Transcript for Changing rainfall patterns in Victoria – a weather system perspective

Transcript. Acacia Pepler- Changes in rainfall patterns

[Speaker: Jacqui Lloyd (DEECA)]

So I can see you down on the screen. Thank you, Acacia.

Dr Acacia Pepler, senior research scientist at the Bureau of Meteorology. Her research focuses on how weather systems such as lows and fronts are changing in a changing climate and the implications for extreme rainfall and other hazards. Acacia, thank you for presenting to us.

[Speaker: Acacia Pepler (BOM)]

And can you hear me in the room?

[Speaker: Jacqui Lloyd (DEECA)]

Yes, we can.

[Speaker: Acacia Pepler (BOM)]

Excellent. I'm looking forward to someone changing the slides for me. All right.

So yes, I'm going to be talking about a weather system perspective on changing rainfall patterns in Victoria. And this is going to bring together both some work that was done in VicWaCI mostly a few years ago now, as well as how in the Australian Climate Service we are takingsome of the things that we developed in VicWaCI and then moving them more into the projection space.

And so there are lots of different ways to break down rainfall into different types of weather systems.
And you have heard of a lot of different weather systems from different people. For the purposes of this talk, I'm going to use the same way of defining the weather systems that influence Victorian rainfall that we used in VicWaCI.

And so this from a paper that was published in 2020. And So what we looked at was three main weather systems for Victorian rainfall. In the top left here I've got a synoptic chart showing a low-pressure system. So there's an area of low pressure than the surrounding background atmosphere with these closed contours around it. These are often called extratropical cyclones.

So I'll probably call these cyclones while I'm talking. And when they happen on the East Coast, they can also be called East Coast lows. The second weather system I'll talk about is in the bottom left. So this is a cold front. And a cold front is generally an area where we have to transition between sort of warmer air masses and more north-westerly winds to colder air masses and more south-westerly winds. And on a synoptic chart, they look like that blue line of the little triangle sticking out of it. And the third with the system that we're going to be talking about a little bit at least is thunderstorms, which are shown by a radar image on the right here. Now these are the very different spatial scale to the other with the systems. They tend to be much smaller and they're also tend to be shorter lived.

And so for this particular study, what we actually looked at were the environments favourable for thunderstorm activity. And you'll be familiar with that kind of approach if you've looked at the Bureau's severe thunderstorm warnings, where there's often like a yellow blob on the map, which is the area where severe thunderstorms might occur based on the environmental conditions.

But of course not at every point in that area will actually have a thunderstorm. So when actual thunderstorms develop, you see it on the yeah, on the radar, or you'll see that sort of red or black oval that the forecasters draw on the map. Importantly, as part of this, we're also looking at the interactions between these different weather systems.

Because obviously, as you can see in that figure from the top left, a lot of low-pressure systems have fronts within them, while other parts of cold fronts, such as in the bottom left are well away from a low-pressure system. And they can be quite different in their impacts.

And both of these systems tend to interact with thunderstorms. So for instance, we'll often have really severe cold fronts, we'll can have thunderstorms developing within them and causing the most significant rainfall. And we saw that in some of the impacts, for instance, in the Black Saturday event.

And the reason we talk about these different weather systems is each of them have different spatial scales. They've got different seasonality, they've got different patterns of what sorts of rainfall they cause and potentially different changes.

And by breaking down our rainfall into these different weather systems, ideally it helps us understand a bit more about what's happening in the current and the future, but also helps us know what we need to understand to have a better idea of why things are changing and what we have confidence in.

So I'll move to the next slide. And on this slide, I'm just looking at, again using the starter set for Victoria here, I've broken down all the rainfall in Victoria by month. On the top plot is the mean rainfall across Victoria, and it's broken down into these different 7 weather systems.

So the Blues and purples are lows and fronts on their own, and you can see that they have a much larger contribution to rainfall in the winter than they do in the summer. And so that explains a large proportion of why Victoria has the seasonality it does in total rainfall. The yellows here are thunderstorms on their own, which aren't particularly important for Victorian rainfall, except they're in parts of the northern border. And then in the Reds and oranges we have those times where a thunderstorm coincides of a lower front. And so those are really important and they happen throughout the year. And now on the bottom right, I've shown the same thing but for heavy rainfall, which here I'm calling days above 25 millimetres.

