Biotic indices are used by environmental managers to assess the biological condition of benthic habitats in response to environmental stressors. Successful application depends on availability of reference data characterizing minimally degraded conditions, but long-term monitoring data needed to define such references are often lacking. Mollusks are good surrogates for benthic communities, and their geohistorical record—dead shell remains preserved beneath the seafloor and time-averaged over decades to centuries—can help fill these data gaps. Here, we develop an initial framework for a mollusk-only biotic index applicable to the geohistorical record, using benthic survey data of living mollusks from the mid-Atlantic coastal region of the United States. Machine learning was used to evaluate biotic metrics—taxonomic diversity, pollution tolerance/sensitivity, and biological traits—for their ability to distinguish sites classified a priori as Reference or Degraded. The final model—geoAMBI—is a modified logistic regression version of the AZTI-Marine Biotic Index (AMBI), calculated using reversed rank-order abundance, which is operationally robust to taphonomic biases in the geohistorical record. Before geoAMBI can be applied to the geohistorical record in the mid-Atlantic region or adapted for use elsewhere, the most immediate need is to assess how time averaging and other preservational biases affect its performance.

Pteropods are marine planktonic snails that are used as bioindicators of ocean acidification due to their thin, aragonitic shells, and ubiquity throughout the world’s oceans; their responses include decreased size, reduced shell thickness, and increased shell dissolution. Shell dissolution has been measured with a variety of metrics involving light microscopy, scanning electron microscopy (SEM), and computed tomography (CT). While CT and SEM metrics offer high resolution imaging, these analyses are cost- and time-intensive relative to light microscopy analysis. This research compares light microscopy, CT, and SEM shell dissolution metrics across three pteropod species: Limacina helicina, Limacina retroversa, and Heliconoides inflatus. Sourced from multiple localities, these specimens lived in tropical to subpolar environments and were exposed to varying aragonite saturations states due to oceanographic differences in these environments. Specimens were evaluated with light microscopy for the Limacina Dissolution Index (LDX), with SEM for percent of pristine shell coverage and maximum dissolution type, and with CT for whole-shell thickness. LDX and the percentage of pristine shell determined via SEM were highly correlated in all three species’ datasets. For L. retroversa, LDX was also significantly correlated to SEM maximum dissolution type. Although the genera Heliconoides and Limacina have different shell microstructures, the relationship between LDX and SEM dissolution did not vary by species. The CT metric for shell thickness was not significantly correlated to any other dissolution metrics for any species. However, severely dissolved areas apparent in SEM were visually discernible in CT thickness heatmaps. While CT may not detect minor shell dissolution, previous studies have used CT to detect reduced calcification in response to ocean acidification. SEM is ideal for detecting the onset of dissolution, but SEMing large numbers of specimens may not be practical due to monetary and time constraints. LDX, on the other hand, is a fast and cost-effective metric that is strongly correlated with SEM metrics, regardless of the oceanographic conditions that those species experienced. These results suggest that an efficient ocean acidification monitoring strategy is to evaluate all pteropod specimens via LDX and to then SEM a subset of those specimens.

The inconclusive category in forensics reporting is the appropriate response in many cases, but it poses challenges in estimating an "error rate". We discuss the use of a class of information-theoretic measures related to cross entropy as an alternative set of metrics that allows for performance evaluation of results presented using multi-category reporting scales. This paper shows how this class of performance metrics, and in particular the log likelihood ratio cost, which is already in use with likelihood ratio forensic reporting methods and in machine learning communities, can be readily adapted for use with the widely used multiple category conclusions scales. Bayesian credible intervals on these metrics can be estimated using numerical methods. The application of these metrics to published test results is shown. It is demonstrated, using these test results, that reducing the number of categories used in a proficiency test from five or six to three increases the cross entropy, indicating that the higher number of categories was justified, as it they increased the level of agreement with ground truth.

It is widely known that leverage reduces the cost of capital of a firm in a perfect capital market save for corporate income taxes. This result, however, rests on a cost of capital measure which is not fit for this purpose. When the correct measure is used, we find that leverage reduces the after-tax cost of capital only if the interest tax shields reduce the risk of the firm. This means that the after-tax cost of capital of a firm is independent of leverage if the firm adopts a leverage policy based on a target debt-value leverage ratio.

