African Institute for Mathematical Sciences South Africa

The emergence of exascale computing and artificial intelligence offer tremendous potential to significantly advance earth system prediction capabilities. However, enormous challenges must be overcome to adapt models and prediction systems to use these new technologies effectively. A recent WMO report on exascale computing recommends “ urgency in dedicating efforts and attention to disruptions associated with evolving computing technologies that will be increasingly difficult to overcome, threatening continued advancements in weather and climate prediction capabilities. Further, the explosive growth in data from observations, model and ensemble output, and post processing threatens to overwhelm the ability to deliver timely, accurate, and precise information needed for decision making. AI offers untapped opportunities to alter how models are developed, observations are processed, and predictions are analyzed and extracted for decision-making. Given the extraordinarily high cost of computing, growing complexity of prediction systems and increasingly unmanageable amount of data being produced and consumed, these challenges are rapidly becoming too large for any single institution or country to handle. This paper describes key technical, and budgetary challenges, identifies gaps and ways to address them, and makes a number of recommendations.

In nature, the two-dimensional (2D) profiles of fruits from many plants often resemble ellipses. However, it remains unclear whether these profiles strictly adhere to the ellipse equation, as many natural shapes resembling ellipses are actually better described as superellipses. The superellipse equation, which includes an additional parameter n compared to the ellipse equation, can generate a broader range of shapes, with the ellipse being just a special case of the superellipse. To investigate whether the 2D profiles of fruits are better described by ellipses or superellipses, we collected a total of 751 mature and undamaged fruits from 31 naturally growing plants of Cucumis melo L. var. agrestis Naud. Our analysis revealed that most adjusted root-mean-square errors (> 92% of the 751 fruits) for fitting the superellipse equation to the fruit profiles were consistently less than 0.0165. Furthermore, there were 638 of the 751 fruits (ca. 85%) with the 95% confidence intervals of the estimated parameter n in the super-ellipse equation not including 2. These findings suggest that the profiles of C. melo var. agrestis fruits align more closely with the superellipse equation than with the ellipse equation. This study provides evidence for the existence of the superellipse in fruit profiles, which has significant implications for studying fruit geometries and estimating fruit volumes using the solid of revolution formula. Furthermore, this discovery may contribute to a deeper understanding of the mechanisms driving the evolution of fruit shapes. ARTICLE HISTORY

Specialisation enhances the efficiency of plant–pollinator networks through the exchange of conspecific pollen transfer for floral resources. Floral resources form the currency of plant–pollinator interactions, but the understanding of how floral resources affect the structure of plant–pollinator networks remains modest. Previous theory predicts that optimally foraging animal species will specialise to improve resource acquisition under high resource availability. Although floral resource availability depends on both the plant production and animal consumption of the resources, previous work has assumed that production and availability are equivalent. This potentially may have led to erroneous inferences on the effect of resource availability on specialisation. We develop a mutualistic Lotka–Volterra consumer‐resource model to investigate the influence of floral resource availability on plant–pollinator network structure. The model incorporates animal adaptive foraging behaviour, floral resource dynamics, and density‐dependent dynamics. Specialisation, nestedness and modularity of simulated networks generated from the model under a wide range of parameters were explained using the generalised linear model. We found that the distinction between floral resource dynamics and plant density dynamics was necessary for partial specialisation of plant–pollinator networks. This is because floral resource dynamics constrained animal preference due to its depletion by animal species. Floral resource abundance had a positive effect on network specialisation, but animal density had a negative effect on network specialisation. Floral resource dynamics thus play key roles in the structure of plant–pollinator networks, distinctive from plant species density dynamics.

The heat kernel associated with a discrete graph Laplacian is the basic solution to the heat diffusion equation of a strict graph or network. In addition, this kernel represents the heat transfer that occurs over time across the network edges. Its computation involves exponentiating the Laplacian eigensystem with respect to time. In this paper, we expand upon this concept by considering a novel network-theoretic approach developed in recent years, which involves defining the k-path Laplacian operator for networks. Prior studies have adopted the notion of integrating long-range interactions (LRI) in the transmission of “information” across the nodes and edges of the network. Various methods have been employed to consider long-range interactions. We explore here the incorporation of long-range interactions in network analysis through the use of Mellin and Laplace transforms applied to the k-path Laplacian matrix. The contribution of this paper is the computation of the heat kernel associated with the k -path Laplacian, called the generalized heat kernel (GHK), and its employment as the basis for extracting stable and useful novel versions of invariants for graph characterization. The results presented in this paper demonstrate that the use of LRI improves the results obtained with classical diffusion methods for networks characterization.

