3.1. Spatial Distribution Characteristics of Human–Land–Water
Based on the third water resource survey and regulation in the Yunnan Plateau, combined with precipitation and runoff observation data, the spatial distributions of water resources per capita and per mu water resources of cultivated land at the township level in the province are shown in
Figure 3. It is evident that the water crisis areas with per capita water resources of <500 m
3/people are concentrated in Dianchi Lake, Fuxian Lake, Qilu Lake, and other surrounding areas (
Figure 3a). Areas between the resource-type and the general water shortage-type are mainly distributed south of Kunming, east of Yuxi, north of Honghe, central Chuxiong, east of Dali, and west of Qujing, all located in central Yunnan Province. Other areas in the province that greatly exceed the water balance line of 3000 m
3/people are regarded as relatively rich in water resources.
In terms of water resources per mu of cultivated land (
Figure 3b), the spatial distribution trend is similar to that of the per capita water resources. It indicates that the lake basin area of the central Yunnan Plateau, where cultivated land is concentrated and contiguous, has the lowest water resources per mu. The northwestern and southern regions of the Yunnan Plateau, with sparse population and fragmented cultivated land, have high water resources per capita and high water resources of cultivated land per mu. The water resources in the central Yunnan Plateau are inherently restricted owing to the agglomeration effect of population and industry in the Central Yunnan Plateau Economic Zone (Kunming as the center), together with the natural endowment of water resources and cultivated land. The contradiction between water demand and supply within the area has existed for a long time, which has resulted in the area having the highest density of water network projects.
Correlation analysis was conducted on the spatial change trends of indicators such as population, cultivated land, water resources, and storage capacity in various elevation areas of the Yunnan Plateau, as shown in
Figure 4. Population increases with expansion of cultivated land (
Figure 4a), with a correlation coefficient as high as 0.90 (significance level,
p < 0.01), indicating that cultivated land is a vital and fundamental resource for human survival. The regulating storage capacity index, which reflects the human transformation of nature and abundant–dry changes in river runoff, has an approximately linear relationship with population growth (
Figure 4b), with a correlation coefficient of 0.86 (significance level,
p < 0.01). The main reason is that humans need to build reservoirs to effectively utilize water resources. The larger the population, the greater the required regulating storage capacity. The spatial relationship between cultivated land and water resources also shows a linear growth trend (
Figure 4c), with a correlation coefficient of 0.81 (significance level,
p < 0.01), indicating that the capacity constraints of regional water resources are prerequisite for expansion of cultivated land, while cultivated land resources have become the basic condition for human survival and development.
3.2. Spatial Distribution Features of Precipitation
The Yunnan Plateau is a typical plateau mountainous area, for which runoff mainly relies on precipitation. Hence, the spatiotemporal patterns of variation in precipitation are broadly consistent with the variation characteristics of runoff and regional water resources. The third Chinese water resource survey revealed that water resources were severely depleted during 2009–2012 owing to four years of severe drought in southwestern China. Analysis of water resources from 1956 to 2015 revealed that the total amount of water resources in the province was 214.1 billion m3, which was 3% less than the average value from 1956 to 2000. However, owing to the effects of topography, monsoon sources, and the local environment, the change in water resources varies across different regions.
The long-term trend was analyzed using observed precipitation data from the 1950s to 2022 from eight meteorological stations in Kunming, Zhaotong, Mengzi, Dali, Jinghong, Shangri-La, Yuanjiang, and Yuanmou, obtained since their establishment. Precipitation in Kunming, the provincial capital, showed a slight trend of reduction overall, but with notable annual variation. In the past 20 years, two periods of reduced precipitation with one intervening period of increased precipitation were observed, i.e., 1999–2012, 2018–2022, and 2013–2017, respectively. These characteristics of periodic continuous abundant precipitation or dry periods are prominent. In the remaining areas, annual precipitation also showed a slight trend of reduction in Mengzi and Shangri-La, whereas annual precipitation declined notably in Dali, Zhaotong, and Jinghong.
Further analysis of the Gini coefficients of the monthly precipitation distribution balance within a year revealed that the average Gini coefficient of the eight stations is 0.48–0.56 and that the average asymmetry coefficient (
S) is 0.85–0.94, as shown in
Figure 5. Generally, the Gini coefficient of each station presents a trend of decline. The asymmetry coefficients show a trend of increase in Kunming, Zhaotong, Dali, Yuanjiang, and Yuanmou, whereas they show a gradual reduction for the other three stations. It shows that the balance of precipitation distribution within a year gradually increases and that the asymmetry weakens. However, the interannual changes are large, and the abundant–dry difference of rainfall affected by the monsoon climate does not change, resulting in increase or decrease in runoff. Therefore, it is necessary to rely on water storage projects such as lakes and artificial reservoirs to reduce the abundant–dry difference in natural water to satisfy the demands for water associated with daily life and industrial/agricultural production.
