Risk assessment

Risk assessment#

In this notebook, the calculation of risk associated with the variable under examination will be shown. Pandas dataframes will be presented which, in the example, show nine airports - the operator will need to modify the list of airports by adding/modifying names. The same concept applies to exposure and vulnerability indicators - indicators will be presented that can be expanded or modified by the user. The only mandatory thing is that the framework was conceptualized for at least two airports

Preparation work#

Import libraries#

import os

import matplotlib.pyplot as plt
import numpy as np
import pandas as pd
import scipy.spatial
import xarray as xr
import contextily

import plotly.express as px

Area of interest#

Specify the same name as in the the hazard assessment to continue with the results of these notebooks.

region_name = 'IT'

Path configuration#

# Path to the folder containing the NetCDF files
indicators_path = f'../data_{region_name}/indicators/uerra'

# Folder to save the csv file where average exposure and vulnerability are going to be stored
csv_folder_path = f'../data_{region_name}/indicators/vulnerability_exposure'
os.makedirs(csv_folder_path, exist_ok=True)

Define the airports#

List the airports and their coordinates.

airports = [
    'Milano Malpensa', 'Milano Linate', 'Bergamo Orio al Serio',
    'Roma Fiumicino', 'Roma Ciampino', 'Napoli Capodichino',
    'Palermo Punta Raisi', 'Catania Fontanarossa', 'Cagliari Elmas'
]

Tip

Enter the locations of your choice here. At least two locations must be defined for the methodology to work.

def to_dataframe(dct):
    idx = pd.Index(airports, name="airport")
    return pd.DataFrame(dct, index=idx, dtype=float)

# List coordinates in the order of the locations from above (°N, °E)
coords = to_dataframe({
    "latitude": [45.63, 45.45, 45.67, 41.80, 41.80, 40.88, 38.18, 37.47, 39.25],
    "longitude": [8.73, 9.28, 9.71, 12.25, 12.59, 14.29, 13.10, 15.07, 9.06]
})

# Check that the coordinates are correctly associated with the airports
coords
latitude longitude
airport
Milano Malpensa 45.63 8.73
Milano Linate 45.45 9.28
Bergamo Orio al Serio 45.67 9.71
Roma Fiumicino 41.80 12.25
Roma Ciampino 41.80 12.59
Napoli Capodichino 40.88 14.29
Palermo Punta Raisi 38.18 13.10
Catania Fontanarossa 37.47 15.07
Cagliari Elmas 39.25 9.06 
fig, ax = plt.subplots(1, 1)
ax.scatter(coords.longitude, coords.latitude, color="k")
ax.set_xlabel("Longitude [°E]")
ax.set_ylabel("Latitude [°N]")
contextily.add_basemap(ax, crs="EPSG:4326", attribution=False)
../../../../_images/5c3f7237721f2b41636f29d1345add5a0bbf51fe0f43346f4b18608d813580be.png

Component normalization#

The aim of normalization processes is to transform the values of these indicators, which are measured at different scales and in different units, into comparable values considered on a common scale (Master Adapt, 2018). We used the Min-Max method to transform all values to scores in a range from 0 to 1, where the value 0 represents the optimal level, while the value 1 reflects the most critical estimates:

def normalize_columns(df):
    return (df - df.min()) / (df.max() - df.min())

Hazard: temperature and precipitation#

Load the hazard indicator data produced by the hazard assessment notebook and extract the information for the airport locations specified above.

