On the evening of 8 August 2026, floodwater on Dr. A. Santos Avenue rose from shin-deep to a motorist's waist in under four hours. There was no tropical cyclone anywhere inside Philippine waters. Not a typhoon, not a storm, not even a depression.
That is the finding Parañaque should sit with. The bar for closing this road with waist-deep water is no longer an extraordinary event. It is an ordinary monsoon evening falling on ground that is already saturated, and the habagat runs from June to September every year. On the radio the next morning, the first thing the anchor put to the district's congressman was that residents were complaining “parang ganyan na naman”. There it goes again.
Every figure below is measured, sourced and reproducible. It is published in the hope that the next version is shorter, because the problem was fixed rather than re-documented.
Nothing was inside the Philippine Area of Responsibility on 8 August. The habagat was being enhanced by Typhoon Dolphin, sitting roughly 1,095 km northeast of extreme northern Luzon, outside PAR, and never forecast to enter it.
Between 6 and 7 p.m., Alabang recorded 42.4 mm and BF Homes 37.6 mm, both above the 30 mm/hour red-warning line. Alabang's three-hour total of 97.3 mm cleared the 65 mm red criterion on its own.
BF Homes took 148 mm between 4 and 7 August. The 8 August rain fell on saturated ground and part-full drains. The same rainfall on a dry week would not have produced the same flood.
Manila Bay was at its lowest tide of the day in the early afternoon and still well below high water through the evening peak. The outfalls were open. The water still backed up, which points squarely at drainage capacity.
This is the finding that matches what residents already say. At 2:53 p.m. MMDA logged Villanueva Creek overflowing and 8 inches of water on Dr. A. Santos Avenue. At that moment the nearest gauge was recording 5.6 mm/hr and the next one 4.6 mm/hr, both below PAGASA's yellow threshold of 7.5 mm/hr. In the fifteen hours leading up to that point neither Parañaque gauge had recorded an hour above 11.7 mm: no orange hour, no red hour, and the road was already impassable. The evening burst that followed was 6.7 times the rate at which the drainage had already failed.
July 2025 dropped about twice as much total water on Parañaque (552 mm vs 257 mm), spread over half again as many raining hours (124 vs 85). Yet 8 August 2026 beat it on the measures that actually cause flash flooding: a higher one-hour peak (26.9 vs 24.6 mm) and a higher three-hour peak (53.3 vs 46.7 mm), and at Alabang a three-hour peak of 97.3 mm against 65.5 mm. The clearest measure is concentration: in 2026 the worst three hours carried 20.7% of the entire event's rain, against 8.5% in 2025. Less water, packed into far less time, and with no cyclone in PAR.
The day had two separate downpours, not one. The late-morning pulse produced the flooding MMDA logged and the news reported. After a lull around 4 p.m., a second and much heavier burst arrived at dusk. That is the one that put water up to people's waists, and almost nobody reported it.
The rain gauges and the cameras were recording independently, and they line up to the minute. The heaviest rainfall hour of the day ran 6 to 7 p.m. Dr. A. Santos Avenue was filmed waist-deep at 6:37 p.m., thirty-seven minutes into it. The road was still impassable an hour later at 7:37 p.m., and passable again by 8:08 p.m. Nothing here rests on recollection: every flood time is a timestamp on a video, and every rainfall figure is a gauge reading.
“Hindi gumalaw ang mga kotse dahil hanggang bewang umano ’yung baha.”
Fourth Estate, Parañaque · video report by Rona Amparo, The Philippine STAR
The severity gap between the two downpours is stark. The midday flooding MMDA recorded on Dr. A. Santos Avenue was 8 inches, about 0.2 m. The evening flooding on the same avenue was waist-deep, roughly 1.0–1.2 m. That is about five times deeper, on the same road, under four hours later.
Almost all news coverage of the flooding falls between 11:12 a.m. and 4:23 p.m. The heaviest rain of the day fell after the last of it. The evening flood, the one that actually closed the road, was documented almost entirely by residents filming it on their phones. Without those posts the worst hours of 8 August would have left very little public record at all.