A lack of location-specific, long-term data is a common obstacle to assessing trends in condition of coastal habitats over time. Without historical monitoring records or other documentation, filling such data gaps can be difficult, but sedimentary records such as death assemblages (DAs; the accumulated, identifiable remains of organisms that lived in or near the habitat in the past) are relatively untapped, location-specific archives of ecological information from the past. In 2018, the Florida Department of Environmental Protection and the Paleontological Research Institution began a collaboration to study the use of oyster reef (Crassostrea virginica) DAs to address monitoring information gaps for oyster size. To-date, our project has sampled DAs from over 30 intertidal oyster reefs around Florida, radiocarbon dated most of the samples, and measured over 26,000 oyster shells. In the process, we found that C. virginica DAs are recent and high-resolution archives, with most samples from 15-35cm burial depth dating to within the last 80 years. We also developed a model to combine the DA data with real-time monitoring data on live oyster sizes from the same reefs to estimate reef- and locality-level size trends from as early as the 1960s to the present. This information is adding temporal context for our overwhelmingly short (~5-10 years) and recent (many post-2010) time series of live C. virginica size data. This case study demonstrates the potential utility of DA data for supplementing real-time monitoring data during the assessment and management of coastal habitats.

Using paleoecological data to inform resource management decisions is challenging without an understanding of the ages and degrees of time-averaging in molluscan death assemblage (DA) samples. We illustrate this challenge by documenting the spatial and stratigraphic variability in age and time-averaging of oyster reef DAs. By radiocarbon dating a total of 630 oyster shells from samples at two burial depths on 31 oyster reefs around Florida, southeastern United States, we found that (1) spatial and stratigraphic variability in DA sample ages and time-averaging is of similar magnitude, and (2) the shallow oyster reef DAs are among the youngest and highest-resolution molluscan DAs documented to date, with most having decadal-scale time-averaging estimates, and sometimes less. This information increases the potential utility of the DAs for habitat management because DA data can be placed in a more specific temporal context relative to real-time monitoring data. More broadly, the results highlight the potential to obtain decadal-scale resolution from oyster bioherms in the fossil record.

Incorporating paleontological data into the methods and formats familiar to conservation practitioners may facilitate greater use of paleontological data in conservation practice. Benthic indices (e.g., Multivariate-AZTI Marine Biotic Index; M-AMBI) utilize reference conditions for monitoring ecological conditions. However, reference conditions from monitoring records are limited in temporal scope and often represent degraded conditions, which can cause inaccurate assessments of ecological quality. Paleontological data, such as molluscan death assemblages, have potential to provide long-term, location-specific reference conditions, which are otherwise inaccessible to decision-makers. Here we use simulations of living communities under constant and changing environmental conditions to evaluate the capacity of death assemblage reference conditions to replicate M-AMBI values when used in place of reference conditions from the living communities. Reference conditions from all death assemblage scenarios successfully replicated correct remediation decisions in most simulation runs with environmental change and stability. Variations in M-AMBI values were due to overestimated species richness and Shannon entropy values in the death assemblages and effects of changes to these parameters varied across scenarios. Time averaging was largely beneficial, particularly when environmental change occurred, and short-term observations of the living communities produced incorrect remediation decisions. When the duration of time averaging is known, death assemblages can provide valuable longer-term perspectives with the potential to outperform temporally constrained baseline information from monitoring the living community.

ABSTRACT Incorporating paleontological data into the methods and formats already familiar to conservation practitioners may facilitate greater use of paleontological data in conservation practice. Benthic indices (e.g., Multivariate - AZTI Marine Biotic Index; M-AMBI) already incorporate reference conditions and are a good candidate for integration. In simulations of living communities under constant and changing environmental conditions, we evaluate the capacity of death assemblage reference conditions to replicate M-AMBI values when used in place of reference conditions from the final ten generations of the simulation or all five hundred simulated generations. Reference conditions from all death assemblage scenarios successfully replicated correct remediation decisions in the majority of simulation runs with environmental change and stability. Variations in M-AMBI values were due to overestimated richness and diversity in the death assemblages but effects of changes to these parameters varied across scenarios, emphasizing the importance of evaluating multiple metrics. Time averaging was largely beneficial, particularly when environmental change occurred and short-term ecological observations (ten generations) produced incorrect remediation decisions. When the duration of time averaging is known, death assemblages can provide valuable long-term perspectives with the potential to outperform temporally constrained baseline information from monitoring the living community. Supplementary material All R code used to produce the simulation, analyze outputs, and create figures is available at: https://doi.org/10.5281/zenodo.6355921 . The simulated data is also available at this location. Supplementary figures and analyses referred to in the text are available at the end of this document.

