Title: Progress tracking is not a realistic use for typical stream monitoring approaches
Dan Haug, April 2021
Abstract
Statistical power analysis can help stakeholders set realistic expectations for stream monitoring, and ensure that enough data is collected to produce conclusive results. Using nitrate and total phosphorus data from Ioway Creek in north-central Iowa, we determined that at least 12 years of biweekly nitrate sampling would be needed to reliably determine whether the project was on-track to meet the 20-year goals in the watershed management plan. For phosphorus, 18 years of biweekly data would be needed. Without some way of controlling for variability (i.e. a paired watershed study) or a plan to sustain monitoring for at least a decade, monthly and biweekly stream monitoring efforts will not be effective for progress tracking, although the data may have other applications. To help illustrate the concept of sampling error for stakeholders who do not have a background in statistics, we pulled four monthly subsets out a weekly, long-term dataset collected by the City of Ames in the South Skunk River. This approach informed a 10-year monitoring plan created by Story County, Prairie Rivers of Iowa, and eight other partners.