By Simon Kruger, WMG Project Communications Officer

At the WMG Spring Field Day on 17 September 2026, WMG Project Communications Officer Simon Kruger trialled a different style of field day session at the WMG RiskWi$e N banking trial at Velyere Farm, Dandaragan. Rather than simply standing in front of the plots and presenting the trial treatments and results, the session was designed to get the 92 people in attendance moving, making a prediction and thinking about the information behind their decision.

Participants were first asked to back one of six canola plots to produce the highest yield this year, based only on what they could see. They were then given three additional pieces of information about the site and treatments, revealed one at a time. After each reveal, they could stay with their original choice or move to another plot if the new information changed their thinking.

The interactive session was co-designed with assistance from Dr Brendan Brown of CSIRO, who leads the behavioural science work within RiskWi$e. The aim was not to find the ‘right’ answer on the day, but to explore how additional information changed people’s predictions, which information mattered to them, and how differently they weighed the same evidence.

This article runs the same exercise ‘on paper’. The canola has not been harvested, so nobody knows the answer, WMG included. At each stage, pick the plot you would back and note the letter before reading on.

Where do your decisions sit?

Before looking at the plots, the group lined up according to how they make N decisions, with gut feel, past experience and rules of thumb at one end and test results and calculations at the other. About 20 per cent placed themselves at the analytical end, 70 per cent used a mix of analysis and intuition, and 10 per cent leaned mainly on gut feel.

Before reading on, place yourself somewhere on the same line and think about why.

Round 1: Look only

These photos were taken on 5 October, about two and a half weeks after the field day, so the canola is further through pod fill than it was on the day.

The site was sown to wheat in 2025 and is in canola this year. On the day, the only other thing the group heard was a working definition of N banking: applying more N than the current crop needs, on the chance it pays back in yield or remains in the soil for the next crop.

While the field day group could walk through and inspect each plot firsthand, the same challenge applies here: based on what you can see, which plot would you back?

The field day group spread across all six. Plot B drew the most support with 22 people, followed by F with 20, E with 18 and C with 16. Plots A and D had eight each.

Round 2: Site background

The trial was established in 2025. The soil is a deep red brown sandy loam, with sand to more than 2 m. From April to October 2025 the site received 526 mm, against a long-term average of 425 mm. The season followed a very dry summer, with a wet July and August and a dry finish. Before sowing in 2025, the soil held 72 kg N/ha to 50 cm.

With that in mind, stay with your plot or move to another.

Round 3: N applied in 2025

On the day, each plot’s details sat under its marker sheet and were uncovered at this point.

Every plot received 10 kg N/ha as urea at sowing in 2025. Plot D, the seeding N only treatment, received nothing further. The other five went on to receive between 125 and 240 kg N/ha in total, applied as follows.

Plot	Treatment	At sowing	~4 weeks post sowing	~8 weeks post sowing	Total
A	Grower practice	10	40	75	125
B	Yield Prophet, decile 5	10	0	180	190
C	N bank, economic optimum	10	0	140	150
D	Seeding N only	10	0	0	10
E	Grower experiment	10	80	150	240
F	High N bank	10	0	230	240

*All figures kg N/ha, applied as urea to wheat.

Plots A and E split their in crop N across two applications, with E receiving twice the in crop N of grower practice. Plots B, C and F received theirs in a single application at first node. E and F reached the same total of 240 kg N/ha by different routes.

Now that you know how much N went on and when, stay or move.

At the field day, approximately 30 of the 92 people, about a third, changed plots after hearing the rates.

Round 4: What was left after harvest

Each plot was soil sampled in January 2026, after the wheat was harvested, by Ethan Sirr of Moora Soil and Plant Contracting.

	A	B	C	D	E	F
N applied in 2025	125	190	150	10	240	240
Soil mineral N, 0 to 50 cm	59	98	113	39	144	137
Soil mineral N, 0 to 90 cm	80	158	146	48	201	188

*kg N/ha, nitrate plus ammonium. These are results from six individual plots, one per treatment, chosen for the exercise. They are not treatment averages.

The two plots that received 240 kg N/ha held the most mineral N after harvest, and plot D, with 10 kg N/ha, held the least. Against the 72 kg N/ha measured to 50 cm before sowing in 2025, plots A and D finished lower, while the other four finished higher. This is consistent with some of the N applied to the wheat in those four plots remaining in the profile in January, although these figures cannot show how much of the rest was taken up, lost or tied up.

Knowing some of that N was still in the soil, stay or move.

Approximately 15 people changed plots at this stage, although most of the group stayed with their previous choice.

