target_0fev / AutoGluon ↗Jena Climate & Weather Telemetry
High-resolution meteorology benchmark spanning 8,784 hourly observations across 21 atmospheric telemetry variables from a single sensor station. Designed for day-ahead 24-hour multivariate weather forecasting to support renewable energy dispatch and environmental operations.
Business Objective: Jena Climate & Weather Telemetry
01Business Context
Grid operators, municipal utilities, and agritech platforms require precise 24-hour day-ahead atmospheric forecasts to schedule wind and solar generation dispatch, manage district heating thermal inertia, and plan agricultural spraying windows.
02Analytical Objective
Forecast 21 interdependent atmospheric and thermodynamic state variables over an hourly 24-step horizon (H=24) using historical multi-sensor telemetry.
03Economic & Decision Impact
Accurate day-ahead atmospheric forecasting minimizes costly ancillary grid reserve activations and imbalance penalties from wind/solar ramping errors. Under-forecasting temperature or wind triggers thermal overload or unhedged peak-load spot purchases, while over-forecasting yields curtailed renewable capacity and stranded reserve commitments.
04Temporal Dynamics & Covariate Information Flow
Solar zenith angle and clear-sky solar radiation profiles serve as deterministic calendar drivers; dynamic barometric pressure gradients and wind vector shifts govern incoming frontal transitions.
Classical univariate statistical baselines (AutoARIMA, ETS) capture diurnal oscillations for thermodynamic targets like temperature but fail on complex non-linear wind dynamics and cross-variable physical relationships. Gradient boosted decision trees (LightGBM lag models) perform robustly with cross-lagged feature engineering across atmospheric density and pressure, while zero-shot time series foundation models (Mistral's Timestral, MOIRAI, TimesFM) excel at capturing multivariate cross-attentions and zero-shot out-of-sample weather shifts without task-specific retraining.
Target & Feature Column Definitions
Exhaustive business definitions, measurement units, roles, and target variables across all 23 columns.
Known Future Covariates vs. Historically Observed Covariates
Multi-step forecasting models must respect the arrow of time and real-world information availability. FEV benchmarks and Carla strictly distinguish between features known across the forward prediction window versus realizations available only up to the forecast origin.
Scheduled promotions (such as Rossmann's Promo), calendar events (school/state holidays), and operational forecasts (such as Nord Pool's day-ahead grid load and wind power forecasts). These are deterministic or published before market clearing, allowing foundation models to legitimately condition on them across future steps.
Stochastic operational observations (such as Rossmann's actual customer footfall Customers, Rohlik's actual precipitation and snowfall, or sudden warehouse shutdowns). Because these are only realized in real-time, future values must never enter a prediction to prevent lookahead target leakage.
| Col | Variable / Feature Name | Role | Type | Unit / Scale | Description & Business Meaning |
|---|---|---|---|---|---|
| C | target_0Barometric Pressure | TARGET | NUM | mbar | Atmospheric surface pressure recorded at station elevation |
| D | target_1Air Temperature | TARGET | NUM | deg C | Ambient dry-bulb air temperature measured at 2 meters elevation |
| E | target_10Air Density | TARGET | NUM | g/m**3 | Atmospheric air density computed from pressure, temperature, and moisture content |
| F | target_11Wind Speed | TARGET | NUM | m/s | Mean horizontal wind speed measured by anemometer at mast height |
| G | target_12Maximum Wind Speed | TARGET | NUM | m/s | Peak wind gust velocity observed during the hourly recording interval |
| H | target_13Wind Direction | TARGET | NUM | deg | Compass azimuth direction from which the wind is blowing |
| I | target_14Precipitation | TARGET | NUM | mm | Total liquid precipitation accumulated over the hourly interval |
| J | target_15Rain Duration | TARGET | NUM | s | Cumulative duration of rainfall events observed within the hour |
| K | target_16Shortwave Solar Radiation | TARGET | NUM | W/m**2 | Downwelling shortwave solar irradiance incident on a horizontal surface |
| L | target_17Photosynthetically Active Radiation | TARGET | NUM | umol/m**2/s | Incident photon flux density available for photosynthesis (400-700 nm spectral band) |
| M | target_18Max Photosynthetically Active Radiation | TARGET | NUM | umol/m**2/s | Peak instantaneous photosynthetically active radiation recorded during the hour |
| N | target_19Logger Internal Temperature | TARGET | NUM | deg C | Internal operating temperature of the telemetry data acquisition system |
| O | target_2Potential Temperature | TARGET | NUM | K | Potential temperature adjusted adiabatically to a reference pressure of 1000 mbar |
| P | target_20CO2 Concentration | TARGET | NUM | ppm | Atmospheric mole fraction of carbon dioxide measured at the mast intake |
| Q | target_3Dew Point Temperature | TARGET | NUM | deg C | Temperature to which air must be cooled to become saturated with water vapor |
| R | target_4Relative Humidity | TARGET | NUM | % | Ratio of current absolute vapor pressure to saturation vapor pressure |
| S | target_5Saturation Vapor Pressure | TARGET | NUM | mbar | Maximum vapor pressure possible at prevailing ambient air temperature |
| T | target_6Vapor Pressure | TARGET | NUM | mbar | Partial pressure exerted by water vapor in the atmosphere |
| U | target_7Vapor Pressure Deficit | TARGET | NUM | mbar | Difference between saturation vapor pressure and actual vapor pressure |
| V | target_8Specific Humidity | TARGET | NUM | g/kg | Mass ratio of water vapor to total moist air mass |
| W | target_9Water Vapor Concentration | TARGET | NUM | mmol/mol | Molar concentration of water vapor molecules per mole of moist air |
| A | idStation ID | SERIES ID | CAT | — | Unique identifier for the recording weather station entity |
| B | timestampTimestamp | TIME | TIME | — | ISO 8601 recording time of atmospheric observation at hourly resolution |
Interactive Data Table
Explore rows, feature values, and target labels for Jena Climate & Weather Telemetry.
