Missing zones

[fssem1]

Hi Sara, I believe I know what the “blank” circuli are. These look like the "plus" circuli, or circuli formed after the final annulus on the scale. It is the spring growth in freshwater after winter ended and before the fish emigrated as a smolt. The best way to confirm this is to look at an image and compare the count of circuli in each year of growth (numbered) and the plus growth (not numbered) on the image and in the measurements. I would check for the presence of plus circuli on his image, but the screen shot does not show the first few columns, one of which is the image name. Going out on a limb, I am remembering that most of the scales, especially the older-age fish, have “plus” circuli. Lorna

From: Miller, Sara E (DFG) Sent: Monday, April 8, 2019 11:41 AM To: Wilson, Lorna I (DFG) lorna.wilson@alaska.gov Subject: FW: Coho Scale Analysis Issue

From: Justin Priest jtpriest@alaska.edu Sent: Sunday, April 7, 2019 4:39 PM To: Priest, Justin Tm (DFG) justin.priest@alaska.gov Cc: Miller, Sara E (DFG) sara.miller@alaska.gov Subject: Re: Coho Scale Analysis Issue

After updating my code implementing Ben's readability suggestions, and fixing the code to not depend upon reshape2, the code line numbers are now off from what I wrote earlier. It should be lines 118-134; you'll see the "Sara look here" comment. After updating everything today, I think everything comes out in the wash and it doesn't matter as far as I can tell (maybe a few edge cases that I can't find?). But still, it would be good to know for sure. I can easily write a line to filter out blank Zones.

To visualize the problem a little better, here's a screenshot of the CSV file:

You can see that there are several distances without corresponding zones.

Justin

On Fri, Apr 5, 2019 at 4:27 PM Priest, Justin Tm (DFG) justin.priest@alaska.gov wrote: Hi Sara,

I have encountered an issue in the code and I think that I'll need your help to me decide which is correct. When I run a "melt" vs "gather, I'm coming up with slightly different results. Basically my code has Circulus 23&24 included while the original code doesn't. I can't find where the orig code specifically calls to exclude these circuli so I am guessing that it was a welcome bug or an unintentional result (or I'm missing something somewhere which is very possible). It wasn't an issue until I was calculating the cumulative sum of things. Here are snapshots of the code output below. I also flagged the sections in the files if you want to run it yourself to view it (Line 149 in LDA_SEM and Line 102-119 in LDA_JTP_DataImport).

Are these circuli supposed to be excluded? If so, I can filter them out easily.

My code:

Original code:

[fssem1]

So, I think the original data had a zone associated with this "plus" group. So in my script I removed all data with zones > then the age of the fish. This dataset seems to have a blank zone for the plus group. I think you should 1) Verify these are plus group zones from the scale images Lorna gave you. If you need to have her confirm something-please email her 2) If these are plus groups, you should be able to remove the data since "Variables should only measure to the end of a zone, not to the end of the plus growth (end of scale) since this distance is variable and occurs after annuli have formed.”

[lornaiwilson]

About (1): Justin and I talked about how to compare measurements to the image files, where overlays are available. For the couple of images we talked through, these are plus group zones. Justin has the capability to confirm these for other scale measurements.

About (2): One suggestion (that may have been mine) is to not consider "plus" circuli because the fish has completed its fresh water age-specific growth by the end of its final winter in fresh water. But, considering that there may be smolt age-specific information in the "plus" growth (my impression si that generally, older smolt have less "plus" circuli) another suggestion is to consider using "plus" circuli data to help to estimate age. It may be good to consider for this alternate suggestion that the data should be there for the age estimation process, at least for smolts and adults. We will need to assume smolts have formed all the "plus" circuli they will have because they are out-migrating (which can be confirmed with other smolt characteristics, e.g. absence of parr marks/silvery color). Adults will have all freshwater scale characteristics.