1 Introduction

The natural flow of a river system performs a crucial role in the watercourse maintaining aquatic habitat and ecosystem services [1]. A river system can be affected by both natural and artificial influences that will lead to alterations in flow. Artificial influences mainly include land use change due to urbanization, construction of canals, and most prominently the building of dams [2]. Dams were/are built all over the world for numerous purposes such as flood control, irrigation, hydropower development, navigation, recreation, water supply, etc. On the contrary, dams create a major influence on the discharge by altering the magnitude, frequency, and timing of the maximum and minimum flows. Climatic changes, and seasonal effects specific to some regions of the world determine the role of reservoirs. Rivers in southeastern and eastern Asian countries (which receive monsoon rainfall) have high temporal variations in discharge during the dry and wet seasons. In these regions dams have been constructed to store rainwater received during wet seasons and to utilize it during dry seasons [3]. In Sri Lanka, ancient kings had built dams in the dry zone of the country for agricultural purposes. These huge reservoirs were connected through canals starting from small tanks considering the slope and spatial features of the land. The aim of the creators of these gigantic structures was to ensure that every tiny rain droplet was redistributed without flowing directly to the sea (Ancient water tanks of Sri Lanka to adapt to a changing climate, 2023).

Most of the reservoirs and tanks built during ancient times Sri Lanka are still utilized for agricultural uses [4]. The country experiences two main rainfall seasons: northeast monsoon prevailing from December to February and southwest monsoon from May to September [5]. The main cultivation seasons named Maha and Yala fall in comparison to the northeast and southwest monsoon seasons, respectively [6]. Despite a rich spatiotemporal rainfall over the year, northwest and southeast parts of the country are drier than the rest of island. Nevertheless, agricultural activities are functioning in these areas with the help of the irrigation system [7].

In the early years of 2000s several studies were carried to understand food supply and water security situation in the southern and southeastern parts (Ruhuna Basin) of Sri Lanka [8,9,10]. These were focused on Walawe Ganga, Kirindi Oya, Menik Ganga, and Malala Oya basins. A recent study by the International Water Management Institute (IWMI) suggested that the water resource development plans in the Ruhuna Basin depicted a shift to services and industrial oriented engagements rather than an agricultural movement [8]. This study forecasted that urban and industrial water usage would increase from 10 million cubic meters (MCM) to 100–150 MCM in the coming years. However, the areas are still under the dominant agricultural lands.

Kirindi Oya was immensely utilized with the available water resources available in the basin and Menik Ganga is not adequately discussed for the availability of water resources. Thus, it is open for the exploitation of water resources. Crop production, water resources, and the aspects of water management were discussed in Menik Ganga basin by Weerasinghe et al. [9] in 2002. The water balance was carried out considering the climate inputs, rainfed and irrigated agriculture, water supply extractions, and losses in the basin. It was concluded that among the rivers in Ruhuna basin, Menik basin retained the least amount of water (15% of total runoff) during this period [9].

Consequently, a high caliber requirement was dawned for a trans basin diversion from Menik Ganga to Kirindi Oya [8]. Menik Ganga is generally a drier basin except for the Maha season, and it was mentioned that Menik Ganga demand deficits in the dry season were even higher than that of the Kirindi Oya basin. Additionally, the Kataragama festival, which is a historic religious event in Sri Lanka, also requires satisfactory water for people visiting from all over the country (more than 100,000 people annually). Considering these facts, diverting water from Menik Ganga to Kirindi Oya Basin (specifically to the Lunugamvehera reservoir) is tricky and only should be done in the wet period if necessary [11]. However, the discussions were on the necessity of trans basin diversion from Menik Ganga to Kirindi Oya Basin. The Lunugamvehera reservoir of storage 222 MCM, at which most water of Kirindi Oya is regulated, had a deficiency of water of around 100 MCM in order to cultivate agricultural lands of command area 10,000 ha per year [12]. This is the main reason for the discussions on trans basin diversion.

Among these heated debates the Weheragala Reservoir along the Menik Ganga started its construction in 2005 and was completed in 2009. Thus, understanding the hydrological behavior and hydrologic alterations are important.

The IHA indices were used for the studies related to Yangtze River in China by Huang et al. [13], Yu et al. [14], Gao et al. [15] and the Duero River Basin in the Iberian Peninsula by Pardo-Loaiza et al. [16]. Alteration of flows due to the reservoirs was assessed in these studies.

