A data cube is a multi-dimensional structure for organizing and analyzing data across various dimensions (e.g., time, location). A semantic data cube enhances this by adding metadata and relationships, allowing the request of data. In the Sen2Cube.at semantic EO data cube it is possible to create and execute semantic models.

One of the objectives of the data cube is to study identifiable phenomena in the real world. In my case, I consider lakes because they are water objects with specific properties of color, texture, and compactness. I associated 4 categories with water: Deep water or shadow, Shallow water or shadow, Turbid water or shadow, and Salty shadow water.

Additionally, lakes remain in areas considered flat. For that reason, I added the slope of less than 2 characteristics. This aspect is based on the topography using a digital elevation model and is valid for the entire time span (unlike satellite images, which are stacked upon each other).

As a result, the model shows two layers: one with the count of pixels classified as lakes and another with the percentage of lakes over space. It also includes a time series of the percentage of pixels classified as lakes. The model was applied in four different regions in Austria to validate satisfactory results.