<p>Long-term time series of coastal bathymetric data are indispensable for analysing morphological resilience to climate change. Despite the increasing popularity of utilizing high-resolution digital elevation models for coastal management, potential errors in analysing the long-term trend of bed elevation change from historical datasets have attracted little attention. Here, we demonstrate that inconsistency in the spatial resolution of small-scale topographic features characterized by sharp bathymetric gradients could produce errors in the mean elevation that in individual years could exceed the sea level change rate and result in a statistically significant bias in the trend of the mean elevation. Neglecting this inconsistency may lead to a misconception of coastal resilience to sea level rise. We provide an analytical method to identify such inconsistency in bathymetric datasets and a homogenization method to minimise the errors. Our methods are broadly applicable for analysing time series of geospatial data and quantitative assessment of data consistency.</p>