The article discusses how an artificial neural network model was used to predict water levels in a small urban river affected by tides, atmospheric pressure, wind, and river discharge. The study focused on the Kanda River in Tokyo and found that atmospheric pressure can cause the river water level to rise by about 7cm for every 5hPa increase, regardless of river discharge, with 1m/s wind speed and north wind direction. The researchers were able to create an accurate rating curve for the tidal river, which is crucial for flood control in urban areas near river mouths.