Water infiltration rate in the Kathmandu Valley of Nepal amidst present urbanization and land-use change

Groundwater is the primary source of water for 2 billion people around the world (Alley et al. 2002). Despite its importance, most aquifers are experiencing rapid rates of groundwater depletion (Konikow & Kendy 2005). Due to combined anthropogenic and environmental processes, there is a high complexity of groundwater management which once degraded is difficult to repair (Jakeman et al. 2016). There are many factors that amplify its degradation, namely population growth, land-use and land-cover (LULC) change, climate change, and others (Gosling & Arnell 2016; Weber & Sciubba 2018; Olivares et al. 2019). However, change in land-cover is expected to lower recharge and deplete groundwater levels even more (Scanlon et al. 2005; Mishra & Kumar 2015). Hence, in order to ensure the sustainability of groundwater, a better understanding of the impact of land-use and the amount of water recharge is needed.

There are different approaches to assessing the impact of LULC change on groundwater recharge. For example, experimental methods such as isotopic tracers (Wang et al. 2008), statistical approaches like water-table fluctuation analysis (Moon et al. 2004), and numerical methods like water balance simulation (Batelaan et al. 2003) are in use. However, the methods are costly and time constraining. Effective infiltration is yet another type of method used mainly for the permeability of surface deposits (Paczynski 1995; Stasko et al. 2012).

Infiltration is one of the important contributing factors to groundwater recharge. Estimation of infiltration in the field has always proved to be difficult. Therefore, most of the available estimates are based on theoretical calculations, considering parameters like slope, characteristics of soils, amount and duration of rainfall, runoff, etc. Whereas, some approaches have been adopted to quantify rate in the field using portable rainfall simulators (Harden & Scruggs 2003): double-ring infiltrometer (Osuji et al. 2010; Wang et al. 2018; Shrestha & Kafle 2020), mini-disc infiltrometer (Kumar et al. 2021), single-ring infiltrometer (Verbist et al. 2010), artificial precipitation simulator (Wang & Zhang 1991), run off-on ponding techniques (Bobe 2004), etc. Nevertheless, the infiltration method is advantageous as it incorporates other processes such as the movement of water within the soil (Turner 2006), soil physical properties (Walker et al. 2006; Rashidi & Seyfi 2007), surface soil compaction (Yimer et al. 2008), vegetation coverage, and types (Molina et al. 2007). Although, the infiltration method is sensitive, it is easy to understand and cost-effective and easy mobility makes it useful to researchers in developing countries.

In general, population growth and urbanization are two main drivers for increasing water demand globally (Bradley et al. 2002; McDonald et al. 2011); meanwhile South-Asian countries are suffering from the reduced groundwater infiltration with increased concrete pavements. Meanwhile, countries in the South-Asian region are highly dependent on groundwater. The problem is huge in those areas where the watershed is isolated, with groundwater being a major component of water resources. The Kathmandu Valley (KV) of Nepal is no exception, having an isolated watershed combined with a population boom and change in land-use patterns.

The sole domestic water supplier in the KV, the Kathmandu Upatyaka Khanipani Limited (KUKL), fulfils 60–70% of the water demand in the dry season and nearly half in the wet season using groundwater sources, causing water scarcity and increased dependency on shallow and deep groundwater (Shrestha & Shah 2014). Consequently, this valley has witnessed water-table decline due to over extraction (Shrestha 2009; KVWSMB 2012; Gautam & Prajapati 2014). Due to unprecedented LULC change in the valley, the recharging areas are also being affected, as it transforms permeable land (open vegetated areas) to impervious (concrete buildings and infrastructure) (Zhou et al. 2013) resulting in increased surface runoff and loss of groundwater recharge areas (Lamichhane & Shakya 2019). The decrease in infiltration amidst over-extraction of groundwater is making the valley vulnerable to land subsidence (Pandey et al. 2010; Gautam & Prajapati 2014).

Based on the groundwater recharge potential, the KV is divided into three different groundwater districts, i.e., Northern, Central, and Southern districts. Among the districts, the Northern District has comparatively high groundwater recharge potential (Shrestha & Shah 2014; Dahal et al. 2019; Shakya et al. 2019). Different groundwater modeling has been adopted to study hydrology and groundwater dynamics of the KV (Sonaje 2013; Dahal et al. 2019; Lamichhane & Shakya 2019, 2020). However, there seems to be very scarce previous works on groundwater recharge based on the field experiments. Since the land-use pattern of the KV is changing rapidly, the infiltration capacities of the critical locations considered as the groundwater recharge are to be considered for the future land-use pattern and water management practice. The present work aims to determine the infiltration rate using simple cost-effective infiltration methods in changing land-use in the northern recharge areas (JICA 1990) of the KV. Depending upon land-use, soil types, soil texture, and geological formation, this study is aimed at understanding the relationship between infiltration rate and soil parameters along with estimating the groundwater recharge volume by predicting the land-use change pattern for 2010, 2020, and 2030 in the KV for effective water management.

