Simon J. Bonner
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Simon J. Bonner has a total of 31 co-authors. Here are the top co-authors by number of shared publications:
- Alexandru Draghici
Department of Statistical and Actuarial Sciences, University of Western Ontario, London, Canada
3 shared publications Douglas G. Woolford
Department of Statistical and Actuarial Sciences University of Western Ontario London Ontario Canada
2 shared publications- Johanna de Haan‐Ward
Department of Statistical and Actuarial Sciences University of Western Ontario London Ontario Canada
2 shared publications - Laura L. E. Cowen
Department of Mathematics and Statistics University of Victoria Victoria British Columbia Canada
2 shared publications - Wei Zhang
Department of Statistics, University of Auckland, Private Bag 92019, Auckland, New Zealand
2 shared publications - Andreu Rotger
Animal Demography and Ecology Unit, Instituto Mediterráneo de Estudios Avanzados (Consejo Superior de Investigaciones Científicas, Universitat de les Illes Balears)
1 shared publication
Latest Publications
- Open Access
Testing hypotheses about correlations between brain activation patterns
Many functional magnetic resonance imaging (fMRI) studies conclude that two conditions engage “overlapping, yet partly distinct” patterns of activation. Yet, there is currently no commonly accepted method for determining the extent of this overlap. While correlations between activation patterns can serve as a measure of their correspondence, empirical correlations are strongly biased towards zero due to measurement noise, preventing their use in testing hypotheses about the actual degree of pattern correspondence. In this paper, we derive the maximum-likelihood estimate for the correlation of the true (noise-less) activation patterns and examine its behavior in the low signal-to-noise regime that is typical for fMRI studies. We show that although the maximum-likelihood estimate corrects for much of the influence of measurement noise, it is ultimately biased. We examine different ways of drawing inferences about the size of the underlying true correlations. We find that a subject-wise bootstrap on the maximum-likelihood group estimate performs best over the tested conditions. We extend the proposed method to test more general hypotheses about the representational geometry of activation patterns for more conditions, and highlight best practices, as well as common pitfalls and problems, in testing such hypotheses.
Open AccessTesting hypotheses about correlations between brain activation patterns
Many functional magnetic resonance imaging (fMRI) studies conclude that two conditions engage “overlapping, yet partly distinct” patterns of activation. Yet, there is currently no commonly accepted method for determining the extent of this overlap. While correlations between activation patterns can serve as a measure of their correspondence, empirical correlations are strongly biased towards zero due to measurement noise, preventing their use in testing hypotheses about the actual degree of pattern correspondence. In this paper, we derive the maximum-likelihood estimate for the correlation of the true (noise-less) activation patterns and examine its behavior in the low signal-to-noise regime that is typical for fMRI studies. We show that although the maximum-likelihood estimate corrects for much of the influence of measurement noise, it is ultimately biased. We examine different ways of drawing inferences about the size of the underlying true correlations. We find that a subject-wise bootstrap on the maximum-likelihood group estimate performs best over the tested conditions. We extend the proposed method to test more general hypotheses about the representational geometry of activation patterns for more conditions, and highlight best practices, as well as common pitfalls and problems, in testing such hypotheses.
