Abstract

Water availability plays a key role in securing agricultural production and sustaining the income of farming households. Nepal is one of the countries most dependent on agriculture; more than 80% of the population works in agriculture, contributing to 35% of its total gross domestic product (GDP). As one of the longest rivers in Nepal, Koshi River is one of the main water supplies for agricultural activities. In recent years, due to the population growth and the climate change, there has been increasing stress on the water resources in Koshi River basin. Therefore, a comprehensive investigation of water availability in the basin area is required, prior to an effective strategy for water resources allocation and management. In this study, we provide a quantitative assessment of available water resources in Koshi River basin and highlight the trend of water availability for agricultural use. Moreover, we discuss the potential water-related risks for farming households in the basin area. The contribution of this study is to provide the basis for efficient water management strategies in Koshi River basin.

Introduction

The management and conservation of water resources are essential to maintain the balance of the hydrological cycle, ensuring water availability for future generations. Understanding the quantity, quality, and accessibility of freshwater is fundamental for water risk management and sustainable water use, which would further benefit economic growth and social stability (Scarascia et al., 2006). Water resources play a vital role in human life and agriculture irrigation, especially for agriculture-reliant countries and regions (Liu et al., 2017). Irrigated farmland, accounting for less than 18% of total arable land, contributes to approximately 40% of the world's foodstuffs, which also has a significant role in meeting the projected world food demand (Postel, 1999). Worldwide, over 70% of the freshwater is used for irrigation and the rest is for meeting other water demands (UNWWDR, 2003; Ayars et al., 2006).

Due to climate change, water scarcity has become an increasing problem throughout the world (Schütze & Wagner, 2016). More and more regions where the demands have outstripped the water supply have suffered from severe or chronic water shortages. In particular, the water-related issues have caused more serious damage to some of the agriculture-reliant countries where more water is required to support the rapidly growing population and the booming agriculture-oriented economy (Li & Huang, 2011). The increasing water demand exceeds the water availability, resulting in a growing severity of water stress. The water stress leads to the overexploitation of existing water resources, deterioration of the water availability and water quality, and further degrades soil quality. In addition, water stress could also escalate conflicts and tensions within social interaction. For agriculture-reliant regions, it is expected that the shortage of water resources would cause even more severe consequences, and the degradation of water availability would introduce a greater negative socioeconomic impact on the farming population. Thus, it is extremely necessary to assess and evaluate the water availability in the regions most dependent on agriculture in order to support an effective water management (Mizyed, 2009).

The availability of water resources for agricultural production has been widely discussed around the world. Dinar (1994) studied energy costs and water resources, in particular, groundwater availability and quality for agriculture in California. In his study (Dinar, 1994), the proposed model considered a dynamic relationship among water level, aquifer quality, water management, and water-related technology to maximize long-run private profits. For regional agricultural water management, Rosenzweig et al. (2004) developed a comparative framework that integrates water availability, agriculture production, and irrigation technology with demographics, economics, and climate conditions. They estimated the future water availabilities while considering the capability of ecosystem services under changing climate conditions, and compared their results in many regions throughout the world, including northern Argentina, southeastern Brazil, northeastern China, the Hungarian and Romanian parts of the Danube basin, and the US cornbelt (Rosenzweig et al., 2004). The International Programme for Technology and Research in Irrigation and Drainage (IPTRID) and the Italian National Committee of the International Commission on Irrigation and Drainage (ICID) combined their efforts in documenting water resources in Italy with special attention devoted to the availability of agricultural water (Scarascia et al., 2006). Almas & Scholz (2006) emphasized the technical issues among agricultural activities in relation to the water sectors of Yemen and proposed a water resource conservation strategy which might lead to a more sustainable social structure and more effective agricultural policies in order to solve the water-related food crisis in Yemen. Their study pointed out that both surface water and groundwater in the Yemen area have been exploited beyond the level of recharging, and 93% of the freshwater is used for traditional agricultural production which does not apply sustainable irrigation techniques (Almas & Scholz, 2006). Setegn et al. (2011) used a physical model based on the Soil Water Assessment Tool (SWAT) to assess the impact of climate change on water availability in the Lake Tana basin, Ethiopia, where the local economy mainly relies on agriculture. A dramatic change in regional hydrologic cycles as well as in local water availability was predicted in their study (Setegn et al., 2011). Amisigo et al. (2015) assessed the impact of projected climate change on water availability and crop production in the Volta Basin and the southwestern and coastal basin systems of Ghana, and they found that not all water demands (municipal, hydropower, and agriculture) met the water supply. They suggested adopting groundwater as an additional source of water supply and drawing up an effective water resources management plan in the catchments to balance demand and supply, in order to ensure a sustainable socioeconomic development (Amisigo et al., 2015). For supporting water management, Wan et al. (2016) investigated the trend of water availability over the past five decades and analyzed the spatial distribution of drought occurrences in these regions based on the time series of regional water datasets. Moreover, they assessed the decadal variability of agricultural production in response to water stress in the karst regions of South China. Their results indicated that the study area was facing great challenges in allocating water resources against drought (Wan et al., 2016). Throughout the above-mentioned studies, it can be found that climate change, population growth, and economic development are the key factors that affect the availability of agricultural water.

A least developed country (LDC) is characterized by the facts of endemic poverty, human resources weaknesses, and economic vulnerability, and those facts lead to the instability of its agricultural production. The current LDCs generally rely predominantly on the agrarian economy that grows in an unhealthy environment with underdeveloped irrigation strategies, inconsistent agricultural policies, poor environmental stewardship, insufficient investments, non-inclusive and non-participatory governance, vulnerability to climate change, and weak research capacity regarding technical challenges (Musa, 2009). Due to the uniqueness of geography and climate conditions, as well as low household incomes and the pattern of industrialization, the economy of LDCs is sensitive to changes regarding the temperature pattern and precipitation condition, and vulnerable to the extreme weather events (Sovacool et al., 2012a, 2012b). LDCs are particularly susceptible to water stress due to the underdeveloped agricultural activities as well as the socioeconomic structure (Miyan, 2015).

