Abstract
We developed independent predictive disturbance models for a full regional data set and four individual ecoregions (Full Region vs. Individual Ecoregion models) to evaluate effects of spatial scale on the assessment of human landscape modification, on predicted response of stream biota, and the effect of other possible confounding factors, such as watershed size and elevation, on model performance. We selected macroinvertebrate sampling sites for model development (n = 591) and validation (n = 467) that met strict screening criteria from four proximal ecoregions in the northeastern U.S.: North Central Appalachians, Ridge and Valley, Northeastern Highlands, and Northern Piedmont. Models were developed using boosted regression tree (BRT) techniques for four macroinvertebrate metrics; results were compared among ecoregions and metrics. Comparing within a region but across the four macroinvertebrate metrics, the average richness of tolerant taxa (RichTOL) had the highest R2 for BRT models. Across the four metrics, final BRT models had between four and seven explanatory variables and always included a variable related to urbanization (e.g., population density, percent urban, or percent manmade channels), and either a measure of hydrologic runoff (e.g., minimum April, average December, or maximum monthly runoff) and(or) a natural landscape factor (e.g., riparian slope, precipitation, and elevation), or a measure of riparian disturbance. Contrary to our expectations, Full Region models explained nearly as much variance in the macroinvertebrate data as Individual Ecoregion models, and taking into account watershed size or elevation did not appear to improve model performance. As a result, it may be advantageous for bioassessment programs to develop large regional models as a preliminary assessment of overall disturbance conditions as long as the range in natural landscape variability is not excessive.
Highlights
Understanding the effects of human land use modification on stream biota, the processes that cause these effects, and the various spatial and temporal scales at which these effects and processes operate are fundamental goals of bioassessment in stream ecology
We developed our boosted regression tree (BRT) models using a multi-stage process: (1) BRT models were first run with all watershed, riparian, and natural landscape variables only, with the top 10 variables in the variable relative importance list retained for further analysis, (2) BRT models were run with all hydrologic runoff based variables only, with the top 10 variables in the variable importance list retained for further analysis, and (3) BRT models were run combining the top 10 variables from steps 1 and 2
Given the use of the EPT Richness (EPTR) metric as a component of many multimetric indices in the northeast, it was unexpected that it would have the lowest R2 across all regions for BRT models when compared to the other metrics, except in the Northeastern Highlands (NE_High), where it had the second lowest value
Summary
Understanding the effects of human land use modification on stream biota, the processes that cause these effects, and the various spatial and temporal scales at which these effects and processes operate are fundamental goals of bioassessment in stream ecology. Ode et al [14] found that macroinvertebrate indices developed at large regional scales, such as the western U.S, had lower precision in California than California-based indices They found that the larger scale indices were influenced by two natural gradients that did not affect the statewide indices. Seelbach et al [15], on the other hand, found that regions too large can sometimes give misleading results when strong natural gradients at larger scales are mismatched with ecological-scale responses Their streamflow models across three states in the midwestern U.S showed that more rainfall in the southern portion of their region created lower stream baseflows (a nonsensical relationship) and that higher northern baseflows were the result of very permeable glacial deposits that are variable across smaller scales that were not accounted for by the large scale models [15]. Current ecosystem theory indicates that models at smaller scales should allow for more insight and interpretation of disturbance related processes or mechanisms that are likely to operate at smaller watershed and site specific scales [5],[6],[16],[17]
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
