Machine learning and hurdle models for improving regional predictions of stream water acid neutralizing capacity

Abstract

[1] In many industrialized regions of the world, atmospherically deposited sulfur derived from industrial, nonpoint air pollution sources reduces stream water quality and results in acidic conditions that threaten aquatic resources. Accurate maps of predicted stream water acidity are an essential aid to managers who must identify acid-sensitive streams, potentially affected biota, and create resource protection strategies. In this study, we developed correlative models to predict the acid neutralizing capacity (ANC) of streams across the southern Appalachian Mountain region, USA. Models were developed using stream water chemistry data from 933 sampled locations and continuous maps of pertinent environmental and climatic predictors. Environmental predictors were averaged across the upslope contributing area for each sampled stream location and submitted to both statistical and machine-learning regression models. Predictor variables represented key aspects of the contributing geology, soils, climate, topography, and acidic deposition. To reduce model error rates, we employed hurdle modeling to screen out well-buffered sites and predict continuous ANC for the remainder of the stream network. Models predicted acid-sensitive streams in forested watersheds with small contributing areas, siliceous lithologies, cool and moist environments, low clay content soils, and moderate or higher dry sulfur deposition. Our results confirmed findings from other studies and further identified several influential climatic variables and variable interactions. Model predictions indicated that one quarter of the total stream network was sensitive to additional sulfur inputs (i.e., ANC < 100 meq L À1 ), while <10% displayed much lower ANC (<50 meq L À1 ). These methods may be readily adapted in other regions to assess stream water quality and potential biotic sensitivity to acidic inputs.

Multivariate Condition Assessment of Watersheds with Linked Micromaps

“…We used the spsurvey R package to estimate the medians for the target population of streams, with 95% confidence bounds, for the standardized GLIMPSS, habitat, and seven of the nine water chemistry variables of the 25 WVDEP subbasins . For pH, we estimated the 25th percentile for the subbasins because acid mine drainage and acid deposition have impacted some streams in West Virginia (Petty et al, 2010;Merovich et al, 2013;Povak et al, 2013). For fecal coliform, we estimated the proportion of sites that exceeded the criterion of 400 colonies/100 mL (WVDEP, 2008).…”

Contribution of private gardens to habitat availability, connectivity and conservation of the common pipistrelle in Paris

“…The gbm approach was relevant in this study because it can handle a large number of predictors – even collinear ones – and deal with spatial autocorrelation effectively, and it also has a strong predictive performance (Elith, Leathwick, & Hastie, 2008). Because the data contained a lot of zeros (130 points out of a total of 224 points), we employed a Hurdle modeling approach to account for zeroinflation (Potts & Elith, 2006; Povak et al, 2013). This approach is consistent with previous studies modeling bat passes (Aurelie Lacoeuilhe, Machon, Julien, Le Bocq, & Kerbiriou, 2014; Vandevelde et al, 2014).…”

Downstream Warming and Headwater Acidity May Diminish Coldwater Habitat in Southern Appalachian Mountain Streams

“… Model future JMMST associated with July atmospheric warming of 2 and 4°C by ( i ) developing a predictive threshold model to identify stream reaches with high sensitivity to warming air temperatures, which were defined by stream and air temperature correlations ( Fig 2b ); and ( ii ) developing a predictive model for high-sensitivity stream sites to derive continuous estimates of the strength of stream water and air temperature correlations ( Fig 2c ). Refine previously published regional stream ANC estimates [ 23 ] by correcting for model bias and re-projecting a continuous ANC surface across the study region. Map refined ANC and JMMST estimates to quantify potential loss of stream habitat for acid-sensitive coldwater species.…”

“…Refine previously published regional stream ANC estimates [ 23 ] by correcting for model bias and re-projecting a continuous ANC surface across the study region.…”

“…The combined effect of thermal habitat loss and stream acidification presents a conundrum for watershed managers tasked with identifying suitable habitats for coldwater species (e.g., brook trout [ Salvelinus fontinalis ]) [ 10 , 21 ]. In general, low-order, higher-elevation streams are most susceptible to acidification; in the absence of adequate buffering capacity, otherwise suitable habitat is rendered uninhabitable in the headwaters of susceptible streams [ 22 , 23 ]. In contrast, thermal habitat for coldwater species becomes less suitable (i.e., warmer) as elevation decreases.…”