As one of the most water-abundant countries, Nepal has plenty of water resources. However, the water in Nepal is unevenly distributed in terms of both space and time. Spatially imbalanced and poorly managed water resources characterize the socioeconomic development in Nepal (Duncan et al., 2013). Concerning the social structure, 83% of Nepalians live in rural societies located in remote areas (Rautanen et al., 2014). Small-scale (average 0.7 ha) subsistence farming is the mainstay of Nepal's economy, which contributes to nearly 36% of Nepal's GDP and employs 78% of workers (Palazzoli et al., 2015). In Nepal, as much as 76% of total arable land is not properly irrigated, and the crop productivity is significantly lower than other South Asia countries. Nepal suffers from food security issues since it relies heavily on imports from India (Suhardiman et al., 2015). Over the past decade, national population expansion and global climate change have increased the pressure on water resources in Nepal. On the other hand, water management is still underdeveloped. As a consequence, several medium-size river basins in Nepal have deteriorated regarding water availability (Pandey et al., 2012), and therefore, an effective water management strategy is urgently required.

The Koshi River basin is a transboundary basin; it crosses China, Nepal, and India. Originating from glaciers and snow-covered regions of China and the alpine regions of Nepal, the area of Koshi River watershed is 69,000 km2, of which, Nepal's territory is 30,700 km2 and accounts for 21% of Nepal's total land area. Koshi River has a large annual average discharge, but there is a significant seasonal variability from flooding to extremely dry due to the temporal pattern of the local precipitation. This creates a challenge to the agricultural activities within the Koshi River basin. To alleviate the seasonal water imbalance, it is necessary to assess the water availability as well as the water demand (both agricultural and non-agricultural use) over the basin area, which is fundamental for the development of water management facilities that can store water during surplus periods and supply for the dry (Chinnasamy et al., 2015). An effective water management system could not only benefit the current productivity but also help to solve the potential challenge introduced by the growing population and economic activities (Jia & Jiang, 2000). The aim of this paper is (1) to give a quantitative assessment of available water resources in Koshi River basin, (2) to clarify the trend of water availability for agricultural use, and (3) to assess the potential water-related risks for farming households.

Data and methods

Data

In this study, the investigation of water availability for agriculture was mainly based on meteorological, hydrological, and socioeconomic data. The meteorological data consist of monthly average precipitation and temperature from 1981 to 2012, acquired from ten weather stations. In terms of the hydrological dataset, river runoff was measured at three stations (Jalbire in upstream, Barhabise in midstream, and Dalalghat) from 1988 to 2012, and the evapotranspiration rate over the same period was mapped by remote sensing technique. To fuse point datasets with the map of evapotranspiration rate, meteorological and river runoff data have been interpolated. The socioeconomic indices are extracted from the Nepal Social-Economic Report 2011–2013, including food supply, agricultural product yield, animal husbandry yield, irrigation area, property rate, agricultural water facility, land-use development, and so forth (Table 1).

Table 1.

Analysis index system of agricultural water resources availability.

