Impacts of Land Use Changes on Hydrological Regime – A Case Study...

Impacts of Land Use Changes on Hydrological Regime – A Case Study of Randenigala and Kotmale Catchments in Sri Lanka

SHARE

Ranjith Premalal De Silva
Department of Agricultural Engineering
University of Peradeniya, Sri Lanka

Madusha Chandrasekera
Mahaweli Authority of Sri Lanka
Damsite, Polgolla, Sri Lanka

Introduction
The perennial water of river Mahaweli that flows from the central hills in the wet zone to the plains of the dry zone has a potential of over thousand megawatt of hydropower from its annual discharge of 53200 MCM. In 1968, with the help of UNDP and FAO funding, a major, multi-purpose development plan named as Mahaweli Development Programme (MDP) was initiated and expected to cover a period of 30 years (Mahaweli, 1986). Subsequently, this programme was accelerated as Accelerated Mahaweli Development Programme (AMDP) in order to complete within 6 years targeting hydropower generation of 605 MW and irrigation of 167,000 ha of agricultural land in the development process. Under this programme, seven reservoirs across the Mahaweli river system were constructed. The investment on the AMDP was Rs. 5200 million. The success of this massive investment depends on the sustainability of the water resources in the catchment areas of these reservoirs. The catchment areas for these reservoirs are collectively called as Upper Mahaweli Catchment (UMC) based on the elevation profile of the catchment.

As a result of the development programme, considerable changes in the structure and composition of the land use and land cover in the UMCA have been very obvious during the last three to four decades. This was further enhanced through the revised government policies moving towards a free market economy (De Silva, 1993). According to Tschakrt & Decurtins (1989), pressure on vacant and less suitable areas, lack of appropriate new settlement alternatives, vulnerability of ecosystem due to high intensity of rainfall and high share of steep areas, constant changes of economic and market environment, lack of integrated and coordinated land use planning are some reasons of rapid depletion of catchment resource base. Critics now argue that these changes have adversely affected the hydrological regime of these catchments resulting diminishing river flows. Further, it is also pointed out that the productivity of the agricultural land has been diminishing due to severe land degradation. Frequent landslides and earthslips damage the infrastructure and threaten the human lives. In this situation, in order to resolve present problems and avoid a future crisis, a comprehensive assessment of land use changes, its spatial distribution and its impact of hydrological regime is required and accordingly, suitable remedial methodologies should be employed for the sustainable utilization of the land and water resources of the catchment.

Objectives
The basic objective of the study was to investigate the spatio-temporal information on the status of land use/ cover changes and corresponding hydrological variability at the sub catchment level and evaluate the impacts of these changes on hydrological regime of the catchments.

Study Area
Two sub catchments of the UMCA namely Randenigala and Kotmale catchments were selected for this study as these two sub catchments represent the diverse hydrological environments and considerable land use changes during the AMDP period. . The location of these two sub catchments within UMC is shown in Figure 01. The Randenigala sub catchment mainly belongs to intermediate zone according to the classification based on rainfall while Kotmale sub catchment belongs to wet zone.

Randenigala sub catchment lies on parts of Kandy and N-Eliya districts and covers 447.84 sq. km area, which is about 14.8% of the total extent of UMC. The average annual rainfall ranges from 1250 mm to 3000 mm and the elevation ranges from 160 m to 2000 m. The capacity of the Randenigala reservoir is 860 MCM at full supply level and installed hydropower generation capacity is 122 MW. Kotmale sub catchment is located in N-Eliya district and covers 571.20 sq. km, which is about 18.8% of the total extent of UMC. The average annual rainfall ranges from 2000 mm to above 4500 mm and the elevation ranges from 600 m to 2524 at Pidurutalagala. The capacity of the Kotmale reservoir is 174 MCM and hydropower generation capacity is 134 MW.

