Brian J Reich

Marine protected areas (MPAs) are widely used for ocean conservation, yet the relative impacts of various types of MPAs are poorly understood. We estimated impacts on fish biomass from no-take and multiple-use (fished) MPAs, employing a rigorous matched counterfactual design with a global dataset of >14,000 surveys in and around 216 MPAs. Both no-take and multiple-use MPAs generated positive conservation outcomes relative to no protection (58.2% and 12.6% fish biomass increases, respectively), with smaller estimated differences between the two MPA types when controlling for additional confounding factors (8.3% increase). Relative performance depended on context and management: no-take MPAs performed better in areas of high human pressure but similar to multiple-use in remote locations. Multiple-use MPA performance was low in high-pressure areas but improved significantly with better …

The Supplementary Material consists of three appendices. Appendix A goes over the some properties of the PMM, and an overview of the variable importance measure used in the text. Appendix B presents supplementary simulation studies detailing the performance of the PMM in various density estimation and parameter estimation scenarios. Appendix C consists of additional results from the HCDN data analysis, including MCMC convergence, model comparison and model fit results, and selected results from analyzing the extremal streamflow data in its original scale.

Spatially dependent data arises in many applications, and Gaussian processes are a popular modeling choice for these scenarios. While Bayesian analyses of these problems have proven to be successful, selecting prior distributions for these complex models remains a difficult task. In this work, we propose a principled approach for setting prior distributions on model variance components by placing a prior distribution on a measure of model fit. In particular, we derive the distribution of the prior coefficient of determination. Placing a beta prior distribution on this measure induces a generalized beta prime prior distribution on the global variance of the linear predictor in the model. This method can also be thought of as shrinking the fit towards the intercept‐only (null) model. We derive an efficient Gibbs sampler for the majority of the parameters and use Metropolis–Hasting updates for the others. Finally, the method is …

Public health data are often spatially dependent, but standard spatial regression methods can suffer from bias and invalid inference when the independent variable is associated with spatially-correlated residuals. This could occur if, for example, there is an unmeasured environmental contaminant. Geoadditive structural equation modeling (gSEM), in which an estimated spatial trend is removed from both the explanatory and response variables before estimating the parameters of interest, has previously been proposed as a solution, but there has been little investigation of gSEM's properties with point-referenced data. We link gSEM to results on double machine learning and semiparametric regression based on two-stage procedures. We propose using these semiparametric estimators for spatial regression using Gaussian processes with Mat\`ern covariance to estimate the spatial trends, and term this class of estimators Double Spatial Regression (DSR). We derive regularity conditions for root- asymptotic normality and consistency and closed-form variance estimation, and show that in simulations where standard spatial regression estimators are highly biased and have poor coverage, DSR can mitigate bias more effectively than competitors and obtain nominal coverage.

Transfer learning (TL) has emerged as a powerful tool to supplement data collected for a target task with data collected for a related source task. The Bayesian framework is natural for TL because information from the source data can be incorporated in the prior distribution for the target data analysis. In this paper, we propose and study Bayesian TL methods for the normal-means problem and multiple linear regression. We propose two classes of prior distributions. The first class assumes the difference in the parameters for the source and target tasks is sparse, i.e., many parameters are shared across tasks. The second assumes that none of the parameters are shared across tasks, but the differences are bounded in -norm. For the sparse case, we propose a Bayes shrinkage estimator with theoretical guarantees under mild assumptions. The proposed methodology is tested on synthetic data and outperforms state-of-the-art TL methods. We then use this method to fine-tune the last layer of a neural network model to predict the molecular gap property in a material science application. We report improved performance compared to classical fine tuning and methods using only the target data.

BackgroundPer- and polyfluoroalkyl substances (PFAS) are persistent organic pollutants with emerging environmental and regulatory concerns.ObjectivesThis study aimed to estimate the burden of PFAS exposures through ground water on the incidence of chronic health conditions among Medicare beneficiaries aged 65 years and older.MethodsWe estimated PFAS groundwater concentrations for every ZIP code tabulated area (ZCTA) in California counties where 25 percent or more of the population’s drinking water was derived from groundwater. We calculated the annual incidence of non-cancer chronic health conditions among 1,696,247 Medicare beneficiaries aged 65 and older by residential ZCTA over the seven-year study period (2011-2017). A Poisson regression model was used to estimate associations between PFAS groundwater concentration and chronic condition incidence with an offset for the number of beneficiary-years at risk and adjusting for bias due to non-random sampling of wells, use of groundwater for drinking water, demographic characteristics, and lung cancer incidence as a control for smoking.ResultsResults suggest an association between a 10 ng/L increment in PFAS contaminated groundwater and chronic health conditions including hypertension (+1.15%, 95% confidence interval (CI) 1.01, 1.30), chronic kidney disease (+0.83%, 95% CI 0.68, 0.99) and cataracts (+1.50%, 95% CI 1.35, 1.66).DiscussionThis small increment in the incidence rate would produce an additional 1,700 new cases of hypertension each year in the study population.