The Water Evaluation And Planning System (WEAP) is such a model which is highly used in hydrological studies to evaluate water availability and thus to arrange planning applications [17,18,19,20]. Abdi and Ayenew [21] researched the sub basins of a WEAP model for the Central Rift Valley basin in Ethiopia [21].

Even though there are many studies in the literature, these models have limited applications in the Sri Lankan context. The integration of the Soil Water Assessment Tool (SWAT) and WEAP models was incorporated in a study for the Malwathu Oya Basin of Northern Sri Lanka by Kaushalya and Hemakumara [22]. Results showcased that the construction of reservoirs in the lower Malwathu Oya Basin and restoration would reduce the demand scarcities existing in the basin [22].

Contemplating the above-mentioned past studies, the flow alteration must have occurred downstream of the Weheragala Reservoir in Menik Ganga basin, after the commencement of the dam. Therefore, it is necessary to understand these hydrological changes in a river basin that flows along a drier area. This would help the planners to arrange the water budget in a sustainable manner to allocate the necessary water for irrigation and other purposes. Thus, this research study presents a comprehensive analysis of flow alterations that occurred in Menik Ganga in the presence of the Weheragala Reservoir. The WEAP model was used to simulate river flow with reservoirs and considering demands. The research aim was achieved by analyzing the IHAs using the discharge data at Kataragama station (1990–2019) for pre- and post-dam periods and then comparing the IHAs of simulated flows in WEAP for the post-dam period (2010–2019) with observed discharge data for the period.

2 Study area and data

2.1 Study area

The Menik Ganga is located between 6.36° N–6.93° N and 81.15° E–81.54° E (refer to Fig. 1). It is one of the major river basins in the Ruhuna Basin of the southeastern part of Sri Lanka. The Menik Ganga has a catchment area of 1272 km2. The river starts from the Passara hills, flows past Kataragama, and reaches the sea at the Yala National Park. The basin receives rainfall mainly from the northeast monsoon from December to February in which annual rainfall accounts for 1496 mm and a dry period prevails from June to September. Generally, the mean seasonal rainfall for the Maha season is twice the rainfall received in the Yala season [8]. More than 50% of the catchment is covered by forests which is mainly due to the Yala National Park (594 km2 of an area of Yala National Park located within the Menik Ganga basin) [11].

Fig. 1
figure 1

The Menik Ganga basin map

Full size image

As it is stated in Sect. 1, trans basin diversion from Menik Ganga to Kirindi Oya Basin was initiated with the construction of the Weheragala Reservoir in 2005. Figure 1 illustrates the trans basin diversion from one reservoir to the other. The Weheragala Reservoir has a capacity of 64 MCM and the project was completed in 2009 to store and divert the excess water of Menik Ganga to Lunugamvehera Reservoir. The Lunugamvehera Reservoir is another reservoir constructed along Kirindi Oya in 1985. The project was proposed to benefit 8000 farmer families in the agricultural lands [23].

2.2 Data sources

The streamflow gauge in Menik Ganga basin is stationed at Kataragama and the flow data were gathered 1990–2019 from the Department of Irrigation, Sri Lanka. We obtained data from two rainfall stations in the basin: Thissamaharamaya and Lunugamvehera. Temperature data were collected for Monaragala and Hambantota stations for the same period from the Department of Meteorology, Sri Lanka.

The Menik Ganga basin consist of cultivation from both rainfed and irrigated agriculture. Paddy is the main crop in the basin among other home garden crops. The agricultural extents of crops of rice, rubber, coconut, sugarcane, and home garden in Menik Ganga basin before the construction of the Weheragala Reservoir were 10,600, 1000, 1250, 1500, and 18,500 (in ha) respectively [9]. The crop coefficients were gathered from the Food and Agriculture Organization (FAO) crop coefficient tables considering the growth stages and days relevant to the stages. The daily effective crop coefficients were obtained considering the coincidence of the growth stages of crops.

