Population And Poverty Data

DATA SOURCE: POPULATION

Population source data:
Historical Reference Period (1995-2014) is derived from:
Center for International Earth Science Information Network - CIESIN - Columbia University. 2018. Gridded Population of the World, Version 4 (GPWv4): Population Density, Revision 11. Palisades, New York: NASA Socioeconomic Data and Applications Center (SEDAC). 
https://sedac.ciesin.columbia.edu/data/collection/gpw -v4 

Projection data is derived from:
Jones, B., and B. C. O'Neill. 2020. Global One-Eighth Degree Population Base Year and Projection Grids Based on the Shared Socioeconomic Pathways, Revision 01. Palisades, New York: NASA Socioeconomic Data and Applications Center (SEDAC).

https://doi.org/10.7927/m30p-j498 . as an advancement from: Jones, B., and B. C. O'Neill. 2016. Spatially Explicit Global Population Scenarios Consistent with the Shared Socioeconomic Pathways. Environmental Research Letters, 11 (2016): 084003.

https://doi.org/10.1088/1748 -9326/11/8/084003 

CCKP Data Reference: Climate Change Knowledge Portal: Projected Population And Poverty,

0.25-Degree DOI:https://doi.org/10.57966/7r80-cc19 

Population data as gridded product or at subnational level are represented as either population count or population density (persons/km2) for “present” (~2010-2015) and as projected in the SSPs for selected time intervals. Note, the population density is calculated based on the grid size and grid contribution to a subnational polygon. These calculations might suffer from lack of precision for small spatial entities (particularly small islands).

Thresholds per grid-cell or aggregation area were set by population count or density (count: 1’000, 10’000, 100’000, 1’000’000; density: 1, 10, 100, 1’000).

Age Pyramid source data:Samir, K. C., & Lutz, W. (2014). The human core of the shared socioeconomic pathways: Population scenarios by age, sex and level of education for all countries to 2100. Global Environmental Change, 378.https://doi.org/10.1016/j.gloenvcha.2014.06.004issn:09593780 

Age- and sex- specific proportions are taken from the IIASA database of human capital reconstruction and projections. The method used for carrying out projections by age, sex, and educational attainment level is a generalization of the standard cohort-component method of population projections. This standard method is based on the fact that the age group a in year t will be a+x in year t+x (it is the same birth cohort, i.e., group of people born in the same year) after adjusting for the effects of mortality and migration and applying fertility rates to derive the number of births (the three components of population change).

DATA SOURCE: POVERTY

Poverty source data:World Bank Data Catalog: International Poverty Line - Global Subnational Poverty Atlas GSPA https://datacatalog.worldbank.org/search/dataset/0042041 

Poverty is represented as a percent of population below a given poverty classification: $1.90, $3.20, $5.50/day, as per World Bank definitions.

Natural Hazards

DATA SOURCE: EM-DAT

Source data:https://www.emdat.be/ 

Credits: EM-DAT: The OFDA/CRED International Disaster Database – www.emdat.be– Université Catholique de Louvain – Brussels – Belgium.

Description: EM-DAT contains essential core data on the occurrence and effects of over 18,000 mass disasters in the world from 1900 to present. The database is compiled from various sources, including UN agencies, non-governmental organizations, insurance companies, research institutes and press agencies.

Variables presented: Top Disasters; Number Killed; Number of Affected; Average Annual Disaster Occurrence by Type.

Tropical Cyclones

Cyclones are powerful, rotating storms that form over warm tropical and subtropical oceans and generally move from East to West before turning towards higher latitudes. These cyclones are known as Hurricanes in the Atlantic and Northeast Pacific basins, and as Typhoons in the Northwest Pacific. Cyclones pose a significant threat upon landfall, causing heavy rain, strong winds, flooding, and widespread damage, which can degrade water quality, spread disease, and destroy infrastructure. In addition to the immediate physical destruction, the aftermath often leads to long-term challenges, such as disruption of essential services, economic loss, and environmental degradation. To mitigate these impacts and support recovery, building broad resilience is essential. In most tropical places the occurrence of these storms in any one place is still rare. Therefore, the historical record, commonly going back several decades to a century, is generally too short to allow for a robust estimation of recurrence intervals of these storms. Such historical uncertainty can be reduced somewhat using models, where large ensembles of tropical cyclones can be generated. CCKP is offering these products as first examples that are designed to assist the stakeholder community in engaging with Tropical Cyclone data, forming a better understanding of their statistical properties today, and then working in a scenario context to consider possible changes in the future.