The study area encompasses cultivated land, built-up areas, and non-cultivated land. The cultivated land covers seasonal crops, irrigated, or non-irrigated farms with main seasonal crops including paddy, wheat, barley, potato, chili, onion, garlic, maize, etc. The built-up or settlement areas cover all types of impervious (or very little pervious) land consisting of paved streets, residential buildings, highways, and commercial areas. The non-cultivated land includes areas with natural vegetation and regeneration, and open areas.

The method of determining infiltration rate using a double-ring infiltrometer has been adopted by many researchers in different regions of the world (Bean et al. 2004; Igboekwe & Adindu 2014; Lamichhane & Shakya 2019; Mahapatra et al. 2020), making it cost-effective and result orientated.

The fit of Equation (1) was carried out in simple linear regression in MS-Excel.

The undisturbed soil samples from the sampling points were collected using a soil auger at depth of 0–20 cm. The upper horizon of soil has direct exposure to natural and anthropogenic changes along with high organic content and nutrient reserves (Tiwari et al. 2006). Hence, for the present study, soil-depth up to 20 cm was considered. Soil samples were separately collected for calculating bulk density and moisture content, pH, texture, and organic matter. For the physical and chemical properties, sampled soils were air dried for 2 weeks, gently crushed, and stored in clean polythene bags which were later passed through a 2-mm sieve for further lab analysis.

The various soil parameters were determined using standard methods. The bulk density was determined using the core sampler with measurement (radius and height of core being 1.8 and 4 cm, respectively) (Grossman & Reinsch 2002) and moisture content determined using the oven dry method. The soil pH was determined using a 1:5 soil–water ratio with a Milwakee pH probe. Likewise, organic carbon was determined using Walkley and Black wet oxidation methods (Nelson & Sommers 1982). Organic matter was calculated by a factor of 1.72 (Van Bemmelen's Correction Factor) (Waxman & Stevens 1930). The soil texture was determined using the Bouyoucos hydrometer method (Gee & Bauder 1986).

The data collected from the field were arranged, organized, and analyzed using Spearman's rank correlation, Mann–Whitney U test, and Kruskal Wallis test. Multivariate analysis such as principal component analysis (PCA) (Tiwari et al. 2006; Abdel-Fattah et al. 2021) and cluster analysis (Cupak et al. 2017) were carried out. All analyses were carried out in MS-Excel, IBM SPSS Statistics 23.0, and ArcGIS 10.2.1.

The present study revealed that the steady infiltration rate varies considerably from 0.01 to 37.2 cm/h with an average of 7.3 cm/h (Figure 4). Setiawan et al. (2019) reported steady infiltration rate variation from 5.4 to 63.93 cm/h in the Lombok Island, Indonesia. Other researchers (Chen et al. 2014; Patle et al. 2018; Wang et al. 2018) also reported the similar findings. Meanwhile, various researchers reported infiltration rates different from the current study. The various infiltration rates are presented in Table 1. These kinds of variation in steady infiltration rate are presumed to be from the root and faunal microspores that exist in association with the land-use and variety of crop species (Harden & Scruggs 2003). Also, adoption of varying ring type and its diameter made it different from other studies’ findings.

In terms of steady state in the present study, it was achieved maximally between the time period 102 and 121.8 min (1.7 and 2.03 h, respectively) (Figure 3). Diamond & Shanley (2010) reported the steady infiltration rate at 4.5–7.5 h in winter and from 6.5 to 8.5 h in summer. The reason for the differential time required to achieve steady state may be due to different spatial and temporal conditions.

According to Kohnke (1968) classification, in the present study, the infiltration rates fall in different categories, i.e., 3.5% in very slow, 9.4% in slow, 25.9% in slow to medium, 23.5% in medium, 11.8% in medium to fast, 20% in fast, and 5.9% in very fast. The interpolated map (Figure 4) shows high infiltration rate in the Sankhu area (blue color). Areas like Gagalphedi, Budhanilkantha, and some parts of Tarkeshor (green color) have medium to fast infiltration rates. The areas such as Gokarna, Tokha and Tarkeshwor (yellow color) have medium infiltration rates (Figure 4). Dahal et al. (2019) reported similar results from the northern part of the KV which includes areas like Lapsephedi, Gagalphedi, Nayapati, Sundarijal, Chapali, Bhadrakali, Budhanilkantha, and Baluwa, validating the results of the current study.

The results showed a significant positive correlation (r = 0.287, 0.01) between infiltration rate and organic matter, whereas other parameters showed no significant correlation (Table 2). Several studies also found a significant positive relation between infiltration rate and organic matter (Osuji et al. 2010; Haghnazari et al. 2015; Wang et al. 2018). Soil organic matter was considered to be the chief component for determining soil quality (Tiwari et al. 2006), hence the soils with high infiltration rates are of good quality index due to high organic content. During the dry season, soil moisture decreases resulting in an increase in absorption capacity (sorptivity) of soil and cumulative infiltration (Yasin & Ghazal 2021). However, in the present study, there seems no apparent impact of soil moisture content on the basic infiltration.

The result of the Kaiser Meyer Olkin (KMO) measure of sampling adequacy and Bartlett's sphericity test during the PCA gave a value 0.613. The eigenvalue greater than 1 was considered for the number of components that study the variation and resulted in three components from nine variables which contributes 65.71% of the total variance. According to factor loading, PC1, PC2, and PC3 explained 34.15, 19.71, and 11.84% of the total variance, respectively (Table 3).

