19 October 2026 · Breaking

POST body parameters, global coverage, 2025 data, typical meteorological year, and OpenAPI schemas

On 19 October 2026 the web API moves all POST parameters into the request body, extends spatial coverage to global land (including the Arctic and Antarctic) plus 25 km into the sea, adds 2025 to the available time period, introduces a typical meteorological year endpoint, and refreshes OpenAPI response schemas and examples.

1. POST parameters in the request body (Breaking!)

POST calculation endpoints no longer read input from the query string. Every parameter that used to be a query argument such as location, time window, surface, photovoltaic options and others, must be sent in the request body as form fields (multipart/form-data). A horizon profile file, is also a body part as before.

Query parameters on POST will not be applied! and clients that still call POST …?latitude=…&longitude=… will break. Field names are unchanged and match the GET query parameters; only the location in the HTTP request changes. This applies on all POST endpoints. See the OpenAPI schema for each operation’s requestBody.

GET endpoints are unaffected and still use query parameters. This follows the REST convention: GET keeps parameters on the URI; POST sends them in the request body.

2. Spatial coverage

The previous service envelope was roughly 75°N–60°S. Coverage now includes global land from 90°N to 90°S, plus 25 km into the sea from the coastline (ERA5, ERA5LandPlus, horizon profile etc).

Use the interactive coverage map to see where each irradiance source applies:

• SARAH-3
• ERA5 / ERA5LandPlus provide the expanded Arctic and Antarctic coverage.

2.1 New polar region coverage

The following studies investigated air temperature, wind speed, downward longwave radiation (STRD), and downward shortwave radiation from ERA5 and ERA5-land and a summary is presented:

• Pernov et al. (2024): Over 17 continental Arctic observatories (60.9–82.5°N, 6 to 20 years per station). Hourly ERA5 surface pressure R² = 0.98 ± 0.02, 2 meter air temperature R² of 0.94 ± 0.02 and RMSE = 3.5 °C, wind speed R² = 0.49 ± 0.12 and RMSE = 2.8 m s⁻¹, downward shortwave radiation R² = 0.87 ± 0.03 and RMSE = 60 W m⁻² (slope 0.88), and finally relative humidity R² = 0.42 ± 0.20.
• Delhasse et al. (2020): PROMICE stations on the Greenland Ice Sheet (61.0–79.9°N; 2010–2016). Daily ERA5 2 meter air temperature: annual mean bias ≈ 0 °C, RMSE 3.1 °C, R = 0.97. Wind speed: annual mean bias −1.0 m s⁻¹, RMSE 2.2 m s⁻¹, R = 0.85. Downward longwave radiation (STRD): annual mean bias −16 W m⁻², RMSE 23 W m⁻², R = 0.94. Downward shortwave radiation: annual mean bias −3.0 W m⁻², RMSE 27 W m⁻², R = 0.98 (summer mean bias −3.7 W m⁻², RMSE 42 W m⁻², R = 0.91).
• Chen et al. (2026): 51 AWS on the Greenland Ice Sheet (~61–80°N; 2018–2022). ERA5 monthly 2 meter air temperature (spring) mean bias is 2.3 °C, RMSE 2.6 °C (R ≈ 1.00). Wind is the weak ERA5 variable (monthly spring R = 0.37). Monthly STRD (spring) mean bias is −21 W m⁻², RMSE 26 W m⁻² (R = 0.82); hourly STRD RMSE is 28–30 W m⁻² (R up to 0.86). Monthly downward shortwave radiation mean bias is −1.7 W m⁻² in spring and −17.5 W m⁻² in summer (R = 0.98); daily shortwave correlation stays above 0.97 except in summer. ERA5-Land is weaker than ERA5 for 2 meter air temperature and surface pressure (ERA5 monthly temperature bias stays below 2.5 °C). ERA5-Land wind speed correlation is inconsistent, and summer downward shortwave radiation is systematically underestimated.
• Jia et al. (2024): 7 AWS on the Zhongshan–Dome A transect, East Antarctica (~69–80°S). ERA5-Land downward shortwave radiation is the most accurate product they tested (R = 0.988, RMSE = 24 W m⁻², mean bias −1.7 W m⁻²). Station mean bias ranges from +9.9 W m⁻² at coastal Zhongshan to −19 W m⁻² inland; RMSE is largest at the coast and decreases inland. ERA5 overestimates downward shortwave radiation (the only product in their set that does not underestimate). ERA5-Land, ERA5 and MERRA-2 are more accurate in spring and summer.
• Rowe et al. (2025): Escudero Station, King George Island (62.2°S; 2017–2023). ERA5 monthly downward shortwave radiation mean bias is +38 to +50 W m⁻² in summer (near zero in winter). Monthly STRD mean bias is −18 to −22 W m⁻² in summer and −16 W m⁻² over the year. Hourly downward shortwave and longwave fluxes together have R = 0.72.
• Sato et al. (2026): R/V Shirase over the Southern Ocean (December 2022 and March 2023). ERA5 overestimates the frequency of low-level clouds with bases below 1 km. ERA5, MERRA-2 and CAM-ATRAS all underestimate STRD.
• Tetzner et al. (2019): 13 AWS in the southern Antarctic Peninsula and Ellsworth Land (71.3–77.5°S; 1982–2017). ERA5 2 meter air temperature R = 0.98; coastal mean bias −0.5 °C, escarpment +0.1 °C. Wind speed mean bias −1.2 m s⁻¹; R up to 0.95. No shortwave or STRD evaluation. ERA5-Land is not evaluated.

