Seasonality of cholera in Kolkata and the influence of climate.

BACKGROUND: Cholera in Kolkata remains endemic and the Indian city is burdened with a high number of annual cases. Climate change is widely considered to exacerbate cholera, however the precise relationship between climate and cholera is highly heterogeneous in space and considerable variation can be observed even within the Indian subcontinent. To date, relatively few studies have been conducted regarding the influence of climate on cholera in Kolkata. METHODS: We considered 21 years of confirmed cholera cases from the Infectious Disease Hospital in Kolkata during the period of 1999-2019. We used Generalised Additive Modelling (GAM) to extract the non-linear relationship between cholera and different climatic factors; temperature, rainfall and sea surface temperature (SST). Peak associated lag times were identified using cross-correlation lag analysis. RESULTS: Our findings revealed a bi-annual pattern of cholera cases with two peaks coinciding with the increase in temperature in summer and the onset of monsoon rains. Variables selected as explanatory variables in the GAM model were temperature and rainfall. Temperature was the only significant factor associated with summer cholera (mean temperature of 30.3 °C associated with RR of 3.8) while rainfall was found to be the main driver of monsoon cholera (550 mm total monthly rainfall associated with RR of 3.38). Lag time analysis revealed that the association between temperature and cholera cases in the summer had a longer peak lag time compared to that between rainfall and cholera during the monsoon. We propose several mechanisms by which these relationships are mediated. CONCLUSIONS: Kolkata exhibits a dual-peak phenomenon with independent mediating factors. We suggest that the summer peak is due to increased bacterial concentration in urban water bodies, while the monsoon peak is driven by contaminated flood waters. Our results underscore the potential utility of preventative strategies tailored to these seasonal and climatic patterns, including efforts to reduce direct contact with urban water bodies in summer and to protect residents from flood waters during monsoon.

Proliferation of atmospheric datasets can hinder policy making: a data blending technique offers a solution.

The proliferation of atmospheric datasets is a key outcome from the continued development and advancement of our collective scientific understanding. Yet often datasets describing ostensibly identical processes or atmospheric variables provide widely varying results. As an example, we analyze several datasets representing rainfall over Nepal. We show that estimates of extreme rainfall are highly variable depending on which dataset you choose to look at. This leads to confusion and inaction from policy-focused decision makers. Scientifically, we should use datasets that sample a range of creation methodologies and prioritize the use of data science techniques that have the flexibility to incorporate these multiple sources of data. We demonstrate the use of a statistically interpretable data blending technique to help discern and communicate a consensus result, rather than imposing a priori judgment on the choice of dataset, for the benefit of policy decision making.

Current and future trends in heat-related mortality in the MENA region: a health impact assessment with bias-adjusted statistically downscaled CMIP6 (SSP-based) data and Bayesian inference.

BACKGROUND: The Middle East and North Africa (MENA) is one of the regions that is most vulnerable to the negative effects of climate change, yet the potential public health impacts have been underexplored compared to other regions. We aimed to examine one aspect of these impacts, heat-related mortality, by quantifying the current and future burden in the MENA region and identifying the most vulnerable countries. METHODS: We did a health impact assessment using an ensemble of bias-adjusted statistically downscaled Coupled Model Intercomparison Project phase 6 (CMIP6) data based on four Shared Socioeconomic Pathway (SSP) scenarios (SSP1-2·6 [consistent with a 2°C global warming scenario], SSP2-4·5 [medium pathway scenario], SSP3-7·0 [pessimistic scenario], and SSP5-8·5 [high emissions scenario]) and Bayesian inference methods. Assessments were based on apparent temperature-mortality relationships specific to each climate subregion of MENA based on Koppen-Geiger climate type classification, and unique thresholds were characterised for each 50 km grid cell in the region. Future annual heat-related mortality was estimated for the period 2021-2100. Estimates were also presented with population held constant to quantify the contribution of projected demographic changes to the future heat-mortality burden. FINDINGS: The average annual heat-related death rate across all MENA countries is currently 2·1 per 100 000 people. Under the two high emissions scenarios (SSP3-7·0 and SSP5-8·5), most of the MENA region will have experienced substantial warming by the 2060s. Annual heat-related deaths of 123·4 per 100 000 people are projected for MENA by 2100 under a high emissions scenario (SSP5-8·5), although this rate would be reduced by more than 80% (to 20·3 heat-related deaths per 100 000 people per year) if global warming could be limited to 2°C (ie, under the SSP1-2·6 scenario). Large increases are also expected by 2100 under the SSP3-7·0 scenario (89·8 heat-related deaths per 100 000 people per year) due to the high population growth projected under this pathway. Projections in MENA are far higher than previously observed in other regions, with Iran expected to be the most vulnerable country. INTERPRETATION: Stronger climate change mitigation and adaptation policies are needed to avoid these heat-related mortality impacts. Since much of this increase will be driven by population changes, demographic policies and healthy ageing will also be key to successful adaptation. FUNDING: National Institute for Health Research, EU Horizon 2020.

