Multiyear simulation 2003-2022

[sumanager56]

I tried running the simulation with the current parameter using old synthetic multiyear data (generated using old version of actual 2021 data). Rainfed simulation seems to underestimate yield

[sumanager56]

Hi Dr. Shortridge, @julieshortridge, I used the updated synthetic data to run the simulation from 2003-2022. I also compared the 2021 result with the one from the previously calibrated model - the results look similar (Figure 1).

Figure 1. Simulation results for the year 2021 using the synthetic 2021 met data

Simulation results for the years 2003-2022 using the synthetic met data.

Irrigation was triggered when the soil water content is 70% of the total available water content, which is, I believe, what Laljeet used. These results are close to his reported values - total Irrigation applied was 150mm and 50 mm, respectively, in 2020 and 2021. N application is 180kg/ha

Figure 2

*Irrigation threshold=0.7Total Available Water Capacity**

[julieshortridge]

Thanks Suman, these are certainly looking better. The 70% threshold was kind of like the first step in his process (he was looking at soil moisture readings and forecasted rainfall simultaneously to decide when/how much to irrigate), but usually irrigation wasn't applied until later, so I'd use a threshold of 50% PAWC as the irrigation trigger.

I think at this point it might make sense to run three different management scenarios, particularly since we're seeing greater N leaching in the non-irrigated treatment, due to low plant growth presumably:

1. Rainfed
2. "Precision irrigation" - trigger at 50%, irrigate to 85% PAWC
3. "Conventional/heavy irrigation" - trigger at 60%, irrigate to 100% PAWC

[sumanager56]

@julieshortridge, Dr. Shortridge, I just tried plotting the yield data from USDA/NASS and it does seem like there's a similar depression in yields through the years 2005-2010 and 2020. Also, the yield started rising in 2010 and remained almost consistent until 2018 like in our simulated plot. 2019 data is missing

[julieshortridge]

Ok, thanks - this seems reasonable and was exactly what I was hoping to see.

[sumanager56]

Thanks, Dr. Shortridge. I did try to run some other management scenarios - the overall plot of irrigated vs non irrigated looked similar- however, the irrigation frequency seems quite reasonable (less frequent, heavier depth) when I trigger irrigation at 50% and irrigating upto 85% PAWC (around 25mm below FC). Here's how the irrigation events for 2020 and 2021 looks like.
Crop Name, Date, Day(crp), Day(cum), Irr(cm),Solute GMaizeD ,2020-06-11 , 53, 6372, 3.4, 0.000E+00 GMaizeD ,2020-07-05 , 77, 6396, 3.3, 0.000E+00 GMaizeD ,2020-07-13 , 85, 6404, 3.7, 0.000E+00 GMaizeD ,2020-07-19 , 91, 6410, 2.7, 0.000E+00 GMaizeD ,2020-07-24 , 96, 6415, 3.3, 0.000E+00 GMaizeD ,2020-07-30 , 102, 6421, 3.3, 0.000E+00 GMaizeD ,2021-05-24 , 35, 6719, 3.3, 0.000E+00 GMaizeD ,2021-06-30 , 72, 6756, 3.2, 0.000E+00

Irrigation triggered at 50%, irrigating to 85% PAWC

The depth of irrigations are close to Laljeet's result (4 events of around 2.5 cm in 2020 (July); 3 events of around 1.5 cm in 2021 (July)).

"Conventional/heavy irrigation" - trigger at 60%, irrigate to 100% PAWC

Crop Name, Date, Day(crp), Day(cum), Irr(cm),Solute GMaizeD ,2020-06-06 , 48, 6367, 4.0, 0.000E+00 GMaizeD ,2020-07-03 , 75, 6394, 4.6, 0.000E+00 GMaizeD ,2020-07-13 , 85, 6404, 4.9, 0.000E+00 GMaizeD ,2020-07-22 , 94, 6413, 5.2, 0.000E+00 GMaizeD ,2020-07-31 , 103, 6422, 1.8, 0.000E+00 GMaizeD ,2020-09-07 , 141, 6460, 4.6, 0.000E+00 GMaizeD ,2021-05-20 , 31, 6715, 4.3, 0.000E+00 GMaizeD ,2021-06-22 , 64, 6748, 3.4, 0.000E+00 GMaizeD ,2021-07-26 , 98, 6782, 0.3, 0.000E+00

[julieshortridge]

Thanks Suman. I was looking back through my notes from our meeting last November and the scenario we considered exploring is a calendar-based irrigation, where the irrigation applied is just the crop water demand minus any rainfall received in the previous week. So I think to do this, you would need to first look at the rainfall data and then once per week, sum the rainfall that occurred over the prior seven days and compare to crop ET demand (just use a basic K_c crop coefficient approach, don't use the actual ET from SWAP. From this, you should get an irrigation schedule and then you'll have to specify that by date for SWAP. Could you try this and see if the results end up differing from the sensor-based irrigation?

