A Dam Reaching Its Spilling Level Is Not the Dam Failing. It Is the Dam Working, and That Is Exactly Why the Town Below It Floods.
On the morning of 10 August a reservoir on the northern edge of Metro Manila was sitting at 80.06 metres. Its spilling level is 80.15. By nine that evening it read 80.27, which is to say it had gone over, and the people living along the river underneath it had been told so in the afternoon.
Nothing broke. Nothing failed. The structure did precisely what it was built to do, and that is the part that is worth understanding, because a dam going over the top is one of the few weather stories where the reassuring explanation and the dangerous outcome are the same sentence.
The photographs here are stock images used for illustration. They are not pictures of the dams or rivers named in the text.
The numbers from that day
La Mesa Dam sits above Quezon City. It is not a flood control structure in the way people assume, it is a water supply reservoir, and its critical level is 79.50 metres with a spilling level of 80.15.
At ten in the morning on 10 August it was at 80.06, already past critical and nine centimetres short of going over. It breached during the afternoon. At nine in the evening it was at 80.27 and by eleven it had eased to 80.22.
Nine centimetres is not much of a margin, and the residents downstream got told about it in roughly that order.
The warnings named the communities along the Tullahan River. Fairview, Forest Hills Subdivision, Quirino Highway, Sta. Quiteria and San Bartolome in Quezon City, then Caloocan, Valenzuela, Malabon and Navotas further down. Anyone who knows Metro Manila will recognise that list as the low ground.
Meanwhile Angat Dam, the one that supplies most of the capital's drinking water, went from 161.81 metres at eight in the morning on 9 August to 170.47 metres at eight the following morning. That is an 8.66 metre rise in a single day. Ipo Dam below it moved from 100.63 to 100.74 in the same period.
Of the nine dams the weather bureau monitors, six were rising and four were near spilling.
What spilling level actually means
It is a design number, not a failure number.
Every reservoir has an elevation above which it cannot hold any more water, and the structure is built so that when the water reaches that elevation it goes out over a spillway or through gates rather than over the embankment itself. That is the entire point of the spillway. Water arriving at the reservoir has to leave the reservoir, and the engineer's job was to decide where.
So "the dam has breached its spilling level" reads like a catastrophe headline and describes a dam behaving correctly. What it tells you is not that the dam is in danger. It tells you that everything now arriving in the catchment above the dam is going straight through to the river below, unbuffered, at whatever rate it happens to be arriving.
The dam has stopped being a sponge and become a pipe.
There is a second number that matters more to anyone downstream, and it is the gate opening. Operators frequently start releasing before the reservoir is full, precisely so they are not forced to release everything at once later. A controlled release announced in advance is a much better outcome than an uncontrolled one, and it is also the release that most often catches people out, because it happens on a day the operator chose rather than a day the weather chose.
A dam does not delete water, it reschedules it
This is the thing almost everyone gets wrong, and it is worth being blunt about.
A reservoir upstream of your town does not reduce the total volume of water heading your way over the course of a wet fortnight. Rain that falls in the catchment leaves the catchment. What the reservoir does is change the timing, holding back the peak and letting it out more slowly, which turns a violent flood into a longer, lower, more manageable one.
That is a genuine and enormous benefit, and it works right up until the reservoir has no headroom left. Past that point the buffering stops. The next hour's rain becomes the next hour's release, and the town below sees the storm's actual intensity for the first time.
Which is why the worst downstream day is very often not the worst rainfall day. It lags. The catchment fills first, the reservoir fills second, the river rises third, and by the time the water is in the streets the rain that caused it fell somewhere else, a day or two earlier, possibly in weather you never experienced.
People standing in a flooded street under a clearing sky reliably conclude that the flood is unexplained. It is not. It is on a delay.
Why this month in particular
Because the rain never stopped long enough for anyone to draw down.
A reservoir operator managing a single typhoon has a fairly clean problem. The storm arrives, the storm leaves, the inflow spikes and then falls away, and there are dry days afterwards to bring the level back to something comfortable. A two-week monsoon surge does not give you those dry days. Inflow stays elevated, the level ratchets up, and every decision gets made with less margin than the last one.
That is how you end up with an 8.66 metre rise in twenty four hours at Angat. Not one dramatic hour, just relentless input into a catchment that was already saturated and could not absorb any of it.
It also explains why four dams were near spilling at once. They are all fed by the same weather. There is no diversification in a regional monsoon, and a system that is fine when one reservoir is stressed behaves differently when several are stressed on the same afternoon.
How to read a dam bulletin without a hydrology degree
There are only three numbers and one announcement, and the announcement is the important part.
Current level against spilling level tells you the headroom. If those two numbers are within a few tens of centimetres of each other, the reservoir has effectively stopped buffering and you should be reading the rest of the bulletin as though the dam were not there.
Rate of change over twenty four hours tells you how fast the situation is moving. A reservoir a metre below spilling that rose eight metres yesterday is in a different position from one a metre below spilling that rose ten centimetres.
The gate announcement is what you act on. Operators publish how many gates are open and to what height, and downstream local governments issue their own warnings off the back of it. That warning names streets and barangays. The reservoir number does not.
And check which river you are actually on. A dam matters to the valley below it and to nowhere else, which sounds obvious and is routinely ignored by people who see a dam story in the news and assume it applies to the whole city.
