Highly Adaptive Electron Density Profile Model Applied to Ionograms for Real Time Monitoring

1
Highly Adaptive Electron Density Profile Model
Applied to Ionograms for Real Time Monitoring

• Carlo Scotto
• Istituto Nazionale di Geofisica
• Italy

2
Introduction

• The INGV developed an ionosonde, called AIS
  (Advanced Ionospheric Sounder), (Zuccheretti et
  al., 2003 Bianchi et al., 2003). Together with
  the ionosonde the INGV developed Autoscala, a
  computer program for the automatic scaling of
  critical frequency foF2 and MUF(3000)F2 from
  ionograms (Scotto and Pezzopane 2002, Pezzopane
  and Scotto 2004, Pezzopane and Scotto 2005).
  Autoscala was recently extended by the addition a
  routine for the automatic scaling of sporadic-E
  layer (Scotto and Pezzopane, 2007) and of F1
  layer (Pezzopane and Scotto, 2007).

3

• The electron density profile model, used in this
  work can be divided into the following two
  regions 
• a) The E valley and the E bottom side
• b) The F2, F1 and intermediate regions.

4
The E valley and the E bottom side. Polynomials
are used to connect the anchor points C, D, E and
F.
5
The F2, F1 and intermediate regions. It is used
the Reinisch and Huang (2000) formulation.
6
The model used in this work we have 12 parameters
7
What is the right profile?

• The problem is to choose the right value for each
  parameter.
• At the beginning we choose a ltltbasegtgt value
  evaluated using the data automatically given by
  Autoscala and the appropriate formulations given
  by literature.
• Then we adjust these values.

8
For each parameter Pi (i1,2,.12) describing
the profile we consider an interval centred on
the ltltbasegtgt value of the parameter and having
width 2? Pi . We look for the ltltoptimizedgtgt
value inside this interval. Therefore we
havePibase -? Pi ? Pioptimized ? Pibase
? Pi .

9
For each of the 12 parameters we are going to see
how to choose the ltltbasegtgt value. These ltltbasegtgt
values, as we have already said, must be
subsequently adjusted.
10
1. Maximum electron density of F2 layer
The maximum electron density NmF2base is
obtained from the critical frequency f0F2 given
by Autoscala.
11
2. The height of the F2 cusp
Autoscala is able to recognize a significant part
of the F2 trace. This part is sufficient to
calculate MUF(3000)F2. So the height of the F2
cusp can be estimated using
12
3. Electron density of F1 layer

• The electron density of F1 layer NmF1baseis
  obtained from the critical frequency f0F1 given
  by Autoscala.

13
4. Shape parameter B1
For the shape parameter B1 it is assumed
B1base 3. Numerical experiments showed that
it is a quiet accurate value to successfully
perform the adjustement procedure.
14
5. D1 shape parameter of F1 layer
The D1 shape parameter of Renisch and Huang
(2000) is set to D1base1.75 if Autoscala
detected an F1 cusp D1base0.35 if Autoscala
did not detect an F1 cusp.
15
6. Shape parameter B0

• For the shape parameter B0base two formulations
  are necessary
• If Autoscala detected an F1 layer.
• If Autoscala did not detect any F1 layer.

16
6a. Shape parameter B0, without F1 layer
17
6b. Shape parameter B0, with F1 layer
18
7. Electron density of E layer

• For the electron density of E layer we introduce
  a ltltnewgtgt model of the critical frequency f0E .
  This model usese ltlttraditionalgtgt formulas the
  work of Titheridge (2000).

19
7. Model for f0E, to be used to calculate
NmEbase
20
8. The height of the E layer

• It is set
• hmEbase 110 km.

21
9.The Bradley and Dudeney width of the E layer
(modelled as a parabola)

• It is assumed
• ymEbase 15 km
• It appear to work slightly better than the
  original ymEbase 20 km, Bradley and Dudeney
  (1973)

22
10. Width of E valley

• The width of E valley is calculated by a
  model based on incoherent scatter radar
  measurements (Mahajan et al., 1997) . According
  to this model the width of E valley can be
  expressed as it follows

23
11. The depth of E valley

• The depth of E valley ?NvEbase is
  calculated by the model of Mahajan et al. (1997).

24
12. The height of E valley

• For the height of E valley hvEbase we use the
  model of Mahajan et al. (1997).

25
(No Transcript)
26
How to compare the artificial ionogram to the
recorded one?

• If we found some reliable points belonging to
  the recorded ionogram we could use the mean
  square deviation.
• Autoscala can reliably give the needed points.

27
Reliable points without F1
28
Reliable points with F1 layer
29
Example nightime ionogram
30
Example daytime ionogram
31
Real time ionograms automatically scaled by
Autoscala and (now) the real time profile
generated by HAIP accessible at http//ionos.ingv
.it/Gibilmanna/latest.html
32

• Thank you.