Observed Exposure of Population and Gross Domestic Product to Extreme Precipitation Events in the Poyang Lake Basin, China

: Based on the observation data from the Poyang Lake Basin (China), an extreme precipitation event (EPE) is deﬁned as that for which daily precipitation exceeded a threshold of 50 mm over a continuous area for a given time scale. By considering the spatiotemporal continuity of EPEs, the intensity–area–duration method is applied to study both the characteristics of EPEs and the population and gross domestic product (GDP) exposures. The main results are as follows. (1) During 1961–2014, the frequencies and the intensities of the EPEs are found to be increasing. (2) The annual area impacted by EPEs is determined as 7.4 × 10 4 km 2 with a general upward trend of 400 km 2 / year. (3) The annually exposed population is estimated as 19% of the entire population of the Basin, increasing by 1.37 × 10 5 / year. The annual exposure of GDP is 8.5% of the entire GDP of the Basin, increasing by 3.8 billion Yuan / year. The Poyang Lake Basin experiences serious extreme precipitation with increasing trends in frequency, intensity, and exposure (for both GDP and population). It is imperative that e ﬀ ective disaster prevention and reduction measures be adopted in this area to mitigate the e ﬀ ects of extreme precipitation.


Introduction
From 1880 to 2012, the global average temperature has risen by 0.85 • C. With global warming, the probability of occurrence of extreme precipitation events (EPEs) has increased regionally and globally [1]. Extreme precipitation is one of the most severe disasters affecting China, causing 37.2% of economic losses and 11.7% of casualties related to meteorological disasters from 1984 to 2014 [2]. In the Poyang Lake Basin, the situation regarding extreme precipitation events is even more serious. From 1984 to 2014, 65% of economic losses and 83% of casualties related to meteorological disasters are attributable to extreme precipitation and derivative disasters [3]. Therefore, it is essential to understand the spatiotemporal distribution and evolution of extreme precipitation events in this area.
The risks posed by global warming, driven by continual industrialization, have become a major challenge to global security and development. The severity of the effects of extreme events depends not only on the actual extremes but also on the degree of exposure and vulnerability. Here, exposure refers to the impact of the adverse effects of extreme events on the population, gross domestic product (GDP), and other aspects [4,5]. One of the main reasons for the growth in economic losses is the increase of the human and economic assets exposed to extreme events [6]. In China, because of the

Basic Geographic Information Data
The basic geographic information dataset is provided by Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences (RESDC) (http://www.resdc.cn). China's 1:100,000 scale land use status remote sensing monitoring database is currently the most accurate land use remote sensing dataset, and plays an important role in the national land resources survey, and in hydrological and ecological research. The land use types include farmland, forest land, grassland, water area, residential land, and unused land.

Population and GDP Data
The population and GDP data are collected to study the population exposed to EPEs. Yearly population and GDP data for the Poyang Lake Basin are derived from the Jiangxi Province Statistical Yearbook (1984-2014), which comprises county-level statistical data (Figures 2 and 3). The population has grown from 34.5 million (1984) to 45.4 million (2014), with increasing rate 0.34 million/year [28].
Without considering the inflation, the GDP has increased from 16.9 billion Yuan (1984) to 1571.5 billion Yuan (2014), with an increase rate of 43.6 billion/year. Here, considering the changes of consumer price index (CPI), the GDP data are normalized to 2014 based on the index. The GDP has increased from 88.6 billion Yuan (1984) to 1571.5 billion Yuan (2014), with an increase rate of 44.7 billion/year. The Poyang Lake Basin is divided into 87 grids (0.5° × 0.5°), consistent with the precipitation dataset (Crosses signs in Figure 1). Similarly, 0.5° × 0.5° gridded GDP and population datasets for each year are produced in terms of the ratio between grid and county areas.

Basic Geographic Information Data
The basic geographic information dataset is provided by Data Center for Resources and Environmental Sciences, Chinese Academy of Sciences (RESDC) (http://www.resdc.cn). China's 1:100,000 scale land use status remote sensing monitoring database is currently the most accurate land use remote sensing dataset, and plays an important role in the national land resources survey, and in hydrological and ecological research. The land use types include farmland, forest land, grassland, water area, residential land, and unused land.