There was no tropical cyclone inside the Philippine Area of Responsibility on 8 August. Not a typhoon, not a tropical storm, not even a tropical depression. It still produced red-level rainfall over Metro Manila.
“Walang bagyo” does not mean “walang baha.” A quiet PAR map is not an all-clear. If a cyclone is sitting anywhere between northern Luzon and Okinawa during the June–September monsoon, it can drive heavy rain over Metro Manila without ever entering Philippine waters, and no storm signal will ever be raised, because signals are only issued for cyclones inside PAR. On 8 August, the warning that mattered was the rainfall warning, not a storm signal.
The rain came from the habagat, the southwest monsoon, the ordinary rain-bearing wind of the Philippine wet season. On its own the habagat gives grey skies and steady showers. What turns it dangerous is enhancement: a tropical cyclone somewhere else in the western Pacific tightens the pressure gradient across the South China Sea and drags a far wetter, faster stream of air over Luzon. The storm never has to come near us.
On 8 August that enhancing system was Typhoon Dolphin, about 1,095 km northeast of extreme northern Luzon, comfortably outside PAR, and PAGASA said it was unlikely to enter. Dolphin's only direct effect on Philippine soil was some rain over Batanes and the Babuyan islands from its trough. Everything that fell on Parañaque came from the monsoon that Dolphin was pulling in behind it.
This pattern is well established in the research literature: cyclones tracking through the northwest-Luzon-to-Okinawa corridor are a recognised driver of enhanced habagat rainfall over Luzon. It is predictable in principle. It is also exactly the case that storm-signal-based public messaging handles worst.
Metro Manila's reference point for monsoon flooding is 19–25 July 2025, when the habagat was enhanced by three cyclones inside PAR: Crising, Dante and Emong. That episode killed 34 people nationally and caused about ₱5 billion in damage. By total rainfall, 2026 looks mild beside it. By intensity, it does not.
| Measure | Jul 2025 Better Living | Aug 2026 Better Living | Jul 2025 Ayala Alabang | Aug 2026 Ayala Alabang |
|---|---|---|---|---|
| Event total | 552.2 mm | 257.3 mm | 608.1 mm | 410.0 mm |
| Hours with rain | 124 | 85 | 126 | 92 |
| Max 1 hour | 24.6 mm | 26.9 mm | 29.2 mm | 42.4 mm |
| Max 3 hours | 46.7 mm | 53.3 mm | 65.5 mm | 97.3 mm |
| Max 6 hours | 78.7 mm | 72.4 mm | 117.6 mm | 126.2 mm |
| Max 24 hours | 210.6 mm | 118.4 mm | 218.4 mm | 232.2 mm |
Read the two halves of that table against each other. In 2025 the rain kept falling for 124 hours at Better Living, five straight days of soaking. In 2026 it rained for 85 hours and delivered less than half the water. But the 1-hour and 3-hour peaks were both higher in 2026, and at Alabang the 2026 event beat 2025 on every window, including the 24-hour total.
One number captures the difference. In 2025, the worst three hours of a seven-day event carried 8.5% of its rain: the water arrived evenly, spread thin. In 2026 the worst three hours carried 20.7% at Better Living and 23.7% at Ayala Alabang. The rain was roughly 2.4× more concentrated into its peak.
This matters because urban drainage fails on intensity, not on volume. A drain sized for 30 mm in an hour copes with 200 mm spread over a day and fails at 42 mm in sixty minutes. The 2025 event was a slow flood of accumulation. 8 August 2026 was a flash flood, and it was produced without any cyclone in Philippine waters, which the 2025 event could not claim.
Rainfall explains why it was wet. It does not explain why Dr. A. Santos Avenue went waist-deep. Four things do, and only the first is weather.
BF Homes recorded 25.6, 38.4, 45.2 and 39.1 mm on 4, 5, 6 and 7 August: 148 mm in four days. Saturated soil absorbs almost nothing, so a far greater share of the 8 August rain became surface runoff than would have on a dry week. The same 147 mm falling on dry ground would have produced a visibly smaller flood.