And so again, you can see that those lows and fronts are more in the winter than in the summer. And here you can see that front only, which is the sort of pale blue aren't very important for heavy rainfall.
You really need thunderstorms associated for front for it to view produce heavy rain, but those red Reds and oranges. So those combined thunderstorms with the synoptic scale, scale weather are particularly important for those heavy rainfall events and especially during that summer. So now that contribution is why we have a much more even seasonal distribution of heavy rainfall compared to what we see for total rainfall.

Alright, next slide. So we will also use these weather systems to try to help understand the drivers of the past changes in Victoria. And so I'm here, I'm showing April to October rainfall anomalies in Victoria. And for this study, we compared 2 periods marked by these grey lines. So 1979 to 1996 is the first line. It's a pretty average. In Victoria and then 1997 to 2015 is the second line and that's was a very dry period. In Victoria, it's missing the last 10 years. But as you can see, generally speaking, the last decade has also continued to be dry in that April to October period.

So on the right, I'vegot a similar plot. So again, for every month here, I'm showing the difference in rainfall between that second period, 97 to 2015 and the first period. And I've broken it down into the weather systems, but I've simplified it a bit.

So here blue is those low front combinations again, yellow is thunderstorm only and red is those reddish orange colours, those times where thunderstorm combines of a lower affront. And it's pretty easy to see that most of the decline is in April to October. We already know that, but that most of that decline is coming due to rain due from lows and fronts as the driver of that decline. And so this is due both to a decrease in the frequency of low-pressure systems in southern Australia, but also there's a decrease in the amount of rainfall produced by both lows and by fronts. And so in VicWaCI, I won't go into the details, we did ;look into why fronts are producing rain less rainfall than they used to.

All right, next slide. So that's the results from VicWaCI in which we use these weather systems to understand what's happened in the past in Victoria. So then the question is, can these weather systems also tell us something useful about future changes in Victoria? And here I'm going to go a lot more into ACS work, including some preliminary work.

So next slide. So here I've got and thanks to Francis for explaining all the regional clone models.
So I don't have to do that. But here I'm using the same 39 regional models that Francis mentioned showing on the left the projected change in May to October. Rain in Southeast Australia as a millimetres per degree of warming across our ensemble. On the right is the same for November to April. And dots indicate we're fewer than 80% of our regional models agree on the direction of the change.

And so on the left-hand side in May to October, we've got fairly good confidence that rainfall is going to decline in Victoria and particularly in the southern parts of the state, but that more uncertainty in the northwest. But on the other hand, we've got very little confidence in the summer, which is on the right where there's some sort of possible decreases in the South of Victoria, possible increases in the north, but generally little model agreement.

So moving to the next slide, part of this difference in the confidence between these different seasons I think can go back to what confidence we have in changes in weather systems. And a lot of this in particular is low pressure systems. So these figures here are from a paper that we published last year.

Again, ACS work looking at how the frequency of low-pressure systems is projected to change over the coming century from our regional cloud models. The left shows a map of the average change in the number of days of a low. Again, dots are where there's no model confidence. So you can see that there's a lot of model agreement that the number of lows is going to decrease in frequency across Southern Australia in a warmer world.

On the right is a different way of showing that. So I've taken that purple box, I've calculated an average. And so each of these box plots is for a different 20 year. So 2010 to 2030 is the first, then 2020 to 2040 and so on and so forth. And so it's the anomaly in each regional model represented by the little dots compared to the 1980 to 2009 mean. And so under high emission scenarios SSP 370, by the middle of the century pretty much no models sync will have above average numbers of lows. And then by the end of the century, we've got these very large declines in the frequency of low-pressure systems on the order of about 30%.

So we move to the next slide. We're also starting to look, and this is very preliminary work. It hasn't been published yet. We're also starting to do the same thing with cold fronts. And so for this one, I particularly need to thank Irena Rudeva and some of her team who've been actually doing all the front tracking. I've just taken her output and looked at it.