Classification First-level indices Second-level indices Weights Definition and assignment 
Domestic water (H) Population size (0.31) Number of rural settlements (H1) 0.302 Rural households (HH) 
Population density (H2) 0.525 Population density (P/km2
Population growth rate (H3) 0.173 Population growth rate (%) 
Socioeconomic status (0.079) Poverty rate (H4) 0.418 The proportion of the poor (%) 
Social development ranking (H5) 0.262 According to the regional social development level, the rank of prosperity is divided into 1–5 levels. The higher the level of social development, the higher the level (dimensionless quantity) 
Wealth gap index (H6) 0.321 The Gini index is used to measure the income gap of residents in a district. Its value is between 0 and 1. The closer the Gini index approaches 1, the more unequal the income distribution (%) 
Drinking water conditions (0.104) Water supply coverage (H7) 0.271 The ratio of the number of household water supply to the total number of households (%) 
Sanitation coverage (H8) 0.729 The ratio of the number of health facilities to the total number of households (%) 
Water facilities (0.507) Tap/Piped drinking water (H9) 0.154 Number of households using tap water/pipelines (HH) 
Tubewell/Hand pump of drinking water (H10) 0.381 Number of households using hand pump/tube wells (HH) 
Covered well of drinking water (H11) 0.128 Number of households using covered wells (HH) 
Uncovered well of drinking water (H12) 0.202 Number of farmers who depend on pond water (HH) 
River/Stream of drinking water (H13) 0.134 Number of farmers who depend on rivers/streams (HH) 
Production water (C) Land use development (0.475) Cultivated land reclamation rate (C1) 0.284 The proportion of cultivated land to county area (%) 
Utilization of paddy fields (C2) 0.114 Ratio of paddy field area to cultivated land area (%) 
Farming ratio (C3) 0.176 Paddy field area and dryland area ratio (%) 
Irrigation rate (C4) 0.346 The ratio of actual irrigated area to cultivated area (%) 
Land leasing rate (C5) 0.153 Percentage of tenants and landowners (%) 
Crop yield (0.397) Food crop production (C6) 0.304 Bean production, including rice, corn, millet, wheat, barley (Mt) 
Economic crop production (C7) 0.137 Including oilseeds, potatoes, tobacco, sugar cane production (Mt) 
Meat production (C8) 0.107 Including meat, eggs, wool, fish production (Mt) 
Livestock number (C9) 0.140 The sum of the quantities of livestock. Cattle = 1.0; buffalo = 0.8; sheep = 0.6; chicken, duck = 0.2; others = 0.1 (in number) 
The yield ratio of crop species (C10) 0.030 Ratio of food crops to economic crops (%) 
Food crop area ratio (C11) 0.128 The proportion of food crops to the total area of food crops and cash crops (%) 
Economic crop area ratio (C12) 0.613 The proportion of cash crops to the total area of food crops and cash crops (%) 
Water resources (0.128) Per capita water resources (C13) 0.221 The ratio of total water resources to population in a district (m3/P) 
Perennial average runoff (C14) 0.284 District annual runoff for many years (m3/s) 
Perennial average precipitation (C15) 0.167 District annual average annual rainfall (mm) 
Classification First-level indices Second-level indices Weights Definition and assignment 
Domestic water (H) Population size (0.31) Number of rural settlements (H1) 0.302 Rural households (HH) 
Population density (H2) 0.525 Population density (P/km2
Population growth rate (H3) 0.173 Population growth rate (%) 
Socioeconomic status (0.079) Poverty rate (H4) 0.418 The proportion of the poor (%) 
Social development ranking (H5) 0.262 According to the regional social development level, the rank of prosperity is divided into 1–5 levels. The higher the level of social development, the higher the level (dimensionless quantity) 
Wealth gap index (H6) 0.321 The Gini index is used to measure the income gap of residents in a district. Its value is between 0 and 1. The closer the Gini index approaches 1, the more unequal the income distribution (%) 
Drinking water conditions (0.104) Water supply coverage (H7) 0.271 The ratio of the number of household water supply to the total number of households (%) 
Sanitation coverage (H8) 0.729 The ratio of the number of health facilities to the total number of households (%) 
Water facilities (0.507) Tap/Piped drinking water (H9) 0.154 Number of households using tap water/pipelines (HH) 
Tubewell/Hand pump of drinking water (H10) 0.381 Number of households using hand pump/tube wells (HH) 
Covered well of drinking water (H11) 0.128 Number of households using covered wells (HH) 
Uncovered well of drinking water (H12) 0.202 Number of farmers who depend on pond water (HH) 
River/Stream of drinking water (H13) 0.134 Number of farmers who depend on rivers/streams (HH) 
Production water (C) Land use development (0.475) Cultivated land reclamation rate (C1) 0.284 The proportion of cultivated land to county area (%) 
Utilization of paddy fields (C2) 0.114 Ratio of paddy field area to cultivated land area (%) 
Farming ratio (C3) 0.176 Paddy field area and dryland area ratio (%) 
Irrigation rate (C4) 0.346 The ratio of actual irrigated area to cultivated area (%) 
Land leasing rate (C5) 0.153 Percentage of tenants and landowners (%) 
Crop yield (0.397) Food crop production (C6) 0.304 Bean production, including rice, corn, millet, wheat, barley (Mt) 
Economic crop production (C7) 0.137 Including oilseeds, potatoes, tobacco, sugar cane production (Mt) 
Meat production (C8) 0.107 Including meat, eggs, wool, fish production (Mt) 
Livestock number (C9) 0.140 The sum of the quantities of livestock. Cattle = 1.0; buffalo = 0.8; sheep = 0.6; chicken, duck = 0.2; others = 0.1 (in number) 
The yield ratio of crop species (C10) 0.030 Ratio of food crops to economic crops (%) 
Food crop area ratio (C11) 0.128 The proportion of food crops to the total area of food crops and cash crops (%) 
Economic crop area ratio (C12) 0.613 The proportion of cash crops to the total area of food crops and cash crops (%) 
Water resources (0.128) Per capita water resources (C13) 0.221 The ratio of total water resources to population in a district (m3/P) 
Perennial average runoff (C14) 0.284 District annual runoff for many years (m3/s) 
Perennial average precipitation (C15) 0.167 District annual average annual rainfall (mm) 

HH, households; Mt, ton; P/km2, per square kilometer of population.

Indexing

In Nepal, the agricultural water system is relatively underdeveloped, and therefore, we adopted a simple indexing method proposed by Shi & Lu (2001). Our index system classifies water use into two categories – domestic and production. In the domestic-water category, population, socioeconomic status, drinking water condition, and water facilities are considered as the level-1 indices. The level-1 indices of the production-water category are land-use development, crop yield, and water resource index. Each level-1 index further cascades into several level-2 indices in order to increase the resolution of this analysis. All the indices were standardized into the 0–1 range based on a min–max normalization. Their priorities were described by weights calculated from their entropy.

Modeling

In other to discover the relationship between water availability and livelihood adaptation, logistic regression models were applied. The dependent variable is one of the indices referring to livelihood adaptation, and the independent variables are water availability indices. The model is given as (Greene, 2003): 
formula
(1)
where is intercept, stands for the regression coefficient for kth independent variable, and stands for the probability of the occurrence of livelihood adaptation.

Characteristics of agricultural water resources in Koshi River basin

Background

Koshi River basin is located in Himalayas, 26°47′ ∼ 29°12′ N and 5°22′ ∼ 88°21′ E, covering a total area of 8.75 × 104 km2. The southern Koshi River basin (around 3.95 × 104 km), which belongs to Nepal, consists of five watersheds – Poiqu, Pumqu, Sun Koshi, Arun, and Tamur. The confluence is at Dhankuta where five water bodies merge into the Koshi River (Figure 1). Koshi River basin has a rapid elevation drop from north to south. The temperature and climate condition, therefore, transit rapidly from the cold north to the warm south, and the vegetation type varies correspondingly. Koshi River drains into two relatively populated and well-developed regions in Nepal. From east to west, there are 28 districts in the basin, including Ilam, Panchthar, Taplejung, Sunsari, Morang, Dhankuta, Bhojpur, Terhathum, Sankhuwasabha, Saptari, Siraha, Khotang, Okhaldhunga, Solukhumbu, Dhanusa, Mahottari, Sarlahi, Sindhuli, Ramechhap, Dolakha, Kavrepalanchok, Bhaktapur, Kathmandu, Lalitpur, Sindhupalchok, Rautahat, Bara, and Makwanpur.

