search

tetratech / baytrends

Evaluate long-term trends in the Chesapeake Bay using a Generalized Additive Modeling (GAM) approach
https://tetratech.github.io/baytrends/
12 stars 4 forks source link

gamPenaltyCrit logic error #85

Closed jharcum07 closed 4 months ago

jharcum07 commented 4 months ago

RK: When I ran my R scripts to create the seasonal trend graphics for CHLA for trends through 2023 I got errors. I did not have a problem with creating the summer DO graphs. I think perhaps it has to do with something that has changed between last year's Baytrends v2.0.9 and this year's version Baytrends v2.0.11. I think that because I tried running the same script with last year's dataset (through 2022) and I got the same error that I get when I tried updating to the new dataset (through 2023).

image

Sometimes the 'length = 8'

Perhaps links to the trends where the output includes F-stat maybe unreliable

I ended up running each station individually for each of the three time periods and separately for spring and summer to see what stations would run. And then I ran them as just gam2 or gam4 if I still didn't get them to run when the script tried to run both gam models. The attached spreadsheet shows the matrix of all timeperiods by station to note which would or would not run.

jharcum07 commented 4 months ago

waiting for testing feedback from RM

jharcum07 commented 4 months ago

should be good to go