Reliable estimates of mean concentrations and loads of pollutants in rivers have become increasingly important for management purposes, particularly with the implementation of the European Union Water Framework Directive (WFD). Here, the Numedalslågen River (5500km2) in southern Norway was used as a case study to evaluate the effects of sampling frequency on mean concentrations and estimated loads of suspended particulate matter (SPM). Daily monitoring data from five seasons (April/June–October/November 2001–2005) were analyzed, and three different load estimation techniques were tested: rating curves, linear interpolation, and the ratio method. The reliability of mean SPM concentrations improved with increasing sampling frequency, but even weekly sampling gave error rates as high as 70% in seasons with elevated sediment loads. Load estimates varied considerably depending on both the sampling frequency and the calculation method used. None of the methods provided consistently good results, but the lowest error rate was achieved when using the rating curve on data from fortnightly sampling and additional sampling during floods. Sampling at monthly intervals gave the highest error rates and cannot be recommended for any of the calculation methods applied here. SPM concentrations were correlated (r2>0.5) with arsenic, lead, nickel, orthophosphate, and total phosphorus in the Numedalslågen River. Therefore, the current findings may also have implications for substances other than SPM. The discussion considers examples from actual use of infrequently collected data, and it is advised that managers account for uncertainties in both concentration means and load estimates when assessing the state of a water body or planning mitigation measures.