impact-based severe weather warning system: a paradigm
TRANSCRIPT
Impact-based Severe Weather Warning System: A Paradigm Shift, example from
South Africa
Elizabeth Webster (Presenter)
Dr Eugene Poolman
South African Weather Service
IBF Workshop – Honiara 2019
04/10/2019Doc Ref no: FCAST-PRE-2018 IDRR Symposium-
20181107-001.12
Outline of presentation
1. History of SAWS Early Warning System
2. Current Warning System
3. Impact-Based Warnings
4. Impact-Based SWWS project in South Africa
5. How will Impact-Based warnings work?
6. Future work
7. Questions
04/10/2019 5
South African Weather Service Act (2001) as amended in 2013:
o Sets the environment for services based on free and cost recovery principles
oMandates SAWS to issue severe weather warnings
oDefines the role of SAWS as the single authoritative voice of
severe weather-related warnings in South Africa
oMandates SAWS to be the custodian of climate data information
Disaster Management Act (2002) and Disaster Management Framework (2004)
-identifies the role of SAWS in support of the national DRR environment
SAWS’ contributions to DRR: Legislation framework
04/10/2019 6
Forecasters:
• Monitor weather hazards using radar, satellite, observations and numerical weather prediction models
• Provide weather outlooks in DRR advisory forums and as requested
• Issue warnings for severe weather hazards to DRR structures and the general public (through media)
• Participate in JOCs prior and duringevents, providing specialised weather information
• Monitor progression of severe weather events, update warnings and shareinformation with DRR structures
• Continuously enhance the SWWS in collaboration with DRR structures
Severe Weather Warning Service (SWWS) within Multi-Hazard Early Warning System (MEWS)
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 7
History: Public Warning Hazards: (up to 2012)
FCAST-PRE-20130226.001.1 7
HAZARD DEFINITION1 Extremely heat Maximum temperature forecast 40°C and above2 Very cold Maximum temperatures 10°C and below and/or Minimum -10°C below3 High discomfort Discomfort Index (D.I.), meeting or exceeding 38°C (or 100 °F)4 Heat wave 3 consecutive days with maximum temperature to be more or equal to 5°C higher than
the average maximum of the hottest month for the particular station in the FPG region/s 5 Gale force winds
and strongerAverage wind speed of more than 34knots (62 km/h) or gusts in excess of 44knots for land based regions
6 Fire Danger Rating
If the fire danger rating is high or extreme according to the NFDRS work instruction
7 Heavy rain 50 mm or more within 24 hours 8 Flash Flood Flash flood as defined by SAFFG work instruction
Warning : Within three hours from time of forecastWatch : 3 to 6 hours from time of forecastAdvisory: Beyond 6 hours from the time of forecast
9 Snow Sufficient snow to cause significant traffic danger and/or disruptions in mountain passes, major roads and/or highways and/or populated areas
10 Severethunderstorms
Severe Thunderstorm in line with the USA NSSL definition with one or any combination of the following:- Hail of greater than 19mm diameter or large amounts of small hail - Tornadoes (any), - Wind gusts 50kts or more, in association with a thunderstorm
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 8
Severe Weather Warning Levels (from 2013):Reduced number of hazards to 7
No AlertSpecial Weather Advisory
Weather WatchWeather Warning
Be Aware! Be Prepared! Take Action!
Nohazardous weather expected
Early warning of potential hazardous weather
Weather conditions are likely to become hazardous
Hazard is already occurring or is imminent
2 to 6 days 1 to 3 days Next 24 hours
Intense large weather systemsExtreme heatHeat Waves Strong windsHighdiscomfort
Heavy rainDisruptive snowGalesHigh wavesSevere T/SFloodingVeldfireconditions
Heavy rainDisruptive snowGalesHigh wavesSevere T/SFloodingVeldfire conditions
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 9
• >3 days
Special Weather Advisory
• 2-3 days
Watch
• 24 hours
Warning
‘Current’ Warning System
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 10
‘Current’ Warning System
Hazard Alert Level Valid From (SAST) Valid To (SAST)Heavy rain Warning 22/03/18 06h00 22/03/18 23h00
is expected over the Gert Sibande District tomorrow.
Description: Flooding / Heavy RainFlooding occurs when water overflows its normal channels such as streams and storm water drains. It can occur with
prolonged period of rain, with continuous heavy falls or in the form of flash floods which are usually associated with
severe thunderstorms. Heavy rain may also result in river flooding causing damage downstream to areas that may receive
no rainfall at all during the flooding event.
