← Back to catalogue
Research draft

weather forecasting

vr.tr.weather-forecasting · ACT.ACT

Enable an agent to identify a weather forecasting activity, assess the fitness and uncertainty of its predictions, and determine how those predictions may be issued, revised or used.

Thing Registry Activities and processes

Research draft, second pass

A second pass drafted this model: the structure a model of this thing needs, and what is known about it in the world. The line under this one says how the second half was obtained - researched against sources, or recalled without web access, in which case nothing here was read anywhere and every claim is a lead to verify. Unreviewed either way.

recalled by Codex without web access - no source was read

Researched by: Codex

Purpose and description

Enable an agent to identify a weather forecasting activity, assess the fitness and uncertainty of its predictions, and determine how those predictions may be issued, revised or used.

Weather forecasting is the activity of estimating future atmospheric conditions for specified places and times using observations, physical understanding, numerical or statistical models, and expert interpretation.

It can be Specify a forecast request with atmospheric targets, location, valid period and decision-relevant thresholds.; Check whether observations and guidance are sufficiently current and applicable to generate or update a forecast.; Compare forecast methods or providers at matched targets, lead times and verification conditions.; Express predictions as values, probabilities or scenarios while preserving their uncertainty and provenance.; Issue, revise, supersede or withdraw a forecast under the applicable publication authority.; Verify archived forecasts and identify conditions requiring recalibration or method review..

Distinguishing features

A forecast predicts conditions at a valid time later than its information cutoff; an observation reports conditions already measured.

An operational forecast must be available before its predicted conditions occur; a retrospective forecast recreates an earlier prediction setting and must be identified separately.

Weather forecasting targets atmospheric conditions for specified future periods; climate projection targets longer-term climate behaviour under stated assumptions.

A forecast estimates conditions or event probabilities; a warning additionally applies an issuing authority's hazard criteria and communication responsibilities.

A numerical model run becomes usable forecast guidance only when its initialization, target times, variables and interpretation are identifiable.

Scope

+ Forecast targets, geographic coverage, issue times, valid times and intended uses

+ Observational inputs, initial conditions and their suitability for the forecast

+ Prediction methods, model guidance and forecaster interpretation

+ Forecast uncertainty, calibration and limits of predictability

+ Forecast issuance, updates, provenance and communication

+ Verification against observations and relevant reference forecasts

- Atmospheric processes themselves, which belong to weather and atmospheric-system models

- Long-term climate projections and climate-change scenarios

- Design and maintenance of observing instruments and computing infrastructure

- Meteorology as a discipline and meteorologist as an occupation

- Emergency response decisions and protective actions taken using forecasts

- Hydrological, agricultural or economic impact predictions beyond the atmospheric forecast

Characteristics

Forecast target
Atmospheric variable or event, including units, threshold and temporal aggregation where applicable Distinguishes predictions such as instantaneous temperature, accumulated rainfall and probability of a thunderstorm.
Geographic support
Point, route, grid cell, area or atmospheric volume; coordinates and elevation where relevant Determines where the prediction applies and which observations can fairly verify it.
Forecast timing
Information cutoff, initialization time, issue time and valid instant or interval, each with time zone Separates what was known when the prediction was made from when it was delivered and when it applies.
Forecast lead time
Minutes, hours or days, with the reference time explicitly named Supports meaningful comparisons of forecast difficulty and usefulness.
Spatial and temporal resolution
Metres or kilometres; minutes or hours; distinguish computational resolution from published detail Helps identify unsupported local precision and mismatches with the user's decision scale.
Prediction method
Numerical prediction, statistical prediction, machine learning, extrapolation, human synthesis or a documented combination Identifies how evidence becomes a forecast and which limitations require examination.
Uncertainty representation
Point estimate, quantiles, interval, event probability, ensemble scenarios or qualitative confidence Determines which interpretations and downstream decisions the forecast supports.
Input readiness
Ready, degraded, blocked or unknown, under declared coverage, latency and quality criteria Makes missing or stale observations visible before forecasts are issued.
Forecast release status
Draft, issued, superseded, withdrawn or expired Prevents users and agents from treating obsolete predictions as current guidance.
Verification performance
Named metric and value, with target, lead time, evaluation period, sample size and reference forecast Makes performance claims interpretable and limits their transfer to untested conditions.

Also called

nowcastingSpecial weather statementSPECIspecial weather reporttropical cyclone forecastingEnsemble forecastingSailing weather predictionTerminal aerodrome forecastPrecipitation Nowcast

Where this came from

wikidata · CC0 1.0

Source repository Registry entry point

Drafted structure

Bundle to layer to finding to question, as the second pass will find it: 7 bundles · 13 layers · 20 findings · 32 questions.

Forecast target and use Defines the atmospheric prediction being requested and the setting in which it will be interpreted.

A forecast cannot be assessed without knowing exactly what, where and when it predicts.

Atmospheric target

Identifies variables, events and their spatial and temporal support.

Target definition

Record the predicted quantity or event with enough detail to distinguish superficially similar forecast products.

