Every year, communities, farms, utilities, insurers, and governments face the same question: how much precipitation will we get this year? The answers they receive are rarely useful. Most seasonal precipitation forecasts offer little more than a coin flip — equal chances of above, below, or near normal. That is not a forecast. It is an acknowledgment that we don’t know.

ALICE changes that.

‍What ALICE Does

ALICE — Atmospheric Linkages Informing Community Expectations — is a proprietary seasonal precipitation forecasting system developed by Weather Tools, Inc. It forecasts total precipitation for a twelve-month water year (WY), issued just one month into that water year. Each forecast is expressed as a specific twenty-percentage-point range relative to normal — for example, 80–100% of normal, or 120–140% of normal. Forecasts are issued once and never revised.

That last point matters more than it might seem. Conventional seasonal outlooks update monthly and frequently shift direction. An operator who commits resources based on an October outlook may find the January revision pointing the other way. ALICE eliminates that variable. Once a forecast is issued, it stands — giving decision-makers the stability to commit to a plan rather than react to a moving target.

Why It Is Different

Most seasonal precipitation outlooks rely on historical climate patterns or large-scale climate indices like El Niño and La Niña. These approaches assume that what happened recently will continue, or that a single ocean temperature index can predict complex regional precipitation. In practice, agencies frequently issue “Equal Chances” outlooks — communicating no directional signal at all.

ALICE takes a fundamentally different approach. Rather than asking “what has the climate done before under similar conditions?” ALICE asks “what is the atmosphere telling us right now?” The atmosphere integrates everything upstream — ocean heat, circulation patterns, energy balance — into a single observable state. ALICE reads that state directly, at the ground level, using standard meteorological instruments. The result is a forecast grounded in what the atmosphere has already committed to, not in what historical analogs suggest might happen.

ALICE reads the atmosphere after its state is committed but before that state has expressed as weather. This is the window between cause and effect.

What this means in practice: when ALICE forecasts 60–80% of normal, you can plan for a below-normal year while the agencies are still saying “Equal Chances” — and you know that number won’t change on you next month.

The Track Record

Credibility in forecasting is earned one verified prediction at a time. ALICE’s track record now includes 25 forecast-verification pairs across three continents: California, Australia, and Israel. The same analytical framework — with no geographic adjustment — has produced verified forecasts across 8 climate regimes.

Of the 25 forecasts issued within ALICE’s design envelope, 20 were direct hits, 3 were near-misses (directionally correct but boundary-adjacent), and one fell just outside the expected range, with below-normal conditions expected but near-normal precipitation observed. That is 96% directional accuracy. One additional forecast was issued deliberately at a site outside the design envelope — Merimbula, Australia, where bimodal precipitation makes the signal unreadable — confirming the domain boundary precisely where the theory predicted ALICE would not work.

California: Ten Consecutive Water Years

The most complete record is California’s Core Water Supply region, where ALICE technology provided 10 consecutive water year forecasts. Observed outcomes have ranged from 45% of normal (exceptional drought) to 171% of normal (exceptional flood) — all correctly characterized in advance. Of those years, eight fell squarely within the forecasted range, one missed by just 6 percentage points, and the current year is pending.

‍

‍

During the same period, the National Weather Service issued “Equal Chances” outlooks in eight of ten years, communicating no directional signal. In the remaining two years it indicated a probability of below normal with no magnitude guidance. ALICE provided a specific twenty-percentage-point magnitude range in every year.
‍

The Signature Forecast: Water Year 2022–23

California experienced three consecutive drought years from 2019-2022. In WY 2022-23 the National Weather Service issued “Equal Chances” for Northern California and “Below Normal” for Southern California. The prevailing expectation across the water management and forecasting communities was continuation of the drought into a fourth year.

ALICE projected 20–40% above normal — a contrarian call with no institutional support from any conventional forecasting agency.

By late January 2023, California had received only about 40% of its normal water year precipitation, with conditions not yet aligning with the wet forecast. 

ALICE held. With late-season storms, California ultimately received 37.7% above normal precipitation — squarely within ALICE’s forecast range. The ALICE forecast was issued one month into the water year and was never adjusted. 

Australia: The Global Test Tube

Australia served as the test of whether ALICE’s methodology could work beyond California. In a single observational year (2016–17), ALICE was tested at eight sites spanning six distinct climate regimes across the continent.  Seven of those eight sites produced direct hits using the same framework verified in California; Merimbula, the sole bimodal site in the test network, confirmed the domain boundary, as ALICE correctly failed to produce a reliable signal where the theory predicted it would not work.

