With global temperature rise and worsening drought in some regions

water consumption has become another critical environmental concern in data center planning

alongside carbon emissions. This paper proposes a risk management planning method for data center microgrids that coordinates electricity

water

and computing resources to address economic planning under multiple uncertainties with water constraints. First

a cooling water consumption model is established via heat conduction equations

limiting water use through excess penalty pricing. Three uncertainties-data center workload

photovoltaic power

and outdoor temperature-are transformed into a univariate form for conditional value at risk (CVaR) using normalization and multipliers. A distributionally robust optimization-CVaR (DRO-CVaR) model is then constructed to measure potential risks by adjusting confidence parameters

balancing robustness and economy. Case studies demonstrate that the proposed water footprint model

by incorporating temperature impacts and penalty constraints

effectively mitigates excess water consumption

while introducing probability parameters for uncertainty distribution enables confidence level adjustment

providing a proactive risk management strategy for decision-makers.