Distributed modelling of water resources and pollute transport in Malwathu

had about 5~7 times the dry season value of the total suspended solids (TSS) in the streams, and in both seasons, the modelled TSS, and were within the ranges of the previously published results. This study will be continued in the future to analyse possible hydrological and material transport related scenarios to identify best water resources management practices and to pragmatically cope with the excess fertiliser usage, an issue commonly found in most of the similar catchments.


INTRODUCTION
Water, being a vital natural resource to sustain all life forms on earth, has now become a limited resource due to the adverse impacts of various natural and anthropogenic causes. Due to the increasing population and rapid urbanisation, the demand for water has been increasing drastically. Further, the quality of the available fresh water resources has been deteriorating mainly due to pollution created by the anthropogenic activities in many rivers in developing countries, which has become a threat to the ecosystem and to human health.
Malwathu Oya basin is the second largest catchment among the river basins in Sri Lanka (3284 km 2 ), and one of the most widely used sources of water for irrigation, water supply and other diversions in the North Central stressed river basin due to over exploitation and water pollution. The basin experiences water scarcity during the dry periods and the downstream areas of the catchment was observed in 2011, 2014 and 2016 ( . feed Nachchaduwa, Tissa wewa and Basawakkulama reservoirs. The basin is distinct as it possesses a large number of tanks, from small size village tanks to comparatively fair-sized tanks, from which the rainfall ). However, the water holding capacity to aquatic weeds, eutrophication and heavy siltation. Hence, most of the tanks are facilitating only for Maha season and lack of irrigation water has become a problem in the Yala season (Gunarathna & Kumari, 2014).
Anthropogenic activities cause stresses on Malwathu et al., 2014). With respect to irrigation-related water quality of the reservoirs, Nachchaduwa has exceeded the threshold value of electrical conductivity (0.750 mScm -1 ) and salinity (0.500 gL -1 ) (Silva, 2004). Water of the nearby et al., 2014). The stream has been (mis) used as a site for waste disposal in the urban area and water becomes polluted easily (De Alwis, 2006). , and Cl -(March-April) in the paddy growing areas have been observed, along with the highest mean dissolved oxygen (DO) and turbidity values observed in the northeast monsoon season (December-February) in Malwathu Oya. The chemical fertiliser application with the paddy seasonal variations observed in the water quality et al., 2014). In addition the outlet of the main stream from the Anuradhapura city and the for total dissolved solids, pH, electrical conductivity, DO level, , and urban land use on water pollution in Malwathu Oya (Madushanka et al., 2015). In addition, the fate and fertilisers applied in paddy lands and other crop areas in the upstream catchment areas remain unresolved, and suspected to be a cause of the high prevalence of chronic kidney disease of unknown aetiology (CKDu) in the area (Munasinghe et al., 2015). Therefore, water resources management alternatives (to overcome issues occurring due to water scarcity, climate change impact, etc.) and measures to address issues related to degrading water quality (caused due to excessive use of fertilisers and agro-chemicals etc.) in the basin should be studied, while the fate of pollutants including their conveyance and spatial and temporal accumulation patterns should also be investigated by studying the dispersal and accumulation behaviour of these elements after they have been added to the crop can take account of temporal and spatial variations of all variables and parameters involved in the basic watershed. In addition, the used parameters are physically measurable. Therefore, they give a detailed and potentially more correct description of the hydrological processes in the watershed than empirical and conceptual hydrological models. Today, there exist several popular models of this type, like SHE, IHDM, SWAT, MIKE SHE (Jia et al et al., 2001b). detailed energy balance analysis in hydrological modelling physics-based models involve detailed consideration of energy transfer processes, use of sub-grid heterogeneity of land use, application of generalised Green-Ampt model to save computation time and its potential use for various scenario analyses. It has been further improved by coupling a soil erosion-transport model to introduce a particle-bound pollutant component (Rajapakse et al., 2010), and by adding simulation of multi-layered successfully applied to river basins in Japan, Korea and China (Jia et al., 2001a;2001b;2007;Rajapakse et al., 2010;Cunwen et al., 2011). Detailed descriptions of model development, hydrologic and material transport modelling procedures and input/output data, are given in the literature [Jia et al. (2001a;2001b;2005) and Rajapakse et al. (2010)].
model to the Nachchaduwa sub-catchment to assess the current status of the basin concerning the water resources management and pollute transport.

Study area
Nachchaduwa catchment was selected as the study area since it is the uppermost sub-catchment in the Malwathu have to be considered. Further, because of the location modelled.
Nachchaduwa reservoir is managed by the Department of Irrigation and it is fed by the Malwathu Oya and the feeder canal from Kala wewa. It has an irrigable area of 2833 ha and a catchment area of 598.74 km 2 .
Nachchaduwa catchment consists of several land use types, which include chena, forests, home gardens/ gardens, other cultivations, paddy, rock, scrub land, as well as water bodies. According to the digital maps prepared by the Survey Department of Sri Lanka in year 2001, the soil types of the catchment are mainly composed terrain) in the vicinity of the stream paths and reddishbrown earths and low humic gley soils everywhere else.
The entire catchment has been delineated into three sub-catchments according to the terrain and stream path distribution (Figure 1).