OBSERVED CYCLONES

DATA SOURCE: IBTRACS

Original Citations:
Knapp, K. R., M. C. Kruk, D. H. Levinson, H. J. Diamond, and C. J. Neumann, 2010: The International Best Track Archive for Climate Stewardship (IBTrACS): Unifying tropical cyclone best track data. Bulletin of the American Meteorological Society, 91, 363-376.
doi:10.1175/2009BAMS2755.1 

Gahtan, J., K. R. Knapp, C. J. Schreck, H. J. Diamond, J. P. Kossin, M. C. Kruk, 2024: International Best Track Archive for Climate Stewardship (IBTrACS) Project, Version 4r01. since1980. NOAA National Centers for Environmental Information. doi:10.25921/82ty -9e16[access date 24th of October 2024].

Original data access: NOAA's International Best Track Archive for Climate Stewardship (IBTrACS) data, https://www.ncei.noaa.gov/products/international -best-track-archive accessed on 24th of October 2024

Description: The International Best Track Archive for Climate Stewardship (IBTrACS) offers a global historical reference dataset for observed tropical cyclones around the world (Knapp et al. 2010). While the archive contains entries in some basins going back over 100 years, the period after 1980 – the satellite era – offers a more stationary observing system across all tropical ocean basins. The data selected here covers the period 1980-2023 (see Gahtan et al., 2024), drawing on wind speed data provided by the USA National Hurricane Center (1-minute sustained wind speed).

'USA_wind' combines data from the National Hurricane Center (NHC), the Join Typhoon Warning Center (JTWC), the Central Pacific Hurricane Center (CPHC), and includes data for the WMO Regional Specialised Meteorological Center at Miami and Honolulu (operated by NOAA).

Summary of data available at CCKP:
Currently, CCKP only displays IBTrACS data. Products are not available for download. Please use this link for data access: https://www.ncei.noaa.gov/products/international -best-trackarchive 

The historical record, commonly going back several decades to a century, is generally too short to allow for a robust estimation of proper recurrence intervals. Using probabilistic products based on large numbers of model simulated cyclones can help fill gaps in the record (see CHAZ data below). 

MODELLED CYCLONES

DATA SOURCE: CHAZ MODEL

Original Citation:

Lee, C.-Y., Tippett, M. K., Sobel, A. H., & Camargo, S. J. (2018). An Environmentally Forced

Tropical Cyclone Hazard Model. Journal of Advances in Modeling Earth Systems, 10(1), 223– 241. https://doi.org/10.1002/2017MS001186 

Lee, C.-Y., S.J. Camargo, A.H. Sobel, and M.K. Tippett (2020). Statistical-dynamical downscaling projections of tropical cyclone activity in a warming climate: Two diverging genesis scenarios. J. Climate, 33, 4815-4834. http://doi.org/10.1175/JCLI -D-19-0452.1 

Original data access:October 2023 (provided directly by Columbia University)

Description:

The Columbia HAZard Model (CHAZ, Lee et al., 2018) simulates potential cyclone distributions across ocean basins and their landfall by generating a comprehensive synthetic catalog of cyclone tracks based on outputs from 12 CMIP6 models, thus offering a broader perspective than observational data alone. Additionally, CHAZ was used to project future cyclone activity based on outputs from the same 12 CMIP6 models under the SSP2-4.5 scenario, centering a 30-year projection period around the year 2050. The CHAZ products used here are based on the genesis configuration using column relative humidity (CRH, see Lee et al. 2020). CCKP is offering these products as first examples that are designed to assist users in engaging with Tropical Cyclone data, forming a better understanding of their statistical properties today, and then working in a scenario context to consider possible changes in the future. CCKP plans to integrate additional models, scenarios, and time periods in the future.

Summary of data available at CCKP:

Presented at grid-level, and aggregated by ocean basin, exclusive economic zone, and by territory/country representing landfall. For ocean basins, a time series and daily climatology of storm occurrence within the seasonal cycle is available.

