Limits...
Two-step forecast of geomagnetic storm using coronal mass ejection and solar wind condition.

Kim RS, Moon YJ, Gopalswamy N, Park YD, Kim YH - Space Weather (2014)

Bottom Line: To forecast geomagnetic storms, we had examined initially observed parameters of coronal mass ejections (CMEs) and introduced an empirical storm forecast model in a previous study.However, the latter produces better forecasts for 24 nonstorm events (88%), while the former correctly forecasts only 71% of them.We then performed the two-step forecast.

View Article: PubMed Central - PubMed

Affiliation: Astronomy and Space Program Division, Korea Astronomy and Space Science Institute Daejeon, South Korea.

ABSTRACT

To forecast geomagnetic storms, we had examined initially observed parameters of coronal mass ejections (CMEs) and introduced an empirical storm forecast model in a previous study. Now we suggest a two-step forecast considering not only CME parameters observed in the solar vicinity but also solar wind conditions near Earth to improve the forecast capability. We consider the empirical solar wind criteria derived in this study (B z  ≤ -5 nT or E y  ≥ 3 mV/m for t≥ 2 h for moderate storms with minimum Dst less than -50 nT) and a Dst model developed by Temerin and Li (2002, 2006) (TL model). Using 55 CME-Dst pairs during 1997 to 2003, our solar wind criteria produce slightly better forecasts for 31 storm events (90%) than the forecasts based on the TL model (87%). However, the latter produces better forecasts for 24 nonstorm events (88%), while the former correctly forecasts only 71% of them. We then performed the two-step forecast. The results are as follows: (i) for 15 events that are incorrectly forecasted using CME parameters, 12 cases (80%) can be properly predicted based on solar wind conditions; (ii) if we forecast a storm when both CME and solar wind conditions are satisfied (∩), the critical success index becomes higher than that from the forecast using CME parameters alone, however, only 25 storm events (81%) are correctly forecasted; and (iii) if we forecast a storm when either set of these conditions is satisfied (∪), all geomagnetic storms are correctly forecasted.

No MeSH data available.


Related in: MedlinePlus

Prediction capabilities (PODy and CSI) according to the disturbances, Bz and Ey, and durations. Blue lines delineate the 3 h duration of Bz disturbances, and sky blue lines are the 2 h durations. Red lines delineate the 3 h duration of Ey disturbances, and orange lines are the 2 h durations.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4508932&req=5

fig04: Prediction capabilities (PODy and CSI) according to the disturbances, Bz and Ey, and durations. Blue lines delineate the 3 h duration of Bz disturbances, and sky blue lines are the 2 h durations. Red lines delineate the 3 h duration of Ey disturbances, and orange lines are the 2 h durations.

Mentions: However, GT criteria was originally proposed for intense storms; therefore, it may be too strict to distinguish moderate geomagnetic storms, which are the focus of the present study. For example, among 31 storms in our data set, only 12 events satisfy this condition for Bz (12/31, 39%), and 19 storms do not. In a similar way, 17 storm events cannot satisfy the Ey condition. Thus, we feel that we need a new solar wind criteria for practical usage in the forecast of moderate storm. To select Bz and Ey criteria as well as their durations for moderate storms, we use contingency tables, which have been widely used in the meteorological forecasting literature. These tables can provide us with information about the success or failure of the forecasting experience. General form and detailed explanation of the contingency table and its statistical parameters can be found in Smith et al. [2000]. For several cases of Bz, Ey, and durations, we compared “probability of detection yes (PODy)” and “critical success index (CSI).” PODy is the proportion of correctly forecasted events among the observed storms, and CSI is the proportion of correctly forecasted storm events among those that were either predicted or observed. Table 2 and Figure 4 show PODy and CSI as the storm prediction capabilities according to each criteria. Note that the better forecasts are indicated by statistical values that are closer to 1.0. As shown in the figure, for the moderate storms with the minimum Dst less than −50 nT, PODy and CSIs have clear tendency to decrease when the criteria become higher.


