This GoldSim model uses a DLL to download a 7-day weather forecast for a specified location within the U.S. using latitude and longitude inputs. The following information is downloaded and saved in GoldSim when you run the model: 

• Maximum Daily Temperature (F) for next 7 days
• Minimum Daily Temperature (F) for next 7 days
• Probability of Rain (%) for next 7 days
• Rain Amount (in) for next 3 days

Below is a screen capture of the model interface, which allows you to provide the lat/long inputs and check the box to enable a fresh download. This example is showing the latitude and longitude for Seattle Washington, USA.

Since the online weather data is only updated about 1 time per hour, it is likely that you don't need to perform a new download each time you run the model. When you check the box "Enable Web Download" to get new forecast data, an XML file will also be in the model directory (see "NDFD_WebData.xml"). This file is used for subsequent model runs should you un-check the "Enable Web Download" box. Note that if you have the box checked, a download will occur for every simulation (and realization, if running Monte Carlo!). When you run the model, please make sure that the simulation start date is the same day as the forecast start date (should be the same as your system's local time).

Note: The model requires that you set the start date of the simulation to today's date and the end date to 7 days after for the web download to operate correctly.

Temperature forecasts are validated each month by the National Weather Service and the results are published on their website (see hyperlink "NDFD Verification Definitions" below). For temperature forecasts, the mean absolute error (MAE) is calculated and published on a monthly basis. Some MAE values for the past year were evaluated at various locations throughout the U.S. and an approximate set of values are defined based on this review and used as input for this model. The MAE is used as the variance from the most likely value using a triangular distribution where the most likely is the forecast value, the minimum is the forecast - MAE and maximum is forecast + MAE.

Read more about the API used to access the data from here: National Digital Forecast Database (NDFD) REST Web Service.

Download the Model Files: