Forecast Fields

Variable: Weather variable being predicted. A full list below-

• Precipitation: Precipitation anomalies expressed as standardized units or units of mm/month.
• Temperature: Temperature anomalies expressed as standardized units or units of degrees Celsius.
• Wind: Wind velocity anomalies expressed as standardized units or units of m/s
• Solar: Solar irradiance expressed as standardized units or units of W/m^2.
• Solar Power: Solar power anomalies expressed as percentage from normal. Indicates the percentage anomaly from the 1980-2010 baseline, Output is in fractional units.
• Wind Power: Wind power anomalies expressed as percentage from normal. Indicates the percentage anomaly from the 1980-2010 baseline, Output is in fractional units.
• Degree Days: Measures energy demand for heating or cooling based on temperature deviations from 65°F. HDDs are for days below 65°F, and CDDs are for days above. Output is in HDD/CDDs per day, averaged over a specified period (forecast_window). Negative values represent CDDs, positive values represent HDDs.

Points: Latitude and longitude coordinates. To retrieve global data, remove the points parameter and use the string "global" instead.

Data_type: Specifies the type of processing applied to the data-

• standardized: All variables range between approximately -3 and 3, allowing for direct comparison on a similar scale.
• unstandardized: Values are presented in physical units (e.g., cm, Celsius).
• percent_anomaly: Indicates the percentage anomaly from the 1980-2010 baseline, Output is in fractional units. Temperature anomalies use 65°F as the baseline for heating degree days and cooling degree days.

window_start: The start date of the forecast, use the start date in the format "YYYY-MM-DD" for backtesting, "MM-DD" for accuracy.

years_in_future: A value of 0 indicates a date within the current calendar year. A value of 1 indicates a date between 1 and 2 calendar years in the future.

forecast_window: Forecast duration in months

Issuance_Date: Used for backtesting and accuracy pulls, Shows the date at which the user wants to see what forecasts would have been on a specific date of issuance, e.g. “I have to put in bids on May 15, so I want to see what forecasts and accuracy would be if I have to issue forecasts on this date”, Use format of MM-DD for Accuracy endpoint, "YYYY-MM-DD" for Backtesting endpoint

Metric: type of accuracy metrics that are returned in Accuracy Endpoint-

• Anomaly Correlation Coefficient (ACC): measures how well predicted anomalies match observed anomalies; higher ACC values indicate better accuracy and are presented in raw form. In this case, metric is using .5 Quantile.
• Continuous Ranked Probability Skill Score (CRPSS): Evaluates the accuracy of probabilistic forecasts by comparing the predicted probability distribution to the observed outcomes. A higher CRPSS indicates better forecast performance. CRPSS is shown as a percentage improvement over climatology (i.e., how much better the forecast is compared to assuming weather conditions will remain the same as in the past). Metric uses all of the Probability Density Function (PDF).
• Mean Absolute Error (MAE): Calculates the average absolute differences between predicted and observed values. A lower MAE indicates better forecast accuracy. MAE is shown as a percentage improvement over climatology (i.e., how much better the forecast is compared to assuming weather conditions will remain the same as in the past). In this case, metric is using .5 Quantile.

Quantiles: The quantiles to be sampled from the forecast data. Values must be between 0.01 and 0.99. A quantile of 0.50 represents the 50th percentile of the forecast distribution.