Selection of parameter values in environmental models using sparse data: A case study