• Spatial resolution: 0.5 o x 0.5o(50km x 50km)
• Time period historical scenario: 1951-2014
• Future scenarios: SSP2-4.5
• Time period projected scenario: 2035-2064

Recommended Use: To evaluate the probability of cyclones at various wind speed categories and assess how these probabilities are projected to change in the future. Due to the limited temporal and thus spatial coverage (lots of gaps) of these observational data, we motivate the use of large, model-based ensembles to fill these historical gaps and to provide a more comprehensive view of actual risk distribution. Applying the identical approach to the evolving climate of future climate pathways, here SSP2-4.5, projections are offered of possible future activity. Currently the results are based on a single Tropical Storm and impacts simulation model, albeit driven by a broad array of 12 different GCMs. Caution should be exercised when interpreting future projections.

Cyclones classification:
We classify tropical cyclones using the US National Hurricane Center's Saffir-Simpson Hurricane Scale, based on maximum (sustained) wind speeds. We do not consider storms below 34 knots (often referred to as Tropical Depressions).

• Tropical Storm: 34 to <64 knots
• Category 1: 64 to <83 knots
• Category 2: 83 to <96 knots
• Category 3: 96 to <113 knots
• Category 4: 113 to <137 knots
• Category 5: >=137 knots

Cyclones are classified in 6 categories, as described above, with category names 'tcts' for tropical storms, 'tccat1' for cyclones of category 1, and so on 'tccat2', 'tccat3', 'tccat4', and 'tccat5'.

Caveats for future projections:
CHAZ (Lee et al., 2018) is one model out of a class of statistical-dynamical downscaling tools that are being developed and continuously refined in the scientific community. Ideally, one would use multiple models with a broad array of configurations to fully sample the space of tropical cyclone dynamics and its related statistics. As of now, coordinated collections of projections based on such models in a similar way as the CMIP6 experiments are not available yet. Additionally, there remain significant gaps in the fundamental understanding of various aspects of the science of Tropical Cyclone dynamics as well as global and basin-scale climate change boundary conditions that render any available projections of change fundamentally uncertain. While some derived tendencies might be more robust than others, for example there appears to be reasonably strong agreement that the strongest Tropical Cyclones might become more likely and intense, any projected overall frequency changes in the total occurrence are to be used with caution. Based on potential trends in global dynamics, some basins might see increases while others might experience a decrease in future number of storms. Small differences can potentially lead to a change in sign of outcomes. As with other projections, ensembles might be more robust than individual model outcomes.

It is worth noting that individual CHAZ simulations were based on the projected climate from 12 different global CMIP6 models. Large ensembles were run for both historical and one future period based on SSP2-4.5 boundary conditions. Simulations were performed for different ocean basins where tropical cyclones are occurring today. It might be possible that under future climates, some basins, such as the South Atlantic, might become more or less prone for the occurrence of these storms. However, these were not simulated here, likely because of a lack of opportunity to calibrate the simplified Tropical Cyclone model. Other limitations arise from the uncertainties from within model parameters. For example, the CHAZ model has been used in other configurations in the genesis part of the model (using saturation deficit rather than column relative humidity), which resulted in different outcomes (see Lee et al., 2020; Fosu et al., 2024)[1]. While in some basins the

cyclone frequency appears to scale reasonably well with prevalent surface temperatures, in others (especially the Atlantic), the response is more complex as multiple dynamical factors in atmosphere and ocean as well as presence of aerosol can be affecting cyclone formation by enhancing or offsetting each other.

Overall, the user is advised to use the projections with caution and regard the products as possible outcomes. The goal of this collection is to allow users to familiarize themselves with the type of probabilistic data and then to engage with the scientific community to better establish an idea of which elements of projections might be more robust than others in their region of interest.

Caveats in Regional Statistics:
While cyclone counts across ocean basins have been validated and bias-corrected against observed values, regional statistics may still show discrepancies. Models provide greater precision in calculating probabilities, but users should recognize that the numbers may systematically differ from actual occurrences, either higher or lower. Generally, the larger the region, the more accurate the statistics. This is particularly important when considering landfall data, and even more so for small islands, where model uncertainties tend to be more pronounced.

DATA PROCESSING STEPS

WORKFLOW FOR CYCLONES – CHAZ MODEL

1. Original Data: The original CHAZ data was provided with storm locations and maximum winds every six hours.

   Gridded Data: These data points were then projected onto a global grid with a resolution of 0.5x0.5 degrees (approximately 50 km x 50 km).

