senesced leaves present on the plant

[edenny]

Do we need a stage for senesced leaves (i.e. dead and still on the plant)? Phenophase in question is “All leaves withered”, when all leaves lose their green color but do not abcise, as in forbs.

[ramonawalls]

I am surprised to find that we don't have classes for senesced leaves still attached to the plant. Yes, I think we should add those. We can track the term creation here.

[edenny]

Looking back at the PhenoObs dataset, they have a phenophase that could include colored, dried, or dropped leaves. Is there a way to include senescing and senesced leaves under a trait that includes both?

[ramonawalls]

The use case here is that someone may not know if a leaf is completely dead or not. We can make a senescing leaf the parent class and create a more specific class for dead leaf, which is when it is dry and cannot revive.

[ramonawalls]

I reviewed the PO term for senescing true leaf:

A 'unfolded true leaf' that is in the 'vascular leaf senescent stage'.

And vascular leaf senescent stage:

A vascular leaf development stage (PO:0025570) that begins with the formation of a abscission zone at the base of a vascular leaf and ends with leaf separation and death.

[ramonawalls]

The definition encompasses dead and still on the plant, so what we need is a way to describe leaves that have already fallen. Such a definition would not be appropriate for PO (we already had a lot of discussion about plant structures that "are or were part or a plant", and we ended up making those classes in PPO instead of PO.

I suggest adding a class for "fallen vascular leaf" defined as "A vascular leaf that has senesced and fallen from the plant" plus a class for "vascular leaf fallen stage".

Then we can make a stage called "vascular leaf senescent or fallen stage" which is the union classe (with an OR) of "vascular leaf senescent stage" and "vascular leaf fallen stage".

@edenny can you confirm if that will work for you?

[edenny]

@ramonawalls, We already have a trait for abscised leaves, so I think leaves that have fallen from the plant is already covered. The case here is leaves still on the plant that have not abscised but are clearly dead and dried. Makes sense that these are part of leaf senescent stage, but can/should we include a specific class for dead leaf still attached under leaf senescent stage?

[ramonawalls]

@edenny thanks for pointing out the class for senesced leaves. I guess I was suggesting that we did not need a specific class for leaves that are dead but not senesced, because the border between dying and dead is hard to tell, while the boarder between attached to the plant and fallen is easy. Nonetheless, if you have a good use case for it, we can include it, it just would not be appropriate for the PO. So, the choice is between fine-grained classification versus botanical accuracy. As you are closer to the research, I leave the choice in your hands.

[edenny]

@ramonawalls The case here is a common phenophase for "All leaves withered", often used for forbs and grasses. Leaves don't abscise, but rather just dry up in place or lay on the ground still attached at the roots. I think using {senescing leaves} is misleading because there is absolutely no life left in them.

Would it be accurate to use {true leaves absent} for this phenophase? Or {unfolded true leaves absent}? The intent is to indicate no live leaves left on the plant, but I am not sure whether "true leaves" in the PO/PPO includes dead as well as live leaves.

If neither of those trait absences are appropriate, then perhaps would should add something for dead leaves still on the plant.

[jdeck88]

a farmer would just say the grass is ready for baling....

…

On Mon, Jan 22, 2024 at 9:34 AM edenny ***@***.***> wrote: @ramonawalls <https://github.com/ramonawalls> The case here is a common phenophase for "All leaves withered", often used for forbs and grasses. Leaves don't abscise, but rather just dry up in place or lay on the ground still attached at the roots. I think using {senescing leaves} is misleading because there is absolutely no life left in them. Would it be accurate to use {true leaves absent} for this phenophase? Or {unfolded true leaves absent}? The intent is to indicate no live leaves left on the plant, but I am not sure whether "true leaves" in the PO/PPO includes dead as well as live leaves. If neither of those trait absences are appropriate, then perhaps would should add something for dead leaves still on the plant. — Reply to this email directly, view it on GitHub <#72 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAIZ3RIXL47GI6C2RMML4GLYP2PKNAVCNFSM6AAAAAA5GTMHMSVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMYTSMBUGQ4DKMZVGA> . You are receiving this because you are subscribed to this thread.Message ID: ***@***.***>

-- John Deck (541) 914-4739

[ramonawalls]

Decision per discussion with Ellen today.