Fig. 1.

The location of the Koshi River basin.

Fig. 1.

The location of the Koshi River basin.

Basic characteristics of water resources

Temperature

According to the meteorological data, the annual average temperature of Koshi River basin ranges from 8 °C and 15 °C in the region above the elevation of 3,500 meters above mean sea level (MAMSL), 10–22 °C in the region between 1,000 and 3,500 MAMSL, and 15–30 °C below 1,000 MAMSL (Figure 2). For the past 50 years, the average, minimum, and maximum annual temperature has shown an upward trend with increases of 0.87 °C, 0.31 °C, and 1.47 °C, respectively (Gao, 2012). There are two warming periods that started at the end of the 1960s and 1990s, where the 1960s period showed the most extreme temperature variation (Figure 2). From the 1960s the temperature of the upper Koshi River basin has been gradually warming up, coincident with the waning trend that occurred in Tibet (Cai et al., 2003). In terms of the magnitude of temperature rise, the trend of minimum annual temperature was greater than the trend of maximum and average annual temperature. This reflected a gradual reduction in daily temperature range, which occurred after the 1960s.

Fig. 2.

Perennial monthly average temperature in the Koshi River basin.

Fig. 2.

Perennial monthly average temperature in the Koshi River basin.

Precipitation

In Koshi River basin, there is a huge spatial variation regarding precipitation from north to south due to the rapid elevation drop. According to the analysis of meteorological stations, above 3,500 MAMSL, the yearly precipitation is lower than 1,000 mm (Figure 3). It rises up to 1,500–2,500 mm as the elevation goes down to 1,000 MAMSL, where the historical monthly precipitation can reach about 500 mm in July. The basin area under 1,000 MAMSL has a yearly precipitation of 1,000–1,500 mm.

Fig. 3.

Perennial monthly average precipitation in the Koshi River basin.

Fig. 3.

Perennial monthly average precipitation in the Koshi River basin.

Evapotranspiration

Koshi River basin has a yearly potential evapotranspiration of about 1,069 mm. In terms of the daily evapotranspiration rate, as the dry period starts from October to the next April, it would have an average of 2.15 mm/d and a minimum of 1.6 mm/d in December and January. During the wet season, the average evapotranspiration can range from 4.09 mm/d to 5.6 mm/d. Due to the surface soil properties in mountainous areas, the yearly potential evapotranspiration volume goes down as the elevation increases (Figure 4). The yearly potential evapotranspiration is greater than 1,100 mm in the regions below 1,000 MAMSL, and it is less than 560 mm in the regions above 5,000 MAMSL. The upper Koshi River basin has a yearly potential evapotranspiration of about 435 mm, the middle has about 960 mm, and the lower about 1,171 mm.

Fig. 4.

The change of potential evapotranspiration with altitude in the Koshi River basin.

Fig. 4.

The change of potential evapotranspiration with altitude in the Koshi River basin.

Water resources

Over the past few decades, the average annual runoff in Koshi River basin has been about 1,419 m3/s, the total runoff volume approximately 1,600 × 108 m3 (Deng & Zhang, 2014), and the total volume of sand about 5,420 × 104 m3 (Hu et al., 2012). In terms of the spatial distribution of runoff depth, it gets deeper from north, where the basin is on the southern slope of Himalayas, to south. The northern basin has an average annual runoff depth of about 600–800 mm, runoff of 217 m3/s, and water resource volume of 0.684 billion m3. Gradually moving to the south, the average annual runoff depth increases to about 1,400 mm, runoff gradually increases to about 736 m3/s, and water resource volume increases to about 2.125 billion m3. The basin has a consistently high monthly runoff depth in the summer–fall season between May and September, and it peaks in August at around 574 mm. The shortest monthly runoff depth is around 327 mm, which might be reached in the cold-and-dry season from October to the next April (Figure 5).

Fig. 5.

The perennial monthly average runoff depth in the Koshi River basin.

Fig. 5.

The perennial monthly average runoff depth in the Koshi River basin.

Per capita water resource

The total population in Koshi River basin is around 1.291 million, about a quarter of Nepal's total population. The average water resources is about 6,161 m3 per person, which is slightly higher than the average water resources in Tibet, China. The upper basin is rich in water due to multiple sources including glaciers, snow, and groundwater. Around 1.09 million residents living in the area results in an average of 17,027 m3 water per person. Due to the topological conditions, groundwater use is over 1.5 of the consumption of surface water, resulting in more water wells and pumping facilities being used in the upper basin agricultural water system. Although the population density in the middle basin is around 200 person/km2, residents are highly concentrated in Kathmandu, Bhaktapur, Lalitpur, and nearby regions. This causes the water resources per person in remote areas (around 6,436 m3/p) in the middle basin to be almost three times higher than the urban areas. The smallest water volume per person is in Kathmandu, reaching around 441 m3/p. The middle basin agricultural water management is mainly based on storage irrigation system, which is similar to the upper basin. Groundwater use is 1.25 of the surface water. The lower basin (lower than 1,000 MAMSL) is a flat floodplain with the biggest population density, 600 person/km2. This is a relatively rich region in Nepal but it has the smallest average water accessibility per person, which is around 3,377 m3/p (Figure 6). The irrigation water use is mainly based on surface water, and groundwater use is three-quarters of the total surface water use.

Fig. 6.

Map of per capita water resources.

Fig. 6.

Map of per capita water resources.

Results and discussion

Domestic water

The domestic water use was analyzed from four different aspects – population size, socioeconomic status, drinking water conditions, and water facilities. The population size has a spatial pattern that increases as elevation drops. Above 3,500 MAMSL, the population size is lower than 30,000 with a density of about 30–40 person/km2, and the population growth rate close to 0% from the negative side (Figure 6). In the region between 1,000 and 3,500 MAMSL, the total population size is around 80,000 with a density between 100 and 200 person/km2. There is a strong negative population growth of about −1%. The region below 1,000 MAMSL has a population of over 100,000, and the population density is around 500–650 person/km2. The population is growing at an average rate of 1–2.5% (Figure 6). The local population hotspots are located in Kathmandu, Lalitpur, and their surrounding areas in the middle basin. Their local population density ranges from 1,000 to 4,400 person/km2 with a rapid growth of 3–4.7%.