- 11 July 2025
- Published by Springer Nature in Journal of Agricultural, Biological and Environmental Statistics
- Vol. 31 (4) , 970-991
Modelling Dependent Fates of Long-Term Mates in Mark-Recapture Studies Using Conditional Data Methods
- 11 July 2025
- Published by Springer Nature in Journal of Agricultural, Biological and Environmental Statistics
- Vol. 31 (4) , 970-991
Modelling Dependent Fates of Long-Term Mates in Mark-Recapture Studies Using Conditional Data Methods
- Open Access
- 9 July 2025
- Published by Wiley in Environmetrics
- Vol. 36 (5)
Occupancy Modeling for Rare Species Using Large Datasets: A Subsampling Approach
Citizen science monitoring programs, such as the Breeding Bird Survey, provide a wealth of data for understanding species abundance and distribution. However, traditional approaches for occupancy modeling of rare species can be difficult to apply to large, imbalanced datasets. We propose a new method for occupancy modeling where the original dataset is subsampled seasonally, keeping all sites with at least one detection along with a random sample of sites with no detections. Occupancy models cannot be fit directly to these subsampled data because the assumption of binomial sampling no longer holds. However, we show that the occupancy probability is adjusted by an offset, meaning inference on the effects of predictors is still valid. We propose a method for model fitting via direct maximum likelihood and demonstrate via simulation that this leads to computational gains. We illustrate our method using data on Canada Warblers (Cardellina canadensis) from the Breeding Bird Survey in Ontario, Canada from 1997 to 2018, where 95% of sites have zero detections annually, demonstrating that we can accurately estimate the occupancy and detection parameters, including estimating the effects of habitat covariates, using just 10% of the original dataset.
- 9 July 2025
- Published by Wiley in Environmetrics
- Vol. 36 (5)
Open AccessOccupancy Modeling for Rare Species Using Large Datasets: A Subsampling Approach
Citizen science monitoring programs, such as the Breeding Bird Survey, provide a wealth of data for understanding species abundance and distribution. However, traditional approaches for occupancy modeling of rare species can be difficult to apply to large, imbalanced datasets. We propose a new method for occupancy modeling where the original dataset is subsampled seasonally, keeping all sites with at least one detection along with a random sample of sites with no detections. Occupancy models cannot be fit directly to these subsampled data because the assumption of binomial sampling no longer holds. However, we show that the occupancy probability is adjusted by an offset, meaning inference on the effects of predictors is still valid. We propose a method for model fitting via direct maximum likelihood and demonstrate via simulation that this leads to computational gains. We illustrate our method using data on Canada Warblers (Cardellina canadensis) from the Breeding Bird Survey in Ontario, Canada from 1997 to 2018, where 95% of sites have zero detections annually, demonstrating that we can accurately estimate the occupancy and detection parameters, including estimating the effects of habitat covariates, using just 10% of the original dataset.
- Open Access
- 22 April 2025
- Published by Wiley in Canadian Journal of Statistics
- Vol. 53 (3)
Predicting rare events using training data from stratified sampling designs, with application to human‐caused wildfire prediction
Response‐based sampling is often used in modelling rare events from large, imbalanced data for efficiency. When modelling the event with logistic regression, the sampling design may be adjusted for using sampling weights or an offset. We propose a stratified sampling design for modelling rare events with large data which improves on previous methods by providing unbiased estimates of the standard errors of the coefficients in a multiple logistic regression scenario. We use multiple intercepts to model the incidence in the sampled data, then adjust each intercept via a stratum‐specific offset. Our simulations provide no evidence of bias in the estimated logistic regression coefficients or their standard errors. We apply this method to spatio‐temporal, fine‐scale human‐caused fire occurrence modelling for a region in northwestern Ontario, Canada, illustrating how the stratified sampling approach results in more locally precise estimates of fire occurrence.
- 22 April 2025
- Published by Wiley in Canadian Journal of Statistics
- Vol. 53 (3)
Open AccessPredicting rare events using training data from stratified sampling designs, with application to human‐caused wildfire prediction
Response‐based sampling is often used in modelling rare events from large, imbalanced data for efficiency. When modelling the event with logistic regression, the sampling design may be adjusted for using sampling weights or an offset. We propose a stratified sampling design for modelling rare events with large data which improves on previous methods by providing unbiased estimates of the standard errors of the coefficients in a multiple logistic regression scenario. We use multiple intercepts to model the incidence in the sampled data, then adjust each intercept via a stratum‐specific offset. Our simulations provide no evidence of bias in the estimated logistic regression coefficients or their standard errors. We apply this method to spatio‐temporal, fine‐scale human‐caused fire occurrence modelling for a region in northwestern Ontario, Canada, illustrating how the stratified sampling approach results in more locally precise estimates of fire occurrence.
