search

APSIMInitiative / ApsimX

ApsimX is the next generation of APSIM
http://www.apsim.info
Other
134 stars 162 forks source link

New Sorghum model in PMF #572

Closed aldoherty closed 1 year ago

sarahcleary commented 4 years ago

@jbrider - is this ready for review?

peter-devoil commented 4 years ago

Am I right to find that when sowing sorghum; row spacing is in m, while wheat (and others?) is in mm?

jbrider commented 4 years ago

There was an if statement that checked if the row spacing was > 1000 it would divide it by 1000... which was just ugly. I'm guessing it was there to be consistent with sowing scripts... which we didn't realise at the time.

Sounds like we should put it back?

On Wed, Nov 27, 2019 at 5:21 PM Peter de Voil notifications@github.com wrote:

Am I right to find that when sowing sorghum; row spacing is in m, while wheat (and others?) is in mm?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/APSIMInitiative/ApsimX/issues/572?email_source=notifications&email_token=AAE3G7CP6X43EGSN53ITSKTQVYNYVA5CNFSM4BWGKYT2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEFIRY5Q#issuecomment-558963830, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAE3G7ADOSPBBXXVEJLAMJTQVYNYVANCNFSM4BWGKYTQ .

hol430 commented 4 years ago

I vote remove the check (or keep it removed if it's already gone). If we're using the wrong units to sow the plant then that's a problem with our usage of the sowing rule, not a problem with the model.

hol353 commented 4 years ago

We can't have the units for row spacing in m in sorghum and mm in every other crop model. If it was my model I'd change the row spacing units to mm.

jbrider commented 4 years ago

easy change is to add the hack back in, but I agree it would be better to change the row spacing calcs to use mm.

On Thu, Nov 28, 2019 at 8:10 AM Dean Holzworth notifications@github.com wrote:

We can't have the units for row spacing in m in sorghum and mm in every other crop model. If it was my model I'd change the row spacing units to mm.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/APSIMInitiative/ApsimX/issues/572?email_source=notifications&email_token=AAE3G7HTW2NPM2MORWBLVJ3QV3V3RA5CNFSM4BWGKYT2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEFK34HQ#issuecomment-559267358, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAE3G7DSAUOGMB7YGALJOOLQV3V3RANCNFSM4BWGKYTQ .

hol430 commented 4 years ago

I'll change sorghum to use mm when I get the chance. Seems like the best solution.

peter-devoil commented 4 years ago

Would be good to include units when reporting it. ATM it's unitless, which only adds to the confusion.

sarahcleary commented 4 years ago

@jbrider - is this ready for review?

jbrider commented 4 years ago

yes

On Thu, Dec 5, 2019 at 2:11 PM sarahcleary notifications@github.com wrote:

@jbrider https://github.com/jbrider - is this ready for review?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/APSIMInitiative/ApsimX/issues/572?email_source=notifications&email_token=AAE3G7EAVMMVTWW2R77YWGTQXB5NJA5CNFSM4BWGKYT2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEF7NIZY#issuecomment-561960039, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAE3G7BPPDS4B2X7SJN2GSDQXB5NJANCNFSM4BWGKYTQ .

sno036 commented 4 years ago

I am getting the error below with the Sorghum validation simulation

image

hol430 commented 4 years ago

Have you run all simulations in the file?

sno036 commented 4 years ago
sno036 commented 4 years ago

In reply to @hol430 - Yes

jbrider commented 4 years ago

Thanks Val - appreciate the feedback.

Working on examples and more sensibility tests at the moment. There should be 1 large one in there, will see where it got to.

It doesn't intercrop above ground yet - as the old model didn't either. Is that's what Micromet is supposed to do?

It should intercrop with roots ok due to the way that is implemented through the strangely named algorithm that makes debugging much more painful.

On Wed, Dec 11, 2019 at 2:30 PM Val Snow notifications@github.com wrote:

  • needs a simulation example
  • Surface OrganicMatter is sitting at the God level - is the C/N balance working properly?
  • where are the sensibility tests?
  • all the models need to intercrop so I dropped a ryegrass model into the simulation with a large initial biomass. It should slowed the Sorghum development because it would have shaded it but that didn't happen. Has the interface been implemented properly? The Ryegrass is intercepting 60% of the radiation initially. (I tried to use Slurp for this but something is wrong with the cultivar thing with two crops in the same zone)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/APSIMInitiative/ApsimX/issues/572?email_source=notifications&email_token=AAE3G7BMWFM2KLJ37CPUNI3QYBUF5A5CNFSM4BWGKYT2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGR226I#issuecomment-564374905, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAE3G7GTML5X7UTHCX3LZA3QYBUF5ANCNFSM4BWGKYTQ .

sno036 commented 4 years ago

Sorry - was in a bit of a rush yesterday (and grumpy for various reasons). Heading into a series of meetings soon so the attached is not well documented sorry.