To further evaluate performance in a Mediterranean climate, Adelaide was tested across five consecutive years (WY 2017–21). Across those years, there was one directional miss—in WY 2020–21, ALICE predicted 60–80% of normal, while observed precipitation fell within 90–110%. This remains the only directional miss in ALICE’s entire design-envelope record.

The Australian record demonstrated something important: ALICE is not a California-specific tool. The same atmospheric reading that works in the Pacific-Rim Mediterranean climate of the Central Valley also holds across tropical monsoon (Darwin, Broome, Cairns), arid interior (Alice Springs), subtropical (Brisbane), and classical Mediterranean regimes (Perth, Adelaide).

Israel: Cross-Continental Training

The Israeli verification tested an additional dimension of ALICE’s portability. At the Hahoresh station in the central Levant, ALICE produced four verified forecasts in the Eastern Mediterranean climate regime — a regime not represented in the California or Australian test networks.

What made the Israeli result particularly notable was how it was achieved: the forecasts were trained using data features from California. Cross-continental model training worked because both regions share a sufficiently similar Mediterranean climate with a winter wet season and a summer dry season.  The atmospheric signal that ALICE reads is governed by the same underlying physics in both regions, even though how that signal shows up in weather patterns differs. Four hits in four years suggests that the signal ALICE measures is not regionally specific — it is globally coherent.

Where ALICE Works

ALICE has been verified across 8 climate regimes on 3 continents: Pacific-Rim Mediterranean, Eastern Mediterranean, Mediterranean, Tropical Monsoon, Tropical Wet-Dry, Tropical Wet, Arid/Semi-Arid, and Subtropical Seasonal. Australia served as the test tube — six of these eight regimes were verified on a single continent in a single observational year. The Eastern Mediterranean regime was verified through the Israeli record.

Each new site requires local calibration. The atmosphere carries a site-specific signature shaped by local topography, proximity to moisture sources, and surrounding land use. Cross-regional training can accelerate deployment when climates overlap — as it did between California and Israel — but each site ultimately builds its own observational history. Where existing weather stations have adequate historical records, an initial skill assessment can begin immediately with no new infrastructure required.

How ALICE Is Used

ALICE delivers seasonal precipitation intelligence to organizations whose decisions depend on knowing how much water is coming. The applications span industries and scales, but they share a common thread: the cost of being wrong about water far exceeds the cost of the forecast.

Water Resource Management. Reservoir operators, water utilities, and state agencies manage systems designed for average conditions in a climate that rarely delivers average years. Knowing whether the coming season will bring 60% or 140% of normal precipitation — with eleven months still ahead — transforms every decision in the chain: when to release stored water, how much to allocate, whether to activate conservation protocols, and how to communicate with the public.

Agriculture and Food Security. For farmers, the coming season’s rainfall determines planting decisions, irrigation budgets, and crop selection. For governments, it determines food security planning across entire regions. In areas where hundreds of millions of people depend on seasonal monsoon or winter rainfall, an accurate precipitation forecast months in advance is the difference between preparedness and crisis.

Retail Sourcing. Major grocery chains and restaurant groups source heavily from regions where seasonal water stress directly affects produce availability, quality, and pricing. An early signal about water-year trajectory helps procurement teams diversify suppliers, lock in forward pricing, or adjust sourcing strategies before the market tightens. ALICE’s early issuance aligns with the forward procurement cycles that drive supply chain decisions months before product reaches the shelf.

Insurance and Reinsurance. The parametric insurance industry prices risk based on measurable weather outcomes. Today, those policies are priced using historical climatology because no reliable seasonal forecast exists at the specificity required for trigger design. ALICE changes that equation. With a verified precipitation forecast issued before the coverage period begins, underwriters can price risk with forward-looking information rather than backward-looking averages. 

Energy Markets and Hydropower. Hydropower generation is directly governed by precipitation. A seasonal precipitation forecast with eleven months of lead time provides energy traders, utilities, and grid operators with the information they need to position ahead of the market. This insight extends to natural gas demand, renewable energy planning, and commodity trading strategies tied to weather outcomes.

Government and Treaty Negotiations. Seasonal precipitation drives the highest-stakes water policy decisions on earth: international river basin treaties, flood control operations, and drought emergency declarations. When one party in a water-sharing negotiation has a reliable precipitation forecast and the other does not, the informed party holds a decisive advantage.

See If ALICE Can Help

If your organization makes decisions shaped by seasonal precipitation — whether you manage water systems, price insurance products, operate agricultural supply chains, trade energy commodities, or negotiate water policy — ALICE may be able to give you information that no one else can provide.

The first step is a conversation. We will assess whether your region falls within ALICE’s verified climate regimes, whether adequate historical weather data is available, and what a deployment timeline would look like. In many cases, an initial skill assessment can begin within weeks using existing station data.

‍