Population and GDP Data
The population and GDP data are collected to study the population exposed to EPEs. Yearly population and GDP data for the Poyang Lake Basin are derived from the Jiangxi Province Statistical Yearbook (1984-2014), which comprises county-level statistical data (Figures 2 and 3). The population has grown from 34.5 million (1984) to 45.4 million (2014), with increasing rate 0.34 million/year [28].
Without considering the inflation, the GDP has increased from 16.9 billion Yuan (1984) to 1571.5 billion Yuan (2014), with an increase rate of 43.6 billion/year. Here, considering the changes of consumer price index (CPI), the GDP data are normalized to 2014 based on the index. The GDP has increased from 88.6 billion Yuan (1984) to 1571.5 billion Yuan (2014), with an increase rate of 44.7 billion/year. The Poyang Lake Basin is divided into 87 grids (0.5 • × 0.5 • ), consistent with the precipitation dataset (Crosses signs in Figure 1). Similarly, 0.5 • × 0.5 • gridded GDP and population datasets for each year are produced in terms of the ratio between grid and county areas.

Spatializion Method for Population
Population is an important factor in vulnerability assessment of the disaster carrying capacity of rainstorm and flood disasters, and fine spatial distribution information of population is an important basis for vulnerability assessment. To distinguish towns with an agricultural population and non-agricultural population, the study area is divided into urban areas and rural areas.
This study analyzed the correlation between the population and the land area of each town, and obtained the land type factors that affect the population distribution of cities and townships, and then established a spatial model of urban and rural population based on land type. The general formula for the population spatialization model is as follows: where Pi denotes the total population of town I, aj is the population distribution factor of the land use type j, xj is the land use area of type j, n is the number of land use types that affects the population distribution, and Bi is the intercept. This study simulated the population space with the land use data and the results of the simulation revealed that the population is mainly concentrated in the Central Plains and southern

Spatializion Method for Population
Population is an important factor in vulnerability assessment of the disaster carrying capacity of rainstorm and flood disasters, and fine spatial distribution information of population is an important basis for vulnerability assessment. To distinguish towns with an agricultural population and non-agricultural population, the study area is divided into urban areas and rural areas.
This study analyzed the correlation between the population and the land area of each town, and obtained the land type factors that affect the population distribution of cities and townships, and then established a spatial model of urban and rural population based on land type. The general formula for the population spatialization model is as follows: where Pi denotes the total population of town I, aj is the population distribution factor of the land use type j, xj is the land use area of type j, n is the number of land use types that affects the population distribution, and Bi is the intercept. This study simulated the population space with the land use data and the results of the simulation revealed that the population is mainly concentrated in the Central Plains and southern

Spatializion Method for Population
Population is an important factor in vulnerability assessment of the disaster carrying capacity of rainstorm and flood disasters, and fine spatial distribution information of population is an important basis for vulnerability assessment. To distinguish towns with an agricultural population and non-agricultural population, the study area is divided into urban areas and rural areas.
This study analyzed the correlation between the population and the land area of each town, and obtained the land type factors that affect the population distribution of cities and townships, and then established a spatial model of urban and rural population based on land type. The general formula for the population spatialization model is as follows: where P i denotes the total population of town I, a j is the population distribution factor of the land use type j, x j is the land use area of type j, n is the number of land use types that affects the population distribution, and B i is the intercept. This study simulated the population space with the land use data and the results of the simulation revealed that the population is mainly concentrated in the Central Plains and southern mountain areas  Figure 4). This population distribution characteristic may cause heavy casualties in the mountain area by the EPEs.
Atmosphere 2020, 11, x FOR PEER REVIEW 5 of 15 mountain areas ( Figure 4). This population distribution characteristic may cause heavy casualties in the mountain area by the EPEs.

Spatialization Method for GDP
In addition to evaluating the vulnerability of the population, another important indicator that needs to be considered in the meteorological disaster risk assessment is the economic development of the region. The GDP, used in most countries and regions of the world, can reflect the full social and economic activities. According to the historical development of social production activities, the division of industrial structure usually divides GDP into three industries [28]. We established the expression model of GDP data space based on the spatial pattern of land use by studying the key factors influencing the development and distribution of GDP from the various industries.

Primary Industry Model
Primary industry usually includes the four branches of agriculture, forestry, animal husbandry, and fisheries. This study analyzed the correlation between the added value of agriculture, forestry, animal husbandry and fishery in each town and the land use type, and obtained the influencing factors of agriculture, forestry, animal husbandry, and fisheries in GDP development. We established the primary industry GDP spatial distribution model and GDP value is expressed by G as follows: where G1j is the GDP of primary industry of town j, G , G , G , and G are the GDP of agriculture, forestry, animal husbandry, and fisheries at town j, respectively;

Spatialization Method for GDP
In addition to evaluating the vulnerability of the population, another important indicator that needs to be considered in the meteorological disaster risk assessment is the economic development of the region. The GDP, used in most countries and regions of the world, can reflect the full social and economic activities. According to the historical development of social production activities, the division of industrial structure usually divides GDP into three industries [28]. We established the expression model of GDP data space based on the spatial pattern of land use by studying the key factors influencing the development and distribution of GDP from the various industries.

