SOStephen Omondi

Stephen Omondi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Stephen Omondi has a total of 14 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    How well can we forecast local magnetic ground perturbations with existing space weather monitoring resources?

    by Stephen Omondi, Spencer Mark Hatch, Andreas Kvammen, Magnar Gullikstad Johnsen, Mathew J. Owens, Kristian Solheim Thinn, Rodrigo López

    In this study we examine how a deep-learning based forecast of local, ground-based geomagnetic field variations trained on solar wind parameters available in real time might be improved by including information contained in an accurate forecast of solar wind conditions. This is accomplished using a long short-term memory (LSTM) model together with magnetic field measurements made at the Rørvik magnetometer station in Mid-Norway. We use Advanced Composition Explorer (ACE) satellite measurements of solar wind and interplanetary magnetic field (IMF) conditions at the first Sun-Earth Lagrange point, and historical lists of coronal mass ejection (CME) impacts at Earth to train and validate the LSTM model. We find that accurate information about the IMF B z component and solar wind speed are important for obtaining a reasonably accurate ( r 2 ≥0.5) forecast of local geomagnetic activity over forecasting horizons beyond ∼3 h. Information about CME arrival time is only important when accompanied by accurate, relatively high-resolution information about IMF B z . In the absence of the latter, CME arrival time information does not contribute to model performance. This empirical result amounts to a quantitative demonstration of the widely recognized impact of IMF orientation on CME geoeffectiveness. This result also highlights that new innovations, probably in the form of new prediction capabilities of conditions in interplanetary space, will be required to produce accurate forecasts of local geomagnetic disturbances beyond a forecast horizon of 1 h.

  • Open Access

    Reply to RC2

    by Spencer Mark Hatch, Stephen Omondi, Andreas Kvammen, Magnar Gullikstad Johnsen, Mathew J. Owens, Kristian Solheim Thinn, Rodrigo López

    In this study we examine how a deep-learning based forecast of local, ground-based geomagnetic field variations trained on solar wind parameters available in real time might be improved by including information contained in an accurate forecast of solar wind conditions. This is accomplished using a long short-term memory (LSTM) model together with magnetic field measurements made at the Rørvik magnetometer station in Mid-Norway. We use Advanced Composition Explorer (ACE) satellite measurements of solar wind and interplanetary magnetic field (IMF) conditions at the first Sun-Earth Lagrange point, and historical lists of coronal mass ejection (CME) impacts at Earth to train and validate the LSTM model. We find that accurate information about the IMF Bz component and solar wind speed are important for obtaining a reasonably accurate (r² ≥ 0.5) forecast of local geomagnetic activity over forecasting horizons beyond ~ 3 h. Information about CME arrival time is only important when simultaneously accompanied by accurate, relatively high-resolution information about IMF Bz. In the absence of the latter, CME arrival time information does not contribute to model performance. This empirical result amounts to a quantitative demonstration of the widely recognized impact of IMF orientation on CME geoeffectiveness. This result also highlights that new innovations, probably in the form of new prediction capabilities of conditions in interplanetary space, will be required to produce accurate forecasts of local geomagnetic disturbances beyond a forecast horizon of 1 h.

  • Open Access

    Reply to RC1

    by Spencer Mark Hatch, Stephen Omondi, Andreas Kvammen, Magnar Gullikstad Johnsen, Mathew J. Owens, Kristian Solheim Thinn, Rodrigo López

    In this study we examine how a deep-learning based forecast of local, ground-based geomagnetic field variations trained on solar wind parameters available in real time might be improved by including information contained in an accurate forecast of solar wind conditions. This is accomplished using a long short-term memory (LSTM) model together with magnetic field measurements made at the Rørvik magnetometer station in Mid-Norway. We use Advanced Composition Explorer (ACE) satellite measurements of solar wind and interplanetary magnetic field (IMF) conditions at the first Sun-Earth Lagrange point, and historical lists of coronal mass ejection (CME) impacts at Earth to train and validate the LSTM model. We find that accurate information about the IMF Bz component and solar wind speed are important for obtaining a reasonably accurate (r² ≥ 0.5) forecast of local geomagnetic activity over forecasting horizons beyond ~ 3 h. Information about CME arrival time is only important when simultaneously accompanied by accurate, relatively high-resolution information about IMF Bz. In the absence of the latter, CME arrival time information does not contribute to model performance. This empirical result amounts to a quantitative demonstration of the widely recognized impact of IMF orientation on CME geoeffectiveness. This result also highlights that new innovations, probably in the form of new prediction capabilities of conditions in interplanetary space, will be required to produce accurate forecasts of local geomagnetic disturbances beyond a forecast horizon of 1 h.