Despite some local variations, the spatial pattern of socioeconomic status in Koshi River basin closely reflects the population spatial distribution. The weight of poverty rate in socioeconomic factors reached a high value of 0.418 (Table 1). Poverty rate is also relatively high in areas with large population density. Population density and poverty rate have similar distribution trends, which indicates that these two factors play a key role in household water use. Statistical yearbook analysis results show that the upper basin has a prosperity rank of 40, a poverty rate of 25%, and a Gini index around 3. The middle basin has a prosperity rank of 15–40, a poverty rate of 15–25%, and a Gini index between 1.5 and 3. There is a big variation of the socioeconomic status in the lower Koshi River basin, where the prosperity rank jumps from 2 to 62, the poverty rate ranges from 7 to 40%, and the Gini index is 1.5 to 10. This is because of the interaction between population pressure and the accessibility of natural resources. Relatively, Lalitpur and Kathmandu have healthier economies in the middle basin, while Ilam is the richest area in the lower basin.

Thanks to the efforts made by Nepal's government, the accessibility of drinking water has been substantially improved, in particular for farmers. Around 70–95% farms have already been covered by drinking water facilities (UNWWDR, 2003). However, the water quality is still an issue due to the lack of observation data. Around 85% of farms around the main cities are covered by water cleaning services, i.e., the urban domestic water in Kathmandu, Lalitpur, Bhaktapur, Siraha, Sunsari, and Morangthe is filtered and disinfected drinking water. Treated water is allocated to the above cities and regions by building water pipes or carrying water by trucks. On the other hand, remote areas have very little coverage of water cleaning plants. Some low-elevation regions in the lower basin can only receive 17% coverage of the clean water supply. Most of the residents above 1,000 MAMSL have access to piped-in water. For those high-elevation regions with no coverage of piped-in water, water supply mainly comes from springs, but 80% of low-elevation residents obtain drinking water from wells.

In summary, water facility and population size are the two most critical factors that affect the accessibility of households to their domestic water. Although the weight of poverty rate and sanitation coverage factor at the second level is very high, even exceeding the weight of all the other factors, because they are lower than water facility and population size indices at the first level, they are generally regarded as the second most important influencing factor. To improve accessibility, it is important to distribute the water facilities according to the topographical conditions, control the population, and efficiently manage water supply.

Production water

Three main driving factors – land use development, crop yield, and water resources – have been considered to study agricultural production water accessibility. Koshi River basin is overpopulated and lacking in usable land, and most of the land is used for agriculture production. In recent years, even land with a slope greater than 25° has been exploited to produce crops. According to the per capita water resource and perennial average runoff, we estimated the total water resources at county level. Combining our calculation with the LUCC map, we further estimated the unit water accessibility for both paddy field and dry land. Due to the topographical condition, the upper Koshi River basin above 3,500 MAMSL has a water resource volume of 0.684 billion m3 (Figure 6), and the cultivated land reclamation rate is at the lowest level throughout the basin. Dry land is the main cultivated land type and paddy field accounts for a relatively low proportion. The water availability of dry land and paddy field is 0.069 billion m3 and 0.017 billion m3, respectively. In the upper Koshi River basin, the irrigation rate is between 8% and 30% and land leasing rate is below 1%. In the middle Koshi River basin, the cultivated land reclamation rate is higher than the upper basin but lower than the plain areas, mostly between 40% and 60%. Some local areas in the middle basin reach more than 70% in terms of land reclamation rate. The middle Koshi River basin has a water resource volume of 1.66 billion m3, while the water availability of dry land is 0.801 billion m3. Despite the high water availability of dry land, the irrigation rate in this area is relatively low. Most of the irrigation rates are below 20%, with only a few areas reaching 30%. In the middle Koshi River basin, most of the land leasing rate is less than 5%; only in the capital and its surrounding areas does it rise to about 30%. Due to good natural geographical conditions and very high population density, the total amount of water resources is as high as 2.125 billion m3 and cultivated land reclamation rate has reached over 80% in the lower Koshi River basin. The water availability of dry land and paddy field reach 0.62 and 1.07 billion m3 separately and the irrigation rate is also increased significantly to 60–90%. In the lower Koshi River basin, most of the land leasing rate is more than 50% and in some regions over 90%.

Crop yield reflects the effectiveness and efficiency of an irrigation system and water management strategy. The upper basin has a relatively low crop yield, but there is around 33.35% of agriculture land being used for growing economic plants. The total water used for economic plant growing is approximately 0.228 billion m3. The local yield ratio of crop species is relatively low, indicating the water used to grow crops is less than economic plants. In the middle basin, food crop area ratio is 88–93%, which indicates this region is still dominated by food crops. The floodplain below 200 m sea level in the lower basin is the main cropland in Nepal, where food crop area ratio is 70–89%. However, the economic crop production is still far greater than the food crop production. It is worth mentioning that higher elevation areas in the lower basin have a food crop production less than the floodplain, while meat production and livestock are equivalently high.

The per capita water resources in the Koshi River basin is over 1,000 m3/p which is extremely abundant. However, water resources management is inefficient, and a serious spatial imbalance is demonstrated. The upper basin has the richest water resource accessibility, but it also has the most inefficient water management and limited irrigation facilities for supporting agricultural activities. The middle basin has a yearly average precipitation of around 1,500–1,800 mm, and the agricultural land is mostly dry land. Many problems have occurred over recent years, especially in those remote areas. The major issues include deforestation and a lack of irrigation facilities, which lead to low crop production. The areas with a better economy, especially the Kathmandu region, have a limited amount of water access. However, there is no issue regarding irrigation since industrial production dominates this region. The floodplain in the lower basin has a tropical–subtropical climate with a large amount of annual precipitation as well as evapotranspiration. The residents on the floodplain have the highest water accessibility in terms of per capita water resources, drinking-related and producing-related water facilities. Of course, both food crop production and economic crop production are higher than other regions in the basin.