The spatial changes in the mean values of the Gini coefficient and the asymmetry coefficient were further analyzed, referring to the monthly precipitation observational data of 70 meteorological stations at various latitudes and longitudes and elevations in the Yunnan Plateau from the 1950s to 2022, as shown in
Figure 6. Reflecting the different sources of water vapor for precipitation on the Yunnan Plateau, the data were divided on the basis of two warm and humid airflows: the southeast monsoon in the North Pacific Ocean and the southwest monsoon in the Bay of Bengal in the Indian Ocean.
Figure 6a shows that irrespective of whether the stations are in the southeast monsoon area or the southwest monsoon area, the Gini coefficient of monthly precipitation for each meteorological station increases with increasing latitude. The correlation coefficients are as high as 0.849 (significance level,
p < 0.01) in the southeast monsoon region and 0.715 (significance level,
p < 0.01) in the southwest monsoon region. Similarly (
Figure 6b), the Gini coefficients of monthly precipitation in the southeast monsoon area and the southwest monsoon area increase with increasing elevation; however, the correlation coefficient is 0.432 (significance level,
p < 0.01), which is less than that associated with the change in latitude. The underlying reason is the geomorphic features of the Yunnan Plateau. The terrain of the Yunnan Plateau gradually declines from the northwest to the southeast. Thus, the southeast and southwest warm and humid airflows flow northward and westward along the southwestern and southeastern valleys, in an area broadly divided by the Yuanjiang-Honghe River and the Ailao Mountains. To the east of the boundary lies the source of the warm and humid airflow current in the North Pacific Ocean, and to the west is the water vapor source in the Bay of Bengal in the Indian Ocean. From the south to the north, the transport distance of both major water vapor sources is lengthened with increasing latitude, and thus the water vapor content available to produce precipitation gradually diminishes [
35]. Not only is the total annual precipitation gradually declining, but the uniformity of the water vapor distribution within a year is also getting worse, which leads to the gradual annual increase in the Gini coefficient of precipitation. The uplift of local terrain at the various meteorological stations can promote the formation of precipitation, i.e., the precipitation within a certain range (such as the windward slope and leeward slope of mountains) increases with the increase in elevation [
29]. It weakens the imbalance caused by the gradual weakening of water vapor with elevation during the long-distance transport process, which means the Gini coefficient of the overall monthly precipitation during a year on the Yunnan Plateau is not as closely related to the increase in elevation as it is to the increase in latitude. It is also evident from
Figure 6a,b that at the same latitude or elevation, the Gini coefficient of precipitation change in the southwest monsoon area is higher than that in the southeast monsoon area. Because the Yunnan Plateau is at the end of the range of influence of the East Asian monsoon in summer, it can be attributed to the fact that the source of the rainfall-producing water vapor is weaker than that of the southwest monsoon region of the Indian Ocean, and, accordingly, the magnitude of its change is much smaller. As show in
Figure 6c, irrespective of whether the stations are in the southeast monsoon area or the southwest monsoon area, the Gini coefficient of monthly precipitation shows no clear pattern with change in longitude, indicating that the intra-annual variation in precipitation is not significantly influenced by longitude, which is consistent with the research results of Liu et al. [
36].
Particularly in the areas of Gongshan, Weixi, and Fugong in the northwest of the Yunnan Plateau, owing to the transport of Indian Ocean water vapor from the Enmeikai River to the northeast and “the rainfall in spring” [
37] flood season during March–April annually, two peaks in precipitation are evident during the year. This feature markedly slows the phenomenon of precipitation depletion in the dry season. The Gini coefficients of the annual monthly precipitation in these places are 21–42% lower than those in other places at the same latitude.
Deqin is the area with the highest elevation and latitude in the northwest of the Yunnan Plateau. The water vapor source of precipitation, whether it be the southeast monsoon or the southwest monsoon, is at the furthest transport distance. The precipitation in this area is only equivalent to that of dry and hot valleys such as the Yangtze River valley, Yuanjiang-Honghe River valley, and other rainless areas. Affected by the Qinghai–Tibet Plateau and the Himalayas, the Gini coefficient of monthly precipitation in the Deqin area is only 0.484.
