Is high latitude phenology working correctly?

[wwieder]

@swensosc and I have noticed this clear line in vegetation, with dead vegetation to the south of ~65 N

Here's another look at the issue, highlighting low LAIs a then end of a pSASU spinup
in LMWG_dev#164.

Given this threshold, we initially blamed the calculation of critical day length being calculated here and pasted below that is used for offset, but we think is is being calculated correctly (ranging from 11-15 hours between 45-65 deg)

    case(critical_daylight_depends_on_lat)
        ! Critical daylength is higher at high latitudes and shorter
        ! for temperatre regions
        crit_daylat=params_inst%crit_dayl_at_high_lat-params_inst%crit_dayl_lat_slope* &
                    (critical_offset_high_lat-abs(grc%latdeg(g)))
        if (crit_daylat < crit_dayl) then
           crit_daylat = crit_dayl !maintain previous offset from White 2001 as minimum
        end if

My next question was if offset trigger is being used correctly (code here) but this also seems OK with ws_flag set to zero after the summer solstice occurs

               ! only begin to test for offset daylength once past the summer sol
               if (ws_flag == 0._r8 .and. dayl(g) < crit_daylat) then
                  offset_flag(p) = 1._r8
                  offset_counter(p) = ndays_off * secspday
                  prev_leafc_to_litter(p) = 0._r8
                  prev_frootc_to_litter(p) = 0._r8
               end if
            end if

This makes me wonder if it's potentially an phenology onset issue? Namelist default are using onset_thresh_depends_on_veg = .true., which I think triggers this code. Nothing odd jumps out when I look at this code, so maybe more on this to come on this... but anyone who's interested in or familiar with the phenology code, feel free to help.

[wwieder]

I found this note in Leah's paper:

"Based on observations at high latitudes, 15 h is a more accurate timing for senescence above 65∘ N (REFS). We scale the photoperiod threshold linearly along a latitudinal gradient from 65∘ N until ∼11 h at 45∘ N such that the temperate latitudes retain the offset timing determined by White et al. (1997)"

Which made be wonder if we should replace this equation:

    if (crit_daylat < crit_dayl) then
           crit_daylat = crit_dayl !maintain previous offset from White 2001 as minimum
    end if

with this, to bound crit_daylay between 11 and 15 hours:

crit_daylat = max(crit_dayl, min(crit_daylat, params_inst%crit_dayl_at_high_lat))

I'm not sure it will necessarily solve the dead veg line we're seeing south of 65N, but it would help the code be consistent with Leah's 2021 paper?

[wwieder]

Two more thoughts on ways to address this:

• set critical_offset_high_lat = 60 to see if we can get longer growing seasons by delaying offset)?
• decrease ndays_on for arctic grasses (and shrubs?), I think Katya found that alpine tundra greened up much faster that we're currently allowing it to in CLM, which makes sense given their short growing season.
  • Default in CLM6 is 30 (although Katya's paper claims it was 10 in CLM5?)
  • Katya used 21 (moist meadow) - 25 (dry meadow) - 28 (wet meadow)
• check Jmax calculations in LUNA. There's more in Leah's paper about
  • "Day length scaling for photosynthetic parameters" and
  • "Temperature acclimation" (which I think we chose not to implement)

[ekluzek]

#1409 talks about this same issue and has some proposed solutions.

[ekluzek]

@wwieder is this something that's too big of answer change to address now? Is this a ctsm6.1 research question to address?

[wwieder]

@wwieder is this something that's too big of answer change to address now? Is this a ctsm6.1 research question to address?

Good question, @ekluzek, I'm not really sure how big of an answer change this would make. Tests in LMWG_dev#168 will start providing insights. Also nice find digging up #1409!