Methodology and Results

  1. Hydrological Data
    Monthly incremental inflow data were collected for the period of 1950 to 1996 to investigate the water yield and yielding patterns. Prior to the reservoir construction (before 1985), stream flow measurements were available and after reservoir construction (after 1985), water yield data were obtained from reservoir water balance. Monthly rainfall data were also collected for a period of 40 years for 14 stations at Kotmale and 7 stations at Randenigala. These rainfall records were also separated into two periods i.e. before and after the AMDP. In addition, data for every decade were analyzed separately.

    For each period, the river flow and rainfall data series were divided into four seasons; namely North East monsoon (NEM, Dec-Feb), First Inter-monsoon (FIM, Mar-Apr), South West monsoon (SWM, May-Sep) and Second Inter-monsoon (SIM, Oct-Nov). Areal rainfall for each subcatchment for each season was calculated based on the coverage of Thyestean polygon. In addition, dry weather flow and wet weather flow were analyzed separately with corresponding areal rainfall data for these periods.

  2. Time Series Analysis
    Trends and Mean shifts of rainfall and river flow time series were examined through a comprehensive time series analysis. Linear rend analysis revealed that there are no linear trends in the hydrological time series although in few occasions, outstanding deviations from stationary was manifested in monthly time series. These deviations were observed for both periods before and after AMDP and hence, no statistical inference can be made on these deviations. Mann-Whitney test was employed to find the presence of mean shifts in the data series. Again, no consistent shift in the mean is displayed by any of the time series; however, data series of some of the decades show marginal deviations. These variations could be attributed to isolated cases of measurement errors or faulty observations, which were commonly experienced during the data collection process.
  3. Detection of Land Use Change
    The first land use inventory of Sri Lanka has been recorded in 1956-1961 period at the scale of 1:63,360 based on aerial photographs acquired in 1956 at 1:40,000 (Chandrasekera, 2000). These land use and cover data for 1956 were available from the Hunting Survey maps, which were digitized to Workstation Arc/Info version 7.12. In addition, TIN module and GRID modules of Arc/Info software was used to identify sub catchment boundaries from the digitized contours and spot heights. .

    Present land use and cover status was identified through the classification of IRS LISS II imagery acquired in 1992 and air photos taken between 1992 and 1995. The scale of the map is 1; 10,000 and the land use/ cover was identified based on 10 broad classes. These classes were determined considering the hydrological importance and hydrological response of each category.

    Because of the differences in the definition of classes in 1956 and 1992 maps, reclassification of the classes were required to bring these two maps to a comparable level. Further, due to the different mapping scales, elimination technique was used to remove areas smaller than the minimum mapping parcel size. Based on the 1956 map, minimum mapping parcel size was determined as 1 ha.

    The scale of the changes of land use and cover categories were assessed and change detection maps for each and every land use category were prepared in GIS. The generalized spatial distribution of land use changes in Randenigala and Kotmale sub catchments are shown in Figure 02 and Figure 03, respectively. The land use change statistics are given in Table 01 and Table 02 for Randenigala and Kotmale, respectively.

Table 01. Land Use Changes from1956 – 1992 in Randenigala Sub Catchment

Land Use 1956 1992 Change
Extent ha % Extent ha % %
Tea 6638.36 14.82 3933.28 8.78 -6.04
Other Perennials 2749.17 6.14 2896.9 6.46 0.32
Paddy 3756.06 8.39 3423.17 7.64 -0.75
Annual Crops 15024.9 33.55 8184.63 18.30 -15.25
Grassland 3318.91 7.42 1119.13 2.50 -4.92
Scrubland 77.95 0.17 4111.58 9.18 9.01
Forest Plantations 342.02 0.76 2221.74 4.95 4.19
Natural Forest 12511 27.94 14812.9 33.10 5.16
Urban/Unproductive 50.04 0.11 1329.87 2.95 2.84
Water Bodies 315.17 0.70 2750.71 6.14 5.44
Total 44783.9 100 44783.9 100

