Forecasts do not tell a utility what to do. They describe what might arrive, with timing and intensity that narrow as an event approaches and remain uncertain until it does. The operational value of a forecast depends on whether the thresholds for action were set before the weather turned, because the hours between a warning and a runoff event are rarely enough to write a procedure from scratch.
The current Southwest example shows the pattern. The Guardian, in a report by Coral Murphy Marcos and agencies dated October 25, 2026, described more inclement weather for a rain-soaked Four Corners region following evacuations and rescues. The report states that "Another round of heavy showers and thunderstorms were expected to hit the south-west US on Sunday." A utility in that position is not facing a single event. It is facing a sequence in which runoff, turbidity, and power reliability may all change on short notice, and each change calls for a different response.
Turn the forecast into named triggers
A useful trigger list links a measurable condition to a defined action. Rising raw water turbidity can trigger more frequent jar testing and a review of coagulant dose. A forecast of heavy rain over the source watershed can trigger a check on intake position, reservoir draw strategy, and whether any pretreatment step will be bypassed or added. Each trigger should name the observation that confirms it, the person who acts, and the record that shows the action was taken. Without that last item, a sound response may be impossible to reconstruct later.
Chemical and fuel stock carry tradeoffs
Pre-buying treatment chemicals is the most visible response and not always the cheapest. Liquid products have supplier-stated storage limits, and stock held in a warm or poorly ventilated room can degrade before it is used. Overstocking also ties up working capital and can fail when delivery routes are cut by the same weather that drove demand. Relying instead on a delivery promise written in a fair-weather contract shifts the risk to the supplier and to whatever terms govern supply interruptions. Neither option is free. The right stock level depends on the plant's historic usage during storm events, which some utilities may not have recorded in a form they can retrieve quickly.
Standby generators raise a similar question. Fuel stored for long periods can degrade, and generator runtime at the plant's actual load should come from a current load study rather than the nameplate rating. A start test run before the storm shows whether the unit responds, but it does not show how long the fuel will last under sustained demand.
Sampling and notification thresholds
The sampling plan should change before the storm, not after the first complaint. Increased sampling during an event shows whether treatment is keeping pace, but laboratory results arrive after turnaround time has passed. A notification decision may therefore need to rest on field indicators and the trend in online instruments. Set those thresholds with the state regulator's requirements in view, because a decision made on a field reading may need to be documented and confirmed afterward. The forecast cannot settle that question, and it should not be treated as if it could.
What the forecast cannot tell you
Forecast confidence generally falls as lead time grows, and rainfall totals across a watershed can differ sharply from totals at a single gauge. A forecast of more storms does not establish which tributary will carry the load or how quickly a reservoir will respond. Treat the forecast as a reason to check thresholds, not as a substitute for observed data. The thresholds worth writing are the ones that still hold when the forecast proves wrong in either direction.