Other than that derived from water vapor associated with the East Asian monsoon, precipitation in the areas of Weixin, Zhenxiong, and Yanjin in the northeast of the Yunnan Plateau is also caused by the uplifting of evaporation owing to high temperatures and to increased water vapor in the northern Sichuan Basin moving along the Wumeng Mountains. This represents a greater source of precipitation than that in other northern areas of Yunnan Province, and it greatly promotes even distribution of precipitation throughout the year. The Gini coefficients of monthly precipitation are also 11–28% lower than those in other places at the same latitude.
As shown in
Figure 6d,e, for different monsoon climate regions, the variations in the monthly precipitation asymmetry coefficient with elevation and latitude are chaotic, indicating irregularity. However, it generally increases from the west to the east with increasing longitude (
Figure 6f), where the correlation coefficient reaches 0.541 (significance level,
p < 0.01), i.e., the asymmetry of the annual distribution of precipitation in the west of the Yunnan Plateau is lower than that in the eastern and central parts. The main reason is that the water vapor source on the Yunnan Plateau in the late rainy season every year follows the basic pattern of that of the East Asian monsoon, i.e., gradual weakening and retreat, followed by the advance of the southwest monsoon. Thus, it affects the trend of the asymmetry of the annual distribution of precipitation in relation to longitude.
3.3. Spatial Distribution Features of Interconnected Lake–River Water Systems
Based on the first national water conservancy survey and the third water resource survey, the actual water supply volume of the interconnected water system in different elevation regions was studied. It should be noted that the water supply volume in this study refers to the actual supply volume of water storage projects with a certain spatiotemporal regulation ability (including large and middle-sized reservoirs, plateau lakes, and comprehensive utilization projects of large and middle-sized hydropower stations) and the interconnected lake–river water systems. The water supply volume is subdivided into five categories for statistical survey: urban, rural, industrial, agricultural irrigation, and ecological. The actual water supply volume of the interconnected water system in different elevation regions is shown in
Figure 7a. The water supply of projects between 1200 and 2200 m accounts for 82.9% of the total water supply of all projects. Among them, the water supply of interconnected water systems in the range of 1800–2000 m is largest, accounting for approximately 25.3% of the total water supply of the project, followed by the range of 1600–1800 m, which accounts for approximately 20.1% of the total.
The elevation distribution of water supply of projects is broadly consistent with that of the population in the Yunnan Plateau. In terms of water use, the trends of urban, industrial, and agricultural water supply in different elevation ranges are similar to the trend of the total water supply, which is mainly distributed in the range of 1800–2200 m. Many important cities and towns are located in the two elevation ranges, as are the core economic zones of central Yunnan Province, including Kunming, Qujing, Yuxi, Chuxiong, Dali, Zhaotong, Baoshan, Lincang, and other prefectures (cities). In particular, industrial water use accounts for a large proportion in the elevation range of 1200–1400 m, which encompasses the locations of important smelting and mining industrial parks such as Honghe and Wenshan in southeastern Yunnan Province. The abovementioned elevation range is also the main water-receiving area of the water diversion project of central Yunnan Province. The rural water supply varies little in the different elevation ranges, which shows that the rural population lives in scattered areas owing to the different living habits and farming practices of the various ethnic groups that live in Yunnan Province. Mountain ethnic groups such as the Miao, Tibetan, and Dulong, mid-mountain ethnic groups such as the Yi and Zhuang, and ethnic groups in the dam area such as the Dai, Hui, and Bai exhibit little difference in terms of distribution at various elevations. The rural water supply in Yunnan Province is mainly from springs or from mountain streams near villages. The ecological water supply, available only within the elevation range of 1600–1800 m, is mainly associated with the Niulan River–Dianchi Lake water diversion project, the Fuxian Lake diversion, and the Xingyun Lake ecological water diversion project.
