Innovative Methods for Multi-source Air Dust Pollution Monitoring using Remote Sensing: A Comparative Study of Chalkidiki (Greece) and Ihalainen (Finland)

The air we breathe is a complex mixture of gases and tiny solid or liquid particles suspended within it. Some of these particles occur naturally, kicked up by wind from deserts or released by volcanoes, while others are created by human activities like burning fuel, driving cars, or digging into the earth for minerals. When these particles are small enough to be inhaled, they can irritate the lungs, trigger asthma, and contribute to serious heart and lung diseases over time. Because these particles are invisible to the naked eye and can travel long distances, tracking them is difficult. Scientists have traditionally relied on physical sensors placed on the ground to measure pollution, but these sensors are expensive to build and maintain, leaving vast areas of the world without any direct measurements. To fill these gaps, researchers have turned to satellites orbiting overhead and powerful computer models that simulate how the atmosphere moves and changes. These tools offer a way to see pollution across entire regions, but they are not perfect; satellites can be blocked by clouds, and computer models often smooth out the sharp, local details of pollution that happen right next to a factory or a mine.

This study set out to test how well these high-tech tools work when applied to two very different mining regions: one in the hot, dry climate of Greece and another in the cold, forested landscape of Finland. The researchers focused on a specific type of pollution called PM10, which refers to dust particles smaller than ten micrometers—roughly the width of a human hair. They wanted to know if data from satellites and computer models could accurately track the dust generated by mining operations in these contrasting environments, and whether these methods could help predict future air quality. By comparing the satellite and model data against actual measurements taken by ground sensors at both sites, the team aimed to create a reliable, scalable method for monitoring dust pollution in areas where traditional monitoring networks are sparse or non-existent.

The researchers examined two distinct locations to see how different climates and mining types affect dust levels. The first site was in Chalkidiki, Greece, an area with a Mediterranean climate characterized by hot, dry summers and complex terrain. Here, mining operations extract metals like gold, silver, lead, and zinc. The second site was in Ihalainen, Finland, a region with a cold, boreal climate where a company mines limestone and produces lime. While the Greek site deals with metal-bearing dust and frequent intrusions of sand from the Sahara desert, the Finnish site deals with carbonate dust and seasonal issues like road dust that becomes loose when the snow melts. The team gathered data from 2013 to 2024, combining hourly readings from twelve ground stations in Greece and seven in Finland with data from two major sources: the Copernicus Atmosphere Monitoring Service, which uses computer models to estimate pollution levels, and MODIS, a satellite instrument that measures how much sunlight is blocked by dust in the atmosphere.

When the researchers compared the ground measurements with the satellite and model data, they found that the results varied significantly depending on the location and the tool used. In Greece, the ground sensors recorded much higher levels of dust, with concentrations reaching up to 347 micrograms per cubic meter, compared to a maximum of 132 micrograms per cubic meter in Finland. The ground data also showed that pollution in Greece was highly uneven; some areas near the mines had much higher dust levels than others just a few kilometers away. In Finland, the pollution was more uniform across the region. The computer model from the Copernicus service proved to be the most reliable tool for tracking changes over time. It matched the patterns seen on the ground quite well, capturing the rise and fall of dust levels with a strong correlation. However, the model tended to underestimate the highest peaks of pollution, likely because its grid cells were too large to capture the intense, localized bursts of dust right next to the mining equipment. The satellite data, which measures the total amount of dust in a column of air from the ground to space, was less effective. It often missed data due to clouds or struggled to distinguish between dust high in the atmosphere and dust near the ground, leading to weaker connections with the actual ground measurements.

The study also looked at how pollution changed over the long term and whether it could be predicted. By breaking down the years of data into trends and seasonal patterns, the researchers observed that air quality has been gradually improving at both sites. In Finland, the decline in dust levels was particularly noticeable, with ground measurements showing a drop of about 5 to 10 micrograms per cubic meter over the study period. In Greece, the trend was more stable, with a slight decrease overall. The seasonal patterns were distinct for each location: Finland saw its highest dust levels in the spring when melting snow exposed dry road dust, while Greece saw peaks in the summer and early autumn, driven by dry conditions and dust blowing in from North Africa. To test if these patterns could be used to forecast future air quality, the team tried different mathematical approaches. They found that for predicting long-term trends, a statistical model designed to handle seasonal cycles worked best. For predicting what would happen in the very near future, such as the next day, a machine learning method that learned from past data patterns performed better.

The findings suggest that while satellites and computer models cannot replace the precision of ground sensors, they are powerful tools for filling in the gaps where sensors are missing. The computer models, in particular, offer a consistent picture of air quality across large regions and over long periods, making them valuable for understanding the overall impact of mining on the environment. However, because these models smooth out the details, they may miss the worst spikes in pollution that occur right next to a mine. The study concludes that the best approach is to combine the broad view from the models with the specific, accurate data from ground stations. This hybrid method allows for a more complete understanding of dust pollution, helping to protect public health and guide environmental policies even in remote or difficult-to-monitor areas. The research demonstrates that with the right combination of tools, it is possible to track and manage air quality in mining regions effectively, regardless of whether the climate is hot and dry or cold and snowy.