Integrated geohazard assessments to aid resilience of hydropower infrastructure

Just after 9 pm on the 5th July 2016 in northern Nepal, the ground began to shake, there was a sudden pungent smell of mud, and then a wall of water and debris came crashing through the evening darkness and struck the hydropower project’s headworks.  The 20-minute-long flash flood was surcharged with debris, including massive boulders 8m by 4m by 2m, weighing in excess of 150 tonnes, that were carried by the flow. The flood hit and overwhelmed the headworks diversion dam [Figure 1]. It stranded big boulders on top of the dam and jammed the biggest boulders into the top of the radial-gated sluices, which quickly blocked [Figure 2].  The flow split into three: one part overtopped the diversion dam (which was designed for this purpose); another swirled around the right abutment causing major erosion and ultimately complete failure; and the third, still heavily debris charged overwhelmed the desilting basin and explosively blew its retaining wall 50m across the river.  The deluge continued downstream until it reached the powerhouse site, where it rapidly raised the level of the river channel and inundated the building flooding it with water and silt. 

I have been involved since mid- 2017 in helping to understand what had caused the flood in the Upper Bhote Koshi River Valley and how it had become so damaging to the Bhote Koshi Power Company’s hydropower scheme.  It appears that the heavy rain over the few days prior to the flood had led to debris flows and landslides upstream in the Chinese part of the catchment.  A small debris flow triggered an outburst flood from a small glacial lake, which fed into the local rivers engorged by the heavy rain.  Local Chinese authorities stated that a landslide had also occurred, and that, in all likelihood, a short-lived dam had formed in the river causing perhaps more than 1-2Mm3 of water to impound behind it.  When this landslide dam failed, it released the catastrophic flash flood, with a discharge rate estimated to be ~2800m³/sec at the downstream headworks site and over 2300m³/sec at the powerhouse site.  The ground vibration and noise were from the boulders entrained in the flow crashing percussively against each other.  Locals reported that the noise, which continued through the night, was louder than what they had experienced during the two devastating earthquakes that struck the area in April and May 2015.

Natural disasters

I had previously been involved for The World Bank in forensic reviews of other natural disasters that had caused serious damage to hydropower infrastructure in the Himalayas.  Over US$3 billion in damage has occurred to hydropower schemes across the Hindu Kush-Karakoram-Himalayan Region in just four years (2014-17).  This included the cloudburst-generated floods across Uttarakhand Region in northern India, that killed over 5000 people and destroyed the 400MW Vishnuprayag HEP headworks.  I also examined the Jure landslide in the Sun Koshi River valley [Figure 3], Nepal, that overwhelmed a village killing 156 people and damaged the Sun Kosi headworks. The landslide formed a 40m high dam across the river, and formed an impoundment lake 3km long that completely submerged the entire powerhouse of the 2.6MW Sunkoshi Small hydro scheme as well as a local community [Figure 4]. A further six hydropower schemes (totalling 65MW) were also adversely affected by damaged transmission lines and headworks. 

I have also spent two years studying the glacial and geological hazards within the entire Upper Indus Basin upstream of the proposed 7.5GW Diamer Basha Dam project in northern Pakistan. When reviewing the work that had been undertaken over the previous two decades, along with documents for many other hydropower projects, it became obvious that a traditional hydropower approach to this scheme was entirely inadequate given the dynamic geological environment in which it might be built.

From the work I was commissioned to do as part of The World Bank’s recent initiative on Disaster Risk Management, coupled with other hydropower projects across the Hindu-Kush-Karakoram-Himalayan Region, I could see a pattern emerging where a comprehensive review had not been undertaken of what hazards existed in the respective upstream catchments of hydropower projects damaged by natural calamities.  The traditional approach for the hydropower industry is to commission separate studies, such as on: 

• Seismic Risk (i.e. will a proposed structure withstand a design earthquake?).
• Hydrological Risk Assessment (with there be too little or too much water during a project’s life time, including taking Climate Change into account?)
• Geotechnical Baseline Report (i.e. will the designed structure’s foundations be adequate and will the local site be geotechnically sound?).
• Glacial Hazard Assessment (i.e. is there a possibility of a Glacial Lake Outburst Flood (GLOF) from upstream?) Indeed even in heavily glacierised regions such as the Himalayas, this is rarely commissioned. 

Often these studies are undertaken over 10 or more years as the project goes through the various stages of pre-construction development (pre-feasibility studies, feasibility studies, Environmental Impact Assessment, detailed Engineering Design, and so on), during which time the underpinning science and engineering design methodologies can change.  For example, Dr Martin Wieland, Chairman, Committee on Seismic Aspects of Dam Design (International Commission on Large Dams; ICOLD) stated in December 2017 that “the pseudo-static analysis method [of seismic risk analysis for large dams] and the representation of the seismic hazard by a seismic coefficient are considered obsolete or even wrong and, therefore, shall no longer be used.”.  He went on to encourage “all dam owners to check the seismic safety of their dams and to use modern seismic risk methods for the dynamic analysis of dams.”.  

