Frequently Asked Questions

ForecastAdvisor was created by ForecastWatch, another valuable member of the Intellovations family. It is a product for everyone who is interested in the what the weather will be. It answers not only the question "What will the weather be?" by providing the National Weather Service Digital Forecast Database forecast, but also "Who provides the most accurate weather forecast for my city?". It does this in a clean, easy-to-read format not cluttered with graphics and ads. ForecastAdvisor .
The overall accuracy percent is computed from the one- to three-day out accuracy percentages for high temperature, low temperature, icon forecast precipitation (both rain and snow), and text forecast precipitation (both rain and snow). Temperature accuracy is the percentage of forecasts within three degrees. Precipitation accuracy is the percentage of correct forecasts. The forecasts are collected in the evening.
High-quality observations are essential to producing meaningful, useful data. Observations are retrieved from quality-controlled local information provided by the National Climactic Data Center (NCDC) and from other meteorological agencies whose data has met specific quality standards. The quality control process on this dataset includes more than 50 validity checks, which include extreme value checks, internal (within observation) consistency checks, and temporal (versus another observation for the same station) continuity checks. Observations that don’t pass these checks are flagged as invalid.
Best efforts are made to ensure that provider forecasts are for the appropriate, corresponding observation location. Latitude/longitude coordinates or actual weather station names are commonly used to ensure that forecasts cover the location being analyzed. This process is applied to 990 U.S. locations. Forecasts are typically retrieved from providers’ public websites and APIs. Forecasts for a given location are retrieved at the same time from all providers, and location order is randomized each day.
A series of audits are performed to ensure that forecasts and observations that are highly improbable or obviously not valid are excluded from analysis. This is accomplished through the use of pre-set thresholds that allow for identification of observations and forecasts falling outside reasonable ranges. Examples of data excluded from analysis include:
  1. a forecast that’s considered impossible, such as a prediction of -200°F for a particular location, or
  2. observations or forecasts that are highly unlikely — such as a temperature of 27°F in Los Angeles. The goal is to eliminate from analysis forecasts that are obviously wrong, but not forecasts that are simply bad.
A sophisticated system is used to parse and interpret text and icon forecasts. For example, if a text forecast calls for "a mix of sun and clouds," logical assumptions are made about what that means regarding the degree of cloud cover. Similar logic is applied to icons depicting varying degrees of sunshine and clouds. The text and icon interpretation system allows forecasts to be placed into canonical categorizations, which assist with analysis and scoring.
ForecastWatch calculates the accuracy, skill, and quality of weather forecasts. We collect over 40,000 forecasts each day from Accuweather, the National Weather Service, and others for 990 U.S. cities and 1300 international cities and compare them with what actually happened. With this information, we help weather forecast companies, businesses, and individuals:
  • Evaluate weather forecast providers
  • Improve decision-making where weather forecasts are used as input
  • Improve weather forecasts by providing useful feedback
  • Improve the quality of weather forecast websites
  • Provide additional data for decision and risk analysis
Weather plays an important role in our economy. Hundreds of thousands of businesses and individuals use weather forecasts every day to make decisions that have real economic impact. Hundreds of thousands more businesses and individuals could use weather forecasts as part of their daily planning process if the weather impacts on their businesses or the risks of using weather forecasts could be quantified. But until ForecastWatch, nobody had real data on who is the most accurate and skillful.
There are many reasons for this. One reason is that because ForecastWatch collects forecasts in the evening, the low temperature forecasts are for 12 hours later than the high temperatures. Therefore, the low temperature forecasts are for a half-day further out than the highs. Think of it this way. The high temperature usually occurs in the late afternoon, around the time ForecastWatch collects its forecasts. So a one-day out high temperature forecast will occur in about 24 hours. The low temperature forecast usually occurs in the early morning. So a one-day out low temperature forecast for tomorrow actually will occur in the morning of the day after tomorrow, about 12 hours after the high.
Accuracy measures how close a weather forecast is to what actually happens. But it doesn't tell you anything about how skillful that weather forecaster is. That is an important difference. Because if you never predict rain, in most parts of the country you will have an accuracy of 70% or so. Skill measurements try to quantify how much better a weather forecast is than chance, or unskilled forecasts like climatology or persistence forecasts. ForecastWatch can help you create an accuracy or skill measurement that will help you evaluate or quantify the value of your weather forecasts.
You might be interested in the more advanced services ForecastWatch provides! In addition to providing the percentages used here at ForecastAdvisor, we also do advanced accuracy and skill calculations. Climatology weather forecasts are forecasts which predict that the weather will match climate averages. Persistence weather forecasts are forecasts that predict that tomorrow, and the next day, and so on will be exactly like today. These are considered "unskilled" forecasts, as it doesn't take any skill to produce. They are used to help calculate the skill of weather forecasts produced by meteorologists.