Background & Summary

Globally, water has become one of the key issues for humanity to survive and thrive. It is estimated that 61% of Earth’s seasonal changes in surface water storage occur in human-managed reservoirs1. However, the extent of reservoir storage changes remains unknown due to a lack of reservoir-level observations2. The impact of human activities on changes in surface water storage is significant compared to natural variability and has a strong regional dimension3,4. In particular, certain arid and semi-arid regions are increasingly stressed by economic development, growing populations, and climate change1. In the period to 2050, the global water deficit will be about 3241.9 km3/year, with severe water shortages concentrated in arid and semi-arid regions such as the western United States, northern China, and the Middle East5.

Reservoirs contribute significantly to maintaining ecological balance, promoting economic development, and improving people’s lives, and are one of the essential infrastructures of modern society. Reservoirs help to regulate the amount of water in time and space to meet human needs and alleviate water scarcity. Reservoirs significantly impact the assessment of water scarcity, especially in small and fragmented areas6. Access to hydropower can accelerate de-carbonization by reducing the capacity and generation of natural gas plants. At the level of deep de-carbonization, reservoir hydropower can effectively contribute to carbon neutrality7. However, today, reservoirs worldwide are abandoned for various reasons, from poor design and mismanagement to social, economic, and environmental issues, particularly as climate change alters precipitation patterns, affecting reservoir total storage capacity and water resource stability8,9. In the Tarim Basin in northwest China, the reservoir surface area is increasing, but evapotranspiration is also increasing. Extreme weather events such as droughts and floods can also affect the operation of reservoirs, significantly increasing the risk of reservoir failure. The abandonment of reservoirs can be a waste of water resources, an environmental degradation, and an economic waste.

In arid zones, water scarcity severely limits socio-economic development. These areas face high evaporation rates, soil erosion, and fragile ecosystems10, restricting agricultural production, industrial activities, and urban growth11, all of which depend on reliable water supplies. In northwest China, water shortages have been shown to stifle economic advancement and heighten poverty risks, particularly where agriculture is the primary livelihood12. Reservoirs can mitigate seasonal precipitation variability and drought-related water pressures through effective storage and management13.Nonetheless, managing these reservoirs presents unprecedented challenges in the face of global climate change14,15.

The study area, located in China’s arid zone (Fig. 1), encompasses the Xinjiang Uygur Autonomous Region, the Hexi Corridor in Gansu Province, and the Alashan Plateau in western Inner Mongolia. China is home to over 98,000 reservoirs and dams across diverse landscapes16; however, the study area contains only 947 reservoirs, representing less than 1% of the national total. Despite their scarcity, these reservoirs are vital for local economic development17,18.

Fig. 1
figure 1

(a) Overview map of the study area. (b–e) The four largest reservoirs by capacity within the study area, image from Google Earth.

Full size image

China’s water resources are distributed unevenly in both space and time, leading to pronounced supply-demand conflicts. Establishing historical and socio-economic datasets on reservoir construction in China’s arid regions is therefore critical for effective water resource management, climate change adaptation, ecological protection, disaster risk assessment, and sustainable development.

Currently, several reservoir datasets are available for research purposes. These include the Global Reservoir and Dam Database (GRanD), which provides geographical locations, basic attributes, and environmental information for reservoirs and dams worldwide, making it widely applicable in water resource management, ecological studies, and assessments of climate change impacts19; the Global GEoreferenced Database of Dams (GOODD), which focuses on geographical reference information for global dams, encompassing basic characteristics and operational status to create a comprehensive dataset that supports water resource and environmental policy formulation20; the Georeferenced Global Dams And Reservoirs (GeoDAR), which offers geographical information about dams and reservoirs, primarily aimed at visualizing and analyzing the spatial distribution of water resources, thus facilitating studies on reservoir locations and aquatic landscape assessments21; the China Reservoir Dataset (CRD), which specifically collects data on reservoirs in China, including size, type, and distribution, making it vital for understanding water resource and reservoir management in the country22; and the Reservoir Dataset in China (Res-CN), which also concentrates on reservoirs within China, but with a focus on detailed hydrometeorological time series data and landscape characteristics23. While these databases primarily focus on the geographical locations, basic attributes, and environmental impacts of reservoirs and dams, they generally lack detailed historical information regarding the construction of reservoirs and comprehensive data collection related to socio-economic factors, particularly concerning the relationships between construction time series, regional development, and water resource management. This limitation restricts a comprehensive understanding of how reservoirs in arid regions contribute to local economic and ecological sustainability.