Table 02. Land Use Changes from1956 – 1992 in Kotmale Sub Catchment

Land Use 1956 1992 Change
Extent ha % Extent ha % %
Tea 32108.9 56.21 24955.6 43.69 -12.52
Other Perennials 1959.06 3.43 2072.32 3.63 0.20
Paddy 1164.35 2.04 806.17 1.41 -0.63
Annual Crops 191.45 0.34 2040.19 3.57 3.23
Grassland 2506.33 4.39 3628.06 6.35 1.97
Scrubland 35.33 0.06 1950.97 3.42 3.35
Forest Plantations 1960.6 3.43 4079.67 7.14 3.71
Natural Forest 16531.6 28.94 14610.9 25.58 -3.36
Urban/Unproductive 388.69 0.68 1864.8 3.26 2.58
Water Bodies 273.79 0.48 1111.35 1.95 1.47
Total 57120.1 100 57120.1 100

    Since 1956, tea, annual crops, forest scrublands and water bodies at Randenigala show considerable changes and only significant change in land use in Kotmale is tea. Land use change detection maps revealed that the tree cover has increased by 8.45% in Randenigala catchment due to the establishment of forest sanctuaries under the AMDP while 8.62% reduction of tree cover is found in Kotmale catchment as a result of the nationalization policy of the tea industry. The change of tree cover in opposite directions in these two catchments resulted an interesting approach to evaluate the impacts of tree cover change on hydrology of these catchments. In addition, 15.25% reduction in annual crop cultivation was observed at Randenigala catchment and an increase of 3.23% of annual crops was found at Kotmale. Further, a 4.92% decrease in grasslands was found at Randenigala while it was 1.97% increase at Kotmale. The direction of change of the other land use categories was the same for both catchments.
  1. Land Use changes and Hydrological Response
    The hydrological responses to land use changes were investigated through the analysis of rainfall runoff relationships. There were no obvious impacts of the change of tree cover or any other land use changes on the river flow during rainy seasons. The analysis of rainfall runoff correlation coefficient and runoff to rainfall ration do not show the effects of changing tree cover and other marginal land use changes during the runoff generation process. However, there were obvious deviations identified in the dry weather flow during the last decade in both catchments. This indicates that there is an interaction of the land use and rainfall in determining the hydrological response. The increase of the dry weather flow in Randenigala could be related to the increase of the tree cover and the reduction in canopy cover could be attributed to the decrease in dry weather flow at Kotmale. Further studies revealed that in addition to the buffering action of forest to produce delayed flows, there is a significant contribution of water from the fog interception at high altitudes where dense tree cover is present. Fog interception at very high altitudes in Horton Planes was found to be reaching 48% of the total precipitation during the dry months. This study provides conclusive evidence that the increase in tree cover would positively contribute to the water yield in the catchments in addition to its protective role of the environment.

Discussion and Conclusions
Considerable land use changes were found in both catchments due to the AMDP, which was implemented within a relatively short time. However, hydrological responses have not been adversely affected due to the changes in the land use. It can also be argued that the time span of the data series is not sufficient to manifest any real changes of the hydrological regime. Rainfall runoff relationships do not display any deviations from the stationary process. However, this study needs to be continued to include more time series data and the next stage of development should focus on GIS modeling approach in order to simulate the catchment dynamics for different land use scenarios. A comprehensive spatial modeling environment for hydrological modeling would provide information for land use planning and water resources management to ensure the sustainability of the benefits of development programme and the catchment environment.

References

  • Chandrasekera, C.M.M.M.K (2000). Investigation of Hydrological Responses to Land Use Changes in Two Sub Catchments of the Upper Mahaweli Catchment, Unpublished M.Phil Thesis, Postgraduate Institute of Agriculture, University of Peradeniya, Sri Lanka.
  • De Silva, W.P.R.P. (1993). Determination of Water Resources Sustainability of the Upper Mahaweli Catchment by Time Series Analysis, Tropical Agriculture Research, Vol 5, pp 11-14.
  • Mahaweli (1986). Projects and Programme,Ministry of Mahaweli Development, Colombo, Sri Lanka.
  • Tschakert, H & Decurtins, S. (1989). The Project Concept of Watershed Management in the Upper Mahaweli Basin. Hydrology of the Natural and Man-made Forests. Sri Lankan- German Upper Mahaweli Watershed Management Project, Kindly, Sri Lanka.