2. Temporal Interpolation: This data is then interpolated to half-hour intervals to create a continuous center track line on the grid.

3. Footprint: A footprint is subsequently applied to capture the full extent of the cyclones, as cyclones are typically larger than 0.5 degrees (roughly 50 km). Applying the full footprint is particularly important for small islands to avoid underestimating the cyclone's impact. The footprint is based on modeled horizontal wind profiles and latitude, which use a dual-exponential decay function derived from 380 observed storms, as detailed by Willoughby et al. (2006) 11 :

   

   With:

   

   

   The resulting Vo is the modeled wind maximum velocity at each radius raway from the center, calculated on a 7x7 matrix (350km diameter) surrounding the storm's center in our case. If the storm is larger, the outer part of the storm is excluded, though this is rare. The approximation needs two inputs, Vmax - the maximum wind speed from CHAZ - and φ - the storm's latitude. Rmax represents the radius at which velocity is Vmax.Here, we assume that the center grid-cell is the only cell with highest winds of V max even if the radius is >25km (25km refers to half of the 0.5deg pixel), which happens for weaker tropical cyclones. Then, X2, which refers to the rapid decay length away from the storm's center, is assumed to be fixed at 25 km (deviating from Willoughby et al., 2006). X 1 represents the slower decay length of the wind profile in the outer bands of the tropical cyclone and can be calculated with the formula above.

   Here is a schematic illustrating the process of transforming the original CHAZ data into a pixelized version with a footprint: a) the original CHAZ points, b) temporal interpolation to complete the track, c) addition of the footprint, and classification into different categories (represented by colors). 

    

   

    
4. Multiple Track Generation: These tracks are generated for all cyclones across each simulation that consist of 6 separate basin results, 40 ensemble members per global driving model (12 CMIP6 models), and future scenario SSP (historical and SSP2-4.5).
5. Bias-Correction: In order to achieve annual counts similar to observations, CHAZ estimates the number of effective years that the simulations cover (as a form of bias correction). Two different numbers are applied: one for the open ocean and a separate number for landfalls.
6. Calculation Of Statistical Products
   1. DENSITY MAP:For each cyclone, the spatial wind calculations are converted to storm categories using the Saffir-Simpson scale. Then, for each map pixel, we count the number of cases in each storm category across all simulations and divide that count by the number of effective years of simulation, resulting in a map of storm density by category (number of cyclones per pixel per year). Each storm is only counted once per grid cell. Note: For the density map, we apply the ocean basin bias correction, including over land areas.

   2. SPATIAL STATISTICS: For each stack of simulated cyclones, we calculate occurrence statistics for each basin, Exclusive Economic Zone (EEZ), and landfall events:

      1. OCEAN BASINS: We analyze the peak intensity of each cyclone to determine its timing and ocean basin attribution (so each cyclone is only linked to one ocean basin for statistics). This enables us to count the number of storms per basin per year. The width of the footprint has no impact on basin-level statistics. The effective years of simulation is applied as bias correction for “values per year”.
      2. ii. EXCLUSIVE ECONOMIC ZONES (EEZs): We analyze whether a cyclone track intersects with a given EEZ. This includes all grid cells of a cyclone and thus includes the broader footprint of the storms. An EEZ intersection statistic will be determined by the highest storm category from all overlapping cells with the area of an EEZ. The footprint is particularly important here, especially for small EEZs. The effective years are applied as bias-correction. iii.LANDFALL: Similarly, we count the number of storms that intersect with the official country or territory polygon (ADM0). As with EEZs, the highest category of overlapping grid cells are counted. The footprint is also particularly relevant when calculating landfall statistics. The separate effective years for landfall are applied for biascorrection.

      NOTE:The counts and annual exceedance probabilities for ocean basins, EEZs, and landfalls are based on counts of storms. Each storm only gets counted once. Therefore, these products are not spatial averages of the gridded fields shown in maps where the same cyclone may be counted in each pixel it intersects.

7. Ensembles:The process is repeated across all simulated tropical cyclones, for each of the 6 basin simulations with CHAZ, and across all 40 ensembles from any single driving GCM. Then, across all 12 CMIP6 models, a multi-model ensemble is formed, which is represented by the median value as well as the 10th and the 90th percentile to reflect crossmodel variability.