Parañaque flooding is often explained away as a Manila Bay tide problem, and sometimes that is right. On 8 August it was not. Low tide fell at 2:26 p.m. at 0.07 m, the lowest water of the day, and the tide was still rising well below high water through the 6–8 p.m. peak. The drainage outfalls to the bay were as open as they get.
That makes the flooding more concerning, not less. The system failed on a favourable tide. Had this rainfall arrived near the 4:31 a.m. high water of 1.21 m, the outfalls would have been throttled and the same storm would have flooded deeper and drained slower.
PAGASA's red rainfall warning went out at 2 p.m., about four hours before the peak. The forecast was not the failure. Villanueva Creek had already overflowed by 2:53 p.m., MMDA had a crew and a mobile pump on Dr. A. Santos Avenue, and the road still went waist-deep by 7 p.m. What ran out was the capacity of the channels and drains to move water, not the time to prepare.
Residents in this part of Parañaque will tell you the road floods even when the rain is unremarkable. The gauge record for 8 August supports them, and it does so at the one moment that can be pinned to an independent report. Here is what was actually falling when MMDA logged that first 8 inches.
| Station | km to Fourth Estate | That hour | PAGASA band | Since midnight |
|---|---|---|---|---|
| BF Homes | 1.68 | 5.6 mm/hr | below yellow | 53.2 mm |
| Ayala Alabang | 2.64 | 11.2 mm/hr | yellow | 90.0 mm |
| Better Living | 3.14 | 4.6 mm/hr | below yellow | 37.7 mm |
At the two gauges inside Parañaque the rain was falling below PAGASA's lowest warning threshold when the creek gave way. Nor was this the tail of something fiercer: across the fifteen hours from midnight to 2:53 p.m., the heaviest hour either Parañaque gauge saw was 11.7 mm, inside the yellow band. No orange hour, no red hour. The drainage on Dr. A. Santos Avenue had already failed before any rainfall worth warning about arrived.
The contrast with what came later is the whole point. Those same two gauges went on to record 37.6 and 26.9 mm in the 6 p.m. hour, roughly 6.7 times the rate at which the road had already gone under.
It does not say 5.6 mm in an hour floods this road. Rain does not arrive one hour at a time. By 2:53 p.m. the catchment had taken 53 mm since midnight on top of 148 mm in the preceding four days, and the channels were carrying that already. What it does say is that the failure point was reached on ordinary monsoon rain, with nothing on the gauges that any warning system would have flagged. The extreme evening burst did not create this problem. It measured it.
Parañaque sits low and flat on the Manila Bay coastal plain, where land subsidence compounds an already marginal gradient, conditions under which drainage depends on pumping and on clear channels rather than on slope. DPWH has been surveying long-term interventions along this exact corridor, and dredging of clogged waterways in the city began in May 2026.
The day after the flood, Rep. Brian Yamsuan of Parañaque's 2nd District described the drainage problem on DZMM. His account is not measurement and this report does not treat it as such. It is worth setting beside the gauge data anyway, because he names specific physical obstructions, and because the areas he lists are the ones our stations cover.
“Maynilad may proyekto, biglang nakaharang pala do'n sa drainage 'yung mga pipes nila… Tapos merong mga drainage naman na may mga biga, kasi 'yung mga overpass na ginawa do'n sa areas namin, kailangan ng pundasyon, nalagyan ng biga. So, instead na dumaloy nang deretso 'yung tubig, kalahati lang or minsan 1/4 lang 'yung dinadaluyan.”
Rep. Brian Yamsuan · DZMM, 9 August 2026
In English: utility pipes laid across drain lines, and foundation beams poured for overpasses, leaving some channels carrying water through a half or a quarter of their designed cross-section. If that is accurate, it is a capacity ceiling built into the infrastructure itself. It would not show up in any rainfall record, and no amount of forecasting lead time would move it.