AMBI and M-AMBI are widely used biotic indices for assessing the ecological quality status of benthic macroinvertebrate communities in estuarine and coastal soft-bottom habitats. Identifying the species needed for estimating these indices, however, is both expensive and time-consuming, and requires a high degree of taxonomic expertise. The use of proxy taxa as a means of subsampling the target community may save time, resources, and the breadth of taxonomic expertise needed. Our study used macroinvertebrate benthic survey data from the Atlantic Coast of the United States to test the fidelity of molluscs as proxies of the whole community. We calculated the AMBI and M-AMBI scores for both the molluscan and whole communities and then adjusted the molluscan-only index scores to that of the whole community using the linear relationship between the two communities within a Bayesian framework. We found that the mollusc-only AMBI approach underperformed at classifying the ecological quality of the whole community, particularly regarding sample sites classified as needing remediation. The low performance of the mollusc-only AMBI approach is likely due to the dearth of molluscs with high environmental stress tolerances. In contrast, the mollusc-only M-AMBI outperformed AMBI at classifying ecological quality. The M-AMBI linear model correctly classified nearly all of the adjusted mollusc-only sample sites needing remediation. The increased efficacy of mollusc-only M-AMBI may be due to the incorporation of species richness and diversity into the index, as both metrics were highly correlated between the molluscan and whole communities. Mollusc-only M-AMBI did have some drawbacks, however, with fidelity decreasing as ecological quality decreased. Overall, our study highlights the potential utility of a mollusc-only approach for assessing the ecological quality of estuarine and coastal soft-bottom habitats.

The effects of overdispersion and zero inflation (e.g., poor model fits) can result in misinterpretation in studies using count data. These effects have not been evaluated in paleoecological studies of predation and are further complicated by preservational bias and time averaging. We develop a hierarchical Bayesian framework to account for uncertainty from overdispersion and zero inflation in estimates of specimen and predation trace counts. We demonstrate its application using published data on drilling predators and their prey in time-averaged death assemblages from the Great Barrier Reef, Australia. Our results indicate that estimates of predation frequencies are underestimated when zero inflation is not considered, and this effect is likely compounded by removal of individuals and predation traces via preservational bias. Time averaging likely reduces zero inflation via accumulation of rare taxa and events; however, it increases the uncertainty in comparisons between assemblages by introducing variability in sampling effort. That is, there is an analytical cost with time-averaged count data, manifesting as broader confidence regions. Ecological inferences in paleoecology can be strengthened by accounting for the uncertainty inherent to paleoecological count data and the sampling processes by which they are generated.

Clinical concept extraction often begins with clinical Named Entity Recognition (NER). Often trained on annotated clinical notes, clinical NER models tend to struggle with tagging clinical entities in user queries because of the structural differences between clinical notes and user queries. User queries, unlike clinical notes, are often ungrammatical and incoherent. In many cases, user queries are compounded of multiple clinical entities, without comma or conjunction words separating them. By using as dataset a mixture of annotated clinical notes and synthesized user queries, we adapt a clinical NER model based on the BiLSTM-CRF architecture for tagging clinical entities in user queries. Our contribution are the following: 1) We found that when trained on a mixture of synthesized user queries and clinical notes, the NER model performs better on both user queries and clinical notes. 2) We provide an end-to-end and easy-to-implement framework for clinical concept extraction from user queries.

We show the impact on transit network connectivity of a major network redesign using a comprehensive connectivity measure. This measure captures the quality of service impact – the difference between the actual accessibility and the designed accessibility as a function of space and time. The former is influenced by on-time performance and ability to make transfers. The measure is unique in that it incorporates both spatial and temporal aspects of the transportation network, so that the effects of the network geometry, transportation services schedules, and operational performance (in the form of service reliability / schedule adherence) are all appropriately reflected in this unified measure. We employ a spatial statistical model to tie aggregate ridership to average connectivity across transportation analysis zones. The statistically significant relationship indicates that, at least in aggregate, connectivity and demand are linked. Using spatio-temporal clustering, we show where, when, and how accessibility is impacted by a significant network redesign.

In the paleoecological literature, drilling frequency the percent of specimens in a prey taxon with complete drill holes is commonly interpreted as an indication of a predator's preference. Such taxon-specific drilling frequencies are often compared with one another and related to underlying prey characteristics such as cost-benefit ratio or body size. Although this approach can demonstrate a predator's relative preference for one prey type over another, it fails to consider whether predation on any prey type is greater than would be expected by a predator without preference. Here we develop a null model for evaluating predator preference in paleoecology by considering drilling frequency in the framework of Manly's alpha, which is a well-established model in the ecological literature that has been used to evaluate predator preferences. In effect, Manly's alpha normalizes taxon-specific drilling frequencies with all potential prey types in the community, allowing for a statistically rigorous test of the null hypothesis that drilling predation, and therefore predator preference, on any given prey type is equivalent to all other available prey types in the community. After discussing the statistical basis for the model, we demonstrate the model's utility by applying it to a published dataset of drilling predation on Pliocene bivalves from Langenboom, Netherlands. The Manly's alpha approach, which uses the same data (i.e., drilling frequencies) that are commonly collected by paleoecologists, provides a null hypothesis to more rigorously assess predator preferences and inherently includes community context for predator-prey interactions in the fossil record.

