Geomagnetic activity nowcast - short explanation - suomeksi

This is an empirical method based on near-real time solar wind observations and ground magnetometer data.

Solar wind data

The US Advanced Composition Explorer (ACE) observes solar wind parameters the Lagrange 1 libration point which is a point of Earth-Sun gravitational equilibrium about 1.5 million km from Earth and 148.5 million km from the Sun. We use near-real time data of ACE provided by the NOAA/Space Weather Prediction Center.

We use 1-minute data of the latest hour of the interplanetary magnetic field (IMF) Bz (northward component) and solar wind velocity and proton density. We apply the following criteria to estimate solar wind activity based on the hourly average:

This is a simplified version of the corresponding table for the ground magnetic field. The parameter limits are based on data of 1996-2011. Example: if the average value of the solar wind velocity of the latest hour is 500 km/s then activity is at the "increased" level. Note: we consider that only negative IMF Bz may indicate activity. So all positive Bz values correspond to a quiet level. However, if the velocity of the solar wind is high then there may be geomagnetic activity even with positive Bz.

Since we use near-real time data in an automatic routine, there is no strict control of possibly erroneous values. The number of available 1-minute values of the latest hour may be less than 60.

Ground magnetic field

We use twelve IMAGE magnetometer stations with a real-time access. We consider the range (RX) of the north component of the magnetic field (BX). RX is simply defined by RX = max(BX)-min(BX), and is given in nT. In the forecast service, we calculate RX for 1-hour sequences.

We characterise the magnetic activity level following the NOAA space weather scale for geomagnetic storms. The original definition is based on the number of storm events when a given level of the 3-hour Kp index is met during a solar cycle (11 years). This leads to the following percentages:

The total number of 3-hour sequences in 11 years is 32120. NOAA defines scales G1-G5, and we have extended them downwards to less active non-storm events (G0-, G0, G0+). According to these selections, about 8% of all events belong to storms. Using the percentages of the last column, we then determine limits of RX corresponding to each scale at the selected IMAGE sites. The limits vary so that an equal disturbance as measured in nT may correspond to a higher activity class in southern Finland than in northern Finland.

Forecast of the magnetic activity level for the next hour is based on the average value of the solar wind magnetic field (IMF Bz) and velocity (V) of the latest hour. Using data of 1996-2011, we have derived a simple statistical relationship between RX and IMF Bz and V for each UT hour. We give the limits within which RX will be during the next hour at a 90% probability. In the historical data, RX has been smaller than the given lower limit in 5% of cases, and larger than the given upper limit in 5% of cases, respectively.

About interpretation

The forecast is produced automatically, so problems in recordings or data transfer may lead to missing or erroneous forecasts. The method tries to detect clear errors automatically, but there is still a risk of some bias. A rule of thumb is that the ground magnetic field is more variable during active solar wind conditions. The ground field is typically more variable in the north (Kilpisjärvi) than at the southern sites (Nurmijärvi and Hankasalmi), and the variability is larger during the night than daytime.

The most important solar wind parameter is the northward magnetic field (IMF Bz). Magnetic storms typically occur during negative enough IMF Bz. The solar wind velocity is also then generally large, but the importance of the density is smaller.

Feedback to ari.viljanen (at) fmi.fi