OverseerFM Irrigation discussion

Good discussion. In response to Marcelo, unfortunately all three modellers came up with different numbers which was significant for the property under discussion, given they had a reduction target to meet.

It might be very insightful to carry out a test case study among ourselves to see how consistent we are 😊 However think about what a trigger of 60% looks like in the paddock and how confident you would know that was actually happening – does that figure apply at the start of a rotation or the end? If the goal is to keep soil moisture above 60% what is the soil moisture at the start of the rotation and which number should be entered into Overseer as the trigger – the start of the rotation or the end?

Another separate point, a few have raised the importance that the irrigation data is entered based on consideration of the long term climate date. In Overseer, you cannot do this. You cannot see, or consider, the monthly rainfall or potential evapotranspiration values estimated from the long-term annual climatic averages.

Good afternoon all,

1)      While I agree with much of what has been said already, I do not see any solutions.

2)      Full disclosure, I am currently part of the ECan Overseer audit process that sometimes requires people to adjust irrigation regimes to align with contented limits, so I accept that I am part of the problem. 

3)      With regard to the proposed changes:

- No longer show amount of water applied through irrigation

I do not support this. I consider that understanding the effect improving irrigation management has on reducing drainage and in turn N loss, is a very significant part of improving behaviour.  

- Could have some representation of soil moisture and drainage over time

A good idea, but only if the modelled irrigation regime aligns with something like the average annual depth applied.

- Only allow one management type per irrigator for the year.

Indifferent.

- For FF, limit number of changes that can be made in a year (maybe 4)

It may be a perversion of the intent, but this approach is common and quite necessary (PC 2 etc) and so I do not support limiting the number of changes. Mainly, because I see little benefit.

- Add more help and warnings to input screens.

A good idea, but only if we are all working to towards the same and end and accept the inherent trade-offs.

4)      With regard to the particular concerns raised in the paper from OML.:

We have found that users are trying to adjust these inputs on a monthly basis to get the same amount of water applied (estimated by the model) as was measured on the actual irrigator for that year.

There are two reasons for this:

-          It is assumed that it will provide a more accurate representation of the farm

-          They have been told that it is a requirement of regulation

a)       Personally, I have not seen any modelling based around the irrigation demands for a single year, but I have heard some discussion and witnessed questions being asked that suggested it does happen, which is a concern.

b)      More education may help, but there needs to be clarity on what the ‘solution’ is.

c)       Irrigation is a very significant driver of N loss locally and if the desired outcome is that farmers can understand and relate to what has been modelled, then clear irrigation inputs are vital.

d)      Most farmers now have some understanding of the amount of water they typically use.

e)      So, to encourage buy-in and understanding, the irrigation water use estimated by the Overseer modelling should be displayed and it needs to align with reality.

f)        Most farmers begrudgingly accept the fact that irrigation inputs for any one season are not useful or appropriate. We can explain that to them.

g)       However, they will not take seriously the outcomes of the Overseer modelling if the estimated annual water use looks nothing like their records.  

5)      The ‘regulation’ comment is far more interesting because it is at least partially true and not easily remedied.

a)       I agree that there are distinct limits on the ability to model irrigation regimes within Overseer, particularly when shorter time-frames are considered (i.e. monthly).

b)      But, doing anything else will have consequences and presents risk as raised by Arron.

6)      As Alister Metherell rightly points out, in many regards a nutrient budget prepared using inputs that relate to any single year is problematic, but it is not uncommon.

a)       For example, unusually high stocking rates supported by favourable weather events are as likely to be included as are low stocking rates due to unfavourable weather events or other external factor.

7)      I think the issue of regulation and the awkwardness of trying to align irrigation with regulation is best considered in two parts:

a)       Baseline

b)      Current.

8)      Baseline NBs:

a)       Rightly or wrongly, most of us work within a Policy framework that uses some reference point or baseline.

b)      These baselines need to correctly reflect all verifiable input data such as stock numbers and stock performance, fertiliser inputs, supplements imported or removed, and irrigation inputs.

c)       All of these parameters can vary year to year, yet we enter them into a model against long term average climate information.

d)      If the stocking rate and stock performance in a baseline year was artificially low due to some unfavourable weather event, then that is unfortunately what we are stuck with.   

e)      Likewise, if a resource consent held in the baseline period limited, say stock numbers, then we are expected to model within those stated limits.

f)        Ultimately, whatever is modelled must be lawful.

g)       This gets rather awkward when limitations on irrigation takes are considered.

h)      A consequence of the proposed changes to OverseerFM is that it would not be possible to apply these limitations.

