11_11 Blogging 2

Communicating hazard and risk: 

Communicating hazard and risk Institute for Hazard, Risk and Resilience My experiences of blogging Timeline Good and the bad Attabad So why blog? Why, and what, you shouldn’t blog

Started as:: 

Started as: On 16 th December 2007

Why start a blog?: 

Why start a blog? Experiment interested to see what would happen Interested in possibilities for communication of science Raising awareness of own research Advertising / disseminating the Landslide Database Fun?

Topics: 

Topics Landslide events Research papers – reviews Conference reports Events and conferences Rapid reports of disasters Landslide videos Presentations and seminars Etc

Did it work? Citations according to Web of Science: 

Did it work? Citations according to Web of Science

PowerPoint Presentation: 

Attabad

PowerPoint Presentation: 

Slope identified as unstable in 2003

PowerPoint Presentation: 

Collapse event: 4 th January 2010

PowerPoint Presentation: 

1. Progressive failure over >7 years, scar height = 1175 m 2. Catastrophic failure event in Jan 2010 with no trigger 3. Emplacement of rockslide onto lake sediments triggered secondary high velocity mudflow event Cross valley runout = 1300 m Deposit height = 120 to 200 m

1858 landslide dam: 

1858 landslide dam

Wenchuan Earthquake barrier lake risk table: 

Wenchuan Earthquake barrier lake risk table Yang et al 2010

Data driven flood estimation: 

Data driven flood estimation Typical values of maximum discharge using this approach: 14,000 cumecs (Costa and Schuster 1986) 12,000 - 26,000 cumecs ( Walder and O’Connor 1997)

First awareness: 

First awareness

Field visit: 

Field visit Approached by Focus Humanitarian Assistance in February 2010 Undertook field visit to Pakistan in March 2010 including a trip up to the dam Presented outcomes and wrote a report for the authorities Presented to the National Disaster Management Agency and NGOs / International Agencies in Islamabad

Global distribution of landslide dams Ermini and Casagli 2003: 

Global distribution of landslide dams Ermini and Casagli 2003

Attabad landslide: 

Attabad landslide

1858 landslide dam: 

1858 landslide dam

Longevity of dams Ermini and Casagli 2003: 

Longevity of dams Ermini and Casagli 2003

Dam failure - floods: 

Dam failure - floods

Dam break flood, Peru: 

Dam break flood, Peru

Wenchuan Earthquake experience: 

Wenchuan Earthquake experience 257 barrier lakes 32 had potential for outburst flooding 17 required emergency action All were successfully mitigated through spillway construction that initiated outburst floods No loss of life due to precautionary evacuations

Water splash: 

Water splash Erosion from splash Probable second phase

PowerPoint Presentation: 

Landslide dam Scarp 800m 620m Landslide stats:- Length: 800 m Width: 620 m Potential lake volume: 315 m m 3 Volume : 20 million m 3 Lake

Wenchuan Risk Table Yang et al 2010: 

Wenchuan Risk Table Yang et al 2010

Current engineering works: 

Current engineering works Spillway construction is the correct approach Excavation rates are impressively high Plan to construct channel c. 30 m deep and 40 m wide Bed of channel will be lined with boulders to prevent erosion Intention is to ensure that collapse does not occur

Future scenarios: 1. Successful spillway: 

Future scenarios: 1. Successful spillway Pros: No outburst Cons: Loss of large section of KKH Isolation of upstream communities Future potential instability through earthquake or flood event

2. Spillway erosion and failure: 

2. Spillway erosion and failure Pros Hazard will reduce after flood If outburst is slow, limited downstream damage KKH can be reopened quickly Preparation for flood is possible Cons Potential for severe flooding downstream Sedimentation at Tarbela

Likelihood of spillway failure?: 

Likelihood of spillway failure? Positives: Wide channel Boulders in lower materials Negatives Materials (especially the lacustrine deposit) >2000 cumecs summer floods Height of dam Potential for downstream scour

3. Landslide into lake triggers failure: 

3. Landslide into lake triggers failure Comparatively rare, but this triggered failure of the 1858 landslide dam at Salmanabad Landslide into the lake → wave that overtops the dam → rapid collapse Probability of landslide increases as lake depth increases Probability of wave overtopping increases as lake depth increases

4. Piping / Seepage failure: 

4. Piping / Seepage failure Comparatively rare Water passage through dam creates a pipe that causes rapid collapse Quite unlikely in this case due to presence of mudflow deposit

