Data in Brief 54 (2024) 110502

Contents lists available at ScienceDirect

Data in Brief

journal homepage: www.elsevier.com/locate/dib

Invited Data Manuscript

A dataset to measure global climate physical

risk
Kun Guo a,b, Qiang Ji c,d, Dayong Zhang e,∗
a
School of Economics and Management, University of Chinese Academy of Sciences, Beijing 100190, China
b
Research Center on Fictitious Economy & Data Science, Chinese Academy of Sciences, Beijing 100190, China
c
Institutes of Science and Development, Chinese Academy of Sciences, Beijing 100190, China
d
School of Public Policy and Management, University of Chinese Academy of Sciences, Beijing 100049, China
e
Research Institute of Economics and Management, Southwestern University of Finance and Economics,
Sichuan 611130, China

a r t i c l e i n f o a b s t r a c t

Article history: Extreme climate events have become more frequent and have
Received 8 April 2024 had serious impacts on the global community. Consequently,
Revised 25 April 2024 the risk associated with climate change has gained increasing
Accepted 30 April 2024 attention and has been considered as a new source of risk
Available online 4 May 2024
factors. To understand the socio-economic impacts of this
Dataset link: Climate Physical Risk Index
new risk, systematically measuring risk around the world is
(CPRI) (Original data) critical for researchers and policymakers. Building on daily
observations from meteorological stations, a Climate Physi-
Keywords: cal Risk Index (CPRI) dataset is constructed for 170 coun-
Climate risk
tries, paying special attention to four extreme climate events:
Extreme temperature
extreme low temperature (LTD), extreme high temperature
Extreme rainfall
Extreme drought
(HTD), extreme rainfall (ERD), and extreme drought (EDD). A
comprehensive index of climate physical risk for each coun-
try has also been constructed, covering the period from 1993
to 2023. The dataset will be updated regularly. Subnational
indices or more detailed regional indices are available upon
request.
© 2024 The Author(s). Published by Elsevier Inc.
This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/)

∗
Corresponding author.
E-mail address: [email protected] (D. Zhang).

https://doi.org/10.1016/j.dib.2024.110502
2352-3409/© 2024 The Author(s). Published by Elsevier Inc. This is an open access article under the CC BY license
(http://creativecommons.org/licenses/by/4.0/)
2 K. Guo, Q. Ji and D. Zhang / Data in Brief 54 (2024) 110502

Specifications Table

Subject Climate and Environmental Finance

Specific subject area An interdisciplinary subject area connecting climate change and environmental
issues with financing and investment issues.
Type of data Table (Excel file)
Data collection Our CRPI indices are constructed based on daily observations from
meteorological stations worldwide. Four extreme climate events are considered
and then combined to form a general index for each country. These events are
extreme low temperature (LTD), extreme high temperature (HTD), extreme
rainfall (ERD), and extreme drought (EDD). The first step is to define the
“extreme” level. Following the literature, historical information, such as
observations from 1973 to 1992, are used to determine the 10th and 90th
percentiles of temperature, the 95th percentile of rainfall, and the 5th
percentile of humidity, which are then used as thresholds. The total number of
days in each of the four categories that exceed these thresholds in each year is
calculated from 1993 to 2023. The four sub-indices are then standardized
through a min-max approach (explained later). The mean value of these
standardized sub-indices is used to represent the country specific CPRI.
Data source location All raw data are downloaded from the National Oceanic and Atmospheric
Administration (NOAA), Washington, DC, U.S.
Data accessibility Repository name: figshare
Data identification number: 10.6084/m9.figshare.25562229.v1
Direct URL to data: https://doi.org/10.6084/m9.figshare.25562229.v1