def load_all(files):
    """Load data based on a name-to-file mapping"""
    das = []
    for name, file in files.items():
        path = os.path.join(indicators_path, file)
        if name == "*":
            ds = xr.open_dataset(path)
            das.extend(ds.data_vars.values())
        else:
            da = xr.open_dataarray(path)
            das.append(da.rename(name))
    return xr.merge(das, compat="override")


def extract_points(ds, points):
    """Extract data from the dataset for a set of points"""
    grid = np.column_stack([
        ds["latitude"].values.flatten(),
        ds["longitude"].values.flatten()
    ])
    tree = scipy.spatial.cKDTree(grid)
    _, closest_idx = tree.query(coords.values, k=1)
    idx_y, idx_x = np.unravel_index(closest_idx, ds["latitude"].shape)
    return (
        ds
        .isel(
            x=xr.DataArray(idx_x, dims="airport"),
            y=xr.DataArray(idx_y, dims="airport")
        )
        .drop_vars(["latitude", "longitude"])
        .assign_coords({"airport": coords.index})
        .to_dataframe()
    )

Temperature#

data_temp = load_all({
    "Days Above 35°C": "Temp_DaysAbove35.nc",
    "Days Above 40°C": "Temp_DaysAbove40.nc",
    "Days Above 45°C": "Temp_DaysAbove45.nc",
})

indicators_temp = extract_points(data_temp, coords)
indicators_temp
Days Above 35°C Days Above 40°C Days Above 45°C
airport
Milano Malpensa 0.000000 0.000000 0.0
Milano Linate 0.533333 0.000000 0.0
Bergamo Orio al Serio 0.433333 0.000000 0.0
Roma Fiumicino 0.166667 0.000000 0.0
Roma Ciampino 3.400000 0.000000 0.0
Napoli Capodichino 3.433333 0.000000 0.0
Palermo Punta Raisi 1.266667 0.100000 0.0
Catania Fontanarossa 8.933333 1.133333 0.2
Cagliari Elmas 6.700000 0.266667 0.0

Combine the indicator columns into a single indicator for the temperature hazard by taking the mean over all columns at each location.

hazard_temp = normalize_columns(indicators_temp).mean(axis=1, skipna=True)

# Quick bar chart of the temperature hazard indicator
hazard_temp.plot.barh(title="Hazard: Temperature", xlim=(0, 1))

<Axes: title={'center': 'Hazard: Temperature'}, ylabel='airport'>
../../../../_images/538fd5239dbd7c2920b980af715855e0bb9f9b29ac9700f815a456f563776930.png

Precipitation#

Repeat the steps for the precipitation indicators:

data_precip = load_all({
    "*": "Precip_return_levels_gumbel.nc",
    "Precip_P99": "Precip_P99.nc",
    "Precip_P999": "Precip_P999.nc"
})
# Remove some left-over coordinates so they don't show up in the indicators table
data_precip = data_precip.drop_vars(["x", "y", "time", "quantile"], errors="ignore")

indicators_precip = extract_points(data_precip, coords)
indicators_precip
return_period_2_y return_period_5_y return_period_10_y return_period_20_y return_period_30_y return_period_50_y return_period_100_y return_period_150_y return_period_200_y return_period_500_y Precip_P99 Precip_P999
airport
Milano Malpensa 77.240410 97.857880 111.508438 124.602386 132.135010 141.551163 154.251877 161.657974 166.906250 183.601303 48.059845 94.042389
Milano Linate 56.678417 70.069450 78.935478 87.439987 92.332420 98.448204 106.697311 111.507561 114.916313 125.759735 34.706484 63.180023
Bergamo Orio al Serio 49.203640 59.805862 66.825455 73.558823 77.432358 82.274467 88.805618 92.614090 95.312943 103.898109 33.550508 51.886977
Roma Fiumicino 50.875408 69.739494 82.229156 94.209549 101.101570 109.716927 121.337532 128.113785 132.915726 148.190964 27.939688 60.487923
Roma Ciampino 43.058388 56.674309 65.689232 74.336563 79.311157 85.529648 93.917290 98.808327 102.274323 113.299850 26.153749 52.537727
Napoli Capodichino 60.708721 85.161652 101.351616 116.881424 125.815338 136.983170 152.046585 160.830429 167.055054 186.855865 35.195312 84.431564
Palermo Punta Raisi 56.251366 81.191704 97.704376 113.543732 122.655724 134.046158 149.409836 158.368759 164.717453 184.912949 35.360001 88.106552
Catania Fontanarossa 58.424789 85.944069 104.164230 121.641449 131.695663 144.263901 161.216263 171.101578 178.106750 200.390549 26.499609 81.429756
Cagliari Elmas 32.551186 44.475163 52.369877 59.942673 64.299118 69.744881 77.090256 81.373528 84.408836 94.064301 17.672501 37.748158 
hazard_precip = normalize_columns(indicators_precip).mean(axis=1, skipna=True)