9)      Current NB’s

 a)       The issues with current NB’s can be similar to baselines, but typically farmers now have some form of Soil Moisture Monitoring (SMM) and modelling this introduces its own challenges.

b)      Often where an irrigation management regime involving SMM is entered into Overseer, the estimated water use is often far less than records (or logic) would suggest.

i)        That is, a farmer uses SMM with, say, a 50% depletion trigger and applies 15 mm per pass.

ii)       Assuming that they have followed this regime for a while, they will have a reasonable understanding of what they typically apply.

iii)     In my experience, when these management parameters are entered into Overseer the resulting estimated annual water use is often much less than the farmers records and/or the IrriCalc/IrriMap demand estimate, and so the farmer does not buy in to the outcome.

iv)     These figures are unlikely to ever align perfectly, but if the average depth applied on your pastoral property (averaged over 5 or 10 years) is 400 mm and the modelled outcome is 25% or 50% less than that, then we lose buy in and have a credibility issue.    

c)       From what I see, modellers then adjust the triggers used to try an achieve an annual applied depth that makes more sense as assessed against long term records and/or IrriCalc/IrriMap guidance.

d)      That seems like a logical response to me and I do the same in some situations.

10)   Modelling Irrigation with regard to consent limits

a)       Irrigation takes often have limits, both annual volume limits and some limit on the rate of take.

b)      Although it is a little awkward, a common ‘lawfulness test’ is to ensure that the modelled irrigation volume is within the consented limit, in the same way that cow numbers are checked against stated limits.

c)       However, during the baseline period many contents had limits on rate of take but had no annual volume limits.

d)      In these situations, the test applied is whether the monthly depths applied are achievable given the effective rate of take limit (generally calculated as mm/day).

e)      This is probably at the outer limits of what can be meaningfully derived from the Overseer modelling, but I am not aware of any other means of giving effect to the limits and constraints that apply.

f)        If the amount of water applied through irrigation is no longer displayed, then these limits could not be represented in the modelling.

g)       A Policy problem rather than a model problem, but still a problem.

h)      It becomes a question for the regulators.

i)        Are the regulators willing to ignore these limits given that applying them or assessing them through Overseer analysis is fraught with difficulties?

j)        If so, what are the consequences?

k)       Are the regulators willing to ignore these limits given that applying them or assessing them through Overseer would become impossible under the proposed changes?

l)        If so, what are the consequences?

11)   Irrigation takes with constrained reliability

a)       Consider irrigation takes with constrained reliability, such as groundwater takes with Adaptive Management conditions or surface water takes with minimum flow conditions.

b)      I have seen various attempts to give effect to these limits ranging from just ignoring them entirely, to tweaking irrigation management parameters on a monthly basis to produce a monthly applied depth that aligns with the irrigation typically possible in that month when the take is averaged over the relevant period

c)       If the amount of water applied through irrigation is no longer displayed, then these limits could not be represented in the modelling.

d)      Are the regulators willing to ignore these limits given that applying them or assessing them through Overseer analysis is fraught with difficulties?

e)      If so, what are the consequences?

f) Are the regulators willing to ignore these limits given that applying them or assessing them through Overseer would become impossible under the proposed changes?

g)        If so, what are the consequences?

12)   FEP auditing as a solution:

a)       FEP auditing can perhaps assess whether the irrigation regime entered into Overseer aligns with what the regime the farmer tries to use, this is only really applicable to current NB’s.

b)      Baseline NB’s are not meaningfully assessed as part of the FEP Audit process and so FEP auditing is not going to address this issue.

c)       The other reason I am not convinced that FEP auditing is the complete solution to this problem, is that as stated previously, often when an irrigation management regime is entered into Overseer, the estimated water use is often far less than records (or logic) would suggest.

13)   The question that arises is, are FEP auditors to check that the irrigation management parameters align with how the farmer tries to manage the irrigation system or are they checking that the modelled annual use aligns with the records available?

a)       My understanding is that FEP auditors are currently guided to look at the irrigation management regime only and not compare the modelled annual depth/volume with the water take records.

b)      While I understand this approach, it seems a bit of an anomaly.

c)       For example, while an FEP auditor may ask a farmer about their nitrogen fertiliser regime, their response is not accepted as proof.  Instead fertiliser application records are also checked.

d)      It seems to me that the primary reason this check is not conducted in regard to irrigation is because of the poor alignment between the modelled outcomes and reality (reality being defined as either averaged long term records and/or the IrriCalc/IrriMap demand estimates).