Two-step forecast of geomagnetic storm using coronal mass ejection and solar wind condition.

Kim RS, Moon YJ, Gopalswamy N, Park YD, Kim YH - Space Weather (2014)

Prediction capabilities (PODy and CSI) according to the disturbances, Bz and Ey, and durations. Blue lines delineate the 3 h duration of Bz disturbances, and sky blue lines are the 2 h durations. Red lines delineate the 3 h duration of Ey disturbances, and orange lines are the 2 h durations.
© Copyright Policy - open-access
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4508932&req=5

fig04: Prediction capabilities (PODy and CSI) according to the disturbances, Bz and Ey, and durations. Blue lines delineate the 3 h duration of Bz disturbances, and sky blue lines are the 2 h durations. Red lines delineate the 3 h duration of Ey disturbances, and orange lines are the 2 h durations.
Mentions: However, GT criteria was originally proposed for intense storms; therefore, it may be too strict to distinguish moderate geomagnetic storms, which are the focus of the present study. For example, among 31 storms in our data set, only 12 events satisfy this condition for Bz (12/31, 39%), and 19 storms do not. In a similar way, 17 storm events cannot satisfy the Ey condition. Thus, we feel that we need a new solar wind criteria for practical usage in the forecast of moderate storm. To select Bz and Ey criteria as well as their durations for moderate storms, we use contingency tables, which have been widely used in the meteorological forecasting literature. These tables can provide us with information about the success or failure of the forecasting experience. General form and detailed explanation of the contingency table and its statistical parameters can be found in Smith et al. [2000]. For several cases of Bz, Ey, and durations, we compared “probability of detection yes (PODy)” and “critical success index (CSI).” PODy is the proportion of correctly forecasted events among the observed storms, and CSI is the proportion of correctly forecasted storm events among those that were either predicted or observed. Table 2 and Figure 4 show PODy and CSI as the storm prediction capabilities according to each criteria. Note that the better forecasts are indicated by statistical values that are closer to 1.0. As shown in the figure, for the moderate storms with the minimum Dst less than −50 nT, PODy and CSIs have clear tendency to decrease when the criteria become higher.

Bottom Line: To forecast geomagnetic storms, we had examined initially observed parameters of coronal mass ejections (CMEs) and introduced an empirical storm forecast model in a previous study.However, the latter produces better forecasts for 24 nonstorm events (88%), while the former correctly forecasts only 71% of them.We then performed the two-step forecast.

View Article: PubMed Central - PubMed

Affiliation: Astronomy and Space Program Division, Korea Astronomy and Space Science Institute Daejeon, South Korea.

ABSTRACT

To forecast geomagnetic storms, we had examined initially observed parameters of coronal mass ejections (CMEs) and introduced an empirical storm forecast model in a previous study. Now we suggest a two-step forecast considering not only CME parameters observed in the solar vicinity but also solar wind conditions near Earth to improve the forecast capability. We consider the empirical solar wind criteria derived in this study (B z  ≤ -5 nT or E y  ≥ 3 mV/m for t≥ 2 h for moderate storms with minimum Dst less than -50 nT) and a Dst model developed by Temerin and Li (2002, 2006) (TL model). Using 55 CME-Dst pairs during 1997 to 2003, our solar wind criteria produce slightly better forecasts for 31 storm events (90%) than the forecasts based on the TL model (87%). However, the latter produces better forecasts for 24 nonstorm events (88%), while the former correctly forecasts only 71% of them. We then performed the two-step forecast. The results are as follows: (i) for 15 events that are incorrectly forecasted using CME parameters, 12 cases (80%) can be properly predicted based on solar wind conditions; (ii) if we forecast a storm when both CME and solar wind conditions are satisfied (∩), the critical success index becomes higher than that from the forecast using CME parameters alone, however, only 25 storm events (81%) are correctly forecasted; and (iii) if we forecast a storm when either set of these conditions is satisfied (∪), all geomagnetic storms are correctly forecasted.

No MeSH data available.


Related in: MedlinePlus