Precautions: Flooding / Heavy RainIf possible stay indoors and off the roads, avoid crossing rivers and swollen streams where water is above your ankles. If
trapped in flooding in a vehicle, abandon it and climb to higher ground. In buildings, move valuables to a safe place above
the expected flood level. Switch off electricity at the supply point to the building. In rural areas protect/relocate animals
to a safe place on higher ground. Abandon your home immediately if evacuation is recommended, before access is cut off
by flood water. NEVER drive on a road covered by water. You do not know how deep it is or if the road has been washed
away. If the vehicle stalls, leave it immediately and seek higher ground. Be especially cautious at night when it's harder to
recognize flood dangers. Listen to the radio or TV for warnings and obey the instructions from disaster management
officers.
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 11
‘Current’ Warning System
- Only information is that heavy rain is
expected
- Questions???
- What is heavy rain?
- 50mm in 24 hour period?
- Heavy downpour?
- Standard warning
- Not clear
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 12
Moving from:
What the weather will be: (Meteorological thresholds)
- 50mm in 24 hours- 35 knot winds
What the weather will do: (Impact Warnings)- Roads flooded
- Communities cut off
To:
What is Impact-Based forecasting?
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 13
Why the change?
General public survey (2016 and 2017)• Lack of understanding (Special Weather Advisory, Watch
and Warning)• Level of response required for each alert not known• Translations
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 14
Subsequent SA Disaster Management Survey and meetings (in 2013)
• Challenge from Disaster Managers:�Clear requirement of matter-of-fact information, i.e
non-scientific and to the point – minimize need for complex interpretation
�The need for improving communication to users to enable their effective decision making
�“Just tell me what is going to happen, when, where and how serious it is”
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 15
Variation in vulnerability across South Africa
Densely populated areas
Large open areas Large rural areas
Main cities
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 16
SA: Country with large climatic diversity• Essentially a dry country: Rainfall vary from <100 to >1200 mm per year
• Summer severe convective rainfall over north-eastern two-thirds
• Tropical rainfall (tropical lows) Jan-Mar in north-eastern provinces
• Winter frontal weather over south-western parts
• Cut-off lows lead to many flooding events over southern half of country
- Winter frontal rainfall- Strong coastal wind
Winter snow
- Summer Severe Thunderstorms- Tropical rainfall systems
2019/10/04 17
Impact-based Severe Weather Warning System
Scenario 1: 23 Feb 2010
- Road breached - Several sheep drowned
Scenario 2:8 June 2011
- Hundreds displaced- Communities threatened- Severe damaged
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 18
Why Impact-based forecasting in South Africa?
• Huge variation in vulnerability of communities across country (from densely populated cities, vast rural communities)
• Severe weather conditions leading to serious impacts in many places and limited impacts elsewhere
Impact-based Severe Weather Warning System
Requirement:�Need improved communication to users to enable their effective decision making�“Just tell me what is going to happen, when, where and how serious it is”
Involves a paradigm shift:•From weather threshold based warnings (50mm in 24hrs)•To warnings based on variations in vulnerability - impact levels and the likelihoodof those impacts to occur
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 19
Current Warning: Heavy rain of more than 50mm in 24 hours is expected
We need to change from this to….
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 20
Warning: 1. Orange warning for rain with a medium likelihood of significant impacts2. Yellow warning for rain with a high likelihood of minor impacts
1
2
An impact-based Early Warning System
1
2
4 Year Project: Developing an
Impact-Based Severe Weather Warning Service
• Aim: Based on WMO recommendations, to develop, test and implement an Impact-based forecasting and warning service in South Africa, through close collaboration between forecasters and disaster managers
• Associated projects and sponsors:
� SAWS and NDMC (Developing Impact-based SWWS)UK Newton Fund WP 4 (High Impact Weather)USAID/NOAA (Advancing IDSS in SA)
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 22
Progress of ImpB SWWS in SA
OVERVIEW GANTT CHART: IMPACT-BASED SWWS (May 2018)