All simulations should have MicroClimate (yes - I used the old name) in them and preferably at the God level. SoilArbitrator deals with below-ground competition.

Apologies for the sad state of the simulation! Test.zip

Have just seen that @hol430 posted a fix for the cultivar issue in #4517

jbrider commented 4 years ago

@Snow, Val Val.Snow@agresearch.co.nz I spoke to Hamish about implementing MicroClimate and unfortunately it won't work at this stage with the Sorghum model. We implement the evapotranspiration (I think I got that right?) in a different way that is not compatible with how MicroClimate manages it.

While I don't think we will be able to resolve this in the short term, it is our goal to find a way to solve this issue in the long term.

On Thu, Dec 12, 2019 at 6:12 AM Val Snow notifications@github.com wrote:

Sorry - was in a bit of a rush yesterday (and grumpy for various reasons). Heading into a series of meetings soon so the attached is not well documented sorry.

  • "Solo" is just one of the validation simulations - I know zilch about Sorghum beyond that it makes gluten-free wheatbix so didn't want to much around with dates etc.
  • "WithRyeGrass" has an AGPRyegrass in the zone and it should compete with the Sorghum but I am not convinced that it is. Ryegrass is not the best test here because it need to be harvested to keep it growing actively and I can't get it to initialise at a high initial biomass (issue #4516 https://github.com/APSIMInitiative/ApsimX/issues/4516)
  • "SlurpSolo" was a start of replacing Ryegrass with something that we can control the shading etc. properly
  • "WithSlurp CultivarFail" was popping Slurp into "Solo" but something has gone wrong in that it will only find cultivars for one of the crops - I've put in an issue ( on this as well #4517 https://github.com/APSIMInitiative/ApsimX/issues/4517)
  • "GrazedBySheep" was the start of a test of the implementation of the interface that all crop models need to have for 'damage' (stock, pests, ...). Something is not right here in that the stock aren't actually going into the Sorghum paddock (not sure if this is my fault but the same script work with wheat and should work here) so the growth of the grazed ("Paddock") and not grazed zones should be the same but they are vastly different.

All simulations should have MicroClimate (yes - I used the old name) in them and preferably at the God level. SoilArbitrator deals with below-ground competition.

Apologies for the sad state of the simulation!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/APSIMInitiative/ApsimX/issues/572?email_source=notifications&email_token=AAE3G7HAXCY2JXWGKUFBRRLQYFCUXA5CNFSM4BWGKYT2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEGUNL6I#issuecomment-564712953, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAE3G7AJ5WFLFPP7UX6F5BLQYFCUXANCNFSM4BWGKYTQ .

sno036 commented 4 years ago

@jbrider - thanks for that clarification. I guess this will need more thought as the Next Gen design has always been centered on the use of MicroClimate to do the above-ground "competition" including considering SurfaceOrganicMatter and partitioning to potential evaporation from the soil surface. All the other crops require it and I suspect that there will be challenges in doing rotations if Sorghum doesn't use MicroClimate. Needs discussion. @hut104, @HamishBrownPFR

sno036 commented 4 years ago

@peter-devoil asked

do you mean rotations or intercrops (or both)? Not many people do anything fancy with sorghum (some grow it beneath trees / windbreaks), but a lot of my colleagues do maize / bean intercrops, and that would be a real headache to a) lose that facility, or b) make it a special case that required specific construction.. P

I meant both. We really don't want to deal with special cases (as much as possible!) as it is hard enough as it is at the moment! MicroClimate and SoilArbitrator will deal with the maize-bean competition in the intercropping - not necessarily perfectly of course but they can always be improved. I was also concerned about being able to do rotations of Sorghum with other crops if Sorghum is interacting differently to other crops. I don't know for sure that there will be a problem but some testing would be needed.

