Detecting and Visualizing Observation Hot-Spots in Massive Volunteer-Contributed Geographic Data across Spatial Scales Using GPU-Accelerated Kernel Density Estimation

Volunteer-contributed geographic data (VGI) is an important source of geospatial big data that support research and applications. A major concern on VGI data quality is that the underlying observation processes are inherently biased. Detecting observation hot-spots thus helps better understand the bias. Enabled by the parallel kernel density estimation (KDE) computational tool that can run on multiple GPUs (graphics processing units), this study conducted point pattern analyses on tens of millions of iNaturalist observations to detect and visualize volunteers’ observation hot-spots across spatial scales. It was achieved by setting varying KDE bandwidths in accordance with the spatial scales at which hot-spots are to be detected. The succession of estimated density surfaces were then rendered at a sequence of map scales for visual detection of hot-spots. This study offers an effective geovisualization scheme for hierarchically detecting hot-spots in massive VGI datasets, which is useful for understanding the pattern-shaping drivers that operate at multiple spatial scales. This research exemplifies a computational tool that is supported by high-performance computing and capable of efficiently detecting and visualizing multi-scale hot-spots in geospatial big data and contributes to expanding the toolbox for geospatial big data analytics.

Land-use history impacts spatial patterns and composition of woody plant species across a 35-hectare temperate forest plot

Land-use history is the template upon which contemporary plant and tree populations establish and interact with one another and exerts a legacy on the structure and dynamics of species assemblages and ecosystems. We use the first census (2010–2014) of a 35-ha forest-dynamics plot at the Harvard Forest in central Massachusetts to describe the composition and structure of the woody plants in this plot, assess their spatial associations within and among the dominant species using univariate and bivariate spatial point-pattern analysis, and examine the interactions between land-use history and ecological processes. The plot includes 108,632 live stems ≥ 1 cm in diameter (2,215 individuals/ha) and 7,595 standing dead stems ≥ 5 cm in diameter. Live tree basal area averaged 42.25 m2/ha, of which 84% was represented by Tsuga canadensis (14.0 m2/ ha), Quercus rubra (northern red oak; 9.6 m2/ ha), Acer rubrum (7.2 m2/ ha) and Pinus strobus (eastern white pine; 4.4 m2/ ha). These same four species also comprised 78% of the live aboveground biomass, which averaged 245.2 Mg/ ha. Across all species and size classes, the forest contains a preponderance (> 80,000) of small stems (<10-cm diameter) that exhibit a reverse-J size distribution. Significant spatial clustering of abundant overstory species was observed at all spatial scales examined. Spatial distributions of A. rubrum and Q. rubra showed negative intraspecific correlations in diameters up to at least a 150-m spatial lag, likely indicative of crowding effects in dense forest patches following intensive past land use. Bivariate marked point-pattern analysis, showed that T. canadensis and Q. rubra diameters were negatively associated with one another, indicating resource competition for light. Distribution and abundance of the common overstory species are predicted best by soil type, tree neighborhood effects, and two aspects of land-use history: when fields were abandoned in the late 19th century and the succeeding forest types recorded in 1908. In contrast, a history of intensive logging prior to 1950 and a damaging hurricane in 1938 appear to have had little effect on the distribution and abundance of present-day tree species. Our findings suggest that current day composition and structure are still being influenced by anthropogenic disturbances that occurred over a century ago.

Lockout, Lockdown and Land Use: Exploring the Spatio-Temporal Evolution Patterns of Licensed Venues in Sydney, Australia between 2012 and 2021 in the Context of NSW Public Policy

The hospitality industry in Sydney, Australia, has been subject to several regulatory interventions in the last decade, including lockout laws, COVID-19 lockdowns and land use planning restrictions. This study has sought to explore the spatial implications of these policies in Inner Sydney between 2012 to 2021. Methods based in spatial analysis were applied to a database of over 40,000 licensed venues. Point pattern analysis and spatial autocorrelation methods were used to identify spatially significant venue clusters. Space-time cube and emerging-hot-spot methods were used to explore clusters over time. The results indicate that most venues are located in the Sydney CBD on business-zoned land and show a high degree of spatial clustering. Spatio-temporal analysis reveals this clustering to be consistent over time, with variations between venue types. Venue numbers declined following the introduction of the lockout laws, with numbers steadily recovering in the following years. There was no discernible change in the number of venues following the COVID-19 lockdowns; however, economic data suggest that there has been a decline in revenue. Some venues were identified as having temporarily ceased trading, with these clustered in the Sydney CBD. The findings of this study provide a data-driven approach to assist policymakers and industry bodies in better understanding the spatial implications of policies targeting the hospitality sector and will assist with recovery following the COVID-19 pandemic. Further research utilising similar methods could assess the impacts of further COVID-19 lockdowns as experienced in Sydney in 2021.