[sumanager56]

Thanks Suman. I was looking back through my notes from our meeting last November and the scenario we considered exploring is a calendar-based irrigation, where the irrigation applied is just the crop water demand minus any rainfall received in the previous week. So I think to do this, you would need to first look at the rainfall data and then once per week, sum the rainfall that occurred over the prior seven days and compare to crop ET demand (just use a basic K_c crop coefficient approach, don't use the actual ET from SWAP. From this, you should get an irrigation schedule and then you'll have to specify that by date for SWAP. Could you try this and see if the results end up differing from the sensor-based irrigation?

Thanks, Dr. Shortridge. Here, you want me to calculate the reference ET using weather data (penman-Montieth equation) and multiple by Kc for each growing stages, right?

[julieshortridge]

I would actually do something even simpler, and use a crop water use curve (see Figure 13 at the link below). This is closer to what I actually think is the common management practice - growers assume the plant needs a certain amount per week depending on growing season, and then irrigate when the rainfall received is less than that amount. I think the explicit use of reference ET data is pretty uncommon.

https://content.ces.ncsu.edu/soil-water-and-crop-characteristics-important-to-irrigation-scheduling

[sumanager56]

Sure, this seems reasonable and helpful!

[sumanager56]

Hi Dr. Shortridge,

I used the corn water requirement published in the document: "https://extensionpublications.unl.edu/assets/html/g1850/build/g1850.htm" and scheduled calendar-based irrigation by subtracting the total rainfall that occurred in the previous week from the total water requirement for a particular week (varies with the date from planting as in the document). Blue-calendar; Red-rainfed; Black=precision

[sumanager56]

@julieshortridge Hi Dr. Shortridge,

I ran the simulation just for 2021 and summarized the results as discussed yesterday. Please let me know if you want me to add something to this. Precision irrigation - Irrigation triggered at 50% depletion and applied upto 85% of FC. Calendar-based irrigation - Calculated by subtracting crop water requirement at a particular growth stage (weekly sum) from the total rainfall in the previous week.

<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:x="urn:schemas-microsoft-com:office:excel" xmlns="http://www.w3.org/TR/REC-html40">

  | Yield (bu/ac) | N_uptake (kg/ha) | N_leach (kg/ha) | Irrigation (cm) -- | -- | -- | -- | -- Rainfed | 201 | 168 | 94 | 0 Calendar | 191 | 164 | 109 | 23 Precision | 212 | 182 | 87 | 8

Rainfed irrigation plots

Calendar-based irrigation plots

Precision irrigation plots

[julieshortridge]

Thanks Suman. I think this is generally looking reasonable, except the irrigation depth for the calendar treatment is really high. What's the total growing season water demand from the UNL curves you're using? I think we may want to dial that back to adjust for Virginia climate, since it's wetter and more humid here you'd expect the evaporative demand to be lower.

[sumanager56]

Thank you, Dr. Shortridge. The total water demand from Apr-20 to Sep-13 is 36 inches. I will look into some papers and get more reasonable values for Virginia's climate.

[julieshortridge]

Oh, that's very high! I would expect something more like in the lower 20's. Even in the UNL document you link to above it says the range of crop water demand goes from 24" in eastern Nebraska (where it's more humid) to 28 in the more arid western part of the state. We certainly shouldn't be higher than that 24" value.

[sumanager56]

My bad-actually it is 28.9 inches but it is still high. I summed 0.8 inch/day instead of 0.08/day for the first 10 days (I corrected it just in my latest R file but not in my previous Excel sheet which I looked into right now). All the results above are using 28.9 inches.

[sumanager56]

@julieshortridge Hi Dr. Shortridge, I couldn't find the reference to corn water use curve for Virginia but did an estimation using reference ET (pulled from website 2003-2022 for the Suffolk area) and crop coefficients value for different growth stages. Using this, the total crop water requirement for corn was 23.5 inches.

The update yield values are shown below for 2021: <html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:x="urn:schemas-microsoft-com:office:excel" xmlns="http://www.w3.org/TR/REC-html40">

### 2021 Yield values   | Yield (bu/ac) | N_uptake (kg/ha) | N_leach (kg/ha) | Irrigation (cm) -- | -- | -- | -- | -- Rainfed | 201 | 168 | 94 | 0 Calendar | 203 | 175 | 96 | 15.89 Precision | 212 | 182 | 87 | 8

The multiyear graph looks somewhat different to the previous one with comparatively less leaching, more uptake for calendar-based system