Applying an ecosystem services framework on nature and mental health to recreational blue space visits across 18 countries.

The effects of 'nature' on mental health and subjective well-being have yet to be consistently integrated into ecosystem service models and frameworks. To address this gap, we used data on subjective mental well-being from an 18-country survey to test a conceptual model integrating mental health with ecosystem services, initially proposed by Bratman et al. We analysed a range of individual and contextual factors in the context of 14,998 recreational visits to blue spaces, outdoor environments which prominently feature water. Consistent with the conceptual model, subjective mental well-being outcomes were dependent upon on a complex interplay of environmental type and quality, visit characteristics, and individual factors. These results have implications for public health and environmental management, as they may help identify the bluespace locations, environmental features, and key activities, that are most likely to impact well-being, but also potentially affect recreational demand on fragile aquatic ecosystems.

A data integration framework for spatial interpolation of temperature observations using climate model data.

Meteorological station measurements are an important source of information for understanding the weather and its association with risk, and are vital in quantifying climate change. However, such data tend to lack spatial coverage and are often plagued with flaws such as erroneous outliers and missing values. Alternative meteorological data exist in the form of climate model output that have better spatial coverage, at the expense of bias. We propose a probabilistic framework to integrate temperature measurements with climate model (reanalysis) data, in a way that allows for biases and erroneous outliers, while enabling prediction at any spatial resolution. The approach is Bayesian which facilitates uncertainty quantification and simulation based inference, as illustrated by application to two countries from the Middle East and North Africa region, an important climate change hotspot. We demonstrate the use of the model in: identifying outliers, imputing missing values, non-linear bias correction, downscaling and aggregation to any given spatial configuration.

Quantifying Spatio-temporal risk of Harmful Algal Blooms and their impacts on bivalve shellfish mariculture using a data-driven modelling approach.

Harmful algal blooms (HABs) intoxicate and asphyxiate marine life, causing devastating environmental and socio-economic impacts, costing at least $8bn/yr globally. Accumulation of phycotoxins from HAB phytoplankton in filter-feeding shellfish can poison human consumers, prompting harvesting closures at shellfish production sites. To quantify long-term intoxication risk from Dinophysis HAB species, we used historical HAB monitoring data (2009-2020) to develop a new modelling approach to predict Dinophysis toxin concentrations in a range of bivalve shellfish species at shellfish sites in Western Scotland, South-West England and Northern France. A spatiotemporal statistical modelling framework was developed within the Generalized Additive Model (GAM) framework to quantify long-term HAB risks for different bivalve shellfish species across each region, capturing seasonal variations, and spatiotemporal interactions. In all regions spatial functions were most important for predicting seasonal HAB risk, offering the potential to inform optimal siting of new shellfish operations and safe harvesting periods for businesses. A 10-fold cross-validation experiment was carried out for each region, to test the models' ability to predict toxin risk at harvesting locations for which data were withheld from the model. Performance was assessed by comparing ranked predicted and observed mean toxin levels at each site within each region: the correlation of ranks was 0.78 for Northern France, 0.64 for Western Scotland, and 0.34 for South-West England, indicating our approach has promise for predicting unknown HAB risk, depending on the region and suitability of training data.