5. Earthquake-induced failure: 

5. Earthquake-induced failure Vey rare but potentially catastrophic Seismic shaking causes mass movement of landslide material, triggering rapid release Location of landslide is a high seismic risk area – high potential in the long term

Critical concern is failure by erosion of spillway: 

Critical concern is failure by erosion of spillway Current engineering works may be successful in stabilising the dam However, there is a significant risk of an outburst event at this point Failure could be rapid and trigger extensive flooding downstream How big might that flood be? Analysis from known landslide failures Flood modelling NB both flood modelling and data-driven analysis are associated with high levels of uncertainty

Data driven flood estimation: 

Data driven flood estimation Typical values of maximum discharge using this approach: 14,000 cumecs (Costa and Schuster 1986) 12,000 - 26,000 cumecs ( Walder and O’Connor 1997)

Flood modelling: 

Flood modelling Two flood models have been run, based upon a 12,000 cumec peak discharge and a total flood duration of 24 hours BOKU (Austria) model uses Flo-3D University of Newcastle code uses a bespoke hydrodynamic code Base elevation model is SRTM NB this is very crude for this type of exercise Land-use analysed from Landsat data

Suggested alert states: 

Suggested alert states

Alternative approach: 1858 flood outcomes: 

Alternative approach: 1858 flood outcomes

1858 flood wave (From Prof. Ken Hewitt): 

1858 flood wave (From Prof. Ken Hewitt) NB the landslide dam was a little larger than Attabad 10 – 20 m above peak summer flows at Gilgit , 20 m at Chilas , 15 m at Attock , 10-15 m at Tarbela Induced extensive erosion of river terraces Reverse wave travelled 50 km up Kabul River NB this is not the flood that killed >1500 Sikh soldiers of the British Army That occurred in 1841 from a landslide dam from the Lichar Spur of Nanga Parbat

Preparedness: 

Preparedness Sensible precaution would be to prepare for an outburst flood when the water reaches the spillway (March / April?) Precautionary evacuations for those in potential direct path of a wave in upstream sections ( Attabad to Gilgit ) Preparation to move for those from Gilgit to Tarbela Who to move?

60 m river terrace appears not to have been affected by 1858 flood: 

60 m river terrace appears not to have been affected by 1858 flood

The 60 m terrace: 

The 60 m terrace 60 m above current flow = 30-50 m above peak summer flood Terrace deposits indicate that it predates the 1858 flood event This could be considered to be the safe level (but NB erosion and landslides)

Who might be affected by a flood?: 

Who might be affected by a flood? Those located within 50 m of the current flow level Those close to terrace edges Those living on or near to landslides Those located on bridges prone to overtopping or scour

Known landslide sites: 

Known landslide sites Between Attabad and Gilgit there are a number of known active landslides Great care is needed to ensure that the populations are safe There are further landslide sites downstream of Gilgit

Proposed actions 1: monitoring the current state: 

Proposed actions 1: monitoring the current state Monitoring is needed of: Seepage and movement of the downstream face of the landslide Is seepage developing? Is the water dirty (indicates scour)? Is there a downstream slope failure developing? The rate of rise of the lake level and the rate of inflow to allow forecasting of the date of flow in the spillway We should be careful to give realistic estimates The stability of the slope at Attabad Safety of the workers, integrity of the channel The state of the slopes above the lake Potential for failure Look for development of cracks, increased rockfall rate

2. Development of a series of alert states: 

2. Development of a series of alert states Level 1: landslide aware (current state) Inform affected population of potential for a flood Evacuation plans developed and disseminated Monitoring of dam state Level 2: Landslide alert Lake level is approaching spillway Precautionary evacuations for most endangered populations to Gilgit 24 hour monitoring, rapid dissemination plan in place

PowerPoint Presentation: 

Level 3: Landslide warning Water flow on spillway Evacuation of potentially evacuated population to Gilgit Population downstream made aware of potential need to move at short notice Level 4: Severe landslide warning Landslide starts to develop rapid erosion or downstream face collapse Evacuation of all population below 60 m level Closure of KKH – all vehicles and people moved to higher levels Emergency relief plan initiated

Staged evacuation plan: 

Staged evacuation plan At alert state 2: All population below 60 m above current river from Attabad to Gilgit All population on highly endangered terrace edges At alert state 3: All population on known landslides between Attabad and Gilgit All population below 60 m between Gilgit and Chilas ? At alert state 4: All population below 60 m between Attabad and Tarbela All population on endangered terrace edges and known landslides

