Plan the whole system
Water supply is a chain of interdependent capacities. An available source does not guarantee that water can be treated, transferred and delivered at the required pressure. Start with a clear service area, assessment period and level of service. Map both normal operation and the assets needed when a component is unavailable.
Five connected planning questions
- Source: How much water is available through a dry sequence, after abstraction rules and environmental requirements?
- Treatment: Can the plant reliably treat the expected raw-water quality and peak throughput?
- Storage: What volume is operationally usable above the minimum supply level?
- Transfer: Can trunk mains and pumps refill storage within their operating window?
- Distribution: Can customers receive adequate pressure and water quality across demand conditions?
Separate the modelling scales
A water resources model tracks availability, allocation and storage over time. A hydraulic network model tests flow, head, controls and distribution behaviour. Connect their assumptions, but do not treat a catchment water balance as proof of network capacity. Explore resource modelling in Source and distribution modelling in EPANET.
Turn analysis into a decision
Record the binding constraint in each scenario. An augmentation may require treatment capacity, power supply and transfer works as well as a new source. Compare a coordinated programme with demand reduction and changes to operating rules. Preserve the assumptions that connect each model.
Draw boundaries around the quantities
Distinguish source abstraction, raw-water transfer, treatment feed, net treated production, network input and customer consumption. Each boundary can have losses, returns and storage changes. A demand estimate expressed at the customer boundary is not automatically the required abstraction. Conversely, a source yield is not the same quantity as net dependable production.
For example, if customers require 8 ML/day, distribution real losses are assumed to be 10% of network input, and treatment recovers 95% of its feed, the simplified feed requirement is 8 / 0.90 / 0.95 = 9.36 ML/day. This assumes the percentages refer to those specific boundaries and that there are no other transfers or storage changes. Adding 10% and 5% to customer demand gives a different result.
Assess dependable capacity at the governing condition
A plant rated at 12 ML/day may only provide 8 ML/day during a difficult raw-water event. A transfer main that can deliver 10 ML/day over 24 hours provides less if its pump has only a 16-hour operating window. A service reservoir can buffer a temporary difference but cannot remove a persistent daily deficit.
Prepare a table for each scenario showing source availability, net treatment output, transfer capability, usable storage and distribution performance. Identify the limiting part and the dependencies that could make limits occur together. A storm-related power failure and high-turbidity event are not necessarily independent.
Use capacity evidence appropriate to each component. Source yield comes from resource analysis; process capacity requires treatability and availability evidence; hydraulic delivery requires a suitable network model. Preserve the interfaces between them, including time steps and demand assumptions.
A worked sequence from event to decision
- Set a peak-day customer requirement of 8 ML and calculate the corresponding upstream requirements using the adopted loss and recovery definitions.
- Apply an adverse source-quality event that reduces net plant output to 7 ML/day. Compare this with the required network input rather than customer use alone.
- Calculate the deficit over time and identify the genuinely usable storage available above operating limits.
- Test when another source or demand response can activate, including mobilisation, treatment and transfer constraints.
- Run the resulting network state to check whether the remaining volume can reach affected customers at the adopted service level.
The sequence makes the decision explicit. Additional raw-water availability will not solve a treatment bottleneck during the event. A treatment intervention may not solve a constrained trunk transfer. Match the option to the actual mechanism.
Link the supply plan to the learning chapters
The detailed learning path starts with physical and chemical quality and microbial barriers, then follows catchment protection, groundwater and the treatment processes. Finish with distribution integrity to reconnect treatment performance with customers.
For each project, keep a assumptions register with the value, units, source, date, confidence, owner and sensitivity. Prioritise investigations where uncertainty could change the preferred option. A precise calculation using an unsupported assumption is still an uncertain result.
Sources & further reading
- EPANET: water distribution modelling ↗US Environmental Protection Agency · 2020 release; living resource
External source · Checked 24 September 2026 - Source: integrated water resource modelling ↗eWater Toolkit · Living technical overview
External source · Checked 24 September 2026
Source findings are distinguished from editorial interpretation. Apply current local criteria and project evidence when making engineering decisions.