Abstract: This study explores the possibility of improving the accuracy of numerical modelling of seawater temperature through data assimilation of remote sensing observations and in-situ measurements. The study focuses on modelling the excess temperature caused by the discharge of cooling water from power plants into coastal seas. A three-dimensional hydrodynamic model is constructed to simulate the mixing and transport of the thermal effluents into a semi-enclosed, shallow bay. Due to the difficulties in accounting for the different spatial and temporal scales associated with the thermal plume in one integral model run, practices of coupling near-field and far-field models are commonly used. This study proposes a framework to improve the coupled model by combining two levels of thermal infrared remote sensing data: satellite and low-altitude drone imageries, along with in-situ observations to derive an improved representation of the trajectory of the thermal plumes in dynamically active, coastal waters.