Retrieve HydroForecast Inputs from the API: Weather and Satellite Data

API Data Sources and Variables

HydroForecast models rely on high-quality environmental data to generate accurate streamflow forecasts. This document provides an overview of the data sources and variables available through our API for:

• Flow Forecasts
• Flow Observations
• Weather Forecasts & Observations
• Surface Observations
• Historical Percentiles of Observed Data
• Unit Systems

This guide is intended for users working with our API who want to understand what data is available for a given site, and how it's structured. For detailed API usage, refer to our API setup documentation.

Flow Forecasts

HydroForecast produces probabilistic streamflow forecasts for short-term and seasonal horizons. These forecasts include confidence intervals to capture uncertainty. The below quantiles are available by default via the API. If your organization needs additional quantiles, contact us to get them set up.

Sources & columns for short-term flow forecasts:

• hydroforecast-short-term
  • Available columns:
    • discharge_q0.01
    • discharge_q0.025
    • discharge_q0.05
    • discharge_q0.1
    • discharge_q0.25
    • discharge_q0.5 (median)
    • discharge_mean
    • discharge_q0.75
    • discharge_q0.9
    • discharge_q0.95
    • discharge_q0.975
    • discharge_q0.95

Sources & columns for seasonal flow forecasts:

• hydroforecast-seasonal
  • This source includes the following discharge quantiles as available columns:
    • discharge_q0.01
    • discharge_q0.025
    • discharge_q0.05
    • discharge_q0.1
    • discharge_q0.25
    • discharge_q0.5 (median)
    • discharge_mean
    • discharge_q0.75
    • discharge_q0.9
    • discharge_q0.95
    • discharge_q0.975
    • discharge_q0.95
    • discharge_mean_all_traces
    • discharge_median_all_traces
  • You may also see columns labeled with integers (1, 2, …, 43). These represent individual forecast traces – each one showing a possible future streamflow scenario generated by the seasonal ensemble model.
• Some seasonal customers may also have hydroforecast-seasonal-upsampled available, which contains the same quantile columns as hydroforecast-seasonal, but does not include individual traces or trace statistics.

In some cases, we display flow forecasts from U.S. River Forecast Centers (RFCs) on the HydroForecast Dashboard. These forecasts are not available at all sites, and we don’t currently support full API integrations with every RFC. If you see RFC forecasts displayed for your site in the Dashboard, contact us to learn more and get the correct source id to use in your API queries.

Flow Observations

Streamflow observations are used for both training and evaluation of HydroForecast models. These data may come from public agencies or customer-provided feeds.

Observation availability varies by site. To determine what flow observation sources are available at your site or set up a new API connection for observation data, reach out to us.

Weather Forecasts & Observations

HydroForecast uses a range of global and regional weather models. Each source provides a consistent set of variables.

Sources for weather forecasts & observations:

• ecmwf-hres
• gefs
• gfs
• hrrr [available within CONUS bounds only]

Available columns for all weather sources:

• precipitation Precipitation rate at surface
• temperature Temperature at 2m
• specificHumidity Specific humidity at 2m
• wind 10m wind speed (u & v components combined)
• windU 10m wind velocity, u component
• windV 10m wind velocity, v component
• downwardShortWaveRadiationFlux Solar radiation at surface

Additional columns for HRRR only [available within CONUS bounds]:

• snowDepth
• SWE Snow water equivalent of accumulated snow depth

Note: GEFS data reflects the mean across all ensemble members. Individual member forecasts and additional statistics are not currently available.

Experimental Weather Sources:

Note that these are only available for short-term weather forecasts, but not observations.

• weathernext-graph
  • Available columns:
    • precipitation
    • temperature

Surface Observations

These remote sensing datasets provide information about surface conditions like vegetation, snowpack, and snow cover.

Surface observation sources & available columns:

• modis-ndsi
  • Available column: ndsiRolling16DayMedian
• modis-ndvi
  • Available column: ndviRolling16DayMax
• snodas [available within CONUS bounds only]
  • Available column: snowWaterEquivalent

Historical Percentiles of Observed Data 

Historical percentiles are also available for some of our observed sources, and can be queried using the same / observations route with the below source names.

Typically the mean, median, 5th, 25th, 75th, and 95th quantiles are available for a subset of variables; querying for any of the below sources without specifying the columns will return all the columns that are available for that source.

Flow observation historical percentiles

Available sources:

• historical-percentile-daily-gauge-observation [at sites where gauge data is available]
• historical-percentile-monthly-gauge-observation [at sites where gauge data is available]

Weather historical percentiles

Available sources:

• historical-percentile-daily-gefs [at sites where we have a seasonal model set up]
• historical-percentile-monthly-gefs [at sites where we have a seasonal model set up]

Surface observation historical percentiles:

Available sources:

• historical-percentile-daily-modis-ndsi
• historical-percentile-daily-modis-ndvi
• historical-percentile-daily-snodas [available within CONUS bounds only]

Unit Systems

The API supports multiple unit systems. Use the unit_system query parameter to control output units. If not specified, the API defaults to Metric.