- Open Access
- 1 July 2024
- Published by Wiley in Ecological Solutions and Evidence
- Vol. 5 (3)
Response to ‘Improving abundance estimation by combining capture–recapture and presence–absence data: Example with a large carnivore’
A key area of research in ecological statistics involves combining data sources from multiple streams to improve population estimates. One such model attempts to integrate capture–recapture and presence–absence data to estimate the population size of Eurasian lynx in the Jura Mountains, eastern France. This model has been observed to underestimate population sizes. We conducted an extensive simulation study to evaluate the model's performance. We describe our methods for generation of simultaneous capture–recapture and presence–absence data and demonstrate that the model is flawed. Finally, we give an outline of our hypothesis on why the model underestimates population sizes.
- 1 July 2024
- Published by Wiley in Ecological Solutions and Evidence
- Vol. 5 (3)
Open AccessResponse to ‘Improving abundance estimation by combining capture–recapture and presence–absence data: Example with a large carnivore’
A key area of research in ecological statistics involves combining data sources from multiple streams to improve population estimates. One such model attempts to integrate capture–recapture and presence–absence data to estimate the population size of Eurasian lynx in the Jura Mountains, eastern France. This model has been observed to underestimate population sizes. We conducted an extensive simulation study to evaluate the model's performance. We describe our methods for generation of simultaneous capture–recapture and presence–absence data and demonstrate that the model is flawed. Finally, we give an outline of our hypothesis on why the model underestimates population sizes.
- Open Access
- 26 October 2023
- Published by Wiley in Environmetrics
- Vol. 35 (1)
On the identifiability of the trinomial model for mark‐recapture‐recovery studies
Continuous predictors of survival present a challenge in the analysis of data from studies of marked individuals if they vary over time and can only be observed when individuals are captured. Existing methods to study the effects of such variables have followed one of two approaches. The first is to model the joint distribution of the predictor and the observed capture histories, and the second is to draw inference from the likelihood conditional on events that depend only on observed predictor values, called the trinomial model. Previous comparison of these approaches found that joint modelling provided more precise inference about the effect of the covariate while the trinomial model was less prone to issues of model mis‐specification. However, we believe that an important issue was missed. We show through mathematical analysis and numerical simulation that the trinomial model is not identifiable when the predictor has no effect on the survival probability. This also causes inferences from the trinomial model to be imprecise when the effect of the covariate on the survival probability is small. We also analyse data on the effect of body mass on the survival of meadow voles to demonstrate the importance of this issue in real applications.
- 26 October 2023
- Published by Wiley in Environmetrics
- Vol. 35 (1)
Open AccessOn the identifiability of the trinomial model for mark‐recapture‐recovery studies
Continuous predictors of survival present a challenge in the analysis of data from studies of marked individuals if they vary over time and can only be observed when individuals are captured. Existing methods to study the effects of such variables have followed one of two approaches. The first is to model the joint distribution of the predictor and the observed capture histories, and the second is to draw inference from the likelihood conditional on events that depend only on observed predictor values, called the trinomial model. Previous comparison of these approaches found that joint modelling provided more precise inference about the effect of the covariate while the trinomial model was less prone to issues of model mis‐specification. However, we believe that an important issue was missed. We show through mathematical analysis and numerical simulation that the trinomial model is not identifiable when the predictor has no effect on the survival probability. This also causes inferences from the trinomial model to be imprecise when the effect of the covariate on the survival probability is small. We also analyse data on the effect of body mass on the survival of meadow voles to demonstrate the importance of this issue in real applications.