4. Warning system and communications: 

4. Warning system and communications Active monitoring of dam state is essential Focus is locating a team with CCTV, but this needs government level assistance Need for active communication system to Tarbela Focus has undertaken considerable work between Attabad and Gilgit , but none below this point This does not mean that the risk downstream of Gilgit is low There is an urgent need to consider the population below Gilgit

All clear and long term monitoring: 

All clear and long term monitoring Erosion often starts some days after initial flow Need an expert group with agreed criteria to give all clear signal Need to consider slope failures above the lake Need for long term monitoring and agreed evacuation plans, and engineering works Critical points after initial emergency will be: Peak summer flow 2010 Exceptional flow events (e.g. 1:10 year, 1:100 year floods) During a seismic event

So what happened?: 

So what happened? Nothing, except some rather half heatered attempts to build a spillway....

Communications: 

Communications

Monitoring website: 

Monitoring website Data collected every day by in field team Posted online each morning Brief interpretation provided, including an estimated date of overtopping (correct to within a few days) Commentary via the main blog

Management of the hazard: 

Management of the hazard Initial management response was the construction of a spillway. Original intention: 30 metres deep Actual depth at time of overtopping: 15 metres Final spillway was very narrow and unlined

Evacuations: 

Evacuations After the landslide, Focus installed sirens in 15 villages People in most hazardous areas relocated into camps Boat service established on lake, but enormous hardship upstream One month before overtopping NDMA evacuated 15,000 people downstream of barrier

Seepage: 

Seepage

Seepage development: 

Seepage development

Prediction of date of overtopping: 

Prediction of date of overtopping NDMA consistently predicted the date of overtopping incorrectly (repeatedly predicted overtopping too early) However, final overtopping occurred later than expected because the lacustrine silts deformed, raising the floor of the channel

Water level rise: 

Water level rise

Overtopping behaviour: 

Overtopping behaviour

Evolution of spillway discharge: 

Evolution of spillway discharge

Spillway discharge – depth relationship: 

Spillway discharge – depth relationship

Spillway evolution: 

Spillway evolution

Problems: 

Problems Difficult relationship with NDMA Difficult to coordinate different points of view Direct criticism from some other academic parties: “You know already that your blog has been taken offline in Pakistan. It's because of the controversy on data you created, and because of comments you made or quoted from the press (I experienced how aggressive the local press is, they try to find a gap or error where they can creating a controversy, it's their business). I don't have your old blogs anymore (and had access to part of them only), but you made comments on the 'head of NDMA' and on 'the international expert'. These comments didn't help the situation. Blasting is not an easy task and you stirred the fire openly, as well as the 'suggestions' made to show the public that you know better what needs to be done... I had to learn in Gilgit that FOCUS / AKDN is running a parallel communication line, not shearing their observations with the officials nor with me (being on the official side), and you are a part of this system. Why? They are collaborating perfectly on the social rescue / recovery side upstream and downstream, but why this secretiveness on the Attabad site itself? Is it necessary that Focus is doubling the action there? How about 'transparent communication strategy', as you request.”

Communicating risk to the public via a blog: 

Communicating risk to the public via a blog Be professional – in particular try to ensure that the information you provide helps not hinders Be measured but realistic It does not help to hype the risks Be aware that journalists may use the material, and will often pick up on the sensational aspects Explain concepts carefully E.g. the difference between hazard and risk, or the difference between forecasting and prediction, is not clear to many people Use photographs and diagrams Invite contributions from the community, but vet them carefully Don’t duck key issues (climate change), but don’t get drawn into unwinnable discussions Correct errors Don’t blog primary research before publication

Advantages and disadvantages of blogging: 

Advantages and disadvantages of blogging Hugely raised profile (personal and institutional) Gatekeeper role Invites to attend meetings and conferences First to know Set the agenda (in part at least) Invitations to join international bodies Discomfort at conferences Precludes work on court cases Time commitment Danger of tripping up Value compared with conventional activities Misinformation Bar to publication

Some great blogs (in my opinion):: 

Some great blogs (in my opinion): Deltoid: http://scienceblogs.com/deltoid/ Eruptions: http://www.wired.com/wiredscience/eruptions Celebrating Science: http://celebrating-science.blogspot.com/ Geology in Motion: http://www.geologyinmotion.com/ Highly Allocthonous : http://all-geo.org/highlyallochthonous/ Real Climate: http://www.realclimate.org/ Climate etc: http://judithcurry.com/ Watts up with That: http://wattsupwiththat.com/ Climate Audit: http://climateaudit.org/ And some not so great (mostly climate) ( aagain in my opinion only):