He also says both creeks were dredged three to five metres before this event, Villanueva among them. Set that against our timeline and it produces the sharpest finding in this section: Villanueva Creek overflowed at 2:53 p.m. anyway. The channel had been deepened, and it still could not carry 8 August. What he blames is what refilled it.
“Alam mo ba nung nagde-dredge kami diyan sa Villanueva Creek, alam mo sa nadedredge namin diyan? Refrigerator, kama, sofa!”
Rep. Brian Yamsuan · DZMM, 9 August 2026
Dredging raised capacity; rubbish took it back. He describes trash traps silting up again after the work, informal settlements along the creek banks obstructing flow, and two mobile pumping stations plus a mini amphibious dredger deployed after the flood, one of them sent to Pascual Compound in Barangay San Antonio, the same barangay as Fourth Estate, where water would not recede on its own.
Yamsuan states that following “'yung nangyaring iskandalo nung nakaraang taon”, the scandal of the previous year, no flood control projects were released to any district, leaving DPWH central office carrying the work alone. This report has no way to verify that. It is recorded here because it is the district representative's own explanation for why long-term drainage work in Parañaque has stalled, and because it is checkable against public budget records by anyone who wants to.
Read plainly, his account and our gauges agree on the conclusion and differ only in what they can prove. The rainfall data shows a drainage system that failed on a favourable tide with four hours of warning. He describes a system carrying water through a quarter of its designed section, refilling with refrigerators, and waiting on a national budget line. Those are the same story told from two ends.
This report is built on personal weather stations: amateur gauges, not calibrated to any common standard, and none trustworthy alone. They are also the only way to see this event at all, because PAGASA operates no rain gauge in Parañaque. So the question is not whether they are as good as official instruments. It is whether they agree with the official record where the two overlap. That is testable, and we tested it.
| Station | Location | km to Fourth Estate | Day total | Status |
|---|---|---|---|---|
| IMUNTI12 | BF Homes, Parañaque | 1.68 | 147.3 mm | Used, primary |
| IMETROMA22 | Ayala Alabang, Muntinlupa | 2.64 | 224.0 mm | Used, primary |
| IPARAAQU3 | Better Living, Brgy. Don Bosco, Parañaque | 3.14 | 117.3 mm | Used |
| ITAGUI1 | Villa Juliana, Brgy. Calzada Tipas, Taguig | 9.82 | 134.4 mm | Set aside, too far |
| IPASIG9 | Brgy. Manggahan, Pasig | 16.73 | 76.4 mm | Set aside, too far |
| ILASPI36 | Estrella, Las Piñas | 3.53 | 4.3 mm | Excluded, gauge failure |
The nearest official synoptic station to Parañaque is NAIA (WMO 98429), whose observations are published through the global SYNOP network. Its daily totals are reported for the 24 hours ending at 00:00 UTC, which is 8 a.m. Philippine time. To compare like with like, the station records below have been rebased onto that same 8 a.m. to 8 a.m. clock rather than local midnight.
| Window ending | NAIA official | BF Homes | Ayala Alabang | Better Living | Taguig | Pasig |
|---|---|---|---|---|---|---|
| 5 Aug | 23.0 | 25.1 | 21.6 | 25.9 | 24.4 | 19.9 |
| 6 Aug | 50.0 | 38.9 | 49.3 | 43.7 | 53.3 | 51.5 |
| 7 Aug | 40.0 | 51.6 | 58.4 | 40.6 | 46.8 | 32.5 |
| 8 Aug | 41.0 | 39.1 | 66.0 | 34.8 | 56.1 | 31.8 |
| 9 Aug flood night | 173.0 | 140.7 | 213.1 | 109.2 | 122.2 | 63.2 |
On every one of those five days the official figure falls inside the spread of the amateur gauges. There is no systematic bias in either direction: NAIA reads higher than some stations and lower than others, and which ones change from day to day. That is what a healthy network looks like, and it is the basis on which the rest of this report asks to be believed.
