Implications of storage state behaviour of village tanks in adaptation to climate change, Sri Lanka

Following Decision 1/CP.21 of the Conference of the Parties, Sri Lanka has identified Nationally Determined Contributions for adaptation to climate change. Restoration and rehabilitation of all abandoned tanks is one such contribution in the irrigation sector. The country has around 13,600 working village tanks of ancient origin. These restored tanks provide irrigation and other water needs at different degrees while a large number of tanks remain abandoned. However, regional storage behaviour of the restored tanks is not adequately understood due to the non-existence of a methodology for storage data collection and assessment. The study presents a statistical approach to assess regional storage behaviour of the tanks with data collected using five storage states method. The storage data of 573 working village tanks in southern Sri Lanka covering three years were analyzed, and revealed a high temporal variation of storage behaviour and low resilience to recover from either dryness or failure for providing irrigation water issues. However, higher time-reliability of water-existence below sluice sill level indicates their potential for providing social and environmental needs. Such assessments facilitate identifying real-time management measures and reviewing policy on restoration of similar reservoirs as an adaptation option for climate change in any region.

for 60% of the tank irrigation in the country (Nagarajan ).
In Sri Lanka, around 13,600 small scale reservoirs termed village tanks exist in working condition with periodic rehabilitation (DoAD ). The village tanks in Sri Lanka fall under the category of the minor irrigation works which are defined as those having command areas up to 80 ha (GoSL ). Around 25% of Sri Lanka's paddy production comes from minor irrigation systems, and the livelihood of people relying on these systems highly depends on the water availability of these tanks (Bronzoni ). However, they have been experiencing water scarcities affecting their agricultural performance. The command areas of most of them are more than the tanks can support, and the cropping performance realized has been far below the optimum (Dharmasena ; Somasiri ; MIWR&DM ). The situation has been further aggravated by varied catchment yields that depend on factors such as catchment shape and size, surface cover, soils, land use, rainfall intensity-duration relationship, drainage density and the rainfall season (Somasiri ). In this background, frequent water shortages and seasonal crop failures have been reported making the people depending on them more vulnerable in terms of both food security and their livelihoods (MIWR&DM ).
The effects of climate change have become an added stressor on these systems, and higher temperatures, reduced rainfall and higher evaporation with an increased frequency of extreme events affect the spatial and temporal freshwater availability in the country (Eriyagama et al. ). The projections on climate change indicate that the effects will further be increased and water and agricultural sectors will be more vulnerable (Collins et  However, the storage behaviour of restored village tanks has not been adequately understood either spatially or temporally due to the absence of data on storage behaviour. Although a water balance modelling approach has been adopted for a tank series to determine the water availability (Jayatilaka et al. ), such an approach will not facilitate generalization of the results for a region. The performance evaluation of water resource systems in terms of reliability, resilience and vulnerability can generally be carried out if measured data are available (Asefa et al. ).
However, the village tanks in Sri Lanka do not have a storage data monitoring or recording method, and the storage data collection method generally used for larger reservoirs is not practical in village tanks due to the large number and their vast distribution. Hence, the storage behaviour of village tanks on a regional basis has not been assessed. Therefore, it is essential to assess and adequately understand the spatial and temporal storage behaviour of the working village tanks to rationalize the restoration of abandoned tanks as an adaptation option to address the uncertainties and challenges associated with the effects of climate change.
This article presents the regional and temporal storage behaviour of working village tanks during three years with observed storage data of 580 such tanks in southern Sri Lanka. It also discusses the implications of regional storage behaviour of the tanks in terms of storage reliability and resilience for providing irrigation water through sluice issues and the existence of water in the tanks for other needs.

Geographical and agro-ecological scope
The geographic scope of the study area covers the Hambantota Administrative District. The District has 642 working village tanks out of 13,578 in Sri Lanka, and it is the fifth  the 75% expectancy-value of annual rainfall ranges from 1,100 to 1,400 mm, and the wet zone in the district consists of WL2a agro-ecological region where the 75% expectancyvalue of annual rainfall ranges from 2,400 to 2,800 mm. such as the storage state data and agro-ecological regions.

Data collection and analysis
Since recorded storage data of the village tanks do not exist, the five storage state assessment method was adopted in this study, and data collection was commenced in October 2014.   Tables 1 and 2 respectively.
The strength of association between the variables can be evaluated using the contingency coefficient, which ranges between 0 and 1. The zero indicates that there is no association between the two variables, and values close to 1      (Table 3).
The neutral transition probabilities of the tanks in each state range from 0.43 to 0.77, and the states dry and storage below sluice sill level show higher neutral transition probabilities. The higher neutral transition probabilities of the states dry and storage below sluice sill level imply low ability to improve storage state of the tanks when they go dry or fail to provide sluice issues. Further, the occurrence of gradual  transitions of the storage states in the tanks is higher than the rapid transitions.

Seasonal behaviour of storage state transitions
The scatter plots between Yala and Maha seasons (Figure 7 (Figure 7(b)). The respective regression equations are given in Table 4.
The gradient of the linear regression equations allows the comparison of general storage transition behaviours between two cropping seasons for management purposes.
The coefficient of regression is relatively higher in the quad-

Seasonal cropping indices of the tank systems
The seasonal cropping index is defined as the ratio of the area cropped to the command area of the tank during a season, and it varies from zero to 1. Cropping indices of  prepared for the times of water shortages to intervene with appropriate remedial measures. Further, the identification of storage behaviour for a longer period will also facilitate the planning of cropping patterns, improving water management and identification of modernization needs of the tank systems.
In addition to the managerial insights, the storage behaviour revealed by the study indicates that there could be limitations in adaptive capacity of the tanks for the projections of climate change. The smaller size and shallow depths of the tanks prevent storing more water during high-intensity rainfalls and the existence of water due to high evaporation caused by increased temperatures. Further, the occurrence of high-intensity rainfalls increases the possibility of tank breaching, and the limitations in increasing the spillway discharge capacities will aggravate the situation making these tank systems more vulnerable. Accordingly, there is potential for further reduction of storage reliability and resilience while increasing the vulnerability of the tank systems with the projections of climate change.
The extension of such evaluations to other areas will increase the data availability for the management and policy formulation. Storage state observations at short time intervals, such as weekly or at times of state changes for longer durations with beneficiary participation, will expand the scope of the analysis further. The developed methodology of the statistical approach can be applied to assess the spatial and temporal storage behaviour of regionally distributed small scale reservoirs where the scale type storage measurements are neither possible nor necessary.
Such evaluations are crucial for an adequate understanding of the real potential of small scale reservoirs in adaptation to climate change, and reviewing and updating the existing policies on the restoration of these reservoirs to make it a better option for adaptation in any region of the world.