Water-resource managers are facing unprecedented challenges in accommodating the large uncertainties associated with climate change in their planning decisions. Integration of climate risk information is a pre-requisite for water resources planning under a changing climate, yet this information is often presented outside the decision-making context and in a way which is not relevant for the decision at hand. Furthermore, there is a lack of approaches that explicitly evaluate the impact of nonstationary climate change on decision-relevant metrics and variables. This thesis describes novel methods for incorporating uncertain information on climate change in water resources decision-making and estimating climate change-related risks in water resources systems. The main hypotheses of this thesis are that: (1) shifting away from planning approaches based on abstract supply-demand balance metrics towards risk-based approaches that quantify the frequency and severity of observable outcomes of concern to water users, such as water shortages, can help decision-makers establish preferences among actions and identify cost and climate risk reduction trade-offs (2) adopting risk-based planning methods allows water managers to characterize and account for different sources of uncertainty in the water planning process and to understand their impact on outcomes of value and decisions. To test these hypotheses, this thesis presents an analytic approach for (1) incorporating nonstationary climate change projections and other uncertain factors related to demand changes into water resources decision-making, (2) understanding trade-offs between benefits of climate risk-reduction and cost of climate change adaptation, and (3) characterizing water supply vulnerability to unprecedented drought conditions. The approach is applied to London's urban water supply system located in the Thames river basin, south-east of England. Results from this thesis demonstrate how a systematic characterization of uncertainties related to future hydro-climatic conditions can help decision-makers compare and choose between a range of possible water management options and decide upon the scale and timing of implementation that meet decision-makers' risk tolerability. Additionally, results show the benefits of combining climate information with vulnerability analysis to test decisions' robustness to unprecedented drought conditions. The application of the proposed methods to the London urban water supply system suggests that the risks of exceeding reliability targets in the future will increase if no further supply or demand side actions were to be taken. Results from the case study also show that changes in demand due to population growth could have greater impacts on water security than climate change and that small reductions in climate-related risk may come at significantly higher costs. It should be stressed that the results from the case study are based on a simplified representation of London's water supply system and that they should be further tested with the full system model employed by the water utility which implements more complex operational rules.