Long and dry: Testing the vulnerability to drought of London’s water supply system

Introduction

Drought is one of the most serious hazards facing water supply systems. Analysis of the vulnerability to drought of a water system is challenging because historical observations are often too short to capture significant drought events. To overcome this challenge, water managers have traditionally studied vulnerability to drought using stochastic stream flow models or, more recently, output from climate models. However, the uncertainties in climate models and their inability to represent long-term persistence mean that limited insight into system response can be obtained. In this paper we present a sensitivity-based approach to testing the vulnerability of water resources systems to intense and persistent drought conditions. We develop a stochastic stream flow generation technique to synthesize a large ensemble of stream flow series with different degrees of persistence and drought characteristics, which we use to test the system’s vulnerability to drought and we compare different water management options on the basis of the robustness to drought conditions.

Methodology

Hydrological droughts result from extended periods of dry conditions, where dry conditions are typically defined as periods where stream flow is below a predefined threshold. In this study we define drought using a monthly Q75 threshold (i.e., the flow is exceeded 75% of the time). Here, we introduce a stochastic framework that synthesizes stream flow time series with different levels of drought duration and deficit. The temporal dependence structure of the time series—controlling drought duration—is represented and perturbed using copulas, and very low flow occurrence—controlling drought deficit—is represented and perturbed using importance sampling. The dependence parameter of the copula function is perturbed to generate stream flow series with longer droughts.

Results and Conclusions

The 100-year long monthly stream flow series with the selected drought duration–deficit properties were used as inflow inputs to the London water supply system model. A stakeholder-defined threshold related to the percentage of usable capacity in the reservoirs is used to quantify the response of the water system to different drought conditions. Figure 1 provides a way of visualizing the sensitivity of the system to different drought scenarios and provides insight into the system’s workings under drought conditions that are beyond the historical record. The system shows a higher sensitivity to drought deficit than to drought duration, suggesting that short intense drought events are more likely to bring the system into an unsatisfactory state (i.e., below the stakeholder defined vulnerability threshold).

Figure 1. Sensitivity of the London water system to drought duration and deficit (click on the image to enlarge).

Please note these Proceedings have received limited or no review from supervisors and IIASA program directors, and the views and results expressed therein do not necessarily represent IIASA, its National Member Organizations, or other organizations supporting the work.

Last edited: 29 September 2015