Ivan Galkin 1 Artem Vesnin 1 Xueqin Huang 2 Alexander “Sasha” Kozlov 1 Paul Song 1 Bodo Reinisch 1,2 1 University of Massachusetts Lowell, Space Science Laboratory 2 Lowell Digisonde International, LCC For an analy alysi sis o of real al-time IR IRI M Map aps s of foF2 an and h d hmF mF2 14 th International Ionospheric Effects Symposium Alexandria ● Virginia ● May 12, 2015 Session 1a | Ionospheric & Space Weather Models
IES 2015 Outline Assimilation Techniques for IRI [model of Ne] 1. GIRO: source of real-time ionogram-derived data 2. IRI empirical model formalism 3. NECTAR assimilative “model morphing” technique = IRTAM: real-time global nowcasting 4D Data Assimilation (4DDA): 24 hour context Session 1a | Ionospheric & Space Weather Models foF2 and hmF2 maps vs TEC maps Alexandria ● VA ● May 12, 2015 GAMBIT Database and Explorer Public access to IRTAM results and analysis tools Open source (2015) Outlook: where do we go from here? IRTAM versus physics-based assimilative modeling Spatial prediction capability: covariances? Time forecast capability
2015 1. Global Ionosphere Radio Observatory IES Real-time GIRO ionosondes, ~50 locations Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015
IES 2015 IRI Real-Time Extension Height of maximum F2 Layer Ionization (hmF2) + = Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015 Global hmF2 “Climatology” IRI Ionosonde Network Real-Time hmF2 Global hmF2 Weather • Credits to the Real-Time IRI Task Force (2009)
IES 2015 IRTAM 24-hour History Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015
IES 2015 IRTAM Deviation Maps HOW IONOSPHERE IS DIFFERENT FROM ITS QUIET-TIME STATE Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015 Is this real?
IES 2015 IRTAM complementary to TEC maps Substorm March 17, 2015 23:22UT ∆TEC ∆ f o F2 ∆ h m F2 Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015 100s sensors 10s sensors (no interpolation) TEC maps courtesy Madrigal Node at MIT Haystack Observatory; [Coster et al., 2008]
IES 2015 Under the Hood: IRI Formalism Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015
IES 2015 Assimilation Technique NECTAR: Non-linear Error Compensating Technique for Associative Restoration Keep IRI expansion basis unchanged Jones-Gallet geographic functions Gk – 76 functions Session 1a | Ionospheric & Space Weather Models 6 th order of diurnal harmonic analysis – 13 functions Alexandria ● VA ● May 12, 2015 Compute new coefficients C ij that minimize data- model error Disseminate new 988 coefficients to existing IRI users Small, efficient, no changes to the IRI engine necessary
IES 2015 4D Data Assimilation (4DDA) One nowcasting calculation = match of 24-hour history of data with 24-hour model = 24-hour 4DDA scheme Overdeter ermi mined ned in n ti time (96 e (96 ti times es, 13 13 coef efficien ents) Session 1a | Ionospheric & Space Weather Models Und nder erdeter ermined ed i in n spa pace (40 e (40 observato tories, 76 76 coeffici ficients) Alexandria ● VA ● May 12, 2015 Not t a cla classic K Kalman filt filter m mated t to o a fir first-princip ciple les model… l… Ionospher ere i in ter n terms of i its ts “ “Eigen func ncti tions” describing the the essenc ence o e of its ts ti timeline b e beha havior
IES 2015 Empirical vs Physics-Based Assimilation COMPLEMENTARY TECHNIQUES EMPIR IRICA CAL PHYSIC SICS-BASED ED 3DDA 4DDA Next time Next time 50 x 1 50 x 96 data points data points Session 1a | Ionospheric & Space Weather Models INITIALIZATION PREDICTION FORECAST UPDATE UPDATE Alexandria ● VA ● May 12, 2015 output output Model 988 C ij climatology drivers nowcast nowcast (24 hours) Represents processes yet to be understood Describes system in terms of constituent processes
IES 2015 24-hour Temporal Harmonics Expansion 4DDA approach is robust to autoscaling blunders Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015 Eglin AFB foF2 series courtesy AFWA NEXION program
IES 2015 Near Future: Global Realistic Ionosphere using GIRO Real-time Data Feeds To raytracing foF2 Session 1a | Ionospheric & Space Weather Models Background 3D Realistic Alexandria ● VA ● May 12, 2015 Ionosphere hmF2 TID modulation FAS
IES 2015 GAMBIT Database and Explorer http://giro.uml.edu/GAMBIT Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015 IRTAM foF2 and hmF2 timelines since 2000 + real-time data with 10 min delay Public access to IRTAM retrospective and current results
IES 2015 GAMBIT Explorer Global Assimilative Model for Bottomside Ionospheric Timelines Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015 http://giro.uml.edu/GAMBIT
IES 2015 Outline Assimilation Techniques for IRI [model of Ne] 1. GIRO: source of real-time ionogram-derived data 2. IRI empirical model formalism 3. NECTAR assimilative “model morphing” technique = IRTAM: real-time global nowcasting 4D Data Assimilation (4DDA): 24 hour context Session 1a | Ionospheric & Space Weather Models foF2 and hmF2 maps vs TEC maps Alexandria ● VA ● May 12, 2015 GAMBIT Database and Explorer Public access to IRTAM results and analysis tools Open source (2015) Outlook: where do we go from here? IRTAM versus physics-based assimilative modeling Spatial prediction capability: covariances? Time forecast capability
IES 2015 Future work: Covariance study ∆ f o F2 ∆ f o F2 Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015 vs. Fast attenuation Slow attenuation (small covariance) (large covariance)
IES 2015 IRTAM resolution = IRI resolution? Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015
IES 2015 Managing the Day Boundary Issue Inertia of the model? 13 million data fits Average Error [Vesnin, UML, 2014] Session 1a | Ionospheric & Space Weather Models Average error of IRTAM matching source data Alexandria ● VA ● May 12, 2015 SOLUTION: 24-hour expansion using 21 hours of data
IES 2015 IRTAM in Forecast Mode Session 1a | Ionospheric & Space Weather Models Alexandria ● VA ● May 12, 2015
IES 2015 Outlook URSI INAG Working Group G.1 actively pursues real-time ionosonde network operation IRI-based Real-Time Assimilative Model is in operation since January 2013; IRTAM validation in progress RETID (USA) and Net-TIDE (Europe) projects support TID detection and evaluation using Digisonde data Session 1a | Ionospheric & Space Weather Models Lowell GIRO Data Center builds a public, open-source Alexandria ● VA ● May 12, 2015 environment for realistic ionosphere nowcast based on ionosonde data feeds Cooperations with CEDAR Madrigal, NASA CCMC and VWO, European ESPAS are good opportunities to provide single-stop data dissemination portals for realistic ionosphere nowcast Acknowledgements: AFRL SBIR “RETID”, NATO SfP 984894. Lowell DIDBase
Recommend
More recommend
