Honey bee (Apis mellifera) colony loss is a widespread phenomenon with important economic and biological implications, whose drivers are still an open matter of investigation. We contribute to this line of research through a large-scale, multi-variable study combining multiple publicly accessible data sources. Specifically, we analyzed quarterly data covering the contiguous United States for the years 2015-2021, and combined open data on honey bee colony status and stressors, weather data, and land use. The different spatio-temporal resolutions of these data are addressed through an up-scaling approach that generates additional statistical features which capture more complex distributional characteristics and significantly improve modeling performance. Treating this expanded feature set with state-of-the-art feature selection methods, we obtained findings that, nation-wide, are in line with the current knowledge on the aggravating roles of Varroa destructor and pesticides in colony loss. Moreover, we found that extreme temperature and precipitation events, even when controlling for other factors, significantly impact colony loss. Overall, our results reveal the complexity of biotic and abiotic factors affecting managed honey bee colonies across the United States.

Studying and measuring accessibility to care services has become a major concern for health care management, particularly since the global financial collapse. This study focuses on Tuscany, an Italian region, which is re-organizing its inpatient and outpatient systems in line with new government regulations. The principal aim of the paper is to illustrate the application of GIS methods with real-world scenarios to provide support to evidence-based planning and resource allocation in healthcare. Spatial statistics and geographical analyses were used to provide health care policy makers with a real scenario of accessibility to outpatient clinics. Measures for a geographical potential spatial accessibility index using the two-step floating catchment area method for outpatient services in 2015 were calculated and used to simulate the rationalization and reorganization of outpatient services. Parameters including the distance to outpatient clinics and volumes of activity were taken into account. The spatial accessibility index and the simulation of reorganization in outpatient care delivery are presented through three cases, which highlight three different managerial strategies. The results revealed the municipalities where health policy makers could consider a new spatial location, a shutdown or combining selected outpatient clinics while ensuring equitable access to services. A GIS-based approach was designed to provide support to healthcare management and policy makers in defining evidence-based actions to guide the reorganization of a regional health care delivery system. The analysis provides an example of how GIS methods can be applied to an integrated framework of administrative health care and geographical data as a valuable instrument to improve the efficiency of healthcare service delivery, in relation to the population’s needs.

High-dimensional classification studies have become widespread across various domains. The large dimensionality, coupled with the possible presence of data contamination, motivates the use of robust, sparse estimation methods to improve model interpretability and ensure the majority of observations agree with the underlying parametric model. In this study, we propose a robust and sparse estimator for logistic regression models, which simultaneously tackles the presence of outliers and/or irrelevant features. Specifically, we propose the use of $L_0$-constraints and mixed-integer conic programming techniques to solve the underlying double combinatorial problem in a framework that allows one to pursue optimality guarantees. We use our proposal to investigate the main drivers of honey bee (Apis mellifera) loss through the annual winter loss survey data collected by the Pennsylvania State Beekeepers Association. Previous studies mainly focused on predictive performance, however our approach produces a more interpretable classification model and provides evidence for several outlying observations within the survey data. We compare our proposal with existing heuristic methods and non-robust procedures, demonstrating its effectiveness. In addition to the application to honey bee loss, we present a simulation study where our proposal outperforms other methods across most performance measures and settings.

Rising temperatures worldwide pose an existential threat to people, properties, and the environment. Urban areas are particularly vulnerable to temperature increases due to the heat island effect, which amplifies local heating. Throughout the world, several megacities experience summer temperatures that stress human survival. Generating very high-resolution temperature forecasts is a fundamental problem to mitigate the effects of urban warming. This paper uses the Analog Ensemble technique to downscale existing temperature forecast from a low resolution to a much higher resolution using private weather stations. A new downscaling approach, based on the reuse of the Analog Ensemble (AnEn) indices, resulted by the combination of days and Forecast Lead Time (FLT)s, is proposed. Specifically, temperature forecasts from the NAM-NMM Numerical Weather Prediction model at 12 km are downscaled using 83 Private Weather Stations data over Manhattan, New York City, New York. Forecasts for 84 h are generated, hourly for the first 36 h, and every three hours thereafter. The results are dense forecasts that capture the spatial variability of ambient conditions. The uncertainty associated with using non-vetted data is addressed.