<p>Support for using diatom DNA metabarcoding in stream monitoring and assessment programs continues to grow, but its use often requires storing samples for months prior to extraction and analysis. Freezing without preservatives is a cheap, simple way to store and handle samples, but little is known about how this could affect results. Here, we conducted a wastewater dosing experiment in artificial streams under low and high light conditions and a watershed survey of wadeable streams to determine stressor effects on diatom assemblages and to test whether DNA extracted from samples that were fresh, frozen, or frozen and stored for 1 or 3 months led to differences in assemblage structure and diatom-environment relationships. In each study, assemblages from frozen samples were similar regardless of storage duration, but each was significantly different from assemblages in fresh samples. Cell lysis caused by freezing likely increased the DNA yield from smaller species and those with stronger frustules less susceptible to breakage by microbead beating. Regardless of being fresh or stored frozen, assemblages in wastewater dosed mesocosms were significantly different from those in controls, assemblages in low light mesocosms were significantly different from those in high light mesocosms, and relationships between assemblage structure and watershed percent agriculture were similarly strong. These results highlight the responsiveness of diatoms to environmental changes associated with human activities, support their usefulness as indicators in monitoring and assessment programs, and show how, when necessary, the long-term storage of frozen samples does not significantly affect interpretations of DNA metabarcoding results.</p>