Methodology
Kpler provides short-term forecasts for dispatchable generation sources, including:
Nuclear
Fossil gas
Hard coal
Lignite (brown coal)
Hydro reservoir
These machine learning models complement our weather-driven fundamentals (wind, solar, and load) and simulate how dispatchable generation responds to market conditions and residual demand
They’re built on the same weather and infrastructure backbone, leveraging ECMWF and GFS (operational and ensemble runs) and covering up to 14 days ahead - see horizons and publication times in the weather inputs section.
We have one model per fuel type, trained across multiple bidding zones. This setup allows the models to generalize shared behaviors while capturing regional specifics. Key inputs include:
Residual demand (load minus renewable generation, using our own fundamental forecasts)
Fuel prices (gas, coal, carbon)
Fuel availability
Each model is retrained daily to incorporate the latest system and market conditions.
Coverage
Country | Bidding Zone | Fossil gas | Hard coal | Lignite | Hydro reservoir | Nuclear |
AT | AT | ✅ |
|
| ✅ |
|
BA | BA |
|
| ✅ | ✅ |
|
BE | BE | ✅ |
|
|
| ✅ |
BG | BG | ✅ |
| ✅ | ✅ | ✅ |
CH | CH |
|
|
| ✅ | ✅ |
CZ | CZ |
|
| ✅ | ✅ | ✅ |
DE | DE-LU | ✅ | ✅ | ✅ | ✅ |
|
DK1 | DK1 | ✅ |
|
|
|
|
DK2 | DK2 | ✅ |
|
|
|
|
ES | ES | ✅ | ✅ |
| ✅ | ✅ |
FI | FI | ✅ |
|
|
| ✅ |
FR | FR | ✅ |
|
| ✅ | ✅ |
GE | GE | ✅ |
|
| ✅ |
|
GR | GR | ✅ |
| ✅ | ✅ |
|
HR | HR | ✅ |
|
| ✅ |
|
HU | HU | ✅ |
| ✅ |
| ✅ |
IE | IE |
|
|
|
| ✅ |
IT | IT |
|
|
|
| ✅ |
IT-Calabria | IT-Calabria | ✅ |
|
|
| ✅ |
IT-Centre-Nort | IT-Centre-Nort | ✅ |
|
|
| ✅ |
IT-Centre-South | IT-Centre-South | ✅ |
|
|
| ✅ |
IT-North | IT-North | ✅ |
|
|
| ✅ |
IT-Sicily | IT-Sicily | ✅ |
|
|
|
|
IT-South | IT-South | ✅ |
|
|
|
|
LT | LT |
|
|
|
| ✅ |
LV | LV |
|
|
|
| ✅ |
MD | MD | ✅ |
|
|
|
|
NL | NL | ✅ | ✅ | ✅ |
| ✅ |
NO1 | NO1 |
|
|
| ✅ | ✅ |
NO2 | NO2 |
|
|
| ✅ | ✅ |
NO3 | NO3 |
|
|
| ✅ | ✅ |
NO4 | NO4 |
|
|
| ✅ | ✅ |
NO5 | NO5 |
|
|
| ✅ | ✅ |
PL | PL | ✅ | ✅ | ✅ | ✅ | ✅ |
PT | PT | ✅ |
|
| ✅ | ✅ |
RO | RO | ✅ |
| ✅ | ✅ | ✅ |
RS | RS |
|
| ✅ | ✅ | ✅ |
SE1 | SE1 |
|
|
| ✅ |
|
SE2 | SE2 |
|
|
| ✅ |
|
SE3 | SE3 |
|
|
| ✅ |
|
SE4 | SE4 |
|
|
| ✅ |
|
SI | SI |
|
| ✅ |
| ✅ |
SK | SK |
|
|
| ✅ | ✅ |
UK | UK | ✅ |
|
|
| ✅ |
XK | XK |
|
| ✅ |
|
|
Access
API
Forecasts by run date: returns all four forecast runs (00, 06, 12, 18) for a given date.
Ensemble statistics: for a given run date, provides statistical summaries (min, q25, median, mean, q75, max) calculated across ensemble members from EC ENS and GFS ENS
FTP
Path : 4_Supply/Forecast/Generation/COR_E/{Country}/yyyy/mm
File naming convention: CORE_SUPPLY_FORECAST_Generation_CORE_{Country}_{FuelType}_Hourly_{EC|GFS}_{OP|ENS|ENS_SCENARIOS}_yyyymmdd{00|06|12|18}.csv
For ensemble forecasts, two files are published for each run:
One
{ENS}file containing the control scenario “0”One
{ENS_SCENARIOS}file containing all other ensemble members