Primary Industry Model
Primary industry usually includes the four branches of agriculture, forestry, animal husbandry, and fisheries. This study analyzed the correlation between the added value of agriculture, forestry, animal husbandry and fishery in each town and the land use type, and obtained the influencing factors of agriculture, forestry, animal husbandry, and fisheries in GDP development. We established the primary industry GDP spatial distribution model and GDP value is expressed by G as follows: where G 1j is the GDP of primary industry of town j, G agr j , G for j , G ani j , and G fis j are the GDP of agriculture, forestry, animal husbandry, and fisheries at town j, respectively; g agr j , g for j , g ani j , and g fis j are the unit area GDP of agriculture, forestry, animal husbandry, and fisheries at town j, respectively; A agr ij , A for ij , A ani ij , and A fis ij denote the area of the ith land use type at town j that affects the development of agriculture, forestry, animal husbandry, and fishery industries, respectively; k, l, m, and n are the number of land use types that affect the development of agriculture, forestry, animal husbandry, and fishery industries, respectively.

Second Industry Model
The secondary industry is the industrial sector that processes the products (raw materials) provided by the primary industry and the third industry. It includes mining, manufacturing, electricity, gas, and water production and supply, and construction. Therefore, it is necessary to establish the towns with second industry GDP statistics and land use types based on correlation analysis. The secondary industry GDP spatial distribution model is as follows: where G 2 j denotes the GDP of the secondary industry at town j, g ind j denotes unit area GDP of the secondary industry at town j, A ind ij denotes the area of the ith land use type at town j that affects the development of secondary industry, and n denotes the number of land use types affecting the secondary industry.

Third Industry Model
Based on the same correlation analysis as the above method, the GDP spatial distribution model of the third industry (or service industry) is established as follows: where G 3 j denotes the GDP of the third industry at town j, g ser j denotes unit area GDP of the third industry at town j, A ser ij denotes the area of the ith land use type at town j that affects the development of the third industry, and n denotes the number of land use types affecting the third industry.

GDP Model
The GDP spatial distribution model of the first, second, and third industries were integrated, and the GDP spatial distribution model was obtained as follows: According to the above methods and statistical yearbook, the GDP spatial model of three major industries is calculated, and the spatial distribution of GDP in the three major industries of Poyang Lake Basin is obtained. As shown in the Figure 5, the GDP in Poyang Lake is mainly concentrated around the Poyang Lake, and the GDP in the plain area is higher than that in the mountain areas, which corresponds to the actual situation.

Intensity-Area-Duration Method
This study adopted the IAD method that linked three important features of extreme events: intensity, impact area, and duration [29]. Contiguous grid points with daily precipitation >50 mm over a given time scale and a continuous area were selected as an extreme event. The mean precipitation of the extreme event was selected as the intensity in this method. The IAD method can be used to both study the simultaneous changes in intensity and impact area over a given duration, and analyze the most severe regional extreme precipitation events by plotting an envelope curve. The required steps are as follows [17,20,21].
(1) Determination of the range of extreme events. First, the given time scale is selected as 1 day and ≥2 days and the intensities are calculated for the different time periods for all grid points. The grid point with the highest intensity is regarded as the "center with highest intensity" of a regional extreme precipitation event ( Figure 6a). Second, among the surrounding eight grid points, the one with the second highest intensity is identified to establish the "center with second highest intensity" (Figure 6b). Note that the intensity and coverage of an extreme event in Figure 6b is the mean intensity and amalgamated area of the continuous grids concerned. Third, among the grid points surrounding the "center with second highest intensity," the one with the third highest intensity is identified. All those grids with precipitation greater than the threshold are then determined and combined into a regional extreme precipitation event (Figure 6c,d). Fourth, another "center with highest intensity" is determined and the above steps are repeated, until all the regional extreme precipitation events are accounted for over a given time scale (Figure 6e).
(2) Establishment of the IAD curve. All the points that denoted recorded extremes of intensity and corresponding coverage are linked into a curve to reflect the intensity-coverage relationship. Intensity-coverage curves are constructed for all events within the same given time scale. The points with the highest intensity of the different impacted areas were linked to form an envelope curve, that is, the IAD curve (Figure 6f). The IAD curve reflects the highest intensity that extreme precipitation events could reach over a given time scale for areas with different impact levels.

