Bathymetry strongly affects nearshore hydrodynamics that drive sediment transport. However, recent bed level information is not always available, especially not simultaneously resolved in space and time. Time series of radar backscatter images can be used to retrieve bed levels in the intertidal area with the so-called Temporal Waterline Approach (TWA) by Bell et al. [2016], yielding a two weeks average bathymetry at spatial resolutions in the order of a few meters. Bed levels are estimated for each radar cell individually by correlating temporal patterns of wet-dry transitions from the radar with hypothetical patterns found in the water level record for different tidal levels. In this study data recorded by a marine X-band radar located at Bunker Hill station (Sylt, German Wadden Sea island) is used to improve TWA, validate the improvement as well as to show its potential in application. A survey was conducted in May 2018 to provide ground truth. Two modifications are made, that improve robustness and reliability of TWA’s output by adding a second threshold and semi-automating the quality procedure. Root mean square error between radar-derived bathymetry (RDB) and ground truth for the intertidal area is 43 cm (bias 12 cm) on a 7.5 m squared grid. Sensitivity analysis of four parameters (i.e. vertical resolution, inclusion of wave induced effects on water level, length and position of analysis window, thresholds in quality check) reveals, that TWA is most sensitive to the choice of thresholds. Sensitivity analysis is utilized to tune TWA to achieve full coverage of the intertidal area (zero coverage elsewhere) with sufficient accuracy. Additionally it is found that at costs of accuracy the period for which averaged bed levels are retrieved can be shortened from 14.5 to 8.3 days. This is promising, because RDB potentially is able to observe short-term changes in the order of a few days (e.g. storm events). To demonstrate the value of the improved TWA for bathymetry surveys in intertidal areas the method is applied to a 3.2 km stretch of beach in front of Bunker Hill station, capturing the beach state variations in the course of 2017. During wind sea dominated season (winter) a ridge-runnel type of beach is present, that changes to a skewed transverse bar and rip system during swell dominated season (summer). The method is sufficiently accurate and robust for application and proves its potential to reveal new insights, e.g. indications on how the transition between two beach states evolves. In coastal science RDB could serve as a tool to improve existing theories by providing detailed information, that reveals underlying processes. In coastal engineering it fills gaps in monitoring schemes, for example to evaluate effects of nourishment strategies.