Sample demographic regression model coefficients

Source: R/demographicCoefficients.R, R/getCoefs.R, R/sampleCoefs.R

Select the regression coefficient values and standard errors for the desired model version (see popGrowthTableJohnsonECCC for options) and then sample from the Gaussian distribution for each replicate population. demographicCoefficients is a wrapper around getCoefs(), which selects coefficients and sampleCoefs(), which samples coefficients, for both the survival and recruitment models.

demographicCoefficients(
  replicates,
  modelVersion = "Johnson",
  survivalModelNumber = "M1",
  recruitmentModelNumber = "M4",
  useQuantiles = TRUE,
  populationGrowthTable = popGrowthTableJohnsonECCC
)

getCoefs(populationGrowthTable, resVar, modelVersion, modNum)

sampleCoefs(coefTable, replicates)

Arguments

replicates

integer. Number of replicate populations.

modelVersion

character. Which model version to use. Currently the only option is "Johnson" for the model used in Johnson et. al. (2020), but additional options may be added in the future.

survivalModelNumber, recruitmentModelNumber

character. Which model number to use see popGrowthTableJohnsonECCC for options.

useQuantiles

logical or numeric. If it is a numeric vector it must be length 2 and give the low and high limits of the quantiles to use. If useQuantiles != FALSE, each replicate population is assigned to a quantile of the distribution of variation around the expected values, and remains in that quantile as covariates change. If useQuantiles = TRUE, replicate populations will be assigned to quantiles in the default range of 0.025 and 0.975.

populationGrowthTable

data.frame.popGrowthTableJohnsonECCC is included in the package and should be used in most cases. A custom table of model coefficients and standard errors or confidence intervals can be provided but it must match the column names of popGrowthTableJohnsonECCC. If the table does not contain the standard error it is calculated from the confidence interval.

resVar

character. Response variable, typically "femaleSurvival" or "recruitment"

modNum

character vector. Which model number(s) to use see popGrowthTableJohnsonECCC for typical options.

coefTable

data.table. Table must have columns "Coefficient" for the name of the coefficient, "Value" for the value of the coefficient and "StdErr" for the standard error of coefficients. Typically created with getCoefs()

Value

For demographicCoefficients a list with elements:

  • "modelVersion": The name of the model version

  • "coefSamples_Survival" and"coefSamples_Recruitment": lists with elements:

    • "coefSamples": Bootstrapped coefficients with replicates rows

    • "coefValues": Coefficient values taken from populationGrowthTable

    • "quantiles": A vector of randomly selected quantiles between 0.025 and 0.975 with length replicates

For getCoefs: a named list with one element per model version. The names are modelVersion_modNum_Type. Each element contains a data.frame that is a subset of populationGrowthTable for the selected model

For sampleCoefs a list with elements:

  • "coefSamples": Bootstrapped coefficients with replicates rows

  • "coefValues": Coefficient values taken from populationGrowthTable

Details

Each population is optionally assigned to quantiles of the error distributions for survival and recruitment. Using quantiles means that the population will stay in these quantiles as disturbance changes over time, so there is persistent variation in recruitment and survival among example populations. See demographicRates() for more details.

References

Johnson, C.A., Sutherland, G.D., Neave, E., Leblond, M., Kirby, P., Superbie, C. and McLoughlin, P.D., 2020. Science to inform policy: linking population dynamics to habitat for a threatened species in Canada. Journal of Applied Ecology, 57(7), pp.1314-1327. https://doi.org/10.1111/1365-2664.13637

Examples

# sample coefficients for default models
demographicCoefficients(10)
#> $modelVersion
#> [1] "Johnson"
#> 
#> $coefSamples_Survival
#> $coefSamples_Survival$coefSamples
#>        Intercept        Anthro Precision
#>  [1,] -0.1336595 -0.0007624518  59.53215
#>  [2,] -0.1409718 -0.0008269774  60.53790
#>  [3,] -0.1331624 -0.0008968514  59.79600
#>  [4,] -0.1279467 -0.0009351104  54.49847
#>  [5,] -0.1402076 -0.0007126604  67.70302
#>  [6,] -0.1341783 -0.0009816173  68.66600
#>  [7,] -0.1278602 -0.0008103753  58.57169
#>  [8,] -0.1476828 -0.0007057689  79.48561
#>  [9,] -0.1480838 -0.0008060019  56.88737
#> [10,] -0.1379074 -0.0010090037  61.13307
#> 
#> $coefSamples_Survival$coefValues
#>    Intercept Anthro Precision
#>        <num>  <num>     <num>
#> 1:    -0.142 -8e-04  63.43724
#> 
#> $coefSamples_Survival$coefStdErrs
#>      Intercept      Anthro Precision
#>          <num>       <num>     <num>
#> 1: 0.007908163 0.000127551  8.272731
#> 
#> $coefSamples_Survival$quantiles
#>  [1] 0.4472222 0.9750000 0.6583333 0.3416667 0.7638889 0.5527778 0.0250000
#>  [8] 0.1305556 0.2361111 0.8694444
#> 
#> 
#> $coefSamples_Recruitment
#> $coefSamples_Recruitment$coefSamples
#>        Intercept      Anthro fire_excl_anthro Precision
#>  [1,] -1.0385914 -0.01784110     -0.005095197  16.24928
#>  [2,] -0.9935816 -0.01546560     -0.007507857  17.86046
#>  [3,] -1.0024445 -0.01887748     -0.006386528  18.14406
#>  [4,] -1.0387195 -0.01414084     -0.004974346  18.99112
#>  [5,] -1.0087392 -0.01477727     -0.012179255  21.82254
#>  [6,] -1.1043363 -0.01951234     -0.011224992  24.66386
#>  [7,] -0.9196172 -0.01576567     -0.008414955  15.99503
#>  [8,] -1.1234125 -0.01502318     -0.004762866  17.96014
#>  [9,] -1.0537040 -0.01606396     -0.005006864  20.54785
#> [10,] -0.8653840 -0.01818452     -0.010078639  22.75818
#> 
#> $coefSamples_Recruitment$coefValues
#>    Intercept Anthro fire_excl_anthro Precision
#>        <num>  <num>            <num>     <num>
#> 1:    -1.023 -0.017          -0.0081  19.86189
#> 
#> $coefSamples_Recruitment$coefStdErrs
#>     Intercept      Anthro fire_excl_anthro Precision
#>         <num>       <num>            <num>     <num>
#> 1: 0.06122449 0.001530612      0.002040816  2.228655
#> 
#> $coefSamples_Recruitment$quantiles
#>  [1] 0.4472222 0.2361111 0.6583333 0.1305556 0.7638889 0.0250000 0.9750000
#>  [8] 0.8694444 0.5527778 0.3416667
#> 
#> 

