Just out of curiousity...when random noise is added to the survey data in the function atlantis_tracer_survey_select the sigma argument is squared to produce the err_variance object within the function on line132 (which is then used as the sd argument of rnorm). I'm just wondering why this term is squared as rnorm takes standard deviation (which is the square root of variance) as the argument for how much variation to introduce in the normal distribution.
Shouldn't the sigma argument either be input directly into rnorm? Or alternatively the user could be asked for variance and then take the sqrt of that and put into rnorm
Just out of curiousity...when random noise is added to the survey data in the function
atlantis_tracer_survey_select
thesigma
argument is squared to produce theerr_variance
object within the function on line132 (which is then used as the sd argument ofrnorm
). I'm just wondering why this term is squared asrnorm
takes standard deviation (which is the square root of variance) as the argument for how much variation to introduce in the normal distribution.Shouldn't the sigma argument either be input directly into
rnorm
? Or alternatively the user could be asked for variance and then take thesqrt
of that and put intornorm