Trace elements play a major role in biogeochemical cycles and oceanographic processes. To determine trace element concentrations, the dissolved and particulate phase are usually separated by filtration. However, the frequently used membrane filtration as well as sample storage can bias the dissolved elemental concentrations by adsorption or desorption/contamination. We present a comparison of two filtration methods for coastal and estuarine water samples (pressure filtration with Nuclepore™ polycarbonate filters, vacuum filtration with DigiFILTER™s) applied to aliquots of a large‐volume coastal water sample that were stored at ‐18 °C or 4 °C for up to nine weeks. The filtrates were analyzed by seaFAST‐ICP‐MS for dissolved Cd, Ce, Co, Cu, Dy, Er, Eu, Fe, Ho, La, Mn, Mo, Nd, Pb, Pr, Sm, Tb, U, V, W, Y and Zn. The filtration blanks of DigiFILTER™s (0.0006 ± 0.0010 ng L‐1 for Ho to 110 ± 180 ng L‐1 for Zn) were sufficiently low for quantification of all analyzed elements with good repeatability, enabling a fast and reliable filtration of large sample sets of coastal water. However, the findings also highlight the need to measure procedural blanks including the filtration instead of only the instrument blanks to validate results. Measured concentrations of both filtration methods did not differ significantly for Cd, Cu, Mo, U, V, W, Zn but for other investigated elements, the ratio between both methods was up to 1.8 for Ce and 4.1 for Fe. Within nine weeks of storage, the elemental concentrations decreased significantly, resulting in losses of 20 % Mn in frozen samples and 63 % Pb, 64 % Co and 93 % Mn in cooled samples.