GAIM: Ionospheric Modeling

Transcription

1 GAIM: Ionospheric Modeling J.J.Sojka, R.W. Schunk, L. Scherliess, D.C. Thompson, & L. Zhu Center for Atmospheric & Space Sciences Utah State University Logan, Utah Presented at: SDO EVE 2008 Workshop Virginia Tech, Blacksburg, Virginia October 6-9, 2008

2 O + H + GAIM s ionosphere Today it is mainly an assimilative F-layer, a physics based E-region, and an empirical D- region. EVE euv-soft x-ray spectra will provide real-time D- and E- region solar input and realistic F-layer background ionosphere. Flares would be a new key addition to GAIM.

3 GAIM Assimilates Multiple Data Sources Data Assimilated Exactly as They Are Measured Bottomside N e Profiles from Digisondes (30) Slant TEC from more than 1000 Ground GPS Receivers N e Along Satellite Tracks (4 DMSP satellites) Integrated UV Emissions (LORAAS, SSULI, SSUSI, TIP) Occultation Data (CHAMP, IOX, SAC-C, COSMIC)

4 Assimilation of SDO EVE into GAIM Physics-based Ionosphere Model has three layers that have different time constants and dominant processes Each layer will need separate assimilation considerations Kalman Filter Technique is currently used, but solar spectra will probably be handled differently CASS-USU will be the test location for SDO-EVE realtime technique development and testing for GAIM

5 Gauss-Markov Kalman Filter Model (GAIM-GM) GAIM-GM is an AFWA operational specification & forecast model It is a global model It runs on 2 CPU USU-ONR-AFWA are studying how to include real-time solar euv soft X-ray spectra as an input

6 GAIM-GM physics is the Ionosphere Forecast Model (IFM) Global physics-based model Provides background ionosphere km 15 - minute output cadence O +, H +, NO +, N 2+, O 2+, T e, T i Only use N e as the assimilation parameter! Kalman solves for deviations from background

7 GAIM-GM Example Global Mode November 16, 2003 GPS Ground TEC measurements from more than 900 GPS Receivers ( SOPAC Data Archive) Includes Receivers from: IGS CORS EUREF and others

8 Gauss-Markov Kalman Filter Reconstruction Physics-Based Model Without Data QuickTime and a Cinepak decompressor are needed to see this picture. Kalman Filter More than 3000 Slant TEC Measurements are assimilated every 15 minutes.

9 Reconstruction of the 3-D N e Distribution QuickTime and a Cinepak decompressor are needed to see this picture.

10 Why SDO EVE spectra? There already exist several models of the solar euv soft X-ray spectrum. Some of these have dynamic spectra based on different solar proxy measurements. Every ionospheric model can be calibrated to give the observed electron density.. hence all of these models can and are used. But they do not necessarily reproduce solar euv soft X-ray variability. The following TDIM plots extend a Chris Smithtro EVE workshop presentation from a few years ago.

11 Millstone Hill location Total Electron Content (the F-region) Solar Minimum and Low geomagnetic activity E-region density USU TDIM model equivalent to IFM, each color is same run with different solar spectral model. Spectral model key ( OK if you can not read this)

12

13

14 The order has changed for the different models between the E-region and F-region.

15 What we expect from EVE Over many years an excellent relative accuracy This needs to be both in time (minutes to years) and more importantly wavelength The absolute spectrum is of scientific value, but in GAIM the electron density can (will) be calibrated One minute cadence (this would make EVE the best data stream for GAIM)

16 Implementation Concepts As spectra change their relative E and F region responses change, this will be captured in GAIM. E-region is described by a chemistry steady state calculation in IFM. The assimilation schemes do not significantly impact this. Hence real time EVE spectra would drive the day-side E-region. Flares are especially important. F-region requires a different approach since the Ne assimilation is already attempting to produce the correct answer. However at each step, a new ionosphere is created and how this background is projected forward during flares would be dependent on EVE. D-region TBD, however for communication applications this is a crucial future GAIM direction.

17 Bob, this would be the end of talk

18 Gauss-Markov Kalman Filter Example Regional Mode 3-D Ionospheric N e Reconstruction over North America Large Geomagnetic Storm on November 20-21, 2003 GPS Ground TEC Measurements from more than 300 GPS Receivers over the continental US and Canada 2 Ionosondes at Dyess and Eglin Observe large TEC Enhancements over the Great Lakes during November 20, 2000 UT.

19 NOAA CORS Data 332 Sites Dual-frequency Receivers Slant TEC

20 Physics-Based Model (IFM) QuickTime and a Cinepak decompressor are needed to see this picture. Kalman Filter Reconstruction About 2000 Slant TEC Values are Assimilated every 15 min

21 Full Physics Kalman Filter Model Specification & Forecast of the Global Ionosphere 30 CPU

22 Full Physics Kalman Filter Model Ensemble Kalman Filter Physics-based Ionosphere-Plasmasphere Model Same 5 Data Sources as Gauss-Markov Model Altitude, Latitude, Longitude Grids Set by User

23 Global Ionosphere-Plasmasphere Model (IPM) 3-D Time-Dependent Parameters NO +, O 2+, N 2+, O +, H +,He + T e, T i u,u Grid System Global Regional Localized 90-30,000 km Realistic Magnetic Field (IGRF)

24 Longitudinal Resolution Resolution is Externally Adjustable Operational Mode: Global: ~ 7.5 o Regional: ~ 1 o 30 Global Simulations are Launched at Each Assimilation Time Step

25 Full Physics-Based GAIM Model Ionospheric Drivers are determined via an Ensemble Kalman Filter Global Run at Mid and Low Latitudes 3-D Electron Density Reconstruction Neutral Wind and Electric Field

26 Full Physics-Based GAIM Model Several Days in March/April of 2004 Geomagnetically Quiet Period Data Assimilated Slant TEC from 162 GPS Ground Receivers Use Ionosonde Data for Validation

27 Full-Physics-Based Kalman Filter Example GPS/TEC Data: Slant TEC Values have been mapped to the Vertical Direction QuickTime and a BMP decompressor are needed to see this picture. GAIM Specification of Global TEC Distribution

28 Comparison with Ionosonde Data K MH I QuickTime and a BMP decompressor are needed to see this picture. PR A AI Ionosonde Data were NOT assimilated!

29 Global Meridional Wind Obtained from GAIM Horizontal Wind Model (HWM) Full Physics-Based Data Assimilation Model Southward Northward

30 USU Physics-Based Data Assimilation Models 1. Kalman Filter Models of the Ionosphere o o Gauss-Markov Model Full Physics Model 2. Ensemble Kalman Filter Model of High- Latitude Electrodynamics 3. Ensemble Kalman Filter Model of the Thermosphere

31

32 GAIM Basic Approach We use a physics-based ionosphere or ionosphereplasmasphere model as a basis for assimilating a diverse set of real-time (or near real-time) measurements. GAIM provides both specifications and forecasts on a global, regional, or local grid. Global Regional Local