# Quick bar chart of the precipitation hazard indicator
hazard_precip.plot.barh(title="Hazard: Precipitation", xlim=(0, 1))

<Axes: title={'center': 'Hazard: Precipitation'}, ylabel='airport'>
../../../../_images/2004fddd88f1b1f07e570d74c553669c68c9825d222920a02d725a337894fcbd.png

Exposure: air and land side components#

The various airports components, divided into landside and airside areas, were considered as exposed samples in accordance with De Vivo et al. (2022).

The air side components include the structures used for the movement of aircraft, such as runways, taxiways, tower, and aprons:

Air side

Unit of measurement

Runways

Area [m²]

Taxiways

Area [m²]

Tower

Height [m]

Apron

Area [m²]

The land side components refer to the public access areas such as offices, terminals, airports access systems and parking areas:

Land side

Unit of measurement

Terminal

Area [m²]

Offices and other buildings

Area [m²]

Airport accesses systems

Area [m²]

Carparks

Number

Define the values for each indicator and airport: (insert None where data is not available)

# Specify the values in the order of the locations defined above
air_land_side_components = to_dataframe({
    'Runways': [470400, 159742, 145530, 396000, 9900, 118260, 292650, 109575, 126000],
    'Taxiways': [250.5, 156, 183, 300, 301, 99, 368, 159, 604],
    'Tower': [80, 47, 30, 56, 60, 40, 25, 12, 28],
    'Apron': [1319000, 387000, 190000, 797250, None, 200000, 148000, 180000, 156000],
    'Terminals': [315000, 85050, 53025, 354300, 20950, 30700, 35400, 43310, 41290],
    'Offices_and_other_buildings': [26715, 13527, 7075, 29520, 5950, 1915, 4180, 6920, 5030],
    'Airport_accesses_systems': [31660, None, None, None, None, None, None, None, None],
    'Carparks': [15000, 3000, 8000, 20100, 1220, 1500, 1364, 1800, 2133]
})

# Check that values are associated correctly to the locations
air_land_side_components
Runways Taxiways Tower Apron Terminals Offices_and_other_buildings Airport_accesses_systems Carparks
airport
Milano Malpensa 470400.0 250.5 80.0 1319000.0 315000.0 26715.0 31660.0 15000.0
Milano Linate 159742.0 156.0 47.0 387000.0 85050.0 13527.0 NaN 3000.0
Bergamo Orio al Serio 145530.0 183.0 30.0 190000.0 53025.0 7075.0 NaN 8000.0
Roma Fiumicino 396000.0 300.0 56.0 797250.0 354300.0 29520.0 NaN 20100.0
Roma Ciampino 9900.0 301.0 60.0 NaN 20950.0 5950.0 NaN 1220.0
Napoli Capodichino 118260.0 99.0 40.0 200000.0 30700.0 1915.0 NaN 1500.0
Palermo Punta Raisi 292650.0 368.0 25.0 148000.0 35400.0 4180.0 NaN 1364.0
Catania Fontanarossa 109575.0 159.0 12.0 180000.0 43310.0 6920.0 NaN 1800.0
Cagliari Elmas 126000.0 604.0 28.0 156000.0 41290.0 5030.0 NaN 2133.0

As for the hazard, normalize all components and combine into a single indicator for the exposure by taking the mean over all components for each location.

air_land_side_normalized = normalize_columns(air_land_side_components)

# Combine into a single indicator
exposure = air_land_side_normalized.mean(axis=1, skipna=True)

# Quick bar plot of the exposure indicator
ax = exposure.plot.barh(title="Exposure", xlim=(0, 1))
../../../../_images/bc7743a29ced07a9eae4d48d06c330cfbc96b8adfe011eeac296f32ed8f211d2.png

Vulnerability#

Vulnerability reflects “the propensity or predisposition of a system to be adversely affected. It encompasses a variety of concepts and elements including sensitivity or susceptibility to harm and lack of capacity to cope and adapt opportunities, or to respond to consequences” (IPCC, 2022).

