Vol. 2 No. 1 (2026)
Sastra Journal : Math Earth
We explore the inherent mathematical beauty and underlying numerical laws governing our planet. This collection of papers posits that Earth is the ultimate laboratory for applied mathematics, examining phenomena from the logarithmic spirals of weather systems to the statistical distribution of seismic events. The contributors bridge the gap between abstract number theory and tangible environmental realities, revealing how humanity's efforts to measure, model, and conserve the planet are inextricably linked to the language of mathematics. It is an invitation to view the world not just through observation, but through quantification.
Published: 2025-08-10
Articles
Stochastic Differential Equations in Global Epidemiology: Modeling Pathogen Spread via Climate Vectors
Abstract
The interconnectedness of global travel and accelerating climate change has fundamentally altered the spread of infectious diseases. This paper moves beyond traditional SIR/SEIR models by introducing a framework based on stochastic differential equations (SDEs) to model the global dispersion of vector-borne pathogens. Our 'Math Earth' approach integrates large-scale climate datasets (temperature and precipitation anomalies) with global air traffic network data. This allows us to model the changing habitat suitability for vectors like the *Aedes aegypti* mosquito and predict new regions at risk for outbreaks of diseases such as dengue and Zika. The stochastic component of the model effectively captures the inherent randomness of transmission events and climate fluctuations. The findings provide a probabilistic risk map, highlighting emerging hotspots and demonstrating the power of advanced mathematical modeling in global public health preparedness and climate-resilient disease control strategies.
Mathematical Modeling of Subsurface Aquifer Dynamics: A Case Study
Abstract
The sustainable management of groundwater resources is a critical challenge for arid and semi-arid regions worldwide. This study presents a comprehensive mathematical model to simulate subsurface aquifer dynamics. We employ a finite element method (FEM) to solve the governing partial differential equations (PDEs) of porous media flow, incorporating factors such as hydraulic conductivity, anisotropic soil properties, and variable recharge rates from precipitation and irrigation. By assimilating satellite-derived data (e.g., GRACE) and in-situ well-level measurements, our model provides a high-resolution spatiotemporal forecast of aquifer depletion and land subsidence. The results offer a robust tool for policymakers to evaluate the long-term impacts of different water management strategies, such as managed aquifer recharge (MAR). This research underscores the essential role of applied mathematics in quantifying and preserving one of Earth's most vital, yet invisible, resources.
Fractal Geometry in Geomorphology: Quantifying Coastline Complexity and Erosion Patterns
Abstract
Mandelbrot's observation that coastlines are not smooth but possess self-similar, fractal properties revolutionized geomorphology. This study moves beyond simple fractal dimension (D) calculations to develop a multi-scalar framework for quantifying coastline complexity and its relationship to erosion and deposition patterns. Using high-resolution satellite imagery (LiDAR and multispectral data), we apply wavelet-based fractal analysis to measure how the fractal dimension changes across different spatial scales. This multi-scalar signature allows us to differentiate between coastlines dominated by various geomorphic processes (e.g., tectonic uplift, deltaic progradation, tidal erosion). We establish a robust mathematical correlation between a declining fractal dimension and increased coastal vulnerability to sea-level rise. This 'Math Earth' approach provides a novel, quantitative tool for coastal zone management, enabling the identification of erosion hotspots and the optimization of coastal defense strategies by treating the coastline as the complex dynamical system it is.
A Review of Inverse Problems in Seismic Tomography: Mathematical Approaches to Imaging Earth's Interior
Abstract
Understanding the structure of Earth's deep interior relies on solving highly complex inverse problems using seismic wave data. This systematic review synthesizes mathematical and computational advancements in seismic tomography. We trace the evolution from linear, ray-based methods to the current frontier of Full-Waveform Inversion (FWI), a non-linear optimization challenge that leverages the complete information in seismic waveforms. The paper critically analyzes the mathematical underpinnings of each approach, including the choice of objective functions, regularization techniques (e.g., Tikhonov, total variation), and the optimization algorithms (e.g., L-BFGS, adjoint-state methods) required to navigate the high-dimensional, non-convex search space. We also explore the role of high-performance computing in making these methods feasible. By examining the successes and persistent challenges, such as cycle-skipping, this review provides a comprehensive roadmap for mathematicians and geophysicists working to create higher-fidelity images of Earth's mantle and core.
Non-Linear Wave Dynamics: Modeling Rogue Waves in Open Ocean Environments
Abstract
Rogue waves—extreme, unexpected, and highly localized surface waves—pose a significant threat to maritime navigation and offshore structures. For decades, their existence was debated, as traditional linear wave theory deemed them statistically impossible. This research applies principles of non-linear mathematical physics to model the genesis and dynamics of these extreme events. We focus on the non-linear Schrödinger (NLS) equation and its extensions, such as the Dysthe equation, which incorporate modulational instability as a primary mechanism for rogue wave formation. By simulating complex, multi-directional wave fields (sea states), our model demonstrates how energy from surrounding waves can become rapidly focused into a single, massive pulse. This work validates the NLS framework against real-world buoy data and satellite observations of rogue wave events. This study bridges the gap between theoretical mathematics and physical oceanography, providing a predictive mathematical tool to assess the probability of rogue wave occurrence in specific ocean basins.
A Review of Coupled General Circulation Models (GCMs): The Mathematical Core of Modern Climate Prediction
Abstract
The foundation of modern climate science rests on General Circulation Models (GCMs), massive computational systems that simulate the Earth's climate. This comprehensive review examines the mathematical core of these models, focusing on the coupled systems of partial differential equations that govern the planet. We analyze the formulation of the atmospheric component (based on the primitive equations, a form of the Navier-Stokes equations for fluid dynamics), the oceanic component (similarly based on fluid dynamics but with different boundary conditions), and the sea-ice and land-surface sub-models. The paper delves into the critical role of 'parameterization'—the mathematical approximation of physical processes, such as cloud formation and radiative transfer, that are too small or complex to be resolved directly. We discuss the numerical methods used to solve this stiff, multi-scale system, and the mathematical challenges in ensuring model stability and energy conservation over long-term simulations. This review serves as a 'Math Earth' primer for researchers, illustrating how climate prediction is fundamentally an applied mathematics problem.
Zero-Day Exploit Detection: A Behavioral Analysis Approach
Abstract
Traditional signature-based antivirus and intrusion detection systems are fundamentally reactive, proving ineffective against zero-day exploits for which no signature yet exists. This research addresses this critical gap by proposing a proactive detection framework based on behavioral anomaly analysis. Our model does not rely on prior knowledge of attack vectors. Instead, it establishes a baseline of normal system behavior by monitoring a high-dimensional feature set, including system call frequencies, network traffic patterns, and memory allocation changes. We employ an unsupervised machine learning model, specifically an isolation forest combined with a Long Short-Term Memory (LSTM) autoencoder, to identify subtle deviations from this baseline that are indicative of malicious activity. The system was trained on a large dataset of benign software behavior and subsequently tested in a high-fidelity sandbox environment against a suite of known and novel exploit kits. Our framework demonstrated a 98.2% detection rate for zero-day-style attacks while maintaining a low false positive rate of 1.5%, offering a resilient and adaptive defense mechanism for modern cybersecurity challenges.