Incidental capture of non-target species poses a pervasive threat to many marine species, with sometimes devastating consequences for both fisheries and conservation efforts. Because of the well-known importance of vocalizations in cetaceans, acoustic deterrents have been extensively used for these species. In contrast, acoustic communication for sea turtles has been considered negligible, and this question has been largely unexplored. Addressing this challenge therefore requires a comprehensive understanding of sea turtles’ responses to sensory signals. In this study, we scrutinized the avenue of auditory cues, specifically the natural sounds produced by green turtles ( Chelonia mydas) in Martinique, as a potential tool to reduce bycatch. We recorded 10 sounds produced by green turtles and identified those that appear to correspond to alerts, flight or social contact between individuals. Subsequently, these turtle sounds—as well synthetic and natural (earthquake) sounds—were presented to turtles in known foraging areas to assess the behavioral response of green turtles to these sounds. Our data highlighted that the playback of sounds produced by sea turtles was associated with alert or increased the vigilance of individuals. This therefore suggests novel opportunities for using sea turtle sounds to deter them from fishing gear or other potentially harmful areas, and highlights the potential of our research to improve sea turtles populations’ conservation.

The inequality in leaf and fruit size distribution per plant can be quantified using the Gini index, which is linked to the Lorenz curve depicting the cumulative proportion of leaf (or fruit) size against the cumulative proportion of the number of leaves (or fruits). Prior researches have predominantly employed empirical models-specifically the original performance equation (PE-1) and its generalized counterpart (GPE-1)-to fit rotated and right-shifted Lorenz curves. Notably, another potential performance equation (PE-2), capable of generating similar curves to PE-1, has been overlooked and not systematically compared with PE-1 and GPE-1. Furthermore, PE-2 has been extended into a generalized version (GPE-2). In the present study, we conducted a comparative analysis of these four performance equations, evaluating their applicability in describing Lorenz curves related to plant organ (leaf and fruit) size. Leaf area was measured on 240 culms of dwarf bamboo (Shibataea chinensis Nakai), and fruit volume was measured on 31 field muskmelon plants (Cucumis melo L. var. agrestis Naud.). Across both datasets, the root-mean-square errors of all four performance models were consistently smaller than 0.05. Paired t-tests indicated that GPE-1 exhibited the lowest root-mean-square error and Akaike information criterion value among the four performance equations. However, PE-2 gave the best close-to-linear behavior based on relative curvature measures. This study presents a valuable tool for assessing the inequality of plant organ size distribution. K E Y W O R D S Akaike information criterion, close-to-linear behavior, Gini index, goodness of fit, Lorenz curve, parameter-effects curvature

We analyze the impurity effects on the existence of localized structures, modulational instability (MI) and energy localization in a 1D quantum diatomic Klein–Gordon chain containing a mass defect. We have found that the impurities significantly affect the vibrational properties of crystals by modifying the distribution of normal mode frequencies. The MI investigation shows that, on the one hand, the mass defect around its critical value can dramatically influence the instability areas and the growth rate; on the other hand, the system exhibits an intrinsic asymmetry property that influences the localization and propagation of the impure mode. The conditions of stability of the impure mode are presented and a scattering phenomenon is observed when the impure mode is unstable. This phonons scattering phenomenon can be controlled either by the mass impurity �m1, by the coupling term c and by the pure mass m2. The accuracy of the analytical analysis has been performed by the numerical simulations and an excellent agreement has been observed. Furthermore, we have found through the energy localization that the impure mode can develop a local accumulation of energy known as self-trapping phenomenon.