Correcting delayed reporting of COVID-19 using the generalized-Dirichlet-multinomial method.

The COVID-19 pandemic has highlighted delayed reporting as a significant impediment to effective disease surveillance and decision-making. In the absence of timely data, statistical models which account for delays can be adopted to nowcast and forecast cases or deaths. We discuss the four key sources of systematic and random variability in available data for COVID-19 and other diseases, and critically evaluate current state-of-the-art methods with respect to appropriately separating and capturing this variability. We propose a general hierarchical approach to correcting delayed reporting of COVID-19 and apply this to daily English hospital deaths, resulting in a flexible prediction tool which could be used to better inform pandemic decision-making. We compare this approach to competing models with respect to theoretical flexibility and quantitative metrics from a 15-month rolling prediction experiment imitating a realistic operational scenario. Based on consistent leads in predictive accuracy, bias, and precision, we argue that this approach is an attractive option for correcting delayed reporting of COVID-19 and future epidemics.

Household cooking fuel estimates at global and country level for 1990 to 2030.

Household air pollution generated from the use of polluting cooking fuels and technologies is a major source of disease and environmental degradation in low- and middle-income countries. Using a novel modelling approach, we provide detailed global, regional and country estimates of the percentages and populations mainly using 6 fuel categories (electricity, gaseous fuels, kerosene, biomass, charcoal, coal) and overall polluting/clean fuel use - from 1990-2020 and with urban/rural disaggregation. Here we show that 53% of the global population mainly used polluting cooking fuels in 1990, dropping to 36% in 2020. In urban areas, gaseous fuels currently dominate, with a growing reliance on electricity; in rural populations, high levels of biomass use persist alongside increasing use of gaseous fuels. Future projections of observed trends suggest 31% will still mainly use polluting fuels in 2030, including over 1 billion people in Sub-Saharan African by 2025.

Associations between green/blue spaces and mental health across 18 countries.

Living near, recreating in, and feeling psychologically connected to, the natural world are all associated with better mental health, but many exposure-related questions remain. Using data from an 18-country survey (n = 16,307) we explored associations between multiple measures of mental health (positive well-being, mental distress, depression/anxiety medication use) and: (a) exposures (residential/recreational visits) to different natural settings (green/inland-blue/coastal-blue spaces); and (b) nature connectedness, across season and country. People who lived in greener/coastal neighbourhoods reported higher positive well-being, but this association largely disappeared when recreational visits were controlled for. Frequency of recreational visits to green, inland-blue, and coastal-blue spaces in the last 4 weeks were all positively associated with positive well-being and negatively associated with mental distress. Associations with green space visits were relatively consistent across seasons and countries but associations with blue space visits showed greater heterogeneity. Nature connectedness was also positively associated with positive well-being and negatively associated with mental distress and was, along with green space visits, associated with a lower likelihood of using medication for depression. By contrast inland-blue space visits were associated with a greater likelihood of using anxiety medication. Results highlight the benefits of multi-exposure, multi-response, multi-country studies in exploring complexity in nature-health associations.

Environmental DNA reveals links between abundance and composition of airborne grass pollen and respiratory health.

