Please login first
Causality Inference and Predictive Modeling for Mitigating Atmospheric Pollution in Green Ports: A Castelló Port Case Study
* , , ,
1  Centro Tecnológico Naval y del Mar
Academic Editor: Francisco Falcone


The growth of commercial activities and the need for competitiveness in the global market have prompted ports worldwide to pursue optimization and cost reduction while mitigating adverse environmental effects. Among these effects, air pollution from emissions generated by maritime traffic and the dispersion of particles, as well as water pollution from spills, pose significant challenges. In response, the concept of Green Ports has emerged, placing sustainability at the core of port development and operations, encompassing social, economic, and environmental aspects. The aim is to achieve a balanced approach that fosters competitiveness, integration with the host city, and environmental stewardship. This paper focuses on addressing a real pollution problem faced by Castelló Port (APC), specifically atmospheric pollution resulting from the dispersion of suspended particles during the loading and unloading of solid bulk cargo on ships at the port's docks (CS05, CS09, CS06 and CS26).

The primary objective of this study is to anticipate episodes of atmospheric pollution related to cargo handling activities and assess the quantitative causality between these variables. We employ causality inference and predictive methods based on time series analysis to investigate the applicability and validity of these techniques in a real-world problem setting. Concretely, causality models such as CCM, CMI and PCMCI along with forecasting models such as ARIMA, SARIMA and LSTM are used. Data provided by the APC for the years 2019-2021 encompass both port operations (cargo handling) and air quality parameters collected through the air quality monitoring network (PM2.5, PM10, wind direction, wind speed, maximum wind speed, temperature, relative humidity, precipitation).

By analyzing the available data and applying predictive models, we aim to gain insights into the occurrence of atmospheric pollution episodes resulting from cargo handling activities. Additionally, we seek to quantify the causal relationships between these variables, providing valuable information for decision-making processes. Thereby, this study contributes to the understanding of the feasibility and effectiveness of predictive techniques in addressing real pollution challenges in port environments.

Results show that the studied cargo handling in the studied docks are particularly influential on the PM measurements. Concretely, bulk discharges from CS05 and CS09 stand out as the most responsible variables for the PM dynamics (PM2.5, particularly). In addition, regarding causality times, significant causal relationships appear for short (1h), intermediate (5h) and long (10h) times. Regarding forecasting results, whilst classic predictive models such as ARIMA and SARIMA have shown proper results, LSTM networks provide more accurate results, achieving an accuracy above 85% on the environmental parameters.

Keywords: Green Ports; sustainability; atmospheric pollution; predictive modeling; time series forecasting; causality; air quality monitoring; cargo handling; Castelló Port.