Environmental monitoring using remote sensing data and machine learning techniques
Thesis event information
Date and time of the thesis defence
Place of the thesis defence
University of Oulu, Tönning-sali L4 (Linnanmaa Campus)
Topic of the dissertation
Environmental monitoring using remote sensing data and machine learning techniques
Doctoral candidate
Master of Science Amirhossein Ahrari
Faculty and unit
University of Oulu Graduate School, Faculty of Technology, Water, Energy and Environmental Engineering
Subject of study
Environmental Engineering
Opponent
Professor in Civil Engineering (Water Resources and Remote Sensing) Hamidreza Norouzi, New York City College of Technology
Custos
Professor in Water Resources and Environmental Engineering Ali Torabi Haghighi, University of Oulu, Water, Energy and Environmental Engineering Research Group
Tracking of Environmental Changes using Satellite Data and Statistics
Our world is changing quickly under the pressures of climate change, agriculture, and growing demand for water. To respond to these challenges, we need accurate information about how our land and water resources are evolving. Gathering this information on the ground is not always possible, but satellites can give us a wide, detailed view of the Earth. This science, called remote sensing, allows us to monitor temperature, soil moisture, vegetation, and water bodies from space.
Still, satellite data often comes with problems. Clouds can block the view, measurements may be too coarse to capture local changes, and gaps appear in long-term records. My research set out to improve the quality of satellite-based environmental monitoring by applying statistical analysis. With these methods, I made the data clearer, more complete, and more useful for studying the links between climate, water, and agriculture.
One part of my work focused on improving satellite data itself. For example, I sharpened soil moisture maps so that smaller-scale changes could be seen, which is especially useful for farming and drought monitoring. I also filled in missing land temperature data caused by cloud cover, using statistical models that recreate smooth and reliable records over time. To track water more directly, I developed a tool called WaTSat, which automatically measures lakes and other water bodies from satellite images with high accuracy.
The second part of the research looked at what this improved data can reveal about the relationship between climate and water resources. I studied long-term changes in Iran and Finland—two very different environments. In Iran, the results showed how rising temperatures and agricultural demands are stressing already limited water supplies. In Finland, where lakes are abundant, the analysis revealed how climate shifts are altering water patterns and soil conditions.
The results were clear: the enhanced methods produced more accurate data, reducing errors and increasing reliability. The water monitoring tool, for example, was over 90% accurate in tracking lake changes. By combining satellite observations with statistical analysis, the study uncovered patterns and connections that match well with earlier research, giving confidence in the findings.
In the end, this research demonstrates how satellites and statistics together can provide powerful insights for managing natural resources. They allow us to track water and soil conditions more precisely, understand how climate and farming interact, and prepare for the risks of a changing environment. The lessons from both Iran and Finland show that these approaches can support more sustainable management of water and agriculture, which are essential for adapting to climate change.
Still, satellite data often comes with problems. Clouds can block the view, measurements may be too coarse to capture local changes, and gaps appear in long-term records. My research set out to improve the quality of satellite-based environmental monitoring by applying statistical analysis. With these methods, I made the data clearer, more complete, and more useful for studying the links between climate, water, and agriculture.
One part of my work focused on improving satellite data itself. For example, I sharpened soil moisture maps so that smaller-scale changes could be seen, which is especially useful for farming and drought monitoring. I also filled in missing land temperature data caused by cloud cover, using statistical models that recreate smooth and reliable records over time. To track water more directly, I developed a tool called WaTSat, which automatically measures lakes and other water bodies from satellite images with high accuracy.
The second part of the research looked at what this improved data can reveal about the relationship between climate and water resources. I studied long-term changes in Iran and Finland—two very different environments. In Iran, the results showed how rising temperatures and agricultural demands are stressing already limited water supplies. In Finland, where lakes are abundant, the analysis revealed how climate shifts are altering water patterns and soil conditions.
The results were clear: the enhanced methods produced more accurate data, reducing errors and increasing reliability. The water monitoring tool, for example, was over 90% accurate in tracking lake changes. By combining satellite observations with statistical analysis, the study uncovered patterns and connections that match well with earlier research, giving confidence in the findings.
In the end, this research demonstrates how satellites and statistics together can provide powerful insights for managing natural resources. They allow us to track water and soil conditions more precisely, understand how climate and farming interact, and prepare for the risks of a changing environment. The lessons from both Iran and Finland show that these approaches can support more sustainable management of water and agriculture, which are essential for adapting to climate change.
Last updated: 28.8.2025