Trait diversity, including trait turnover, that differentiates the roles of species and communities according to their functions, is a fundamental component of biodiversity. Accurately capturing trait diversity is crucial to better understand and predict community assembly, as well as the consequences of global change on community resilience. Existing methods to compute trait turnover have limitations. Trait space approaches based on minimum convex polygons only consider species with extreme trait values. Tree‐based approaches using dendrograms consider all species but distort trait distance between species. More recent trait space methods using complex polytopes try to harmonise the advantages of both methods, but their current implementation has mathematical flaws. We propose a new kernel integral method (KIM) to compute trait turnover, based on the integration of kernel density estimators (KDEs) rather than using polytopes. We explore how this approach and the computational aspects of the KDE computation can influence the estimates of trait turnover. The novel method is compared with existing ones using justified theoretical expectations for a large number of simulations in which the number of species and the distribution of their traits is controlled for. The practical application of KIM is then demonstrated using data on plant species introduced to the Pacific Islands of French Polynesia. Analyses on simulated data show that KIM generates results better aligned with theoretical expectations than other methods and is less sensitive to the total number of species. Analyses for French Polynesia data also show that different methods can lead to different conclusions about trait turnover and that the choice of method should be carefully considered based on the research question. The mathematical properties of methods for computing trait turnover are crucial to consider because they can have important effects on the results, and therefore lead to different conclusions. The novel KIM method provided here generates values that better reflect the distribution of species in trait space than other methods. We therefore recommend using KIM in studies on trait turnover. In contrast, tree‐based approaches should be kept for phylogenetic diversity, as phylogenetic trees will then reflect the speciation process.

We aim to explore what processes dominate community assembly of dragonflies (Odonata: Anisoptera) and damselflies (Odonata: Zygoptera) by differentiating the environmental and geographical drivers behind compositional turnover of narrow‐ranged versus widespread species. In this way, we further aim to describe patterns of species incidence and compositional turnover to expand upon the body of knowledge related to understanding biodiversity patterns and processes. We explored species turnover of dragonflies and damselflies separately, using zeta diversity to measure compositional turnover among multiple assemblages. Narrow‐ranged and widespread species within each suborder showed similar drivers. Specifically, both narrow‐ranged and widespread dragonflies show rapid turnover with small shifts in annual mean temperature, temperature seasonality and annual precipitation, whereas for damselflies, the major driver for turnover is distance between sites followed by climatic variables. Our results therefore show that odonate turnover is largely driven by climate, although the limited dispersal capabilities of damselflies also influences community assembly. Climate change could cause major changes in composition of odonates, presenting a challenge for conservation planning in Africa as species assemblages that were previously conserved may no longer be protected if their ranges shift outside protected areas. For damselflies, adaptation is a major concern, and with their limited dispersal capabilities and climate sensitivity, they may not be able to migrate effectively in response to changing climate conditions. The underlying assembly processes do not differ considerably for narrow‐ranged and widespread species within each suborder, suggesting that conservation planning tailored to each suborder may be sufficient in Africa.

Aim A pervasive negative relationship between the species richness of an assemblage and the mean global range size of the species it contains has recently been identified. Here, we test for an effect of habitat patch size on the mean landscape‐scale incidence (estimating local range size) of constituent species independent of variation in richness. Location Global. Time Period Contemporary. Major Taxa Studied Various. Methods We devised a new standardized patch‐scale metric, Mean Species Landscape‐scale Incidences per Patch (MSLIP). Positive values indicate a patch contains more widespread (high‐incidence) species than expected relative to their frequency in the landscape; negative values indicate the presence of more narrow‐range (low‐incidence) species. For 202 metacommunity datasets (archipelagos, habitat islands and fragments) we regressed MSLIP on patch size, testing four hypotheses: (i) no effect (zero slope), (ii) more widespread species in smaller patches (negative slope), (iii) more narrow‐range species in smaller patches (positive slope), (iv) a composite effect (curvilinear/unimodal). Results There was a generally negative curvilinear relationship between MSLIP and patch size, consistent with smaller and larger patches respectively containing more high‐ and low‐incidence species than expected. At intermediate patch sizes, mean species incidence approximates that of the overall landscape. Relative richness mediated this relationship; species‐rich sites of any size increasingly favoured low‐incidence species, species‐poor sites high‐incidence species. Results were consistent for taxonomic groups, and metacommunity types, except aquatic habitat islands. Main Conclusions The dependence of mean species incidence on both patch size and relative richness has implications for biodiversity and community assembly theory, or when seeking to understand ecological processes sensitive to species composition (e.g., ecosystem functioning, trophic web structures). We propose the ‘patch size where the MSLIP regression intersects zero’ as a benchmark distinguishing smaller from larger patches, with the expectation small patches preferentially support more high‐incidence and large patches more low‐incidence species.

