Opinion Print edition: 2026-11-16

Tarbela’s T-5 accidents—what to do now?

Published Updated

Monday’s T-5 incident at Pakistan’s biggest Dam at Tarbela is the second in the last 15 months. On 9th November 2026, during excavation of 2nd Bench Connecting Tunnel, weak rock mass collapsed, water from Intake pit entered the tunnel and the under-construction powerhouse.

Reportedly, one worker was swept away, but luckily rescued. All workers were evacuated, and the work was suspended indefinitely. Here, it is important to note that WAPDA’s statement after the incident informed that “main tunnel, powerhouse, tailrace were undamaged”. However, the inquiries in the offing will check, and then we would know the facts. Earlier in July 2026, the then inquiry into the cofferdam collapse held the contractor, consultant and WAPDA jointly responsible for unauthorized design changes after contract. An omission of biblical proportions! Both of these disasters have made big headlines with a section of the media portraying the latest as the end of the Dam itself.

Surely, it is not so, but still all the above is considered as a classic mega-project failure pattern. Experts opine that for any mega project needs a 7-layer defence. Further saying that any dereliction would lead to more such incidents and which could further delay the mega project’s completion. The first being the principle relating to geotechnical studies, which suggests that old data can never be trusted. Here, it is important to understand that the T-5 failed on “weak rock mass”.

It must be understood that for Tunnel-5—originally a low-level irrigation outlet that was converted into a power tunnel after much controversy—the geology changes every 10 meters. Consequently, there was a mandatory requirement to probe 30 meters ahead of the tunnel face and conduct 3D geological mapping. This mapping needed to be performed by an independent geologist rather than just the contractor’s representative, whose qualifications had never been validated. Furthermore, real-time adjustments to the rock support design had to be implemented using the Q-system or RMR, as these requirements cannot be determined solely on the basis of the BOQ.

The second requirement was to ensure Design Integrity without any post-contract juggling. In this regard, the earlier July inquiry proved that the Contractor replaced the proposed RCC cofferdam with a weaker rock-fill structure without approval. This occurred at a time when the original project cost of PKR 82 billion had reportedly escalated to a staggering PKR 317 billion—a 285 percent increase. Any design change should have gone through the Design Review Board, alongside the Employer’s representatives, the Consultant, and an Independent Checker (the Lender’s Engineer, which is the World Bank in this case). In fact, work should have stopped if the change was not approved, yet it remains unclear why the alteration was allowed to proceed.

Next comes the issue of the stability of all protective structures, including the bund and cofferdam—especially since any breach in the protective bund could become an entry point for water. Experts consider that bunds / cofferdams must be designed for 1-in-100-year flood and also with a 1.5 factor of safety, with filter criteria and additional grout curtain. In addition, there has to be daily monitoring through piezometers and the state of art settlement markers.

Fourthly, construction supervision must be assured, including as a third-party check. Unfortunately, this is normally shunned by contractors and, sadly, by consultants too. This should also be the time to ensure detailed Contractor QA/QC and Consultant supervision, including third-party QA by the Employer/WAPDA. Additionally, there must be no excavation without a permit-to-work signed jointly by the Geologist, Safety Manager, and Shift In-Charge at the dam site. This kind of stringent three-way check is considered very important.

Fifthly, there must be serious emergency preparedness. Although the evacuation was successful and saved all workers’ lives, strict safety measures must become the norm for all underground work. These should include two escape routes, life jackets in the intake pit area, flood alarm sirens, muster points above the flood level, and daily toolbox talks on potential inundation. Additionally, dewatering pumps with 200% capacity should always be on standby. We further see that there could be gaps in the contractor management – specially, relating to the Chinese JV issue – the main contractor where language and possibly safety culture gap is a real threat. Consequently, we should provide bilingual HSE manuals and empower the WAPDA safety officer to stop work—along with a penalty clause for violations equal to 2 percent of the IPC, rather than just a warning.

Lastly, we have to look into possible gaps in both governance and transparency. Although, a hi-level inquiry has been ordered, an inquiry after the incident is too late. As such, there is a need to make the lender’s engineer report public on monthly basis. We need to install 24/7 CCTV cameras on the Dam site, arrange drone monitoring of bund and intake pit, etc. No work should be undertaken at night in case of any critical excavation, unless the CEO of the project’s SPV permits.

Here, it is also pertinent to provide details of the T-5 project when it originally started in August 2021. The details indicate that Tarbela’s T-5 tunnel will generate 1,530 MW (three units of 510 MW each) of hydel power by the rescheduled commencement date of May 2027. Additionally, the project will protect against Tarbela’s sediment delta—thereby extending the lifespan of the reservoir—provide greater flexibility for power generation, ensure timely irrigation releases, and produce an expected 1,347 GWh of annual energy. The original PC-I cost of PKR 82.4 billion, approved in 2016, was also noted. Furthermore, the project is expected to generate a large volume of carbon credits.

Further, it was stated in a report of 2024 that the T-5 tunnel stands closed for 33 months with effect from September 29, 2024. That a further closure too would be needed beyond the above 33 months is another story.

Turning to the lessons learned, the Tarbela T-5 debacle teaches us that in megaprojects, concrete fails less often than management does. It must be understood that the rock did not collapse that day; rather, our supervision collapsed first in July when we allowed unauthorized design changes. So, the need is to fix supervision and once that is done any possible issue of geology will be manageable. Furthermore, safety and security at the dam site will be assured. Similar to the power sector, protection modules must trip before any equipment gets damaged.

Copyright Business Recorder, 2026

Engr Tahir Basharat Cheema

The writer is President IEEEP