Foundation Models vs Traditional ML & Baselines (41 Methods)
Empirical evaluation directly from the AutoGluon fev benchmark suite. Global aggregate ranking: The win rates and skill scores below represent cross-dataset performance across all 100 standardized tasks in fev-bench, rather than results isolated to Jena Climate & Weather Telemetry.
| Rank | Method / Model | Family | Organization | SQL Win Rate?Average win rate (%) against all competing models across the benchmark tasks under Scaled Quantile Loss. | SQL Skill?Normalized skill score relative to the Seasonal Naive baseline (0% = baseline, higher is better). | MASE Win Rate?Average win rate (%) under Mean Absolute Scaled Error. | FEV Latency?FEV benchmark reference latency measured in server Python (median seconds per 100 series). Independent in-browser WebGPU execution for Carla is reported in the telemetry card above. | Link |
|---|---|---|---|---|---|---|---|---|
| #1 | Timestral-209M | Foundation Model | Mistral AI | 87.7% | +50.4% | 86.1% | 0.51s | Repo ↗ |
| #2 | Timestral-123M (Carla Engine) | Foundation Model | Mistral AI | 86.9% | +50.2% | 85.0% | 0.32sPython (FEV) | Repo ↗ |
| #3 | TimeRSI | Closed-Source API | GaTech & SamayLabs | 84.8% | +51.2% | 83.8% | 2512s | Repo ↗ |
| #4 | t0-gamma | Foundation Model | The Forecasting Company | 84.4% | +49.4% | 81.8% | 4.38s | Repo ↗ |
| #5 | Timestral-28M | Foundation Model | Mistral AI | 79.2% | +48.4% | 77.0% | 0.11s | Repo ↗ |
| #6 | timesfm-3 | Foundation Model | 78.6% | +49.0% | 77.0% | 3.41s | Repo ↗ | |
| #7 | ChakraTS | Baseline | Community | 78.2% | +48.1% | 79.5% | 14.22s | Repo ↗ |
| #8 | Lingjiang2.0 | Baseline | Community | 73.9% | +47.0% | 71.3% | 8.89s | Repo ↗ |
| #9 | Chronos-2 | Foundation Model | AWS | 70.4% | +47.3% | 68.4% | 0.84s | Repo ↗ |
| #10 | timeh2o-v1-tiny | Foundation Model | H2O.ai | 69.9% | +45.5% | 67.1% | 0.19s | Repo ↗ |
| #11 | Timestral-5M | Foundation Model | Mistral AI | 69.0% | +46.2% | 67.7% | 40ms | Repo ↗ |
| #12 | Toto-2.0-2.5B | Baseline | Community | 67.5% | +44.4% | 65.7% | 5.58s | Repo ↗ |
| #13 | t0-beta | Foundation Model | The Forecasting Company | 67.4% | +46.6% | 64.5% | 2.29s | Repo ↗ |
| #14 | Toto-2.0-1B | Baseline | Community | 66.4% | +44.4% | 65.6% | 2.48s | Repo ↗ |
| #15 | TiRex-2 | Foundation Model | NX-AI | 65.8% | +45.5% | 63.4% | 0.27s | Repo ↗ |
