The Deep Space Climate Observatory (DSCOVR), can measure the strength and speed of the solar wind in space, which enables us to predict geomagnetic storms that can severely impact important systems like GPS and electrical power grids on Earth. It acts like a sensor buoy at sea that warns of an oncoming tsunami.
DSCOVR can warn forecasters 15 to 60 minutes before a storm of particles and magnetic field, known as a coronal mass ejection (or CME) reaches earth. However, it continues to operate past its expected lifetime and produces occasional faults that may themselves be indicators of space weather (the launch was in Feb. 11 2015, the arrival l at Sun–Earth L1 Lagrange point in Jun 8, 2015, and was expected to last 5 years). The instrument onboard DSCOVR that measures the solar wind's magnetic field continues to function very well, the instrument that measures the solar wind density, temperature, and speed has lost sensitivity and experiences faults and anomalies from time to time.
DSCOVR orbits about a million miles from Earth in a unique location called Lagrange point 1 (L1), which allows it to hover between the Sun ☀️ and our planet 🌎. From that vantage point, DSCOVR measures the plasma that may cause geomagnetic storms hours before it reaches us– ideally providing an early warning of what’s coming our way. The time that it takes for that plasma to reach Earth and trigger a geomagnetic storm might be anywhere from about 15 minutes to a few hours.
DSCOVR uses 2 main space weather instruments:
- Magnetometer: Measures solar wind magnetic field vector. Given dataset has 3 features: x, y and z.
- Faraday Cup: Provides real-time measurement of solar wind proton density, speed, velocity, temperature, etc... NASA and NOAA (SWPC and OSPO) are actively monitoring Faraday Cup behavior in various solar wind conditions, and developing or updating flight software solutions to optimize the instrument behavior. The most recent modification was performed in August 2020, and additional work is ongoing. Given dataset has 50 features.
DSCOVR uses measurements of the solar wind density, temperature, speed, and magnetic field to run computer simulations of the Earth's magnetic field and atmosphere. Based on those simulations, NOAA forecasts when a geomagnetic storm will occur and how strong it will be. The strength of the geomagnetic storm is measured on a scale called the Planetary K-index (Kp). The K-index is a scale ranging from 0 to 9, with higher values indicating more severe geomagnetic disturbances. Here's what the different K-index values represent:
Kp = 0: Very quiet geomagnetic conditions.
Kp = 1: Quiet geomagnetic conditions with minimal disturbance.
Kp = 2: Slightly disturbed geomagnetic conditions.
Kp = 3: Unsettled geomagnetic conditions.
Kp = 4: Active geomagnetic conditions.
Kp = 5: Minor geomagnetic storm.
Kp = 6: Major geomagnetic storm.
Kp = 7: Severe geomagnetic storm.
Kp = 8: Very severe geomagnetic storm.
Kp = 9: Extremely severe geomagnetic storm.
As a part of the NASA Space Apps Challenge, the project must predict the risk of geomagnetic storms on a 3 hour basis on Earth following the K-Index Scale. This project uses Machine and Deep Learning techniques to develop models that can improve the actual predictions given by NOAA (National Oceanic and Atmospheric Administration)
- EDA: This exercise has been iterative with and executed simultaneously with the Data Cleaning and Feature Engineering ones, allowing to identify potential improvements that have been implemented in the later stages, and so on. The EDA starts some statistics to later on, display several visualizations with the aim of detecting patterns and correlation between Kp and Ap and the vectors measures of PlasMAG and the different FC measurements. Found in notebook: 1_Initial_EDA_DSCOVR_data.ipynb
- Data Wrangling: Considering time constraints and the current intervals of Kp, team has decided to aggregate the data transforming it from minutes to 3 hour interval. Several statistical features covering the measurements that have taken place inside this 3-hours have been created. Many of the rows are NaN but they are registered as 0. They could be an anomaly, and they can also be indicator of space weather changes, so the approach of discarding the rows, or filling the NaN with other statistical measurements has been rejected by the group. Instead, we have transformed this null existance in a categorical feature, creating new features for each of our variables as ‘Proportion of NaN values’. The period where DSCOVR was shut down (from 27 june 2019 to 2 march 2020) has been discarded. More information could be found in notebook: 2_Cleaning_and_engineering_DSCOVR.ipynb
- Model Development: In our work, we have tried both regression and classification models. The model presented here is a binary classification model that constitutes the first step towards building more complex models. The target variable Kp has been classified into higher or equal to 5 (1) in order to indicate an incoming solar storm, and less than 5 (0). More information can be found in notebook 3
Here's a brief high-level overview of the tech stack the project uses:
The data has been splitted into train and test. Train dataset consists of the first 80% of the time series data from 2016. Test dataset consists of the last 20% of the data till 2023.