# try a different model
demographicCoefficients(10, modelVersion = "Johnson", survivalModelNumber = "M1",
                        recruitmentModelNumber = "M3")
#> $modelVersion
#> [1] "Johnson"
#> 
#> $coefSamples_Survival
#> $coefSamples_Survival$coefSamples
#>        Intercept        Anthro Precision
#>  [1,] -0.1478660 -0.0009110440  56.49724
#>  [2,] -0.1446051 -0.0006600159  53.13350
#>  [3,] -0.1384197 -0.0009008546  73.85101
#>  [4,] -0.1431392 -0.0006623174  59.66967
#>  [5,] -0.1404038 -0.0009753110  69.12058
#>  [6,] -0.1360657 -0.0006962018  63.91080
#>  [7,] -0.1542562 -0.0008695985  65.19231
#>  [8,] -0.1530940 -0.0007732620  65.08848
#>  [9,] -0.1553243 -0.0008528537  62.93344
#> [10,] -0.1581460 -0.0006450591  69.57432
#> 
#> $coefSamples_Survival$coefValues
#>    Intercept Anthro Precision
#>        <num>  <num>     <num>
#> 1:    -0.142 -8e-04  63.43724
#> 
#> $coefSamples_Survival$coefStdErrs
#>      Intercept      Anthro Precision
#>          <num>       <num>     <num>
#> 1: 0.007908163 0.000127551  8.272731
#> 
#> $coefSamples_Survival$quantiles
#>  [1] 0.7638889 0.4472222 0.8694444 0.3416667 0.9750000 0.0250000 0.6583333
#>  [8] 0.1305556 0.5527778 0.2361111
#> 
#> 
#> $coefSamples_Recruitment
#> $coefSamples_Recruitment$coefSamples
#>        Intercept  Total_dist
#>  [1,] -0.8062151 -0.01563943
#>  [2,] -0.9196684 -0.01593913
#>  [3,] -0.8979511 -0.01562644
#>  [4,] -1.0171600 -0.01356781
#>  [5,] -0.9280347 -0.02011052
#>  [6,] -0.9018135 -0.01677620
#>  [7,] -0.8550586 -0.01417455
#>  [8,] -0.9719659 -0.01426752
#>  [9,] -0.9180666 -0.01365285
#> [10,] -0.8895341 -0.01191941
#> 
#> $coefSamples_Recruitment$coefValues
#>    Intercept Total_dist
#>        <num>      <num>
#> 1:    -0.956     -0.015
#> 
#> $coefSamples_Recruitment$coefStdErrs
#>    Intercept  Total_dist
#>        <num>       <num>
#> 1: 0.0619898 0.001530612
#> 
#> $coefSamples_Recruitment$quantiles
#>  [1] 0.3416667 0.7638889 0.6583333 0.0250000 0.5527778 0.2361111 0.9750000
#>  [8] 0.4472222 0.8694444 0.1305556
#> 
#> 

getCoefs(popGrowthTableJohnsonECCC, "femaleSurvival", "Johnson", "M1")
#> $Johnson_M1_National
#>    modelVersion responseVariable ModelNumber     Type Coefficient    Value
#>          <char>           <char>      <char>   <char>      <char>    <num>
#> 1:      Johnson   femaleSurvival          M1 National   Intercept -0.14200
#> 2:      Johnson   femaleSurvival          M1 National      Anthro -0.00080
#> 3:      Johnson   femaleSurvival          M1 National   Precision 63.43724
#>         StdErr lowerCI upperCI
#>          <num>   <num>   <num>
#> 1: 0.007908163  -0.158 -0.1270
#> 2: 0.000127551  -0.001 -0.0005
#> 3: 8.272730950      NA      NA
#> 

cfs <- getCoefs(popGrowthTableJohnsonECCC, "recruitment", "Johnson", "M3")

sampleCoefs(cfs[[1]], 10)
#> $coefSamples
#>        Intercept  Total_dist
#>  [1,] -0.9578218 -0.01492273
#>  [2,] -0.9135661 -0.01039179
#>  [3,] -0.9276099 -0.01584458
#>  [4,] -0.9451325 -0.01377702
#>  [5,] -0.9568116 -0.01266597
#>  [6,] -1.0399361 -0.01389490
#>  [7,] -0.9200718 -0.01369685
#>  [8,] -1.0040460 -0.01466352
#>  [9,] -1.0560931 -0.01433763
#> [10,] -1.0387943 -0.01374982
#> 
#> $coefValues
#>    Intercept Total_dist
#>        <num>      <num>
#> 1:    -0.956     -0.015
#> 
#> $coefStdErrs
#>    Intercept  Total_dist
#>        <num>       <num>
#> 1: 0.0619898 0.001530612
#>