Last month I wrote up a June that broke the El Niño script and gave us 95.9 mm. July went and did the opposite, harder. The gauge collected 7.8 mm across the whole month, which makes it the driest July I have in the archive, and the biggest single fall was 2.6 mm on the 21st. Not one day in July managed 3 mm. What surprised me when I started to analyze it properly is that the air over us was not dry at all. There was moisture here the whole month. It just had nowhere to go.

What actually happened, in numbers

The station recorded 7.8 mm across 10 days that registered anything at all. Mean daytime top 23.5°C, mean overnight low 8.0°C. Warmest day 28.1°C on the 3rd, which for a July in this valley is a proper number. Coldest morning 3.4°C on the 26th. Six days made 25°C or better, and six mornings dropped below 5°C.

Mean pressure for the month came in at 1025.0 hPa, and the average daily peak was 1027.2 hPa. Both are the highest I have for any July. That single number explains more of this month than anything else I looked at, adn I will come back to it.

How it compares with every other July on the record

The median July here is 45.6 mm and the average is 64.8 mm. At 7.8 mm we got 17% of the median. The nearest thing to it in my record is July 2009 at 9.5 mm, then 2014 at 12.8 mm and 2016 at 19.2 mm. At the other end of the scale sit 2008 with 169.7 mm and 2022 with 149.6 mm, so this month is a long way from anything normal.

The statistic that got my attention is not the total though. It is the wettest day column. In July 2026 the biggest fall of the month was 2.6 mm. Go through every other July in the table and the smallest “wettest day” you can find is 5.1 mm in 2009. Every other July, even the dry ones, managed at least one day where something organised itself into rain. This July, nothing did. Ten days of the tipping bucket going tick, tick, and then stopping.

*2016 covers 27 of 31 days. The blank cells for 2020, 2021 and 2024 are deliberate. Those years came into the database from an older import where the monthly total is sound but the day by day split is not, so I will not publish rain day counts I do not trust. Temperature averages use days 1 to 30 for every year, sicne the 31st of this July was still running when I wrote this. No station data at all for 2013 and 2017 to 2019.

Record warm days sitting on top of completely normal nights

This is the split that makes July 2026 interesting rather than just dry. The daytime average of 23.5°C beats the previous best on my record, 22.9°C in 2023, and sits 2.4°C above the station July average of 21.1°C. That is a big margin for a monthly mean.

Now look at the blue line. The overnight average was 8.0°C against a station July average of 7.9°C. One tenth of a degree. Six mornings under 5°C, where the station average for July is just under six of them. Dead normal, in a month where the sky was clear for most of it and clear July nights in this valley are supposed to be cold ones.

My first thought was that the warm airmass simply lifted everything, days and nights together. Actually that is wrong, and the daily data shows why. Strip out the six days from the 17th to the 22nd and the remaining 24 mornings averaged 6.8°C, which is 1.1°C below the station average. The clear ridging nights were cold, exactly like they should be. One grey week in the middle carried the monthly mean back to normal all by itself.

Day by day: one grey week in an otherwise blue month

The first half of the month was a straight run of ridging weather. Clear, mild days in the 22 to 25°C band, cold starts, nothing in the gauge. The 13th and 14th gave 4.8 and 4.5°C mornings and the whole first fortnight produced 1.8 mm.

Then from the 17th the mood changed. Moisture came in on a southeasterly feed off the water, cloud filled in, and the mornings jumped to 15.1°C on the 19th and 15.3°C on the 20th. The 20th was the strangest day of the month: it started at 15.3°C and topped out at 17.3°C. A two degree range across a whole day, which means the cloud never once broke. The 21st gave us the big 2.6 mm.

After the 23rd it cleared out again and the cold came straight back. The 25th, 26th and 27th ran 3.9, 3.4 and 4.2°C, and I would put money on there having been ground frost in the low corners of both paddocks on the 26th, even though the screen at 1.2 metres only got to 3.4°C. I did not walk out to control it at dawn, so treat that one as my read and not a measurement.

The part that does not add up, and where the real answer is

Here is where I had to stop and rethink. A month this dry should be a month of dry air. It was not. The mean dew point for July 2026 was 10.8°C, which ties with 2023 for the highest July dew point on my record. Mean humidity ran 81%, second highest in the record. Afternoon humidity bottomed out around 55% on average, where July 2014 dropped to 38% and July 2025 to 45%. In plain terms: this July had more water in the air than most, not less.

Look at the shaded window on the chart. Dew point climbed from 7.7°C on the 15th up to 14.8°C by the 22nd, the moistest air of the whole month, and while it did that the pressure stayed pinned between 1024 and 1028 hPa. That combination is the whole story. The moisture arrived. The pressure meant nothing could lift it.

Under a strong ridge the air aloft is sinking, and sinking air warms and puts a lid over the top of the boundary layer. Moisture coming in underneath that lid gets trapped in the lowest kilometre or so. You get exactly what we got: heavy dew, fog on the flats, grey mornings, drizzle that wets the deck and evaporates by ten. What you do not get is rain, because rain needs the air to be lifted deep enough for the water to condense and keep condensing. Nothing lifted it. Ten days of drizzle, 7.8 mm total, and a wettest day of 2.6 mm is what a capped, moist boundary layer looks like on a rain gauge.

So why did the ridge sit on us for a whole month?

Three things, in rough order of how much I think each one mattered.