Same information, different decisions

Some people stayed with their first choice through every reveal, while others moved as new information arrived. Given the same three cards, people weighed them differently and reached different conclusions.

Three cards are not enough to make a case for or against N banking, and the session was not designed to make one. The intent, in line with the RiskWi$e approach, was to slow the decision down, work out which information matters, and make the best-informed call available. That does not guarantee the right result, because the season still has a large say, but it should give more confidence in the decision, make it easier to explain why it was made.

If you played along, compare the plot you picked from the photos with the one you finished on, and think about which piece of information moved you, if any did, and why?

The 2025 wheat result

After the session, a couple of growers and advisers asked how the wheat yielded in 2025. Treatment average yields ranged from 4.5 to 4.9 t/ha, with no statistically significant differences in yield or grain protein between treatments. Across rates from 10 to 240 kg N/ha, the trial did not detect a yield or protein response in its first season. The 2026 canola result is not yet known, and WMG will report it once the plots are harvested and analysed.

Benchmark your own decision making

A Decision Benchmarking survey, shared with WMG by Dr. Brendan Brown of CSIRO, compares your decision default settings with those of more than 1,000 other growers who have completed it, and returns a personalised benchmark.

Complete the survey at https://farmvoice.au?module=decision_benchmarking or scan the QR code below.

Whether you joined us at the field day or played along through this article, we hope the exercise prompted you to think about how you make decisions and what information carries the most weight. More information does not guarantee the right outcome, but taking the time to seek out the information that matters can help make a better-informed decision and give you greater confidence in why you made it.

If you would like to learn more about the RiskWi$e decision-making work, or have a decision you would like to explore, get in touch with WMG Project Communications Officer Simon Kruger.

Acknowledgements

Thanks to Chris Spencer, farm manager at Velyere, for hosting the trial and the field day group, to Kalyx for managing the plots, and to WMG’s Chanel Duggan and Gabby Carrivick, the volunteers who held the plot markers on the day and everyone who interacted and got involved.

RiskWi$e is GRDC’s National Risk Management Initiative, a five-year, $30 million investment running from 2023–2028 that aims to help Australian grain growers strengthen decision making under risk and uncertainty for enduring profitability.

RiskWi$e helps growers understand the trade-off between potential reward and downside risk across a range of farm management decisions. The initiative brings together 26 grower, adviser and extension groups, 10 research partners and six Action Research Groups to help Australian grain growers and advisers get decisions ‘more right more often’.

CSIRO leads the initiative nationally, with the Grower Group Alliance leading in WA.

By Simon Kruger, WMG

Nitrogen is one of the harder inputs to get right, because much of the decision has to be made before the season shows what it will do. Growers with later applications still to come face that squarely by mid-July. Should they stick to the plan, add more, or pull back? If the season is running well, the extra nitrogen may pay. If it’s not, it may be money spent on a crop that can’t use it. This in-season call is harder to bring evidence to than the autumn budget, because it depends on how the rest of the season plays out, and nobody yet knows that. Nothing removes the uncertainty, but some tools can help put a shape around it.

One of the tools that can be of use with this decision point is the Rainfall to Date tool from the Department of Primary Industries and Regional Development (DPIRD). It’s free, WA-based, and shows where a season sits now against the historical record, then sketches what a poor, average or good finish might look like from here. WMG’s interest is in putting it in front of growers and thinking through what it might offer the top-up decision.

Where three stations sit as of mid-July

The Rainfall to Date tool draws on DPIRD’s network of weather stations across the state. It plots cumulative growing-season rainfall for a chosen station against that station’s own record from 1975 onward. The three solid green lines are the historical decile 1, 5 and 9 tracks. Decile 5 is the median, the line half of all years on record fell below and half above. Decile 1 is the level only the driest 10% of years fell below, and decile 9 the level only the wettest 10% exceeded. The purple line is the rainfall received so far this season. From the current date, three lighter blue lines project what a decile 1, decile 5 and decile 9 finish would look like from here. The historical tracks are in green, the projections from today in blue.

Figure 1. Example of how to interpret Rainfall to Date Graph

Run for three DPIRD stations across the region on 14 July 2026, the tool shows a season tracking below to around average across the board.

Figure 2. Badgingarra, Rainfall to Date, 14 July 2026

Figure 3. Moora, Rainfall to Date, 14 July 2026

Figure 4. Dandaragan West, Rainfall to Date, 14 July 2026

Badgingarra had received 199.8mm of growing-season rainfall, sitting on decile 5, the median, with a summer at decile 4 behind it.

Moora had received 136.3mm, sitting at decile 4 and the lowest of the three in absolute terms, on a summer at decile 4.