  1. Does the forecast concern a value, accumulation, extreme or threshold event, and what units and aggregation define it? definition
  2. Does the target apply to a station location, an entire area, any point within an area or a specified atmospheric layer? boundary

Forecast time and decision

Connects prediction timing to the intended use.

Usable forecast window

Record when the forecast becomes available, when it applies and how much advance notice its user requires.

  1. What are the information cutoff, issue time and valid period, and which reference time defines lead time? measurement
  2. What decision must this forecast support, and by what deadline must it be available to affect that decision? action
Observations and initial state Captures the evidence used to characterize current atmospheric conditions.

Forecast fitness depends partly on the availability, quality and representativeness of its starting evidence.

Observing inputs

Records the observing sources actually available to the forecasting activity.

Observation coverage and quality

Record input origin, observation time, arrival time, quality checks and consequential gaps.

  1. Which station, radar, satellite, upper-air or other observations were actually available before the forecast information cutoff? provenance
  2. Which missing, delayed or suspect observations affect coverage of the target location or developing weather system? measurement

Initial condition construction

Identifies how observations and prior estimates establish the starting atmospheric state.

Analysis and boundary inputs

Record the analysis or other starting representation, including external boundary inputs where the method requires them.

  1. Which analysis, observation sequence or other starting state initialized the prediction, and how was it constructed? provenance
  2. Do inconsistencies or stale initial or boundary inputs require delaying the run, using a fallback or flagging degraded guidance? action
Prediction and interpretation Describes how atmospheric evidence is transformed into forecast guidance and a final prediction.

The agent must distinguish raw method output from the adjustments and judgments embodied in an issued forecast.

Predictive method

Identifies the prediction system and its applicability to the requested target.

Method configuration and fit

Record the method version, configuration and evidence of suitability for the phenomenon and lead time.

  1. Which numerical, statistical, machine-learning or extrapolation system produced the guidance, using which version and configuration? provenance
  2. What evidence supports this method's use for the requested weather phenomenon, terrain, spatial scale and lead time? boundary

Guidance to forecast

Tracks blending, correction, localization and human interpretation.

Forecast transformations

Record how the issued forecast differs from its input guidance and why those differences were introduced.

  1. Which bias corrections, downscaling steps, guidance blends or forecaster adjustments changed the original output? provenance
  2. When guidance disagrees about storm track, onset time or intensity, how is the final forecast selected or expressed as alternatives? action
Uncertainty and predictability Makes the forecast's range of possible outcomes and limits of confidence explicit.

An agent needs to distinguish a useful estimate from an unsupported claim of precision.

Probability and scenarios

Defines the interpretation and construction of probabilistic or scenario-based guidance.

Uncertainty semantics

Record precisely what a probability, interval or scenario represents and how it was derived.

  1. For each probability or interval, what event, location, valid period and coverage interpretation does it describe? definition
  2. Was uncertainty derived from ensemble members, calibrated statistical output, historical errors or expert judgment? provenance

Limits of confidence

Identifies unresolved timing, location and intensity uncertainty relevant to use.

Decision-relevant uncertainty

Record which uncertainties can change the user's interpretation and what evidence could reduce them.

  1. How much uncertainty concerns event occurrence, location, timing and intensity, and which components are quantified? measurement
  2. Which new observations or forecast cycles could justify narrowing the range, changing confidence or revising the preferred scenario? action
Issuance and revision Controls how forecast products become available and remain interpretable through updates.

Even technically sound guidance can mislead if its authority, age, scope or replacement status is unclear.

Forecast product

Connects scientific content to a traceable issued product.

Release identity and meaning

Record the issuer, product version, issue time and wording needed to interpret the prediction correctly.

  1. Who issued this forecast, which release does it belong to and which guidance supports it? provenance
  2. Does the product communicate a forecast, a hazard watch or an official warning, and whose criteria establish that distinction? boundary

Update lifecycle

Tracks scheduled cycles and event-driven corrections.

Revision and supersession

Record when changing evidence requires an update and how recipients identify the current forecast.

  1. What observed developments, guidance changes or discovered errors trigger an unscheduled forecast revision? action
  2. How are superseded, withdrawn and expired forecasts identified while retaining their original content for verification? action
Verification and improvement Assesses predictions against outcomes and translates demonstrated weaknesses into changes.

Forecast quality must be evaluated under explicit comparison conditions rather than inferred from isolated successes.

Verification design

Defines the observations, matching rules and reference predictions used for evaluation.

Fair outcome comparison

Record how archived forecasts are paired with suitable observations without introducing information unavailable at issuance.

  1. Which observations verify the target, and how are spatial support, accumulation windows, missing data and observation uncertainty handled? measurement
  2. Which persistence, climatological or operational reference forecast provides a meaningful comparison at the same lead time? boundary

Performance and correction

Interprets verification results for the forecast's intended use.

Conditional skill and remediation

Record performance by relevant conditions and the evidence required to change the forecasting process.

  1. Which metrics assess bias, error, probabilistic calibration or event detection, and how do results vary by lead time, season, region and event rarity? measurement
  2. What demonstrated performance failure warrants recalibration, revised communication, a method change or a restriction on use? action
Evidence and external alignment What the world already says about this thing, gathered so the model can be checked against it.

