JCJason Conner

Jason Conner

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Jason Conner has a total of 21 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Framework to Investigate the Effects of Observation Error on Neural Network Predictions of Fish Age

    by Derek W. Chamberlin, Thomas E. Helser, John D. Brogan, Irina M. Benson, Jason Conner, Andrew T. Chin, Christopher M. Gburski, Mary Elizabeth Matta, Julie A. Pearce, Kali R. Stone, et al.

    Technological innovations for predicting fish age represent a paradigm shift from conventional age estimation methods used in fisheries science. Recently developed secondary methods rely on models trained on conventional age estimates, derived from subjective interpretation of growth patterns and a biological property of the fish to predict age. Hence, quantifying the error that propagates from these models and the conventional age estimates on which they are trained is critical to fully account for uncertainty in predicted ages used in fisheries stock assessments. We review the development of three secondary age prediction methods: image analysis, epigenetics and Fourier transform near‐infrared spectroscopy (FT‐NIRS); the use of artificial intelligence (AI) within each method; and present an approach to examine the effects of ageing error on AI age prediction models. As a case study, we conducted an empirical study, coupled with simulation, to quantify the effects of ageing error on the performance of a multimodal convolutional neural network (MMCNN) model used to predict eastern Bering Sea walleye pollock ( Gadus chalcogrammus ) ages from otoliths analysed using FT‐NIRS. Results indicated repeatability of predicted ages was high between instrument operators, while adding ageing error resulted in a slight decrease in model performance from R 2 = 0.92 and CV = 7.6% to R 2 = 0.87 and CV = 10.2% on test datasets. Our results also suggest the MMCNN model is robust to noise in calibration age data and model performance may be better than performance metrics indicate when the model is evaluated against data with increased error.

  • Open Access

    The generalized gamma is a flexible distribution that outperforms alternatives when modelling catch rate data

    by Jillian C Dunic, Jason Conner, Sean C Anderson, James T Thorson

    Fisheries scientists use regression models to estimate population quantities, such as biomass or abundance, for use in climate, habitat, stock, and ecosystem assessments. However, these models are sensitive to the probability distribution used to characterize observation error. Here, we introduce the generalized gamma distribution (GGD), which has not been widely used in fisheries science. The GGD has useful properties: (1) it reduces to the lognormal distribution when the shape parameter approaches zero, (2) it reduces to the gamma distribution when the shape and scale parameters are equal, and (3) the coefficient of variation is independent of the mean. We assess the relative performance and robustness of the GGD to estimate biomass density across different observation error types in a simulation experiment. When fit to data generated from the GGD, lognormal, gamma, and Tweedie families, the GGD had low bias, high predictive accuracy, and appropriate confidence interval coverage under most scenarios. Finally, we fit spatiotemporal index standardization models using the R package sdmTMB to 14 species from three trawl surveys from the Gulf of Alaska and coast of British Columbia, Canada. When the Akaike information criterion weight was compared among fits using the lognormal, gamma, and Tweedie families, the GGD was the most commonly selected model. When compared via cross validation, predictive performance favoured the GGD for most (75%) Gulf of Alaska species but was less clear for the smaller British Columbia surveys. We conclude that the GGD is a flexible and effective choice for modelling fisheries-independent survey data, offering several advantages over commonly used distributions. Given its strong performance in both simulation and empirical analyses, we recommend that practitioners consider the GGD when selecting error distributions for these models.


  • The sensitivity of fisheries-independent survey indices to decisions of sampling design and intensity and the mitigation of biased precision estimators for systematic sampling

    by Jason Conner, Stan Kotwicki, Kotaro Ono, Lewis A.K. Barnett

    Fisheries-independent surveys are used to track trends in fish stocks globally. Survey alterations occur for a myriad of reasons (e.g., vessel availability, changes in fish distribution, funding shortfalls). Understanding the sensitivity of survey estimates to changes in sampling is pivotal to sustainable management. We present a case-study of an annual multispecies survey. Simulating distributions for four species with a spatiotemporal model, we evaluated simple random, stratified random, and systematic grid sampling designs across four sampling intensities. The systematic design yielded higher-precision estimates at all sampling intensities. However, the commonly used standard error estimator thereof resulted in mean biases from 24% to 63%. We evaluated two alternative standard error estimators that successfully mitigated these biases. Our results indicate that estimates from systematic survey designs may be robust to decreases in sampling intensity, and that analyses such as integrated stock assessments which use these estimates may mitigate model misspecification by applying alternative variance estimators.


  • Reductions in sampling effort for fishery-independent age and length composition: balancing sampling efficiency, data uncertainty, and workforce health

    by Peter-John F. Hulson, Benjamin C. Williams, Meaghan D. Bryan, Jason Conner, Matthew R. Siskey

    Unavoidable survey effort reduction has become a reality that must be accounted for in fisheries stock assessment. In addition, negative consequences to survey staff health due to repetitive motion injuries are becoming increasingly costly for managing agencies. In this study, we evaluated the outcomes of reductions in age and length data used in fisheries stock assessment models. The main goal was to determine whether sampling can be reduced to a level that does not excessively increase data uncertainty, yet provides a reduction in repetitive motions that can cause injury to survey staff. We found that reducing length sampling to a maximum of 100–150 fish sampled per haul (either sex-specific or combined sex) provides length composition data for which the uncertainty is not appreciably increased, and it has minimal effect on the uncertainty in age composition data that is subsequently expanded from this subsampled length frequency data. The method employed here, and the results presented, can aid management agencies to balance the magnitude of data collection and subsequent consequences to fisheries stock assessment models.