Relationships and Thresholds for Continuous Dissolved Oxygen Variables, Nutrient Effects, and Biological Attributes in Northeast Illinois Rivers and Streams

A suite of continuous D.O. statistics was generated from the available Illinois statewide continuous database from which minimum daily, maximum daily, mean daily, and maximum daily diel swing values were derived. Steps were taken to reduce the influence of extreme daily minimum or maximum values hence the 5^th and 95^th percentile values were used as representing more consistent, representative, and biologically meaningful exposures to low or high D.O. levels. Data was limited to the July 15-September 30 period to ensure that the most biologically and D.O. limiting period was captured, i.e., the seasonal period that coincides with the highest ambient temperatures, low summer-early fall flows, low minimum D.O. levels, and the widest diel swings. These statistics were plotted against the Illinois fIBI and mIBI to visually reveal variables that are meaningfully associated with biological assemblage performance. Statistical analysis included Pearson correlation and Random forest regression tree (RF) analyses at sites that had continuous D.O. and paired habitat and biological data statewide. Weighted Stressor Values (WSVs) were derived for continuous D.O. variables for fish species and macroinvertebrate taxa. Species and taxa were ranked along each D.O. variable and the 20^th (tolerant) and 80^th (sensitive) percentile ranks were used to identify the most sensitive and tolerant species/taxa for each D.O. variable. Minimum D.O. as represented by the 5^th percentile value was the most explanatory of all of the D.O. variables examined. Because of the lack of association between the maximum D.O. or maximum diel D.O. swing and the fIBI or mIBI, these statistics were not by themselves predictive of aquatic life impairment unless they were associated with low minimum D.O. values. As a result Weighted Stressor Values (WSVs) were not derived for either fish species or macroinvertebrate taxa for these variables. No significant correlation between chlorophyll a measures and the fIBI and mIBI was found by this study. The lack of significant correlation may be the result of generally low benthic chlorophyll a values compared to much higher literature values that are considered to be excessive. This is consistent with other Illinois studies that found similarly low benthic chlorophyll a measures than might be expected based on elevated nutrient concentrations. Minimum D.O. thresholds, represented by the 5^th percentile D.O. statistic for fish and macroinvertebrates can be used for stressor identification purposes. This measure was also incorporated into a modification of the Nutrient Ranking Index (NRI) that was previously designed to help identify when nutrients are likely to impair aquatic life. The identification of nutrients as major causes of aquatic life impairment is complex in urbanized watersheds with unrelated co-occurring stressors. Salinization in the form of elevated chloride levels can influence nutrient availability. Modified stream geomorphology and physical habitat quality can also influence the dynamics about how nutrients affect D.O. and aquatic life. In this study, QHEI and several of its metrics showed threshold relationships that influenced minimum D.O. concentrations at degraded sites. Therefore these co-occurring factors need to be considered when evaluation or predicting the influence of nutrients on the D.O. regime.

Authors