When I get a bit more time I'll try to get back to the grazing test as it is important that all the models interact with Stock in the same way - no if this crop then do this way (ill)logic - and something wasn't working properly on my first go. The issue with that simulation might not have been necessarily a Sorghum issue - I need to spend more time with it I think but swapping from Wheat grazign ot Sorghum grazing should be dead simple (in theory!).

sarahcleary commented 4 years ago

@hut104 and @HamishBrownPFR - the @APSIMInitiative/reference-panel have proposed you as reviewers of this model. Please let me know if you are able to conduct this review. As you are aware, we are conducting the process solely via GitHub. When you are ready to review, please refer to the Reviewer's Instructions and commence review: http://www.apsim.info/Documentation/TechnicalandDevelopment/ReviewersInstructions.aspx Please let me know if these Instructions need to be amended in anyway.

sarahcleary commented 4 years ago

@hut104 and @HamishBrownPFR - can you please confirm you are willing to undertake this review?

sarahcleary commented 4 years ago

@hut104 and @HamishBrownPFR have confirmed they will undertake the review. Commence this month.

jbrider commented 4 years ago

After having a conversation with @Huth, Neil (A&F, Toowoomba) Neil.Huth@csiro.au I will be making some changes to improve the models ability to integrate with other crops. The model currently performs the same as it did in Apsim Classic.

On Mon, Feb 3, 2020 at 12:18 PM sarahcleary notifications@github.com wrote:

@hut104 https://github.com/hut104 and @HamishBrownPFR https://github.com/HamishBrownPFR have confirmed they will undertake the review. Commence this month.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/APSIMInitiative/ApsimX/issues/572?email_source=notifications&email_token=AAE3G7DTXNSPMRX6MXCNGMTRA55HHA5CNFSM4BWGKYT2YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEKSJBFY#issuecomment-581210263, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAE3G7DIJVMPXRKQLOM2JT3RA55HHANCNFSM4BWGKYTQ .

sarahcleary commented 4 years ago

@jbrider - can you please provide an update in the next RP meeting?

sarahcleary commented 4 years ago

@jbrider - can you please provide an update in the next RP meeting?

sarahcleary commented 4 years ago

@jbrider - are you able to provide an update? are you still working on updates from @hut104 and @HamishBrownPFR's feedback?

jbrider commented 4 years ago

no change as I have been doing project work. Hope to get back to it soon.

JJguri commented 4 years ago

We are calibrating the Sorghum model for forage crops under Argentinian environments. We started working from the apsimx file which is under review. We have several measured datasets at the field level and remote sensing datasets which provides also observations for the calibration. Although most of the output variables were easy to find, we are still looking for the following ones:

Could someone help us?

jbrider commented 4 years ago
JJguri commented 4 years ago
hut104 commented 4 years ago

1) Allocation will include DM retranslocated and/or reallocated from senscence and so will be greater than actual growth. When demands are low, it may also be lower than potential photosynthesis (ie DM supply). I think the bigger issue will be how you compare one model estimate of daily growth with another (ie the remote sensing is an estimate using a model based on spectral properties etc - not an actual measurement). 2) Height should be cultivar specific and so you will need to incorporate this into your cultivar specification. It should also be realistic as it will affect water demand and light interception calculations when using micromet, especially in complex planting geometries. 3) We have no mechanistic models for DM content in most models. 4) Dead leaf number should be able to be estimated from SLAI and TLAI in the sorghum model if you invert the plant leaf area equations. 5) In most crops, the ratio of water demand (ie potential ET) and actual ET is Leaf.FW. Sorghum uses a different approach for ET and so may be different, but the standard within APSIM is the FW property. Alternatively you could use the water demand property of leaf to get the potential plant ET. 6) Panicle weight will need to be determined according to that used in your field protocol. It may be simply Rachis weight if that was the meaning in the field. Otherwise add Rachis to Grain.

From: Jonathan Ojeda notifications@github.com Sent: Sunday, 21 June 2020 10:04 AM To: APSIMInitiative/ApsimX ApsimX@noreply.github.com Cc: Huth, Neil (A&F, Toowoomba) Neil.Huth@csiro.au; Mention mention@noreply.github.com Subject: Re: [APSIMInitiative/ApsimX] New Sorghum model in PMF (#572)

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/APSIMInitiative/ApsimX/issues/572#issuecomment-647059613, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AC2UVWSLTYSVIPRXC5ZLR23RXVFCTANCNFSM4BWGKYTQ.

jbrider commented 4 years ago

Thanks @hut104.