A model that cannot be lined up against existing standards, identifiers and practice cannot be adopted by anyone who already uses them.

Reported evidence

Findings from the breadth pass, kept separate from the structural claims.

Check these first

Recalled without web access and unsourced; every item is a lead to verify.

  • This describes weather forecasting as an activity, consistent with ACT.ACT, rather than a forecast document or the broader discipline of meteorology.
  • Forecast-horizon terminology varies; seasonal products concern statistical tendencies rather than precise daily weather.
  • Standards are named from recall; applicable editions, provisions, and national implementation require verification.
  1. Which of these check these first hold for the sense of weather forecasting this model covers, and on what evidence? provenance

Kinds and varieties

Recalled without web access and unsourced; every item is a lead to verify.

  • Nowcasting
  • Short-range forecasting
  • Medium-range forecasting
  • Extended-range forecasting
  • Seasonal forecasting
  • Impact-based forecasting
  1. Which of these kinds and varieties hold for the sense of weather forecasting this model covers, and on what evidence? provenance

Standards and regulation

Recalled without web access and unsourced; every item is a lead to verify.

  • World Meteorological Organization Technical Regulations (WMO-No. 49): international meteorological practices and service requirements.
  • World Meteorological Organization Manual on Codes (WMO-No. 306): standardized representations and exchange formats for meteorological information, including GRIB and BUFR.
  • International Civil Aviation Organization Annex 3: meteorological service requirements for international air navigation.
  1. Which of these standards and regulation hold for the sense of weather forecasting this model covers, and on what evidence? provenance

Real-world use

Recalled without web access and unsourced; every item is a lead to verify.

  • Issuing warnings and supporting emergency preparation for hazardous weather.
  • Planning aviation, maritime navigation, and surface transport.
  • Scheduling agricultural operations and managing irrigation and frost risk.
  • Predicting electricity demand and weather-dependent renewable generation.
  • Planning construction, outdoor work, and public events.
  1. Which of these real-world use hold for the sense of weather forecasting this model covers, and on what evidence? provenance

Typical measurements

Recalled without web access and unsourced; every item is a lead to verify.

  • Forecast lead time - Minutes to months, depending on the product; longer leads generally describe aggregated conditions rather than specific local weather events - minutes, hours, days, or months
  • Forecast probability of a specified event - 0-100 - %
  • Brier score for a binary probabilistic forecast - 0-1; lower is better - dimensionless
  • Root mean square error of temperature forecasts - Nonnegative; useful benchmarks depend on location, season, lead time, and verification data - K or °C for temperature differences
  1. Which of these typical measurements hold for the sense of weather forecasting this model covers, and on what evidence? provenance

Failure modes and hazards

Recalled without web access and unsourced; every item is a lead to verify.

  • Incomplete or inaccurate observations produce errors in estimated initial atmospheric conditions.
  • Model resolution and approximations miss or misrepresent local terrain effects, convection, clouds, and precipitation.
  • Chaotic error growth limits detailed prediction as lead time increases.
  • Poorly calibrated probabilities or underdispersive ensembles understate uncertainty.
  • Missed events, false alarms, and ambiguous communication can cause harmful decisions or reduce trust in warnings.
  1. Which of these failure modes and hazards hold for the sense of weather forecasting this model covers, and on what evidence? provenance

Regional variation

Recalled without web access and unsourced; every item is a lead to verify.

  • Observation coverage and computing resources vary, affecting forecast quality and service availability.
  • Dominant hazards differ across tropical, polar, coastal, mountainous, and continental regions.
  • Warning thresholds, terminology, units, and issuing authorities differ by jurisdiction.
  1. Which of these regional variation hold for the sense of weather forecasting this model covers, and on what evidence? provenance

Neighbouring kinds and how to tell them apart

Recalled without web access and unsourced; every item is a lead to verify.

  • Meteorology - Meteorology studies the atmosphere broadly; weather forecasting specifically estimates future atmospheric conditions.
  • Weather observation - Observation measures existing or past conditions; forecasting estimates conditions that have not yet occurred.
  • Numerical weather prediction - Numerical weather prediction computes atmospheric evolution with models; forecasting also includes interpretation, statistical processing, communication, and other methods.
  • Climate projection - Climate projection describes possible climate evolution under specified forcing scenarios; weather forecasting predicts forthcoming conditions from the current atmospheric state.
  • Weather warning - A warning communicates an anticipated hazard and its significance; a forecast can describe weather without issuing a warning.
  1. Which of these neighbouring kinds and how to tell them apart hold for the sense of weather forecasting this model covers, and on what evidence? provenance

What the second pass must settle

  • Where should this registry place the boundary between weather forecasting, subseasonal prediction and seasonal climate outlooks?
  • Should retrospective forecasts and reforecasts be instances of this activity with a distinct mode, or belong to a neighbouring evaluation model?
  • Which authoritative definitions should govern precipitation probability and area-based event probabilities across providers?
  • Which verification metrics, reference forecasts and minimum evidence requirements are appropriate for each intended use and rare-event category?
  • Should provider-specific watches and warnings remain linked products, or should their issuance activities have separate registered models?
Draft as JSON