Vulnerability is composed of sensitivity and adaptive capacity.

Sensitivity#

Sensitivity is intended as the degree to which the system is (positively or negatively) affected by climate variability or climate change.

Indicator

Unit of measurement

Soil Sealing

Area [%]

Passenger

Number

Building in bad condition

Number or Absence/Presence

Age building

Age

Air traffic

Number

Parking accesses

Number

Staff work outside airport

Number

Define the sensitivity indicators and their values: (use None where data is not available)

sensitivity_indicators = to_dataframe({
    'Soil_sealing': [1235, 300, 300, 1590, 133, 217, 391, 225, 246],
    'Passengers': [27000000, 7000000, 13857257, 43532573, 5879496, 10860068, 7018087, 10223113, 4747806],
    'Buildings_bad_conditions_or_maintenance': [None, None, None, None, None, None, None, None, None],
    'Age_buildings': [73, 82, 83, 63, 105, 111, 61, 97, 84],
    'Air_traffic': [201050, 85730, 95377, 309783, 52253, 82577, 54243, 73494, 39691],
    'Parking_accesses': [None, None, None, None, None, None, None, None, None],
    'Staff_work': [1997.1, 855.9, 1044, 2363.9, 1013.1, None, 283, 183, 146]
})

sensitivity_normalized = normalize_columns(sensitivity_indicators)
sensitivity_normalized
Soil_sealing Passengers Buildings_bad_conditions_or_maintenance Age_buildings Air_traffic Parking_accesses Staff_work
airport
Milano Malpensa 0.756349 0.573735 NaN 0.24 0.597422 NaN 0.834618
Milano Linate 0.114619 0.058069 NaN 0.42 0.170457 NaN 0.320078
Bergamo Orio al Serio 0.114619 0.234872 NaN 0.44 0.206174 NaN 0.404888
Roma Fiumicino 1.000000 1.000000 NaN 0.04 1.000000 NaN 1.000000
Roma Ciampino 0.000000 0.029179 NaN 0.88 0.046510 NaN 0.390955
Napoli Capodichino 0.057653 0.157594 NaN 1.00 0.158783 NaN NaN
Palermo Punta Raisi 0.177076 0.058535 NaN 0.00 0.053878 NaN 0.061770
Catania Fontanarossa 0.063143 0.141172 NaN 0.72 0.125154 NaN 0.016682
Cagliari Elmas 0.077557 0.000000 NaN 0.46 0.000000 NaN 0.000000

Adaptative capacity#

These indicators are defined by the classes reported below and included in De Vivo et al. (2022). If the indicator is present or absent, its value will be 0.9 or 0.1, respectively.

Risk awareness: Initiatives for mitigation to climate change#

  1. Adherence to “NetZero2050” Programme (neutrality 3+)

  2. Certification ISO 50001 and initiatives of “energy saving”

  3. Neutrality 3+ Airport carbon accreditation

  4. Neutrality 4+ Airport carbon accreditation; EP-100 intelligent use of energy from “The Climate Group”- Certification ISO 50001

  5. Level 2 Airport carbon accreditation, environmental improvement program

  6. Commitment to obtaining certfiication

Initiatives for adaptation to climate change#

  • Efficient drainage system

    1. Present

  • Insurance policy for extreme events

    1. Absent

  • Monitoring and alarm system

    1. Present

    2. Present (but alert wind shear)

    3. Absent

  • Bioinfiltration and permeable pavements

    1. Present

    2. Absent

  • Risk awareness (and initiatives of mitigation)

