search

GLEON / LakeMetabolizer

Collection of Lake Metabolism Functions
16 stars 10 forks source link

2 Errors in metab.bayesian() example #30

Open rBatt opened 10 years ago

rBatt commented 10 years ago

https://github.com/GLEON/LakeMetabolizer/blob/master/man/metab.bayesian.Rd#L59

1) k.gas needs to be divided by the number of obs per day 2) wnd[,2] needs to be scaled to wind at 10m

k.gas needs to be scaled to the size of the time step. Right now that's the K in units of m/day, needs to be m/5 minute or w/e. Just divide by the number of time steps per day.

I have some functions I've been using to calculate the size of a time step. There are a couple ways someone could go about doing this (different time step sizes for obs, or just 1 for the whole time series?).

Either way, I think this example is wrong as-is. There might be other errors.

jordansread commented 10 years ago

I don't think that is a good idea. Units of k are Length/Time, and those functions should output in that format so they are still useful if used outside of the package.

Jordan Read, PhD | U.S. Geological Survey | Center for Integrated Data Analytics 608-821-3922 | http://cida.usgs.gov/people/jread.html

On Jun 11, 2014, at 12:32 PM, Ryan Batt wrote:

https://github.com/GLEON/LakeMetabolizer/blob/master/man/metab.bayesian.Rd#L59

1) k.gas needs to be divided by the number of obs per day 2) wnd[,2] needs to be scaled to wind at 10m

k.gas needs to be scaled to the size of the time step. Right now that's the K in units of m/day, needs to be m/5 minute or w/e. Just divide by the number of time steps per day.

I have some functions I've been using to calculate the size of a time step. There are a couple ways someone could go about doing this (different time step sizes for obs, or just 1 for the whole time series?).

Either way, I think this example is wrong as-is. There might be other errors.

— Reply to this email directly or view it on GitHub.

rBatt commented 10 years ago

@jread-usgs : The k functions are fine. But the example is wrong.

jordansread commented 10 years ago

Thanks @rBatt I get it now.

lawinslow commented 10 years ago

We either need to have the metabolism functions assume m/day and do the conversion themselves, or assume m/timestep and force the user to do the conversion. I would probably prefer the former. k/timestep can get confusing and hard to work with.

-Luke