Radionuclides in the environment require regular monitoring to establish activity concentrations and make reliable dose estimates. This study examines the variability in selected radionuclide activity concentrations at a variety of spatial scales and assesses the resulting errors in dose estimates. The Ribble Estuary, UK provides an ideal location to conduct non-destructive tests (β-dose rate measurements) due to the discharge of 234Th and $^{234\textrm{\tiny{m}}}$Pa from the BNFL fuel fabrication plant at Springfields (Lancashire, UK). This estuary also receives Sellafield derived radionuclides after transport through the Irish Sea. Results indicate that significant errors in external beta and gamma dose assessments may result from the use of single sediment activity concentration or dose values to represent an area of contaminated sediment. For example, in an area only 1m × 1m square, a field coefficient of variation for beta dose may be as high as 25%, even when there are no visible textural or compositional variations to account for this, and this increased to 83% at the 100×100m scale. Erroneous conclusions on the nature of change in time-series measurements can be made if local field variability has not been assessed and allowed for when planning the monitoring programme.