Dandaragan West had received the most rain of the three at 209.1mm, yet sits at decile 4, just below Badgingarra despite the higher total, on a summer at decile 4.

The three stations tell a reasonably consistent story this year. The season is running a little below to around its long-term middle by mid-July. Dandaragan West is the one worth pausing on. It had received more rain than Badgingarra, 209.1mm against 199.8mm, yet sits a decile lower. The decile is measured against each station’s own record rather than an absolute total, so the same amount of rain can read differently from one location to the next depending on what that station usually receives.

There is still a question worth sitting with, even in a year where the three stations sit close together. The tool compares each season against the record of the station nearest to it, which may be a more useful reference than a district average or a general sense of how the year is going. Which station sits closest to a grower’s own property, and how well its rainfall reflects what has fallen on the paddock, are the grower’s own to judge.

Reading the finish from here

The three blue projection lines are where the tool earns its place in a nitrogen conversation. Each extends the season forward from today’s rainfall to a historical finish. The bottom line runs to a decile 1 finish, the middle to a decile 5 finish, the top to a decile 9 finish. Between them they set out a question a grower is already weighing. If the rest of the season comes in poor, average or good, where does the year end up?

Reading the projected finishes off the three graphs, rounded to the nearest sensible figure, gives a feel for the range at each station by early November.

Badgingarra, sitting at 199.8mm now, projects to somewhere around 300mm on a poor finish, about 450mm on an average one, and close to 585mm if the rest of the season is wet.

Moora, the driest of the three at 136.3mm, projects to around 215mm on a poor finish, roughly 300 to 330mm on an average one, and about 360mm on a good one.

Dandaragan West, at 209.1mm now, projects to around 380mm on a poor finish, roughly 460mm on an average one, and around 520mm on a good one.

Two things in those ranges are worth considering.

The gap between a poor and a good finish at each station is wide. At Badgingarra it is most of 300mm, at Dandaragan West around 140mm, at Moora around 150mm. Even at the station sitting on the median, the finish is a long way from decided. Most of the season, on these numbers, is still to come.

The second is that all three stations are starting from a similar place, a little below to around their long-term middle. None of them has banked the kind of wet start that would take the pressure off a nitrogen decision, and none is so dry that the season is already lost. That leaves each of them with a genuinely open range, where a poor finish and a good one are both still in reach.

From a rainfall range to a nitrogen question

The step from a rainfall finish to a nitrogen rate is not a clean one, and it is worth being upfront about why. Rainfall is one input into yield, not the whole of it. Stored soil moisture at the break, soil type and water holding capacity, nitrogen already in the profile, and the timing of the rain through the season all sit between a millimetre total and a tonne of grain. The tool measures rainfall and nothing else. It can’t tell a grower their yield potential, and it certainly can’t tell them a rate.

What it can do is give some shape to the range a crop might be growing into, which is the thing the top-up decision is really about. A later nitrogen application assumes the crop will have the yield potential to use it. The wider and lower that range looks, the more that assumption is exposed. The narrower and higher it looks, the safer it becomes.

Set against the three stations, the picture this year is more consistent than it sometimes is. None of the three has the kind of wet profile that would make a later application look low risk, and none is so far behind that the case for holding back is clear cut. Each sit in the middle ground where the decision is genuinely open.

Moora is the driest of the three, and the one where the downside case is closest to hand. A decile 1 finish from 136.3mm would leave the season well short, and that is the situation where extra nitrogen is most at risk of going onto a crop that cannot convert it. A grower on that country might reasonably want to see more of the season before committing, and to look hard at stored moisture and crop condition rather than rainfall alone.

Badgingarra and Dandaragan West sit a little more comfortably, closer to their long-term middle, but not by enough to change the shape of the decision much. The range of finishes at both is wide enough that a later application could pay in a good year or be stranded in a poor one. The rain each has received matters less than the rain still to come, which no one can yet see.

None of that is a recommendation. It is the same tool read across three stations that happen to be telling a similar story this season, showing a decision that is open at all three rather than settled at any.

The trade-off both ways

The reason this decision is hard isn’t just the uncertainty. It’s that the two directions don’t cost the same, and they don’t cost the same in every season.

Underdo the nitrogen on a season that turns out well, and the cost is forgone yield. The crop had the potential to do more, and the nitrogen was not there to let it. That is a real cost, but an opportunity one. The grain could have been grown and was not.