Intensity-Area-Duration Method
This study adopted the IAD method that linked three important features of extreme events: intensity, impact area, and duration [29]. Contiguous grid points with daily precipitation >50 mm over a given time scale and a continuous area were selected as an extreme event. The mean precipitation of the extreme event was selected as the intensity in this method. The IAD method can be used to both study the simultaneous changes in intensity and impact area over a given duration, and analyze the most severe regional extreme precipitation events by plotting an envelope curve. The required steps are as follows [17,20,21].
(1) Determination of the range of extreme events. First, the given time scale is selected as 1 day and ≥2 days and the intensities are calculated for the different time periods for all grid points. The grid point with the highest intensity is regarded as the "center with highest intensity" of a regional extreme precipitation event (Figure 6a). Second, among the surrounding eight grid points, the one with the second highest intensity is identified to establish the "center with second highest intensity" (Figure 6b). Note that the intensity and coverage of an extreme event in Figure 6b is the mean intensity and amalgamated area of the continuous grids concerned. Third, among the grid points surrounding the "center with second highest intensity," the one with the third highest intensity is identified. All those grids with precipitation greater than the threshold are then determined and combined into a regional extreme precipitation event (Figure 6c,d). Fourth, another "center with highest intensity" is determined and the above steps are repeated, until all the regional extreme precipitation events are accounted for over a given time scale (Figure 6e).

Mann-Kendall Test
The nonparametric Mann-Kendall (MK) test [30,31] is widely used to detect trends in time series of extreme precipitation. The MK test has been widely applied in studies of hydrology, meteorological ecology, and the environment to establish whether time series have abrupt changes [32,33]. The MK statistic (MKs) value represents the tendency and significance of the trend. A value of MKs ≥ 1.96 indicates a significant positive trend and a value of MKs ≤ −1.96 represents a significant negative trend (both at the 95% confidence level).

Changes in Frequency of Extreme Precipitation Events
As shown in Figure 7, 1-day EPEs occur 1966 times in the Poyang Lake Basin during 1961-2014; the highest occurrence is in 1999 (55) and the lowest in 1963 (14). The 1990s (1990-1999) and 2010s (2010-2014) are the decades with the highest frequencies: 42 and 41 times yr −1 , respectively. The decades of the 1960s (1961-1969), 1970s (1970-1979), and 2000s (2000-2009) are similar to each other: 32, 34, and 34 times/year, respectively. In general, the occurrence of extreme precipitation events has increased significantly at a rate of 1.5 times decade −1 (significant at the 95% level). (2) Establishment of the IAD curve. All the points that denoted recorded extremes of intensity and corresponding coverage are linked into a curve to reflect the intensity-coverage relationship. Intensity-coverage curves are constructed for all events within the same given time scale. The points with the highest intensity of the different impacted areas were linked to form an envelope curve, that is, the IAD curve (Figure 6f). The IAD curve reflects the highest intensity that extreme precipitation events could reach over a given time scale for areas with different impact levels.

Mann-Kendall Test
The nonparametric Mann-Kendall (MK) test [30,31] is widely used to detect trends in time series of extreme precipitation. The MK test has been widely applied in studies of hydrology, meteorological ecology, and the environment to establish whether time series have abrupt changes [32,33]. The MK statistic (MKs) value represents the tendency and significance of the trend. A value of MKs ≥ 1.96 indicates a significant positive trend and a value of MKs ≤ −1.96 represents a significant negative trend (both at the 95% confidence level).

Changes in Frequency of Extreme Precipitation Events
As shown in Figure 7 times/year, respectively. In general, the occurrence of extreme precipitation events has increased significantly at a rate of 1.5 times decade −1 (significant at the 95% level). Figure 7, 1-day EPEs occur 1966 times in the Poyang Lake Basin during 1961-2014; the highest occurrence is in 1999 (55) and the lowest in 1963 (14). The 1990s (1990-1999) and 2010s (2010-2014) are the decades with the highest frequencies: 42 and 41 times yr −1 , respectively. The decades of the 1960s (1961-1969), 1970s (1970-1979), and 2000s (2000-2009) are similar to each other: 32, 34, and 34 times/year, respectively. In general, the occurrence of extreme precipitation events has increased significantly at a rate of 1.5 times decade −1 (significant at the 95% level).

