Modified Suleiman-Ritchie soil evaporation

[chporter]

Modified as per Ayman Suleiman at Jan 2025 DSSAT sprint.
-- ES occurs from top 15 cm only. Re-instate UPFLOW.
-- Uses MESEV = "M" (for Modified Suleiman-Ritchie)

This runs for all DSSAT experiments and treatments. Soil evaporation is considerably lower than all other methods. I ran the default soil evaporation in each FileX vs R, S, and M method override using external simulation control file.

Average ratio over all experiments / treatments
0.98393 | Ritchie : default
1.183594 | Suleiman-Ritchie : default
0.716105 | Modified Suleiman-Ritchie : default
0.62579 | Modified Suleiman-Ritchie : Suleiman-Ritchie

This needs to be thoroughly vetted by Ayman, Gerrit, Fabio, Bruce before accepting the pull request.

[aymansuleiman]

Dear Mrs. Porter, Many thanks for your trying the soil evaporation from the 15 cm depth only.  I suggest that we try soil evaporation from 30 cm instead from om 15 because the simulated SOIL EVAPORATION WAS low. Kind regards, Ayman On Tuesday, February 25, 2025 at 07:12:40 PM EST, Cheryl Porter ***@***.***> wrote: Closed #471. — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because your review was requested.Message ID: ***@***.***>

[chporter]

@aymansuleiman @GerritHoogenboom
I changed the soil depth to 30 cm and it improved results (i.e., came closer to what we expected), but still only evaporates about 86% on average of Ritchie method and about 72% of the original Suleiman-Ritchie method.
ET_comparison_30.xlsx

[aymansuleiman]

Many thanks for changing the soil depth to 30 cm. For sure we need to change the depth one more time. Please, can you change the depth to 45 m. On Sunday, March 2, 2025 at 07:31:12 PM EST, Cheryl Porter ***@***.***> wrote: @aymansuleiman @GerritHoogenboom I changed the soil depth to 30 cm and it improved results (i.e., came closer to what we expected), but still only evaporates about 86% on average of Ritchie method and about 72% of the original Suleiman-Ritchie method. ET_comparison_30.xlsx — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***> chporter left a comment (DSSAT/dssat-csm-os#471) @aymansuleiman @GerritHoogenboom I changed the soil depth to 30 cm and it improved results (i.e., came closer to what we expected), but still only evaporates about 86% on average of Ritchie method and about 72% of the original Suleiman-Ritchie method. ET_comparison_30.xlsx — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>

[chporter]

I modified the method with the following:

• it's now more generic and the depth of evap can be changed with one line of code to simplify exploring the range of options.
• Modified S-R calls UPFLOW routine to move water up, but also takes water from layers up to specified depth. The evaporation is not (yet) added to UPFLOW which is used to move soil N. We probably want to make this change before finalizing it.
• I corrected an error where the mulch evaporation was not taken into account prior to calling the routine previously, so these results are slightly different.

The 15cm depth and even going up to 30 cm is not enough to increase evaporation to more than the Ritchie method (on average). I also tried 50 cm and it does increase evaporation slightly (+4% over the Ritchie method).

I also tested the model with a depth of 200 cm which should give similar results to the original Suleiman-Ritchie model. It is very similar, with differences due to the addition of UPFLOW.

The attached plot shows how the various methods compare to the Ritchie method for all the DSSAT experiments and treatments.

The source code currently has the 50 cm depth but we need to discuss this prior to merging the pull request.

[aymansuleiman]

Many thanks for you traying the 50 cm depth. I agree with you that 15 and 30 cm depths are not suitable because their evaporation is lower than Ritch method. 50 cm depth is closer as it evaporates on average about 90% of Ritchie method. Using the 50 cm is more appropriate.   On Friday, March 14, 2025 at 12:18:18 AM EDT, Cheryl Porter ***@***.***> wrote: I modified the method with the following: - it's now more generic and the depth of evap can be changed with one line of code to simplify exploring the range of options. - Modified S-R calls UPFLOW routine to move water up, but also takes water from layers up to specified depth. The evaporation is not (yet) added to UPFLOW which is used to move soil N. We probably want to make this change before finalizing it. - I corrected an error where the mulch evaporation was not taken into account prior to calling the routine previously, so these results are slightly different. The 15cm depth and even going up to 30 cm is not enough to increase evaporation to more than the Ritchie method (on average). I also tried 50 cm and it does increase evaporation slightly (+4% over the Ritchie method). I also tested the model with a depth of 200 cm which should give similar results to the original Suleiman-Ritchie model. It is very similar, with differences due to the addition of UPFLOW. The attached plot shows how the various methods compare to the Ritchie method for all the DSSAT experiments and treatments. The source code currently has the 50 cm depth but we need to discuss this prior to merging the pull request. image.png (view on web) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***> chporter left a comment (DSSAT/dssat-csm-os#471) I modified the method with the following: - it's now more generic and the depth of evap can be changed with one line of code to simplify exploring the range of options. - Modified S-R calls UPFLOW routine to move water up, but also takes water from layers up to specified depth. The evaporation is not (yet) added to UPFLOW which is used to move soil N. We probably want to make this change before finalizing it. - I corrected an error where the mulch evaporation was not taken into account prior to calling the routine previously, so these results are slightly different. The 15cm depth and even going up to 30 cm is not enough to increase evaporation to more than the Ritchie method (on average). I also tried 50 cm and it does increase evaporation slightly (+4% over the Ritchie method). I also tested the model with a depth of 200 cm which should give similar results to the original Suleiman-Ritchie model. It is very similar, with differences due to the addition of UPFLOW. The attached plot shows how the various methods compare to the Ritchie method for all the DSSAT experiments and treatments. The source code currently has the 50 cm depth but we need to discuss this prior to merging the pull request. image.png (view on web) — Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you were mentioned.Message ID: ***@***.***>