Understanding spatial effects in species distribution models.

Species Distribution Models often include spatial effects which may improve prediction at unsampled locations and reduce Type I errors when identifying environmental drivers. In some cases ecologists try to ecologically interpret the spatial patterns displayed by the spatial effect. However, spatial autocorrelation may be driven by many different unaccounted drivers, which complicates the ecological interpretation of fitted spatial effects. This study aims to provide a practical demonstration that spatial effects are able to smooth the effect of multiple unaccounted drivers. To do so we use a simulation study that fit model-based spatial models using both geostatistics and 2D smoothing splines. Results show that fitted spatial effects resemble the sum of the unaccounted covariate surface(s) in each model.

A flexible and efficient Bayesian implementation of point process models for spatial capture-recapture data.

Spatial capture-recapture (SCR) is now routinely used for estimating abundance and density of wildlife populations. A standard SCR model includes sub-models for the distribution of individual activity centers (ACs) and for individual detections conditional on the locations of these ACs. Both sub-models can be expressed as point processes taking place in continuous space, but there is a lack of accessible and efficient tools to fit such models in a Bayesian paradigm. Here, we describe a set of custom functions and distributions to achieve this. Our work allows for more efficient model fitting with spatial covariates on population density, offers the option to fit SCR models using the semi-complete data likelihood (SCDL) approach instead of data augmentation, and better reflects the spatially continuous detection process in SCR studies that use area searches. In addition, the SCDL approach is more efficient than data augmentation for simple SCR models while losing its advantages for more complicated models that account for spatial variation in either population density or detection. We present the model formulation, test it with simulations, quantify computational efficiency gains, and conclude with a real-life example using non-invasive genetic sampling data for an elusive large carnivore, the wolverine (Gulo gulo) in Norway.

Benthic animal-borne sensors and citizen science combine to validate ocean modelling.

Developments in animal electronic tagging and tracking have transformed the field of movement ecology, but interest is also growing in the contributions of tagged animals to oceanography. Animal-borne sensors can address data gaps, improve ocean model skill and support model validation, but previous studies in this area have focused almost exclusively on satellite-telemetered seabirds and seals. Here, for the first time, we develop the use of benthic species as animal oceanographers by combining archival (depth and temperature) data from animal-borne tags, passive acoustic telemetry and citizen-science mark-recapture records from 2016-17 for the Critically Endangered flapper skate (Dipturus intermedius) in Scotland. By comparing temperature observations to predictions from the West Scotland Coastal Ocean Modelling System, we quantify model skill and empirically validate an independent model update. The results from bottom-temperature and temperature-depth profile validation (5,324 observations) fill a key data gap in Scotland. For predictions in 2016, we identified a consistent warm bias (mean = 0.53 °C) but a subsequent model update reduced bias by an estimated 109% and improved model skill. This study uniquely demonstrates the use of benthic animal-borne sensors and citizen-science data for ocean model validation, broadening the range of animal oceanographers in aquatic environments.

Combining rapid antigen testing and syndromic surveillance improves community-based COVID-19 detection in a low-income country.

Diagnostics for COVID-19 detection are limited in many settings. Syndromic surveillance is often the only means to identify cases but lacks specificity. Rapid antigen testing is inexpensive and easy-to-deploy but can lack sensitivity. We examine how combining these approaches can improve surveillance for guiding interventions in low-income communities in Dhaka, Bangladesh. Rapid-antigen-testing with PCR validation was performed on 1172 symptomatically-identified individuals in their homes. Statistical models were fitted to predict PCR-status using rapid-antigen-test results, syndromic data, and their combination. Under contrasting epidemiological scenarios, the models' predictive and classification performance was evaluated. Models combining rapid-antigen-testing and syndromic data yielded equal-to-better performance to rapid-antigen-test-only models across all scenarios with their best performance in the epidemic growth scenario. These results show that drawing on complementary strengths across rapid diagnostics, improves COVID-19 detection, and reduces false-positive and -negative diagnoses to match local requirements; improvements achievable without additional expense, or changes for patients or practitioners.

