Climate indices#

The climate component offers an overview of climate indices related to hydro-meteorological hazards based on the most updated information (CMIP6). The challenge is to offer a tool that convey the complexity of climate models into statistics that are easily interpretable by non-climate experts, providign a well-rounded perspective for both space and time dimension. This is reflected on the technical side, by the challenge to reduce huge datasets into manageable pieces.

The climate component provides aggregated statistics at boundary level (country or subnational level) for a selection of:

  1. climate-related hazards

  2. country

  3. time periods

The table summarises the relevant climate indices and related time scale.

Name

Description

Time-scale

R10mm

Days with rainfall > 10 mm [days]

Annual

Rx5day

Maximum 5-day precipitation [mm]

Monthly

R99p

Extremely wet day precipitation [days]

Monthly

CWD

Consecutive Wet Days [days/month]

Annual

CDD

Consecutive Dry Days [days/month]

Annual

slr

Sea Level Rise [m]

Annual

SPEI

Standard Precipitation-EvapoT Index [-]

Annual

Heat

WBGT or UTCI [°C] - bias adjusted

Daily

tmean

Mean surface temperature [°C]

Monthly

Compared to the information directly offered by the CCKP country page, this procedure adds:

  • Linking relevant climate indices to natural hazard occurrance in order to estimate changes in risk drivers over the future

  • Standardisation of anomaly over historical variability as common metric of change

Input data#

  • raster data aggregated across time (20 years windows) for spatial representation

  • csv data aggregated across space (country boundaries) for time-series representation

    • includes ensemble p10 and p90 (variability across models)

Sources of CMIP6 data#

Name

Developer

Description

Data format

Climate Knowledge portal

World Bank

Large selection of climate indices for both trends and extremes

Table, geodata, charts

Climate Extreme Indices (CDS)

UNDRR

Full selection of variables, model members, periods

Geodata

IPCC atlas

IPCC

Selection of climate variables for a range of periods and scenario

Table, geodata, maps, charts

Dimensions:#

  • SSP: SSP1/RCP2.6; SSP2/RCP4.5; SSP3/RCP7.0; SSP5/RCP8

  • Ensemble member: r1i1p1f1 (largest number of models available)

  • Ensemble range: p10, p50, p90

  • Period: {Historical (1981-2015)}, [Near term (2020-2039), Medium term (2040-2059), Long term (2060-2079), End of century (2080-2099)]

  • Time scale: Annual (R10mm, CWD, slr, SPEI); Monthly (Rxday, R99p, tmean); Daily (Heat)

  • Value statistic: {P10, P50, P90, SD}, [P10, P50, P90]

Script setup [WIP]#

  • Navigate to your working directory: cd <Your work directory>

  • Run jupyter notebook

  • Select CCDR-climate.ipynb and chose the indices to plot.

docs/images/ccdr-clim.png

Fig. 33 Starting page for the climate outlook notebook.#

Processing#

  • Runs over one selected country and for a specific set of indices depending on selected hazard

  • Consider four SSP (ex RCP) scenarios (SSP1/RCP2.6; SSP2/RCP4.5; SSP3/RCP7.0; SSP5/RCP8.5)

  • Consider four 20-years periods (near term, medium term, long term, end of century)

  • Calculate median, 10th percentile (p10) and 90th percentile (p90) of standardised anomaly across models in the ensemble (more details)

  • Plot maps and timeseries

  • Exported results as csv (table) and geopackage (vector)

Output presentation#

A) Map output (spatial distribution)

  • Raster data aggregated across time (20 years windows)

Ensemble_mean(Period_mean(anomaly/hist_SD))
Ensemble_p50(Period_mean(anomaly/hist_SD))
../_images/ci_raw.png

Fig. 34 Example of mean standardardised anomaly (ensemble median) plotted for one climate index over Pakistan, period 2040-2060, 3 SSP scenarios - grid data.#

  • ADM1-mean values from raster data

ADM1_mean(Ensemble_p50(Period_mean(anomaly/hist_SD)))
../_images/ci_adm.png

Fig. 35 Example of mean standardardised anomaly (ensemble median) plotted for one climate index over Pakistan, period 2040-2060, 3 SSP scenarios - mean for subnational unit.#

B) Chart output (time-series)

  • Spatial data aggregated for country ADM0 boundaries plotted as chart

Ensemble_p10(ADM0_mean(anomaly/hist_SD))
Ensemble_p50(ADM0_mean(anomaly/hist_SD))
Ensemble_p90(ADM0_mean(anomaly/hist_SD))
../_images/ci_tseries.png

Fig. 36 Example of mean standardardised anomaly (ensemble median) plotted for one climate index over Pakistan, time-series up to 2100, 3 SSP scenarios - mean at country level.#