Seasonal Patterns and Forecasting of CO and Ozone Using Singular Spectrum Analysis in a Tropical Urban Environment

Amaury de Souza — 2026-01-02

Singular Spectrum Analysis (SSA) was applied to daily time series of carbon monoxide (CO) and ozone (O₃) observed between 2000 and 2018 in Campo Grande, MS, Brazil, to identify seasonal patterns, long-term variability, and evaluating the predictive capacity of the technique. The methodology involved the decomposition of the series into structural components and subsequent prediction using the Linear Recurrence Formula (LRF). The analysis revealed strong and persistent annual seasonality for both pollutants, particularly for CO, whose maximum concentrations occur between August and October, coinciding with the dry season and intensified biomass-burning activity. SSA proved effective in extracting low-frequency components, including trend and seasonal cycles, providing a clear representation of the dominant temporal structure of both pollutants. Forecasting results indicated that SSA-LRF successfully reproduced the main seasonal behavior of O₃, while daily prediction skill remained limited, as reflected by negative R² values during the validation period. For CO, the highly irregular and episodic nature of fire-related peaks resulted in larger forecast errors and reduced predictive skill. These results highlight that univariate SSA is more suitable for reconstructing and predicting low-frequency pollutant dynamics than short-term daily variability. The findings demonstrate that SSA is a robust exploratory and decomposition tool for air-quality time series in tropical environments, particularly for identifying seasonal and structural patterns. For operational forecasting of pollutants with strong volatility, such as CO, hybrid approaches combining SSA with statistical or machine-learning models are recommended to improve predictive performance.

Comparative Analysis of Traditional and Machine Learning Models for Rainfall Forecasting in Barishal District of Bangladesh

Shawrab Chandra — 2026-01-05

Rainfall prediction is crucial for agricultural planning and water resource management, as Bangladesh’s agriculture heavily depends on rainfed irrigation. Existing forecasting models are complex and costly, both budgetarily and computationally. As a result, our study evaluates the comparative performance of forecasting models, comprising two traditional time series models (Exponential Smoothing (ES) and Seasonal Autoregressive Integrated Moving Average (SARIMA)), and one machine learning model (Long Short-Term Memory (LSTM)). The monthly rainfall data for Barishal, Bangladesh, spanning the period from 1970 to 2022, were obtained from the Bangladesh Meteorological Department. The models' performance was assessed using root mean square error (RMSE), mean absolute error (MAE), correlation coefficient (R), Nash-Sutcliffe efficiency coefficient (NSEC), and Kling-Gupta Efficiency (KGE). The ES and SARIMA models perform closely. With RMSE, MAE, R, NSEC, and KGE values of 109.35, 73.60, 0.79, 0.62, and 0.74, respectively, the ES model performs better than the SARIMA model. On the other hand, the machine learning model LSTM struggled with the test data, resulting in a higher RMSE (150.34), MAE (100.95), and lower R (0.60), NSEC (0.27), and KGE (0.60) values. This indicates that for the small dataset, the LSTM machine learning model is less effective. Therefore, our suggestion is to use a statistical model, especially the ES model, to forecast monthly rainfall in the Barishal division, as it is effective and computationally efficient. These findings are beneficial for policy development, the pesticide industry, tourism, event management, water conservation, and predicting floods and droughts.

Journal of Atmospheric Science Research

Seasonal Patterns and Forecasting of CO and Ozone Using Singular Spectrum Analysis in a Tropical Urban Environment

Comparative Analysis of Traditional and Machine Learning Models for Rainfall Forecasting in Barishal District of Bangladesh