  • Open Access

    Comment on egusphere-2025-6298

    by Stephen Omondi, Spencer Mark Hatch, Andreas Kvammen, Magnar Gullikstad Johnsen, Mathew J. Owens, Kristian Solheim Thinn, Rodrigo López

    In this study we examine how a deep-learning based forecast of local, ground-based geomagnetic field variations trained on solar wind parameters available in real time might be improved by including information contained in an accurate forecast of solar wind conditions. This is accomplished using a long short-term memory (LSTM) model together with magnetic field measurements made at the Rørvik magnetometer station in Mid-Norway. We use Advanced Composition Explorer (ACE) satellite measurements of solar wind and interplanetary magnetic field (IMF) conditions at the first Sun-Earth Lagrange point, and historical lists of coronal mass ejection (CME) impacts at Earth to train and validate the LSTM model. We find that accurate information about the IMF Bz component and solar wind speed are important for obtaining a reasonably accurate (r² ≥ 0.5) forecast of local geomagnetic activity over forecasting horizons beyond ~ 3 h. Information about CME arrival time is only important when simultaneously accompanied by accurate, relatively high-resolution information about IMF Bz. In the absence of the latter, CME arrival time information does not contribute to model performance. This empirical result amounts to a quantitative demonstration of the widely recognized impact of IMF orientation on CME geoeffectiveness. This result also highlights that new innovations, probably in the form of new prediction capabilities of conditions in interplanetary space, will be required to produce accurate forecasts of local geomagnetic disturbances beyond a forecast horizon of 1 h.

  • Open Access

    Comment on egusphere-2025-6298

    by Stephen Omondi, Spencer Mark Hatch, Andreas Kvammen, Magnar Gullikstad Johnsen, Mathew J. Owens, Kristian Solheim Thinn, Rodrigo López

    In this study we examine how a deep-learning based forecast of local, ground-based geomagnetic field variations trained on solar wind parameters available in real time might be improved by including information contained in an accurate forecast of solar wind conditions. This is accomplished using a long short-term memory (LSTM) model together with magnetic field measurements made at the Rørvik magnetometer station in Mid-Norway. We use Advanced Composition Explorer (ACE) satellite measurements of solar wind and interplanetary magnetic field (IMF) conditions at the first Sun-Earth Lagrange point, and historical lists of coronal mass ejection (CME) impacts at Earth to train and validate the LSTM model. We find that accurate information about the IMF Bz component and solar wind speed are important for obtaining a reasonably accurate (r² ≥ 0.5) forecast of local geomagnetic activity over forecasting horizons beyond ~ 3 h. Information about CME arrival time is only important when simultaneously accompanied by accurate, relatively high-resolution information about IMF Bz. In the absence of the latter, CME arrival time information does not contribute to model performance. This empirical result amounts to a quantitative demonstration of the widely recognized impact of IMF orientation on CME geoeffectiveness. This result also highlights that new innovations, probably in the form of new prediction capabilities of conditions in interplanetary space, will be required to produce accurate forecasts of local geomagnetic disturbances beyond a forecast horizon of 1 h.

  • Open Access

    How well can we forecast local magnetic ground perturbations with existing space weather monitoring resources?

    by Stephen Omondi, Spencer Mark Hatch, Andreas Kvammen, Magnar Gullikstad Johnsen, Mathew J. Owens, Kristian Solheim Thinn, Rodrigo López

    In this study we examine how a deep-learning based forecast of local, ground-based geomagnetic field variations trained on solar wind parameters available in real time might be improved by including information contained in an accurate forecast of solar wind conditions. This is accomplished using a long short-term memory (LSTM) model together with magnetic field measurements made at the Rørvik magnetometer station in Mid-Norway. We use Advanced Composition Explorer (ACE) satellite measurements of solar wind and interplanetary magnetic field (IMF) conditions at the first Sun-Earth Lagrange point, and historical lists of coronal mass ejection (CME) impacts at Earth to train and validate the LSTM model. We find that accurate information about the IMF Bz component and solar wind speed are important for obtaining a reasonably accurate (r² ≥ 0.5) forecast of local geomagnetic activity over forecasting horizons beyond ~ 3 h. Information about CME arrival time is only important when simultaneously accompanied by accurate, relatively high-resolution information about IMF Bz. In the absence of the latter, CME arrival time information does not contribute to model performance. This empirical result amounts to a quantitative demonstration of the widely recognized impact of IMF orientation on CME geoeffectiveness. This result also highlights that new innovations, probably in the form of new prediction capabilities of conditions in interplanetary space, will be required to produce accurate forecasts of local geomagnetic disturbances beyond a forecast horizon of 1 h.