- Open Access
- 17 November 2022
- Published by Wiley in Journal of Animal Ecology
- Vol. 92 (1) , 183-194
Life span, growth, senescence and island syndrome: Accounting for imperfect detection and continuous growth
Small vertebrates on islands are expected to attain a larger body size, and a greater survival than their mainland counterparts. Comparative studies have questioned whether lizards exhibit this set of adaptations, referred to as the 'island syndrome'. We collected data on 730 individuals the endemic Lilford's lizard Podarcis lilfordi throughout a 10-year period on a small island of the Balearic archipelago (Spain). We coupled a growth function with a capture-mark-recapture model to simultaneously estimate size- and sex-dependent growth rate and survival. To put our results into a wider context, we conducted a systematic review of growth, life span and age at maturity in different Podarcis species comparing insular and mainland populations. We found a low average growth coefficient (0.56 and 0.41 year(-1) for males and females to reach an asymptotic size of 72.3 and 65.6 mm respectively), a high annual survival probability of 0.81 and 0.79 in males and females, and a large variability between individuals in growth parameters. Survival probability decreased with body size in both sexes, indicating a senescence pattern typical of long-lived species or in populations with a low extrinsic mortality. Assuming a constant survival after sexual maturity, at about 2 years old, the average life span was 6.18 years in males and 8.99 in females. The oldest animal was a male last captured at an estimated age of >= 13 years and still alive at the end of the study. Our results agree with the predictions of the 'island syndrome' for survival, life span and growth parameters. A comparative analysis of these values across 29 populations of 16 different species of Podarcis indicated that insular lizards grow slower and live longer than their mainland counterparts. However, our data differed from other island populations of the same species, suggesting that island-specific characteristics play an additional role to isolation. Within this study we developed an analytical approach to study the body size-dependent survival of small reptiles. We discuss its applicability to contrast hypotheses on senescence in different sexes of this species, and provide the code used to integrate the growth and capture-mark-recapture models.
- 17 November 2022
- Published by Wiley in Journal of Animal Ecology
- Vol. 92 (1) , 183-194
Open AccessLife span, growth, senescence and island syndrome: Accounting for imperfect detection and continuous growth
Small vertebrates on islands are expected to attain a larger body size, and a greater survival than their mainland counterparts. Comparative studies have questioned whether lizards exhibit this set of adaptations, referred to as the 'island syndrome'. We collected data on 730 individuals the endemic Lilford's lizard Podarcis lilfordi throughout a 10-year period on a small island of the Balearic archipelago (Spain). We coupled a growth function with a capture-mark-recapture model to simultaneously estimate size- and sex-dependent growth rate and survival. To put our results into a wider context, we conducted a systematic review of growth, life span and age at maturity in different Podarcis species comparing insular and mainland populations. We found a low average growth coefficient (0.56 and 0.41 year(-1) for males and females to reach an asymptotic size of 72.3 and 65.6 mm respectively), a high annual survival probability of 0.81 and 0.79 in males and females, and a large variability between individuals in growth parameters. Survival probability decreased with body size in both sexes, indicating a senescence pattern typical of long-lived species or in populations with a low extrinsic mortality. Assuming a constant survival after sexual maturity, at about 2 years old, the average life span was 6.18 years in males and 8.99 in females. The oldest animal was a male last captured at an estimated age of >= 13 years and still alive at the end of the study. Our results agree with the predictions of the 'island syndrome' for survival, life span and growth parameters. A comparative analysis of these values across 29 populations of 16 different species of Podarcis indicated that insular lizards grow slower and live longer than their mainland counterparts. However, our data differed from other island populations of the same species, suggesting that island-specific characteristics play an additional role to isolation. Within this study we developed an analytical approach to study the body size-dependent survival of small reptiles. We discuss its applicability to contrast hypotheses on senescence in different sexes of this species, and provide the code used to integrate the growth and capture-mark-recapture models.