14)   A possible to solution

a)       I wonder if something like the following may be a solution.

b)      What if it was possible to enter an annual applied depth and leave Overseer to then spread this throughout the irrigation season based on local long-term climate data?

c)       That is:

i)        the modeller enters only some very basic parameters about the irrigation system, perhaps only specify system type.

ii)       Then enter an annual depth.

iii)     Overseer does the rest.

d)      This approach would require a justifiable annual depth to be used and this could be easily assessed as part of the FEP audit.

e)      The annual depth applied would either be supported by averaged long term records and/or IrriCalc/IrriMap demand estimates.

f)        There are some complexities to be resolved, such as addressing multiple irrigation systems, soil types and/or climatic variation across a property, but nothing that could not be solved.

i)        Perhaps the depths would need to be generated at the block level?

g)       Determining the annual depth would be the ‘contestable’ part of the process, such as performing a WQN9 calculation.

i)        Through the process of generating these applied depths allowance would be made for system constraints, consent limits and reliability of take.

ii)       Once the ‘scaling factors’ for these various issues were agreed, the process would be come easily repeatable.

h)      Updates to soil parameters etc would pose a few changes, but this nothing new or unique.

Just a thought.

Reuben

Great to see some good discussion here, although perhaps a little concerning that my quick summation of the discussion seems to be that we all agree that some behaviours need to change when it comes to how irrigation systems are being modelled, but don't generally people don't want to change anything within the software? As a software business, we believe that changes to guide better behaviour belong, firstly in the software itself.

If I could make an analogy with another part of the model. Most users accept that you add animals to the farm system and then let the model get on with it. When users try to be very specific about certain animals being on certain blocks at a certain time is when feeding errors and strange results start to occur. In this regard, in normal use, the model goes off and "looks for feed" to meet the level of production which you have already stated... I think we need to get to a similar level of trust around the model "doing it's thing" to match irrigation need.

I think there is a great deal of merit to considering the approach which Leo and the ECAN team have put thought into - that it is the mechanics and management of the irrigation system which should be being audited (and so presumably consented) rather than modelled applications.
If a farmer "improves" their application method - Overseer will reflect that
If a farmer improves their management system - Overseer will reflect that

I wonder also, to Reuben and other's points about getting farmer buy-in, if having the monthly values doesn't actually serve the opposite. They shouldn't reflect what was actually applied that month, because they are what the model needs compared to a 30 year climate datapoint - not a single years weather.
Perhaps another way here would be for us to add the total irrigation volume the model is applying to the new Farm Plan report which we are soon releasing in the software (teaser screenshot below 😉) so that a "sense check" could be performed on the irrigation volume + avg rainfall numbers?

That said, water take should be checked against water take or metering records, not from an Overseer analysis; and if current consenting or auditing guidelines suggest otherwise then those guidelines should be updated.

I'm also interested to better understand this whole "checking in irricalc" idea? what exactly are we checking? given that John's paper comparing the two established there is little difference between the two outside of the climate data which is used and the way in which the models handle that data?

Hi All,

Not sure if I have much more to add, but here are my two cents:

1. Leo's approach to model auditable management decision makes sense, particularly for year-end nutrient budgets

2. The choices available in Overseer all have their place, however I think further guidance is necessary to ensure the most applicable options are used

3. Showing the irrigation inputs by system and/or total volume is useful to sense check the modelling. It is not uncommon for a farmer to say they generally do one thing (i.e. 40 mm/10 days FF), but in reality they do adjust what they do in response to rainfall or low flow restrictions. Solely relying on the modelling therefore gives quite a non-sensical result and checking the modelled inputs against averaged actual water use is useful. Agree adjusting irrigation volumes to perfectly reflect actual inputs is not ideal.

4. I agree with updating the FF/VF/FV/VV terminology - while I understand where you are coming from, it's not actually that intuitive. For instance, I wonder if it would be easier to tick a box called "VRI" instead of having a VF/VV option? Marcelo's common terminology options would be good as well, and may be easier to model the actual decision making process occurring on the property.