The urgent need for effective wildlife monitoring solutions in the face of global biodiversity loss has resulted in the emergence of conservation technologies such as passive acoustic monitoring (PAM). While PAM has been extensively used for marine mammals, birds, and bats, its application to primates is limited. Black‐and‐white ruffed lemurs ( Varecia variegata ) are a promising species to test PAM with due to their distinctive and loud roar‐shrieks. Furthermore, these lemurs are challenging to monitor via traditional methods due to their fragmented and often unpredictable distribution in Madagascar's dense eastern rainforests. Our goal in this study was to develop a machine learning pipeline for automated call detection from PAM data, compare the effectiveness of PAM versus in‐person observations, and investigate diel patterns in lemur vocal behavior. We did this study at Mangevo, Ranomafana National Park by concurrently conducting focal follows and deploying autonomous recorders in May–July 2019. We used transfer learning to build a convolutional neural network (optimized for recall) that automated the detection of lemur calls (57‐h runtime; recall = 0.94, F1 = 0.70). We found that PAM outperformed in‐person observations, saving time, money, and labor while also providing re‐analyzable data. Using PAM yielded novel insights into V. variegata diel vocal patterns; we present the first published evidence of nocturnal calling. We developed a graphic user interface and open‐sourced data and code, to serve as a resource for primatologists interested in implementing PAM and machine learning. By leveraging the potential of this pipeline, we can address the urgent need for effective primate population surveys to inform conservation strategies.

The number and composition of species in a community can be quantified with α-diversity indices, including species richness ( R ), Simpson’s index ( D ), and the Shannon–Wiener index ( H΄ ). In forest communities, there are large variations in tree size among species and individuals of the same species, which result in differences in ecological processes and ecosystem functions. However, tree size inequality (TSI) has been largely neglected in studies using the available diversity indices. The TSI in the diameter at breast height (DBH) data for each of 999 20 m × 20 m forest census quadrats was quantified using the Gini index (GI), a measure of the inequality of size distribution. The generalized performance equation was used to describe the rotated and right-shifted Lorenz curve of the cumulative proportion of DBH and the cumulative proportion of number of trees per quadrat. We also examined the relationships of α-diversity indices with the GI using correlation tests. The generalized performance equation effectively described the rotated and right-shifted Lorenz curve of DBH distributions, with most root-mean-square errors (990 out of 999 quadrats) being < 0.0030. There were significant positive correlations between each of three α-diversity indices (i.e., R , D , and H' ) and the GI. Nevertheless, the total abundance of trees in each quadrat did not significantly influence the GI. This means that the TSI increased with increasing species diversity. Thus, two new indices are proposed that can balance α-diversity against the extent of TSI in the community: (1 − GI) × D , and (1 − GI) × H' . These new indices were significantly correlated with the original D and H΄ , and did not increase the extent of variation within each group of indices. This study presents a useful tool for quantifying both species diversity and the variation in tree sizes in forest communities, especially in the face of cumulative species loss under global climate change.

In nature, the fruit shapes of many plants resemble avian eggs, a form extensively studied as solids of revolution. Despite this, the hypothesis that egg-shaped fruits are themselves solids of revolution remains unvalidated. To address this, 751 Cucumis melo L. var. agrestis Naud. fruits were photographed, and the two-dimensional (2D) boundary coordinates of each fruit profile were digitized. Then, the explicit Preston equation (EPE), a universal egg-shape model, was used to fit the 2D boundary coordinates to obtain the estimates of the EPE’s parameters of each fruit. Under the hypothesis that egg-shaped fruits are solids of revolution, the fruit volumes were estimated using the solid of revolution formula based on the estimated EPE’s parameters. To test whether the fruits are solids of revolution, the fruit volumes were measured by using a graduated cylinder and compared with the estimated volumes using the solid of revolution formula. The EPE was demonstrated to be valid in describing the 2D profiles of C. melo var. agrestis fruits. There was a significant correlation between the measured fruit volumes using the graduated cylinder and the estimated fruit volumes using the solid of revolution formula based on the estimated EPE’s parameters. Acknowledging potential measurement errors, particularly fruit fuzz causing air bubbles during volume measurements, we recognize slight deviations between measured volumes and estimated values. Despite this, our findings strongly suggest that C. melo var. agrestis fruits are solids of revolution. This study contributes insights into the evolutionary aspects of fruit geometries in plants with egg-shaped fruits and introduces a practical tool for non-destructively calculating fruit volume and surface area based on photographed 2D fruit profiles.