Grass (Poaceae) pollen is the most important outdoor aeroallergen,1 exacerbating a range of respiratory conditions, including allergic asthma and rhinitis ("hay fever").2-5 Understanding the relationships between respiratory diseases and airborne grass pollen with a view to improving forecasting has broad public health and socioeconomic relevance. It is estimated that there are over 400 million people with allergic rhinitis6 and over 300 million with asthma, globally,7 often comorbidly.8 In the UK, allergic asthma has an annual cost of around US$ 2.8 billion (2017).9 The relative contributions of the >11,000 (worldwide) grass species (C. Osborne et al., 2011, Botany Conference, abstract) to respiratory health have been unresolved,10 as grass pollen cannot be readily discriminated using standard microscopy.11 Instead, here we used novel environmental DNA (eDNA) sampling and qPCR12-15 to measure the relative abundances of airborne pollen from common grass species during two grass pollen seasons (2016 and 2017) across the UK. We quantitatively demonstrate discrete spatiotemporal patterns in airborne grass pollen assemblages. Using a series of generalized additive models (GAMs), we explore the relationship between the incidences of airborne pollen and severe asthma exacerbations (sub-weekly) and prescribing rates of drugs for respiratory allergies (monthly). Our results indicate that a subset of grass species may have disproportionate influence on these population-scale respiratory health responses during peak grass pollen concentrations. The work demonstrates the need for sensitive and detailed biomonitoring of harmful aeroallergens in order to investigate and mitigate their impacts on human health.

A framework for probabilistic weather forecast post-processing across models and lead times using machine learning.

Forecasting the weather is an increasingly data-intensive exercise. Numerical weather prediction (NWP) models are becoming more complex, with higher resolutions, and there are increasing numbers of different models in operation. While the forecasting skill of NWP models continues to improve, the number and complexity of these models poses a new challenge for the operational meteorologist: how should the information from all available models, each with their own unique biases and limitations, be combined in order to provide stakeholders with well-calibrated probabilistic forecasts to use in decision making? In this paper, we use a road surface temperature example to demonstrate a three-stage framework that uses machine learning to bridge the gap between sets of separate forecasts from NWP models and the 'ideal' forecast for decision support: probabilities of future weather outcomes. First, we use quantile regression forests to learn the error profile of each numerical model, and use these to apply empirically derived probability distributions to forecasts. Second, we combine these probabilistic forecasts using quantile averaging. Third, we interpolate between the aggregate quantiles in order to generate a full predictive distribution, which we demonstrate has properties suitable for decision support. Our results suggest that this approach provides an effective and operationally viable framework for the cohesive post-processing of weather forecasts across multiple models and lead times to produce a well-calibrated probabilistic output. This article is part of the theme issue 'Machine learning for weather and climate modelling'.

Peaks in bat activity at turbines and the implications for mitigating the impact of wind energy developments on bats.

Wind turbines are a relatively new threat to bats, causing mortalities worldwide. Reducing these fatalities is essential to ensure that the global increase in wind-energy facilities can occur with minimal impact on bat populations. Although individual bats have been observed approaching wind turbines, and fatalities frequently reported, it is unclear whether bats are actively attracted to, indifferent to, or repelled by, the turbines at large wind-energy installations. In this study, we assessed bat activity at paired turbine and control locations at 23 British wind farms. The research focussed on Pipistrellus species, which were by far the most abundant bats recorded at these sites. P. pipistrellus activity was 37% higher at turbines than at control locations, whereas P. pygmaeus activity was consistent with no attraction or repulsion by turbines. Given that more than 50% of bat fatalities in Europe are P. pipistrellus, these findings help explain why Environmental Impact Assessments conducted before the installation of turbines are poor predictors of actual fatality rates. They also suggest that operational mitigation (minimising blade rotation in periods of high collision risk) is likely to be the most effective way to reduce collisions because the presence of turbines alters bat activity.

Results from an 18 country cross-sectional study examining experiences of nature for people with common mental health disorders.

