This paper by M. Bruce Beck reviews the role of uncertainty in water quality modeling, focusing on four key areas: uncertainty about model structure, uncertainty in estimated model parameter values, propagation of prediction errors, and experimental design to reduce critical uncertainties. The review is divided into two parts: a shorter, nontechnical version and a detailed technical discussion. The technical part covers identifiability and experimental design, generating preliminary model hypotheses, unconditional uncertainty in unknown field data, identifying elements of model structure, parameter estimation, model verification and discrimination, and prediction error propagation. Beck emphasizes the importance of system identification and the analysis of field data for developing scientific theories about complex environmental systems. The paper also discusses the challenges of model identifiability and the need for new algorithms and methods to address these issues. The review concludes that while significant progress has been made, much remains to be done, particularly in the development of novel algorithms for model structure identification and the integration of artificial intelligence techniques.This paper by M. Bruce Beck reviews the role of uncertainty in water quality modeling, focusing on four key areas: uncertainty about model structure, uncertainty in estimated model parameter values, propagation of prediction errors, and experimental design to reduce critical uncertainties. The review is divided into two parts: a shorter, nontechnical version and a detailed technical discussion. The technical part covers identifiability and experimental design, generating preliminary model hypotheses, unconditional uncertainty in unknown field data, identifying elements of model structure, parameter estimation, model verification and discrimination, and prediction error propagation. Beck emphasizes the importance of system identification and the analysis of field data for developing scientific theories about complex environmental systems. The paper also discusses the challenges of model identifiability and the need for new algorithms and methods to address these issues. The review concludes that while significant progress has been made, much remains to be done, particularly in the development of novel algorithms for model structure identification and the integration of artificial intelligence techniques.