Examining when ice formation and breakup occurs

[dzrobert]

I used to use a command:
iceonoff_mod = mda.lakes::get_ice_onoff(ice_data, get_temp(nc_file))
to get ice information

This now gives me the following error:
Error in ice_runs[, "values"] : subscript out of bounds

Anyway to fix this so that I can compare actual and observed freeze and breakup dates?

[hdugan]

I'm not familiar with the mda.lakes package, but does
get_temp(nc_file) still work on its own?

[jordansread]

Oh, I didn't realize this was an mda.lakes function. That package is pretty unsupported for general use...

The function is here. I'm guessing there is either a change w/ GLM v3 that isn't supported or part of the move to the updated get_var?

[dzrobert]

Yes, that works fine. I was using the following series of commands to compare actual and measured ice freeze and breakup:
iceonoff_mod = mda.lakes::get_ice_onoff(ice_data, get_temp(nc_file))
#MODELED Ice Dates
iceonoff_mod
iceonoff_obs = read.csv('Green_ice19.csv', as.is=TRUE)
iceonoff_obs$on = as.POSIXct(iceonoff_obs$on)
iceonoff_obs$off = as.POSIXct(iceonoff_obs$off)
sprintf('Ice ON: %g mean diff mod - obs', mean(as.numeric(iceboth$on_mod - iceboth$on_obs, units='days')))
sprintf('Ice OFF: %g mean diff mod - obs', mean(as.numeric(iceboth$off_mod - iceboth$off_obs, units='days')))

sprintf('Ice ON: %g mean absolute diff', mean(abs(as.numeric(iceboth$on_mod - iceboth$on_obs, units='days'))))
sprintf('Ice OFF: %g mean absolute diff', mean(abs(as.numeric(iceboth$off_mod - iceboth$off_obs, units='days'))))

[hdugan]

The error is in the mda.lakes::get_ice_onoff function.
It was probably due to an update in the rle2 function.
The function is calling the wrong names from the rle2 output. They need to be
'value'
'start'
'stop'
'length'
NOT values starts stops lengths

[hdugan]

This is a nice function though - maybe it should be in glmtools?

[dzrobert]

For us ice people it would be great. Dale

…

Sent from my iPhone

On Jul 18, 2019, at 3:57 PM, hdugan ***@***.***> wrote: This is a nice function though - maybe it should be in glmtools? — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or mute the thread.

[jordansread]

I'm worried about exposing it in the package because it can give you the wrong answer (it makes assumptions about ice dates relative to peak temperatures). I think it is right 99.9% of the time, but consider it to be more of an analysis method that you'd do on the data coming out of get_ice() (or _var) where you'd have control and visibility in the decision of when to declare on vs off.

Or perhaps we expose it but treat as experimental?

[pchanson]

If it’s included, could it be called an ‘estimator’ or some other similar term that indicates it’s derived from assumed relationships that really are statistical in nature? Alternatively, is there a need for more of a GLM data analysis package, kind of lake the existing lake physics package you created, Jordan? --------------------------------------------- Paul Hanson Distinguished Research Professor UW Center for Limnology From: Jordan S Read <[email protected]<mailto:[email protected]>> Reply-To: USGS-R/glmtools <[email protected]<mailto:[email protected]>> Date: Friday, July 19, 2019 at 8:50 AM To: USGS-R/glmtools <[email protected]<mailto:[email protected]>> Cc: Subscribed <[email protected]<mailto:[email protected]>> Subject: Re: [USGS-R/glmtools] Examining when ice formation and breakup occurs (#251) I'm worried about exposing it in the package because it can give you the wrong answer (it makes assumptions about ice dates relative to peak temperatures). I think it is right 99.9% of the time, but consider it to be more of an analysis method that you'd do on the data coming out of get_ice() (or _var) where you'd have control and visibility in the decision of when to declare on vs off. Or perhaps we expose it but treat as experimental? — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub<#251>, or mute the thread<https://github.com/notifications/unsubscribe-auth/AADDY7PGJEEATLSUWELK67TQAHBD3ANCNFSM4IE4HO4A>.

[dzrobert]

I would be OK working with an experimental package? Right now I am just trying to see if ice on and off dates are close, Rather than looking through the daily ice thicknesses, this was working great. Is there a way just to grab a few lines from mda.lakes that I can run?

[hdugan]

Dale, I forked the mda repo and fixed the function. You can use it here:

install_github('hdugan/mda.lakes')

Don't know if you want a pull request to the main repo