Boosted trees for ecological modeling and prediction

Accurate prediction and explanation are fundamental objectives of statistical analysis, yet they seldom coincide. Boosted trees are a statistical learning method that attains both of these objectives for regression and classification analyses. They can deal with many types of response variables (numeric, categorical, and censored), loss functions (Gaussian, binomial, Poisson, and robust), and predictors (numeric, categorical). Interactions between predictors can also be quantified and visualized. The theory underpinning boosted trees is presented, together with interpretive techniques. A new form of boosted trees, namely, "aggregated boosted trees" (ABT), is proposed and, in a simulation study, is shown to reduce prediction error relative to boosted trees. A regression data set is analyzed using ABT to illustrate the technique and to compare it with other methods, including boosted trees, bagged trees, random forests, and generalized additive models. A software package for ABT analysis using the R software environment is included in the Appendices together with worked examples.     

Forecasting the Probability of Future Groundwater Levels Declining Below Specified Low Thresholds in the Conterminous U.S.

We present a logistic regression approach for forecasting the probability of future groundwater levels declining or maintaining below specific groundwater‐level thresholds. We tested our approach on 102 groundwater wells in different climatic regions and aquifers of the United States that are part of the U.S. Geological Survey Groundwater Climate Response Network. We evaluated the importance of current groundwater levels, precipitation, streamflow, seasonal variability, Palmer Drought Severity Index, and atmosphere/ocean indices for developing the logistic regression equations. Several diagnostics of model fit were used to evaluate the regression equations, including testing of autocorrelation of residuals, goodness‐of‐fit metrics, and bootstrap validation testing. The probabilistic predictions were most successful at wells with high persistence (low month‐to‐month variability) in their groundwater records and at wells where the groundwater level remained below the defined low threshold for sustained periods (generally three months or longer). The model fit was weakest at wells with strong seasonal variability in levels and with shorter duration low‐threshold events. We identified challenges in deriving probabilistic‐forecasting models and possible approaches for addressing those challenges.     

Significant Escherichia coli attenuation by vegetative buffers on annual grasslands

A study was conducted to estimate the retention efficiency of vegetative buffers for Escherichia coli deposited on grasslands in cattle fecal deposits and subject to natural rainfall-runoff conditions. The study was conducted on annual grasslands in California's northern Sierra Nevada foothills, a region with a distinct wet-dry season Mediterranean climate. We used 48, 2.0- by 3.0-m runoff plots to examine the efficacy of 0.1-, 1.1-, and 2.1-m buffers at three land slopes (5, 20, and 35%) and four dry vegetation matter levels (225, 560, 900, and 4500 kg/ha) across 27 rainfall-runoff events during two rainfall seasons. Buffer width treatments were implemented by placement of cattle fecal material containing known loads of E. coli 0.1, 1.1, or 2.1 m upslope of the plot runoff collector. Mean total runoff to total rainfall ratio per plot ranged from 0.014:1 to 0.019:1 and reflected the high infiltration capacity of these soils. Approximately 94.8 to 99.995% of total E. coli load applied to each plot appears to be either retained in the fecal pat and/or attenuated within 0.1 m downslope of the fecal pat, irrespective of the presence of a wider vegetated buffer. Relative to a 0.1-m buffer, we found 0.3 to 3.1 log10 reduction in E. coli discharge per additional meter of vegetative buffer across the range of residual dry vegetation matter levels, land slope, and rainfall and runoff conditions experienced during this project. Buffer efficiency was significantly reduced as runoff increased. These results support the assertion that grassland buffers are an effective method for reducing animal agricultural inputs of waterborne E. coli into surface waters.     

In vitro bioassay for reactive toxicity towards proteins implemented for water quality monitoring

Reactive organic chemicals comprise a large number of compounds with a variety of reactive moieties. While most assays for reactive toxicity focus on DNA damage, reactivity towards proteins can also lead to irreparable damage, but reactivity towards proteins is typically not included in any test battery for water quality assessment. Glutathione (GSH) is a small tripeptide whose cysteine moiety can serve as a model for nucleophilic sites on proteins. GSH is also an important indicator of detoxification processes and the redox status of cells and due to its protective role, depletion of GSH ultimately leads to adverse effects. A bioassay based on genetically modified Escherichia coli strains was used to quantify the specific reactivity towards the protein-like biological nucelophile GSH. The significance of GSH for detoxification was assessed by comparing the growth inhibition induced by reference chemicals or water samples in a GSH-deficient strain to its fully functional parent strain. The GSH deficient strain showed the same sensitivity as the GSH proficient strain to non-reactive and DNA damaging chemicals, but was more sensitive to chemicals that attack cysteine in proteins. The difference in effect concentrations for 50% inhibition of growth assessed as biomass increase (EC(50)) between the two strains indicates the relevance of GSH conjugation as a detoxification step as well as direct reactivity with cysteine-containing proteins. Seven reference compounds serving as positive and negative controls were investigated. The E. coli strain that lacks GSH was four times more sensitive towards the positive control Sea-Nine, while negative controls benzo[a]pyrene, 2-aminoanthracene, phenol, t-butylhydroquinone, methyl methane sulfonate and 4-nitroquinoline oxide showed equal effect concentrations in both strains. Water samples collected across an indirect potable reuse scheme representing the complete water cycle from sewage to drinking water in South East Queensland, Australia were used to evaluate the applicability of the E. coli assay for reactive toxicity in water samples. While the EC(50) values of the GSH+ strain showed similar trends as in other biological endpoints over the various treatment chains, the specific response indicative of protein damage was only observed in samples that had undergone chlorination as a disinfection process. High natural organic matter or other matrix components disturbed the bioassay so much that we recommend it for future routine testing only in tertiary treated water or drinking water.     