  1. you can't access yesterday's values directly - you can use a manager script to store it in a local variable, and then expose a variable that calculates today's value - yesterday's value to get the growth rate. As Neil pointed out, you will need a way to compare the 2 different measures.
  2. At 2700mm, it is not going to match anything in the existing cultivar's - you will need to specify the relationship as Neil suggested as a Cultivar property. The current calculation is a function of the Stem's biomass.
  3. Is the DryMatterContent you're measuring the dry weight of the biomass samples - the model reports all of it's biomass values as the dry weight value.
JJguri commented 4 years ago

Thanks @hut104

1- We have several datasets: field measurements, sensor data from field and sensor data from the satellite. So, the idea is to do three comparisons, APSIM vs field observations, RS model vs field observations and APSIM vs RS model. In the first comparison, we test APSIM, in the second we test the RS model and in the third, we compare both models.

4- Are you talking about SenescedLAI [Leaf].LAIDead and Total LAI [Leaf].LAITotal? I can report these variables and LiveLAI [Leaf].LAI however I cannot find the equations you mentioned to estimate the number of dead leaves, could you please be more specific?

5- For Sorghum you have: [Leaf].ExpansionStress.WaterStressEffect and [Leaf].Photosynthesis.FW instead of a single FW. I realized that [Leaf].WaterAllocation is equal to [Leaf].Transpiration. So which variable is potential transpiration? As you suggested, I checked [Leaf].WaterDemand as an estimation of potential plant ET and it showed really low values to be mm, e.g. for a given day during the season the crop has 4 mm of transpiration and the water demand was 0.23. Which is the difference between [Leaf].WaterDemand and [Soil].SoilWater.Eo if you want to get an estimation of maximum atmosphere demand or potential ET?

Thanks @jbrider

3- The Dry Matter Content is the content of dry matter (from 0 to 100%) in the Aboveground Wt. It is not the weight of the abovegroundWt.

HamishBrownPFR commented 4 years ago
  1. you can't access yesterday's values directly - you can use a manager script to store it in a local variable, and then expose a variable that calculates today's value - yesterday's value to get the growth rate. As Neil pointed out, you will need a way to compare the 2 different measures.

I have a pull request open at the moment that will give the met class a set of Met variables for today and yesterday as well. Not sure why it is failing though (#5323)

JJguri commented 4 years ago
  1. you can't access yesterday's values directly - you can use a manager script to store it in a local variable, and then expose a variable that calculates today's value - yesterday's value to get the growth rate. As Neil pointed out, you will need a way to compare the 2 different measures.

I have a pull request open at the moment that will give them met file a set of Met variables for today and yesterday as well. Not sure why it is failing though

For sure, that will work for climate variables, but what about crop output variables? It can be adapted for these also? We are trying to get the difference between today and yesterday AbovegrounWt as an output in the DailyReport.

hut104 commented 4 years ago

I think the easiest way is to determine the true meaning of the remote sensing estimate of growth rate (eg is it daily photosynthesis, or change in daily above ground biomass – ie the net effect of above ground growth and loss from senescence etc) and then get the existing data from arbitrator if that suits.

From: Jonathan Ojeda notifications@github.com Sent: Monday, 22 June 2020 8:49 AM To: APSIMInitiative/ApsimX ApsimX@noreply.github.com Cc: Huth, Neil (A&F, Toowoomba) Neil.Huth@csiro.au; Mention mention@noreply.github.com Subject: Re: [APSIMInitiative/ApsimX] New Sorghum model in PMF (#572)

  1. you can't access yesterday's values directly - you can use a manager script to store it in a local variable, and then expose a variable that calculates today's value - yesterday's value to get the growth rate. As Neil pointed out, you will need a way to compare the 2 different measures.

I have a pull request open at the moment that will give them met file a set of Met variables for today and yesterday as well. Not sure why it is failing though

For sure, that will work for climate variables, but what about crop output variables? It can be adapted for these also? We are trying to get the difference between today and yesterday AbovegrounWt as an output in the DailyReport.

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/APSIMInitiative/ApsimX/issues/572#issuecomment-647191523, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AC2UVWU7WJOPVCS6YKRDK63RX2E4HANCNFSM4BWGKYTQ.

