Evaluation of GLM and GAM for estimating population indices from fishery independent surveys

Potts, Stephen E.; Rose, Kenneth A.

doi:10.1016/j.fishres.2018.07.016

Cited by 27 publications

(19 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…A major challenge for estimating pre-recruit abundance indices from surveys is to account for complex spatial and temporal variation in pre-recruit abundance (Denson et al, 2017;Potts and Rose, 2018). Variation in abundance across successive juvenile stages could be driven by small scale processes, leading to large spatial discrepancies among juvenile habitats (Scharf, 2000).…”

Section: Discussionmentioning

confidence: 99%

The use and performance of survey-based pre-recruit abundance indices for possible inclusion in stock assessments of coastal-dependent species

Pape

Vermard

Guitton

et al. 2020

ICES Journal of Marine Science

View full text Add to dashboard Cite

We reviewed the use of survey-based pre-recruit abundance indices in short-term recruitment forecasts for fish species relying on coastal habitats at the juvenile stage and that are assessed by ICES. We collated information from stock assessment reports and from a questionnaire filled out by the stock assessors. Among the 78 stocks with juvenile coastal dependence, 49 use short-term forecasts in stock assessment. Survey-based pre-recruit abundance indices were available for 35 of these stocks, but only 14 were used to forecast recruitment. The questionnaire indicated that the limited use of survey-based pre-recruit abundance indices was primarily due to sampling inefficiency, which may preclude reliable recruitment estimates. The sampling is inefficient because the juvenile coastal distribution is outside the geographical area covered by large-scale surveys or targeted coastal surveys are conducted on limited spatial and temporal scales. However, our analysis of the relationship between survey-based pre-recruit indices and assessment-generated recruitment indices revealed that survey-based pre-recruit abundance indices were sufficiently accurate to provide useful information for predicting future recruitment. We recommend expansion of the use of survey-based indices of pre-recruit abundance in stock assessment and recruitment forecasting, and consideration of how to include juveniles in ongoing and future surveys.

show abstract

Section: Discussionmentioning

confidence: 99%

The use and performance of survey-based pre-recruit abundance indices for possible inclusion in stock assessments of coastal-dependent species

Pape

Vermard

Guitton

et al. 2020

ICES Journal of Marine Science

View full text Add to dashboard Cite

show abstract

“…GAM implemented in the mgcv package in R (4.0.3). The most optimum model was selected with the lowest Akaike Information Criterion (AIC) and residual deviance [ 54 ]. The general form is as follows:

where

is the connection function;

is the expected value of the response variable

;

is a constant; and

is a smooth function of the explanatory variable

.…”

Section: Methodsmentioning

confidence: 99%

Section: Model Fitting and Selectionmentioning

confidence: 99%

The Modeling and Forecasting of Carabid Beetle Distribution in Northwestern China

Liu

Wang

et al. 2021

Insects

View full text Add to dashboard Cite

Beetles are key insect species in global biodiversity and play a significant role in steppe ecosystems. In the temperate steppe of China, the increasing degeneration of the grasslands threatens beetle species and their habitat. Using Generalized Additive Models (GAMs), we aimed to predict and map beetle richness patterns within the temperate steppe of Ningxia (China). We tested 19 environmental predictors including climate, topography, soil moisture and space as well as vegetation. Climatic variables (temperature, precipitation, soil temperature) consistently appeared among the most important predictors for beetle groups modeled. GAM generated predictive cartography for the study area. Our models explained a significant percentage of the variation in carabid beetle richness (79.8%), carabid beetle richness distribution seems to be mainly influenced by temperature and precipitation. The results have important implications for management and conservation strategies and also provides evidence for assessing and making predictions of beetle diversity across the steppe.

show abstract

“…At a national level we compare foraging radius distributions for 25 breeding seabird species to extensive aerial surveys conducted over a two-year period. To provide a benchmark for the correlation values between foraging radius distributions and empirical data, we also model distributions from the aerial survey data using generalised additive models (GAMs), incorporating environmental predictors, as this approach is often considered to be the best method for modelling survey data (Booth andHammond 2014, Potts andRose 2018). We discuss the performance of the foraging radius model in comparison to the empirical data and the appropriateness of using this method for assessing seabird distributions under different scenarios.…”

Section: Introductionmentioning

confidence: 99%

Assessing the effectiveness of foraging radius models for seabird distributions using biotelemetry and survey data

et al. 2019

View full text Add to dashboard Cite

Relatively simple foraging radius models have the potential to generate predictive distributions for a large number of species rapidly, thus providing a cost-effective alternative to large-scale surveys or complex modelling approaches. Their effectiveness, however, remains largely untested. Here we compare foraging radius distribution models for all breeding seabirds in Ireland, to distributions of empirical data collected from tracking studies and aerial surveys. At the local/colony level, we compared foraging radius distributions to GPS tracking data from seabirds with short (Atlantic puffin Fratercula arctica, and razorbill Alca torda) and long (Manx shearwater Puffinus puffinus, and European storm-petrel Hydrobates pelagicus) foraging ranges. At the regional/national level, we compared foraging radius distributions to extensive aerial surveys conducted over a two-year period. Foraging radius distributions were significantly positively correlated with tracking data for all species except Manx shearwater. Correlations between foraging radius distributions and aerial survey data were also significant, but generally weaker than those for tracking data. Correlations between foraging radius distributions and aerial survey data were benchmarked against generalised additive models (GAMs) of the aerial survey data that included a range of environmental covariates. While GAM distributions had slightly higher correlations with aerial survey data, the results highlight that the foraging radius approach can be a useful and pragmatic approach for assessing breeding distributions for many seabird species. The approach is likely to have acceptable utility in complex, temporally variable ecosystems and when logistic and financial resources are limited.

show abstract

Evaluation of GLM and GAM for estimating population indices from fishery independent surveys

Cited by 27 publications

References 55 publications

The use and performance of survey-based pre-recruit abundance indices for possible inclusion in stock assessments of coastal-dependent species

The use and performance of survey-based pre-recruit abundance indices for possible inclusion in stock assessments of coastal-dependent species

The Modeling and Forecasting of Carabid Beetle Distribution in Northwestern China

Assessing the effectiveness of foraging radius models for seabird distributions using biotelemetry and survey data

Contact Info

Product

Resources

About