5. Disagree with only having one management option per system. We are encouraging farmers to reduce their trigger point in the fringes of the season to capture more rainfall, with higher trigger points in the peak, only permitting one management option will prevent the ability to model irrigation managed in this way.

6. I like the idea of trending water use/drainage over time, as it might identify issues/inconsistencies with irrigation modelling and/or demonstrate systematic improvement in management decisions.

Some other ideas:

• Whats the chance of Year-End nutrient budgets using year-end climate data? i.e. VCN information for the area? Predictive/Scenario budgets should definitely be 30 year average. Will certainly make life easier if all inputs impacted by climate effects are modelled in the year-end, rather than just some. E.g. stock numbers, fertiliser use etc

• If you do happen to do the above, could Overseer have a running (or 3/5 year rolling) average N, P, GHG losses across all year-ends?

• What if the (averaged) annual water usage volume is an input with the irrigation management rules determining how and when that water was distributed on the property?

Eva

I agree with the general consensus of the document. Please see my comments below on the four points. The main point I think we need is more clarity and definitions around the inputs and the impact of inputs on not only irrigation and nitrogen loss but drainage as well.

In the previous Best Practice Input standards there was a good flow diagram and explanation around the managements, can this please make a comeback with added detail? And this could be a solution to making inputs more consistent.

- No longer show amount of water applied through irrigation.

I disagree with this. I think this is a great discussion point when it comes to farmers who are needing to understand their losses and look at opportunity where they can improve and Irrigation management is one of them. Providing the monthly usage and annual usage is key part to understanding the nutrient budget.

A suggestion could be/would be beneficial for users, farmers and auditors to include the total irrigation for the season sort of like the total tonnages with the crops.

I think it would be a disadvantage if you were to remove this for everyone.

- Could have some representation of soil moisture and drainage over time.

With drainage driving nitrogen loss, providing the information monthly could highlight problem months based on the model. It would be great if we could understand the rainfall as well as the two really go together.

- Only allow one management type per irrigator for the year

Yes, if there is full confidence with the average irrigation management, hence average drainage can be accurately modelled with this approach.

- For FF, limit number of changes that can be made in a year (maybe 4

My question for this is why four? My question with modelling little GMP (Canterbury Focused) we could be requiring individual inputs per month (irrigation season can be up to six months). Can I please gain some more understanding around this and the reasoning behind it?

Overall Farmers need to understand how to reduce nitrogen losses on farm within their Overseer. With OverseerFM being more visual and farmers having access I think providing them with understanding and total amounts used is key. If we start to restrict the tools to model irrigation, we are starting to restrict the opportunity for farmers to demonstrate their reductions and this will benefit the regulatory and monitoring perspective as well.

I think with my earlier statement we require more education and certification to help with the consistency in modelling.

A few thoughts a a bit of a rebuttal.

1 – There are potentially some positives to simply entering irrigation management parameters and letting Overseer ‘do its thing’ rather than assessing the resulting applied depths, but there are many drawbacks too.

2 –The analogy of animal performance parameters was raised as an example of why we should just let leave Overseer to ‘do its thing’. I believe that this example is in fact a very good illustration of why we do not do that and why we should not do that.

3 – The option exists to enter only some very basic parameters about the livestock and therefore to leave Overseer to ‘do its thing’. It is also perfectly valid to enter more detailed parameters to better reflect a particular farm system. This allows modellers to develop models of farm system that make sense to the farmers. This should result in outcomes that better reflect the intricacies of a particular farm system.

4 – To take the analogy further, the more accurate comparison would be that Overseer decided that because you had a certain breed of sheep, the lambs must weigh 30 kg LWT at sale, while the farmer has records showing that over the last 10 years that the lambs have averaged 45 kg LWT. A farmer is not going to buy in to the outcomes of modelling when they can see such large discrepancies.

5 - In my opinion, in many situations the difference between recorded long-term average water use and modelled outcomes when using the same irrigation management parameters regime is very significant and enough to make the results meaningless to farmers. This is not simply a consequence of short-term water use records a versus long term climate data.

6 – To me the choice becomes, do you create the opportunity to better align these outcomes, or do you ignore the discrepancy? I vote that we maintain the option to better align the outcomes, so long as we all do this this is a sensible and similar way.  