None of this is unique to the Philippines, incidentally. Japan releases from its dams on the same logic and publishes the same kind of notice, Korea does it on the Han and the Nakdong, and Laos does it on tributaries of the Mekong with considerably less warning attached. The vocabulary changes and the arithmetic does not.
What does change is how much time you get. A well run system will tell you hours in advance which barangays to move out of. A badly run one tells you when the water is at the door.
So the sentence to watch for is not the one about the dam. It is the one about the gates.
The numbers, and why two dams spilling are not the same event
The argument above is that spilling is a design behaviour rather than a failure. The published figures for the Luzon dams make that concrete, and they also reveal something the news coverage never separates.
Every one of these dams has a normal high water level, which is the level operators aim to stay at or below, and a spilling level, at which water goes over the gates. The gap between the two is the freeboard, and it is the amount of rain the dam can absorb before it has to pass water downstream.
| Dam | Normal high water (m) | Spilling level (m) | Freeboard |
|---|---|---|---|
| Angat, flood season (1 May – 31 Oct) | 210.00 | 217.00 | 7.00 m |
| Angat, outside flood season (1 Dec – 30 Apr) | 212.00 | 217.00 | 5.00 m |
| La Mesa, upper end of normal | 79.00 | 80.15 | 1.15 m |
| La Mesa, lower end of normal | 78.00 | 80.15 | 2.15 m |
| Ipo | 101.10 | 101.10 | 0.00 m |
Ipo Dam has no freeboard at all
Look at the last row. Ipo's normal high water level and its spilling level are the same number: 101.10 metres. There is no gap. Ipo spills the moment it reaches the level it is normally held at.
That is not a flaw. Ipo is a diversion structure rather than a storage reservoir; its job is to pass water on to the treatment plants, not to hold a season's worth of it. But it has a consequence for how you should read a headline.
"Ipo Dam is releasing water" and "Angat Dam is spilling" are reported in the same sentence structure and read identically to an ordinary person. They are not remotely the same event.
| Ipo releasing | Angat spilling | |
|---|---|---|
| What it means | The dam is at its normal level | The dam has absorbed 7 m above its flood-season target |
| How unusual | Routine in the wet season | Requires a very large rainfall event |
| What it tells you about upstream rain | Little | A great deal |
If you are trying to judge whether something serious is happening upstream of Metro Manila, Ipo releasing water tells you almost nothing and Angat spilling tells you a lot. A reader who treats the two alike will be alarmed by the wrong story about half the time.
How far Angat actually is from spilling
Angat's full operating range runs from a minimum operating level of 180.00 metres to a spilling level of 217.00 metres, which is 37 metres of range. That number is the reason "the dam is near critical" headlines are so often misleading: a movement of one or two metres is a few per cent of the range.
Taking a dated reading, Angat stood at 208.24 metres on 5 October 2026.
| Measure | Value |
|---|---|
| Level on 5 Oct 2026 | 208.24 m |
| Against flood-season normal high water (210.00) | 1.76 m below |
| Against spilling level (217.00) | 8.76 m below |
| Position within the 180–217 operating range | About 76% of the way up |
So on that date Angat was simultaneously "below its normal high water level", which sounds like a drought story, and "roughly three quarters of the way up its operating range", which sounds like the opposite. Both are true. Neither is news.
Earlier in the same year, in August 2026, Angat's level rose by about 8 metres during a stretch of monsoon rain. Set that against the 8.76 metres it sat below spilling in October and the scale becomes legible: a single sustained monsoon episode can move Angat by roughly the whole distance between a normal level and a spilling one. That is the honest reason to pay attention to Angat specifically, and it is a much better reason than any headline adjective.
The reading rule this produces
| When you see | Ask | Because |
|---|---|---|
| “Dam releasing water” | Which dam? | Ipo at 101.10 m is normal; Angat at 217 m is not |
| “Near critical level” | How many metres from spilling? | Angat has a 37 m operating range |
| “Above normal high water” | Which season's normal? | Angat's normal is 210 m in flood season, 212 m outside it |
| “Dam overflowed” | Is its freeboard zero? | Ipo's is; La Mesa's is about 1 to 2 m |
The one-line version: ask for the metres. Every one of these dams publishes a number, every number has a published spilling level to compare it against, and the gap between the two is the only part of the story that carries information.
Sources for this section
- Spot.ph, Luzon Dams: Your Guide to Their Spilling Levels and Where Their Water Goes (normal high water and spilling levels for Angat, Ipo, La Mesa, Magat, Pantabangan)
- MWSS, Dam Elevation (daily published levels)
- Tribune, Angat Dam water level near critical as major Philippine reservoirs record declines
- Manila Bulletin, Angat Dam water level up 8 meters amid habagat rains, 10 August 2026
- BusinessWorld, La Mesa Dam overflows, Ipo Dam releases water amid heavy rains
- Presidential Communications Office, Major dams below spilling level despite Nika – PAGASA
The freeboard column, the 37-metre operating range and the percentage position in the range are arithmetic on the published levels rather than separately published figures.
More on typhoons and storms
- PAGASA Rainfall Warning Colours (Yellow, Orange, Red, Class Suspension)
- Luis and Maymay Were Both Gone Within a Day. The Flooding They Left Behind Ran for Two Weeks, and the Thing Actually Doing It Has Its Own Name.
- How Long Power Stays Out After a Typhoon in the Philippines (Transmission, Electric Cooperatives, Restoration Timeline)
Comments
Post a Comment