The advent of 3D Magnetic Resonance Imaging (MRI) has revolutionized medical imaging and diagnostic capabilities, allowing for more precise diagnosis, treatment planning, and improved patient outcomes. 3D MRI imaging enables the creation of detailed 3D reconstructions of anatomical structures that can be used for visualization, analysis, and surgical planning. However, these reconstructions often require many scan acquisitions, demanding a long session to use the machine and requiring the patient to remain still, with consequent possible motion artifacts. The development of neural radiance fields (NeRF) technology has shown promising results in generating highly accurate 3D reconstructions of MRI images with less user input. Our approach is based on neural radiance fields to reconstruct 3D projections from 2D slices of MRI scans. We do this by using 3D convolutional neural networks to address challenges posed by variable slice thickness; incorporating multiple MRI modalities to ensure robustness and extracting the shape and volumetric depth of both surface and internal anatomical structures with slice interpolation. This approach provides more comprehensive and robust 3D reconstructions of both surface and internal anatomical structures and has significant potential for clinical applications, allowing medical professionals to better visualize and analyze anatomical structures with less available data, potentially reducing times and motion-related issues.

Thin spray-on liners (TSLs) have been found to be effective for structurally supporting the walls of mining tunnels and thus reducing the occurrence of rock bursts, an effect primarily due to the penetration of cracks by the liner. Surface tension effects are thus important. However, TSLs are also used to simply stabilize rock surfaces, for example, to prevent rock fall, and in this context crack penetration is desirable but not necessary, and the tensile and shearing strength and adhesive properties of the liner determine its effectiveness. We examine the effectiveness of nonpenetrating TSLs in a global lined tunnel and in a local rock support context. In the tunnel context, we examine the effect of the liner on the stress distribution in a tunnel subjected to a geological or mining event. We show that the liner has little effect on stresses in the surrounding rock and that tensile stresses in the rock surface are transmitted across the liner, so that failure is likely to be due to liner rupture or detachment from the surface. In the local rock support context, loose rock movements are shown to be better achieved using a liner with small Young’s modulus, but high rupture strength.

While there has been great interest in species characteristics that promote invasiveness, still little is known about the characteristics that distinguish invasive from non‐invasive insects. Using a database on the naturalised distributions of alien insects and expert opinions about their impacts, we identified the world's 100 worst invasive insect species. By comparing species characteristics reported in the literature using a meta‐analysis, between the 100 worst invasive species and related non‐invasive species, we found that invasive insects overall have more pathways of introduction, occur in more habitats, have higher fecundities, higher voltinism, more genes, shorted lifespans and faster development from egg to adult. Some of the differences in species characteristics related to propagule pressures, life‐histories and biotic interactions, conditional on whether the non‐invasive species compared is known to be naturalised somewhere, whether the invasive species is globally distributed, and the climatic region of the species. Synthesis and applications . We show for the first time, using a multi‐species comparative approach, that invasive insects differ in several characteristics from related non‐invasive insects. Our results show that invasive species, such as Spodoptera frugiperda , typically are habitat generalists with a high fecundity, a short lifespan and fast development, whereas the importance of female body size and number of enemies are context dependent. Our study can guide and improve existing screening tools for assessing the invasion potential of alien insects.

The postharvest processes of groundnuts often become sources of microbial contamination leading to infections and intoxication. Hence, this study examined the microbial pathogens contaminating groundnuts after harvesting. About 50 samples were randomly collected from four major groundnut-producing towns: Bolgatanga, Chiana, Navrongo, and Bongo, all in the Upper East Region of Northern Ghana, and microbiologically examined using Analytical Profile Index (API® 20E). The results revealed that samples from Bolgatanga were the most contaminated, while Chiana has the least contaminated samples. Several species of bacterial genera such as Staphylococcus, Proteus, Escherichia, Bacillus, and Micrococcus, and fungal genera including Aspergillus, Fusarium, Rhizopus, Mucor, Saccharomyces, and Eurotium were isolated as the main microbial pathogens contaminating the produce. Navrongo and Bolgatanga recorded the highest rate of bacterial species for unshelled (29.5%) and shelled (30.4%) groundnuts, respectively, while Bongo and Bolgatanga registered the highest rate of fungal species under unshelled (32.8%) and shelled (32.6%) groundnuts, respectively. Due to the high levels of microbial contamination of most of the samples and the kind of microbial species involved, proper hygiene standards must be adopted during the postharvest handling of the shelled and unshelled groundnuts.