Striving for consistency in a national assessment: the challenges of applying a reference-condition approach at a continental scale

Large-scale macroinvertebrate assemblage patterns from least-disturbed wadeable stream sites across the 48 contiguous US states

“…In the US, it is much easier to find least-disturbed sites in mountainous ecoregions than lower gradient ecoregions due to their lower levels of overall human disturbance. Herlihy et al (2008) examined the quality of reference sites in the WSA data we used in this analysis and did find variable reference site quality among the nine aggregate ecoregions in the survey. They used a regression approach to adjust reference expectations in ecoregions that showed a relationship between macroinvertebrate condition and human disturbance within reference sites.…”

“…Whittier et al (2007) describe in detail the process for defining leastdisturbed sites for the EMAP-West survey. For the eastern stream sites, a similar process was used to identify least-disturbed sites based on acid neutralizing capacity, sulfate, chloride, total nitrogen, total phosphorus, turbidity, % fine sediment, EMAP riparian disturbance index, and EPA rapid bioassessment protocol habitat score as detailed in Herlihy et al (2008). For our analyses, if sites were sampled several times, only data from the first sample visit were included and sites with macroinvertebrate samples with fewer than 150 invertebrates identified and counted were also deleted.…”

Assessing Stream Ecosystem Condition in the United States

“…From the WSA probability sample of 1392 sites, 334 sites that met a set of ecoregion-specific chemical and physical habitat criteria were selected as reference sites to represent least disturbed conditions within each aggregated ecoregion. Additional reference sites were obtained from a variety of sources with compatible data and methods, yielding a total of 1625 reference sites (see Herlihy et al [2008] for details).…”

“…The second was an observed/expected (O/E) taxa loss index based on predictive models. Reference site distributions for these indicators were used to set thresholds for three condition classes (good, fair, and poor) within each analysis region [Herlihy et al, 2008]. A similar process was used to set condition class thresholds for stressors.…”

Accounting for regional variation in both natural environment and human disturbance to improve performance of multimetric indices of lotic benthic diatoms

“…Site specific models and site classification by either ecoregions or biological typologies that account for natural variability among sites have been used separately in previous studies (Hawkins, 2006;Herlihy et al, 2008;Vander Laan and Hawkins, 2014), but they have seldom been used together or at large spatial scales with an evaluation of their independent and interactive effects on MMI performance. Site specific modeling was used for removing effects of natural gradients in algal MMIs for streams within a diverse group of small ecoregions (Cao et al, 2007) and for lakes across the US without ecoregion grouping (Stevenson et al, 2013).…”