Agricultural water resources availability

According to the weights in Table 1, we calculated the level-1 indices for a total of 29 districts in Koshi River basin. High population size index, 0.3–1, occurs in the Kathmandu region. The lower basin has an index of 0.2–0.3, while the middle and upper basins have an index less than 0.06 (Figure 7). The best economy index also occurs in Kathmandu region, around 0.7–0.9. In spite of some areas in the lower basin having high economy indices, the overall economic performance is slightly lower than the middle basin. The lowest economy index value (less than 0.35) occurs in the upper basin. Corresponding to the economy and population indices, the drinking water condition index peaks in the region around the capital city. Water supplies from springs and ponds are dominant in the upper basin, while the lower basin is mainly based on streams, pipes, and wells. This tells us the spatial differences in water accessibility and quality.

Fig. 7.

Agricultural water availability first-level indexes normalized value (dimensionless quantity).

Fig. 7.

Agricultural water availability first-level indexes normalized value (dimensionless quantity).

In terms of land-use development, the lower basin floodplain has the highest rate of arable land with an index value over 0.4. Kathmandu region has a lesser amount of average land per person but their land-use index is around 0.2–0.5 (Figure 7), which is higher than the average value of 0.1 in the middle basin. There is consistency between crop yield index and land-use index; the lower basin has the highest average crop yield index (over 0.7), while the upper basin has the lowest (less than 0.2). Water resource index is high in the upper basin, over 0.7, but a very limited amount of water has been used for agricultural production. There is a big spatial imbalance of the water resource index over the middle basin. Some districts, such as Sankhuwasabha, have a water resource index of 0.9, while other districts, such as Kathmandu, have an index of 0.15. A similar situation occurs in the lower basin and the index varies from 0.15 to 0.5. This is due to the size of the residential population. In summary, the water use efficiency for agricultural production in the middle basin is higher than the lower basin, while the upper basin is worst (Figure 7).

Similar to the water resource index, strong spatial imbalance of the domestic water index occurs in the middle basin. Kathmandu's index reaches 0.886, but Bhaktapur, Lalitpur in the middle basin, and Sunsari and Morang in the lower basin are between 0.37 and 0.45 (Figure 8), and the districts in the upper basin are lower than 0.15. The overall domestic water index is very low throughout the whole basin, and much effort has to be made to improve domestic water accessibility. The production water index is high in the lower floodplain, and it drops from 0.7 to 0.1 as the elevation increases. The upper basin has relatively smaller values of both domestic and production water index, reflecting its poor economy. The whole basin has sufficient water resources but limited water facilities and inefficient water management systems. Effective establishment of a series of multi-purpose water management systems should aim at improving the quality of life and the environmental conditions of local people and improving rural livelihood choices through rational, equitable, and sustainable water resources planning. Sustainable management of local water resources can be realized by enhancing local rural management capacity. Financial, technical, and managerial support can help to attain clean drinking water, irrigation, and small hydropower facilities, and to provide diversified and available technological options to improve livelihoods (Suhardiman et al., 2015). These methods have been gradually applied to Kathmandu and its surrounding areas in recent years. Although better access to drinking and agricultural water is taking place in the Kathmandu region, there is still great potential to make improvements. The lack of water facilities and effective water management systems are particularly prominent in the middle streams of the Koshi River. These problems will lead to the phenomenon of farmers constantly seeking water. Many water conservation sites have been destroyed and forest coverage is decreasing year by year. In view of these phenomena, we should improve our water management systems and invest in the construction of rural water facilities to preserve, store, and use water, which will lead to the achievement of sustainable use of water resources.

Fig. 8.

Agricultural water type normalized value (dimensionless quantity).

Fig. 8.

Agricultural water type normalized value (dimensionless quantity).

Livelihood adaptation strategy

Major industrial areas are centered on Kathmandu, except for Dolakha industrial district which is located in the upper basin and Siraha in the lower basin. The three districts around Kathmandu are rich in terms of both economy and culture, but the average agricultural land is less than 0.03 hm2 per person and the average agricultural production is less than 200 kg per person. The population density is quite low in Dolakha, resulting in average agricultural land of 0.25 hm2 per person and yield of 300 kg per person. The lower floodplain has high-quality agricultural land but a very dense population. The average agricultural land owned by one person is therefore only 0.13 hm2, but its average yield is over 360 kg per person.

With respect to non-agricultural production, livelihood and production activities vary over the Koshi River basin. Kathmandu is mainly supported by the hospitality industry, which is helped by its natural and cultural attractions. Dolakha is located on the southern slope of the Himalayas, where the weather conditions limit agricultural production. Due to less arable land per capita and relatively low agricultural incomes in the Dolakha area, most farmers go out to work to improve their livelihood. Its economy, therefore, is mainly sourced from non-farming activities, such as tourism-related services, grocery, restaurant, and crafting. Siraha mainly benefits from its relatively well-developed industry, but Siraha is rich in water resources and has the potential to be developed into both an agricultural and non-agricultural area.

