Deck 2 - Lectures 6-10 - Jamie Flashcards
Flood frequency analysis issues
definition and chch example
Assumes that th eobservation record is from homogenous conditions. Means that each flood needs to occur under the same type of conditions.
basin alterations e.g. urbanisation alters the behaviours of flood events and can dramatically reduce the number of years of homongenous data.
Massive changes to the urban environments therefore can only use most recent years of observations to predict flooding. Many places in the world are increasing population and area. Global rural population is going down so urbanisation catchment is a huge issue globally.
Issue globally: The observation record for the entire record should be from homogenous conditions → christchurch city in the 90s was a lot smaller - ground conditions as a result have changed massively in the city, if that was the source of where the catchment was (flooding), then you couldn’t use a record that went back to the 90s for 2021 to determine the flood risk because the ground surface conditions are completely different, retention ponds, housing and roads. Would only be able to use it for the length of time that the modern conditions actually prevail.
Consequences of climate change on events
changing averages, affects extremes of distributions, increase frequency and magnitude (return periods decreasing), therefore increase probability and consequences.
changes the risk and may increase vulnerability (risk management strategy no longer sufficient). data is no longer collected unfer homogenous conditions.
Integrated risk analysis under climate change
changes to weather systems in NZ
- Normal way to have rainfall in winter - a southerly storm off the ocean brings in rain.
- Main events - coming in off the north tasman - summer weather pattern, tropical moisture being sucked down off the queensland coast or south west pacific convergence (off fiji), being fed into the NZ area.
- Isn’t as much moisture in the southern ocean - capacity of the air to hold moisture is directly related to temperature.
- Southerly storm - air mass hitting the country with an average temperature of 5-6 degrees in mid winter has way less moisture available to it than a storm system coming through with a south westerly component however moisture from the sub tropics has been sucked down into the New Zealand region and incorporated into our weather systems - air mass contains significantly more moisture than would normally be available
- Since European settlement and the start of scientific measurements, this type of weather system is almost unheard of in winter - occurs in late summer when ex tropical cyclones are coming out of the north tasma or south west pacific convergence. As a result it’s hard for models to predict this because the historical data doesn’t contain these types of events. So caution is the most important thing you can put into scientific modelling. Need conservative in your risks - put some additional uncertainty to allow for the fact that the model may no longer be accurate because the climate system is not behaving the way that it used to.
Aims and challenges of flood mapping
- to determine the impact area of a flood event - during event for emergency management and post event for damage or impact assesment (e.g. for insurance or rebuilding).
- cause use as a source of data to assess the reliability of flood models
- to generate flood hazard zone maps.
challenges: flood events are usually associated with bad weather - optical imagery.
Using social media to map floods
definition and challenges
“crowdsourcing” of data - publically contributed photos of flooding.
Issues → coverage is often patchy. Depending on where people are and who is photographing what etc. may not get a complete distribution over where its happening.
questionable reliability
Remote sensing of flood inundation
acquisition of data at some distance from the object of interest. the measurement of reflected radioactive energy from the earths surface using spatially seperated opto-mechanical devices located on ground-based, airborn or space-borne platforms. Using drones, planes and helicopters.
advantages - floods are hazardous events, mapping duirng an event needs to be done remotely.
Mapping using drones
issues
- not always good for mapping
- extent of coverage is small - only suitable for small events
Mapping using aerial imagery
definition and issues
digital air photos acquired then georectified onto a map grid, mosaicked together. flood extend then delineated in GIS. Traditional method - aerial photographs. Critical to orthorectify them to get as little distortion as possible.
Hard to get imagery at that quality because the storms are usually occurring when you’re trying to fly the path and you can’t see through the clouds. Putting up an aircraft is an expensive operation and there might not be the ability to put up the aircraft at that time.
Mapping using satellite imagery
definition and issues
optical imagery e.g. landsat tend to be less useful for flood mapping.
clouds tend to obscure floods.