To estimate design-flood quantiles, such as the 100-year flood, the observed annual maximum flood (AMF) series is fitted to a probability distribution and then the design flood quantile is estimated from the fitted distribution. This is because, in most cases, historical records are not long enough to observe rare, design-flood events. Changes in the AMF series, which are usually detected using simple trend tests (eg, Mann-Kendall test (MKT)), are hypothesized to result in changes in design-flood estimates. This hypothesis is tested by using an alternate framework to detect significant changes in design-flood between two periods–rather than changes in the AMF series–and then evaluated using synthetically generated AMF series from the Log-Pearson Type-3 (LP3) distribution due to changes in moments associated with flood distribution. Synthetic experiments show that the MKT does not consider changes in all three …

Perfluoro-3, 5, 7, 9, 11-pentaoxadodecanoic acid (PFO5DoA, DTXSID50723994) is a perfluoroalkyl ether acid (PFEA) produced at a fluorochemical facility (“Fayetteville Works”) in Bladen County, North Carolina. In 2015, PFO5DoA was first identified in Cape Fear River water samples collected downstream of the facility’s wastewater discharge point. 1 Approximately 280,000 people rely on public water sourced from the lower Cape Fear River. 2 The GenX Exposure Study started in 2017 to characterize PFEA exposure in Cape Fear River Basin, North Carolina, residents. We detected three PFEAs—ethanesulfonic acid, 2-[1-[difluoro (1, 2, 2, 2-tetrafluoroethoxy) methyl]-1, 2, 2, 2-tetrafluoroethoxy]-1, 1, 2, 2-tetrafluoro-(also known as Nafion by-product 2, DTXSID10892352); perfluoro (3, 5, 7, 9-butaoxadecanoic) acid (PFO4DA, DTXSID90723993); and PFO5DoA—in blood serum from nearly all 344 participants who …

Per-and polyfluoroalkyl substances (PFAS) are a large class of synthetic, fluorinated chemicals. 1 Wastewater discharges from a fluorochemical manufacturing facility (“Fayetteville Works”) contaminated the lower Cape Fear River in North Carolina with per-and polyfluoroalkyl ether acids (PFEAs), 2 a subgroup of PFAS. The GenX Exposure Study aims to characterize exposure to PFAS in Cape Fear River Basin, North Carolina, residents. The study started in 2017 with Wilmington, North Carolina, residents who were exposed to PFEAs through municipal water derived from the lower Cape Fear River. 3, 4 In some Wilmington serum samples, we identified 2, 2, 3, 3-tetrafluoro-3-((1, 1, 1, 2, 3, 3-hexafluoro-3-(1, 2, 2, 2-tetrafluoroethoxy) propan-2-yl) oxy) propanoic acid (also known as Hydro-EVE, DTXSID60904459), a polyfluoroalkyl ether carboxylic acid generated at Fayetteville Works. This chemical was first identified …

In this paper, we propose a new regime-based model to describe spatio-temporal dynamics of precipitation data. Precipitation is one of the most essential factors for multiple human-related activities such as agriculture production. Therefore, a detailed and accurate understanding of the rain for a given region is needed. Motivated by the different formations of precipitation systems (convective, frontal, and orographic), we proposed a hierarchical regime-based spatio-temporal model for precipitation data. We use information about the values of neighbouring sites to identify such regimes, allowing spatial and temporal dependence to be different among regimes. Using the Bayesian approach with R INLA, we fit our model to the Guanajuato state (Mexico) precipitation data case study to understand the spatial and temporal dependencies of precipitation in this region. Our findings show the regime-based model’s …

Importance Studies elucidating determinants of residential neighborhood–level health inequities are needed. Objective To quantify associations of structural racism indicators with neighborhood prevalence of chronic kidney disease (CKD), diabetes, and hypertension. Design, Setting, and Participants This cross-sectional study used public data (2012-2018) and deidentified electronic health records (2017-2018) to describe the burden of structural racism and the prevalence of CKD, diabetes, and hypertension in 150 residential neighborhoods in Durham County, North Carolina, from US census block groups and quantified their associations using bayesian models accounting for spatial correlations and residents’ age. Data were analyzed from January 2021 to May 2023. Exposures Global (neighborhood percentage of White residents, economic-racial segregation, and area deprivation) and discrete (neighborhood …

Extreme environmental events frequently exhibit spatial and temporal dependence. These data are often modeled using Max Stable Processes (MSPs) that are computationally prohibitive to fit for as few as a dozen observations. Supposed computationally-efficient approaches like the composite likelihood remain computationally burdensome with a few hundred observations. In this article, we propose a spatial partitioning approach based on local modeling of subsets of the spatial domain that delivers computationally and statistically efficient inference. Marginal and dependence parameters of the MSP are estimated locally on subsets of observations using censored pairwise composite likelihood, and combined using a modified generalized method of moments procedure. The proposed distributed approach is extended to estimate inverted MSP models, and to estimate spatially varying coefficient models to deliver …

