Indian sector · Solar Cycle 25
LT — Not fetched API
Live prediction
Manual input
OMNIWeb fetch
History
Research
IMF Bz (reconnection driver)
PPEF estimate (15–20 min lag)
Bz < 0 → eastward penetration
Indian sector local time
EEJ peaks 10–14 LT
Bz GSM
nT
By GSM
nT
Speed V
km/s
Density n
cm⁻³
Pressure
nPa
Ey (×Bz)
mV/m
Predicted SYM-H (+60 min)
nT
Uncertainty (1σ):
Input mode:
Sequence length: min
NOAA data age:
Model: EEJ_LSTM_v3 (SC24)
Station: TIR–ABG (Indian sector)
v5
Parallel prediction (+60 min)
nT
+ F10.7:
Model: SymH_LSTM_v5
SYM-H prediction history (session)
No predictions yet.
Solar wind Bz (last fetch)
No data yet.
Single-point prediction
Ey and pressure auto-computed if blank
Southward = negative. Storm driver.
Quiet ~400, Storm ~700+
Pressure = 1.67×10⁻⁶ × n × V²
DDEF proxy. Active >500 nT.
NASA OMNIWeb historical fetch
Max 2-hr window (120 min)
POST https://api.arpitmann.com/predict/omni
Run a fetch to see results.
Prediction log (session)
No predictions yet.
What is the Equatorial Electrojet?

The Equatorial Electrojet (EEJ) is a narrow ribbon of electric current flowing eastward at ~100 km altitude over the magnetic dip equator — directly above India's southern tip. It forms because the ionosphere at the equator has unusually high electrical conductivity (Cowling conductivity, 5–10× normal), amplifying the normal dynamo current into a concentrated jet ~600 km wide.

During geomagnetic storms, two competing mechanisms disrupt the EEJ: the Prompt Penetration Electric Field (PPEF) — driven by southward IMF Bz, reaches the equator in 15–20 minutes — and the Disturbance Dynamo Electric Field (DDEF) — driven by auroral heating and thermospheric winds, arrives 4–8 hours later. Their interplay produces the complex EEJ variability seen during extreme storms.

We measure EEJ strength using ΔH = H_TIR − H_ALB: the difference in the horizontal magnetic field between Tirunelveli (on the dip equator) and Alibag (off-equator reference). Positive ΔH = normal eastward EEJ. Negative ΔH = Counter Electrojet (CEJ) — a westward reversal that disrupts HF radio, GPS, and navigation systems across India.

May 2024 Extreme Storm — Case Study

SYM-Hmin = −518 nT · CME cannibal event · Strongest storm of Solar Cycle 25

May 2024 EEJ storm analysis showing Bz, solar wind speed, AE index, H-component disturbance, delta-H EEJ index, and SYM-H
Key findings: EEJ peaked at +137.8 nT on May 10 at local noon (07:00 UTC) — the undisturbed diurnal signature. Storm onset May 10 18:00 UTC: Bz plunged to −47.9 nT, V jumped to 900 km/s, AE reached 4098 nT. EEJ became chaotic — rapid oscillations between +120 and −178 nT as PPEF and DDEF competed. SYM-H hit −518 nT on May 11 02:14 UTC. 567 minutes of Counter Electrojet recorded — nearly 10 hours of westward reversal.
ML Model

EEJ_LSTM_v3 — Bi-layer LSTM trained on Solar Cycle 24 storms (2015–2020), evaluated on Solar Cycle 25 (2021–2024). Predicts SYM-H 60 minutes ahead using 2 hours of solar wind context.

Features
12
incl. sin/cos LT
Overall MAE
7.14
nT (SC25 test)
Training storms
98
SC24 2015–2020

⚠ Model trained on SC24. Extreme SC25 events (SYM-H < −300 nT) are outside the training distribution — predictions for such events carry high uncertainty. This is a known limitation and an active research finding.

About this research

This project is part of ongoing M.Sc. research by Arpit Mann (IIT JAM 2025 AIR 465, M.Sc. Astronomy). Data: NASA OMNI 1-minute solar wind (2015–2024), IIG magnetometer network (TIR, ABG, HYB, JAI). Target journal: Space Weather (AGU). Contact: [email protected]