The Salient platform provides access to gridded data for all model outputs (and historical data) on a unified grid.

Figure 1: Map of timescales to available models

Figure 1: Map of timescales to available models

Salient Blend

The Salient Blend model is a multi-model blend of our proprietary climatology, AI models (natively probabilistic and calibrated by design) with properly calibrated dynamical (GEFS and ECMWF) models.

Our blending algorithm objectively assesses which models are most skillful for particular variables, regions, seasons, and lead times. The blending model also dynamically weights the input models based on current climate conditions, which allows the blend forecast to capture signals from each input.

Then it uses that information to appropriately apply grid-point-by-grid-point model weights to combine the AI and NWP models resulting in a 'Salient Blend' forecast that is superior to the individual model forecasts by accounting for each model's varying spatial and temporal strengths.

Model Name

API model: blend

Spatial Resolution

The Salient forecast has a global extent and provides forecasts for latitude-longitude grid points at 0.25 x 0.25 degrees (25 km) resolution.

Salient Blend Temporal Resolution and Update Frequency

The following table provides an overview of the Salient model update frequency and initialization times for each timescale in UTC time. For example, let’s say the Initialization Date -Time is November 1st 00 UTC. A forecast with an Init Date of November 1st will be available on November 2 at 8:00 UTC.

<aside> 💡 Note: The forecast availability is dependent on the availability of 3rd party data at initialization time. If the initialization data is not available, Salient automatically retries downloading the data on a pre-determined schedule, which may cause a delay in forecast availability.

</aside>

Timescale (UI) API timescale Horizon Update Frequency Initialization Time Availability post-initialization
Weekly rolling sub-seasonal Weeks 1-5 Daily 00z + 1 day 08z
Weekly calendar weekly Weeks 1 to 4 or 5 Daily 00z + 1 day 08z
Monthly rolling seasonal 3-30 day periods Weekly Wednesday 00z Thursday 08z
Monthly calendar monthly 2 calendar months Weekly Wednesday 00z Thursday 08z
Quarterly rolling long-range Quarters 1-4 Monthly 15th 00z 16th 08z
Quarterly calendar quarterly Seasons 1 to 3 or 4 Monthly 15th 00z 16th 08z

<aside> 📌

There are rolling and calendar time periods for the timescale parameter.

The rolling periods are calculated by adding the lead time to the init date.

all, subseasonal, seasonal, long-range

timescale Time Period
sub-seasonal 5 - 7 day periods - Rolling 7 days from initialization date, e.g. if the init date is August 15, the week will start on August 15 and end on August 21.
seasonal 3 - 30 day periods - Rolling 30 days from initialization date, e.g. if init date is August 15, then days 1-30 will include August 15.
long-range 4- 90 day periods - Rolling 90 day (starting the month after the init date), e.g. if init date is August 15, then the quarter will begin on Sept 1, i.e. days 1-90 including Sept. 1

The calendar periods are fixed to a constant period definition to enable easier comparisons between model init dates. These are always used in the forecast evolution view.

weekly, monthly, quarterly

timescale Time Period
weekly Complete weeks Monday through Sunday. 5 leads are available on the Monday init date; otherwise there are 4 weeks. ** Available for the Salient forecast, NOAA GEFS and EC ENS.
monthly Complete calendar months so there are generally only 2 full months. About once a year there is only 1 month when the forecast date is right after the first of the month and there are two months that have 31 days in them. ** Only available for the Salient forecast and EC SEAS 5.
quarterly Complete meteorological ** seasons: DJF, MAM, JJA, SON. The init dates before the meteorological quarter (Feb, May, Aug, Nov) have the full four quarters; otherwise there are 3 quarters. ** Only available for the Salient forecast.
</aside>