| #16 | Toto-2.0-313m | Baseline | Community | 65.1% | +44.2% | 63.5% | 1.06s | Repo ↗ |
| #17 | TiRex | Foundation Model | NX-AI | 58.1% | +42.6% | 54.2% | 0.24s | Repo ↗ |
| #18 | Toto-2.0-22m | Baseline | Community | 57.7% | +42.9% | 56.0% | 0.25s | Repo ↗ |
| #19 | TimesFM-2.5 | Baseline | Community | 54.6% | +42.2% | 53.5% | 1.87s | Repo ↗ |
| #20 | TabPFN-TS-3 | Foundation Model | Prior Labs | 52.4% | +43.1% | 51.3% | 235s | Repo ↗ |
| #21 | TS-ICL | Foundation Model | EDF-Lab | 51.2% | +43.1% | 49.8% | 2.49s | Repo ↗ |
| #22 | FlowState | Foundation Model | IBM | 49.3% | +39.9% | 49.9% | 2.29s | Repo ↗ |
| #23 | citras-fm | Foundation Model | Hitachi | 48.8% | +41.2% | 46.9% | 0.60s | Repo ↗ |
| #24 | Toto-1.0 | Baseline | Community | 47.5% | +40.8% | 45.9% | 22.35s | Repo ↗ |
| #25 | TabPFN-TS | Foundation Model | Prior Labs | 47.3% | +41.5% | 45.4% | 109s | Repo ↗ |
| #26 | Toto-2.0-4m | Baseline | Community | 46.2% | +40.7% | 43.6% | 0.20s | Repo ↗ |
| #27 | Moirai-2.0 | Baseline | Community | 41.9% | +39.3% | 40.5% | 0.35s | Repo ↗ |
| #28 | Chronos-Bolt | Foundation Model | AWS | 41.4% | +38.9% | 39.3% | 0.26s | Repo ↗ |
| #29 | TFT | Deep Learning | — | 31.7% | +32.2% | 29.2% | 829s | Repo ↗ |
| #30 | Stat. Ensemble | Baseline | Community | 27.5% | +20.7% | 30.6% | 147s | Repo ↗ |
| #31 | Sundial-Base | Foundation Model | Tsinghua University | 27.4% | +33.4% | 33.6% | 8.01s | Repo ↗ |
| #32 | PatchTST | Deep Learning | — | 26.8% | +30.1% | 25.2% | 664s | Repo ↗ |
| #33 | DeepAR | Deep Learning | — | 26.1% | +29.1% | 23.6% | 1113s | Repo ↗ |
| #34 | AutoARIMA | Statistical Baseline | — | 23.3% | +21.3% | 21.5% | 20.14s | Repo ↗ |
| #35 | AutoETS | Statistical Baseline | — | 20.3% | -26.8% | 22.5% | 3.48s | Repo ↗ |
| #36 | CatBoost | Tree Ensemble | — | 19.2% | +23.0% | 35.1% | 29.48s | Repo ↗ |
| #37 | LightGBM | Tree Ensemble | — | 17.8% | +21.0% | 33.2% | 3.03s | Repo ↗ |
| #38 | AutoTheta | Statistical Baseline | — | 15.6% | +5.5% | 19.8% | 3.28s | Repo ↗ |
| #39 | Seasonal Naive | Baseline | Community | 10.4% | 0.0% | 11.9% | 0.48s | Repo ↗ |
| #40 | Naive | Statistical Baseline | — | 6.7% | -45.4% | 11.0% | 0.47s | Repo ↗ |
| #41 | Drift | Statistical Baseline | — | 6.1% | -45.8% | 10.0% | 0.45s | Repo ↗ |
Variable Schema & Summary Distributions
Observed numerical ranges, category cardinalities, missing rates, and sample values across 8,784 rows.