7 - Many ‘tactical’ choices can be represented in Overseer, such as culling stock early or changing planting dates. These are options farmers can consider and then trial through Overseer which is invaluable.

8 - As improving irrigation management is a major driver of nutrient losses on irrigated properties, it would make sense that it was possible to model tactical changes to irrigation management. This could be changing the regime used in a single month, or simply answering the question ‘what if I improved my management and used X% less water per annum?’

9 - The major challenge seems to be making sure that whatever is entered makes sense given: the climate data used by Overseer, how this data is used, and the decision-making process applied within Overseer. In my opinion, concern about the fit of ‘short-term’ water use records with the long-term average data set is red herring and it is not the major cause of the discrepancies we see.

10 - A single year data set is essentially meaningless, although it is allowed with regard to other inputs (i.e. fertiliser and stock), despite these inputs often also being in conflict with the long-term average climate data.

11 – Inputs averaged over four years (as are quite fashionable locally) are better than a single year data set, but in fairness I have never seen any statistical analysis of how much better. Now, that would be interesting to see...

12 – Many farmers have water use records of more than 10 years in duration. Increasingly these famers have had SMM throughout much of this period and so these records are a very useful guide as to the typical irrigation demand for their farm and farm system, even if it doesn’t align perfectly with the 30 year climate data used within Overseer. In my opinion, use of an averaged 10-ish year duration water use record is better than any current known alternative.

13 – With regard to IrriCalc as a sense check, in my experience, when a location, soil PAW (60cm) and an irrigation system type is entered into Overseer and the same parameters are entered into IrriCalc, the outcomes are very different. Often the modelled applied depth determined by Overseer is 25 to 50% less than the IrriCalc estimate, sometimes more. This is what discourages buy-in and this what turns our efforts into box ticking exercises.

14 - While the Bright-Wheeler paper showed that under very specific circumstances the outcomes can be quite similar, in most situations this does not hold true. Try it yourself and see. Noting also that the use of IrriCalc demand estimates as a check/guide is stipulated by ECan under several sub-regional plans.   

15 - A common rule-of-thumb used locally is that rainfall plus irrigation should sum to about 1,000 to 1,100 mm for established pasture. This obviously varies across Mid-canterbury, let alone wider Canterbury, but it is a handy sense check. Commonly, the annual applied depths resulting from Overseer modelling fails this test, often by a large margin.

16 – So, if the modelled applied depths that result from specifying an irrigation system type and then entering the irrigation management regime to the best of our abilities into Overseer results in an outcome that; fails a basic sense test, as well as making little sense against an averaged 4 to 10+ year duration water use record, and is vastly different to the IrriCalc demand estimate, then I strongly suggest we need to maintain the ability to tweak the inputs.

17 – We should do this in a common manner and in a manner that fits as well as possible with the long-term average data set used by Overseer. Not only this, whatever we do should align with the many other assumptions built into the Overseer. This is only possible with much greater clarity about these assumptions and some training.

18 - For example, I do not have complete faith in the TV3 weather girl and so I may choose to irrigate even though there is potential for some rain in the next few days. I will try to ensure that I leave capacity in the soil to capture the forecast rain. If Overseer determines the soil moisture deficit to be 24 mm while the trigger is a deficit of 25 mm, it seems that Overseer would not apply irrigation because it has perfect knowledge of the typical rainfall that will occur in the next few days based on the way in which it processes the climate data it operates on. Making the decision to irrigate in this situation, while aiming to leave sufficient capacity for the soil to capture the forecast rainfall, seems to meet GMP expectations in my opinion. Overseer is doing something else and effectively holding us all to an even higher standard, which is often utterly unachievable. In my opinion, the assumptions used in the irrigation model within Overseer are a potentially a bigger issue than the lack of fit with the long-term climate data.

19 - The scenario above applies to a system with a short return period and no known system capacity constraints. If, however the return period of the irrigation system is constrained, then the decision-making process gets even more complicated.

20 - Irrigation may well be stopped ahead of forecast rainfall, but irrigation will likely recommence sooner than soil moisture monitoring would suggest. But it may well be that the settings on the irrigator are altered (i.e. turning on all the cams of a Roto-rainer) and, say, half the typical applied depth is applied for some of the first irrigation round. This is all far too subtle to be accurately represented through the irrigation management rules available in Overseer, even on a monthly timestep. So instead, people apply their best judgement to try and create a typical set of management parameters and/or enter parameters that result in something that resembles typical irrigation applied. None of which is perfect, but there are good reasons for trying.   

