In March 2021, large parts of coastal New South Wales received a week's worth of rain that, in places, exceeded anything recorded there since national daily rainfall records began in 1900. It's worth looking at what actually drove it, because the setup was slightly different from the systems that usually flood this part of the coast.
The system
Unlike most major Sydney flood events, this one wasn't a classic East Coast Low. It was driven by a blocking high pressure system parked in the Tasman Sea, which held a low-pressure trough in place off the NSW coast and fed it a sustained stream of warm, moisture-laden air for nearly a week. Because the high refused to move on, the trough and its associated rain bands stayed locked over the same stretch of coastline day after day, rather than sweeping through and clearing the way a typical front does. That persistence, more than the intensity of any single burst of rain, is what made the event so severe.
The rain began on 18 March in the Mid North Coast, spread south to the Illawarra and Sydney by 21 and 22 March, and continued through to 25 March, by which point it had largely cleared. Over that period, the Bureau of Meteorology described the event as "volatile, dangerous and dynamic," and confirmed that the week ending 24 March 2021 was the wettest week recorded anywhere on the NSW coast since national daily rainfall records began in 1900.
How much rain actually fell
The totals were substantial even by flood-event standards. Mount Seaview, in the Hastings catchment on the Mid North Coast, recorded 907mm over the week, close to four times its entire average rainfall for the month of March. Kendall, near Port Macquarie, recorded more than 400mm in just 24 hours between 19 and 20 March.
Closer to Sydney, the totals feeding into the city's water catchments were what actually turned this into a flood emergency for the metropolitan area: 503mm at the Blue Mountains, 348mm at the Upper Nepean, 307mm at Prospect, 281mm at Warragamba, and 260mm in the Shoalhaven, largely concentrated in the 24 hours to 9am on 21 March. That rainfall pushed every one of Sydney's water supply dams toward capacity, with most spilling. Warragamba Dam alone discharged around 500 gigalitres downstream, roughly equivalent to the volume of Sydney Harbour.
What that rain actually did
The Hawkesbury-Nepean river system, which drains much of the rainfall from Sydney's western catchments through a narrow, winding valley northwest of the city, bore the brunt of it. The valley's geography means floodwater can back up at natural choke points rather than draining away freely, a pattern authorities refer to as the bathtub effect, and it produced the system's highest flood levels in more than 30 years, with some measurements the highest since 1961. Around 18,000 people were placed under evacuation order and a further 62,000 under evacuation warning across Sydney, the great majority in the Hawkesbury-Nepean Valley. Further north, the Hastings and Manning Rivers around Port Macquarie and Taree also reached record flood levels.
While the Hawkesbury-Nepean flooding developed over days as the catchment filled, in line with the classic riverine flooding pattern described elsewhere in this guide, the event also produced genuine flash flooding in places, including in parts of Sydney's Parramatta River catchment, where no formal flood warning system existed but rapid, localised flooding still occurred. It's a useful real-world example of how a single weather system can produce both flood types at once, depending on the terrain it's falling on.
Why this event stood out
The Bureau's own forecasting during the event tells part of the story: early modelling on the Wednesday suggested only a small chance of Warragamba Dam spilling. By Thursday, the rainfall forecast had been doubled. By Saturday, it had been doubled again, and the actual rainfall that fell matched that final, far more severe forecast. That escalation across just a few days is a genuine illustration of how difficult extreme, slow-moving rain events like this one can be to forecast precisely in advance, even with only two or three days of lead time, because the persistence of the blocking pattern, not just the rainfall rate itself, was what ultimately decided how bad the event became.