1. Value of the Data

• These data are useful in evaluating the degree of climate physical risk worldwide. Unlike
some existing measures that use natural disasters and their associated economic impacts to
construct such indices (e.g., Global Climate Risk Index | Germanwatch e.V.), the information
here is more consistent, with better accuracy, in higher frequency, and can be extended to
cover smaller regional climate conditions as the raw data is taken from actual observations
by the meteorological stations at daily frequency.
• The core data covers information over 30 years for 170 countries and can be updated eas-
ily to provide a balanced panel. Given that the source is at the station level, our approach
can be easily extended to calculate subnational or regional level of physical climate risks.
Taking China as an example, the indices for 31 provinces and 229 prefectural cities are
calculated.
• Although similar sources of information have been used by individual researchers, a compre-
hensive dataset open to the public and rich enough to capture different categories of climate
events worldwide is yet to be available. This dataset can fill the gap by providing publicly
available assessments of different types of climate physical risks.
• Researchers, business managers, financial analysts, and policy makers can use this dataset
to analyze the macro or micro impacts of climate physical risk on the economy and soci-
ety, thereby actively seeking strategies to mitigate the negative impacts of climate risk on
macroeconomy, financial markets, firms and institutions, and others.

2. Background

Over recent decades, the frequency of extreme climate events such as heatwaves, extreme
precipitation, and droughts has increased substantially [1]. Not only do these events cause di-
rect losses to the global community through the destruction of infrastructure, reduction in pro-
ductions, or increase in human mortality, but they also introduce significant uncertainties and
indirectly lead to serious economic consequences, such as increasing systemic risks in the fi-
nancial system [2–5]. As a consequence, climate-related risks have drawn increasing attention
 K. Guo, Q. Ji and D. Zhang / Data in Brief 54 (2024) 110502 3

Tabel 1
Overview of the content of the repository.

File name Sheet name Sample Indicators

Global Climate Physical GCPRI 170 countries between LTD

Risk Index (GCPRI).xlsx 1993 and 2023 HTD
ERD
EDD
CPRI
Chinese Climate Province level 31 provinces of China LTD
Physical Risk Index between 1993 and HTD
(CCPRI).xlsx 2023 ERD
EDD
CPRI
City level 229 cities of China City
between 1993 and LTD
2023 HTD
ERD
EDD
CPRI

in the literature and also been widely considered by practitioners and authorities. For exam-
ple, the Basel Committee on Banking Supervision (BCBS) has issued principles for the effective
management and supervision of climate-related financial risks.1
Typically, climate risks can be divided into physical risks and transitional risks [6]. While
physical risk refers to the direct climate impacts and is often captured by extreme weather
events or natural disasters [7], transitional risks are uncertainties due to efforts to mitigate cli-
mate changes, e.g., climate policy uncertainties [8].
In practice, meteorological data have been used to study climate impacts [9]. However, a gen-
eral dataset covering climate physical risks worldwide has yet to be developed. Using a natural
disaster database such as EM-DAT is one option [10], but it has notable problems compared
to measurements build on meteorological information. The CPRI (Climate Physical Risk Index)
dataset is constructed based on station-level meteorological observations. It can fill the gap and
provide a flexible panel data easily adapted to different purpose of usage.

3. Data Description

The dataset stored in the repository [11] contains two Excel files, one referring to the
global climate physical risk index (GCPRI) system for 170 countries, and another demonstrat-
ing the capability of extending the indices to sub-national or smaller regional level using
the case of China, which includes Chinese climate physical risk indices for 31 provinces and
229 cities.
Each set of CPRI data contains four sub-indices and a total index, covering the period from
1993 to 2023. The contents of the repository are presented in Table 1. LTD (Extreme Low Tem-
perature Days), HTD (Extreme High Temperature Days), ERD (Extreme Rainfall Days), EDD (Ex-
treme Drought Days) are four sub-indices representing the number of extreme low tempera-
ture days, extreme high temperature days, extreme rainfall days, and extreme drought days in
a country/region during a year, respectively. These indices are standardized using the method
explained below and are then used to construct the general CPRI, which refers to the overall
degree of climate physical risk of a country/region.