This study aims to establish a comprehensive historical and socio-economic dataset on reservoir construction in China’s arid regions, significantly contributing to water resource management, climate change adaptation, ecological protection, disaster risk assessment, and sustainable development. The dataset is compiled from multiple sources, with key information on construction years obtained from water conservancy yearbooks and local histories. Additionally, thorough field research and consultations with local water conservancy departments ensured the accurate collection of reservoir locations and indicators.

This dataset provides a detailed historical record of reservoir construction, covering 669 reservoirs and their socio-economic implications. It serves as a valuable resource for researchers and policymakers focused on water resource management, climate adaptation strategies, and sustainable development initiatives. With a completion rate of 98.53%, the dataset ensures high reliability for various applications.

Given the challenges of reservoir abandonment and the pressures on water sustainability, this dataset offers essential information to enhance reservoir management practices. The integration of historical construction data, socio-economic indicators, and geospatial analysis reflects a multidisciplinary approach that deepens understanding within the field. Researchers and practitioners are encouraged to use this dataset to inform future studies and resource management strategies, addressing water scarcity and promoting sustainable practices in arid and semi-arid regions. Furthermore, it lays a solid foundation for exploring the relationship between reservoir management and regional development.

Methods

Data collection

Reservoirs data

Combines existing data from GRanD24, GOODD25, GeoDAR26, CRD27 and Res-CN28. The Google Earth Engine (GEE) remote sensing platform has been used to collect reservoir location data (longitude, latitude, province, city, county) and reservoir indicators (area, capacity, flow, residence time, and long-term loss estimate). The final number of reservoirs in the dataset was 947 after screening the total reservoir information according to the boundaries of China’s arid zones. Information on each reservoir’s name and construction date was obtained from water conservancy yearbooks and local histories; some missing information was obtained through field research and visits to local water conservancy departments. The final number of reservoirs for which historical information could be accurately determined was 669, representing 70.64% of the total reservoirs, while the number currently operating was 679. A total of 268 dams, mostly unnamed, built many years ago and no longer used, represent 28.29%. This means that 98.53% of currently functioning reservoirs are perfectly informed.

Socio-economic data

Panel data such as population, arable land, number of pumping stations, water-saving irrigation area, hydropower production and output value of agriculture, forestry, animal husbandry and fishery, and Gross Domestic Product(GDP) per capita were obtained from the official websites of the Chinese Ministry of Agriculture29, the State Forestry Administration of China30, the National Bureau of Statistics of China31, the Ministry of Water Resources of China32, the China Meteorological Administration33, and the China Customs34. Spatial data such as population density and land cover were obtained through the interpretation of remote sensing platforms such as GEE and Landsat.

The data used correspond to the years 1990 to 2020, which were collected and analyzed to provide a comprehensive understanding of socio-economic changes over these three decades. This period was chosen due to its significance in China’s economic reforms and water resource management strategies, which have had a substantial impact on agricultural productivity and water usage patterns.

Moreover, certain key statistics, such as the number of pumping stations and the area of water-saving irrigation, are only available within this timeframe. The quality of data during this period is significantly higher compared to previous years, as data collection methods and infrastructure were less developed prior to 1990, leading to unreliable data.

The years after 2020 were excluded primarily due to the impact of the COVID-19 pandemic, which severely disrupted data collection and consistency. The ongoing phase of economic recovery has introduced significant fluctuations in economic conditions and changes in government policies regarding water resource management and agricultural practices.

Data processing

Handling of missing data and outliers

Some values, outliers, and duplicate entries in the integrated reservoir data must be added. These problems mainly included duplicate reservoir information, incorrect or missing names, unrecorded construction time, and latitude, longitude, and reservoir capacity data inaccuracies. To solve these problems, pre-processing and data cleaning were performed on the collected raw data, including checking duplicate entries.