The flood night is the exception worth staring at. The official station recorded 173 mm. Across six gauges in one metropolitan area the range was 63 to 213 mm. That is not instrument error. It is how convective rainfall actually falls, and it is why no single number can honestly be called "the rainfall in Parañaque" that night. It also means the flooding at Fourth Estate cannot be reconstructed from any one gauge.
The flooding was filmed at 14.4609 N, 121.0278 E on Dr. A. Santos Avenue. Distances in the table above are measured from that point, and they decide which gauges this report leans on: BF Homes at 1.7 km and Ayala Alabang at 2.6 km are the two nearest working stations, so they are treated as primary. Better Living, at 3.1 km, is retained for a different reason: it is the only Parañaque gauge with a continuous record going back to the July 2025 habagat, which is what makes the historical comparison possible at all.
Proximity is not the only argument. Fourth Estate floods from Villanueva Creek, so what matters hydrologically is rain falling on that creek's catchment rather than on the nearest rooftop. BF Homes and Ayala Alabang sit on the higher ground draining toward it, which means the two criteria agree here. Where they disagree, catchment should win.
Taguig and Pasig are shown in the cross-check above, where more stations make the validation stronger, but they are set aside everywhere else: both sit well outside the catchment and neither tells you anything about what fell on Villanueva Creek.
ILASPI36 reported 4.3 mm on 8 August while a station 4 km away reported 117.3 mm. That is not a rainfall gradient; that is a broken instrument. The station tracked its neighbours correctly from 1 to 5 August and then diverged sharply from 6 August onward, the signature of a rain gauge that clogged or jammed. It is excluded from every figure in this report. We are publishing this rather than quietly dropping the station, because a reader who checks the raw feed deserves to know why the number they can see does not appear here.
IPASIG9 stopped reporting hourly data after 6 p.m. on 8 August. Its 76.4 mm day total is therefore an underestimate that misses the entire evening burst, and so is its 63.2 mm figure in the cross-check table above. Read that one cell as a floor, not a measurement.
What would close these gaps: barangay-level flood-depth logging against fixed reference marks, an official Parañaque rain gauge with maintained calibration, and drainage and pump-station telemetry that is published rather than held internally.
The evidence points away from forecasting and toward capacity and communication. The warning arrived four hours early; the water still won.
There will be no storm signal for an event like 8 August, because there is no storm in PAR to signal. During June–September, check the rainfall warning colour, not just whether a typhoon is coming. Orange means flooding is a definite threat; red means it is an emergency.
The single largest gap in this report is that the rainfall is measured to the millimetre while the flooding is described in inches from a handful of reports. A standing instruction for barangay staff to record depth, location and time at fixed reference points would make the next review far stronger, and would let the city prove which interventions worked.
8 August failed at roughly 27–42 mm in one hour on a low tide with the outfalls open, and it had already failed hours earlier at under 6 mm in an hour. Those are the two design cases to work from. Prioritise the Dr. A. Santos Avenue corridor and Villanueva Creek, where every flood report from that day was concentrated, and publish the design capacity of the corridor's drains and pumps so the next event can be assessed against it.
148 mm in the four preceding days is why a 147 mm day flooded as badly as it did. A running multi-day rainfall total is a cheap, powerful addition to local flood readiness. It is the difference between "heavy rain expected" and "heavy rain expected on ground that cannot take any more."
| Hour (PHT) | Better Living | Ayala Alabang | BF Homes |
|---|
Rainfall figures are derived from Weather Underground's PWS history endpoints.
Each station reports precipTotal as a running daily counter that resets at local
midnight; hourly values here are consecutive differences of that counter, with resets handled
explicitly. Event statistics (max 1/3/6/24-hour) are maximum rolling sums over the hourly series
across the whole event window, not per-calendar-day maxima.
PAGASA thresholds applied: yellow 7.5–15 mm in one hour; orange 15–30 mm; red above 30 mm in one hour, or rainfall continuing past three hours exceeding 65 mm. Tide figures are predicted heights for Manila. Flood depths are as reported; "waist-deep" is rendered as 1.0–1.2 m for comparison only and is not a survey measurement.