A Critical Review of Assumptions About the Prairie Dog as a Keystone Species

Cynomys spp.) have been labeled as keystone species because of their influence on biological diversity and ecosystem function. However, the validity of several assumptions used to support keystone status is questionable. We review the strength of the evidence and the magnitude of the prairie dog's effects on ecosystem structure and function. We use this review to reevaluate the keystone role for prairie dogs. Our goal is to encourage sound management of the prairie dog ecosystem by improving the ecological foundation of their keystone status. Our review confirms that prairie dogs affect a number of ecosystem-level functions but that their influence on prairie vertebrates may be less than previously suggested. Species richness and abundance patterns were variable among plants, mammals, and birds and were not consistently higher on prairie dog colonies compared to uncolonized areas. In addition, only nine of the 208 species listed in the literature as observed on or near prairie dogs colonies had quantitative evidence of dependence on prairie dogs. Abundance data indicated opportunistic use of colonies for an additional 20 species. A total of 117 species may have some relationship with prairie dogs, but we lacked sufficient data to evaluate the strength of this relationship. The remaining 62 species may be accidental or alien to the system. Despite our conclusion that some prairie dog functions may be smaller than previously assumed, collectively these functions are quite large compared to other herbivores in the system. We suggest that prairie dogs also provide some unique functions not duplicated by any other species and that continued decline of prairie dogs may lead to a substantial erosion of biological diversity and landscape heterogeneity across prairie and shrub-steppe landscapes. Thus, we concur that keystone status for prairie dogs is appropriate and may aid conservation efforts that help protect species dependent on prairie dogs and support other important ecosystem functions.     

Ecolabeled paper towels: consumer valuation and expenditure analysis

Ecolabeled paper towels are manufactured using post-consumer recycled material and sold in markets using a recycle logo. Environmentally conscious consumers purchase these paper towels and thereby contribute to improving environmental quality. In this paper, we estimate the implicit value placed by consumers on ecolabeled paper towels using a hedonic price function and conduct an expenditure analysis using Heckman's selection model. Using the data set from the Internet-based grocery stores called as Peapod we find that some consumers recognize ecolabels on paper towels and place a substantial, positive price premium on them. The expenditure analysis indicates that for the preferred functional form, the demand for ecolabeled paper towels is inelastic for environmentally conscious consumers. The simulated results from the selection model indicate that a small subsidy for ecolabeled paper towels will not substantially change consumers' purchase decisions.     

Impacts of climate change on land-use and wetland productivity in the Prairie Pothole Region of North America

Wetland productivity in the Prairie Pothole Region (PPR) of North America is closely linked to climate. A warmer and drier climate, as predicted, will negatively affect the productivity of PPR wetlands and the services they provide. The effect of climate change on wetland productivity, however, will not only depend on natural processes (e.g., evapotranspiration), but also on human responses. Agricultural land use, the predominant use in the PPR, is unlikely to remain static as climate change affects crop yields and prices. Land use in uplands surrounding wetlands will further affect wetland water budgets and hence wetland productivity. The net impact of climate change on wetland productivity will therefore depend on both the direct effects of climate change on wetlands and the indirect effects on upland land use. We examine the effect of climate change and land-use response on semipermanent wetland productivity by combining an economic model of agricultural land-use change with an ecological model of wetland dynamics. Our results suggest that the climate change scenarios evaluated are likely to have profound effects on land use in the North and South Dakota PPR, with wheat displacing other crops and pasture. The combined pressure of land-use and climate change significantly reduces wetland productivity. In a climate scenario with a +4 °C increase in temperature, our model predicts that almost the entire region may lack the wetland productivity necessary to support wetland-dependent species.     

Effect of manganese on the concentration of amino acids in different regions of the rat brain

Abstract

The present study was designed to determine if chronic exposure of weanlings and adult rats to Mn produces significant alterations in amino acid concentrations in different regions of the rat brain. Weanling (30 day old) and adult (90 day old) male rats were exposed to 10 and 20 mg Mn/kg body weight per day, by gavage, for 30 days. Forty‐ eight hours after the last dose, animals were sacrificed by decapitation and brains were dissected into different regions to determine the concentration of amino acids by HPLC/EC. A dose dependent decrease in body weight gain was found in the adult, but not in the weanling rats. Significant increases occurred in concentrations of aspartate, glutamate, glutamine, taurine and y‐aminobutyric acid (GABA) in the cerebellum of the adult rats dosed with 20 mg/kg per day, Mn. A significant decrease in the concentration of glutamine was observed in caudate nucleus and hippocampus of weanling rats dosed with 10 mg/kg, Mn. These data suggest that chronic Mn exposure can produce a decrease in body weight gain in adult rats and alterations in amino acids in different regions of weanling and adult rat brains.