Due to the limitations of earlier statistical techniques and changes in administrative boundaries, some socio-economic data contained missing values. Interpolation was employed to fill in these missing values because it is a statistical method that estimates missing data based on the trends and patterns of existing points, thus helping to maintain overall data integrity. Moreover, interpolation is widely used for filling in missing values in socio-economic data as it effectively enhances data completeness and analytical accuracy35. However, missing data for large areas were considered invalid and excluded from the analysis. The specific data processing steps are shown in Fig. 2.

Fig. 2
figure 2

Data processing steps.

Full size image

Data classification and processing

Reservoirs are classified into five classes according to the “Classification and Design Standards for Hydraulic and Hydroelectric Power Hub Projects (Hill and Hilly Area Part)” (SDJ12-78) in China’s water conservancy engineering. The specific divisions are as follows:

  • Large (I) type reservoirs have over 1 billion cubic metres of capacity.

  • Large (II) type reservoirs have a total capacity of between 100 million and 1 billion cubic metres.

  • Medium (III) type reservoirs have a total capacity of between 0.01 and 0.1 billion cubic metres.

  • Small (I) type reservoirs have a total capacity of between 0.01 and 100 million cubic metres.

  • Small (II) type reservoirs have a total capacity of less than 0.01 million cubic meters.

We assumed theoretical service ranges based on different reservoir levels and simplified reservoirs’ application function by considering only spatial distances (Table 1). Factors such as reservoir use, water capacity and surrounding traffic influence the actual service area of a reservoir. We use reservoir capacity to measure the service area, normalise the service area of reservoirs with different capacities, and create a reservoir service area buffer using a Arc GIS 10.8 tool with the criterion of 100 km served by 10 m3 capacity.

Table 1 Reservoir classification, scope and purpose.
Full size table

A buffer is an area of a certain width (buffer distance) around a geographic element, which is used to analyse the spatial proximity between the geographic element and the surrounding elements. The area is enclosed by a circle with a radius equal to the buffer distance and centred on the centre of gravity of the geographical location of the reservoir. The buffer zone generation algorithm is the circular step-fitting method, which divides the central angle of the circle into equal parts, replaces the circular arc with a chord of equal length, replaces the curve with a straight line, and gradually approximates the circular arc segments with equal line segments of uniform step length. Eventually, the 1 km population density data, the 30 m land cover shape and the raster file containing the geographic information were extracted through the buffer zones of the different levels of the reservoirs to obtain the population and agricultural land data within the service area of each reservoir. The socio-economic data were classified according to China’s administrative divisions in terms of prefecture-level cities, and time series were also added to classify each indicator in time series through the collected historical data of reservoir construction.

Spatial and temporal distribution of reservoir construction

Based on Fig. 3a, we compiled a temporal construction dataset of reservoirs in China with the following characteristics.

  • Orange (reservoirs built before 1949): 291, with 212 unaccounted for and not maintained by local water bodies. 80.76% are small reservoirs (235 in total). These were the first reservoirs erected, and they may have been positioned in vital places where water demands were initially identified, thereby supporting early agricultural and industrial development.

  • Red (reservoirs built 1949–1965): 218; the number of reservoirs expanded dramatically, particularly from 1958 to 1960, when 132 reservoirs were created in three years, accounting for 60.55 percent of the total for that period. This growth is linked to the industrialization and modernization of agriculture following the creation of New China. After the foundation of New China, water conservation became one of the country’s goals, and the building of a number of water conservancy projects was supported to support population and economic growth.

  • Yellow (reservoirs built 1966–1976): 142 Despite China’s social, economic, and political turmoil, the development of water reservoirs remains a national priority. Farmers in rural regions were actively encouraged and inspired to create small reservoirs and canals, mostly for irrigation and to address the issue of water use on farmland, with data indicating that 86.23% of reservoirs built during the era were small to medium in size.

  • Green (Reservoirs built during 1977–2008): 212, reflecting China’s rapid economic expansion. Between 1977 and 2008, China’s reservoir building underwent extraordinary growth and expansion, marking an important milestone in China’s water conservancy modernization process. By 2000, China had more than 85,000 reservoirs, but reservoir construction in arid areas had not improved considerably.