OUTPUT DATA – CHAZ MODEL

GRIDDED PRODUCTS:
We provide the annual exceedance probability ('aep') and return period ('returnperiod') as multimodel ensemble medians for each of the Saffir-Simpson storm category boundaries. An annual exceedance probability for Tropical Storms of 2.5 means that on average between 2 and 3 tropical storms of at least34kn can be expected each year. These counts can include also storms with much higher winds (TS to Cat5). Annual exceedance probabilities of Category 4 storms, however, do only include storms with maximum winds above 113kn. These numbers are computed for the historical period (1951-2014) and a future period based on the SSP2-4.5 scenario (2035-2064) and presented at an interpolated grid of 0.5x0.5 degrees. We also provide the anomaly products between the projected period and the corresponding historical reference ('faep', for median, p10 and p90 to account for uncertainties across models: note because of the current limitation to 12 models, p10 is represented by the second lowest, and p90 by the second highest of the ranked individual model results). The anomaly is expressed as a ratio or fractional change (future period results divided by the corresponding values in the historical run). A fractional change of less than 1 indicates a decrease in probability (a lower annual exceedance probability) or an increase in return period by that factor, while a fractional change greater than 1 indicates an increase in probability or a decrease in return period. One visualization also converts the change into a percent change of counts at each Tropical Cyclone level.

SPATIALLY AGGREGATED PRODUCTS:
The following spatially aggregated products can be download by users:

• Cyclone occurrence as number counts per year ('counts') and percentages ('tcfraction') for each storm category, as well as annual exceedance probability ('aep') for each category and above (at least this category) as spatially aggregated to global, ocean basin, Exclusive Economic Zone (EEZ), and country/territory (landfalls) for historical and projected SSP2-4.5.
• Annual timeseries of counts over the historical period (1951-2014) have been generated for three broader intensity groups (annual temporal aggregation): tropical storms, minor cyclones (cat 1 and cat 2, 'tcminor') and major cyclones (cat 3 and above, 'tcmajor').
• As a representation of the seasonal cycle of “presence of tropical cyclone activity” in all of the ocean basins, the 5-day moving averaged daily storm counts (normalized to per 100 years) have been aggregated for tropical storms, minor cyclones, and major cyclones.

[10] Fosu, B. O., Sobel, A. H., Lee, C.-Y., Camargo, S. J., Tippett, M. K., Hemmati, M., Drinka, R., Polamuri, S. H.,

Bowen, S. G., & Bloemendaal, N. (2024). Assessing Future Tropical Cyclone Risk Using Downscaled CMIP6

Projections. Journal of Catastrophe Risk and Resilience, 2024(2), https://doi.org/10.63024/dpva -2pa1 

[11] Wiloughby, H.E., Darling, R. W. R., and Rahn, M. E. (2006). Parametric Representation of the Primary

Hurricane Vortex. Part II: A New Family of Sectionally Continuous Profiles. AMS Monthly Weather Review. 134 (4) 1102-1120. https://doi.org/10.1175/MWR3106.1 

Sea Level Change

Rising sea levels contribute to coastal flooding, erosion, and the loss of habitat for both human populations and wildlife. As the climate warms, thermal expansion of seawater and the melting of land-based ice sources - such as glaciers and the Greenland and Antarctic ice sheets - are the primary drivers of sea level rise. In addition, regional variations in sea level are influenced by local factors, such as vertical land motion and oceanic processes like ocean current changes, making it essential to study both global and localized trends. These drivers of sea level change are explained in more detail below.

Understanding past and present sea level changes is crucial for projecting future risks and for developing adaptation and resilience strategies for affected communities. By analyzing historical sea level records and reconstructions (data which combines observed tide gauge measurements with satellite altimetry data to create longer records of sea level changes), scientists are able to track the rate of change and identify trends in different regions of the world. This information can help policymakers and planners understand changing ocean dynamics and anticipate the impacts of future sea level rise to guide decisions on coastal infrastructure, conservation, and disaster preparedness.

Sea Level Change Measurement Methods

The data used to study sea level change is derived from multiple sources, each offering unique insights into global and regional trends.

