We have developed a satellite based water quality prediction model, based on deep networks in the context of a Tübitak project during 2018-2023. Given a multispectral Sentinel-2 image set, the model calculates the quality map of a water body with respect to various parameters. We have selected Lake Balik of the Kizilirmak Delta (a RAMSAR site since 1998, home to 74% of birds species encountered in Turkey), to illustrate its performance. We have achieved high spatial and temporal correlation levels. The article detailing the underlying approach is in "E. Aptoula, S. Ariman, Chlorophyll-a retrieval from Sentinel-2 images using convolutional neural network regression, IEEE Geoscience and Remote Sensing Letters, 2021", "M. Ilteralp, S. Ariman, E. Aptoula, A Deep Multitask Semisupervised Learning Approach for Chlorophyll-a Retrieval from Remote Sensing Images, Remote Sensing, Vol. 14, No. 1, 2022 ", "E. Aptoula, S. Ariman, Hierarchical Spatial-Spectral Features for the Chlorophyll-a Estimation of Lake Balik, Turkey, IEEE Geoscience and Remote Sensing Letters, Vol. 19, 2022 ".

The model has been trained by M. Ilteralp and E. Aptoula, using field samples collected and analyzed by S. Ariman from Samsun University across 3 calendar years. The spatial, short-term as well as long-term temporal generalization performance of the model is encouraging in terms of the development of a national scale water quality monitoring system.