Exposure to natural environments is associated with a lower risk of common mental health disorders (CMDs), such as depression and anxiety, but we know little about nature-related motivations, practices and experiences of those already experiencing CMDs. We used data from an 18-country survey to explore these issues (n = 18,838), taking self-reported doctor-prescribed medication for depression and/or anxiety as an indicator of a CMD (n = 2698, 14%). Intrinsic motivation for visiting nature was high for all, though slightly lower for those with CMDs. Most individuals with a CMD reported visiting nature ≥ once a week. Although perceived social pressure to visit nature was associated with higher visit likelihood, it was also associated with lower intrinsic motivation, lower visit happiness and higher visit anxiety. Individuals with CMDs seem to be using nature for self-management, but 'green prescription' programmes need to be sensitive, and avoid undermining intrinsic motivation and nature-based experiences.

A laboratory investigation into features of morphology and physiology for their potential to predict reproductive success in male frogs.

Amphibian populations are declining globally, however, the contribution of reduced reproduction to declines is unknown. We investigated associations between morphological (weight/snout-vent length, nuptial pad colour/size, forelimb width/size) and physiological (nuptial pad/testis histomorphology, plasma hormones, gene expression) features with reproductive success in males as measured by amplexus success and fertility rate (% eggs fertilised) in laboratory maintained Silurana/Xenopus tropicalis. We explored the robustness of these features to predict amplexus success/fertility rate by investigating these associations within a sub-set of frogs exposed to anti-androgens (flutamide (50 µg/L)/linuron (9 or 45 µg/L)). In unexposed males, nuptial pad features (size/colour/number of hooks/androgen receptor mRNA) were positively associated with amplexus success, but not with fertility rate. In exposed males, many of the associations with amplexus success differed from untreated animals (they were either reversed or absent). In the exposed males forelimb width/nuptial pad morphology were also associated with fertility rate. However, a more darkly coloured nuptial pad was positively associated with amplexus success across all groups and was indicative of androgen status. Our findings demonstrate the central role for nuptial pad morphology in reproductive success in S. tropicalis, however, the lack of concordance between unexposed/exposed frogs complicates understanding of the utility of features of nuptial pad morphology as biomarkers in wild populations. In conclusion, our work has indicated that nuptial pad and forelimb morphology have potential for development as biomarkers of reproductive health in wild anurans, however, further research is needed to establish this.

Multivariate hierarchical frameworks for modeling delayed reporting in count data.

In many fields and applications, count data can be subject to delayed reporting. This is where the total count, such as the number of disease cases contracted in a given week, may not be immediately available, instead arriving in parts over time. For short-term decision making, the statistical challenge lies in predicting the total count based on any observed partial counts, along with a robust quantification of uncertainty. We discuss previous approaches to modeling delayed reporting and present a multivariate hierarchical framework where the count generating process and delay mechanism are modeled simultaneously in a flexible way. This framework can also be easily adapted to allow for the presence of underreporting in the final observed count. To illustrate our approach and to compare it with existing frameworks, we present a case study of reported dengue fever cases in Rio de Janeiro. Based on both within-sample and out-of-sample posterior predictive model checking and arguments of interpretability, adaptability, and computational efficiency, we discuss the relative merits of different approaches.

Desirable BUGS in models of infectious diseases.

Bayesian inference using Gibbs sampling (BUGS) is a set of statistical software that uses Markov chain Monte Carlo (MCMC) methods to estimate almost any specified model. Originally developed in the late 1980s, the software is an excellent introduction to applied Bayesian statistics without the need to write a MCMC sampler. The software is typically used for regression-based analyses, but any model that can be specified using graphical nodes are possible. Advanced topics such as missing data, spatial analysis, model comparison and dynamic infectious disease models can be tackled. Three examples are provided; a linear regression model to illustrate parameter estimation, the steps to ensure that the estimates have converged and a comparison of run-times across different computing platforms. The second example describes a model that estimates the probability of being vaccinated from cross-sectional and surveillance data, and illustrates the specification of different models, model comparison and data augmentation. The third example illustrates estimation of parameters within a dynamic Susceptible-Infected-Recovered model. These examples show that BUGS can be used to estimate parameters from models relevant for infectious diseases, and provide an overview of the relative merits of the approach taken.