parameter optimisaion via factorial

[frank0434]

• Prune the workflow to use only one DUL_LL method - the arbitrary one.
• Define the ranges of U and ConA
• Modify the Base.apsimx
• Decide whether gsmax and R50 need to be included or not.
• Define the ranges of gsmax and R50
• Modify the Base.apsimx

[frank0434]

[frank0434]

Actual soil evaporation is equal to potential evaporation up to a specified limit u (stage I) after which the evaporation rate becomes proportional to the square root of time (stage II):

$cumulative_second_stage = cona * t^0.5$

Methodology in soil-water-solute balance modelling: An evaluation of the APSIM-SoilWat and SWIMv2 models

[frank0434]

A conservative estimate of 2 is recommended for U as a modelling parameter. This is lower than most values commonly in use (3-8), however in agreement with recommendations from Yunusa et al. (1994) for Australian conditions.

ConA - winter: 3.0(±1.3) to 3.7(±1.4)
ConA - summer: 4.6(±0.4) to 5.3(±0.4)

SOIL EVAPORATION – HOW MUCH WATER IS LOST FROM NORTHERN CROP SYSTEMS AND DO AGRONOMIC MODELS ACCURATELY REPRESENT THIS LOSS?

[frank0434]

In New Zealand
U and α are soil texture properties of 9 mm and 4.4 mm/day-1/2 reported for a silt loam soil
(Jamieson et al., 1995a), similar

ES occurred when the SWC content of the top layer was greater than the lower limit. from Richard's thesis

[frank0434]

How to calculate Evaporation? water balance?
ConA must be smaller than U?
No ConA is in charge of the 2nd stage. U is the amount of water control when to go to the 2nd stage.

Solutions - CompositeFactor?

How can I check if the simulation use the correct values?
report only show the name of the value, not the exact values

[frank0434]

2430 simulations if using a setup like the snapshot
Simulations complete [824.97 sec]

[frank0434]

[Soil].SoilWater.SummerU = 1 to 9 step 1
[Soil].SoilWater.SummerCona = 1 to 6 step 0.5

Seems we can remove the combination below simulation ID 2000

Remove 2000

Seems ok to remove 3000

U ranges from 1 to 9
ConA ranges from 4 to 6
This reduced the number of simulations down to 2430 from 5346

The mismatch in the red circle could be caused by Winter u and cona
what is the definition of summer and winter here?
Answers:

 [Description("Start date for switch to winter parameters for soil water evaporation")]
        public string WinterDate { get; set; } = "1-Apr";

[Description("Start date for switch to summer parameters for soil water evaporation")]
        public string SummerDate { get; set; } = "1-Nov";

[frank0434]

with an arbitrary combination parameters:
U = 4
Cona = 4
Gsmax = 0.006
and the 54 KLR with RFV and 0.06 SKL

[frank0434]

with an arbitrary combination parameters:
U = 4
Cona = 4
Gsmax = 0.006
winterU = 1, 10
wintercona = 9, 0
and the 54 KLR with RFV and 0.06 SKL

lower the U and cona during winter definitely improve that funny bit in the red circle.
but guess the definition of summer of winter need to be tested as well?

change the summer definition from 1-Nov to 1-Sep