The HydroSHEDS Digital Elevation Model (DEM) data is based on the high-resolution data gathered from the Space Shuttle flight for NASA’s Shuttle Radar Topography Mission (SRTM). ESA-CCI-LC land cover data originates from the Medium Resolution Imaging Spectrometer (MRIS) satellite and has a spatial resolution of 300 m. The Princeton historical climate data comprises observations to generate global daily data of precipitation, wind, and temperature for 1948–2010 with a 0.25° spatial resolution [24]. The description of data used in this study are given in Table 1 below.

Table 1 Details of data used in this study
Full size table

2.3 Methodology

The methodology consists of two sections. The following subtopics discuss the methodology of WEAP model buildup and IHA analysis for the discharges.

2.3.1 WEAP model buildup and calibration

The WEAP hydrological model was used to implement planning and evaluation of water resources and executing policies. The model has the competence of concentrating on concerns linked to analyzing demands, allocation of water, reservoir management and operations, simulation of discharge, and benefit-cost analyses [24]. In addition, the model is capable to incorporate a variety of hydrological processes contained with water resources and infrastructure management reasonably. Numerous scenarios related to climate and other temporal variants are also able to be set up in the model [25].

The required data were gathered from the sources mentioned in Table 1. A hydrological model was built using the WEAP model for the Menik Ganga basin before the construction of the Weheragala Reservoir by delineating the catchment and adding respective irrigation and water supply demands prior to the reservoir construction. The model was calibrated at the Kataragama gauging station for the period 1990–2009. Then, the model was run for the 2010–2019 duration with ground rainfall climate data. This simulated streamflow was compared with the Kataragama discharge (post-Weheragala Reservoir construction) for IHA analysis (described in Sect. 2.3.2). The flowchart for methodology is illustrated in Fig. 2.

Fig. 2
figure 2

Flowchart of WEAP model development

Full size image

The hydrological model was built using WEAP software for the scenario before the construction of the Weheragala reservoir. The catchment was delineated using the “Automatic Catchment Delineation” mode and downloaded all climate and land use data for respective elevation bands (500 m) existing in the basin as explained in Sect. 2.2. The irrigation and water supply demand was added separately using demand nodes (refer to the red dots) as shown in Fig. 3. Then, the reservoir data and characteristics were included in the reservoir node depicted in a green triangle on the WEAP interface. The discharge data were also added by including a Streamflow gauge node (blue colored) at the outlet of the basin. The connections from the river to the respective demands and vice versa were represented by adding transmission and return flow links respectively. A separate transmission link was added to describe the transferring of water from the reservoirs to irrigation demands. The daily variations of crop coefficients were entered into the model by referring to the crop coefficients and their cultivation periods sourced from Food and Agriculture Organisation (FAO).

Fig. 3
figure 3

Schematic diagram of Menik Ganga WEAP model

Full size image

After this step, the model was calibrated by comparing the flows at the Kataragama gauging station, varying the land use parameters of the soil moisture method. The soil moisture method is used in WEAP model to calculate the discharge and demands variations based on the inflows and outflows of the model. It is basically categorized into two sections namely, the root zone and the deep zone. This was illustrated by a conceptual bucket diagram (refer to Fig. 4). More information on this can be found in Sieber and Purkey [24]. This method transforms the climate data and other demands into runoff through the forms of overflow such as surface runoff, interflow, percolation, and base flow. These runoff forms depend on a variety of land use parameters such as runoff resistance factor, root zone conductivity, soil water capacity, preferred flow direction, deep conductivity, deep water capacity, \({\text{z}}_{1}\), and \({\text{z}}_{2}\). Streamflow is the sum of surface runoff, interflow, and baseflow [24]. These parameters were varied so that they would match with the discharge data (1990–2009) of Kataragama station.

Fig. 4
figure 4

The conceptual diagram for the soil moisture method in the WEAP model

Full size image

The WEAP model was calibrated by analyzing the statistics of coefficient of determination \(({\text{R}}^{2})\), Nash–Sutcliffe model efficiency coefficient (NSE), and Percentage Bias (PBIAS). The coefficient of determination [26], Nash–Sutcliffe model efficiency coefficient [27], and Percentage Bias [28, 29] are defined by the Eqs. 1, 2, and 3.