2015/16 2016/17 2017/18 2018/19
Activities Apr-Sep Oct-Mar Apr-Sep Oct-Mar Apr-Sep Oct-Mar Apr-Sep Oct-Mar
1 Project development
OUTCOME 1: Subjective decision making process
2 Develop Guidelines
3 Prototype test Phase (NFC and NDMC)
4 Summer Pilot Phase (all summer provinces) GT KZN,FS, etc
5 Winter Pilot Phase (WC, EC, KZN) WC, EC WC,EC,KZN
6 Community level stakeholder workshops (2:City and Rural) Jhb, EC
7 Public information and awareness campaigns
8 Operational implementation (preparation)
OUTCOME 2: Objective decision support system
10 Develop scientific DSS vs 1
11 Develop scientific DSS vs 2
General12 Wrap up workshop
13 Project conclusion
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 23
• Prototype developed and tested (2016):
– Forecasting process developed and tested in
case studies
Risk Matrix
Progress of ImpB SWWS in SA
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 24
Forecast Prototype Test Case study: 25 July 2016- Desktop case study
Impacts:
• 5 People died due to flooding
• Cars swept off bridges
• Houses and shops were
damaged due to floods as
well as wind
• Many passes closed due to
snow
Rainfall and Snow Data:
• Paddock, KZN: 351mm in 3
days
• Margate, KZN: 283mm in 3
days
• Tiffindell Ski Resort: Over 1m
snow
Airmass Image, 27 July
03:00 SAST
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 25
Decision Making Process Development
• Prototype developed and tested (2016):
– Forecasting process developed and tested in
case studies
• Stakeholder Workshops
– Covering each of the provinces
– Disaster managers from each District
– Introduce the concept of ImpB SWWS
– Develop and update Impact tables, using their
experience
– Agree on co-ordination mechanisms during pilot
phases
Risk Matrix
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 26
Progress of ImpB SWWS in SAStakeholder Workshops:
• Gauteng (December 2016)
• Western Cape (May 2017)
• Free State, Northern Cape and North West (August 2017)
• KwaZulu-Natal (October 2017)
• Limpopo and Mpumalanga (January 2018)
• Marine Workshop (March 2018)
• Eastern Cape (November 2018)
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 27
Progress of ImpB SWWS in SA
“Other” Stakeholder Workshops:
• Gauteng Outreach Workshop (January
2019)
• Eastern Cape Outreach Workshop
(February 2019)
• Media Workshop (April 2019)
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 28
• Prototype developed and tested (2016):
– Forecasting process developed and tested in
case studies
• Stakeholder Workshops
– Covering each of the provinces
– Disaster managers from each District
– Introduce the concept of ImpB SWWS
– Develop and update Impact tables, using their
experience
– Agree on co-ordination mechanisms during pilot
phases
• Forecaster Training
– Train all forecasters around the country for the
paradigm shift
Risk Matrix
Progress of ImpB SWWS in SA
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 29
• Prototype developed and tested (2016):
– Forecasting process developed and tested in
case studies
• Stakeholder Workshops
– Covering each of the provinces
– Disaster managers from each District
– Introduce the concept of ImpB SWWS
– Develop and update Impact tables, using their
experience
– Agree on co-ordination mechanisms during pilot
phases
• Forecaster Training
– Train all forecasters around the country for the
paradigm shift
• Pilot phases
– Started with one province at a time
– Targeted provinces according to season
Risk Matrix
Progress of ImpB SWWS in SA
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 30
Progress of ImpB SWWS in SA
Pilot phases:
• 2017 Summer Pilot
• 2017 Winter Pilot
• 2017/2018 Summer
• 2018 Winter
• 2018/2019 country-wide
• Present events during pilot phases
• Feedback of events (forecaster)
• Conduct Surveys assess SWWS
• Discuss co-ordination
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 31
Progress of ImpB SWWS in SA- Washout Workshops
How does it work operationally?
1. Forecaster analyses the current and
expected weather conditions
2. Compiles a forecast
3. Considers possible sensitivities and
vulnerabilities
4. Using Impact tables, decides on possible
impact level
5. Contacts disaster manager
6. Issues ImpB weather warnings
7. Verifies and evaluates the impacts with
disaster managers and media reports
Population Density
Sensitivities and vulnerabilitiesForecaster now considers the following
(paradigm shift):
• What areas is the rainfall expected to affect?
• Mainly farmlands?
• Densely populated cities?
• Area prone to these conditions?
• What time is the heaviest rain expected?
• During rush hour?
• Any pre-existing weather conditions?
• Recent rainfall
• Any major events expected to take place?