Critically, no one organisation takes an overview as to how these various aspects of a hydropower scheme interact.  Through my work on the Upper Indus Basin, I could see how these could all be drawn together and built upon in a more holistic and physically meaningful way by undertaking an Integrated Geohazard Assessment (IGA).   This approach has been adopted by The World Bank in its recent Hydropower Sector Climate Resilience Guidelines (Mott MacDonald, 2017) and its separate Nepal Dam Safety Guidelines (Hatch, 2017). 

An IGA involves examining all the reports produced for a given hydropower scheme as listed above and extracting salient information about the underlying geology and neo-tectonic structures, evidence of previous natural disasters (floods, landslides, GLOFs, earthquakes, etc.), and historical and contemporary seismicity.  It then involves reviewing detailed satellite imagery of an upstream catchment to identify pre-historical as well as historical geomorphic scars associated with previously unreported events so that a picture can be compiled as to the range and scale of geohazards that might exist within the catchment.  This is referred to as Event Mapping. 

For instance, in the Upper Indus Basin in Pakistan, features known as Massive Rock Slope Failures (MRSFs) had been identified around 2005, but their role in shaping the behaviour of entire catchments had not been appreciated previously.  A MRSF typically is a major slope instability that might be of the order of 2km across, more than 2km in height and involve 0.5->5km³ of material, with many subordinate landslides typically on its lower slopes. These features are not limited to the Karakoram but can also be found across the Hindu Kush-Himalayan region as well [Figure 5].  

Recognition of these features is critical in understanding the relationships between small-scale landslides and the much larger feature on which they are located and how they are all inter-related.  This can have major implications for hydropower schemes and access roads; recognising and monitoring such structures can be critical.  

The next stage is to analyse each event in terms of the processes at play.  All too often, hazard assessments are thought of erroneously as discrete single processes, e.g. an earthquake, a cloudburst, a landslide.  I have found that sites prone to disasters are often pre-conditioned by earlier events over geological time through to the present, and that the processes that occur are rarely discrete.  They more commonly comprise a complex multiple-process cascade of events that can magnify what might start out as seemingly innocuous to become devastating downstream.  For instance, earthquakes can trigger river-blocking landslides that that when they fail lead to landslide dam outburst floods, and it is the flood that impacts a hydropower scheme, not the effects of the earthquake.  Also, a cloudburst, prolonged monsoon rains, or typhoon can trigger landslides with the same kinds of devastating effects, such as in the Typhoon Morakot in August 2009 in Taiwan.  Heavy rain initiated a major landslide that dammed the Chi-San River.  Its collapse after only one hour led to 398 deaths in Hsiaolin Village, and to a debris-charged flash flood that flowed downstream wreaking havoc. 

So geological processes can lead to hydrological catastrophes, and hydro-meteorological processes can lead to major slope instabilities and geological disasters.  A carefully constructed Integrated Geohazard Assessment can tie together these different strands so that the style of geohazards can be better understood and a refined Disaster Risk Management produced.  This leads to a more resilient hydropower scheme that is better able to withstand the dynamic processes within an upstream catchment as well as at the project site.  Ultimately, this risk reduction approach can save HEP scheme owners/developers substantial amounts of money through reduced ongoing maintenance, physically more robust and better designed schemes, more appropriate insurance policies (with perhaps lower premiums).  It provides advice about ongoing monitoring of sensitive constituents within a hydropower scheme as well as within the upstream catchment.  

It can be conjectured that, had an Integrated Geohazard Assessment been undertaken at Oroville in the US, the previously unnoticed ‘mischaracterised’ ground conditions might have been identified, and appropriate monitoring and remediation measures put in place to avoid dam failure. Through better Emergency Response Plans, speed of recovery is commonly achieved meaning that a project can get back to generating electricity and hence revenue more quickly. 

Demand

Owners/developers for too many hydropower projects are paying lip service to consideration of what processes could negatively impact their hydropower scheme.  It is seen as a ‘tick-box’ exercise to satisfy a Financier’s Technical Adviser’s concerns so that access to substantial funds is granted. Financiers, investors, and insurers, to name but three interested stakeholders, should ask whether the hazards that could impact any scheme have been thoroughly investigated and, if not, they should demand that an Integrated Geohazard Assessment is undertaken as part of both best practice and sensible due diligence.  

The author is Professor John M. Reynolds, Owner/Managing Director, Reynolds International Ltd, Suite 2, Broncoed House, Broncoed Business Park, Mold, CH7 1HP, UK. Email: jmr@reynolds-international.co.uk

www.reynolds-international.co.uk