JJguri commented 4 years ago

I think the easiest way is to determine the true meaning of the remote sensing estimate of growth rate (eg is it daily photosynthesis, or change in daily above ground biomass – ie the net effect of above ground growth and loss from senescence etc) and then get the existing data from arbitrator if that suits. From: Jonathan Ojeda notifications@github.com Sent: Monday, 22 June 2020 8:49 AM To: APSIMInitiative/ApsimX ApsimX@noreply.github.com Cc: Huth, Neil (A&F, Toowoomba) Neil.Huth@csiro.au; Mention mention@noreply.github.com Subject: Re: [APSIMInitiative/ApsimX] New Sorghum model in PMF (#572) 1. you can't access yesterday's values directly - you can use a manager script to store it in a local variable, and then expose a variable that calculates today's value - yesterday's value to get the growth rate. As Neil pointed out, you will need a way to compare the 2 different measures. I have a pull request open at the moment that will give them met file a set of Met variables for today and yesterday as well. Not sure why it is failing though For sure, that will work for climate variables, but what about crop output variables? It can be adapted for these also? We are trying to get the difference between today and yesterday AbovegrounWt as an output in the DailyReport. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#572 (comment)>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AC2UVWU7WJOPVCS6YKRDK63RX2E4HANCNFSM4BWGKYTQ.

The true meaning of the remote sensing estimate is the monthly actual aboveground growth rate (kg DM/ha/day).

I reported allocated DM doing the sum of different plant components and using the AllocatedDM for each of them. I compared cumulative TotalGrowthRate = [Leaf].Allocated.Wt + [Stem].Allocated.Wt + [Grain].Allocated.Wt + [Rachis].Allocated.Wt + [Root].Allocated.Wt and [Arbitrator].DeltaWt, and they are the same. So I discounted the [Root].Allocated.Wt from TotalGrowthRate and I got the AboveGroundGrowthRate in blue. Untitled

hut104 commented 4 years ago

OK, so we are talking about growth rate, not a change in biomass. So what you suggest probably be more of what you need rather than using changes in above ground biomass, because the later can also include losses of biomass due to senescence/detachment and also effective loss of mass in oil crops if DM is retranslocated into organs with an oil content and hence a higher carbon density. Growth is not the only process that can cause a change in mass. I would continue looking into outputs along the lines of what you suggest.

Neil.

From: Jonathan Ojeda notifications@github.com Sent: Monday, 22 June 2020 4:13 PM To: APSIMInitiative/ApsimX ApsimX@noreply.github.com Cc: Huth, Neil (A&F, Toowoomba) Neil.Huth@csiro.au; Mention mention@noreply.github.com Subject: Re: [APSIMInitiative/ApsimX] New Sorghum model in PMF (#572)

I think the easiest way is to determine the true meaning of the remote sensing estimate of growth rate (eg is it daily photosynthesis, or change in daily above ground biomass – ie the net effect of above ground growth and loss from senescence etc) and then get the existing data from arbitrator if that suits. From: Jonathan Ojeda notifications@github.commailto:notifications@github.com Sent: Monday, 22 June 2020 8:49 AM To: APSIMInitiative/ApsimX ApsimX@noreply.github.commailto:ApsimX@noreply.github.com Cc: Huth, Neil (A&F, Toowoomba) Neil.Huth@csiro.aumailto:Neil.Huth@csiro.au; Mention mention@noreply.github.commailto:mention@noreply.github.com Subject: Re: [APSIMInitiative/ApsimX] New Sorghum model in PMF (#572https://github.com/APSIMInitiative/ApsimX/issues/572) 1. you can't access yesterday's values directly - you can use a manager script to store it in a local variable, and then expose a variable that calculates today's value - yesterday's value to get the growth rate. As Neil pointed out, you will need a way to compare the 2 different measures. I have a pull request open at the moment that will give them met file a set of Met variables for today and yesterday as well. Not sure why it is failing though For sure, that will work for climate variables, but what about crop output variables? It can be adapted for these also? We are trying to get the difference between today and yesterday AbovegrounWt as an output in the DailyReport. — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub<#572 (comment)https://github.com/APSIMInitiative/ApsimX/issues/572#issuecomment-647191523>, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AC2UVWU7WJOPVCS6YKRDK63RX2E4HANCNFSM4BWGKYTQ.

The true meaning of the remote sensing estimate is the monthly actual aboveground growth rate (kg DM/ha/day).

I reported allocated DM doing the sum of different plant components and using the AllocatedDM for each of them. I compared cumulative TotalGrowthRate = [Leaf].Allocated.Wt + [Stem].Allocated.Wt + [Grain].Allocated.Wt + [Rachis].Allocated.Wt + [Root].Allocated.Wt and [Arbitrator].DeltaWt, and they are the same. So I discounted the [Root].Allocated.Wt from TotalGrowthRate and I got the AboveGroundGrowthRate in blue. [Untitled]https://user-images.githubusercontent.com/41041705/85253967-93776b00-b4a2-11ea-89aa-2728ad9f313c.jpg

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHubhttps://github.com/APSIMInitiative/ApsimX/issues/572#issuecomment-647306368, or unsubscribehttps://github.com/notifications/unsubscribe-auth/AC2UVWSEFGC4V6HOXOM2GX3RX3ZCFANCNFSM4BWGKYTQ.