- Open Access
- 2 November 2022
- Published by Oxford University Press (OUP) in Biometrics
- Vol. 79 (3) , 2732-2742
Latent Multinomial Models for Extended Batch-Mark Data
Batch marking is common and useful for many capture–recapture studies where individual marks cannot be applied due to various constraints such as timing, cost, or marking difficulty. When batch marks are used, observed data are not individual capture histories but a set of counts including the numbers of individuals first marked, marked individuals that are recaptured, and individuals captured but released without being marked (applicable to some studies) on each capture occasion. Fitting traditional capture–recapture models to such data requires one to identify all possible sets of capture–recapture histories that may lead to the observed data, which is computationally infeasible even for a small number of capture occasions. In this paper, we propose a latent multinomial model to deal with such data, where the observed vector of counts is a non-invertible linear transformation of a latent vector that follows a multinomial distribution depending on model parameters. The latent multinomial model can be fitted efficiently through a saddlepoint approximation based maximum likelihood approach. The model framework is very flexible and can be applied to data collected with different study designs. Simulation studies indicate that reliable estimation results are obtained for all parameters of the proposed model. We apply the model to analysis of golden mantella data collected using batch marks in Central Madagascar.
- 2 November 2022
- Published by Oxford University Press (OUP) in Biometrics
- Vol. 79 (3) , 2732-2742
Open AccessLatent Multinomial Models for Extended Batch-Mark Data
Batch marking is common and useful for many capture–recapture studies where individual marks cannot be applied due to various constraints such as timing, cost, or marking difficulty. When batch marks are used, observed data are not individual capture histories but a set of counts including the numbers of individuals first marked, marked individuals that are recaptured, and individuals captured but released without being marked (applicable to some studies) on each capture occasion. Fitting traditional capture–recapture models to such data requires one to identify all possible sets of capture–recapture histories that may lead to the observed data, which is computationally infeasible even for a small number of capture occasions. In this paper, we propose a latent multinomial model to deal with such data, where the observed vector of counts is a non-invertible linear transformation of a latent vector that follows a multinomial distribution depending on model parameters. The latent multinomial model can be fitted efficiently through a saddlepoint approximation based maximum likelihood approach. The model framework is very flexible and can be applied to data collected with different study designs. Simulation studies indicate that reliable estimation results are obtained for all parameters of the proposed model. We apply the model to analysis of golden mantella data collected using batch marks in Central Madagascar.
- 14 September 2022
- Published by Wiley in Environmetrics
- Vol. 34 (2)
Shooting for abundance: Comparing integrated multi‐sampling models for camera trap and hair trap data
Abundance estimation is a vital goal in wildlife monitoring. Camera‐traps are a tool to survey wildlife populations noninvasively and can be used for abundance estimation if individuals are identifiable. However, for species without individual identification characteristics, camera‐trap surveys have often been combined with some other survey method such as capture‐recapture (CR, using traditional tags or DNA through hair snags or scat) to inform an integrated model. We discuss and apply two integrated models involving presence‐absence data from camera traps and CR data from hair traps to compare bias and precision to estimate the population density of grizzly bears of the central Rocky Mountains of Alberta, Canada. Unlike many other studies, we found that integrating presence‐absence data with CR data does not improve the precision of the density estimates. The possible reasons for such results are discussed in detail.
- 14 September 2022
- Published by Wiley in Environmetrics
- Vol. 34 (2)
Shooting for abundance: Comparing integrated multi‐sampling models for camera trap and hair trap data
Abundance estimation is a vital goal in wildlife monitoring. Camera‐traps are a tool to survey wildlife populations noninvasively and can be used for abundance estimation if individuals are identifiable. However, for species without individual identification characteristics, camera‐trap surveys have often been combined with some other survey method such as capture‐recapture (CR, using traditional tags or DNA through hair snags or scat) to inform an integrated model. We discuss and apply two integrated models involving presence‐absence data from camera traps and CR data from hair traps to compare bias and precision to estimate the population density of grizzly bears of the central Rocky Mountains of Alberta, Canada. Unlike many other studies, we found that integrating presence‐absence data with CR data does not improve the precision of the density estimates. The possible reasons for such results are discussed in detail.
