Dam collapse: Was it human error?

Dam collapse: Was it human error?

The collapse of the Saddle Dam D of the Xe-Pian Xe-Namnoy Power Co a month ago killed 39 people, with moire than 1,100 still officially missing. (Photo provided)
The collapse of the Saddle Dam D of the Xe-Pian Xe-Namnoy Power Co a month ago killed 39 people, with moire than 1,100 still officially missing. (Photo provided)

The July 23 collapse of one of the saddle dams of the 410-megawatt Xe-Pian-Xe-Namnoy hydropower complex, situated on the mountaintop in the southeast section of Bolaven Plateau in southern Laos, demands a full clear answer as to what actually happened.

As in-depth investigation, led by the Lao government and possibly other parties involved, is perhaps already or may soon be under way, one aspect of the issue worth being looked upon in greater detail is rainfall.

Available satellite rainfall data can help establish whether the dam breach was caused by heavy rain, as claimed by the project's developers, or human factors. I have analysed such data and found the heavy rainfall last month was predictable and close attention should be paid to the human factor.

The dam complex is designed to collect rainwater from three catchments -- Xe-Namnoy, Xe-Pian and Houay Makchan -- with the main reservoir built in the first and largest catchment. Last month, continuous heavy downpour caused substantial water overflow at the main reservoir. After that, the excess water brought about a major surge and flooding in low-lying areas of Sanamxay district of Attapeu province, resulting in more than 30 deaths, nearly a hundred still missing, and thousands of people affected and evacuated. Those who live along the Sekong River in Cambodia's Stung Treng and Kratie provinces have also been affected but to a much lesser degree.

The dam construction began in 2013 or early 2014 and it was planned to be officially operational by next year. Reportedly, the construction was near completion prior to the incident, with water having filled the main reservoir several months before.

The Lao government confidently points to the likely cause of the incident being poor construction of the dam while some sceptics are ready to blame weather anomalies or even go as far as climate change. In fact, project development companies from South Korea, Thailand and Laos appear to have suggested heavy rain as the main cause of the incident.

In general, when it comes to design, pre-operational planning and management of a dam, rainfall is inevitably the most important factor to be taken into account. Given the current context, to assess whether heavy rainfall events existed over the period leading to the dam breach and how extreme it was is indeed relevant and crucial.

Further, have such events happened in the past? This is the important question and the answer to it could potentially exclude or point to heavy rain as being the natural factor behind the catastrophe, as opposed to a non-natural or man-made counterpart.

Advances in rainfall estimation by space-based remote sensing have benefited climate and hydrological research carried out by scientists and engineers for many years. Here, I have utilised satellite rainfall data to find out about the incident since there is no observed rainfall data at the dam area available.

Firstly, I reviewed daily rainfall data sets which are publicly accessible and then used the National Aeronautics and Space Administration's (Nasa) Tropical Rainfall Measuring Mission's (TRMM) data which covers information dating back to the early 2000s. The TRMM data is acceptable in terms of quality when compared to observed rainfall data at the government's official weather station in Pakse, a major southern city in Champasak province, 90 kilometres west of the dam area.

In southern Laos, it is not uncommon to have annual rainfall of as much as 2,000 millimetres. The peaks are usually in July and August, suggesting heavy rain is more often seen during these two months. Like most part of the Indochinese Peninsula, the rainy season of the region spans between May and October, contributing 80-90% of the total annual precipitation volume.

Since the dam complex is situated on a plateau, a steep change in elevation could possibly make rainfall more concentrated in the dam site than other low-lying areas.

My study looked into TRMM rainfall data at the Xe-Namnoy catchment during the pre-monsoon and monsoon months from March to October since 2001 to this year. I also compared different event types based on rainfall accumulation over a period of one to 15 days to represent the immediate, short-, and long-term impact of rainfall. In each type, all events were ranked by magnitude and their corresponding return periods were estimated. The term "return period" simply means how long on average a particular event takes to recur, repeat, or come back.

The study finds that not only did heavy rain exist over the course of two weeks prior to the July 23 incident but it was also extreme, with the July 21-22 downpour being the most intense. The historical rainfall patterns, especially the return periods, suggests that such extremes are expected to recur every five years or even more frequently. From a climatological point of view, this is considered a frequent recurrence of a climatic condition.

By looking at the top three precipitation levels during the 18-year period, each of them took place at least a couple of years apart. This implies the July 2018 heavy rain cannot be viewed as a "first-timer".

My analysis concludes that July 2018 did see extreme heavy rainfall but these are considered as predictable. It provides basic evidence and sufficient logic to rule out heavy rain, or weather/climate, as the culprit for the dam breach. In turn, it tips the scale toward the human factor in the dam development processes encompassing dam design, construction, and pre-operation dam management and supervision.


Kasemsan Manomaiphiboon is Associate Professor at the Joint Graduate School of Energy and Environment, King Mongkut's University of Technology Thonburi.

Do you like the content of this article?
COMMENT (7)