Learn more

See below to access the cyberinfrastructure that underlies the VERA Challenge, helpful tutorials for participating the VERA Challenge, and papers related to VERA.

Tutorials

We provide an introductory tutorial for submitting to the Challenge: https://github.com/LTREB-reservoirs/vera4cast-example.

We provide videos of the tutorial to guide you through the submission process. The videos include an:

Anatomy of a VERA forecast

Setting up your R Environment

Generating a Forecast Part 1

Generating a Forecast Part 1

Automating Your Forecast

Python Tutorial

Click here an introductory tutorial for submitting to the Forecting Challenge using Python

Research from the Virginia Ecoforecast Reservoir Analysis team

Publications about forecasting and modeling

Accessing NOAA Global Ensemble Forecasting System weather forecasts

Your forecast model may require weather as a driving variable. We provide functions in the vera4castHelpers package that access NOAA weather forecasts that have already been downloaded and standardized for the reservoirs. To install the vera4castHelpers use

remotes::install_github("ltreb-reservoirs/vera4castHelpers")

Using the package, 31 ensemble member forecasts are provided at 1 hr time resolution

Variables and units are:

  • air_temperature (K)
  • air_pressure (Pa)
  • relative_humidity (proportion)
  • surface_downwelling_longwave_flux_in_air (W/m^2)
  • surface_downwelling_shortwave_flux_in_air (W/m^2)
  • precipitation_flux (kg/(m^2 s))
  • eastward_wind (m/s)
  • northward_wind (m/s)
weather <- vera4castHelpers::noaa_stage2(start_date = "2022-04-01")
df_future <- weather |> 
  dplyr::filter(site_id == "fcre") |>
  dplyr::collect()

The data frame has following columns:

site_id: string : NEON site ID
prediction: double : forecasted value
variable: string : weather variable
horizon: double : number of hours in the future
family: string: class of uncertainty (ensemble)
parameter: int32 : ensemble member number
reference_datetime: timestamp[us, tz=UTC]: datetime of horizon 0
datetime: timestamp[us, tz=UTC] : datetime of forecast

“Historical weather”

Stage 3 can be viewed as the “historical” weather for site as simulated by NOAA GEFS. Stage 3 is useful for model training because it ensures that the magnitude and variability of the weather data used to train your model is similar to that in the NOAA GEFS weather forecast you may use as inputs to your forecast.

Stage 3 uses CF variable names and 1 hr interval

  • air_temperature (K)
  • air_pressure (Pa)
  • relative_humidity (proportion)
  • surface_downwelling_longwave_flux_in_air (W/m^2)
  • surface_downwelling_shortwave_flux_in_air (W/m^2)
  • precipitation_flux (kg/(m^2 s))
  • eastward_wind (m/s)
  • northward_wind (m/s)
weather_stage3 <- vera4castHelpers::noaa_stage3()
df_historical <- weather_stage3 |> 
  dplyr::filter(site_id == "fcre") |>
  dplyr::collect()

Stage 3 has the following columns

site_id: string : NEON site ID
prediction: double : forecasted value
variable: string : weather variable
family: string: class of uncertainty (ensemble)
parameter: int32 : ensemble member number
reference_datetime: timestamp[us, tz=UTC]: always NA in Stage3 datetime: timestamp[us, tz=UTC] : datetime of forecast