Twenty-four districts in Koshi River basin are dependent on agriculture, which accounts for 74.1% of the total population. Due to the simple cultivation structure, we divided agricultural production into crop and non-crop (economic plants). During the modeling, we encoded pure non-crop districts as 0, the districts with food crop production to economic crop production ratio greater than 1.5 as 1, and the pure economic crop district as 2 to implement logistic regression (Table 2). From the regression result, we found that population density, cultivated land reclamation rate, utilization rate of paddy field, and river/stream source of drinking water have significant influence on estimating livelihood strategies in food crop-dominated regions. On the other hand, establishment of livelihood strategies is highly correlated with population density, tap/piped and well-sourced drinking water. We found that population stress forces households to conduct more non-crop activities (Su et al., 2016). Moreover, a better development and management of water accessibility would benefit the income of farming households.

Table 2.

Multivariate logistic regression results of crop types and livelihood strategies.

Classification Indicator type Food crops are dominant
 
Economic crops are dominant
 
Regression coefficients B Exp(B) Regression coefficients B Exp(B) 
Population size Population density −2.364** 0.053 −4.024*** 0.334 
Water facilities Tap/Piped source of drinking water 10.479 3.557 3.387** 5.168 
Uncovered well source of drinking water −9.585 0.687 2.071** 0.107 
River/Stream source of drinking water 0.981* 2.666 −3.344 0.035 
Land use development Cultivated land reclamation rate 2.259* 0.104 2.838 17.086 
Utilization rate of paddy field 2.303* 10.004 −6.096 0.002 
Classification Indicator type Food crops are dominant
 
Economic crops are dominant
 
Regression coefficients B Exp(B) Regression coefficients B Exp(B) 
Population size Population density −2.364** 0.053 −4.024*** 0.334 
Water facilities Tap/Piped source of drinking water 10.479 3.557 3.387** 5.168 
Uncovered well source of drinking water −9.585 0.687 2.071** 0.107 
River/Stream source of drinking water 0.981* 2.666 −3.344 0.035 
Land use development Cultivated land reclamation rate 2.259* 0.104 2.838 17.086 
Utilization rate of paddy field 2.303* 10.004 −6.096 0.002 

Goodness of fit x2 = 65 5.104**, the coefficient of determination is Nagelkerke R2 = 0.632.

The model treats non-agricultural patterns as reference numbers.

*, **, ***, respectively, 0.1, 0.5, and 0.01 significance level.

Conclusion

Due to the lack of water facilities and inefficient management, the overall quality and accessibility of drinking water is relatively low in Koshi River basin compared to other regions. Some of the basin areas have very limited drinking water with less than 5 liters per day for one person. In spite of the average water resource owned by each person in Koshi River basin being quite high (9,117 m3/P), the average water consumption is extremely low (154 m3/P). Moreover, over 25% of residents have no access to clean drinking water. This situation gets worse when moving from the lower to the upper basin. In the upper basin, there is very limited coverage of water treatment plants as well as tap water. Seventy-three percent of people in the upper basin live on exposed spring water with simple piping. The monthly average water consumption in this region is around 4 tons in each family. However, such a water supply generally cannot meet consumption in the dry season. To support some of the upper areas, several international non-government organizations have installed many wells for supplying drinking water. In contrast, Kathmandu has a relatively complete water supply system, and over 70% of residents have access to tap water and 14% are supplied by tanks. The monthly average water consumption in Kathmandu region is around 6 tons for each family. Nevertheless, there is an imbalance between water consumption and water supply. According to the statistics (http://smartpaani.com/the-water-situation-in-kathmandu-valley), Kathmandu water demand is 36 × 104 t each day, which is substantially higher than the surface water supply 9 × 104 t in the dry season and 14 × 104 t in the wet season. The consequence is that groundwater has been overexploited. The follow-up issue is that the groundwater in the Kathmandu area has been found to have a high concentration of iron, and therefore, is not safe to drink before treatment. The lower basin has low population density and a relatively high average water resource owned by each person. In recent years, local government has made great efforts to improve the coverage of water access facilities in the lower basin. So far, tap water supply has already covered 40% of residences and more than 80% of farming households have a backyard water well for drinking supply. It is worth mentioning that those farmers in the lower basin who have a water well consume the highest volume of water – over 8 tons of water per month. Drinking water consumption is highly related to local population size and the accessibility of water supply. In the upper and middle basins, population size exhibits a strong positive relationship with water plant coverage, while the lower basin population positively relates to the water wells. Although there is spatial variation in the types of water supply, it still proves that the water supply facility has a strong influence on drinking water accessibility. Overall, Koshi River basin has an average drinking water coverage of 77%, with few districts lower than 65%.

According to the study based on our index system, we found that only Kathmandu region has good accessibility to domestic water while for other districts it is quite low. This tells us that the overall domestic water supply in Koshi River basin is insufficient. The scarcity of domestic water has particularly occurred among the agricultural population. On the other hand, the accessibility of production water is sufficient in the lower basin, but gets worse as the elevation increases. Although Koshi River basin has a great amount of water resources, the inefficient water management strategy has introduced many issues related to water availability, and therefore, the basin economy is vulnerable to climate change. In particular, due to the lack of water-related facilities, the farming households in the Koshi River basin are extremely susceptible to the change in weather conditions. It is important to adopt new engineering solutions and technical advances in order to adapt to the changing climate. Moreover, optimizing agricultural water use and planting multiple crop types can also help agricultural households to face the challenges introduced by climate change. In addition, we should improve water management systems and invest in the construction of rural water facilities to preserve, store, and use water, which will achieve sustainable use of water resources in a least developed country.

Acknowledgments

This study was supported by the National Natural Science Funds of China (grant no. 41871357), the Sichuan Basic Science and Technology Project (grant no. 18YYJC1148), the Branch of Mountain Sciences, Kathmandu Center for Research and Education, CAS-TU, Chengdu, China (grant no. Y8R3310310), the Hundred Young Talents Program of the Institute of Mountain Hazards and Environment (grant no. SDSQB-2015-02), the One-Three-Five Project of Chinese Academy of Sciences (grant no. SDS-135-1708), and the Science and Technology Service Network Program of the Chinese Academy of Sciences (grant no. Y8R2020022). We also thank the Geography Department of Nepal Tribhuvan University and the Joint Research Center for China-Nepal Geography for providing data.