  • Blue reservoirs (constructed after 2008): 84. Reservoir construction in dry zones slowed dramatically during this period due to issues with early design and construction, as well as the instability of reservoir storage and water resources caused by specific climatic circumstances. This phase of water engineering in dry zones focused on reservoir repair, maintenance, and operation.

Fig. 3
figure 3

Reservoir, Population, and Land Cover in Arid Regions (a) Spatial and temporal distribution of reservoir construction time in China’s arid regions (b) Spatial distribution of reservoirs of different levels in China’s arid regions and hotspots of reservoir capacity (c) Spatial distribution of population density in China’s arid regions in 2020 (d) Fig. Spatial distribution of 30 m land cover in China’s arid regions in 2020 Spatial distribution map.

Full size image

The dataset shows that the construction of reservoirs in China’s arid zones can generally be divided into phases of rapid growth, stable growth, and high-quality growth. The dataset documents reservoir construction during three major periods: the initial phase before 1949, the rapid growth phase from 1949 to 1965, the period between 1966 and 1976, the stable growth phase from 1977 to 2008, and the focus on maintenance after 2008.

According to Fig. 3b, we have compiled a dataset of the distribution status of reservoirs by level in the arid zone of China. The dataset includes four large Type I reservoirs: Jilintai First Class Reservoir (capacity 2.53 billion km³), Kalasuk Reservoir (capacity 2.419 billion km³), Chavqi Sea Reservoir (capacity 1.694 billion km³), and Sansheng Gong (capacity 1.231962 billion km³). Jilintai First Class Reservoir and Chavqi Sea Reservoir are geographically located less than 100 km apart, both built next to the Illi River; Kalasuk Reservoir is built next to the Irtysh River (Fig. 1). The dataset also contains 244 medium reservoirs and 441 small Type I reservoirs distributed throughout the region, with small Type II reservoirs more widely dispersed across the study area.According to the reservoir construction time series (Fig. 4), the dataset records the temporal development of reservoirs in the arid region, documenting changes in reservoir number, area, and total storage capacity over time.

Fig. 4
figure 4

Changes in the number, area, and capacity of reservoirs in China’s arid zones in chronological succession. STOR refers to the mean total storage capacity. Area represents the maximum water surface area of the reservoir. Num refers to the total number of reservoirs.

Full size image

Agricultural land and population within the supply area of reservoirs

The dataset includes information on agricultural land and population within the service areas of reservoirs in China’s arid regions. As shown in Figs. 3c and 3d, the spatial distribution data encompasses both population density and land cover information for 2020. The dataset records the relationship between reservoir characteristics and their service areas, as illustrated in Fig. 5 for population served and Fig. 6 for cropland quantity within reservoir service areas.

Fig. 5
figure 5

Relationship between the Population within the Service Area of Reservoirs and Reservoir Level, Area, and Capacity.

Full size image
Fig. 6
figure 6

Relationship of Cropland Quantity within Reservoir Service Area with Reservoir Class, Area, and total storage capacity.

Full size image

Dataset on reservoir capacity and socio-economic change

The study area, in accordance with the different landscapes, which are divided into five zones from west to east: the Tarim Basin, the Junggar Basin, the Tuha Basin, the Hexi Corridor, and the Inner Mongolia Plateau (Table 2), the collected data were divided into five zones to facilitate the analysis of the socio-economic impacts.

Table 2 Meaning of reservoir indicators.
Full size table

The dataset we compiled contains data on the impact of reservoir construction on regional socioeconomic indicators between 2000 and 2020. It includes socioeconomic data on reservoir construction progress, GDP, primary industry output, and GDP per capita, covering several regions, including the Junggar Basin, Tarim Basin, Hexi Corridor, and Inner Mongolia Plateau. According to the data, the development of reservoirs has varied depending on the location. For instance, the Tarim Basin had a surge in reservoir construction between 2007 and 2010, while the Junggar Basin saw a huge number of reservoirs built between 2001 and 2008. The dataset also includes GDP, primary industry output, and GDP per capita fluctuations, as well as the growth rates for reservoir building and capacity. Such data provide valuable information for studying the relationship between reservoir construction and regional economic fluctuations36.

Data Records

Historical Data Set of Reservoir Construction in Arid Regions37 available from the Mendeley Data Repository https://data.mendeley.com/datasets/tytcw3g4w5/1.