  • Open Access

    Forecasting of Auroral Pc5 Pulsations from Solar Wind Parameters Using Machine Learning Approach.

    by Stephen Owino Omondi, Justice Allotey Pappoe, Sebwato Nasurudiin, Ayman Mohamed Mahrous

    The coupling between the magnetosphere and solar wind contributes to the energy, momentum, and mass transfer between the systems. However, geomagnetic pulsations facilitate the continuation of this process in the magnetosphere and the production of discrete auroral arcs. Therefore, remote-sensing the magnetospheric conditions. Data analytics with machine learning (ML) gives insight into scalability, adaptability, and feature extraction compared to traditional empirical models. The availability of big data in the Svalbard network spanning 25years from 1996 motivated the current study. Hence, we present the forecasting of auroral Pc5 pulsations from solar wind parameters using the ML technique. In the training phase, there was a regression of 0.75 and MSE=11.90 nT2. The relationship between Pc5 forecast and observations in low and high geomagnetic activity and solar activity showed good consistency with R=0.76 and MSE= 11.4 nT2. For instance, the model adapted well to the St. Patrick geomagnetic storm of March 17th, 2015 despite uncertainties in the data. In addition, the model also adapted well with stunning performance in all Svalbard observatories with HOP leading with 6949 prediction events and NAL with the least. Thus, this was consistent with previous studies in terms of Pc5 pulsations latitudinal or L-shell dependence. Finally, validation with Kp and F10.7 indices presented excellent coherence between the models. Overall, The ML studied the connection between solar wind and interplanetary magnetic field properties to the ground magnetic field perturbations with good correlation results. Hence, the model will be fit for use by the magnetospheric community for space weather studies.

  • Open Access

    Forecasting of Auroral Pc5 Pulsations from Solar Wind Parameters Using Machine Learning Approach.

    by Stephen Owino Omondi, Justice Allotey Pappoe, Sebwato Nasurudiin, Ayman Mohamed Mahrous

    The coupling between the magnetosphere and solar wind contributes to the energy, momentum, and mass transfer between the systems. However, geomagnetic pulsations facilitate the continuation of this process in the magnetosphere and the production of discrete auroral arcs. Therefore, remote-sensing the magnetospheric conditions. Data analytics with machine learning (ML) gives insight into scalability, adaptability, and feature extraction compared to traditional empirical models. The availability of big data in the Svalbard network spanning 25years from 1996 motivated the current study. Hence, we present the forecasting of auroral Pc5 pulsations from solar wind parameters using the ML technique. In the training phase, there was a regression of 0.75 and MSE=11.90 nT2. The relationship between Pc5 forecast and observations in low and high geomagnetic activity and solar activity showed good consistency with R=0.76 and MSE= 11.4 nT2. For instance, the model adapted well to the St. Patrick geomagnetic storm of March 17th, 2015 despite uncertainties in the data. In addition, the model also adapted well with stunning performance in all Svalbard observatories with HOP leading with 6949 prediction events and NAL with the least. Thus, this was consistent with previous studies in terms of Pc5 pulsations latitudinal or L-shell dependence. Finally, validation with Kp and F10.7 indices presented excellent coherence between the models. Overall, The ML studied the connection between solar wind and interplanetary magnetic field properties to the ground magnetic field perturbations with good correlation results. Hence, the model will be fit for use by the magnetospheric community for space weather studies.

  • Open Access

    Wavelet Analysis for Automatic Detection of Pi-2 Pulsations during Substorm Onset Along the 210$^{\circ}$ Magnetic Meridian

    by Stephen Owino Omondi, Akimasa Yoshikawa, Teiji Uozumi, Yuki Obana, Moe Hayashi, Ayman Mohamed Mahrous

    Ground Pi-2 pulsations comprise superpositions of various modal components of shear and fast Alfven waves, field line resonance, and plasmaspheric resonances. These complex waveforms, hard to resolve with Fourier transforms are successfully characterized by wavelet techniques. Wavelet detection employs decomposition and reconstruction modes to characterize time-frequency components. Hence, suitable for the examination of the locality and complexity of natural signal patterns. The current study presents the automatic detection of Pi-2 pulsations using Daubechies and Morlet wavelet transforms. In the study, distinct Pi-2 events from CPMN stations along 210${^\circ}$ magnetic meridian were detected. Global Pi-2 pulsations with harmonious H oscillations and discrete D bays in the sub-aurora zone suggest a common source with diverse tunneling paths. Scalograms of Pi-2 undulations of the frequency band of 6.7-22 mHz were observed despite different kinds of Pi-2s. Auroral Pi-2s were highly localized in local time with clear H and D bays, implying magnetospheric-ionospheric current couplings. Latitudinal and longitudinal Pi-2 propagations are exemplified by 180${^\circ}$ phase-shift (polarization) in EWA and group delay in the mid-latitudes of the northern hemisphere. Overall, Pi-2 wave power from high to low latitudes declined with peak amplitudes of 15 nT to less than 1 nT, respectively. Finally, external influences from sea currents causing signal attenuation due to the station's proximity to the sea were also identified. To conclude, the accuracy and efficiency of wavelet analysis with no computation hassle render it a valuable tool for the study of space events in the magnetospheric community.