We propose a methodology to identify and characterize different plasma environments observed by magnetospheric mission spacecraft instruments, creating a data-driven magnetosphere atlas. To enable an accurate analysis under highly non-uniform sampling, we design a 3D adaptive octree data structure that refines magnetospheric regions with high sampling counts. A variational autoencoder that leverages the octree structure allows us to automatically identify different regions of the Earth’s magnetosphere. The model also generates per-region feature distributions, improving interpretability by summarizing both typical conditions and variability. We apply this methodology to 10 years of plasma observations from NASA’s Magnetospheric Multiscale (MMS) mission, identify known magnetospheric regions, and create an atlas of plasma environments in space. The resulting atlas can be used to better understand magnetospheric plasma regions, detect anomalies, automatically generate labels for space applications, and study the magnetospheric response across geomagnetic activities.

The Disturbance Storm Time (Dst) index quantifies geomagnetic storm intensity by measuring global magnetic field variations. In this study, we apply interpretable machine-learning (ML) techniques to derive data-driven models describing the temporal evolution of the Dst index. We use historical data from the NASA OMNIWeb database, including solar wind density, bulk velocity, convective electric field, dynamic pressure, and magnetic pressure. We employ KAN networks and the symbolic regression framework PyOperon, based on an evolutionary algorithm, to identify closed-form expressions linking (Formula presented) to key solar wind parameters. The equations obtained via symbolic regression form a hierarchy of complexity levels and capture nonlinear dependencies and threshold effects in Dst evolution. In addition, we use a conventional MLP network as a reference black-box model. We benchmark all ML models against observed Dst data and compare their performance with empirical formulations such as the Burton-McPherron–Russell and O’Brien-McPherron models. The performance evaluation on historical storm events includes the 2003 Halloween storm, the 2015 St. Patrick’s Day storm, a moderate storm in 2017, and the extreme storm of May 2024. The data-driven models, particularly the MLP, demonstrate superior accuracy in most cases. While the symbolic regression expressions provide insight into the underlying physics, the results highlight an intrinsic trade-off between model interpretability and predictive accuracy. This is an extended version of a previous work presented in Markidis et al. (2025) [1].

Particle-in-Cell (PIC) Monte Carlo (MC) simulations are central to plasma physics but face increasing challenges on heterogeneous HPC systems due to excessive data movement, synchronization overheads, and inefficient utilization of multiple accelerators. In this work, we present a portable, multi-GPU hybrid MPI+OpenMP implementation of BIT1 that enables scalable execution on both Nvidia and AMD accelerators through OpenMP target tasks with explicit dependencies to overlap computation and communication across devices. Portability is achieved through persistent device-resident memory, an optimized contiguous one-dimensional data layout, and a transition from unified to pinned host memory to improve large data-transfer efficiency, together with GPU Direct Memory Access (DMA) and runtime interoperability for direct device-pointer access. Standardized and scalable I/O is provided using openPMD and ADIOS2, supporting high-performance file I/O, in-memory data streaming, and in-situ analysis and visualization. Performance results on pre-exascale and exascale systems, including Frontier (OLCF-5) for up to 16,000 GPUs, demonstrate significant improvements in run time, scalability, and resource utilization for large-scale PIC MC simulations.

Lambert's problem is a boundary value problem that arises in preliminary mission design, where orbital transfers must be determined given orbital location and a prescribed time-of-flight. The related equations are non-invertible and have traditionally been solved using iterative methods. In this work, we study the bounded one-revolution elliptic Lambert problem and propose a neural-surrogate approach that combines geometry-aware normalization with the prediction of the transfer semi-major axis. The normalization maps transfer instances with geometry-dependent admissible ranges into a common canonical representation, on which a multilayer perceptron (MLP), DeepONet, and Kolmogorov-Arnold Networks (KAN) are trained and compared. Among the tested models, the MLP achieves the highest predictive accuracy, while the structured architectures provide an alternative view of the normalized solution space. In Earth-Mars and multi-planetary flyby case studies, the surrogates preserve the broad features of the classical solutions and identify promising transfer regions.

Understanding Earth's magnetosphere requires analyzing complex, multiscale interactions captured through diverse in-situ measurements such as particle distributions (e.g., ion spectra and flow velocities), and electromagnetic fields. Modern space missions, including NASA's MMS and THEMIS, produce large volumes of such time-series data, where scientifically relevant events often appear as short-lived, context-dependent outliers. In this work, we investigate the structure of time-windowed, multi-feature datasets from Earth's magnetosphere and show that their information content can be effectively represented using a reduced set of largely time-invariant principal components. Using Principal Component Analysis (PCA) for static datasets and Incremental PCA for streaming observations, we characterize these components and leverage the associated reconstruction error to develop an unsupervised outlier detection method suitable for evolving data distributions. Building on our earlier approach, this method robustly identifies significant plasma phenomena in both structured and streaming-mode measurements. Applied to MMS and THEMIS observations, it successfully recovers known events, including bow shock and magnetopause crossings, foreshock transients, and plasma bubbles, while also highlighting additional scientifically relevant structures. This demonstrates that dimensionality-reduction-based outlier detection provides an effective pathway for automated event identification in complex space plasma datasets. This is an extension of previous work presented in Ekelund et al. (2025).

