Add reallocation of assimilates from the storage roots

lpmorenoc commented 4 years ago

Some ideas:

[x] Define a cultivar coefficient to define the percentage of reallocation from the storage roots to the aboveground growth.
[x] Create more main stems that define a new demand for assimilates to the top growth. UPDATE: Although the idea was tested, that was overestimating LAI because the new main stems created had the same age and branching levels than the current main stems which increased the number of apices and then overestimated the effect of new main stems.
[x] Generate more leaves in existing stems from nodes where old leaves already fell. UPDATE: We decided to just increase the leaf appearance rate without affecting all nodes with leaves that had already died.

lpmorenoc commented 3 years ago

Reallocation index created (orange line) based on the difference of the water stress factor of the last 5 days and the water stress factor 6 to 20 days before: IF (STRESS20W < 0.5 .AND. (SUM(STRESSW(1:5))/5.0) > 0.5) THEN WFGREA = 1 + (SUM(STRESSW(1:5))/5.0)-(SUM(STRESSW(6:20))/15.0) ELSE WFGREA = 1 ENDIF

lpmorenoc commented 3 years ago

The model increases now the reserves after the release of the stress (see graph above) dca5f5f:

However, because the demand for this specific experiment is not high, then there is not change in the LAI or harvest weight:

lpmorenoc commented 3 years ago

If we increase the number of shoots when the water stress relief factor is greater than 1.2 (set that value just for testing) we get this curves: The increase of LAI also generates higher yields:

The water stress factor increase around 300 DAP (due to the increase of leaf growth):

And also the allocation of assimilates to the reserves for aboveground growth at the end of the growing season:

However, the stem growth is not properly represented: I will try to adjust the coefficients and also increase the initial number of shoots (PLPH)

lpmorenoc commented 3 years ago

Increasing the initial shoot number to 2: LAI:

Stem weight:

Harvest weight:

I need to verify the estimated value of the total biomass and the reserves because apparently the reserves are considered twice:

lpmorenoc commented 3 years ago

Using one initial shoot but PHOTO=V and EVAPO=H:

Before (also PHOTO=V and EVAPO=H):

lpmorenoc commented 3 years ago

Before change in leaf appearance: After change in leaf appearance using as multiplicative factor to increase leaf appearance WFGREA:

lpmorenoc commented 3 years ago

If we consider to increase the leaf appearance rate 3 times during the days WFGREA >1

That increase the LAI 250 DAP:

lpmorenoc commented 3 years ago

Increasing the leaf size (*3) during the water stress release in addition to the leaf appearance:

LAI is higher but not as high as expected.

lpmorenoc commented 3 years ago

Branching before any change: Reducing branching time by 3: But it does not have a high effect in the LAI:

lpmorenoc commented 3 years ago

If we increase the number of shoots (+1) when the water release stress factor is greater than 1.5, we overestimate the LAI:

If we modify the increase in the number of shoots as +0.5 when the water release stress factor is greater than 1.5, we still overestimate the number of apices at the end of the growing season:

Currently the shoots have reduced growth being the second shoot of the same size than the first one:

    SHGR(1) 0.0000000E+00   REAL(4) 
    SHGR(2) 1.000000    REAL(4) 
    SHGR(3) 0.9444444   REAL(4) 
    SHGR(4) 0.8888889   REAL(4) 
    SHGR(5) 0.8333333   REAL(4) 
    SHGR(6) 0.7777778   REAL(4) 
    SHGR(7) 0.7222222   REAL(4) 
    SHGR(8) 0.6666666   REAL(4) 
    SHGR(9) 0.6111111   REAL(4) 
    SHGR(10)    0.5555556   REAL(4) 
    SHGR(11)    0.5000000   REAL(4) 
    SHGR(12)    0.4444444   REAL(4) 
    SHGR(13)    0.3888889   REAL(4) 
    SHGR(14)    0.3333333   REAL(4) 
    SHGR(15)    0.2777778   REAL(4) 
    SHGR(16)    0.2222222   REAL(4) 
    SHGR(17)    0.1666667   REAL(4) 
    SHGR(18)    0.1111111   REAL(4) 
    SHGR(19)    5.5555522E-02   REAL(4)

If we reduce the growth of the second shoot as 0.75 of the first shoot:

    SHGR(1) 0.0000000E+00   REAL(4) 
    SHGR(2) 0.7500000   REAL(4) 
    SHGR(3) 0.7083333   REAL(4) 
    SHGR(4) 0.6666667   REAL(4) 
    SHGR(5) 0.6250000   REAL(4) 
    SHGR(6) 0.5833333   REAL(4) 
    SHGR(7) 0.5416666   REAL(4) 
    SHGR(8) 0.5000000   REAL(4) 
    SHGR(9) 0.4583333   REAL(4) 
    SHGR(10)    0.4166667   REAL(4) 
    SHGR(11)    0.3750000   REAL(4) 
    SHGR(12)    0.3333333   REAL(4) 
    SHGR(13)    0.2916667   REAL(4) 
    SHGR(14)    0.2500000   REAL(4) 
    SHGR(15)    0.2083333   REAL(4) 
    SHGR(16)    0.1666666   REAL(4) 
    SHGR(17)    0.1250000   REAL(4) 
    SHGR(18)    8.3333313E-02   REAL(4) 
    SHGR(19)    4.1666627E-02   REAL(4)

We still overestimate the LAI at the end of the growing season using a reduced shoot growth index (SHGR):

lpmorenoc commented 3 years ago

The leaf size is increasing but still have some restrictions due to the WFG: I will remove the effect of water stress during those days.

