IONOSPHERIC SHORT TERM EMPIRICAL FORECASTING MODEL

Transcription

1 IONOSPHERIC SHORT TERM EMPIRICAL FORECASTING MODEL Giorgiana De Franceschi,, Lucilla Alfonsi, Luca Spogli, Claudio Cesaroni, Vincenzo Romano, Marcin Grzesiak, Andrej Wernik PRESENTED BY Paulo Camargo Conference and Trade Fair for Geomatics and Geospatial Solutions 7-9 MAY, Frei Caneca Convention Center - Sao Paulo - Brazil

2 OUTLINE Climatology of scintillation Over Brazil (CIGALA/CALIBRA network) Over the Sao Paulo State (URTKN) Ionospheric short term empirical forecasting model The idea behind and the formulation The model data input Results on test cases Velocity field reconstruction Performance of forecasting Remarks

3 CIGALA/CALIBRA data analysis The analysis of the CIGALA/CALIBRA network data along 2012 by an ad hoc technique, the GBSC, highlights the impact of both the northern and southern crest of the EIA in driving the scintillation phenomena on GNSS signals. The amplitude scintillation patterns follow closely mainly the enhancement of the standard deviation of the ROT, confirming the important role of TEC gradients to understand scintillation occurrence. The most affected regions probed are those in the latitudinal range between 30 S and 10 N and in correspondence of longitudes between 300 E and 330 E with particular strength: over Sao Paulo and Tocantins States (due to the presence of the EIA southern crest) northward of MANA (due to the presence of the EIA northern crest). over POAL, possibly due to the presence of the particle precipitation occurring in the borders of the SAMA. The scintillation occurrence is prevalent between 22 to 04 UT.

4 URTKN data analysis Characterization of the ionosphere over Sao Paulo state in terms of: Calibrated TEC and TEC gradients Location Station Code Lat. N Lon. E Araçatuba SPAR ' ' Campinas SPCA ' ' Cananéia NEIA ' ' Ourinhos OURI ' ' Presidente Prudente PPTE ' ' São José do Rio Preto SJRP ' ' São Paulo POLI ' ' Maringá PRMA Curitiba UFPR Varginha/Cemig MGVA ' ' Inconfidentes MGIN ' ' Amplitude scintillation occurrence. The derivation has been made from the GPS and GLONASS RINEX files (15 s temporal resolution) Data can be downloaded by ftp://geoftp.ibge.gov.br/

5 TECu SPRING: 26 Sep 2012 from 02:20 to 02:30 UT TEC TECu/km MODERATE SCINTILLATION TEC gradient N-S TECu/km TEC gradient E-W CALIBRA MTR

6 URTKN analysis remarks Equatorial Plasma Bubbles often create the conditions to drive scintillations Not only EPB but also electron density enhancements can result into scintillations EPB are better described by the N-S gradients. Not all the mapped ionospheric irregularities produce amplitude scintillation. Amplitude scintillations are almost always found in correspondence with the edges of the irregularities The seasonal variation confirms equinoxes to be most active in recording high level of S 4. Both the overall and the finer URTKN climatology confirm that Sao Paulo State is the best candidate to test the model

7 Model development Scope: to develop an empirical, local, TEC and scintillation short term (seconds-minutes) forecast model Approach: by assuming the TEC temporal gradient is mainly function of the TEC spatial gradient and of the plasma drift velocity. How: by applying the transport theory to derive the plasma drift velocity field from TEC measurements by considering additional source terms such as the plasma density production and loss rate in the continuity equation to take into account the impact of solar radiation. Input: TEC and scintillation parameters Output: Forecasted values of TEC, TEC gradient, scintillation parameters

8 Model development: the idea of modeling The model is based on the transport theory for a scalar field. Equation of continuity for a scalar f provided known velocity field v. Equation of continuity Scalar field Velocity of the field Source term The reconstruction of the velocity field v is performed by fitting it to the time changes of vtec field. vtec is approximated piecewise linearly on the unstructured triangular grid.

9 MODEL development data input- List of the CIGALA/CALIBRA network sites Station ID Location Lat ( N) Lon ( E) PRU1 Presidente Prudente PRU2 Presidente Prudente PRU3 Presidente Prudente GALH Presidente Prudente SJCU San Jose dos Campos SJCE San Jose dos Campos INCO Inconfidentes MAC2 Macaé POAL Porto Alegre UFBA Salvador de Bahia PALM Palmas FORT Fortaleza MAN2 Manaus GPS + GLONASS L1 frequency

10 Model development: velocity field reconstruction Triangulation of the computational domain Reconstructed velocity field 100 m/s This example is by using as input CIGALA/CALIBRA network TEC data at pierce points (points numbered in the plot at left) for the 1 November 2011 at a given time T0.

11 Model results: Scintillation parameters Day 271/ Forecasting horizon: 1 minute - SigmaPhi distribution is slightly narrower than S4, even if the standard deviation is almost the same (different tails). - The bulk of the distribution is between -0.1 and Prediction is harder during local post-sunset (red box) σ Φ Standard deviation 0.03 radians σ Φ S4 Standard deviation 0.03 S4

12 Model results: vtec Day 271/2013 vtec Standard deviation 0.8 TECu vtec - Forecasting horizon: 15 seconds - The bulk of the distribution is between -2 and 2 TECu - Prediction is harder during local post-sunset (red box) - Similar behavior for spectral parameters T and p DAY Summary of model results Standard deviation of the distribution S4 SigmaPhi (radians) TEC (TECu) p T 271/ E-004

13 Model development: Remarks The climatological studies conducted in the framework of CALIBRA allowed to identify the Sao Paulo state area as the best candidate site to test the regional model The idea behind the forecasting is to apply the transport theory for a scalar field to predict TEC, scintillation and signal spectrum parameters. The model has been fed by CIGALA/CALIBRA network data The model was tested by forecasting the parameters for an entire day (271/2013) characterized by strong scintillation conditions. Model is able to give good results in terms of standard deviation of the distribution of the measured-forecasted value The prediction is harder during the local post-sunset hours The prediction strongly depends on the spatial experimental data distribution Extensive test against different scintillation conditions are currently ongoing.

14