$${\text{R}}^{2}=\frac{{\left(\sum[{\text{X}}_{\text{i}}-\bar{\text{X}}][{\text{Y}}_{\text{i}}-\bar{\text{Y}}]\right)}^{2}}{\sum {\left({\text{X}}_{\text{i}}-\bar{\text{X}}\right)}^{2}\sum {\left({\text{Y}}_{\text{i}}-\bar{\text{Y}}\right)}^{2}}$$
(1)
$$\text{NSE}=1-\frac{\sum _{\text{i}=1}^{\text{n}}{\left({\text{X}}_{\text{i}}-{\text{Y}}_{\text{i}}\right)}^{2}}{\sum _{\text{i}=1}^{\text{n}}{\left({\text{X}}_{\text{i}}-\bar{\text{X}}\right)}^{2}}$$
(2)
$$\text{PBIAS}=\frac{{\sum }_{\text{i}=1}^{\text{n}}\left({\text{X}}_{\text{i}}-{\text{Y}}_{\text{i}}\right)}{\sum _{\text{i}=1}^{\text{n}}\left({\text{X}}_{\text{i}}\right)}$$
(3)

where \({\text{X}}_{\text{i}}, {\text{Y}}_{\text{i}},\stackrel{-}{\text{X}},\) and \(\stackrel{-}{\text{Y}}\) are observed discharge, simulated flow, mean of observed discharge and mean of simulated flow respectively. The PBIAS (expressed as a percentage) was used for the analysis of long-term bias in water balance and calculates average tendency of the modeled data to be higher or lower than the observed data.

The model was set up daily and it was run from 1990 to 2009. Another scenario was run for the period of post-construction. The climate data for the period of 2011–2019 was separately added (since Princeton climate data limits up to 2010) to the previously calibrated model and performed simulations. This scenario represents the post-construction of the Weheragala Reservoir but without adding any reservoir regulations or demands in the WEAP model. These simulated flows were considered in the IHA analysis described in Sect. 2.3.2.

2.3.2 Indicators of hydrologic alteration (IHA)

The IHA is extensively used in many studies that investigate the alterations of the river flow due to interventions. The general approaches were defined as identifying the data series for the pre- and post-influence period and calculating the indicators relevant to those periods [30]. IHA statistics are categorized into five groups: magnitude of monthly water conditions, magnitude, and duration of annual extremes, the timing of annual extremes, frequency and duration of high and low pulses, rate, and frequency variations. The hydrologic alteration indicators assessed in this research are mentioned in Table 2.

Table 2 The indicators of hydrologic alterations used in this study
Full size table

The IHA indicators were obtained for two scenarios as given below.

  1. 1.

    Kataragama observed discharge during pre-dam construction (1990–2009) and post-dam construction (2010–2019) periods.

  2. 2.

    WEAP simulated post-dam construction (2010–2019) flow with observed streamflow at Kataragama gauging station (2010–2019).

The first scenario was analysed to investigate alteration of the indicators for pre- and post-dam periods of Kataragama observed discharge. The variation of indicators due to the reservoir construction was investigated from the pre-dam construction period to the post-dam construction period. Thereafter, the IHA of simulated WEAP flow was compared with observed discharge at Kataragama for the post-dam period. Therefore, the changes of indicators can be explored with and without the Weheragala Reservoir for the period of 2010–2019. This was analysed using the percentage difference of IHA of Kataragama flows with respect to simulated flows in WEAP for post reservoir construction period according to the Eq. 4.

$$\%{\text{difference}}=\left(\frac{(\text{Simulated}-\text{Observed})}{\text{Simulated}}\right)\times 100{\%}.$$
(4)

3 Results

3.1 WEAP model calibration

Figure 5 presents the observed and modeled flows at Kataragama station. There is a good match with the discharges in observed and simulated cases. In fact, the peaks are acceptable but with some discrepancies in the year 1990 and years after 2005. These discrepancies in post-2005 might be due to the effect of the construction of Weheragala Reservoir which was started in 2005. Table 3 presents the calibrated parameters for the WEAP model.

Fig. 5
figure 5

Simulated flow vs. observed flow at Kataragama station during 1990–2009

Full size image
Table 3 The calibrated parameters for WEAP model
Full size table

3.2 IHA parameters

Figure 6 illustrates the IHA parameters calculated from the discharge at Kataragama station from 1990 to 2019. Thirty-two indicators were analyzed to study the variation of flow during pre and post Weheragala reservoir construction periods.

