1. Summary of part of L2 session
Tools e.g. for data selection
Products, including models
Mostly based on magnetic data
Both dedicated and comprehensive models
Validation
Enhancing return from mission

2. Quiet time definition
Magnetosphere-ionosphere system reconfigures in minutes
Use SUPERMAG vector data to quantify external field
SMDL characterises field, value in nT
Quiet time – about 50% data overlap with definition used by Nils Olsen for data selection

4. Testing SUPERMAG definition
Compare with previous CHAOS model
Use SMDL to select quiet time data
Same amount of data at high latitudes
Same amount at mid/low latitudes, and ± 20%
Fit as good as or better than for CHAOS (e.g. < 2nT rms misfit to Bϕ)
Larger residuals near poles though overall misfit the same

5. IBI, TEC and FAC
FAC and ionospheric radial currents: good agreement between satellites, and between original and mapped data
Also undertook in-orbit comparison
Comparison between TEC from GPS and electron density measured by Langmuir probe
Obtain VTEC from slant TEC – reasonable agreement with electron density
Ionospheric bubble indices – investigated seasonal variations
Some ionospheric bubbles don’t have a magnetic signature – use electron density instead

6. FAC – Swarm A, B, and C High-Latitude (01 May 2014 – 31 May 2015)
Northern Hemisphere
Southern Hemisphere
Swarm A
Swarm B
Swarm C

7. TEC: Vertical TEC & Electron Density
Swarm A
Swarm B
Swarm C
VTEC Ne
Statistical Distribution of VTEC and Ne: good agreement

8. IBI: Seasonal Variations (18-04 MLT)
Swarm A
Swarm B
Swarm C
Event detection threshold = 0.15 nT
December Solstice
Equinox
June Solstice

9. Ionospheric irregularity index based on plasma density
IBI and Ne (constant and variable) indices
Both: use high-pass filter
IBI Index: constant threshold on magnetic field
Ne Index: threshold on density
Ne Index
Constant threshold: not yet sufficient for quantitative bubble description
Variable threshold
Better coincidence with plasma density maximum: enhanced symmetry about the equator
sensitive to data quality
New approach is necessary
direct detection of depletions and enhancements (work in progress)
Preliminary results of the new approach

10. Dedicated core field model
Based on GRIMM methodology
No observatory data
Data selection went well beyond using data flags
SV agrees with other models to degree 10
Euler angles (estimated from quaternions) show LT dependencies

11. Dedicated lithospheric field model
Several models investigated for removing core field – none perfect
Some along-track filtering, resulting in slight loss of power visible in power spectrum comparisons
Developed new orbit filtering based on similar track recognition
Included gradient data in inversion
Problems near poles – FACs?
Good resolution to degree ~75

12. PRESENTATION AND VALIDATION OF THE DIFI L2 PRODUCT
First DIFI model released to ESA on July 24, 2015
Second DIFI model calculated from satellite data until July 2015 and with new Q matrix based on more recent 1-D conductivity profile (Puethe et al.)
Change of Q matrix has a (very) small effect on the model
Overall performance of the DIFI chain is close to the requirement
DIFI and CI models have noticeable differences (but CI soon to be updated)
DIFI, Spring (Apr 1)
Equivalent current function of the ionospheric primary currents

13. Ocean tides Comparison with numerical model
Used residual night-time data, CHAMP and Swarm
Modelled to degree and order 40
Tidal periods imposed
Investigated M2, S2, K1 and N2
Some non-physical features (signal on land)
Some amplitudes too big

14. Conductivity

15. Comprehensive modelling
Improved lithospheric field in latest version
Described how model was obtained, although not clear why it is an improvement
Scalar and vector sums and differences used (restrictions on use of vector s/d)
Definition of M2 as good from 18 months of Swarm as 10 years of CHAMP
Ionospheric field over-regularised

16. Time-variable gravity field from Swarm GPSR data
Aleš Bezděk
Josef Sebera
Jaroslav Klokočník
Astronomical Institute, Czech Academy of Sciences,
Czech Republic
Swarm 5th Data Quality Workshop,
Institut de Physique du Globe de Paris, France, 7 – 10 September 2015

17. GPS monthly solutions: GRACE A/B & Swarm
KBR monthlies: CSR and GFZ
GPS monthlies: GRACE A/B ( )
GPS monthlies: Swarm (12/2013-3/2015)
Time series of GRACE A/B GPS-based monthly solutions is successfully continued by Swarm
both seasonal and secular variations
Note: this is mm precision in all data, transformations & background models
Swarm GPS orbits are accurate enough to track the mm-level time-variable gravity signal
Other groups (ITSG, AIUB) obtained similar results by using different inversion methods
Grace A/B: Institute of Geodesy (IfG), Graz University of Technology; Swarm monthly solutions: TU Delft

18. Validation Cross-validation between comprehensive and dedicated models
Compare with auxiliary models (e.g. IGRF)
Compare with observatory data

19. BGS Validation Summary
The purpose is primarily to:
ensure no grievous errors
give confidence to ESA and users products have been independently verified
Validation procedure is active and effective
Products deemed valid and suitable for release
HUA
b) (and c)) is key.
BOU
AUX_OBS_2_ residuals; Dedicated MIO models

20. Enhancing return Add-ons and improvements to existing products
Improvements to processing and modelling chains
New products
Lots of possible examples cited by Nils