Land conversion from permeable to non-permeable built-up areas has declined the volume of water that infiltrates into the ground by 8.68 million cubic meters per year (MCM/year) in 2020 and is projected to decline by 7.55 MCM/year in 2030. In 20 years, i.e., from 2010 to 2030, 16.23 MCM/year loss has been estimated (Table 4). The decline in groundwater recharge due to the LULC change is also supported by Lamichhane & Shakya (2020) and Pu et al. (2020). Increase in urban areas has resulted in a decrease in groundwater volume at a rate of 284.34 MCM, a deficit of 115.34 MCM (a reduction in the groundwater levels of 0.1 m/year) in Oaxaca, Mexico (Olivares et al. 2019). Urbanization has led to impervious pavement, causing an effect on surface water systems and groundwater recharge. Decline in rainwater infiltration leads to high runoff, peak discharge flow, then urban flood at downstream areas posing threats to life and property (Pataki et al. 2011; Yao et al. 2015). Urbanization along with climate change increases these issues even more seriously.

The techniques described by Singh et al. (2010), i.e. the ‘recharge from field percolation’ method, has somewhat similar kinds of recharge processes to those adopted in this research. It is based on deep percolation occurring through cultivated areas from irrigated fields and is the main recharge component, contributing 57% of the total recharge.

The estimated volume of groundwater recharge in this study was greater than that estimated by Lamichhane & Shakya (2020). Since this northern groundwater district is a major aquifer for the KV, estimates may be high. This value is assumed to be the maximum in the permeable areas of KV, whereas other studies (Pandey et al. 2013; Shrestha et al. 2017) presented the average value for the whole KV. Moreover, the variation in estimates may be due to differences in the approach used and aquifer layers considered (Shrestha et al. 2012).

The present study has estimated the potential volume of water recharge based on the permeable area in ideal conditions irrespective of any kind of climatological and hydrological factors. However, earlier studies like Gupta et al. (1990); Pandey & Kazama (2011); Shrestha et al. (2017); and Lamichhane & Shakya (2020) have mainly focused on rainfall characteristics and different hydrological factors.

The recharge capacity for the study area was estimated to be uniform for all degrees of slopes; however, Adams et al. (2004) distinguished the probability of recharge based on it as 100, 95, 75, 50, and 25% for slope degree 0–5, 5–10, 10–20, 20–50, and 50–90, respectively. The double-ring infiltrometer used in the present research mainly favors non-sloppy areas, but it has been used constantly in calculating volume for sloppy areas which might have increased the estimation of the results. Permeable areas with mild slopes hence overestimate the volume of water recharge. This kind of overestimation could be reduced using rain simulators or chemical methods in sloppy areas. In addition, the ring size of the infiltrometer also matters for the accuracy of the infiltration rate. The larger rings would provide high accuracy and smaller ring sizes would give less accuracy. Furthermore, to gain high accuracy in the measurement of recharge volume, seasonal study along with morphological characteristics such as slope, elevation, aspect, curvature, and topographic position index (TPI) that impact infiltration rate will add more knowledge of the hydrology of the northern valley for groundwater recharge and supply.

The present study revealed the baseline on the infiltration rate of the northern groundwater district of the KV, which is considered as the groundwater recharge area of the valley, amid the changing land-use pattern. The current study shows a significant impact of land-use on the infiltration rate where higher infiltration rate was observed in the cultivated land compared to the non-cultivated land. Among the different soil properties like pH, moisture, bulk density, organic matter, and soil texture, organic matter was found to impact infiltration rate directly and positively, i.e., land with higher organic matter content has increased infiltration rates and vice versa. In terms of land-use change, 11.78% of permeable land was found to be converted into impervious concrete structures in just a decade (2010–2020), projected to decrease the water volume by 16.23 MCM/year for the period from 2010 to 2030. The double-ring infiltrometer techniques adopted in limited economic conditions in this study were found to be useful in determining the infiltration rate in different land-use types and between various soil components.

The baseline information on the infiltration rate of the study area and estimation of recharge volume with respect to decrease in permeable areas due to urbanization and land-use change are the main outcomes of the present research. The former outcome is important in understanding groundwater reservoirs, designing groundwater management plans, irrigation and drainage systems, and contamination evaluation, whereas the latter is believed to be applied by urban planners to work for the preservation and conservation of permeable areas during urban planning. This field-based assessment coupled with different models of flood helps mitigate and solve urban flooding to some extent, which is emerging in the case of the KV during monsoon. Nevertheless, further field experiments are required to evaluate the impacts of rainfall characteristics in the study area.

Authors are thankful to the Central Department of Environmental Science, Tribhuvan University, Nepal for providing the platform to carry out the present work. The first author expresses sincere acknowledgement to the National Youth Council, Center of Research for Environment, Energy and Water, and Environment and Public Health Organization for providing an M.Sc. research grant. The authors are grateful to Dr Ramesh Sapkota for his valuable support in the statistical analysis.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.