decide what to do about GDD

[lizzieinvancouver]

@alinazeng I am thinking on what to say about the GDD results, but it depends on how we did them. Can you clarify in the text how you did GDD. Specifically:

1. did you calculate it specific per tree or garden or ... ?
2. And what about data before 2011?

I know we discussed, but I have forgotten. Thanks!

[alinazeng]

@lizzieinvancouver Hi Lizzie, sorry for the late response.

did you calculate it specific per tree or garden or ... ?

Recall that I extracted daily temp data from 2011 to 2020 for each provenance and garden (meaning each event doy is associated with the temp on that day), I then calculated GDD using 0 degree Celsius as base, for each provenance and each garden (so now each event doy is associated with the temp on that day as well as the GDD on that day for us to analyze)

And what about data before 2011?

Since we decided to use temp data from 2011 to 2020 for all provenances instead of data of the year in which a study was conducted, I calculated gdd using data in the 2011 to 2020 period too, so there's no data before 2011.

I hope my response above makes sense.

[lizzieinvancouver]

@alinazeng thanks for working with me on this! I am know I am forgetting things you have told me once or twice before.... A couple more:

a) Did we only then include data collected in 2011 or 2020 (I doubt it base on the figures, but I could be wrong)? b) If yes, what percent of data did we exclude? b) If no, did we average 2011-2020 data before taking GDD? If not, how did we match up years for each event?

[alinazeng]

Hi Lizzie, Anytime！ Studies we included were conducted in a wide range of years (from 1960ish to 2016 or something like that. don't quote me on this cuz I don't remember the exact top of my head) Daily climate data wasnt available for many of the years in which the studies were done, therefore, we decided that we would use daily climate data from 2011-2020 for all studies we included, regardless what year they were done. This means that we never matched events with the climate in the days when observations were made (unless the study happens to be within 2011-2020. Our MAT comes from taking the average of daily climate by month, and then by year. I calculated GDD using daily temp data by year. So each provenance should have a GDD on the event day for 2011-2020. Let me know if this makes sense！ Cheers,Alina

On Wednesday, August 2, 2023, 2:09 AM, Elizabeth M Wolkovich @.***> wrote:

@alinazeng thanks for working with me on this! I am know I am forgetting things you have told me once or twice before.... A couple more:

a) Did we only then include data collected in 2011 or 2020 (I doubt it base on the figures, but I could be wrong)? b) If yes, what percent of data did we exclude? b) If no, did we average 2011-2020 data before taking GDD? If not, how did we match up years for each event?

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>

[lizzieinvancouver]

@alinazeng

a) Okay, can you tell me if the following is correct?

We calculated GDD for all data from Jan 1 to the event day of year using average daily climate from 2011-2020? (I am still confused on exactly how GDD was calculated, and for how much of the data we were able to calculate it. So if you can spell out the steps in detail, that would help.)

b) in the paper for MAT, we write:

We estimated the mean annual temperature (MAT) for each provenance using the Climate Information Tool by Food and Agriculture Organization of the United Nations and ClimateWNA ...

When did we use FAO versus ClimateWNA? I think we should use the FAO whenever possible (and give more details on what period it averages over).

[alinazeng]

Hi Lizzie, Sorry I wasn't clear about GDD before, I would say that a) we first calculated GDD for all data from Jan 1 to the event day of year using the daily climate data for 2011, 2012, 2013, 2014... 2020, separately (having 10 GDD numbers), and then took the average of the 10 GDD (this is because it was difficult to get the GDD for the specific years the studies were conducted in, hence why we chose to took the average of the period 2011 to 2020, to be consistent with how we calculated climate overlap.) The calculation itself was just filtering out temperatures below 0 (our base), and added everything else up till the event DOY. Let me know if this makes more sense.

b) We used FAO for both European and North American sites. I also looked at North American sites on ClimateWNA to check that the FAO data was trustworthy (and it was). The historical period both tools average over is from 1961 to 1990, which is consistent with Aitken and Bammals 2016. Cheers,Alina On Tuesday, August 8, 2023 at 07:17:52 AM GMT+8, Elizabeth M Wolkovich @.***> wrote:

@alinazeng

a) Okay, can you tell me if the following is correct?

We calculated GDD for all data from Jan 1 to the event day of year using average daily climate from 2011-2020? (I am still confused on exactly how GDD was calculated, and for how much of the data we were able to calculate it. So if you can spell out the steps in detail, that would help.)

b) in the paper for MAT, we write:

We estimated the mean annual temperature (MAT) for each provenance using the Climate Information Tool by Food and Agriculture Organization of the United Nations and ClimateWNA ...

When did we use FAO versus ClimateWNA? I think we should use the FAO whenever possible (and give more details on what period it averages over).

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: @.***>

[lizzieinvancouver]

b) We used FAO for both European and North American sites. I also looked at North American sites on ClimateWNA to check that the FAO data was trustworthy (and it was). The historical period both tools average over is from 1961 to 1990, which is consistent with Aitken and Bammals 2016.

How wonderfully thorough of you (us)! I am just adjusting this now in the methods.

[lizzieinvancouver]

Okay, so our method for GDD I think could be summarized for a method section as:

We calculated GDD (with a base 0 C and no upper limit) for all data from Jan 1 to the event day of year using average daily climate for each year from 2011-2020 separately, then took the average of these 10 values. Given we could not get daily climate data for all years of data reported (i.e., some years earlier than our standardized fine-scale gridded data provided) we used the same 10 years of all climate data for all events at all sites.

[lizzieinvancouver]

Given the state of the results (seems like we have enough without GDD by MAT or latitude trends) and how we had to calculate GDD (not matching data perfectly by year would introduce some bias in the results we may need to dig deeper into to understand the results), and the complexity of spring phenology (which is not cued just by GDD), I suggest we remove GDD by MAT adn GDD by latitude trends from the ms. We can add them in later if reviewers ask about it.