SMRAI with SuperDARN Data Assimilation for Substorm Analysis

SMRAI with SuperDARN Data Assimilation for Substorm Analysis This repository contains analysis workflows for comparing SMRAI2.1 ionospheric electrodynamics with SuperDARN data assimilation (DA) during isolated substorm events. The notebooks use OMNI solar wind data, SMRAI2 machine-learning models, and SMRAI2.1 potential/electric field outputs to compute AL/AU indices and produce spatial visualizations. Overview SMRAI2: Provides electric potential and Hall conductivity via trained models driven by solar wind inputs (Refer Kataoka et al.,2024). SMRAI2.1: Potential/electric field files are used for auroral current calculations (Refere Nakano et al., 2025). Analysis: AL/AU index time series and multi-panel polar plots for selected substorm intervals (e.g., 2023-10-30, 2023-10-31). Notebooks 1. predict_al_au_figure2.ipynb Purpose: Compute and compare AL/AU indices from SMRAI2.1 control and SuperDARN DA runs, and produce time series comparison plots (Figure 2–style output). Workflow: Downloads OMNI 5-minute data for a user-specified time range. Default dates: 2023-10-30 or 2023-10-31 (configurable in the notebook). Uses SMRAI2 to predict: Hall conductivity Loads electric field files for SMRAI2.1 control (no data assimilation) and SMRAI2.1 with SuperDARN DA (SMRAI2.1_DAssim) Interpolates conductivity fields from the SMRAI2 grid onto the SMRAI2.1 spatial grid for consistent comparison. Computes AL and AU from derived Hall currents in the auroral zone (60–70° latitude band). Produces two AL/AU series: SMRAI2.1 control (without data assimilation) and SMRAI2.1 with SuperDARN data assimilation (SMRAI2.1_DAssim). Time series comparison plots showing the performance of each approach over the selected interval. 2. four_para_plot_figure4.ipynb Purpose: Generate detailed spatial visualizations at selected times for model outputs and SuperDARN coverage (Figure 4–style multi-panel plots). Workflow: You can change the time and date in the notebook to focus on specific substorm phases. Ionospheric potential fields for: SMRAI2.1 SMRAI2.1 with SuperDARN DA Conductivity distributions SuperDARN observational coverage (where DA observations are available) Visual inspection of model outputs and their spatial relationships during chosen substorm phases. Requirements Typical dependencies (install as needed for your environment): Custom modules such as esn_dts_openloop (Refer Kataoka et al, 2024) numpy, matplotlib, pandas pyspedas, pytplot (OMNI and time series) geopack (e.g., for coordinate/IMF handling) scikit-learn (e.g., for PCA/ML components in SMRAI2) Usage Summary AL/AU and time series (Figure 2) Run predict_al_au_figure2.ipynb. Set the date range (e.g., 20231030 or 20231031); the notebook will download OMNI data, run SMRAI2, load SMRAI2.1 fields, interpolate conductivity, compute AL/AU, and plot comparisons. Spatial snapshots (Figure 4) Run all_para_plot_figure4.ipynb. Set the date and time of interest, then execute to generate multi-panel polar plots of potentials, conductivities, and SuperDARN data coverage.