There are numerous models available, with more being developed each year, differing in scale of the modeled landscape and complexity of use and inputs. In relating models to observed conditions, models are calibrated, and model output is compared to field data, historical reports and expected behavior (US EPA, 2006). These comparisons allow the validity

of model output to be assessed and provide “weight-of-evidence” support for the use of the model (US EPA, 2006). A recent study compared four commonly used watershed models, including STEPL, with 30 years buy Ceritinib of monitoring data from a Kansas dam impoundment (Neiadhashemi et al., 2011). When comparing modeled loading with measured results, the study indicated: Gemcitabine The models varied in their ability to replicate measured data; models best conformed to the measured pollutant loading when input data was based on observed local conditions instead of regional defaults;

STEPL performs well in estimating relative contribution from land use but less well in geographically determining major sources of sediment. STEPL is included in the US EPA website as an acceptable watershed-scale model. In Ohio, it was used in conjunction with stream monitoring data to develop the Euclid Creek TMDL watershed plan (Ohio EPA, 2005). The Middle Cuyahoga River study provides an additional example of measured data that supports the strength of the STEPL model, with comparison to a decades-long sediment record Immune system instead of the relatively limited time frame of stream monitoring. Where two distinctly different methodologies compare closely, as with the Middle Cuyahoga study, an understanding of the similarities and differences

in results and assumptions can assist investigators in several ways. First, the similar results help support the validity of both approaches/interpretations. Second, investigators can compare the more easily derived model results for watersheds and subwatersheds having more limited monitoring data with a degree of confidence. For example, pollutant loading model results for other subwatersheds of the Cuyahoga River can be compared with downstream monitoring data to determine the relative contribution from subwatersheds. This could allow watershed managers to target high-sediment yield subwatersheds/land uses for best management practices. Third, the sediment study points to limitations in the modeling process that watershed managers can address by varying assumptions. For instance, the sediment record demonstrates a potential increase in high-flow events, which may increase stream erosion. Watershed managers can easily model several scenarios of pollutant loading with different average precipitation amounts and even an increased amount of gully formation.