The development of a risk of failure evaluation tool for small dams in Mzingwane Catchment, Zimbabwe

Abstract

Small dams in Mzingwane Catchment in southern Zimbabwe are mostly in poor physical condition mainly due to lack of resources for repair and maintenance. Most of these dams are likely to fail thereby adversely affecting water availability and livelihoods in the area. To assist those involved in maintenance, repair and rehabilitation of small dams in resource poor and data sparse areas such as Mzingwane Catchment, a non-probabilistic but numerical risk of failure evaluation tool was developed. The tool helps to systematically, and objectively classify risk of failure of small dams, hence assist in the ranking of dams to prioritise and attend to first. This is important where resources are limited. The tool makes use of factors such as seepage, erosion and others that are traditionally used to assess condition of dams. In the development of the tool, an assessment of the physical condition of 44 (1 medium sized and 43 small dams) dams was done and the factors were identified and listed according to guidelines for design and maintenance of small dams. The description of the extent to which the factors affect the physical condition of small dams was then standardised. This was mainly guided by standard based and risk-based approaches to dam safety evaluation. Cause-effect diagrams were used to determine the stage at which each factor is involved in contributing to dam failure. Weights were then allocated to each factor depending on its stage or level in the process of causing dam failure. Scores were allocated to each factor based on its description and weight. Small dams design and maintenance guidelines were also used to guide the ranking and weighting of the factors. The tool was used to classify 10 dams. The risk of failure was low for one dam, moderate for one, high for four and very high for four dams, two of which had already failed. It was concluded that the tool could be used to rank the risk of failure of small dams in semi-arid areas. The tool needs to be tested in a wider geographical area to improve its usefulness.

Introduction

Worldwide much attention is given to the prevention of failure of medium sized to large dams with little attention being paid to small dams (Orr et al., 2004, Pisaniello et al., 2006). This is because dam failure is normally viewed in the context of the risk that is posed to life and property downstream of dams (Rettermeier et al., 2001). Consequently medium to large sized dams are considered to pose higher risk if they fail as compared to small dams, which do not normally result in huge losses of life and property. In this paper dam failure refers to the inability of a dam to hold water due to breaching or siltation. Risk of failure refers to the possibility of failure. Physical condition refers to the state of the components of a dam such as spillways and dam walls in terms of the presence of deficiencies such as cracks, erosion and seepage. The Water Act of Zimbabwe defines a small dam as a structure which has vertical height of more than 8 m but less than 15 m measured from the non-overflow crest of the wall to the lowest point on the downstream face of such wall, or is capable of storing more than 5000 m³ but less than 1,000,000 m³ of water at full supply level (Government of Zimbabwe, 1998).

Small dams however are very important for ensuring water availability and sustenance of livelihoods for rural communities (Vermillion and Al-Shaybani, 2004, Mushtaq et al., 2007). This is especially so in semi-arid and water scarce areas such as Mzingwane Catchment in Zimbabwe (Sithole and Senzanje, 2006). In Zimbabwe, like in the rest of the world, inadequate care is being given to small dams. A large number of these are in poor physical condition (CARE Zimbabwe, 2002). Without appropriate design, construction and maintenance, small dams eventually fail, depriving the communities and animals of the much-needed water that is vital for sustenance of life. Lack of resources was noted through a socio economic survey by Sithole and Senzanje (2006) as the major reason for small dams not being properly maintained in Mzingwane Catchment. Thus, there is need to prioritise the use of the limited resources are available for dam maintenance. However there is no simple tool that would help achieve this through evaluating and quantifying risk of failure of small dams in the study area. The study was carried out to explore issues that lead to failure of small dams, and contribute towards solving this problem by developing a simple and easy to use numerical risk of failure evaluation tool. The risk of failure evaluation tool brings more objectivity in selecting the dams which are most affected by lack of care – systematically and more objectively classifying the risk of failure of small dams, hence assisting in the selection or prioritisation of dams to attend to first for repairs or maintenance.

Major causes of failure of earth dams worldwide include construction flaws, seepage, piping, overtopping and siltation. Lack of maintenance, as mentioned earlier, is also an important factor causing dam failure (Orr et al., 2004, Pisaniello et al., 2006). Dam safety management programs are usually put in place to try and prevent dam failure. There are two basic approaches to dam safety evaluation; the conventional standards based approach and the risk-based approach (Bowles et al., 1997, Rettermeier et al., 2001). The standards based approaches uses design, construction, inspection and maintenance standards. The evaluations are carried out on a particular dam so as to keep the dam in a condition prescribed, or to check whether the dam is performing according to standards. The approach does not attempt to quantify the risk of failure of a dam (Bowles et al., 1997). The risk of failure defines acceptable safety condition defined using information provided from a risk assessment. Thus this approach quantifies the risk of failure of dams (Bowles et al., 1997). Risk-based approaches are focused towards prevention of loss of life and property. They also rely on quantitative data from regular monitoring and surveillance which is not possible for the small dams in the study area which do not have installed monitoring equipment and regular inspection and maintenance programs in place due to lack of resources. A risk of failure evaluation tool suitable for dams in the study area was therefore developed.

The main aim of the study was to develop a risk of failure evaluation tool for small dams in Mzingwane Catchment, Zimbabwe. In order to develop this tool, qualitative and quantitative assessments of the physical condition of small dams in the study area were done. Factors leading to the observed physical condition of the dams in the study were also identified. The physical condition of the small dams was compared to the recommended small dams design and maintenance guidelines from Zimbabwe and other countries. Data from the assessments of the physical condition of small dams in the study area and literature on risk of failure evaluation were assessed and used to develop the risk of failure evaluation tool for small dams in the study area.

Even though the concept of the Maximum Probable Flood (MPF) was used in assessing the spillway capacity of the small dams, the effect to dam failure of extreme inflows resulting from unexpected extreme rainfall events with a return period exceeding 100 years was not considered in this study. Addressing the limitations of the MPF in predicting extreme flood events resulting from effects of climatic change for example, was also not considered in this study. This was due to the limited availability of the necessary data in the study area as well as the need to make the tool simple. Also, the ranking and weighting of the failure causing factors did not involve the use of probability methods. The overall risk of failure score of an assessed dam is therefore a number that is not probabilistic. The developed tool is therefore considered non-probabilistic but numerical and is meant for use in data sparse locations.