Collection and pre-processing of data
The model input data required for the hydrological component and the material transport component were Nachchaduwa reservoir operation data (sluice release, spill release, water issues, irrigation issues, etc.) were collected from the Department of Irrigation. Land use details and soil types of the catchment were extracted from the spatial maps prepared by the Survey Department in year 2001. Geological and soil layer details were obtained from the borehole data of the construction projects undertaken in the close proximity.
Meteorological data including rainfall, temperature, wind velocity, relative humidity and sunshine hours were collected from the Meteorological Department and checked using the hydrological and statistical data checking procedures. Daily rainfall data for the stations, Anuradhapura, Kahatagasdigiliya, Kekirawa, Maha Thiessen average daily rainfall values were calculated. data. Since hourly data were not available, it was checked whether an improvement could be made by using hourly data generated from daily data, following the Disaggregated Rainfall Method (Bennett et al., 2015). Excel Visual Basic for Applications (VBA) was used for data pre-processing and to prepare the input from year 2008 to 2011 and for validation, data from year 2012 to 2015, were used. checked against the rainfall values to identify the catchment response and impact of reservoir storage. The reservoir) with precipitation was checked. It was noted from the reservoir do not show a good correlation due to lack of reliable reservoir operation related data. The catchment response to the rainfall indicated that the measured spill and total release data are highly regulated calibrated HEC-HMS [Hydrologic Engineering Centre's Hydrologic Modelling System, developed by the United States Army Corps of Engineers (USACE)] model that was developed for this catchment was applied in the present study by incorporating the most suitable parameter values taken from previously published studies (loss method -soil moisture accounting, transform method -constant monthly method) (Hettiarachchi, from the total basin were obtained for a 1 h time interval

Water balance and yield analysis
A situation analysis was carried out by conducting a yield analysis to verify the current water scarce situation in the Nachchaduwa sub-catchment. Irrigation requirement was calculated considering the current practice in the scheme; low land paddy (135 days) for the Maha season and low land paddy (105 days) and other water balance study was carried out considering 75 % the design rainfall values for the reservoir operation study, and the Thiessen average daily rainfall values of collected data. The reservoir operation study model outputs were compared with actual operational data of an alternative cropping option was considered by using low land paddy (105 days) for Maha season and low land paddy (105 days) and OFC for the Yala season, to determine whether an improvement is achievable for the water resources management.

Water quality testing
mainly in dissolved and particulate forms (Hydrologic Engineering Research Team, 2012). Water quality samples were collected throughout the stream cascade covering both dry and wet seasons and they were tested suspended solids (TSS), turbidity, temperature and pH were collected from eight locations throughout the catchment considering reservoirs, stream segments, of the samples were preserved by adding 0.5 mL of concentrated sulfuric (H 2 SO 4 each, were collected from all the locations. The samples simultaneous determination of total nitrogen and total for the concentrations of anions (such as , and ) in ppm by using the 930 Compact IC Flex Ion Chromatography system (Metrohm AG, Switzerland). For checking the -N, the UDK 149 Automatic was used.
, and The colourimeter was used for determining by Nessler's method. All the fertiliser issued by the Anuradhapura DSD were assumed to be applied for paddy. Missing data were and Maha seasons from 2011 to 2016), and the dataset was developed from 2008 Yala to 2016 Yala season. Urea 45 % phosphorus. It was assumed that 30 % and 10 % of the fertiliser amount applied for paddy is equal to the fertiliser amounts applied for other crops and homesteads, month for all the crops were calculated, according to the application patterns of fertilisers relevant to the current practices in the catchment. The applied amounts were compared with the required amounts of fertilisers, which were found in literature (Table 1).  parameters, sub-catchment delineation, land use, meteorological data, soil parameters, river channel element details, aquifer details, initial groundwater levels, etc. have also been prepared for the model runs.

Parameter sensitivity analysis
by varying the parameter values within a pre-determined with the previously published water quality data in the same basin.  operation study model outputs with the actual operational data -alternative cropping pattern

Water balance/yield analysis in Nachchaduwa subcatchment
The results of yield analyses are presented in Figure 3.
the catchment, especially during the dry season extending from April to September. As seen from Figure 3(c), the gap between the demand and supply of water has been that an improvement is achievable for the water resources management. However, even with the alternative cropping option, still there is water scarcity prevalent in the catchment.
According to the water quality test results of the samples that were collected throughout the stream cascade in the dry season (Yala season), the single factor ANOVA test of -N and The -N and -N concentrations were below the minimum measurable limit of the apparatus.

Fertiliser input and dispersal analysis
For paddy, other crops and homesteads, the monthly each month have been calculated (considering the data for years 2008 Yala to 2016 Yala) and compared with (kg/ha) in almost all months for all three types of crops that were considered have exceeded the plant required amounts.
The input fertiliser amounts and the uptake fertiliser amounts from the crops were calculated and the variation of total input and total uptake (from the crops) of N in each year (for the years 2008-2011 are shown), in each in Table 3.