1. Satellite Altimetry Since 1992, satellite altimeters, such as those on NASA’s TOPEX/Poseidon, Jason-1, and Jason-2 satellites, have provided precise global measurements of sea surface height. Using radar pulses, they track changes in sea level across remote and coastal areas, revealing long-term trends and patterns on a global scale.
2. Tide Gauges Tide gauges, in use for over a century, measure sea level relative to fixed land points at coastal locations. While geographically limited, these instruments provide valuable local data on sea level trends and complement satellite observations by capturing regional variations due to land subsidence or ocean dynamics.
3. Probabilistic Reconstruction By combining tide gauge records, satellite data, and geophysical models, researchers create probabilistic reconstructions of sea level changes. These account for uncertainties and quantify contributions from key factors like ice melt, ocean warming, and land motion, producing a comprehensive long-term record.

Key Drivers of Sea Level Change

Sea level rise is driven by several interconnected processes that shape both regional and global patterns. The main drivers include:

1. Ice Sheets and Glaciers
   • Greenland and Antarctic Ice Sheets: Major contributors to global sea level rise as warmer temperatures accelerate ice melting, releasing freshwater into the oceans.
   • Mountain Glaciers: Their retreat adds additional freshwater to the seas, further amplifying global sea level trends.
2. Vertical Land Motion can lead to localized sea level changes and can either exaggerate or mitigate the effects of global sea level rise.
   • Natural Causes: Tectonic shifts or sediment settling can cause the land to rise or sink.
   • Human Activities: Groundwater extraction or resource removal exacerbates subsidence.
3. Sterodynamics
   • Thermal Expansion: Oceans absorb heat from global warming, causing water to expand and raise sea levels.
   • Ocean Currents: Altered by freshwater input from melting ice and changing heat distribution, these shifts create regional sea level variations.
4. Terrestrial Water Storage
   • Groundwater Depletion: Pumping groundwater adds water to the ocean, increasing sea levels.
   • Land Retention: Water stored in reservoirs, soils, or other terrestrial systems can offset sea level rise temporarily.

Together, these drivers interact to produce the observed trends and regional variations in sea level rise, highlighting the complex dynamics of Earth's changing climate system.

HISTORICAL SEA LEVEL CHANGE
 

ORIGINAL DATASETS (consolidated by NASA and made available via NASA Global Sea Level Change)

1. Tide gauge data: Permanent Service for Mean Sea Level (PSMSL)
2. Altimetry data: NASA - Integrated Multi-Mission Ocean Altimeter Data for Climate Research (NASA-SSH)
3. Reconstruction product: Kalman Smoother Sea Level Reconstruction
   • This data product offers a near-global, probabilistic analysis of sea level changes from 1900 to present, combining tide gauge records, satellite data, and geophysical models to fill gaps in long-term observations. By resolving individual processes like ice melt, land motion, and ocean dynamics, it provides valuable insights into historical trends and regional variations, aiding coastal communities in understanding and adapting to rising sea levels.
   • Citation: Dangendorf, S., Sun, Q., Wahl, T., Thompson, P., Mitrovica, J. X., and Hamlington, B.: Probabilistic reconstruction of sea-level changes and their causes since 1900, Earth Syst. Sci. Data, 16, 3471–3494, https://doi.org/10.5194/essd-16-3471-2024, 2024.

AGGREGATED DATASETS FOR CCKP USERS

1. Global Mean Sea Level
2. Gridded Sea Level
3. Tide Gauge Sea Level
4. EEZ Sea Level

For each of the above, the following are available via API:

1. Reconstructed sea level data (annual): 1970 - 2021, baseline 2005
2. Satellite altimetry data (monthly): 1993 - present, baseline 2005
3. Tide gauge data (monthly): 1993 - present, baseline 2005
4. High Tide Flooding (HTF) data: 1970 - present
   • Tracks the number of days per year where sea levels exceed set thresholds (40cm[400mm], 60cm[600mm], 80cm[800mm] above average high tide), indicating increasing flood risk. These thresholds provide a standardized way to assess changes in coastal flooding potential over time. Exceeding thresholds signals higher flood risk, but impacts vary by location. This dataset tracks trends but doesn’t confirm flooding on HTF days.
5. Percent contribution of various processes to total sea level change (average from 1993 - present)

Description Of Climate Indicators