Dissolved organic carbon (DOC) transported by rivers represents an important link between carbon pools of terrestrial and oceanic ecosystems. However, it is unclear how frequent DOC must be sampled to obtain reasonable load estimates. Here, we used continuous records of the specific UV absorption coefficient (SAC) and discharge from a headwater stream at the Ore Mountains (Germany) to calculate load errors depending on DOC sampling frequency. SAC was used as a proxy for DOC. The results show that the load was underestimated by 13-19% with monthly, 10-13% with bi-weekly and 7-9% with weekly DOC samplings, respectively. We conclude that collecting additional data from high discharge events decrease the error significantly.