Fig. 6
figure 6

The IHA parameters for Kataragama gauge station (1990–2019)

Full size image

Overall, it can be observed that IHA indicators varied significantly in the period of post-construction of the Weheragala Reservoir in the Menik Ganga basin. The monthly average discharges from January to May, November, and December showed a decreasing trend from 2010. This decreasing trend gradually reduces from November to May. On the contrary, an increasing trend was observed in the monthly flows from June to October. The increasing trend is lowest in October. A significant increase in minimum flows; 1-day minimum, 3-day minimum, 7-day minimum, 30-day minimum, and 90-day minimum, is represented after 2010. This is exceptional in the year 2018 because of higher zero flows and missing data.

The baseflow indices also demonstrated a similar upsurge as they represent 7-day minimum and annual mean flow. On the other hand, the maximum flows such as 1-day maximum, 3-day maximum, 7-day maximum, 30-day maximum, and 90-day maximum drastically decreased after 2010. The Julian date of 1-day minimum mostly occurred in the months of June and July prior to 2010, and this was shifted to the months of January, February, and March during the period of 2011–2016. The Julian date of 1-day maximum appeared in the months of November, December, and January for both pre- and post-2010, though this occurred in May for 2015 and 2016. The number of counts of low and high pulses took a considerable downturn after 2010. This phenomenon was reflected in both low and high pulse durations as well. The rise and fall rates were reduced by a substantial rate post-2010 with regard to the pre-2010 period.

Figure 7 presents the IHA parameters for the simulated flows from WEAP. The IHA for simulated flows of the WEAP model was obtained for the period 2010–2019, in which regulations of the Weheragala Reservoir or demands during this period were not included in the model (refer to Fig. 7). These indicators were compared with the observed discharges at Kataragama for the same period.

Fig. 7
figure 7

The IHA parameters for the simulated flows from WEAP (2010–2019)

Full size image

The percentage difference of IHA of observed Kataragama discharge with respect to simulated flows in WEAP was obtained for each year in accordance with Eq. 4 (Sect. 2.3.2) and illustrated in Fig. 8.

Fig. 8
figure 8

The percentage difference IHA parameters between the simulated WEAP flow and Kataragama observed discharge (2010–2019)

Full size image

The blue-colored bars represent positive percentage differences whilst dark, orange-colored bars denote the negative percentage differences in Fig. 8. The highest percentage differences were observed in June, July, August, September, and November. Negative percentage differences were detected in most years of June and July and those values are extremely large. 1-day, 3-day, 7-day, 30-day, and 90-day minimums and baseflow index denoted extremely large negative percentage differences. The majority of years in the maximum flows (1-day to 90-day) showed positive differences though the years in which the highest magnitudes of percentage differences were negative. The Julian date of 1-day maximum has a positive difference compared to the observed flow (except in 2015 and 2016 with slight negative differences), which implies that the 1-day maximum occurred late in a specific year for the simulated flow than that of the Kataragama discharge. The Julian date of 1-day minimum has negative differences only in 2010, 2017, and 2019. The low pulse count has negative percentage differences from 2010 to 2013 and then it has become all positive except in 2018. The low pulse duration was observed to have a positive fraction in most of the years. The number of times that high pulses occurred to have positive alterations compared to observed discharge except in 2015 and 2018. These pulses occurred in longer durations in prior reservoir construction scenarios than in the regulated state. This pre-dam commencement scenario was observed to have higher rise and fall rates nevertheless negative alterations of fall rates in 2010, 2012, and 2013.

4 Discussion

The observed discharges at Kataragama station in Menik Ganga were analyzed for pre- and post-Weheragala Reservoir construction periods using IHA indicators. Significant alterations were portrayed for pre- (1990–2009) and post-dam (2010–2019) time periods. The months in the Maha cultivation season; November, December, January, February, and March have shown a significant decreasing trend in the post reservoir constriction period (refer to Fig. 6). Menik Ganga basin receives most rainfall during this season from the northeast monsoon; therefore, expected to generate high flows. Thus, declining in monthly flows in the post-dam period should result from mainly decreasing in high peak flows. This argument is quite obvious from the graphs of 1-day, 3-day, 7-day, 30-day, and 90-day maximums. The months of June, July, August, and September in Yala cultivation season, experienced an increasing trend of flows after 2010. The rain of this season comes from the southwest monsoon, supposed to be a drier period of the year. The low flows occurred mostly in this period of the year, so the increasing tendency of flows should outcome from the increasing of the low flows. This statement is quite noticeable from the illustrations of 1-day, 3-day, 7-day, 30-day, and 90-day minimums. It is further determined by the declination of the baseflow index (represented by a 7-day minimum) in the post-dam period. The decline of maxima and inclining of minima upshot decreasing of both high and low pulses of streamflow after 2010. The reductions of rise and fall rates in post reservoir period illustrate the less fluctuation rates that occurred after 2010.