WEP model results
values of water and heat balance as well as water quality and material transport results for each grid, as outputs. The preliminary results pertaining to the hydrological and material transport processes have been presented in this paper.     The temporal variation of TSS showed a correlation with the Thiessen average rainfall values, with the peaks and troughs of the TSS graph corresponding to the peaks and troughs of the rainfall graph (Figure 7), implying that the rainfall would induce a washout of the solids, and hence adding them into the streams. The mean ± standard deviation, minimum and maximum values of TSS were: 0.90 ± 4.56 mg/L, 0.11 mg/L and 87.40 mg/L in the calibration dry seasons; 4.79 ± 14.88 mg/L, 0.13 mg/L and 148.57 mg/L in the calibration wet seasons; 0.76 ± 3.45 mg/L, 0.09 mg/L and 49.94 mg/L in the validation dry seasons; 5.62 ± 20.38 mg/L, 0.10 mg/L and 304.57 mg/L in the validation wet seasons, respectively. Therefore, on average, the wet season has dry season value of the TSS in the streams.
Several water quality studies have been conducted focusing on the Nachchaduwa catchment (Wijesundara et al et al., 2014). The temporal previous studies.
water quality testing period of this study are illustrated model results for the entire duration of calibration and validation periods, with the entire duration of the three published results.
The three published studies have only measured the water quality as spontaneous measurements, for a shorter duration (only for one-year period), and the water quality parameters under all weather conditions have not been considered (as the sampling has been done once a month results shown are the average of a longer time period (hourly values of all days for a period of four years), and therefore represent a wider and a more reasonable range of values of water quality parameters in the streams. However, it is evident from Figures 8 and 9, that the results.
Further, for the N components for both the calibration values showed the best match with the published results. The published values are given for , therefore, when comparing, the DN component conditions ranges from 0.0023 mg/L to 3.3177 mg/L in the calibration period and in the validation period it ranges from 0.00349 mg/L to 3.6347 mg/L. In the published values, it ranges from 1.05 mg/L to 12.00 et al concentration in Malwathu Oya river has been found to be 15 mg/L in the month of August, by Zoysa and Weerasinghe (2016).
below the maximum value recorded in the basin and below the threshold value of 10 ppm for drinking water (WHO, 2011). the best match with the published results. This could be solids, hence most of the particulate nutrients which were adsorbed to the sediments get washed away, adding them to the waterways (Wijesundara et al., 2012). The The excess fertiliser dispersal (wash out) paths and accumulation hotspots (the locations with the highest that would need urgent attention when implementing best management practices for the fertiliser usage. Further, these hotspots could be compared with the CKDu prone areas in the catchment, to check whether they are spatially correlated.
varied by ± 25 % to check the model sensitivity.
deposit (S max ) and amount of initial deposit (S ini ). It was max Figure 9: validation period; (c) phosphorus components -calibration period; (d) phosphorus components -validation period allowable riverbed deposit (S max ), amount of initial riverbed deposit (S ini ), suspended condition phosphorus transport in forest and urban area (nonpointsource.csv Sensitivity analysis has shown the response it could be assumed that the pollutant loading due to the point sources in this catchment are less dominant than the non-point sources of pollutants since this catchment is not highly urbanised and developed, the results of the sensitivity analysis have shown that the parameters governing both point sources as well as non-point sources

CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE STUDIES
The present study incorporated a detailed modelling approach to the Nachchaduwa sub-catchment to study the water resources management as well as pollutant transport of the river. Only the present condition has been analysed up to now and this study will be continued in the future to analyse possible hydrological and material transport related scenarios in the catchment. The catchment response to the rainfall indicated that the measured spill and total release data are highly regulated needs to be analysed as an ungauged basin with regulated will set the baseline for studying ungauged basins with were reasonably matching with the HEC-HMS model of 0.6875), establishing the suitability of applying the scarcity in the catchment, especially during the dry season (April-September), even after implementing the proposed alternative crop pattern. However, an improvement in water resources management is achievable by choosing alternative crop patterns for this catchment.
A distinct variation of the measured dry and wet season water quality parameters was observed in the amounts (kg/ha) of fertilisers in almost all months for all three types of crops that were considered have to seven times the dry season value of the TSS in the streams, establishing that the high washout of nutrients in the wet season causes an increase in the concentrations of nutrients in waterways. The water quality values of time period, and therefore represent a wider and a more reasonable range of values of water quality parameters in the streams. Nevertheless, in both Yala and Maha and are within the range of the previously published results.
results have a reasonable response to all the parameters to the model results. Although it could be assumed that the pollutant loading due to the point sources in sources of pollutants since this catchment is not a highly urbanised and developed area, the results of the sensitivity analysis have shown that the parameters governing both point sources as well as non-point sources contribute to identify the most critical areas and time periods of the pollutant accumulation, which would be addressed in future studies of this research. This information could be water resources and pollutant transport management of the catchment. and for recommending the best management practices and for coping with the excess fertiliser/agrochemical usage of this catchment in a more pragmatic manner. river basins in Japan, Korea and China in similar studies, the results of this study could also be generalised and are applicable to any similar ungauged basin in this region