When leaves are still attached to the plant but all dead, there are no leaves present. True leaves are a non-cotyledon in the stages emerging to senescing, so once they are senesced, they are no longer "true leaves" (but dead leaves). We will add a comment to explain the use case in grasses and forbs and some trees.

The above paragraph means that plants where all the leaves have died but not fallen off should be scored as "true leaves absent", however, true leaves are also absent before any have formed (after cotyledon emergence but before true leaf emergence).

[ramonawalls]

Renaming issue to encompass any senesced structures.

We already have classes for senesced floral structures. These classes were added so that we could score grasses where the flowers die by stay on the plant, which is really the same use case as described here for leaves. We also have classes for senesced floral structures present, which make no sense to me, because a plant that has not flowers or only living flowers would have no senesced flowers present. I suspect we just added the absent classes following a pattern. I hope we are not using them to map data.

If we use the logic in my previous comment to map grasses with withered floral structures attached as "floral structures leaves absent", then we would have the same confusion I described above, where it could either mean no floral structures have developed yet OR all floral structures have died but remain on the plant OR all floral structures have died and fallen off the plant.

I think we need to hop on another call.

[ramonawalls]

Related to senesced structures, we created the property "is or was part of" in order to described senesced structures. However, the classes discussed in this issue are for senesced structures that are still part of the plant, so they don't need this new property.

@edenny are there phenology classes in our data where people score the percent or count of flowers or leaves that are on the ground? If so, that would call for "is or was part of", but I think we should maybe look to model it another way.

[edenny]

@ramonawalls There are cases, like in PEP, where leaf fall is scored as percentages. But this is mapped as {abscised leaf presence} which presumably already has a property of "was part of" instead of "is part of"....?

Your comment about senesced floral structures above is confusing. I think there are typos in some places (e.g. where you wrote present and meant absent). Could you check the wording on that and edit as necessary?

Looking back at PEP mappings, it seems we used {senesced flowers present} for the phenophase "End of flowering". That phenophase could technically also be mapped as {open flowers absent}, but I think {senesced flowers present} might more clearly imply that flowering has just recently ended which is what the "End of flowering" phenophase is intended to capture. This is useful in those datasets (like PEP) where there are only 2 datapoints--one for the start of flowering (open flowers present) and one for the end of flowering (senesced flowers present)--with nothing noted in-between. It is also useful for cases where the phenophase is phrased as "Last flower" where all but one flower are senesced. In that case "open flowers absent" would not be technically correct.

However, I am not sure if data users would agree on that--might be easier to just use the open flowers present or absent data points? Or maybe it's not a big deal and in the pre-processing for an analysis they could change any "senesced flowers present" to "open flowers absent" if that made it easier for whatever they wanted to do with the data?

Same logic could be applied to "All leaves withered". The phenophase is usually intended to capture the beginning of this stage when the last leaf has recently withered. I am starting to think maybe we should just use [abscised leaves present} for this and make a note for the trait that abscised leaves do not always fully separate from the plant, especially in forbs and grasses. If we think of it that way (as you described the other day-- the nutrients are cut off by an abscission layer but the leaf sometimes remains stuck on the stem) then I guess we will need to give abscised leaves the "is or was part of" property.

I like this solution for the Budburst data because the definition for "All leaves withered" states that "most or all leaves from this season have lost green color and are dried or dead". Since they use the term "most", then {true leaves absent} would not be technically correct in many cases.

[edenny]

That said, maybe I am overthinking this way too much?! We could also just say "End of flowering" or "Last flower" is close enough to {open flowers absent}, and "All leaves withered" or "Last leaf" is close enough to {true leaves absent}.

Maybe it depends on whether we decide to incorporate the upper/lower percents. With them, I would say the lower/upper bounds on {abscised leaves present} are 90% / 100% for "All leaves withered" (between 90 and 100% of leaves are abscised). However without incorporating them, it could be anywhere between 1 and 100% and the data user would have no idea. In that case it would be better to use {true leaves absent}....