Transcript Document
Electric Field Variability and Impact on the Thermosphere Yue Deng1,2, Astrid Maute1, Arthur D. Richmond1 and Ray G. Roble1 1. HAO National Center for Atmospheric Research 2. CIRES University of Colorado and SWPC NOAA MURI,2008 Joule heating calculation: QJ E E 2 2 2 E Codrescu et al., [1995] The quantitative application of GCMs for predictive purposes is limited by uncertainties in the energy inputs How big is the E-field variability and what’s the effect to the energy input? (Codrescu et al., [1995, 2000], Crowley & Hackert, [2001], Matsuo et al., [2003], Matsuo & Richmond [2008] and so on.) Dynamic Explorer 2 Data Set •Time period: August 1981-March 1983 •Ion Drift Meter (IDM) g cross-track ion drift •Retarding Potential Analyzer (RPA) g along-track ion drift •Fluxgate Magnetometer (MAGB) g magnetic field •Low Altitude Plasma Instrument (LAPI) g ion / electron energy flux •IGRF for geomagnetic main field •IMF conditions: hourly averaged •Number of passes: 2895 Empirical Model Empirical model of the high latitude forcing: •Electric potentialb •Magnetic Potentialb •Poynting fluxb •Small scale electric field variabilityb •Auroral particle precipitation a Input to general circulation models Poynting flux empirical Model: E B 0 E B 0 Diff Bt= 5 nT, Equinox, IMF_angle = 1800 Point measurements of E-field and B-field data from the DE-2 satellite. Poynting > ExB ~ Weimer05 Standard deviation of E-Field N (E) ( Ei i 1 DE2 Ei model 2 ) N where E electric field (here Ed1 and Ed2 components) N number of trips EDE2 electric field from DE2 data set Emodel electric field from empirical model Energy distribution (Equinox): E E+varE Poynting Altitude integrated Joule heating and Poynting flux from the topside. E-field variability increases JH significantly. Total Joule heating has a similar distribution as Poynting flux, with some detailed difference at the polar cap, cusp and nightside. Comparison of energy input into GCM: 250 Total energy input [GW] 200 E 150 E+varE 100 Poynting By = 0 Bz= -5nT SW=400km/s HP=30GW 50 0 Summer Equinox Winter The E-field variability increases the energy input by > 100%. The total Joule heating has a good agreement with Poynting flux. The inconsistent particle precipitation makes the JH higher than Poynting flux in the solstice. Temperature response: Polar average (Lat > 47.50) at equinox. E-field variation causes >100 K temperature increase above 300 km. Temperature difference ~ [62 K, 250 K]. Density response: Percentage difference compared with the average E-field case. The difference is close to 30% at 400 km altitude. Conclusion : • The electric field variability increases the Joule heating by more than 100%, and significantly improves the agreement between the Joule heating and Poynting flux. • E-field variation causes >100 K temperature increase at 400 km, and the corresponding percentage difference of density is close to 30%. Future Work: • Develop a consistent particle precipitation model. • Improve the similarity of the total Joule heating and the Poynting flux distributions. • Comparison with observations to evaluate the Poynting flux and E-field variability in the model. Thanks! MURI,2008 Questions? • Q1: Why there was no E-var empirical model before when the idea has been proposed since 1995 and the DE-2 data are there? A. Just a matter of time, funding. • Q2: Why there are no dependence on solar wind velocity and density? A. Maybe in the future, it will be parameterized to IEF instead of IMF. IEF is close to –VxB and the effect of solar wind will be taken into account indirectly. • Q3: Why 50 lat resolution for Poynting model and 20 for others? How about horizontal resolution? A. Possibly Poynting flux needs both E and B. The available data are less. Check with Astrid. Horizontally, the Fourier function has been used for the MLT fit. The latitudinal dependence is presented by the Spherical Cap Function. • Q4: Is the E-var from the empirical model sub-grid? Is it temporally and spatially correlated? A. E-var just shows the difference between the DE-2 observation and empirical average model, and can include both sub-grid and large scale variation. When I implement the E-field variability by switching the sign of the sigma-E every time step, this means it is not temporally correlated. When we the sign in the whole polar region simultaneously, it means it is spatially full-correlated. When I set some phase difference between different latitude and longitude, in some way it is spatially uncorrelated. Questions? (Cond.) • Q5: Does the E-var from the empirical model represent more like spatial variability or temporal variability? A. Technically, it should be both. From the methodology of the processing the data, it represents more about the temporal variability between different satellite orbits. When run this model for a real case, hourly IMF condition will be recommended to use to drive the model, since the average model is binned based on the hourly IMF conditions and the E-field variability model is referred to that average model. If higher frequent IMF data (10 min average) have been used to drive the model, the E-var model should subtract the temporal component between 10min and 1 hour, which has been shown in the average model. • Q6: Why the E-var is maximum in the winter season? A. Usually, the E-var is largest when the conductance is small from the observation. E=J/sigma. When sigma is small, sigma and J are variable, the E can be very variable.