Overdo the nitrogen on a season that turns out poorly, and the cost is different. It’s money spent on a product the crop couldn’t use, and in a dry finish it can be worse than neutral. A crop pushed with nitrogen into a dry spring can run its water down faster and hay off harder than it otherwise would. So, the downside is the input cost, plus in some cases a yield penalty on top.

Those two costs are not symmetric, and which one a grower is more exposed to shifts with the season and with the numbers in front of them. Two things this year press on that balance.

The first is the price of nitrogen. Urea prices spiked sharply earlier in 2026 on the back of global supply disruption, as WMG covered in the May newsletter. Prices have eased from that high since, though the volatility has not gone away, and the direction from here is unsettled. That instability is itself part of the decision. A later application is being weighed against an input cost that may have moved since the autumn budget, in either direction, so the price at the time of the call is worth checking rather than assumed. A higher price raises what is at stake in getting the decision wrong, both the cost of nitrogen that a poor season strands and the value of yield that a good season leaves unfilled. It sharpens the decision rather than settling it, which is the same reason it is worth bringing information to, not just a feel for whether fertiliser seems dear this year.

The second is where a station sits. This year the three stations point much the same way, all tracking a little below to around their long-term middle, so the reading is fairly consistent across the region rather than pulling growers in different directions. Moora, the driest, is where the downside case is closest, since a poor finish from an already low position is the scenario where nitrogen is most likely to be stranded. Badgingarra and Dandaragan West sit a little higher but not far enough to change the picture. In a year like this one, the tool is less about separating one station’s prospects from another and more about showing that none of them has yet had the season it would need to make a later application look comfortable.

That is the whole of what the tool offers this decision. Not an answer, but a clearer view of which way a grower is exposed. Where a season sits locally, and what nitrogen costs at the time of the call, both change the balance of what is at stake without deciding it. The point of setting them out is not to land on hold or push, but to see the exposure clearly enough that the decision is made on information rather than instinct. Even the clearest reading is something to weigh against the crop, the stored moisture and the paddock plan, not a rule to follow.

The limits of the tool

The tool is built to be read with its limits in mind, and DPIRD is clear about them. Rainfall to date is a useful narrative for how a season is tracking, but it isn’t the same as yield potential, and the tool doesn’t claim to be.

The things sitting between rainfall and yield have already been noted: stored moisture, soil type and water holding capacity, nitrogen in the profile, the timing of the rain. A few of them bear especially on how the tool should be read. Summer rainfall gives only a partial signal of stored soil moisture at the break. Non-wetting, waterlogging and leaching on lighter country each separate the rain that falls from the yield it produces, in ways rainfall alone will not show. A high rainfall position is not the same as a high yield, and the tool doesn’t suggest otherwise.

The projections carry their own caveat too. The decile 1, 5 and 9 finishes are drawn from the historical record for each station, not from a forecast. They describe how seasons have finished from a similar mid-July position in the past. They are not a prediction of how this one will, and a season can still do something the record has not seen.

None of that makes the tool less useful for the top-up decision. It sets what the tool is for. It puts a defensible range around a season that is still open, so a grower is weighing a real spread of outcomes rather than a single guess. The narrowing of that range, from rainfall to stored moisture to what the crop is actually showing in the paddock, is work the grower and their agronomist do from here.

What the tool is for

The Rainfall to Date tool will not make the nitrogen call for anyone, and it’s not meant to. What it offers is a way to slow the top-up decision down and put some data behind it. Where does the season sit against the local record, how wide is the range of finishes still open, and which way is a particular paddock exposed if the season runs poor or finishes well.

That distinction, between a good decision and a good outcome, sits at the centre of what RiskWi$e is about. Whether a nitrogen call turns out right is usually only known after harvest, and it depends heavily on how the season finishes, which no one controls in July. What a grower does control is how the decision is made. A call reached by slowing down, looking at where the season sits, and weighing the exposure both ways is a better decision than one made on a feeling about the year, even if the season later turns and the numbers don’t come off. The aim isn’t to be proven right by the outcome – it’s to be able to say the decision was a sound one, based on what could be known at the time.

The tool is free and takes a few minutes to run for the station nearest a grower’s farm. It’s available through the DPIRD website at https://www.dpird.wa.gov.au/online-tools/rain-and-irrigation-calculators/rainfall-to-date-tool/. For growers wanting to think through nitrogen decisions on their own paddocks, talk over how the season is shaping in the West Midlands or learn how they can be involved in RiskWi$e, please get in contact with the team at WMG.

This article was produced through WMG’s involvement in the RiskWi$e National Risk Management Initiative, led nationally by CSIRO and by GGA in WA and funded by the Grains Research and Development Corporation. The Rainfall to Date tool is developed and maintained by DPIRD.