these processes occur and understand the science in the reductionist sense at the small scale. Yet when this well-established knowledge from the laboratory and from textbooks is applied on the watershed scale, the current existing models do not provide the necessary reliability for making water quality decisions.
When well-established knowledge from the lab and from textbooks is applied on the watershed scale, the current existing models do not provide the necessary reliability for making water quality decisions.
Reckhow cited two ways in which to make modest reductions in prediction error in the future: advances in scientific knowledge that lead to increasingly elaborate and detailed models, and better observational data and enhancements in statistical techniques for extracting patterns from the data. The problem with current models is that it is difficult to capture the inherent complexity of an aquatic ecosystem and the extreme variability of nature. Yet predictions are nevertheless needed to guide decision making.
Reckhow suggested that scientists need to employ adaptive management by observing how the actual water body responds, and then use this information to augment the predictive power of the model system. He further noted that it is not improvements in models from better science, more detailed mathematics, or better data that will lead to advancement. It is the integration of the monitoring, associated with the post-implementation response of a real system, with the model forecast—an adaptive framework, in which we learn while doing—that should become common practice for these assessments, concluded Reckhow.