Boulder, Colorado, USA, 22-28 June 2013
By Alex Chartier SSI Fellow and PhD Candidate, Invert centre for imaging science, University of Bath
I met a range of people from the USA's Space Weather community, notably Tony Mannucci (JPL), Ludger Scherliess (Utah State), Tom Gaussiran (UT Austin), Lars Dyrud (Draper), Aaron Ridley (UMich) and Gang Lu (NCAR). I took part in a collaborative Observation System Simulation Experiment planning workshop and agreed to work in collaboration with various groups to make joint comparisons of our different experimental approaches.
I found out about a new scheme to make low-cost ionosondes (ionospheric measurement instruments) using software radio technology. I plan to investigate the use of these instruments for ionospheric imaging in the
algorithms I currently work with. I hope to keep in touch with Juha Vierinen of MIT Haystack and Sixto Gonzalez of Arecibo Incoherent Scatter Radar to make a collaborative paper on the subject.
I was previously unaware of the inclusion of the UMich model, called Global Ionosphere Thermosphere Model in the Data Assimilation Research Testbed. I look forward to collaborating with that group to see what progress we both make using the same algorithm but different models.
In general groups such as the Software Sustainability Institute should focus on helping scientists from different groups and related fields make their models more modular. For example, in a project making a chain of models to represent the Sun-Earth system, it should be easy for a user to swap in/out a different model from the chain to see if it helps the overall results or not. This will allow the overall project to proceed without being held up by failures in a specific area, or being unable to benefit from advances made by a group that was not originally included.
Coupling, Energetics and Dynamics of Atmospheric Regions (CEDAR) is a focused Global Change program sponsored by the USA's National Science Foundation. The primary objective of CEDAR is to develop scientific understanding of the upper atmosphere. The current phase of CEDAR is designed to explore coupling with lower altitudes, solar-terrestrial interactions, polar aeronomy and long-term variations. The CEDAR research program began in 1986 as a grass-roots community initiative. CEDAR meetings have taken place annually for a number of years, with this year's meeting held in Boulder, Colorado and next year's meeting planned for Seattle, Washington. This year's meeting had several hundred attendees and included future planning workshop sessions as well as scientific presentations and extended tutorials. Videos of the key presentations are available on the CEDAR website.
Weekend Session 22-23 June 2013
On the weekend before the CEDAR meeting, a joint session was held between CEDAR and the Geospace Environment Modelling (GEM) program, which had their meeting the previous week. This session was an opportunity for CEDAR and GEM to collaborate and included a number of student tutorial sessions on the
basics of both CEDAR and GEM science.
Roger Varney (NCAR HAO) gave a summary of ionospheric physics. The description began with the temperature and composition of the neutral atmosphere, leading to a discussion of ion production and recombination rates. The higher collisional cross-section of ions compared to electrons leads to much hotter electrons, because they cannot lose their heat to the neutral particles. The discussion of ionospheric energetics led on to a series of ionospheric temperature profiles. In addition to the dynamic pressure, from the neutral species, and the thermal pressure, the magnetic pressure (due to the magnetic field) was introduced. The auroral zone, or cusp region, was introduced as the boundary between open and closed magnetic
field lines. The concept of ExB drift was introduced, as was the Boltzmann equation.
Ethan Miller (JHU/APL) gave an introduction to modelling in aeronomy. Aeronomy is fundamentally an observational science, since the discovery of the ionosphere by Marconi. However, the problem is starved of observations, so people began to create models. The presenter described empirical, first-principles and assimilative models as different classes of model. Issues such as emergence and chaos were discussed with reference to the two-week limit of predictability for tropospheric weather.
Aaron Ridley (Umich) described a number of technical difficulties associated with modelling the ionosphere and thermosphere. Problems included discretizing a continuous problem, grid choices (how to avoid the pole
problem?), assumptions of the balance between gravity and pressure and solving in time – either ignore short time-scale processes, assume things happened at the previous time, or introduce extra complexity but solve
properly at the current time. Every model has assumptions, each one includes or rejects processes based on how important they are perceived to be. Boundary conditions pose a serious problem.
Ludger Scherliess (Utah State) gave a history of data assimilation approaches for upper atmospheric studies. He highlighted AMIE, by Richmond and Kamide (1988), Kalman Filter and 3DVAR approaches to ionospheric and thermospheric data assimilation and data assimilation for the radiation belts.
Tony Mannucci (JPL) gave a description of different data assimilation approaches, emphasising the need for good system boundaries – we need to update the parts of the state that are important to the forward progression of the system.
Joe Huba (NASA Goddard) described his model (SAMI2), which is an open-source ionospheric modelling project.
Astrid Maute (NCAR HAO) introduced the open-source TIEGCM model of the thermosphere and ionosphere.
Monday to Friday
Jorge Chau (Jicamarca Radio Observatory) gave the 2013 CEDAR prize lecture on the puzzling phenemenon of 150 km echoes observed at multiple longitudes in the equatorial latitudes (up to about 100 Magnetic South and a few degrees Magnetic North). These echoes are a daytime phenomenon. They take a necklace
shape in a string-of-pearls formation, with faster pulsations overlaid on these. Only one PhD thesis has been written on this phenomenon, by Fawcett. The phenomenon may be related to meteors entering the Earth's atmosphere.
Phil Richards (George Mason University) described the importance of solar flux in relation to the ionosphere and thermosphere. Solar flux is the primary source of energy for the ionosphere and an important source of energy for the thermosphere. However, solar flux is poorly characterised by sunspot number. The 10.7 cm radiation (F10.7) is reasonably well correlated with the 33.5 cm radiation, but the relationship is not that good in conditions of low radiation. It is better to use MgII to characterise EUV solar flux. In general, it is difficult to characterise EUV emissions with observations, but they are important ionospheric energy inputs. Different observations of EUV flux disagree with each other.
Gary Bust (JHU/APL) ran a workshop on Observation System Simulation Experiments (OSSEs). These experiments work by first simulating a 'truth' ionosphere, then measuring the simulation with the fictional truth and trying to reproduce the truth with the simulated observations. The aim of these experiments is to establish the impact of future observation campaigns on our ability to specify the ionosphere, and to determine the performance of our specification techniques. Gary Bust led a collaborative meeting of
modellers, forecasters and observationalists to find a group approach for using OSSEs in our field. Some people contributed ways of making standard 'truth' model runs that the various assimilation or imaging groups could attempt to reproduce.
Work on DART
Before and during the CEDAR meeting, I was in Boulder to work on the Data Assimilation Research Testbed approach to ionospheric assimilation, using the Thermosphere-Ionosphere-Electrodynamics General Circulation Model (TIEGCM). Gang Lu (NCAR HAO) provided me with her code to run TIEGCM with the
Assimilative Mapping of Ionospheric Electrodynamics (AMIE) high-latitude inputs, which could significantly improve the high-latitude electric field and solar flux external forcing specifications to TIEGCM. I am attempting to reproduce Gang's results before I implement this version of TIEGCM inside DART. I also met with the Tomoko Matsuo and the DART team – Jeff Anderson, Nancy Collins and Tim Hoar – to discuss my results so far. The problem of negative densities or mixing ratios after assimilation of observations is caused by spurious correlations. Jeff suggested I introduce a fictitious vertical location for my vertical column-integrated observations at the height of the peak density. I could then use a vertical localization function to eliminate any correlations far from the location of the observations. Another problem that Jeff highlighted is that my posterior guess looks nothing like any of my prior ensemble members. This is a problem because the
Ensemble Kalman Filter is supposed to select the posterior solution from the ensemble of priors. This problem will be addressed after the vertical localization has been implemented.