And so here I'm just using 14 regional clown models, the ones from the ACS. We don't have the NSW or Queensland models in this data set yet, but similarly I've got the projected change per degree of warming and dots are showing where this little model agreement.

So on the right-hand side in November to April, these models have very strong agreement that cold fronts are going to become less common in the warm season across southern Australia, including Victoria. And on the right, on the left-hand side in May to October, there's a strong degree of model confidence in a decrease near southwestern Western Australia as well as across a lot of sort of northeastern Australia. But there's not a lot of model of confidence on how frontal frequency will change in Victoria.

And so going to the next slide, we can now use these data sets to do the exact same thing that we did in the VicWaCI work. But for the future, we can look at how the cyclone only days, the front only days versus the interactions are changing. This will be much better if we added thunderstorms, but that's something to still to come.

There's a lot of people such as Andrew Dowdy, University of Melbourne and Daniel Udy at the Bureau working on the thunderstorm part. But for now, we can at least look at lows and fronts. And what you can see from the left two plots is that lows and the parts of the front that are embedded in the lows show that very clear decrease across most of southern Australia.

And then on the right-hand plot, if we take out just the parts of the front that aren't so embedded within a low, we can now see that different change. So we've got in parts of sort of southern and Eastern Victoria, the models are starting to suggest an increase in the frequency of these cold fronts on their own compared to the decrease on the cold fronts embedded in the lows, which I think is quite interesting.

So what does this tell us for rainfall? Well, we move to the next slide here. I've now broken down looking at the rainfall on the left. So just like I said before, the May to October change and rainfall per degree of warming, now I'm just showing it for these two sub ensembles. So the seven BARPA models on the top and the seven CCAM models on the bottom, BARPA tends to be a drier feature for the May to October, whereas CCAM was the wettest of all of the models in Victoria. And again, we'd love to have NSW in Queensland, but that's to come.

And so when you compare these two, you can see that BARPA has quite a strong degree of model confidence that rainfall will decline in the future, in May to October in Southern Vic and even in northern Victoria, whereas CCAM does have a decline, but it's less confident.

But now when we apply this weather system framework to it, in the second set of plots, you can see the projected change in rainfall from lows, including those embedded fronts. And what I hope you can see here is that actually there's quite a lot of model agreement in this.

So because the models agree that the frequency of lows is going to decrease, similarly the models agree that the amount of rainfall from lows is going to decrease. And so there's a lot more confidence in CCAM, and the patterns are much more similar between the two models. In the third and fourth sets of plots, we've got the rainfall from fronts on their own as well as the rainfall from other weather systems. BARPA thinks these are also going to decrease. CCAM has no idea what's going to happen to them.

And so this is sort of showing how this with the system framework is helping us to look at what we do know, which is that rainfall from lows will decrease and what are the areas of uncertainty. And it's really what's happening to the rainfall that's not related to a low.

Moving to the next slide, um so, that's what we can say about changes in total rainfall, especially in the cool season. But can the weather system with perspective also help us in understanding extreme rainfall?

Next slide, as Sandra said, we have a projected increase in the intensity of extreme rainfall across Australia and these are from a couple of recent reports. On the right-hand side is from the update to the Australian rainfall and run off on the left-hand side is from the recent national climate risk Assessment. And both are showing this projected increase in the intensity of your extreme rainfall events as the climate continues to warmth as an Australia wide number. With them generally suggesting larger increases for those short duration extremes of an hour or less and smaller rates of intensification for those daily scale extremes.

And on the next slide is some work we've been doing on in the ACS to look at this in a bit more detail. And again, this is unpublished, but we've taken on the X axis a larger range of durations going from one hour to three hour to six hour all the way up to 168 hours, which is 7 days. And then we've calculated on extreme value analysis to look at those annual exceedance probabilities. So there's rarer less than one, one less than annual events and how those are going to change with global warming again across our 39 regional cloud models.

And so here the shades show how much rainfall is expected to intensify per degree of warming and dots show where models don't agree on the sign of the change. And the main takeaway is that we've got this clear sign that there's going to be larger intensification of our short duration extremes consistent with the previous page and also higher rates of intensification for those rare extremes like the one in 20 year events or the one in 100 year event, which I think Sandra mentioned earlier as well.

