Cold-deciduous climate enveloping: What is the correct value for the cndaysleafoff parameter?

[sharma-bharat]

Describe the issue

Relevent code: https://github.com/NGEET/fates/blob/main/biogeochem/EDPhysiologyMod.F90#L1170

Description of the Issue
During spinup (ELM-FATES-CNP) for Oak Ridge site (35° 54' N; 84° 20' W in southeastern United States) with only one cold deciduous PFT, after about 20 years, the GPP, TLAI, LEAFC, etc (Fig. 1) drops to zero does not recover.

Reason for dieback
A chilling event in year 19 triggered early senescence, followed in year 20 by a warm fall that delayed senescence (Fig. 2). Consequently, the total leaf-off duration since last leaf senescence (cndaysleafoff) exceeded 400 days therefore classifying the region as non-cold for cold-deciduous PFTs (see phenology code summary below and relevant documentation).

Summary of phenology code:

1. Cold Leaf-on timing:
   1.1. GDD > GDDcrit (Growing degree days need to be more than a threshold of 0°C)
   1.1.1. GDD is number if days > 0°C
   1.2. Number of chilling days (T<5°C) >= 1
   1.3. Leaves should not be already on
2. Cold Leaf Off timing:
   2.1. Number of cold days (T<7.5°C) in 10 days > 5
   2.2. Total number of leafon > 90 days
   2.3. Leaves should not be already off
3. Leaf Off: Cold Lifespan Threshold / Global implementation modifications
   3.1 If total # leafoff days > 400: declare the area as too warm for cold deciduous plants

Interim Solution:
To address the issue, we increased cndaysleafoff to ensure a more stable phenology pattern.
We updated:
currentSite%cndaysleafoff > 500 ! Set threshold to 500. Implemented in EDPhysiologyMod.F90
This adjustment prevents premature classification of the site as non-cold and restores periodic phenology behavior (Fig. 3).

In our case, the total number of leaf-off days needed to exceed 404 (Fig. 4).

We created this issue to discuss the ideal threshold for this parameter.

Figures:

Fig 1: LeafC drops to zero and never increases i.e. leaves never grow back.

Fig 2: Temperature in year 19 (top) and 20 (bottom). Red dashed vertical lines show 0°C

Fig 3: After the fix, we see normal phenology

Fig 4: Number of days since last senescence.

For additional plots of phenology variables:

• cndaysleafoff > 400
• cndaysleafoff > 500

Special thanks to @walkeranthonyp .

Thanks,
Bharat

Relevant log output

FATES tag

sci.1.68.2_api.31.0.0

Host land model tag

v2.1.0-13087-ge9afb7cb13

Machine

perlmutter

Other supported machine name

No response

Additional context

No response

[rgknox]

Thanks for submitting this @sharma-bharat

[mpaiao]

Thanks for the issue @sharma-bharat. I need to remember what is done for the drought deciduous, but there are similar situations there, in which the climate never gets really dry (or never gets moist). In these situations, the drought deciduous code forces abscission/flushing based on the time since last abscission/flushing event, so plants must abscise/flush leaves at least once a year (or every 13 months). I am wondering if we should something similar for cold-deciduous.