Dani Gamerman’s research while affiliated with Federal University of Rio de Janeiro and other places

Many techniques have been proposed to model space-varying observation processes with a nonstationary spatial covariance structure and/or anisotropy, usually on a geostatistical framework. Nevertheless, there is an increasing interest in point process applications, and methodologies that take nonstationarity into account are welcomed. In this sense, this work proposes an extension of a class of spatial Cox process using spatial deformation. The proposed method enables the deformation behavior to be data-driven, through a multivariate latent Gaussian process. Inference leads to intractable posterior distributions that are approximated via MCMC. The convergence of algorithms based on the Metropolis–Hastings steps proved to be slow, and the computational efficiency of the Bayesian updating scheme was improved by adopting Hamiltonian Monte Carlo (HMC) methods. Our proposal was also compared against an alternative anisotropic formulation. Studies based on synthetic data provided empirical evidence of the benefit brought by the adoption of nonstationarity through our anisotropic structure. A real data application was conducted on the spatial spread of the Spodoptera frugiperda pest in a corn-producing agricultural area in southern Brazil. Once again, the proposed method demonstrated its benefit over alternatives.

Indigenous societies are known to have occupied the Amazon basin for more than 12,000 years, but the scale of their influence on Amazonian forests remains uncertain. We report the discovery, using LIDAR (light detection and ranging) information from across the basin, of 24 previously undetected pre-Columbian earthworks beneath the forest canopy. Modeled distribution and abundance of large-scale archaeological sites across Amazonia suggest that between 10,272 and 23,648 sites remain to be discovered and that most will be found in the southwest. We also identified 53 domesticated tree species significantly associated with earthwork occurrence probability, likely suggesting past management practices. Closed-canopy forests across Amazonia are likely to contain thousands of undiscovered archaeological sites around which pre-Columbian societies actively modified forests, a discovery that opens opportunities for better understanding the magnitude of ancient human influence on Amazonia and its current state

Preferential sampling is a common feature in geostatistics and occurs when the locations to be sampled are chosen based on information about the phenomena under study. In this case, point pattern models are commonly used as the probability law for the distribution of the locations. However, analytic intractability of the point process likelihood prevents its direct calculation. Many Bayesian (and non-Bayesian) approaches in non-parametric model specifications handle this difficulty with approximation-based methods. These approximations involve errors that are difficult to quantify and can lead to biased inference. This paper presents an approach for performing exact Bayesian inference for this setting without the need for model approximation. A qualitatively minor change on the traditional model is proposed to circumvent the likelihood intractability. This change enables the use of an augmented model strategy. Recent work on Bayesian inference for point pattern models can be adapted to the geostatistics setting and renders computational tractability for exact inference for the proposed methodology. Estimation of model parameters and prediction of the response at unsampled locations can then be obtained from the joint posterior distribution of the augmented model. Simulated studies showed good quality of the proposed model for estimation and prediction in a variety of preferentiality scenarios. The performance of our approach is illustrated in the analysis of real datasets and compares favourably against approximation-based approaches. The paper is concluded with comments regarding extensions of and improvements to the proposed methodology.

Detailed knowledge on the effects of air pollutants on human health is a prerequisite for the development of effective policies to reduce the adverse impact of ambient air pollution. However, measuring the effect of exposure on health outcomes is an extremely difficult task as the health impact of air pollution is known to vary over space and over different exposure periods. In general, standard approaches aggregate the information over space or time to simplify the study but this strategy fails to recognize important regional differences and runs into the well‐known risk of confounding the effects. However, modelling directly with the original, disaggregated data requires a highly dimensional model with the curse of dimensionality making inferences unstable; in these cases, the models tend to retain many irrelevant components and most relevant effects tend to be attenuated. The situation clearly calls for an intermediate solution that does not blindly aggregate data while preserving important regional features. We propose a dimension‐reduction approach based on latent factors driven by the data. These factors naturally absorb the relevant features provided by the data and establish the link between pollutants and health outcomes, instead of forcing a necessarily high‐dimensional link at the observational level. The dynamic structural equation approach is particularly suited for this task. The latent factor approach also provides a simple solution to the spatial misalignment caused by using variables with different spatial resolutions and the state‐space representation of the model favours the application of impulse response analysis. Our approach is discussed through the analysis of the short‐term effects of air pollution on hospitalization data from Lombardia and Piemonte regions (Italy).