Model/Technique Registration (*=required):
Information about your group/project:
Name(s) and e-mail (s) (please list primary contact first)*: David Falconer (david.a.falconer@nasa.gov), Igor Khazanov (igor.khazanov@uah.edu)
Associated Institution/Project name/Group name*: NASA/MSFC and University of Alabama Huntsville
Website url(s): http://www.uah.edu/cspar/research/mag4-page
Information about your method:
Forecasting method name*: MAG4 (Magnetogram Forecast)
Shorthand unique identifier for your method: MAG4_FE_1, MAG4_FE+F_1
Short description*:
Measures a free energy proxy from LOS or vector magnetogram and predict event rates for each AR. This can be done using LOS magnetogram (scientifically good to 30 heliocentric degrees), or vector (likely good to 45-60 heliocentric degrees, need to test).
There are four forecasts created from MAG4:
LOS magnetogram free energy only
LOS magnetotrams free energy + previous flares
Vector magnetogram free energy only
Vector magnetotrams free energy + previous flares
References:
Falconer, D.A., Barghouty, A. F., Khazanov, I., and Moore, R.L., 2011, "A Tool for Empirical Forecasting of Major Flares, Coronal Mass Ejections, and Solar Particle Events from a Proxy of Active-Region Free Magnetic Energy", Space Weather 9, S04003.
Falconer, D.A., Moore, R.L., Barghouty, A. F., and Khazanov, I., 2012, "Prior Flaring as a Complement to Free Magnetic Energy for Forecasting Solar Eruptions", Astrophys. J. 757, 32.
Falconer, D.A., Moore, R.L., Barghouty, A. F., and Khazanov, I., 2014, "MAG4 versus alternative techniques for forecasting active region flare productivity",Space Weather 12, 306.
Further Model Details:
(1)* Please specify if the forecast is human generated, human generated but model-based, model-based, or other.
Automated model-based.
(2)* Does your flare prediction method forecast “M1.0-9.9” or “M and above”? Would it be difficult for you to adapt your method from one to another? Which forecast binning method do you prefer?
It does M and above and X and above. By subtracting the rates, you could get M1-M9.9.
MAG4 measures a free-energy proxy and converts that into a predicted event rate
R=AW^B
Where A and B are fitted paramaters from a large sample, and W is the free-energy proxy. All forecasts are round to single digit accuracy.
(3)* How do you specify active regions in your model? Do you relate their location to other schemes such as NOAA or Catania? If so, what is your criterea to relate them?
For LOS version it finds strong magnetic field “islands” and associates them with NOAA AR. Association criteria is based on if a NOAA AR is in, or if not in one “island” close to an “island”.
Vector version uses HARP tiles, and the bitmap, to divide a HARP into individual AR or AR complexes. We use the HARP fits and NOAA association.
(4)* Uncertainties given as an upper and lower bound is an optional field. Does your model provide uncertainties for the forecasted probability?
If yes, what percentiles do you use to determine your upper and lower bound? If you are using the XML format, multiple percentiles of the probability density distribution can be specified.
We use uncertainties, the uncertainties are dominated by uncertainties in A and B fitting parameters. These result in a multiplicative uncertainty in R. When doing full Sun, we add event rates of each AR, and use the uncertainty from the largest predicted event rate, which typically is the dominate term. These uncertainties are propagated into the uncertainties in the probabilities.
(5)* For each forecast, what prediction window(s) does your method use? E.g. Does your method predict for the next 24 hours, 48, and 72 hours or the next 24 hours, 24-48 hours, 48-72 hours? (This information is useful for displaying only comparable methods together).
Our forecast is actually an event rate, that will slowly change as AR evolve and more rapidly change if there is recent flare history for our forecasts based on free-energy proxy and previous flare history. The forecasts curves are derived for 24 hours, but the 48 and 72 hours forecasts will be similar.
(6)* Calibration levels are optional fields. Do you have calibration for the probabilities from your model? E.g. is a 40% forecast a “high” probability for your method?
If yes, please provide the probability ranges for what is considered a “low”, “medium”, or “high” level.
We color code individual AR based on threat level. We have a threat gauge, the yellow level shows the 1 sigma confidence level of our forecast.
For example in the MAG4 model:
Probabilities in range → Level
Prob < 0.02 → 1
0.02 ≤ Prob < 0.18 → 2
Prob ≥ 0.18 → 3