First, a negative Southern Annular Mode. The BOM had the SAM index negative through late July and forecast to stay negative until at least early August. Their own explainer for winter is blunt about what that does to us: in a negative SAM phase the belt of westerly winds expands north towards Australia, and for eastern Australia the northward shift of those westerlies “means less moist onshore flow from the east, and thus decreases rainfall for eastern Australia”. That is the mechanism, straight from the source. Less onshore easterly means the ridge sits further north over southeast Queensland and the moist feed off the Coral Sea gets throttled. The one time in July we did get a proper onshore feed, the 17th to the 22nd, is the only stretch that put anything in the gauge.

Second, El Niño. The BOM has it established since mid June and strengthening fast. As at the week ending 26 July the relative Niño3.4 index was +1.94°C against a threshold of +0.80°C, and the 30 day SOI was −28.0, which is strongly negative. Models are pointing at a strong to very strong event peaking in late spring or summer. A strongly negative SOI means higher pressure over the western Pacific and our side of the country, weaker trade winds, less convection anywhere near us. It stacks the deck toward exactly the ridging pattern we sat under. I will say plainly that El Niño's grip on our winter rain is looser than its grip on our spring, so I rate it as a strong background push rather than the direct cause of any particular dry week.

Third, warm water and a warm world. The BOM has sea surface temperatures above average around most of Australia, reaching 2 to 3°C above average off the New South Wales coast, and globally June 2026 was the warmest June on record at 0.51°C above the 1991 to 2020 average. That is your explanation for the daytime record. Airmasses that reach us have crossed warmer water than they used to, and under clear skies with a subsiding ridge overhead you convert almost all of that into daytime heating. 23.5°C average tops in July follows on naturally from a warm airmass plus sunshine plus nothing to spend the energy on, because with a dry ground there is no water to evaporate and evaporation is what normally holds a daytime maximum down.

The one to watch is the Indian Ocean Dipole. On 26 July the weekly index hit +0.44°C, the first week it has touched the positive threshold. It is still officially neutral, and one week is not an event, but the BOM says a positive IOD may develop through winter and into spring. If that lands on top of a strengthening El Niño, spring gets dry in a way that July was only a warm up for.

What this means on the ground

• Storage. 7.8 mm does nothing for a tank or a dam. June's 95.9 mm was the last real top up you got. Year to date I am on 876.5 mm against 1443.7 mm to the same point last year, so we are running about 40% behind 2025.
• Soil and pasture. Soil temperature at my probe averaged 15.7 to 17.3°C through July, against 14.8 to 16.0°C in the same month last year. That is only two years of comparison so do not read too much into it, but warm soil with no rain is the worst combination for pasture. It keeps the plant ticking over and burning reserves without giving it anything to grow on.
• Frost is still live. A 3.4°C morning on the 26th with the sky wide open is a ground frost in the low spots. August under a ridge like this can go colder than July did. If you have tender stuff in, do not relax yet.
• Fire fuel. June grew a lot of grass. July then dried it standing. That is the curing process running about a month early, and the fuel is sitting there right now in the paddocks and along the road edges. Slash it while there is still some green in the base.
• Watering. With 7.8 mm and warm afternoons, anything you planted in June has had no natural water for four weeks. Check under mulch rather than on top of it. The top 20 mm can look damp from the dew and be bone dry underneath, and in the weekend I found exactly that in the bed nearest the shed.

Get the numbers before the month is over

These monthly write ups are the rear view mirror. If you want the same data while it is still useful, that is what the Weather Reports membership on dayboro.au is for. It runs $3.95 a month, or $39 for the year, and it gets you the member weather reports plus the 7 day and 30 day outlooks, the forecast accuracy scoreboard where I publish where I got it wrong, the weather history reports and the rest of the hub. I built this station because the nearest BOM gauge is in another valley and it does not describe what happens here. That is still the whole point of it.

And if what you actually care about is growing food through a dry spring, go and have a look at gardenbuddy.au. I built that one too. It tracks the daily price of 76 vegetables across Woolworths, Coles, Aldi and IGA, currently sitting on about 107,000 price records, and it tells you which ones are cheaper to grow yourself and when to plant them for your climate zone rather than for somebody's garden in Kent. The price tracker, the planting calendar by postcode, the growing guides and the bed planner are all free to use with no signup. Create an account when you want it to start tracking your own beds and sending you the Monday note on which vegetables just got cheaper. With a positive IOD possibly forming, knowing what to plant and when is going to matter more this spring than it did last.

I will do the same exercise for August at the start of September. If the ridge holds, that write up is going to look a lot like this one, only with worse numbers.

Henk

Method note: Dayboro numbers come from my own station archive. Rainfall totals are full calendar month. Temperature, pressure, dew point and humidity averages use days 1 to 30 so that 2026 compares fairly against the other years. The record covers 2008 to 2026 with no data for 2013 and 2017 to 2019, and the 2020, 2021 and 2024 rainfall arrived from an older import where only the monthly total is reliable. Days with rain counts every day registering 0.2 mm or more. Brisbane comparison figures are BOM station 040913 daily weather observations for July 2026 and the 040913 and 040842 climate averages tables. Climate driver status is the BOM Southern Hemisphere Monitoring update issued 28 July 2026, and the winter SAM quotation is from the BOM's own Southern Annular Mode in winter page. Where something is my interpretation rather than a measurement, I have said so in the text.