| Col | Variable Name | Type | Missing | Distinct | Summary Stats / Distribution | Sample Values |
|---|---|---|---|---|---|---|
| C | target_0TARGET | NUM | 1 (0.01%) | 6,988 | [955.739990234375, 1020.038330078125] μ=990.0 σ=9.2 | 1008.71197509765621008.3566894531251008.3333129882812 |
| D | target_1TARGET | NUM | 1 (0.01%) | 6,797 | [-6.176666736602783, 34.28666687011719] μ=10.8 σ=7.5 | 0.5780000090599060.14166666567325592-0.15333333611488342 |
| E | target_10TARGET | NUM | 1 (0.01%) | 8,305 | [1108.8983154296875, 1318.2733154296875] μ=1210.9 σ=36.4 | 1281.05798339843751282.6933593751284.068359375 |
| F | target_11TARGET | NUM | 1 (0.01%) | 2,848 | [-1664.9066162109375, 12.573333740234375] μ=2.0 σ=17.9 | 0.73199999332427980.78166669607162480.5099999904632568 |
| G | target_12TARGET | NUM | 1 (0.01%) | 3,763 | [0.44333332777023315, 19.198333740234375] μ=3.6 σ=2.3 | 1.36000001430511471.35333335399627691.2300000190734863 |
| H | target_13TARGET | NUM | 1 (0.01%) | 6,857 | [4.771666526794434, 353.8666687011719] μ=176.8 σ=65.9 | 208.83999633789062177.8000030517578169.3633270263672 |
| I | target_14TARGET | NUM | 1 (0.01%) | 58 | [0, 3.700000047683716] μ=0.0 σ=0.1 | 000 |
| J | target_15TARGET | NUM | 1 (0.01%) | 308 | [0, 600] μ=25.9 σ=99.5 | 000 |
| K | target_16TARGET | NUM | 1 (0.01%) | 4,742 | [0, 928.3683471679688] μ=131.4 σ=210.1 | 000 |
| L | target_17TARGET | NUM | 1 (0.01%) | 4,839 | [0, 1823.943359375] μ=259.9 σ=412.2 | 000 |
| M | target_18TARGET | NUM | 1 (0.01%) | 4,874 | [-5947.46142578125, 2304.921630859375] μ=308.7 σ=509.7 | 000 |
| N | target_19TARGET | NUM | 1 (0.01%) | 6,837 | [6.953333377838135, 48.974998474121094] μ=21.5 σ=7.8 | 11.6140003204345712.06333351135253912.008333206176758 |
| O | target_2TARGET | NUM | 1 (0.01%) | 6,847 | [266.4966735839844, 308.7466735839844] μ=284.8 σ=7.6 | 273.0580139160156272.6499938964844272.3599853515625 |
| P | target_20TARGET | NUM | 1 (0.01%) | 3,569 | [-9999, 522.066650390625] μ=417.8 σ=306.3 | 429.17999267578125432.04998779296875434.01666259765625 |
| Q | target_3TARGET | NUM | 1 (0.01%) | 6,396 | [-13.141666412353516, 20.06999969482422] μ=5.4 σ=6.0 | -1.4780000448226929-1.6616666316986084-1.7999999523162842 |
| R | target_4TARGET | NUM | 1 (0.01%) | 5,776 | [21.899999618530273, 100] μ=72.5 σ=19.2 | 86.019996643066487.599998474121188.58333587646484 |
| S | target_5TARGET | NUM | 1 (0.01%) | 6,296 | [3.8450000286102295, 54.11833190917969] μ=14.5 σ=7.6 | 6.3699998855590826.1700000762939456.038333415985107 |
| T | target_6TARGET | NUM | 1 (0.01%) | 5,360 | [2.2149999141693115, 23.523332595825195] μ=9.7 σ=4.0 | 5.4780001640319825.405000209808355.349999904632568 |
| U | target_7TARGET | NUM | 1 (0.01%) | 4,775 | [0, 40.35499954223633] μ=4.8 σ=5.5 | 0.88999998569488530.76499998569488530.6916666626930237 |
| V | target_8TARGET | NUM | 1 (0.01%) | 4,313 | [1.375, 14.930000305175781] μ=6.1 σ=2.6 | 3.3859999179840093.33833336830139163.306666612625122 |
| W | target_9TARGET | NUM | 1 (0.01%) | 5,364 | [2.2116665840148926, 23.786666870117188] μ=9.8 σ=4.1 | 5.4320001602172855.3633332252502445.304999828338623 |
| A | id | CAT | 0 | 1 | — | item_0item_0item_0 |
| B | timestamp | CAT | 0 | 8,784 | — | 2020-01-01T00:00:002020-01-01T01:00:002020-01-01T02:00:00 |
Frequently Asked Questions: Jena Climate & Weather Telemetry
Common questions regarding the Jena Climate & Weather Telemetry dataset, machine learning task formulations, and in-browser tabular inference.
What is the forecasting horizon for Jena Climate & Weather Telemetry?
The standardized benchmark evaluates a 24-hour horizon (H=24) at hourly cadence, mapping directly to day-ahead utility dispatch and meteorological operations.
Can I run zero-server forecasts on this dataset in Google Sheets?
Yes, Carla enables zero-server in-context time series forecasting directly in spreadsheets using Mistral's Timestral foundation model.
Which models perform best on Jena Climate & Weather Telemetry?
Zero-shot foundation models like Mistral's Timestral and MOIRAI alongside tuned multivariate LightGBM models deliver top accuracy by capturing complex thermodynamic and diurnal cross-correlations.
Forecast with Carla: Jena Climate & Weather Telemetry
Open the pre-loaded Google Sheet and let Carla generate multi-step probabilistic forecasts using Mistral's Timestral directly inside your browser with zero server uploads.
Launch Carla
Click once to open the spreadsheet and Carla side panel together.
Review the Setup
Carla auto-detects the target sequence, frequency cadence, and temporal covariates.
Forecast & Calibrate
Generate multi-step quantile forecasts with zero server uploads.
Provenance & Attribution
Max Planck Institute for Biogeochemistry Weather Station Dataset, Jena, Germany (2020).