also be used for the prediction of space weather and long-term evolution of the atmosphere. We also develop our own models from the first physical and dynamical principles, and implement new capabilities and modules such as nudging and nested-grid to the sophisticated community models.
Satellite measurements continuously monitor the Earth's environment from a global view and provide critical parameters such as temperature, winds, composition, ion drifts, influx of energetic particles, and magnetospheric disturbances. We use a wide range of satellite data including those from ACE, DMSP, TIMED/SABER, TIME/TIDI, COSMIC, CHAMP and etc., and plan for the new NASA missions such as GOLD and ICON, for various purposes of our research.
Satellite data are used not only to examine exciting atmospheric phenomena, but also to validate and constrain the numerlcal models for both drivers and outputs.
Compared with satellite data, ground-based measurements such as lidar not only provide high-resolution measurements with large vertical extension, but also probe the regions unreachable by other techniques, which enable new science discoveries. For the middle and upper atmosphere, we use the metal (such as Na and Fe) layers originated from the meteor trails as tracers and apply spectroscopy to infer atmospheric temperature, density, and horizontal and vertical winds.
We also use radar data such as ISR and SuperDARN to obtain electron and ion properties and imager data such as temperature mapper and FPI for the 2-D information.
Data Assimiation and Machine Learning for Space Physics
The atmospheric and space science has entered the era of big data. Data assimilation is a powerful way to combine the various data sources with different resolutions, sampling rates, and coverages seamlessly. Our group is currently implementing Lattice Kriging (LK) method to assimilate satellite, ground-based, and empirical models which provide auroral energy and electric field information. The newly assimilated maps provide more realistic high-latitude drivers for the magnetosphere-ionosphere-thermosphere coupling. By adjusting the number of basis functions for different levels, we achieve the multi-resolution data assimilation with most efficient time expense. The machine learning techniques have also been used to untackle the spatial and temporal evolution of ionospheric variabilities and neural network such as ConLSTM has been used for the prediction of space weather.