Changes in Intensity of Extreme Precipitation Events
During 1961-2014 in the Poyang Lake Basin, the average intensity of 1-day events is 60.9 mm d −1 . Figure 9 illustrates the change in intensity. The 1-day events with the greatest intensity occur in the 1990s with an average intensity 62.0 mm d −1 . The rate of increase in intensity of 0.2 mm decade −1 is not statistically significant.

Changes in Intensity of Extreme Precipitation Events
During 1961-2014 in the Poyang Lake Basin, the average intensity of 1-day events is 60.9 mm d −1 . Figure 9 illustrates the change in intensity. The 1-day events with the greatest intensity occur in the 1990s with an average intensity 62.0 mm d −1 . The rate of increase in intensity of 0.2 mm decade −1 is not statistically significant.

Changes in Intensity of Extreme Precipitation Events
During 1961-2014 in the Poyang Lake Basin, the average intensity of 1-day events is 60.9 mm d −1 . Figure 9 illustrates the change in intensity. The 1-day events with the greatest intensity occur in the 1990s with an average intensity 62.0 mm d −1 . The rate of increase in intensity of 0.2 mm decade −1 is not statistically significant.

Identification of the Most Severe Events
Extreme precipitation has been analyzed previously at station level without consideration of the impact area. The IAD method provides a new perspective for understanding the extreme precipitation events by considering the intensity, duration, and impact area.
Based on the daily precipitation data for the Poyang Lake Basin, 1-day and ≥2-day extreme precipitation events are analyzed. In Figure 12, the envelope comprises the five severest 1-day EPEs (colored dots), and the remaining 1-day extreme precipitation events are plotted below the envelope (colored diamonds). The precipitation event that occurred on 2 May 1994, has a maximum intensity of 208 mm d −1 and it covers 2625 km 2 . The event with the largest impact area occurs on 12 April 2006. It covers an area of 1.47 × 10 5 km 2 and its intensity is around 70 mm d −1 .

Identification of the Most Severe Events
Extreme precipitation has been analyzed previously at station level without consideration of the impact area. The IAD method provides a new perspective for understanding the extreme precipitation events by considering the intensity, duration, and impact area.
Based on the daily precipitation data for the Poyang Lake Basin, 1-day and ≥2-day extreme precipitation events are analyzed. In Figure 12, the envelope comprises the five severest 1-day EPEs (colored dots), and the remaining 1-day extreme precipitation events are plotted below the envelope (colored diamonds). The precipitation event that occurred on 2 May 1994, has a maximum intensity of 208 mm d −1 and it covers 2625 km 2 . The event with the largest impact area occurs on 12 April 2006. It covers an area of 1.47 × 10 5 km 2 and its intensity is around 70 mm d −1 . In Figure 13, the envelope comprises the two most severe ≥2-day extreme precipitation events. The event with the greatest intensity of precipitation (151 mm d −1 ) occurs on 19 June 2010 and it covers an area of 2300 km 2 . The event with the largest impact area occurs on 24 June 2003. It covers an area of 0.68 × 10 5 km 2 and its intensity is around 100 mm d −1 . In Figure 13, the envelope comprises the two most severe ≥2-day extreme precipitation events. The event with the greatest intensity of precipitation (151 mm d −1 ) occurs on 19 June 2010 and it covers an area of 2300 km 2 . The event with the largest impact area occurs on 24 June 2003. It covers an area of 0.68 × 10 5 km 2 and its intensity is around 100 mm d −1 .

Exposure of Population and GDP
Changes in the exposure of the population and GDP to extremes are not simply related to changes of the extreme precipitation events themselves, but they are also dependent on the growth and redistribution of GDP and the population.
The population exposure grows slowly during 1961-2014 at a rate of about 1% yr −1 . From 1984 to 2014, the annual average exposed population is 7.9 million (19% of the entire population of the Poyang Lake Basin), with a rate of increase of 1.37 × 10 5 people yr −1 (passing the 95% significance MK test).