    1. Risk management framework

    2. Development of local (Urban) climate change mitigation and adaptation strategies

    3. Sustainability Report 2019

    4. gesap environmental report

    5. Absent

Guidelines for adaptation plan to climate change#

  1. Plan (PACC) and regional strategy for adaptation to climate change

  2. Regional plan under development

  3. Absent

  4. Sustainable Energy and Climate Action Plan (Covenant of Mayors)

  5. Sustainable Energy and Climate Action Plan (Covenant of Mayors) municipality of Catania

  6. Regional climate change adaptation strategy

Class No.

Description

Value Range

Indicator Range (0-1)

1

Optimal

0 - 0.2

0.1

2

Quite Positive

0.2 - 0.4

0.3

3

Neutral

0.4 - 0.6

0.5

4

Quite Negative

0.6 - 0.8

0.7

5

Critical

0.8 - 1

0.9

Define the adaptive capacity indicators and their values: (use None where data is not available)

adaptive_indicators = to_dataframe({
    'Vegetation_zone': [None, None, None, None, None, None, None, None, None],
    'Initiatives_for_optimization_of_energy ': [0.3, 0.3, 0.3, 0.1, 0.1, 0.3, 0.5, 0.7, 0.7],
    'Green_wall': [0.1, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9],
    'Green_roof': [0.1, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9],
    'Heat_resistent_coating': [0.9, 0.9, 0.9, 0.1, 0.9, 0.9, 0.9, 0.9, 0.9],
    'Insurance_policy_for_extreme_events': [0.9, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9, 0.9],
    'Guidelines_for_adaptation_climate_change': [0.3, 0.3, 0.3, 0.5, 0.5, 0.9, 0.3, 0.3, 0.3],
    'Evening_departures': [None, None, None, None, None, None, None, None, None]
})

# As defined in the example, the indicators are already normalized into the range [0, 1]
adaptive_normalized = adaptive_indicators
adaptive_normalized
Vegetation_zone Initiatives_for_optimization_of_energy Green_wall Green_roof Heat_resistent_coating Insurance_policy_for_extreme_events Guidelines_for_adaptation_climate_change Evening_departures
airport
Milano Malpensa NaN 0.3 0.1 0.1 0.9 0.9 0.3 NaN
Milano Linate NaN 0.3 0.9 0.9 0.9 0.9 0.3 NaN
Bergamo Orio al Serio NaN 0.3 0.9 0.9 0.9 0.9 0.3 NaN
Roma Fiumicino NaN 0.1 0.9 0.9 0.1 0.9 0.5 NaN
Roma Ciampino NaN 0.1 0.9 0.9 0.9 0.9 0.5 NaN
Napoli Capodichino NaN 0.3 0.9 0.9 0.9 0.9 0.9 NaN
Palermo Punta Raisi NaN 0.5 0.9 0.9 0.9 0.9 0.3 NaN
Catania Fontanarossa NaN 0.7 0.9 0.9 0.9 0.9 0.3 NaN
Cagliari Elmas NaN 0.7 0.9 0.9 0.9 0.9 0.3 NaN

Vulnerability calculation#

Combine the sensitivity and adaptive capacity indicators by taking the mean at each location, then calculate a single vulnerability indicator by taking the mean of these two components.

vulnerability = 0.5 * (
    sensitivity_normalized.mean(skipna=True, axis=1)
    + adaptive_indicators.mean(skipna=True, axis=1)
)

ax = vulnerability.plot.barh(title="Vulnerability", xlim=(0, 1))
../../../../_images/c437f5376df00f510b23f288fa0470b335b13209fb9d1463d19ebc0f9151ca42.png

Calculate Risk#

The risk (R) will be computed by multiplication of the three components hazard (H), exposure (E) and vulnerability (V):

