Environmental Modeling: The Past, the Present, and the Future

摘要

The information age has ushered in a global awareness of complex environmental problems that do not respect political or physical boundaries: e.g., climatic change, ozone layer depletion, deforestation, desertification, and nonpoint source (NPS) pollution. The ability to simulate spatially and temporally variable environmental impacts, via mathematical modeling, is an important modern day necessity. Simulation provides the ability to ask and answer "what if questions that can be used to guide decision-making strategies. Uncertainty in simulated assessments is unavoidable and sometimes even undetectable in environmental modeling. Quantification of accuracy and uncertainty in environmental modeling establishes the extent to which simulated results are reliable predictions of observed truth. Uncertainty can be due to model errors produced from simplification of process complexities described by the model. Uncertainty can also result from data error, i.e. the result of errors in the input data or the lack of information. Intrinsic model and data uncertainties have significant practical implications, either affirming or negating the use of predictive outputs from environmental models for guidance and action in a decision-making process. The development, the current utility, and the likely new roles of environmental modeling are discussed. The underlying focus of the paper is the characterization of accuracy and the impact of uncertainty in environmental modeling.