Analyzing multi-featured time series data is critical for space missions making efficient event detection, potentially onboard, essential for automatic analysis. However, limited onboard computational resources and data downlink constraints necessitate robust methods for identifying regions of interest in real time. This work presents an adaptive outlier detection algorithm based on the reconstruction error of Principal Component Analysis (PCA) for feature reduction, designed explicitly for space mission applications. The algorithm adapts dynamically to evolving data distributions by using Incremental PCA, enabling deployment without a predefined model for all possible conditions. A pre-scaling process normalizes each feature’s magnitude while preserving relative variance within feature types. We demonstrate the algorithm’s effectiveness in detecting space plasma events, such as distinct space environments, dayside and nightside transients phenomena, and transition layers through NASA’s MMS mission observations. Additionally, we apply the method to NASA’s THEMIS data, successfully identifying a dayside transient using onboard-available measurements.

Graphics Processing Units (GPUs) have become the standard in accelerating scientific applications on heterogeneous systems. However, as GPUs are getting faster, one potential performance bottleneck with GPU-accelerated applications is the overhead from launching several fine-grained kernels. CUDA Graph addresses these performance challenges by enabling a graph-based execution model that captures operations as nodes and dependence as edges in a static graph. Thereby consolidating several kernel launches into one graph launch. We propose a performance optimization strategy for iteratively launched kernels. By grouping kernel launches into iteration batches and then unrolling these batches into a CUDA Graph, iterative applications can benefit from CUDA Graph for performance boosting. We analyze the performance gain and overhead from this approach by designing a skeleton application. The skeleton application also serves as a generalized example of converting an iterative solver to CUDA Graph, and for deriving a performance model. Using the skeleton application, we show that when unrolling iteration batches for a given platform, there is an optimal size of the iteration batch, which is independent of workload, balancing the extra overhead from graph creation with the performance gain of the graph execution. Depending on workload, we show that the optimal iteration batch size gives more than $1.4 \times$ speed-up in the skeleton application. Furthermore, we show that similar speed-up can be gained in Hotspot and Hotspot3D from the Rodinia benchmark suite and a Finite-Difference Time-Domain (FDTD) Maxwell solver.

Space missions increasingly deploy high-fidelity sensors that produce data volumes exceeding onboard buffering and downlink capacity. This work evaluates FPGA acceleration of neural networks (NNs) across four space use cases on the AMD ZCU104 board. We use Vitis AI (AMD DPU) and Vitis HLS to implement inference, quantify throughput and energy, and expose toolchain and architectural constraints relevant to deployment. Vitis AI achieves up to 34.16 × higher inference rate than the embedded ARM CPU baseline, while custom HLS designs reach up to 5.4 × speedup and add support for operators (e.g., sigmoids, 3D layers) absent in the DPU. For these implementations, measured MPSoC inference power spans 1.56.75 W, reducing energy per inference versus CPU execution in all use cases. These results show that NN FPGA acceleration can enable onboard filtering, compression, and event detection, easing downlink pressure in future missions. 

Artificial Intelligence (AI) has the potential to revolutionize space exploration by delegating spacecraft decisions to onboard AI. The onboard neural-network (NN) will have parameters that can be updated onboard by telecommands after training on ground. However, Satellite uplinks have limited bandwidth and transmissions can be costly. Furthermore, a suboptimal NN will miss valuable scientific data. Smaller networks can therefore decrease the uplink cost and increase the value of the downlinked data. In this work, we evaluate and discuss using reduced-precision and small NNs to reduce the upload time. As an example of a mission where AI could be used, we focus on NASA's Magnetospheric MultiScale (MMS) mission. We showcase how an onboard AI can be used in the Earth's magnetosphere to classify data for selective downlink or recognize a region of interest to trigger a burst-mode, collecting data at a high-rate. Using a simple algorithm, we show the detection of a region of interest in on a stream of classifications. To provide the classifications, we use a Convolutional Neural Network (CNN) trained to an accuracy >94%. We show how the NN can be reduced to a single linear layer without accuracy loss. Thereby, Reducing the upload time by up to 98.9%. Each network can be reduced further by using lower-precision formats, changing the accuracy by less than 0.6 percentage points.