Changes in annual maximum flood (AMF), which are usually detected using simple trend tests (e.g., Mann‐Kendall test (MKT)), are expected to change design‐flood estimates. We propose an alternate framework to detect significant changes in design‐flood between two periods and evaluate it for synthetically generated AMF from the Log‐Pearson Type‐3 (LP3) distribution due to changes in moments associated with flood distribution. Synthetic experiments show MKT does not consider changes in all three moments of the LP3 distribution and incorrectly detects changes in design‐flood. We applied the framework on 31 river basins spread across the United States. Statistically significant changes in design‐flood quantiles were observed even without a significant trend in AMF and basins with statistically significant trend did not necessarily exhibit statistically significant changes in design‐flood. We recommend …

BackgroundFractional exhaled nitric oxide (FeNO) is a marker of airway inflammation. Elevated FeNO has been associated with environmental exposures, however, studies from tropical countries are limited. Using data from the Infants' Environmental Health Study (ISA) birth cohort, we evaluated medical conditions and environmental exposures’ association with elevated FeNO.MethodsWe performed a cross-sectional analysis of 277 women and 293 8-year old children who participated in the 8-year post-partum visit in 2019. We measured FeNO and collected information on medical conditions and environmental exposures including smoke from waste burning, work in banana plantations, and home pesticide use. We defined elevated FeNO as >25 ppb for women and >20 ppb for children. To evaluate factors associated with elevated FeNO, we used logistic regression models adjusted for obesity in women and …

Biological sex and gender are critical variables in biomedical research, but are complicated by the presence of sex‐specific natural hormone cycles, such as the estrous cycle in female rodents, typically divided into phases. A common feature of these cycles are fluctuating hormone levels that induce sex differences in many behaviors controlled by the electrophysiology of neurons, such as neuronal membrane potential in response to electrical stimulus, typically summarized using a priori defined metrics. In this paper, we propose a method to test for differences in the electrophysiological properties across estrous cycle phase without first defining a metric of interest. We do this by modeling membrane potential data in the frequency domain as realizations of a bivariate process, also depending on the electrical stimulus, by adopting existing methods for longitudinal functional data. We are then able to extract the main …

The Cretaceous Period (145 to 66 Ma) was a prolonged warmhouse to hothouse period characterized by high atmospheric CO2 conditions, elevated surface temperatures, and an enhanced global hydrologic cycle. It provides a case study for understanding how a hothouse climate system operates, and is an analog for future anthropogenic climate change scenarios. This study presents new quantitative temperature and precipitation proxy datasets for nine key Cretaceous time slices (Berriasian/Valanginian, Hauterivian/Barremian, Aptian, Albian, Cenomanian, Turonian, Coniacian/Santonian, Campanian, Maastrichtian), and a new geostatistical analysis technique that utilizes Markov Chain Monte Carlo algorithm and Bayesian hierarchical models to generate high resolution, quantitative global paleoclimate reconstructions from these proxy datasets, with associated uncertainties. Using these paleoclimate …

Nag, P., Sun, Y., & Reich, BJ (2023). Spatio-temporal DeepKriging for interpolation and probabilistic forecasting. Spatial Statistics, 100773. https://doi. org/10.1016/j. spasta. 2023.100773. DOI: 10.1016/j. spasta. 2023.100773 Handle: 10754/693089

Adjusting for an unmeasured confounder is generally an intractable problem, but in the spatial setting it may be possible under certain conditions. We derive necessary conditions on the coherence between the exposure and the unmeasured confounder that ensure the effect of exposure is estimable. We specify our model and assumptions in the spectral domain to allow for different degrees of confounding at different spatial resolutions. One assumption that ensures identifiability is that confounding present at global scales dissipates at local scales. We show that this assumption in the spectral domain is equivalent to adjusting for global-scale confounding in the spatial domain by adding a spatially smoothed version of the exposure to the mean of the response variable. Within this general framework, we propose a sequence of confounder adjustment methods that range from parametric adjustments based on …

The Appendix is divided into three sections. The first section details the complete procedure to evaluate the approximation in (5). The second section provides all the details of the optimization procedure for all the non-Bayesian methods considered in the paper. Finally, the last section covers the details of the MCMC details for the Bayesian ethods.

Flexible estimation of multiple conditional quantiles is of interest in numerous applications, such as studying the effect of pregnancy‐related factors on low and high birth weight. We propose a Bayesian nonparametric method to simultaneously estimate noncrossing, nonlinear quantile curves. We expand the conditional distribution function of the response in I‐spline basis functions where the covariate‐dependent coefficients are modeled using neural networks. By leveraging the approximation power of splines and neural networks, our model can approximate any continuous quantile function. Compared to existing models, our model estimates all rather than a finite subset of quantiles, scales well to high dimensions, and accounts for estimation uncertainty. While the model is arbitrarily flexible, interpretable marginal quantile effects are estimated using accumulative local effect plots and variable importance …