lpmorenoc commented 3 years ago

If I increase the potential leaf size 6 times instead of 3 times (similar effect of removing the water stress factor). The LAI improves: If we assume the potential leaf size is the maximum leaf size during the water release stress: Increasing leaf duration: Assuming two main stems from planting and potential leaf size = maximum leaf size during the water release stress::

lpmorenoc commented 3 years ago

If we define WFG as 0 (no stress-in the code 1) the LAI increase more 250 DAP (leaf size as maximum):

lpmorenoc commented 3 years ago

After discussing with James we agreed on:

[x] Modify potential leaf size curve when WFGREA >1 so it restarts from the maximum leaf size (at 900 GDD). this would work even for the days when WFGREA <1 once the model has already have a value of WFGREA >1. Issues:
- What if WFGREA >1 is before 900 GDD? Currently it assumes that will have the maximum leaf size but just for the days when WFGREA >1.
- What if WFGREA >1 happens more than once during the growing season?
[ ] To avoid multiple days with WFGREA >1, add this effect considering 30 days of stress instead of 20.
[ ] Allow reallocation of assimilates for longer time than currently, so WFGREA >1 until the CHO produced by the plant are enough to supply aboveground growth. If they are enough, WFGREA =1.

lpmorenoc commented 3 years ago

I am restarting the curve for leaf size. However, it is restarting every time WFGREA >1:

That could contribute to the high LAI at the end of the growing season:

lpmorenoc commented 3 years ago

Now leaf size curve just restarts once. However, there are still restrictions in the actual leaf size (grey line):

LAI now:

lpmorenoc commented 3 years ago

Modified the potential leaf size curve when WFGREA >1 so it restarts from the maximum leaf size (at 900 GDD). this would work even for the days when WFGREA <1 once the model has already have a value of WFGREA >1. About the issues:

If WFGREA >1 is before 900 GDD, it would just affect the potential leaf size for that specific day when WFGREA >1 but would not affect the whole curve (which is already increasing the leaf size).
If WFGREA >1 happens more than once during the growing season, we do not reset the curve, we just do the reset of the curve once.
For now I will keep the effect considering 20 days instead of 30 but I would check this later.

Reserves are used during the first release of water stress: Now I need to allow the use of the reserves for longer time.

lpmorenoc commented 3 years ago

Allowing the use of the reserves for longer time did not solve the issue because the demand for leaf growth was already reduced by the WFG: If I remove the effect of WFG after WFGREA >1:

New leaf size:

lpmorenoc commented 3 years ago

If the reallocation is avoided the first day the reserves are not used (enough production of assimilates), it just last few days after the release of the water stress and it does not allow the desired increase of the LAI because the water stress factor is reset and then the potential growth is reduced:

lpmorenoc commented 3 years ago

Allowing the model to use the storage root reserves even after 10 days of not being used and also eliminate the water stress factor for those days (so it allows the plant to recover and increase the demand) generates the following results:

However, if we just consider 5 days the reserves are not used the enough time to increase the demand and actual growth of the plant:

However, the simulated branching occurs late:

Reducing the time for branching does not improve the LAI:

lpmorenoc commented 3 years ago

As suggested by James, I removed the restrictions on leaf size due to water stress. I also removed this restrictions in the number of apices because we were underestimating the number of apices. As result: Number of apices:

LAI: We overestimate LAI. I also removed the restriction of water stress on stem weight:

I do not consider it is good idea to remove the restrictions of water stress in leaf size. Also, the information from the experiment we are testing, mentioned a reduction on the leaf age due to water stress. We currently do not have that effect.

lpmorenoc commented 3 years ago

If I reduce the leaf duration during water stress, which actually happened in the trial. I get the following result: It is possible to add a varietal coefficient that defines the sensitivity to leaf fall under water stress. Here an example adding a 0.5 factor instead of the whole effect: This generates an actual reduction of the storage root weight:

lpmorenoc commented 3 years ago

Allowing reallocation of assimilates and no restriction on leaf size due to water stress:

The overestimation of LAI is not related to the overestimation of apices:

I am removing the water stress for a long period:

New curve of leaf size without water restrictions:

lpmorenoc commented 3 years ago

Reducing leaf duration to 1000 degree days generates the following curve for LAI: We are still overestimating the increase on LAI due to the release of water stress. If we check the actual data from the trial that we are using as example (Veltkamp, experiment 1), they mention a reduction in the leaf size during water stress. However, their values are from the total plant leaf area and total number of leaves per plant. They also registered some increase in the leaf area but probably not as high as the one that we are generating: If we do not modify the potential leaf size after the release of water stress we get the following curve:

The reduction on storage root weight is higher at the beginning of the growing season probably due to the demand of aboveground growth:

lpmorenoc commented 3 years ago

If we reduce the potential leaf size to 400 cm2 instead of 500 cm2:

We actually reach the observed values of the trial, without increasing the leaf size; although there is some underestimation between 250-300 DAP after the release of water stress: This is the storage root weight: Stem weight:

lpmorenoc commented 3 years ago

Adding nitrogen:

lpmorenoc commented 3 years ago

We increase leaf size (*2) during the recovery face (days when WFGREA > 1.0). The reverse spill-over is allowed during the whole growing season when there are not enough assimilates for aboveground growth. This can create strong reduction in the storage root weight and it could need further revision (see issue: https://github.com/lpmorenoc/dssat-csm-os/issues/46).

lpmorenoc / dssat-csm-os

Add reallocation of assimilates from the storage roots #24