1
https://www.bis.org/bcbs/publ/d532.htm.
4 K. Guo, Q. Ji and D. Zhang / Data in Brief 54 (2024) 110502

4. Experimental Design, Materials and Methods

4.1. Data collection

The raw meteorological data are sourced from NOAA (the National Oceanic and Atmo-
spheric Administration).2 Daily climatological information from meteorological stations around
the world (excluding special types of meteorological stations such as military bases and airports)
from 1973 to 2023 is collected and used to construct the indices. The original indicators include
daily average temperature, daily rainfall, and daily dew point observations. Here the dew point
data is primarily used to calculate humidity in conjunction with the temperature data, which is
then used to measure the risk of extreme drought. Historical data dating back to 1973 will be
used to calculate the thresholds for extreme values. Specifically, 20 years of historical observa-
tions (from 1973 to 1992) are used for this purpose, therefore the effective CPRI data spans from
1993 to 2023.

4.2. Data preprocessing

Step 1. Missing value processing. If a meteorological station has a large amount of missing
data, it is removed from the sample.
Step 2. Calculating the historical distributions of each indicator between January 1, 1973 and
December 31, 1992. Define Ti10 as the lower 10th percentile of the historical daily average tem-
perature of station i, representing the threshold value for extreme low temperature; Define Ti90
as the 90th percentile of the historical daily average temperature of station i, representing the
threshold value for extreme high temperatures; R95 i
is defined as the 95th percentile of historical
daily rainfall of station i, representing the threshold value for extreme rainfall; Hi5 is defined as
the 5th percentile of historical daily humidity of station i, representing the threshold value for
extreme drought.
Step 3. Counting the number of extreme days for each station from 1993 to 2023 for each
type of event. The extreme low temperature days for station i in year n can be defined as LT Di,n .


365
LT Di,n = LTi,n,t (1)
t=1

1 i f Ti,n,t < Ti10
LTi,n,t = (2)
0 i f Ti,n,t ≥ Ti10
Where Ti,n,t represents the average temperature of station i in year n and day t. Similarly,
extreme high temperature days HT Di,n , extreme rainfall days ERDi,n and extreme drought days
EDDi,n can be calculated.

365
H T Di,n = HTi,n,t (3)
t=1

1 i f Ti,n,t > Ti90
HTi,n,t = (4)
0 i f Ti,n,t ≤ Ti90


365
E RDi,n = ERi,n,t (5)
t=1

1 i f Ri,n,t > R95
i
ERi,n,t = (6)
0 i f Ri,n,t ≤ R95
i

2
https://www.ncei.noaa.gov/data/global- summary- of- the- day/archive/.
 K. Guo, Q. Ji and D. Zhang / Data in Brief 54 (2024) 110502 5


365
E DDi,n = EDi,n,t (7)
t=1


1 i f Hi,n,t < Hi5
EDi,n,t = (8)
0 i f Hi,n,t ≥ Hi5

Where Ri,n,t and Hi,n,t represent rainfall and relative humidity of station i in year n and day
t, respectively.
Step 4. Calculating the annual number of extreme weather days at the regional level. Using
the geographical coordinates of each station, the metrological data can be mapped to a certain
region (country/province/city), so the average number of extreme days of the covered stations
in the region can be calculated. Take the extreme low temperature days LT Dm,n of the region m
in year n for example, the formula is:

1 
M
LT Dm,n = LT D j,n (9)
M
j=1

Where M is the number of stations in the region, and LT Dm,n is the arithmetic average ex-
treme low temperature days of all stations in the area. Similarly, HT Dm,n , ERDm,n and EEDm,n of
the region m can be obtained.

4.3. Computation of the index

Given that the four individual extreme climate measures differ in nature, they are not directly
comparable. The min-max standardization approach is used to process the data to construct a
general index in each category. Taking the extreme low temperature days as an example:
 
LT Dm,n − min LT Dk,l
k=1,··· ,K;l=1··· ,L
LT Dm,n =     × 100 (10)
max LT Dk,l − min LT Dk,l
k=1,··· ,K;l=1··· ,L k=1,··· ,K;l=1··· ,L

Where K is the total number of countries/regions, and L is the total number of years in the
sample set. In this way, the four sub-indices LTD, HTD, ERD and EDD can be obtained.
After standardizing, the four sub-indices are used to construct a general climate physical risk
index for each country/region. Here the simple average is used such as:

CP RIm,n = ω1 LT Dm,n + ω2 HT Dm,n + ω3 ERDm,n + ω4 EDDm,n (11)

In this dataset, ωi is set to 0.25. Obviously, the weights can be changed to serve any individ-
ual purpose in a particular project.