• Concerts
• Sporting events
• Elections
• Initiation
Rainfall Impact TableMinimal Minor Significant Severe
Business as usual
Localised
Business as usual
Localised
Short term strain on emergency
personnel
Widespread
Prolonged strain on emergency
personnel
• Some pooling of water
on roads or in
formal/informal
settlements
• Day to day activities not
disturbed
• Wet roads and reduced
visibility
• Minimal traffic
congestion
• Isolated mudslides and
rockfalls
• Localised flooding of susceptible
formal/informal settlements or
roads, low-lying areas and bridges
• Major roads affected but can be
used, increased travel times
• Difficult driving conditions on dirt
roads
• Minor motor vehicle accidents due
to slippery roads and/ reduced
visibility
• Closure of roads crossing low water
bridges
• Localised and short term disruption
to essential services (water,
electricity, hospitals, schools etc)
• Localised mudslides, rockfalls and
soil erosion
• Localised disruption due to
sinkholes/potholes/blocked drainage
systems
• Isolated cases of breakages of farm
dam walls
• Localised damage to mud-
based/make-shift houses/structures
• Localised disruption of access to
drinking water and damage to crops
• Flooding of roads and
settlements (formal and
informal)
• Danger to life (fast flowing
streams / deep water)
• Displacement of affected
communities
• Some communities temporarily
not accessible/cut-off
• Damage to property,
infrastructure, loss of livelihood
and livestock
• Major disruption of traffic flow
due to major roads being flooded
or closed
• Possible damage to roads and
bridges
• Disruption to essential services
(water, electricity, comms,
schools, etc)
• Mudslides rockfalls and soil
erosion
• Disruption due to
sinkholes/potholes/ blocked
drainage systems
• Isolated incidents of
communicable diseases
• Isolated cases of breakages of
informal/farm dam walls
• Damage to mud-based/make-
shift houses/structures
• Disruption of access to drinking
water and damage to crops
• Widespread flooding of roads and
settlements
• Danger to life (fast flowing streams
/ deep water)
• Large communities not
accessible/cut-off for a prolonged
period
• Widespread displacement of
affected communities
• Widespread damage to property,
buildings and loss of livelihoods
and livestock
• Widespread transport routes and
travel services severely affected
• Major roads and bridges damaged
or washed away
• Widespread, prolonged disruption
to essential services (water,
electricity, comms, schools, etc)
• Widespread mudslides, rockfalls
and soil erosion
• Long term disruption due to
sinkholes/potholes/ blocked
drainage systems
• Widespread incidents of
communicable diseases
• Breakage of dam walls
• Widespread damage to mud-
based/make-shift
houses/structures
• Widespread disruption of access to
drinking water and damage to
crops
Rainfall impact table
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 38
Warning chart for tomorrow
Warning areaL
ikel
ihoo
dV
ery
low
to H
igh
X
Minimal to SevereImpact
• Localised flooding of
susceptible
formal/informal
settlements or roads,
low-lying areas and
bridges
• Major roads affected but
can be used, increased
travel times
• Minor motor vehicle
accidents due to slippery
roads and/ reduced
visibility
• Localised mudslides,
rockfalls and soil erosion
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 39
Warning chart for tomorrow
Warning areaL
ikel
ihoo
dV
ery
low
to H
igh
X X
Minimal to SevereImpact
• Flooding of roads and
settlements (formal and
informal)
• Danger to life (fast
flowing streams / deep
water)
• Displacement of affected
communities
• Damage to property,
infrastructure, loss of
livelihood and livestock
• Major disruption of
traffic flow due to major
roads being flooded or
closed
• Mudslides rockfalls and
soil erosion
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 40
Feedback and Evaluation
• Submission of situation reports (Disaster
Management)
• Media (on the ground evidence, aware of
hoaxes)
• General public (surveys, social media)
Evaluation of
warning
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 41
1. NOT weather threshold based (50mm in 24hrs), but impact level based (Paradigm shift for forecasters)
2. Combining expected • impact level in an area, and • likelihood of impact occurring• = Warning Risk Level (G,Y,O,R)
3. Co-ordination with disaster management is important for orangeand red warnings
Main points to remember
Objective decision support system• To support to forecasters and disaster
managers in warning decision making
• Based on • Objective hazard prediction systems, • GIS-based spatial vulnerability
information at regional and local levels
• Vulnerability data from disaster management particularly at local levels, weather information from SAWS
Total Vulnerability to flash floods based
on Mesozones
Integrated “end-to-end” EWS
Large Scale weather patterns
National, regional, local scale weather impacts
Community, “house” scale social impacts
SAWS monitoring weather &
initiate alerts
PDMCs, MDMCs impact
assessment
Coordination
Warnings issued:Yellow, orange, red
DM structures Media
Rural villages, informal settlements,
Communities
Cities, Towns
Impact-B
ased S
WW
SC
B E
WS
Dissemination platforms
Some Lessons learnt so far
• Start simple – focus on a few hazards with social impacts to disasters only (other sectors can follow later)
• Development and implementation through a phased approach
• Effective coordination between forecasters and disaster managers during a severe weather event is crucial and must be developed at the initial workshops and tuned in pilot test phases
• Importance of involving both forecaster and disaster manager to create dual ownership of warnings
• Need for preparation of effective, actionable, communication to general public in layman’sterminology when going operational
• Include the media once readyDoc Ref no: FCAST-PRE-2019 IBF Intro Honiara
2019
2019/10/04 Templ ref: PPT-ISO-colour.001 Doc Ref no: 46
Conclusion
• Good progress in the project to implement an Impact-based
SWWS in South Africa
� Partnership established between SAWS and NDMC
� Forecasting process developed and tested
� Buy-in of the forecasters
� Pilot phases completed across the country
� Involves disaster management stakeholders, SAWS, DMCs
(national, provincial, local)
� Very positive feedback from disaster managers
• Based on international developments
• It will change the EWS to into a user oriented
risk-based EWS, adapted for South African conditions