References

References
Almas
A. A. M.
&
Scholz
M.
, (
2006
).
Agriculture and water resources crisis in Yemen: need for sustainable agriculture
.
Journal of Sustainable Agriculture
28
(
3
),
55
75
.
Ayars
J. E.
,
Christen
E. W.
,
Soppe
R. W.
&
Meyer
W. S.
, (
2006
).
The resource potential of in-situ shallow groundwater use in irrigated agriculture: a review
.
Irrigation Science
24
(
3
),
147
160
.
Cai
Y.
,
Li
D. L.
,
Tang
M. C.
&
Bai
C. Y.
, (
2003
).
Decadal temperature changes over Qinghai-Xizang Plateau in recent 50 years
.
Plateau Meteorology
5
,
464
470
.
Chinnasamy
P.
,
Bharati
L.
,
Bhattarai
U.
,
Khadka
A.
,
Dahal
V.
&
Wahid
S.
, (
2015
).
Impact of planned water resource development on current and future water demand in the Koshi River basin, Nepal
.
Water International
40
(
7
),
1004
1020
.
Deng
W.
&
Zhang
Y. L.
, (
2014
).
Resources, Environment and Development of Koshi River Basin
.
Sichuan Science and Technology Press
,
Chengdu
,
China
.
Gao
J. G.
, (
2012
).
Land Cover Change and its Relationship with Climate Change in the Koshi River Basin
.
Doctoral dissertation
,
Beijing
,
China
.
Greene
W. H.
, (
2003
).
Econometric analysis
.
Journal of the American Statistical Association
89
(
428
),
182
197
.
Hu
G. S.
,
Chen
N. S.
,
Narendra
K.
&
Han
D. W.
, (
2012
).
On the water hazards in the trans-boundary Kosi river basin
.
Advances in Earth Science
27
(
8
),
908
915
.
Jia
D. L.
&
Jiang
W. L.
, (
2000
).
Comments on improving agricultural water use efficiency
.
Water Saving Irrigation
5
,
18
21
.
Li
Y. P.
&
Huang
G. H.
, (
2011
).
Planning agricultural water resources system associated with fuzzy and random features
.
Journal of the American Water Resources Association
47
(
4
),
841
860
.
Musa
I. K.
, (
2009
).
Priority issues of least developed countries in Africa
.
Irrigation and Drainage
58
,
39
51
.
Palazzoli
I.
,
Maskey
S.
,
Uhlenbrook
S.
,
Nana
E.
&
Bocchiola
D.
, (
2015
).
Impact of prospective climate change on water resources and crop yields in the Indrawati basin, Nepal
.
Agricultural Systems
133
,
143
157
.
Pandey
V. P.
,
Manandhar
S.
&
Kazama
F.
, (
2012
).
Water poverty situation of medium sized river basins in Nepal
.
Water Resources Management
26
,
2475
2489
.
Postel
S.
, (
1999
).
Pillar of Sand: Can the Irrigation Miracle Last?
W.W. Norton
,
New York
,
USA
, p.
313
.
Rautanen
S. L.
,
Koppen
B. V.
&
Wagle
N.
, (
2014
).
Community-driven multiple use water services lessons learned by the rural village water resources management project in Nepal
.
Water Alternatives
7
(
1
),
160
177
.
Rosenzweig
C.
,
Strzepek
K. M.
,
Major
D. C.
,
Iglesias
A.
,
Yates
D. N.
,
McCluskey
A.
&
Hillel
D.
, (
2004
).
Water resources for agriculture in a changing climate: international case studies
.
Global Environmental Change
14
(
4
),
345
360
.
Scarascia
M. E. V.
,
Battista
F. D.
&
Salvati
L.
, (
2006
).
Water resources in Italy: availability and agricultural uses
.
Irrigation & Drainage
55
(
2
),
115
127
.
Schütze
N.
&
Wagner
M.
, (
2016
).
Integrated management of water resources demand and supply in irrigated agriculture from plot to regional scale
.
Proceedings of the International Association of Hydrological Sciences
373
,
51
55
.
Setegn
S. G.
,
Rayner
D.
,
Melesse
A. M.
,
Dargahi
B.
,
Srinivasan
R.
&
Wörman
A.
, (
2011
).
Nile River Basin
.
Springer Science + Business Media
,
Dordrecht
, pp.
241
265
.
Shi
Y. L.
&
Lu
L. S.
, (
2001
).
Agricultural Water Demand and Construction of Water Saving and Efficient Agriculture in China
.
China Water & Power Press
,
Beijing
,
China
.
Sovacool
B. K.
,
Agostino
A. L. D.
,
Rawlani
A.
&
Meenawat
H.
, (
2012a
).
Improving climate change adaptation in least developed Asia
.
Environmental Science & Policy
21
,
112
125
.
Sovacool
B. K.
,
Agostino
A. L. D.
,
Meenawat
H.
&
Rawlani
A.
, (
2012b
).
Expert views of climate change adaptation in least developed Asia
.
Journal of Environmental Management
97
,
78
88
.
Su
Y.
,
Deng
W.
,
Zhang
J. F.
,
Hriday
L. K.
&
Paudel
K. S.
, (
2016
).
Relationship between livelihood capital and livelihood strategies of rural households in Melamchi basin of central mountainous area in Nepal
.
Journal of Mountain Science
34
(
5
),
645
653
.
Suhardiman
D.
,
Clement
F.
&
Bharati
L.
, (
2015
).
Integrated water resources management in Nepal: key stakeholders perceptions and lessons learned
.
International Journal of Water Resources Development
31
(
2
),
284
300
.
UNWWDR (United Nations World Water Development Report)
(
2003
).
Water for People, Water for Life – Executive Summary
.
UNESCO Publishing
,
Paris
,
France
, pp.
17
18
.