The regulating storage capacity and water supply of lakes and reservoirs in each elevation range and the corresponding population and cultivated land resources were further analyzed. The changes in the water supply and regulating storage capacity per mu of cultivated land and per capita with elevation difference are shown in
Figure 7b,c. To determine the value of regulating storage capacity, the regulating storage capacity is used for the reservoir. Lake volume between the highest and lowest operating control water levels is taken for plateau lakes. The storage for comprehensive utilization projects of large and middle-sized hydropower stations is only calculated according to their tasks of supplying water to surrounding industrial and agricultural production. Since 2009, the Yunnan Plateau, especially in central and southeastern areas, has suffered multiple winter–spring–early summer droughts [
38]. With ecological and environmental problems such as declining water quality and falling water levels of regional lakes, Fuxian Lake, Yilong Lake, Xingyun Lake, Chenghai Lake, and other lakes have stopped providing useful water to the surrounding areas. Instead, long-distance ecological water projects are planned to replenish the water in these lakes or replace their water supply tasks. It is evident from
Figure 7b that the regulating storage capacity per capita and per mu of cultivated land decreases from 2716 and 1713 m
3, respectively, in the hot and dry valley areas below 800 m, to 128–332 and 46–175 m
3, respectively, in areas above the elevation of 1000 m. The per capita water supply (
Figure 7c) also drops from 723 to 61–170 m
3 and reaches 236–308 m
3 in three elevation zones: 1200–1400, 1600–1800, and 2400–2600 m. The trend of the water supply per mu of cultivated land (
Figure 7c) is similar to that of the regulating storage capacity, which rapidly drops from 456 to 50–120 m
3 in the dry and hot valley areas below the elevation of 800 m. In alpine mountainous areas above the elevation of 2600 m, only cold-resistant, low-yielding dry crops can be grown, and the average water supply per mu is only 20 m
3. These results quantitatively reflect the topographic water shortage on the Yunnan Plateau, with the characteristics of water resources being in lower regions, while people and cultivated land are in higher regions.
3.4. Spatial Distribution Characteristics of Water Consumers
Among all types of water consumers, the average water transmission distance between the water source and the water consumer is sorted as follows (from shortest to longest): rural < irrigation < industrial < urban. It indicates that the water diversion distance to the water consumer is directly proportional to the affordability of water supply costs by consumers. The relative difficulty of water supply is measured by the water transmission distance required to obtain a water supply (unit: per million m
3), as shown in
Figure 8c. It is evident that the average water transmission distance for a rural water supply is 5.89 km/million m
3, while the absolute distance is only 2–8 km. It indicates that rural water supply is affected by factors such as water supply costs, and that it is typically supplied from the nearest water supply source, which is the most expensive in terms of transportation. The average water delivery distance is 2.02 km/million m
3 for industrial water supply (
Figure 8c), and the absolute value is between 4 and 16 km. The industrial water supply is relatively large within the elevation range of 1200–2200 m (
Figure 8a), which means that the water supply distance is longer. The locations of industries are mostly based on mineral resources, topography, transportation, and other factors, whereas water source conditions are not considered as limiting factors. Therefore, in addition to adjusting and utilizing water supplies such as nearby reservoirs and lakes, many industrial water supplies require long distances for self-provided water. Water supply by pumping projects account for a large proportion of industrial water use. The average water transmission distance for the urban water supply is 1.80 km/million m
3 (
Figure 8c); however, the absolute distance is generally longer, i.e., 6–23 km. This is attributable both to the large demand for urban water usage and to the difficulty of finding suitable water sources. It is also related to the strong water price affordability of the urban water supply. Especially in the elevation range of 2200–2400 m, the average water transmission distance is 23 km in terms of the urban water supply. Two of the targets of the urban and irrigation water supply are Ninglang and Yongsheng Counties, with a corresponding large reservoir (Guanquaping Reservoir) under construction, for which the water transmission distance is up to 52 km. In terms of agricultural irrigation, the average water transmission distance is 0.70 km/million m
3 (
Figure 8c), and the absolute value is 5–11 km. From the perspective of individual projects, few projects have agricultural irrigation water transmission distances of <1 or >10 km. This is because the water supply for agricultural irrigation is relatively concentrated, i.e., it supports irrigation in an area of concentrated cultivated land where the water demand is relatively large. Finding a water source project that can meet the water demand while comprehensively considering factors such as hydrology, geology, and the ecological environment is difficult; thus, the location of its construction is generally not close to the irrigation area. Although the construction of water source projects with agricultural irrigation tasks has national investment subsidies, the cost affordability of agricultural irrigation water supply is still not high, resulting in few long-distance agricultural irrigation projects. Currently, China’s fee revenue for agricultural water accounts for only approximately one third of the production cost of the water supply. In most areas, the price level is lower than the operation and maintenance costs of the irrigation facilities. The inability and unwillingness of water users to pay for a water supply have become constraints in the implementation of policies related to the reform of agricultural water pricing [
39]. The average distance for ecological water replenishment is only 0.296 km/million m
3, but the absolute water diversion distance is 17–116 km.