Exposure of Population and GDP
Changes in the exposure of the population and GDP to extremes are not simply related to changes of the extreme precipitation events themselves, but they are also dependent on the growth and redistribution of GDP and the population.
The population exposure grows slowly during 1961-2014 at a rate of about 1% yr −1 . From 1984 to 2014, the annual average exposed population is 7.9 million (19% of the entire population of the Poyang Lake Basin), with a rate of increase of 1.37 × 10 5 people yr −1 (passing the 95% significance MK test). The ascending order of decades based on the annual population exposure is the 1980s   Unlike the population, the GDP has grown rapidly in the Poyang Lake Basin, with a rate of increase of about 10% yr −1 (1984-2014). During 1984-2014, the annual exposure of GDP is 32.7 billion Yuan (7.7% of the total GDP of the Poyang Lake Basin), with a rate of increase of 3.8 billion Yuan yr −1 (passing the 99% significance MK test). The ascending order of decades based on the annual GDP exposure is the 1980s (1.6 billion), the 1990s (11.45 billion), and the 2000s (59.36 billion), accounting for 6.1%, 10.5%, and 8.4% of the total GDP, respectively. The three years with the largest exposure of GDP are 2013 (136.4 billion), 2012 (136.5 billion), and 2010 (155.6 billion). The three years with the largest proportional exposure of GDP are 1999 (15.2%), 1998 (15.8%), and 2010 (16.5%) (Figure 15).

Discussion and Conclusions
Global warming has already affected the extreme precipitation in China and worldwide during the 20th century [34][35][36] and if it continues it may have further impacts. In this paper, based on the daily precipitation data (1961-2014) from 81 climate stations in the Poyang Lake Basin (China), using the IAD method, the frequency, intensity, and impact area of EPEs in the Poyang Lake Basin is analyzed. Based on gridded population and GDP data , the exposure of the population and GDP to EPEs are discussed. Climate change is not the only reason for the worsening losses from disasters. Even if the EPEs do not change, the disaster losses would increase with the development of the economy. In the future study, we will attempt to distinguish the contribution rate of the climate change and the economic development. Overall, the frequency, intensity, and impact area of EPEs and the exposures of the population and GDP have increased. This means that EPEs have an increasingly negative impact on the Poyang Lake Basin. The rapidly developing economy and the changes in extreme precipitation are the two main reasons for the increase in the exposure of GDP and the population. It is difficult but necessary to seek the primary causes of the increasing exposure. In our future research, the relative rates of the contributions of economic development and climate change will be calculated. In addition, future climate change might lead to increases in the frequency of extreme precipitation events and the occurrence of more severe disasters. The potential impact of future climate change on extreme precipitation events and the socioeconomic situation of the Poyang Lake Basin will be

Discussion and Conclusions
Global warming has already affected the extreme precipitation in China and worldwide during the 20th century [34][35][36] and if it continues it may have further impacts. In this paper, based on the daily precipitation data (1961-2014) from 81 climate stations in the Poyang Lake Basin (China), using the IAD method, the frequency, intensity, and impact area of EPEs in the Poyang Lake Basin is analyzed. Based on gridded population and GDP data , the exposure of the population and GDP to EPEs are discussed. Climate change is not the only reason for the worsening losses from disasters. Even if the EPEs do not change, the disaster losses would increase with the development of the economy. In the future study, we will attempt to distinguish the contribution rate of the climate change and the economic development.
From 1961 to 2014, the 1-day EPEs occur 1966 times with average intensity of 60.9 mm d −1 . The ≥2-day events occur 353 times with average intensity of 67.3 mm d −1 . From 1961 to 2014, the annual impacted area of EPEs is established as 7.4 × 10 4 km 2 (45% of the area of the Poyang Lake Basin) with a general upward trend of 400 km 2 yr −1 .The frequencies, intensities, and impact area of EPEs in Poyang Lake Basin are found to have increased.
During 1961-2014, the annual population exposure is 7.90 million people (19% of the entire population on the Poyang Lake Basin), increasing by 1.37 × 10 5 yr −1 . The three years with the largest exposures of population are 1999 (15.68 million), 1998 (16.26 million), and 2010 (16.36 million), accounting for 36.6%, 38.8%, and 37.1% of the total population, respectively. The annual exposure of GDP is 32.7 billion Yuan (7.7% of the entire GDP of the Poyang Lake Basin), increasing by 3. Overall, the frequency, intensity, and impact area of EPEs and the exposures of the population and GDP have increased. This means that EPEs have an increasingly negative impact on the Poyang Lake Basin. The rapidly developing economy and the changes in extreme precipitation are the two main reasons for the increase in the exposure of GDP and the population. It is difficult but necessary to seek the primary causes of the increasing exposure. In our future research, the relative rates of the contributions of economic development and climate change will be calculated. In addition, future climate change might lead to increases in the frequency of extreme precipitation events and the occurrence of more severe disasters. The potential impact of future climate change on extreme precipitation events and the socioeconomic situation of the Poyang Lake Basin will be studied in future research.