Limitations

One of the main issues with this dataset is the use of simple average of the extreme climate
days across all meteorological stations in a country/region, which essentially assumes that the
importance of different meteorological station is the same. However, some stations may be lo-
cated in the more central areas, whereas others may be located in sparsely populated areas. Of
course, this concern is trivial for a global panel. For individual researchers focusing on a particu-
lar region, population density or night light data can be used to create a weighted average index
to gain more accurate measures. Another problem may require attention is when extending to
smaller regions, the availability of meteorological station data can be a limiting factor for the
minimum size of the regional extension to be considered.
6 K. Guo, Q. Ji and D. Zhang / Data in Brief 54 (2024) 110502

Ethics Statement

The authors have read and followed the ethical requirements for publication in Data in Brief
and confirmed that the current work does not involve human subjects, animal experiments, or
any data collected from social media platforms.

Data Availability

Climate Physical Risk Index (CPRI) (Original data) (figshare).

CRediT Author Statement

Kun Guo: Conceptualization, Methodology, Software, Formal analysis, Visualization, Writing

– original draft; Qiang Ji: Data curation, Investigation, Validation, Supervision, Project admin-
istration, Funding acquisition; Dayong Zhang: Investigation, Supervision, Validation, Writing –
review & editing, Project administration.

Acknowledgments

This work was supported by the National Natural Science Foundation of China [grant num-
bers 72348003, 72022020, 71974159].

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal rela-
tionships that could have appeared to influence the work reported in this paper.

References

[1] World Meteorological Organization (WMO), State of the global climate 2023, https://library.wmo.int/idurl/4/68835
(accessed 7 April 2024).
[2] J. Rising, M. Tedesco, F. Piontek, et al., The missing risks of climate change, Nature 610 (2022) 643–651, doi:10.1038/
s41586- 022- 05243- 6.
[3] K. Guo, Y. Li, Y. Zhang, Q. Ji, W. Zhao, How are climate risk shocks connected to agricultural markets? J. Commod.
Mark. 32 (2023) 100367, doi:10.1016/j.jcomm.2023.100367.
[4] Y. Zhou, S. Wu, Z. Liu, et al., The asymmetric effects of climate risk on higher-moment connectedness among carbon,
energy and metals markets, Nat. Commun. 14 (2023) 7157, doi:10.1038/s41467- 023- 42925- 9.
[5] S. Dietz, A. Bowen, C. Dixon, et al., Climate value at risk’ of global financial assets, Nat. Clim. Change 6 (2016)
676–679, doi:10.1038/nclimate2972.
[6] D. Zhang, Y. Wu, Q. Ji, K. Guo, B. Lucey, Climate impacts on the loan quality of Chinese regional commercial banks,
J. Int. Money Financ. 140 (2024) 102975, doi:10.1016/j.jimonfin.2023.102975.
[7] I. Shala, B. Schumacher, The impact of natural disasters on banks’ impairment flow–Evidence from Germany, J. Clim.
Financ. 6 (2024) 10 0 031, doi:10.1016/j.jclimf.2024.10 0 031.
[8] Y. Ma, Z. Liu, D. Ma, et al., A news-based climate policy uncertainty index for China, Sci. Data 10 (2023) 881, doi:10.
1038/s41597- 023- 02817- 5.
[9] S.O. Lee, N.C. Mark, J. Nauerz, J. Rawls, Z. Wei, Global temperature shocks and real exchange rates, J. Clim. Financ. 1
(2022) 10 0 0 04, doi:10.1016/j.jclimf.2022.10 0 0 04.
[10] Y. Shen, X. Sun, Q. Ji, D. Zhang, Climate events matter in the global natural gas market, Energy Econ. 125 (2023)
106887, doi:10.1016/j.eneco.2023.106887.
[11] K. Guo, Q. Ji, D. Zhang, Climate physical risk index (CPRI), figshare, Dataset (2024), doi:10.6084/m9.figshare.
25562229.v1.