R = H * E * V

risk_table = pd.DataFrame.from_dict({
    "Temperature hazard": hazard_temp,
    "Precipitation hazard": hazard_precip,
    "Exposure": exposure,
    "Vulnerability": vulnerability,
    # Calculate the risk by multiplying the components
    "Temperature risk": hazard_temp * exposure * vulnerability,
    "Precipitation risk": hazard_precip * exposure * vulnerability,
})
risk_table.to_csv(os.path.join(indicators_path, f"risk_{region_name}.csv"))
risk_table
Temperature hazard Precipitation hazard Exposure Vulnerability Temperature risk Precipitation risk
airport
Milano Malpensa 0.000000 0.958177 0.830052 0.516879 0.000000 0.411093
Milano Linate 0.019900 0.418003 0.266326 0.458322 0.002429 0.051023
Bergamo Orio al Serio 0.016169 0.226971 0.200527 0.490055 0.001589 0.022304
Roma Fiumicino 0.006219 0.484378 0.776851 0.687333 0.003321 0.258637
Roma Ciampino 0.126866 0.221027 0.208675 0.484664 0.012831 0.022354
Napoli Capodichino 0.128109 0.820619 0.105080 0.571754 0.007697 0.049303
Palermo Punta Raisi 0.076675 0.786895 0.210126 0.401793 0.006473 0.066435
Catania Fontanarossa 1.000000 0.853814 0.091670 0.489948 0.044914 0.038348
Cagliari Elmas 0.328431 0.000000 0.245209 0.437089 0.035201 0.000000 
ax = risk_table[["Temperature risk", "Precipitation risk"]].plot.barh(title="Risk indicators")
../../../../_images/10cba38db65d8911acb7760197abb7b6e322060fab7685cb47583dc67577f97d.png 
fig, axs = plt.subplots(1, 2)

axs[0].scatter(coords.longitude, coords.latitude, s=risk_table["Temperature risk"]*1000, color="k")
axs[1].scatter(coords.longitude, coords.latitude, s=risk_table["Precipitation risk"]*1000, color="k")

for ax in axs:
    ax.set_xlabel("Longitude [°E]")
    ax.set_ylabel("Latitude [°N]")
    contextily.add_basemap(ax, crs="EPSG:4326", attribution=False)
../../../../_images/bba423a8ee02243f3d9ceac94bab2090dfc38f3093588b97dd5a65b74cbe907e.png

Classify risk#

To evaluate the risk classes for each of the airport the Quantile Method was used. Classifying each airport in terms of quantile with respect to all other airport taken into account

def classify_risk(risk, classes=["Very Low", "Low", "Medium", "High", "Very High"]):
    return pd.qcut(risk.rank(pct=True), q=len(classes), labels=classes)

risk_classes = pd.DataFrame.from_dict({
    c: classify_risk(risk_table[c]) for c in ["Temperature risk", "Precipitation risk"]
})
risk_classes
Temperature risk Precipitation risk
airport
Milano Malpensa Very Low Very High
Milano Linate Low High
Bergamo Orio al Serio Very Low Very Low
Roma Fiumicino Low Very High
Roma Ciampino High Low
Napoli Capodichino High Medium
Palermo Punta Raisi Medium High
Catania Fontanarossa Very High Low
Cagliari Elmas Very High Very Low 
def highlight_class_column(s):
    """Color the table entries based on the risk classes"""
    colors = {
        "Very Low": "background-color: forestgreen",
        "Low": "background-color: limegreen",
        "Medium": "background-color: yellow",
        "High": "background-color: orange",
        "Very High": "background-color: red; color: white"
    }
    return [colors.get(val, "") for val in s]

risk_classes.style.apply(highlight_class_column, axis=0)
  Temperature risk Precipitation risk
airport    
Milano Malpensa Very Low Very High
Milano Linate Low High
Bergamo Orio al Serio Very Low Very Low
Roma Fiumicino Low Very High
Roma Ciampino High Low
Napoli Capodichino High Medium
Palermo Punta Raisi Medium High
Catania Fontanarossa Very High Low
Cagliari Elmas Very High Very Low
Contents