From
Figure 8b, it can be observed that there are differences in the difficulty of water supply to water consumers in different elevation ranges. In the elevation ranges of 2000–2400 m, 2200–2400 m, and 1600–1800 m, the difficulty of water supply ranks from difficult to easy in the following order: rural > industrial > urban > irrigation > ecological. In other elevation ranges, except for the elevations of 800–100 m and >2600 m, in most areas, the difficulty of water supply is sorted as follows (from difficult to easy): rural > urban, industrial > irrigation > ecological. Only the ranking of urban and industrial sectors differs.
As the difficulty of implementing new water source projects gradually increases, and the affordability of water supply costs gradually increases, it is expected that the water transmission distance to each water supply object will also show a trend of increase in the future. The Yunnan Plateau is dominated by large-scale agriculture; however, the steep mountains and canyons pose huge obstacles to the livelihoods and productivity of the population. Particularly for those cities and towns that are developing secondary and tertiary industries, topography has become the primary factor of focus. The nine plateau lakes and the surrounding areas, together with the valleys and basins, have become key areas for the development of cities and industrial parks. However, problems such as water shortages, low per capita arable land, deterioration of river and lake water environments, and the fragile water ecology have become increasingly prominent, resulting in changes in water resource allocation plans for the development of cities and industrial parks and the water supply guarantee measured from adjacent drainage to long-distance cross-drainage water transport. For example, the planned and constructed Central Yunnan Water Diversion Project [
40] is the core axis of the regional water network on the Yunnan Plateau, connecting the main stream of the Yangtze River with 36 counties (cities, districts) in six prefectures (cities) including Dali, Lijiang, Chuxiong, Kunming, Yuxi, and Honghe. Four major water systems, i.e., the Yangtze River River, Lancang River, Yuanjiang-Honghe River, and Pearl River, are connected by the Central Yunnan Water Diversion Project through the interconnection of 108 reservoirs and 111 water plants with the main canal. This system mainly ensures the safety of the water supply for cities, industries, key irrigation areas, and plateau lakes. Large reservoirs such as Deze, Songhuaba, Yunlong, Qingshanzui, Chaishitan, Chemabi, Heitanhe, Haishao, Xiaoshimen, and other large reservoirs represent the primary nodes of the water network in central Yunnan Province, forming connections between the main streams and the tributaries. The network connects rivers with counties and towns, realizes the regulation of abundance and dryness, supports regional mutual assistance, improves the river water ecological environment, and integrates the Yangtze River Economic Belt, the Western Economic Belt, and the national Double T-shaped water network into an organic entity [
13].
According to the distribution of differences in water delivery elevation, the changing patterns of water supply volume of various water consumers were classified statistically, as shown in
Figure 9a. It is evident from
Figure 8b that, irrespective of whether lifting or natural flow water supply, the water supply for agricultural irrigation, rural daily life, and industrial production based on the built interconnected water system is mainly concentrated within the elevation difference of 100 m, accounting for 67.5–72.2% of the total industrial water supply volume. The urban and ecological water supply is relatively concentrated in several elevation ranges, e.g., <50, 150–200, 200–250, and >300 m. Because of long-term water shortages in urban areas and the vulnerability of the water ecology of plateau lakes such as Dianchi Lake, the population is forced to bear greater water supply costs to provide water for urban life and ecological restoration of lakes. From the perspective of water supply objects, the elevation difference of the rural and agricultural irrigation water supply is generally smaller than that of the urban and industrial water supply. This can be attributed to the water transmission distance, because the distance between the objects of rural and agricultural irrigation water supply and the water source is relatively small, and the elevation difference is also relatively small. The water supply efficiency for industrial/agricultural production, urban, and rural areas is at its maximum value for the elevation differences of <50 and 50–100 m, except for the abnormal phenomenon of the ecological water replenishment affected by the Niulan River–Dianchi Lake water diversion project, which has a much larger water supply than other similar projects. Overall, the water supply efficiency gradually diminishes as the elevation difference of water delivery increases. For example, the representative agricultural irrigation water supply accounts for 56.0% of the total water supply, and the water supply efficiency (
Figure 8c) gradually reduces from 1.97 million m
3/km to 0.66–0.96 million m
3/km. Overall, the average water supply efficiency of the interconnected water systems in the Yunnan Plateau is 0.83 million m
3/km, although it varies among the various industries (as shown in
Figure 9b). The ecological water supply efficiency is highest, i.e., 3.38 million m
3/km, followed by that of agricultural irrigation at 1.43 million m
3/km. It is broadly the same for the industrial and urban water supply, i.e., 0.5–0.55 million m
3/km. Rural areas have the lowest water supply efficiency, i.e., only 170,000 m
3/km (
Figure 9c).