Periodic calculation of coastal bathymetries can show the evolution of geomorphological features in active areas such as mesotidal estuary mouths. Bathymetries in shallow coastal areas have been addressed mainly by two technologies, lidar and optical remote sensing. Lidar provides good accuracy, but is an expensive technique, requiring planned flights for each region and dates of interest. Optical remote sensing acquires images periodically but its results are limited by water turbidity. Here we use a lidar bathymetry to compare different bathymetry computation methods using a SPOT optical image from a nearby date. Three statistical models (green-band, PCA correlations, and GLM) were applied to obtain mathematical expressions to estimate bathymetry from that image: all gave errors lower than 1 m in an area with depths ranging from 0 to 6 m. These algorithms were then applied to images from three different dates, correcting the effects caused by different tidal and atmospheric conditions. We show how this allows the study of morphological changes. We discuss the accuracy obtained with respect to the reference bathymetry (0.9 m on average, but less than 0.5 m in low-turbidity areas), the effects of the turbidity on our estimations, and compare both with previously published results. The results show that this approach is effective and allows identification of known features of coastal dynamics, and thus it would be an important step towards short-term bathymetry monitoring based on optical satellite remote sensing.