The predictor variables are all the variables coming from the DSCVR and the feature engineering performed in previous stage plus the historical Kp values. The target values are future Kp values. In this case, the sequence includes 240 previous datapoints for training, and 8 future points for forecasting.
Among several architectures tried, an LSTM-based model is presented. It consists of 100 LSTM units, followed by a dense layer with dropout and a final dense layer with 8 neurons and sigmoid as activation function to predict the probabilities of the class in the next 8 time steps.
Since the output of the model are the probabilities, a user-defined threshold is set based on the training data using ROC-AUC methodology.
Finally, a chart comparing the predicted targets in the train set, and the predicted targets in the test set are shown.
Due to time constraints, really complex models, or accurate finetuning, has been difficult to reach. Also, the final aim is to improve also the front page and stablish and alert system for the stakeholders affected. The team has directed its efforts into preparing the machine learning pipeline and working on the models, as the predictions made are considered more important for this challenge than the way in which they are displayed.
The team has already created a plan that wasn’t implemented on this first iteration due to the time and computational capabilities constraints. The next steps to be performed are as follow:
- Improve the feature engineering: Feature engineering is an iterative process. For us, it implies going back and forth to add/drop features and re-run models to check their impact on performance. Our plan is to add some data that can be also used by NASA (in instance, Canada Solar Flux data of the prior day: not as it is, but prepared as can be seen in notebook: ….). Also, a detailed correlation analysis, application of tree based models for detecting feature importance scores and dimensionality reduction techniques are some of the next steps planned at this point.
- Improve the models: basically on 2 different lines:
- Try more complex models that have not been tried, mainly because they required more running time, so applying them could have been an one shot exercise.
- Improve finetuning parameter of the used models.
- Improve the data-serving: Once improving our forecast for the next 3 days 3-hour intervals, the team would like to improve the visualization style of the page. Also, vinculate the forecast with a mailing list that contains the main stakeholders affected for strong space electromagnetic storms, creating a system of alerts, would be a good option to prevent and mitigate potential risks. Finally, allowing the users to access to the comparison between historical and predicted values (and even making our model open-source) could deal to future iterations that improve the final work (despite the actual deliverables are just a first sketch.
- Khurana, S., & Collado-Vega, Y. M. "Geomagnetic Storms: A Study of the Relationship between Geomagnetic Storms and the Interplanetary Magnetic Field, and Monitoring Geomagnetic Storms in the Ionosphere with GPS Errors
- Matzka, J., Stolle, C., Yamazaki, Y., Bronkalla, O. and Morschhauser, A., 2021. The geomagnetic Kp index and derived indices of geomagnetic activity. Space Weather, https://doi.org/10.1029/2020SW00264
- Lockwood, M. (2022). “Solar Wind—Magnetosphere Coupling Functions: Pitfalls, Limitations, and Applications”.
 Alfonso Hervella (MSc Big Data and AI) 💻📊 |
 Javier Nieves (MSc Big Data and AI) 💻 |
 callysthenes (MSc Big Data and AI) 💻📝 |