21 – While I like the apparent simplicity of auditing against irrigation management parameters I think that it still presents some real challenges. Consider the example above, does the FEP auditor simply ask, ‘How much do you apply per pass? What is your average return period? What are the triggers and targets used?’. If so, then life is easy. Inaccurate but easy.

22 –If the FEP auditor is to ask a few more questions about the ‘howevers’ and the exceptions, then the variations and complications would emerge. In this scenario it is a lot less clear as to exactly how a FEP auditor would determine that the irrigation management regime as entered into Overseer correctly reflects what a farmer typically does.

23 – To give the FEP Auditor the ability to meaningfully answer this question would require all modellers to follow the same process in getting from the myriad of factors a farmer considers throughout the season down to typical management regime that could be entered into Overseer. In this sense I agree with many people who have commented previously. Flow charts, decision trees and significantly more guidance than currently exist would be required to achieve this outcome.

24 – If a FEP auditor has to go beyond the basics and there is not meaningful clarity and therefore consistency in how we get from what we are told by the farmer to set of irrigation management parameters, then in my opinion we end up back we where we are now. That is, we all largely do our own thing and the FEP auditor simply assesses whether a suitably skilled person did the work. I do not see the benefit.

25 - I have spoken with a number of people over the last few days and some interesting issues have emerged. These sensible and suitably skilled people do some quite interesting things with regards to modelling irrigation in Overseer for many seemingly sensible reasons. However, no two people I have spoken do the same thing. Even if they both model irrigation systems using SMM and soil moisture triggers, how they arrive at those triggers and why differ. Therefore, the outcomes differ, and this is a problem. Looking away does not fix it.

26 – For those of you who have stuck it out and read this far, well done.

27 – As I suggested previously, a possible solution would be to add the option of entering an average annual depth applied (perhaps at the block level??). The logic being that this average annual applied depth would then be spread throughout the irrigation season in a manner consistent the climate data used by Overseer.

28 - Arriving at this annual applied depth would be a well-defined, robust and therefore auditable process. Something like WQN 9 perhaps. In this process is where the magic happens and this is how restrictions (say adaptive management or minimum flow conditions) and other constraints are reflected. I.e., annual depth of X is reduced by Y% to reflect the consequence of Adaptive management conditions in Valetta. Just a theory.

29 – Further, it would great to see the outcome of this displayed on a monthly timescale and that these parameters could be altered. For example, an annual average depth of 425 mm is entered, and this equates to 85 mm applied in March. While working with a farmer I can adjust this down, say, 10% to test the consequence improving shoulder season management. How the farmer could achieve that reduction is up to the farmer as it should be. This would only ever be done as a scenario for educational purposes, but I could see it being quite useful.

30 – Theoretically, if in a few years’ time it could be illustrated that the water use had reduced, then a reduced average annual depth could be entered. Thus, progress could be tracked and modelled.

31 – From a puritanical standpoint, so long as any given property is modelled in a consistent manner year-to-year, then much of the variation in how irrigation systems are modelled does not matter. However, most of us work under a set of rules that require assessment and reporting against limits and therefore all analyses must be prepared in a consistent manner. This may be impure, but it is very much reality.

32 – Either way, I endorse more training, more guidance and more clarity about how the irrigation model works. This should improve the consistency of the outcomes achieved. On this basis, the options available to modellers should not be reduced.

And a good night to all.

Reuben

Alan Hawkins

Firstly to answer the questions posed:

Note: these questions are answered as if it’s PC5 applicable N baseline rules

No longer show amount of water applied through irrigation

I do not support this. N leaching is a function of drainage and N concentration.  Climate & soils are set; irrigation is able to be changed.  Irrigation is the main variable determing drainage and we need to see what is behind the N loss figure. 

Could have some representation of soil moisture and drainage over time

Can’t see value in soil moisture as can’t see how it would be reported and what it would mean.  Drainage is a good figure to see.   It offers insight into the water balance (irrigation + rainfall = drainage + ET).    

Only allow one management type per irrigator for the year.

Depends on what the results are being used for.  As a reporting figure see thoughts below.  As a tool to explore management options then yes.

 For FF, limit number of changes that can be made in a year (maybe 4)

Can FF meet GMP requirements? 