A downscaling approach to compare COVID-19 count data from databases aggregated at different spatial scales.

As the COVID-19 pandemic continues to threaten various regions around the world, obtaining accurate and reliable COVID-19 data is crucial for governments and local communities aiming at rigorously assessing the extent and magnitude of the virus spread and deploying efficient interventions. Using data reported between January and February 2020 in China, we compared counts of COVID-19 from near-real-time spatially disaggregated data (city level) with fine-spatial scale predictions from a Bayesian downscaling regression model applied to a reference province-level data set. The results highlight discrepancies in the counts of coronavirus-infected cases at the district level and identify districts that may require further investigation.

How does the community COVID-19 level of risk impact on that of a care home?

OBJECTIVES: To model the risk of COVID-19 mortality in British care homes conditional on the community level risk. METHODS: A two stage modeling process ("doubly latent") which includes a Besag York Mollie model (BYM) and a Log Gaussian Cox Process. The BYM is adopted so as to estimate the community level risks. These are incorporated in the Log Gaussian Cox Process to estimate the impact of these risks on that in care homes. RESULTS: For an increase in the risk at the community level, the number of COVID-19 related deaths in the associated care home would be increased by exp (0.833), 2. This is based on a simulated dataset. In the context of COVID-19 related deaths, this study has illustrated the estimation of the risk to care homes in the presence of background community risk. This approach will be useful in facilitating the identification of the most vulnerable care homes and in predicting risk to new care homes. CONCLUSIONS: The modeling of two latent processes have been shown to be successfully facilitated by the use of the BYM and Log Gaussian Cox Process Models. Community COVID-19 risks impact on that of the care homes embedded in these communities.

Accounting for preferential sampling in species distribution models.

Species distribution models (SDMs) are now being widely used in ecology for management and conservation purposes across terrestrial, freshwater, and marine realms. The increasing interest in SDMs has drawn the attention of ecologists to spatial models and, in particular, to geostatistical models, which are used to associate observations of species occurrence or abundance with environmental covariates in a finite number of locations in order to predict where (and how much of) a species is likely to be present in unsampled locations. Standard geostatistical methodology assumes that the choice of sampling locations is independent of the values of the variable of interest. However, in natural environments, due to practical limitations related to time and financial constraints, this theoretical assumption is often violated. In fact, data commonly derive from opportunistic sampling (e.g., whale or bird watching), in which observers tend to look for a specific species in areas where they expect to find it. These are examples of what is referred to as preferential sampling, which can lead to biased predictions of the distribution of the species. The aim of this study is to discuss a SDM that addresses this problem and that it is more computationally efficient than existing MCMC methods. From a statistical point of view, we interpret the data as a marked point pattern, where the sampling locations form a point pattern and the measurements taken in those locations (i.e., species abundance or occurrence) are the associated marks. Inference and prediction of species distribution is performed using a Bayesian approach, and integrated nested Laplace approximation (INLA) methodology and software are used for model fitting to minimize the computational burden. We show that abundance is highly overestimated at low abundance locations when preferential sampling effects not accounted for, in both a simulated example and a practical application using fishery data. This highlights that ecologists should be aware of the potential bias resulting from preferential sampling and account for it in a model when a survey is based on non-randomized and/or non-systematic sampling.

Identifying prognostic structural features in tissue sections of colon cancer patients using point pattern analysis.

Diagnosis and prognosis of cancer are informed by the architecture inherent in cancer patient tissue sections. This architecture is typically identified by pathologists, yet advances in computational image analysis facilitate quantitative assessment of this structure. In this article, we develop a spatial point process approach to describe patterns in cell distribution within tissue samples taken from colorectal cancer (CRC) patients. In particular, our approach is centered on the Palm intensity function. This leads to taking an approximate-likelihood technique in fitting point processes models. We consider two Neyman-Scott point processes and a void process, fitting these point process models to the CRC patient data. We find that the parameter estimates of these models may be used to quantify the spatial arrangement of cells. Importantly, we observe characteristic differences in the spatial arrangement of cells between patients who died from CRC and those alive at follow up.