More typical problem - can’t see through the clouds. If you are depending on passive systems then your view will be obstructed by the clouds and that will limit it. While giving a good measurement of the broad view, satellites cannot give high resolution measurements on the ground.
Mapping using SAR
definition and advantages
- Synthetic aperture radar
- active remote sensing: sends out its own electromagnetic signal rather than relying on the sun
- Measurements of reflected signals. Generates its own signal - cant get high resolution at night if using the sun.
- cloud penetrating microwave radiation used
- all weather - independent of the sun
- high spatial resolution - depends on what height but from a plane or a drone you get high quality resolution.
- can capture short lived events associated with cloud cover.
SAR how does it work?
Microwave system that pulses out of the satellite platform and measles what comes back. Brightness depends on how rough the surface is. Rough areas reflect more signal than smooth areas. If you have water, water tends to be much smoother than a dry land surface and you will get textured contrast between areas that are not flooded.
flooded areas tend to be smooth and the reflection occurs away from the sensor so consequence is that flooded areas appear to be dark.
Satellite versus airborne SAR
Airborne: multi-frequency, fully polarimetric, ~1m resolution, limited by aircraft range, coverage controllable.
Multi different wavelengths are ideal because then you get multiple information back, vegetation types etc.
Airborne is fully polarimetric - additional information.
Important because it makes it easier to interpret.
satellite - single wavelength, 1 or 2 polarisations, ~25m resolution, remote regions, limited by orbit.
Principles of flood modelling
A flood model is usually developed initially for a past flood event for which observation data are availible.
What elements should be in a model produced to predict a flood extent? Soil types, infiltration into soil for rainfall - the amount and speed at which the water reaches the channels will be completely different if the soil is dry or wet - clay soil will be completely different to gravel.
model predictions compared to observations for verification. model parameters (usually friction) modified using calibration, such that model error is minimised.
input the rain into the model and flows how it should in real life. that is controlled by the slope of the land and the friction of the surface. If you change the friction coefficient in the model the flow will either slow down or speed up the rate at which the water is moving through an area. That will change the timing of the flooding, the elevation of the maximum flood and the timing of the floods going down.
Model validation then should be completed. simulation of a second independent past event for which observation data are availible.
We need to validate the model, the best way to do that is to go to somewhere else for where we also have data and then run the model again, a second event through the model for which observation is also available, if the model is well calibrated the model should predict what is seen in the second event. If not we need to recalibrate so that it still matches the modern data. When we run the model on the second event it gives a better outcome on the second event. Do this until you have tweaked the model that can predict accurately what happens in a second dependent event. Measure the accuracy of the model. The problem is that the observational data is so rare that proper validation of models is often not possible - the second validation step is not completed fully so much larger errors around the models than you would like. More true in NZ than overseas - they have more data to run the models with.
once calibrated and validated the model can be used reliably for the prediction of unseen events.
outputs of the flood inundation model
- Inundation extent: what area will the flood go over?
- Inundation depth - how deep the water is in different places
- Flow rates derived from the slopes and inferred friction coefficient
- Flood timings - compound of the flow rates and landscapes - different landscapes will accelerate or decelerate the flood and by doing that you change the flood peak.
How would the land surface change the flood peak?
Concrete in the river channel versus vegetation in the river channel - when the flood water hits the concrete channel it goes through relatively quickly. Vegetation in the channel slows down the flood water speed - the elevation of the water has a piling up effect because there is an obstruction or greater friction in the channel. Storm water systems in urban systems are designed to optimise the speed at which you get the water away from the urban area but causes floods where the storm water system finishes. If you get your water down too quickly, you get flooding at the lower part of the catchment. So you build retention areas like fake lakes and swamps designed to hold water temporarily when the heavy rain comes in and feed it in more slowly. Intended to stop flooding lower down in the system.
LISFLOOD-FP overview
- raster based model
- requires a DEM, friction derived from landcover
- water added to the channel - rainfall
- raster model discretisation of floodplain and channel topography
- once bankfull depth is exceeded water can flow laterally over adjacent low lying floodplains according to topography and free surface gradient.