Add more help and warnings to input screens.

I dislike popups. No unless something is clearly wrong.

Are the questions above the relevant questions?   Here are some questions I would like answers to:

1)      Is it expected that GMP baseline irrigation would be modelled consistently between Overseer users?  My view is defiantly yes. I.e. for a farm all users should model irrigation the same way.

2)      Can the famer improve irrigation management beyond GMP and that improvement be   modelled and the reductions count towards reductions?  What level of justification is required?  Answers to that question are  not so clear.   If GMP baseline irrigation was 500 mm and the farmer chose to only use 400 mm, then drainage would be reduced and so probably would N loss.  That is real but it gets mixed up with season to season variation.  See comment below about variability of rainfall.

If I asked an informed farmer how Overseer was used, this is what I would hope they would say:

1.       Record N baseline farming practice and calculate an N baseline figure.

2.       Calculate N loss figure for current practice and compare  to N baseline.

3.       Calculate N loss figure for current practice and compare to  N baseline reductions required.

4.       Use Overseer to compare the impact of different management options on N loss.

A collection of comments:

·         It is my view that the N loss figure from Overseer can be used as a relative figure and not as an absolute figure.  If the farms irrigation was modelled a certain way in the baseline period and is modelled the same way in 2020 then the comparison is valid.

·         It is my view that nutrient budgets are a measuring tool for current practice with the ultimate aim to meet water quality targets.  The N loss figure from Overseer is not an absolute figure but a relative figure.  The water quality in the receptors will be the ultimate test. 

·         Is it possible to be rewarded for irrigation to a standard better than GMP?  If long term GMP irrigation is 500mm and the farm is run at 400mm there will be less drainage and less leaching.  Then the question is how do you prove that the reduced irrigation is not a one off?

·         Here are the rainfall figures from the ECan Hurunui SHI site 2006 – 2019.  14 years.  Average rainfall  was 753mm.  Actual years were 676, 529, 990, 692, 616, 979, 980, ,970, 421, 551, 924, 1036, 713 and  474.  That gives some idea of the variability. 

·         I have not teased the figures apart for a while but a few years ago leaching was a function of drainage and concentration.  As you forced a reduction in drainage the concentration went up and overall leaching went down.  My recollection it varied with soil type among other things.  The way to reduce leaching was to reduce drainage.    

·         I do not believe that the absolute figure produced by Overseer is a fair representation of the actual N loss.  That is a discussion that may play out in the PC7 hearings. One document worth reading is Ants Roberts evidence in PC7 under Ravensdown.

·         The ultimate outcome that society  wants is meeting water quality targets. Overseer is a tool to get there.  The measured outcome will be the water quality. As an example take a lowland stream with a mean residence time of 10 years. After 10 years the guess is that you would be much of the way to equilibrium.  The time frame can be relatively short.

·         The requirements to use the model in a regulatory framework identified by the Environment Court* are quoted as:

a.      A consistent approach to model input data and maximising the accuracy of that data

b.      Using only appropriately qualified and experienced experts to run the model for compliance

c.       Establish a clear, efficient, and reliable process to review and update model outputs and management practices at appropriate intervals

d.      Appropriate on-site verification that modelled inputs and outputs are being complied with, in addition to independent peer review of performance

e.      Using the model to predict trends and relative changes in farm management systems, rather than absolute values.

*(Federated Farmers of New Zealand Inc. v Bay of Plenty Regional Council [2019] NZEnvC 136 at [117])

Alan

Very much in support of displaying monthly drainage estimates.

Hi all - thank you every one for your contributions to this discussion - the debate seems to have dried up a wee bit now so we will take some time to review the feedback and comments before we make any decisions about next steps...
In the meantime, we will releasing the beta version of the Scenario Tool and a new Farm Impact Report into OverseerFM very shortly, we look forward to hearing feedback on both.
Again, thanks for your contributions to this discussion - especially Reuben 🙂

Good morning Alastair,

In your first email about this discussion, there was a stated end date. “The discussion will be closed on Friday 23rd October.“

That could be why things have gone quiet.

Probably a good idea to have an end point or some people could really waffle on and on and on…… (tee hee).

This particualr topic is an interesting varation on a theme, what Overseer can or can’t do versus what Overseer is being used for.

I try to do things with Overseer that are an over-reach on a daily basis. The Policy requires it.

Good luck.

Kind regards,

Reuben