All right, moving to the next slide. Unfortunately, one thing we don't have a lot of confidence about is where we're going to see the largest rates of increase, which I know everybody's always asking, like our rainfall is going to intensify more or less in Victoria than they are anywhere else. And unfortunately, this is still a big area of uncertainty.

And this is a figure from a paper there by Rowan Eccles last year. And they look at here the intensification of the 10% AEP of daily rainfall. And they've broken it down into those 4 ensembles. So we've got the Queensland models on the top left, CSIRO on the bottom left, BARPA at the top right and the NSW knock them at the bottom right and they use dots to show where the models do agree. But what you can take away from this is that on the left-hand models we've got largest rates of intensification in southern Australia, including quite robust agreement on intensification in Victoria in the top right. BARPA has its largest rates of intensification in the north and the east of the country.

So in Victoria, we've got confidence in Northern Victoria, but perhaps southern, Southern Vic. And then in NARCliM in the bottom right, we have very little confidence in the direction of change in Victoria compared to much larger rates of increase in Northern Australia.

All right. So on the next slide. So the question now becomes, can weather systems tell us anything to help us like narrow down this range of intensification of rainfall extremes?

And so just before I get into the weather system stuff, I'm going to here quickly look at the change in the number of days of rainfall about 1 millimetre, which is shown here in the left. And annually, you can see that this is robustly again expected to decline in Victoria, especially Southern Vic. And on the right, we've got the projected increase in days above the current 99 point 7th percentile. And so by definition, this occurs once per year. So a .1 increase in the number of days is equivalent to a 10% increase. And so here again, so I'm only using CCAM and BARPA for this. I don't have that NARCliM models included, for instance. But you can again see there's like high rates of intensification across Victoria, especially in Northern Vic.

So moving to the next slide. I've shown those figures on the left now, so we have the context. But now again, I've gone into our fourteen members of regional climate models and looked at that weather system perspective to see if it can tell us anything useful about what's going on with extreme rainfall change compared to changes in rain days and total rain. And so that's what these box plots are. I've broken out change in rainfall into that due to lows, including the front embedded in lows, the contribution from front only days and the contribution from all other weather systems.

And so on the left-hand side of box plots, you can see that we've got very strong agreement that the number of rain days into lows is going down and that's consistent with what I showed before. In the middle box you can see that there's a bit more agreement that front only rainfall will decrease, but that's particularly in BARPA. Whereas in CCAM there's a bit, some of the models have an increase in front only rainfall, which is what we saw on those, you know, changes in total rainfall plots, a few slides back and then for other rainfall there's no clear change either way.

But if we go into the right-hand box plot now I'm looking at how the changes in the 99.7% also those heavy rain days get broken down by weather systems. And in the two right hand boxes of this plot, you can see that generally speaking, the models agree that both front only and other rainfall are likely to become more frequent heavy sources of heavy rain days in the future. And while this says front only for those who are paying attention way back in the start of my talk, front only where the doesn't really cause extreme rainfall. So these are almost certainly fronts of embedded thunderstorms. So this is mostly thunderstorm related rainfall that's expected to increase.

But what I find really interesting in this plot is the box related to the cyclone or the low-pressure system rainfall. And here we've got this huge range where some models are projecting a small decrease in heavy rainfall linked to these low pressure systems and other models expecting a huge increase in the frequency of heavy rainfall linked to these low pressure systems.

So this is suggesting to me that one of the areas that we really need to understand better to understand how extreme rainfall is going to change in Victoria in the future is getting into what's going to happen to those really intense rainfall from cyclones, given that we know that cyclones overall are going to become less Tolman.

So I just moved to my summary slide here. So Victorian rainfall is dominated by lows in fronts, especially in the cool season. But thunderstorms are really important too when they interact with these, especially for heavy rainfall. And the frequency of lows has declined in recent decades and is projected to continue to decline. And this is the most robust cause of projected cool season rainfall declines in Victoria. On the other hand, we have lower confidence in changes in fronts and that contributes to some of our uncertainty and future changes. And yes, extreme rainfall is projected to intensifying future and this is both due to rainfall linked to thunderstorms as well as rainfall from lows.

Page last updated: 09/07/26