JJguri commented 4 years ago

What is specifically the meaning of senRadnCrit in the model? @jbrider @hut104 I am comparing GreenSeeker observed data vs Leaf.CoverGreen and, generally, the model overpredicted for late stages of the crop. I checked CoverGreen an it has really high values (>0.7) for late stages where the crop should be senescing. Looking for senescence in the Leaf model, I found the parameter senRadnCrit with a constant value of 2 MJ m-2. When it is increased, the crop senesces early.

jbrider commented 4 years ago

@JJguri I am assuming you're just looking for a parameter to cause extra senescence at the end of the crop? senRadnCrit is a parameter for senescence caused by shading which is not going to have the affect you're looking for. You could look at the SLN Target - it should be higher for larger plants. From what little experience I have had with forage sorghum, it put's more biomass into the stem than standard sorghum - which would result in more leaf senescence. Do you have biomass measurement's for stem? You could try adjusting NUptakeCease - which will affect the amount of N being taken up post anthesis.

jbrider commented 4 years ago

I have been told that a taller sorghum such as CSH13R would be closer to what you need.

JJguri commented 4 years ago

Thanks @jbrider. I applied the following setup [Leaf].TargetSLN.PreEventValue.FixedValue = 3 for a given cultivar to see the effect of this parameter on leaf senescence. However, cover green maintained constant. I also applied the following code to a given cultivar ([Root].NUptakeCease.FixedValue=50 or 1000) and no change on the cover green was reported. Is there another factor to control the leaf senescence or ratio leaf:stem supply? image

jbrider commented 4 years ago

@JJguri NUptakeCease is in thermal time (570 for buster) - I would have thought the crop should fail using 50. I would probably try reducing in increments of 50 to see if it responds appropriately. I may have to check the SLN parameter as there should have been some response changing the TargetSLN.
How is your phenology looking? Did you have stem measurements?

JJguri commented 4 years ago

@jbrider We have stem data, but we are still processing it. Phenology is looking good, although I just used [Phenology].TTEndJuvToInit.FixedValue and [Leaf].HeightFunction to set up cultivars. We have also leaf number data. I am reporting the number of leaves as [Sorghum].Leaf.LeafNo. How should I set up this parameter for each cultivar? What is the meaning of [Leaf].LeafNumSeed?

jbrider commented 4 years ago

@JJguri I got the NUptakeCease a bit wrong too - 10 is probably appropriate. SLN does have a bug that I am fixing, but it still doesn't seem to bring about the required change. Can you upload your apsimx file so I can have a play with it?

JJguri commented 4 years ago

@jbrider please find the files attached. Have a look at the CoverGreen over predictions highlighted in the figure. All XY figures were created using LICOR and GREENSEEKER data vs APSIM predictions. Sorgo.zip Untitled We were wondering also why CoverTotal is not calculated as the sum of CoverGreen and CoverDead. Accordingly with the SorghumLeaf.cs file it is calculated as follows CoverTotal=1.0 - (1 - CoverGreen) * (1 - CoverDead). We were thinking that maybe we should compare the estimations from RS products (MODIS and Sentinel2) with CoverTotal and not with CoverGreen.

hol430 commented 4 years ago

The CoverTotal implementation isn't specific to sorghum, it appears to be copied from Leaf and SimpleLeaf (might be good to clean this up at somepoint). @hut104 can probably comment on why it's implemented the way it is.

hut104 commented 4 years ago

The equation is correct. If you have 50% green cover and 50% dead cover the answer should be 75% total cover. Think of it this way, 50% is covered by the green cover. The remaining 50% not covered by green, will have 50% of that area covered by the dead. It's similar to combining probabilities etc.

JJguri commented 4 years ago

Thanks @hol430 and @hut104 . The relations between components show two inflection points, one at 70% CoverTotal when CoverGreen stops and CoverDead increases, and another after 90% of CoverTotal when the crops started to translocate and when the senescence rate becomes important and changes between cultivars. For 0.9 of CoverTotal, we can have several scenarios considering different CoverGreen:CoverDead ratios. See an example here: Untitled

jbrider commented 4 years ago

@JJguri Looking at some of the simulations that are producing higher Cover (mainly the LT ones), do you have observed leaf numbers? I had a look at the biomass for each organ - grain looks very high for a forage crop - do you have observations for grain?