Re-evaluation of individual diameter : height allometric models to improve biomass estimation of tropical trees.

Accurate estimation of tree biomass is necessary to provide realistic values of the carbon stored in the terrestrial biosphere. A recognized source of errors in tree aboveground biomass (AGB) estimation is introduced when individual tree height values (H) are not directly measured but estimated from diameter at breast height (DBH) using allometric equations. In this paper, we evaluate the performance of 12 alternative DBH : H equations and compare their effects on AGB estimation for three tropical forests that occur in contrasting climatic and altitudinal zones. We found that fitting a three-parameter Weibull function using data collected locally generated the lowest errors and bias in H estimation, and that equations fitted to these data were more accurate than equations with parameters derived from the literature. For computing AGB, the introduced error values differed notably among DBH : H allometric equations, and in most cases showed a clear bias that resulted in either over- or under-estimation of AGB. Fitting the three-parameter Weibull function minimized errors in AGB estimates in our study and we recommend its widespread adoption for carbon stock estimation. We conclude that many previous studies are likely to present biased estimates of AGB due to the method of H estimation.

The why, what, and how of global biodiversity indicators beyond the 2010 target.

The 2010 biodiversity target agreed by signatories to the Convention on Biological Diversity directed the attention of conservation professionals toward the development of indicators with which to measure changes in biological diversity at the global scale. We considered why global biodiversity indicators are needed, what characteristics successful global indicators have, and how existing indicators perform. Because monitoring could absorb a large proportion of funds available for conservation, we believe indicators should be linked explicitly to monitoring objectives and decisions about which monitoring schemes deserve funding should be informed by predictions of the value of such schemes to decision making. We suggest that raising awareness among the public and policy makers, auditing management actions, and informing policy choices are the most important global monitoring objectives. Using four well-developed indicators of biological diversity (extent of forests, coverage of protected areas, Living Planet Index, Red List Index) as examples, we analyzed the characteristics needed for indicators to meet these objectives. We recommend that conservation professionals improve on existing indicators by eliminating spatial biases in data availability, fill gaps in information about ecosystems other than forests, and improve understanding of the way indicators respond to policy changes. Monitoring is not an end in itself, and we believe it is vital that the ultimate objectives of global monitoring of biological diversity inform development of new indicators.

Functional principal component data analysis: a new method for analysing microbial community fingerprints.

A common approach to molecular characterisation of microbial communities in natural environments is the amplification of small subunit (SSU) rRNA genes or genes encoding enzymes essential for a particular ecosystem function. A range of 'fingerprinting' techniques are available for the analysis of amplification products of both types of gene enabling quantitative or semi-quantitative analysis of relative abundances of different community members, and facilitating analysis of communities from large numbers of samples, including replicates. Statistical models that have been applied in this context suffer from a number of unavoidable limitations, including lack of distinction between closely adjacent bands or peaks, particularly when these differ significantly in intensity or size. Current approaches to the analysis of banding structures derived from gels are typically based on standard multivariate analysis methods such as principal component analysis (PCA) which do not consider structure of DGGE gels but treat the intensity of each band as independent from the other bands, ignoring local neighbourhood structures. This paper assesses whether a new statistical analytical technique, based on functional data analysis (FDA) methods, improves the discriminatory ability of molecular fingerprinting techniques. The approach regards band intensities as a mathematical function of the location on the gel and explicitly includes neighbourhood structure in the analysis. A simulation study clearly reveals the weaknesses of the standard PCA approach as opposed to the FDA approach, which is then used to analyse experimental DGGE data.