LISFLOOD-FP hydraulic model - mass continuity
change in flood depth (h) in a cell in one time step is equal to the sum of flows (q) into and out of the cell into each of the four directions.
Simultaneous measurements as a loss or gain of water between squares.
LISFLOOD-FP hydraulic model - flow rate
formulation based on the manning equation.
A starting speed of the water and the water will be accelerating because of slope change, greater height in the box there is a tendency for the surface elevation difference to change which is what drives the flow - the slope drives the flow. The model is doing instantaneous measurements of the elevation of the water of this box and adding the water in which means there is more force to flow the water downstream down the floodplain. Or there’s less water coming in so the surface elevation has gone down and the flows will decrease or reverse depending on what is happening.
The role of IFRM in the flooding chain
flooding chain definition and IFRM role
The flooding chain: sources (rainfall, wind, waves) what is causing the flooding –> pathways (overtopping, inundation of flood plain) what happens when the flood is underway –> receptors (people, poverty, environment) what is the impact on people, poverty and the environment?
Ifrm - works on risk reduction from the source to the pathway to the receptor.
Reduce flood hazard - restore the wetlands, build green infrastructure. Should restore the wetlands. If market gardens flood, doesn’t matter, doesn’t matter for the wetlands (can absorb the water), if you turn them into housing it puts in a risk situation.
Physical protection - build embankments and flood barriers.
Regulate land use - setback lines, building restrictions and proofing in place.
Raise awareness and preparedness - people need to know whether they are in a hazard zone, liable for flooding or not? Need warning that a flood is coming through, need proper plans for how to deal - evacuation plan.
Need to mitigate residual risk - mitigated but cannot get rid of it. Emergency response, insurance, relief funds and recovery plans.
Flood adaptation in LA
what is going on here?
How much natural water is in the channel? The climate is 20 yrs a day of rain. The normal amount of drainage is pretty dry.
Why do you need such a big channel? It all happens at once - heavy rainfall in the hills around LA, the soils around LA are not normally saturated with water, in NZ the landscapes are designed so that if a flood comes down and the soils become saturated that’s when we get floods occurring.
Causes removal of water from systems: Transpiration - causes the removal of water from systems. For an area in NSW consequence of european settlement - the lakes filled up - not what supposed to find. Trees cut down, The transpiration was reduced because you’ve switched from trees to grass, every time it rained, moisture into the soil instead of being evaporated off it had no where to go. Lakes and swamps build up. Pre european there was a lot less surface water even though the landscape was greener in some sense it was drier.
In this LA system - flashy rainfall, mostly arid - sparse vegetation cover, soils are not thick and slope is steep, all work against storing water in the system. The channels go from dry to raging torrents.
Not aesthetically pleasing, cheap or ecologically sound. Huge engineering work to put these in. nothing natural or organic - not healthy ecosystems. Large push to move away from highly engineered systems to work with nature rather than against it.
Principles of insurance
3 principles
If you own something or want to do something you want to be assured that you can do so safely. Everything has some sort of risk associated with it.
- mishaps and wear - most likely thing to occur is not very much. motor insurance e.g. scratch on a car door
- significant impact - a small but not insignificant chance of something worse. motor insuranve e.g. minor crash
- severe - very small chance of something very bad occuring - sever car crash.
Insurance - how it works - pooling risks
definition and diagram
- The insurance company takes the policies from all the people involved and pools them together. Gives them a larger sum of money to cover all of the concerned.
- Need to generate an amount of premiums/payments where the expected annual losses should be less than the sum of the premiums.
- You also need to take into account the variability around the actual case. Motor accidents are common and a large pool of people who make payments towards those versus a hazard event which occurs once every so many years.
non structural measures - temporary/localised levees
Temporary/ localised levees
– Structures put in place just
before flooding
– Cost effective, but requires
good flood warning
Less intrusive - removed afterwards.