Reservoirs buffer vector data

The data is provided in shapefile format and includes information on the spatial location of the reservoir, time of construction, rating, buffer range, and basic information such as name, province, prefecture, county, area, and reservoir capacity.

Reservoirs information

The data is provided in excel (xls) format, which includes information on the construction time of the reservoir, rating, Buffer range of population and cropland area, and basic information such as name, province, prefecture, county, area, reservoir capacity, and so on. Basic Information.

Socioeconomic data

The data is provided in excel (xlsx) format and includes cumulative reservoir capacity and GDP, GDP of the Primary Industry, and Per Capita Regional GDP for calendar years 1990–2020 for each sector and city.

Technical Validation

In the research, reverse geocoding (RGC) was used to determine the specific geographic location of reservoirs in order to ensure the accuracy of historical data on reservoir construction in arid areas. The specific implementation steps are as follows. Using the known geographic coordinates of the reservoir, reverse geocoding is performed to obtain the administrative division information of the reservoir. The geographical information obtained is then matched and verified with the list of reservoirs, reservoir capacity and area, and other key indicators published by the local administrative districts to ensure the accuracy of the location and attribute data. During the data matching process, two approaches were used to supplement and verify data that differed from or were missing from the reservoir information provided by the administrative districts.

  • Telephone Verification. For missing or inconsistent data, firsthand information was obtained by contacting reservoir flood control officials by telephone to obtain additional or corrected data.

  • Email Verification. An inquiry email was sent to the local water authority, requesting official data and relevant information for cross-validation. This method ensures the accuracy and consistency of spatial location and attribute information for each reservoir in the dataset, providing a reliable database for subsequent studies.

To validate the results with real-world conditions, the obtained data was compared against established records and reports from local authorities to assess discrepancies and confirmation of key attributes for each reservoir. This comprehensive verification process fosters a high degree of confidence in the accuracy of the constructed reservoir database.

In a further effort to refine the identification of reservoir construction history, this study incorporates the following methodology:

  • Historical archives and literature review. To obtain detailed information on reservoirs’ design, approval, and construction, historical archives, engineering reports, and water conservation literature from local governments, archives, and water conservation organizations were consulted.

  • Satellite imagery and remote sensing analysis. Observe changes in the reservoir and its surroundings by comparing satellite images taken at different times to deduce how long the reservoir took to build and expand.

  • Local records and historic maps. To confirm the reservoir’s construction records and alteration information, particularly the relevant sections of the regional waterworks, refer to local records and historical maps.

  • Local news and announcements. Collect local historical news reports and government announcements to provide a public record of key points and important events in the construction of the reservoir.

  • Verbal History. Oral history interviews with participants, managers, and neighbors of the reservoir construction to further identify the early reservoir construction.

  • Geographic Information System (GIS) technology. Integrating historical archives, satellite imagery, local literature, and existing data to cross-validate multi-source data on the history of reservoir construction.

Usage Notes

Recommendations for the use of the dataset

This dataset is suitable for water resource management, hydraulic engineering, and environmental impact studies and contains detailed historical information on the construction of reservoirs in arid areas. To ensure a high level of accuracy and to provide a data structure that researchers can easily reuse, the dataset has been validated against multiple sources. The dataset is provided in shapefile format, and a service-wide buffer has been created based on the reservoir class. Moreover, it is also provided in xls and.xlsx formats, which facilitates the researcher’s direct use in a wide range of data analysis and GIS software, such as Excel, R, Python (pandas, pandas), ArcGIS, QGIS, and others. This software can be used for initial exploration, visualization, and further analysis of the data. Pandas are recommended for reading and processing data, and Geopandas are recommended for spatial data analysis.

Data integration and comparison recommendations

Integration with other geospatial data sets: This data set can be combined with other water management data sets, precipitation data, topography data, etc., to perform integrated analyses. Data can be integrated geographically using tools such as Geopandas or ArcGIS.To increase the robustness of the research, researchers are encouraged to cross-check the data with local authority water resource statistics, other academic research data, or historical weather data. Spatial analysis: To identify the distribution pattern and spatial relationship of reservoirs in the arid zone, the geographic coordinates of the reservoirs can be imported into GIS software to perform spatial cluster analysis.