In its 2012 National Lakes Assessment, the EPA determined that nutrients are the most widespread stressor of US Lakes, with one-third of US lakes containing excess nitrogen and/or phosphorus. In response, states are developing numeric nutrient criteria to regulate the allowable amount of nutrients in surface waterbodies. In this study, we estimate the recreational value of improved water quality as a result of reduced nutrients in Mississippi lakes using a travel cost model. A major challenge, however, is the lack of water quality data for many Mississippi lakes. We introduce a water quality prediction model to impute water quality measures for lakes that lack water quality data. We compare welfare estimates from travel cost models based on the limited available water quality data to models using predicted water quality measures to evaluate the extent to which inadequate data affects welfare estimates for improved water quality.