Managed print service (MPS) is a type of information technology infrastructure service that provides centralized management of companies’ printing device fleets. In this paper, we estimate the provider’s risk preference in MPS using a proprietary data set from Xerox Corporation. We adopt a structural approach in our empirical analysis by modeling the contracting and usage processes of MPS as a two-stage screening game and building econometric models based on the equilibrium contracts and print volumes. Our econometric models have a unique hierarchical structure that allows clustering of printers with the same contracts in the same company, thereby capturing the B2B nature of MPS. We find that Xerox exhibits risk aversion in MPS contracting and provide institutional details of Xerox’s commission holdback policy that may cause the observed risk aversion. In the counterfactual analysis, we demonstrate the significance of the provider’s risk aversion and the implications of the commission holdback policy on equilibrium contracts, the expected earnings of Xerox and customer companies, and their preferences for printer models. The E-companion is available at https://doi.org/10.1287/opre.2017.1673 .

Public transport operations data and in particular fare collection data can be used to reconstruct and analyse mobility patterns. So far, various methods have been proposed and studied in some specific contexts. This paper proposes a general framework for looking at all the core elements of mobility in the various possible operational settings of public transport. It also describes some novel methods for trip alignments, travels’ origin and destination detection and vehicle load estimation. Two use cases illustrate the use of the framework and validate the efficiency of the proposed reconstruction methods.

Global warming, acidification, and oxygen stress at the Paleocene-Eocene Thermal Maximum (PETM) are associated with severe extinction in the deep sea and major biogeographic and ecologic changes in planktonic and terrestrial ecosystems, yet impacts on shallow marine macrofaunas are obscured by the incompleteness of shelf sections. We analyze mollusk assemblages bracketing (but not including) the PETM and find few notable lasting impacts on diversity, turnover, functional ecology, body size, or life history of important clades. Infaunal and chemosymbiotic taxa become more common, and body size and abundance drop in one clade, consistent with hypoxia-driven selection, but within-clade changes are not generalizable across taxa. While an unrecorded transient response is still possible, the long-term evolutionary impact is minimal. Adaptation to already-warm conditions and slow release of CO2 relative to the time scale of ocean mixing likely buffered the impact of PETM climate change on shelf faunas.

In this paper, we aim to quantify uncertainty in short-term traffic volume prediction by enhancing a hybrid machine learning model based on Particle Swarm Optimization (PSO) and Extreme Learning Machine (ELM) neural network. Different from the previous studies, the PSO-ELM models require no statistical inference nor distribution assumption of the model parameters, but rather focus on generating the prediction intervals (PIs) that can minimize a multi-objective function which considers two criteria, reliability and interval sharpness. The improved PSO-ELM models are developed for an hourly border crossing traffic dataset and compared to: (1) the original PSO-ELMs; (2) two state of the art models proposed by Zhang et al. (2014) and Guo et al. (2014) separately; and (3) the traditional ARMA and Kalman filter models. The results show that the improved PSO-ELM can always keep the mean PI length the lowest, and guarantee that the PI coverage probability is higher than the corresponding PI nominal confidence, regardless of the confidence level assumed. The study also probes the reasons that led to a few points being not covered by the Pls of PSO-ELMS. Finally, the study proposes a comprehensive optimization framework to make staffing plans for border crossing authority based on bounds of Pb and point predictions. The results show that for holidays, the staffing plans based on PI upper bounds generated much lower total system costs, and that those plans derived from PI upper bounds of the improved PSO-ELM models, are capable of producing the lowest average waiting times at the border. For a weekday or a typical Monday, the workforce plans based on point predictions from Zhang et al. (2014) and Guo et al. (2014) models generated the smallest system costs with low border crossing delays. Moreover, for both holiday and normal Monday scenarios, if the border crossing authority lacked the required staff to implement the plans based on PI upper bounds or point predictions, the staffing plans based on PI lower bounds from the improved PSO-ELMS performed the best, with an acceptable level of service and total system costs close to the point prediction plans.