- 20 May 2022
- Published by Wiley in Aquatic Conservation: Marine and Freshwater Ecosystems
- Vol. 32 (8) , 1341-1350
New abundance and survival estimates for the critically endangered Taiwanese white dolphin indicate no signs of recovery
Small populations of cetaceans with restricted ranges are particularly vulnerable to anthropogenic disturbances. The assessment of demographic parameters for these populations can provide meaningful insight into the ways in which they are being impacted by different threats. The Taiwanese white dolphin (TWD), Sousa chinensis taiwanensis, is a subspecies endemic to Taiwan and is classified as Critically Endangered. Multistate robust design mark–recapture models were fitted under a Bayesian framework to estimate demographic parameters and trends for the population from 2010 to 2018. Differences in these parameters by age category and the dolphins' use of their habitat were also explored. Abundance estimates varied from a maximum of 72 in 2010 (95% credible interval (CI) [65, 78]) to a minimum of 61 (95% CI [58, 64]) in 2017. In the oldest age category, abundance estimates varied from 10 (95% CI [10, 11]) in 2013 to 13 (95% CI [10, 13]) in 2016. Modelling showed that, over time, the abundance of dolphins in the oldest age group remained stable, while the remainder of the population, which includes reproductive individuals, experienced a steady decline. This decline is bound to affect the demographic recovery of TWD. Spatial heterogeneity indicated that some areas (i.e. with consistent high occurrences) likely represent important areas for biological functions of the TWD. This suggests that the recovery of the population may be enhanced with immediate and focused actions (such as the ban of gillnet fisheries) in these areas. However, with a small and declining population, reducing threats throughout their entire range is needed, including the new offshore windfarm development being constructed in and adjacent to the dolphin's range. Concerted actions to address known threats, continued monitoring of the population, and education of stakeholders are needed to prevent the extinction of the TWDs.
- 20 May 2022
- Published by Wiley in Aquatic Conservation: Marine and Freshwater Ecosystems
- Vol. 32 (8) , 1341-1350
New abundance and survival estimates for the critically endangered Taiwanese white dolphin indicate no signs of recovery
Small populations of cetaceans with restricted ranges are particularly vulnerable to anthropogenic disturbances. The assessment of demographic parameters for these populations can provide meaningful insight into the ways in which they are being impacted by different threats. The Taiwanese white dolphin (TWD), Sousa chinensis taiwanensis, is a subspecies endemic to Taiwan and is classified as Critically Endangered. Multistate robust design mark–recapture models were fitted under a Bayesian framework to estimate demographic parameters and trends for the population from 2010 to 2018. Differences in these parameters by age category and the dolphins' use of their habitat were also explored. Abundance estimates varied from a maximum of 72 in 2010 (95% credible interval (CI) [65, 78]) to a minimum of 61 (95% CI [58, 64]) in 2017. In the oldest age category, abundance estimates varied from 10 (95% CI [10, 11]) in 2013 to 13 (95% CI [10, 13]) in 2016. Modelling showed that, over time, the abundance of dolphins in the oldest age group remained stable, while the remainder of the population, which includes reproductive individuals, experienced a steady decline. This decline is bound to affect the demographic recovery of TWD. Spatial heterogeneity indicated that some areas (i.e. with consistent high occurrences) likely represent important areas for biological functions of the TWD. This suggests that the recovery of the population may be enhanced with immediate and focused actions (such as the ban of gillnet fisheries) in these areas. However, with a small and declining population, reducing threats throughout their entire range is needed, including the new offshore windfarm development being constructed in and adjacent to the dolphin's range. Concerted actions to address known threats, continued monitoring of the population, and education of stakeholders are needed to prevent the extinction of the TWDs.