Analysis of remote sensing images by methods of convolutional neural networks and marked random point fields

The methodology of remote sensing image analysis for detection of dependences in the process of development of biological species is proposed. Classification methods based on convolutional networks are applied to a set of fragments of the input image. In order to increase the accuracy of classification by increasing the training and test samples, an original method of data augmentation is proposed. For a series of images of one part of the landscape, the fragments of images are classified by their numbers, which coincide with the numbers of the previously classified image of the training and test samples which are created manually. This approach has improved the accuracy of classification compared to known methods of data augmentation. Numerous studies of various convolutional neural networks have shown the similarity of the classification results of the remote sensing images fragments with increasing learning time with the complication of the network structure. A set of image fragment centers of a particular class is considered as random point configuration, the class labels are used as a mark for every point. Marked point field is considered as consisting of several sub-point fields in each of which all points have the same qualitative marks. We perform the analysis of the bivariate point pattern to reveal relationships between points of different types, using the characteristics of marked random point fields. Such relationships can characterize dependences and relative degrees of dominance. A series of remote sensing images are studied to identify the relationships between point configurations that describe different classes to monitor their development.

Spatial Point Pattern Analysis Identifies Mechanisms Shaping the Skin Parasite Landscape in Leishmania donovani Infection

Increasing evidence suggests that in hosts infected with parasites of the Leishmania donovani complex, transmission of infection to the sand fly vector is linked to parasite repositories in the host skin. However, a detailed understanding of the dispersal (the mechanism of spread) and dispersion (the observed state of spread) of these obligatory-intracellular parasites and their host phagocytes in the skin is lacking. Using endogenously fluorescent parasites as a proxy, we apply image analysis combined with spatial point pattern models borrowed from ecology to characterize dispersion of parasitized myeloid cells (including ManR+ and CD11c+ cells) and predict dispersal mechanisms in a previously described immunodeficient model of L. donovani infection. Our results suggest that after initial seeding of infection in the skin, heavily parasite-infected myeloid cells are found in patches that resemble innate granulomas. Spread of parasites from these initial patches subsequently occurs through infection of recruited myeloid cells, ultimately leading to self-propagating networks of patch clusters. This combination of imaging and ecological pattern analysis to identify mechanisms driving the skin parasite landscape offers new perspectives on myeloid cell behavior following parasitism by L. donovani and may also be applicable to elucidating the behavior of other intracellular tissue-resident pathogens and their host cells.

Discrete Protein Metric (DPM): A new image similarity metric to calculate accuracy of deep learning-generated cell focal adhesion predictions

Understanding cell behaviors can provide new knowledge on the development of different pathologies. Focal adhesion (FA) sites are important sub-cellular structures that are involved in these processes. To better facilitate the study of FA sites, deep learning (DL) can be used to predict FA site morphology based on limited datasets (e.g., cell membrane images). However, calculating the accuracy score of these predictions can be challenging due to the discrete/point pattern like nature of FA sites. In the present work, a new image similarity metric, discrete protein metric (DPM), was developed to calculate FA prediction accuracy. This metric measures differences in distribution (d), shape/size (s), and angle (a) of FA sites between the predicted image and its ground truth image. Performance of the DPM was evaluated by comparing it to three other commonly used image similarity metrics: Pearson correlation coefficient (PCC), feature similarity index (FSIM), and Intersection over Union (IoU). A sensitivity analysis was performed by comparing changes in each metric value due to quantifiable changes in FA site location, number, aspect ratio, area, or orientation. Furthermore, accuracy score of DL-generated predictions was calculated using all four metrics to compare their ability to capture variation across samples. Results showed better sensitivity and range of variation for DPM compared to the other metrics tested. Most importantly, DPM had the ability to determine which FA predictions were quantitatively more accurate and consistent with qualitative assessments. The proposed DPM hence provides a method to validate DL-generated FA predictions and can be extended to evaluating other predicted or segmented discrete structures of biomedical relevance.

Mapping the Intensity Function of a Non-stationary Point Process in Unobserved Areas

Seismic networks provide data that are used as basis both for public safety decisions and for scientific research. Their configuration affects the data completeness, which in turn, critically affects several seismological scientific targets (e.g., earthquake prediction, seismic hazard...). In this context, a key aspect is how to map earthquakes density in seismogenic areas from censored data or even in areas that are not covered by the network. We propose to predict the spatial distribution of earthquakes from the knowledge of presence locations and geological relationships, taking into account any interactions between records. Namely, in a more general setting, we aim to estimate the intensity function of a point process, conditional to its censored realization, as in geostatistics for continuous processes. We define a predictor as the best linear unbiased combination of the observed point pattern. We show that the weight function associated to the predictor is the solution of a Fredholm equation of second kind. Both the kernel and the source term of the Fredholm equation are related to the first- and second-order characteristics of the point process through the intensity and the pair correlation function. Results are presented and illustrated on simulated non-stationary point processes and real data for mapping Greek Hellenic seismicity in a region with unreliable and incomplete records.