The WEAP model, which was developed in this study, represents the hydrological model for the Menik Ganga basin simulated for the 2010–2019 period without any Weheragala Reservoir regulations and demands that prevailed in that period. This was compared with the observed Kataragama discharge (2010–2019) by calculating the percentage difference of IHA according to Eq. 4. The difference in streamflow altered in post Weheragala construction phase with regard to pre reservoir construction phase is given in Eq. 4. In other words, it illustrates the percentage of flow altered in the basin due to the reservoir regulations and demands that existed during 2010–2019.

The percentage difference of IHAs between pre- and post-dam construction scenarios (2010–2019) showed significant outcomes. The monthly flows of the Maha season showed positive percentage differences for most of the years with regard to the regulated or post-dam construction scenario. The maximum flow decreased due to reservoir construction and its regulations. This was quite obvious in the maxima of flows of 1-day, 3-day, 7-day, 30-day, and 90-day. But this argument is exceptional for some of the first years in November and 1 year (2016) in December. Anyhow, the percentage difference gradually decreases, and it becomes positive (increase in the pre-dam construction than post-construction) for last year in November. In addition, the monthly flows in the Yala season exhibited extremely high negative percentage differences (highest in June and July) over the years.

The observed discharges at Kataragama have a steep increase after 2010 in June and simulated (unregulated Weheragala conditions) showed small amounts of fluctuations of the low flows. Although the magnitude of alteration between regulated and unregulated conditions seemed little, since the unregulated Weheragala scenario displayed very low flows, the percentage variation represents higher alteration (refer to Fig. 9). This increment of low flows in the post-dam construction scenario was further depicted in the majority of minima (1-day, 3-day, 7-day, 30-day, 90-day minimums, and baseflow index).

Fig. 9
figure 9

The observed discharge at Kataragama (left), simulated WEAP modeled flow (middle) and the percentage difference between the simulated and observed streamflows (right)

Full size image

The fluctuations of the flows decreased since the reduction of the high flows and rise of low flows occurred in the post-dam construction scenario, which resulted in decreasing low and high pulses. Similarly, the rise and fall rates also decreased in the post-dam commencement scenario compared to pre reservoir construction scenario as it was indicated by a positive percentage difference in the last two graphs of Fig. 8. An exception was there for fall rates in the years 2010, 2012, and 2013.

The most sensitive IHAs in this study are the monthly flows of June. July and December and minimum flows. Overall, monthly flows of June are the most sensitive IHA in which represents the highest percentage difference.

5 Conclusions and recommendations

This research was conducted to investigate the flow alterations occurred due to the construction of the Weheragala Reservoir in the Menik Ganga basin. The Indicators of Hydrologic Alterations (IHA) were used to analyze the alteration of flow. This research indicated that the construction of Weheragala reservoir resulted in decreasing in high flows in the wetter periods (Maha season) of the year and increasing in low flows in the drier periods (Yala season) of the year. Both low and high pulses were decreased in the post-dam scenario. Furthermore, rise and fall rates also decreased in the downstream of the Menik basin due to the regulations of the Weheragala reservoir. We suggest that any possible diversion from the Weheragala reservoir is supposed to be in Maha season rather than in Yala season, which will be the reason for these results. This study proves that WEAP software can be utilized in research that analyses flow alteration of a river basin over a period of time. In addition, there was an increase in low flows during the drier season of the year. It will be interesting to analyze how these increased low flows in the Yala season would affect the demand deficits prevailing in the Menik Ganga basin. This research only limits of including demands for pre reservoir construction period. A separate study can be conducted considering the reservoir regulations and demand requirements in Menik Basin after the construction of the Weheragala Reservoir.