References

• Pernov, J. B., Gros-Daillon, J., and Schmale, J. (2024). Comparison of selected surface level ERA5 variables against in-situ observations in the continental Arctic. Quarterly Journal of the Royal Meteorological Society, 150(761), 2123–2146. https://doi.org/10.1002/qj.4700
• Delhasse, A., Kittel, C., Amory, C., Hofer, S., van As, D., Fausto, R. S., and Fettweis, X. (2020). Brief communication: Evaluation of the near-surface climate in ERA5 over the Greenland Ice Sheet. The Cryosphere, 14, 957–965. https://doi.org/10.5194/tc-14-957-2020
• Chen, Z.-M., Zeng, Z.-L., Ding, M.-H., and Wang, Y.-Q. (2026). Comprehensive evaluation of multi-source reanalysis datasets for surface atmospheric parameters over the Greenland Ice Sheet. Advances in Climate Change Research, 17(1), 48–64. https://doi.org/10.1016/j.accre.2025.10.007
• Jia, J. J., Zeng, Z. L., Zhang, W. Q., Zheng, X. D., Wang, Y. Q., and Ding, M. H. (2024). The performance of downward shortwave radiation products from satellite and reanalysis over the transect of Zhongshan Station to Dome A, East Antarctica. Advances in Atmospheric Sciences, 41(8), 1574–1588. https://doi.org/10.1007/s00376-023-3136-0
• Rowe, P. M., Zou, X., Gorodetskaya, I., Stillwell, R. A., Cordero, R. R., Bromwich, D., et al. (2025). Comparison of cloud and radiation measurements to models over the Southern Ocean at Escudero Station, King George Island. Journal of Geophysical Research: Atmospheres, 130, e2025JD043563. https://doi.org/10.1029/2025JD043563
• Sato, K., Inoue, J., and Matsui, H. (2026). Evaluation of low-level clouds, temperature, and surface radiation using a model, two reanalyses, and Southern Ocean observations. Geophysical Research Letters, 53, e2026GL123000. https://doi.org/10.1029/2026GL123000
• Tetzner, D., Thomas, E., and Allen, C. (2019). A validation of ERA5 reanalysis data in the southern Antarctic Peninsula – Ellsworth Land region, and its implications for ice core studies. Geosciences, 9(7), 289. https://doi.org/10.3390/geosciences9070289

3. Time period 2014–2025

The service window now is 2014–2025 and start_time, end_time now accept timestamps through 31 December 2025.

4. Typical meteorological year (New feature!)

A typical meteorological year is now a first-class meteorological endpoint:

GET /api/v6/meteorological/typical-meteorological-year

The endpoint can return the assembled year in any of these formats:

The default Finkelstein–Schafer method is ISO-15927-4-ECDF. Its weighting scheme is ISO-TMY (ISO 15927-4:2005). The same method also accepts Sandia (Hall, I. J., Prairie, R. R., Anderson, H. E., & Boes, E. C. (1978). Generation of Typical Meteorological Years for 26 SOLMET Stations. SAND78-1601. Albuquerque, NM: Sandia National Laboratories.) and NSRDB-TMY (Wilcox & Marion, 2008, Table 2-1).

Literature aliases follow Sun et al. (2020), Table 1. Each accepted name is the readable form of a bracket label in that table, for example Pusat2015 ([50]) and Lund1995Pissimanis1988 ([24,32]). The full list is in the OpenAPI schema.

finkelstein_schafer_method=PVGIS-legacy-sorted accepts PVGIS-legacy-ISO or PVGIS-legacy-PV. These follow PVGIS TMY version ≤ 5.3 (Huld et al., 2018).

References

• International Organization for Standardization. (2005). ISO 15927-4:2005. Hygrothermal performance of buildings — Calculation and presentation of climatic data — Part 4: Hourly data for assessing the annual energy use for heating and cooling. https://www.iso.org/standard/41371.html
• Hall, I. J., Prairie, R. R., Anderson, H. E., & Boes, E. C. (1978). Generation of Typical Meteorological Years for 26 SOLMET Stations. SAND78-1601. Albuquerque, NM: Sandia National Laboratories.
• Marion, W., & Urban, K. (1995). User’s Manual for TMY2s (Typical Meteorological Years). NREL/SP-463-7668.
• Wilcox, S., & Marion, W. (2008). Users Manual for TMY3 Data Sets. Technical Report NREL/TP-581-43156, Revised May 2008. https://doi.org/10.2172/928611
• Sun, J., Li, Z., Xiao, F., & Xiao, J. (2020). Generation of typical meteorological year for integrated climate based daylight modeling and building energy simulation. Renewable Energy, 160, 721–729. https://doi.org/10.1016/j.renene.2020.07.024 https://www.sciencedirect.com/science/article/pii/S0960148120311009
• Huld, T., Paietta, E., Zangheri, P., & Pinedo Pascua, I. (2018). Assembling Typical Meteorological Year Data Sets for Building Energy Performance Using Reanalysis and Satellite-Based Data. Atmosphere, 9(2), 53. https://doi.org/10.3390/atmos9020053 https://www.mdpi.com/2073-4433/9/2/53

5. OpenAPI schemas and examples

The schema for this API (this deployment) is:

• OpenAPI JSON
• Swagger UI
• ReDoc

Calculation endpoints now declare Pydantic response_model objects split by theme (performance, power, meteorological, position, surface, topography, coverage). Shared enums live in schemas/common.

200 examples in Swagger / ReDoc are generated from those models, including CSV examples, so the documented payload matches the schema. Error examples (400, 404, 422, 429, 500) are shared.

If you generate clients from OpenAPI, download the JSON above. JSON keys and CSV headers follow the 23 July 2026 snake_case / no-symbol convention; this release does not rename those keys again.

What does not change

• Versioned URL prefix /api/v6
• Existing performance, power, solar-position, and topography paths
• GET query parameter names
• HTTP status codes and error envelope

What you should do

1. Send POST inputs in the request body (form fields), not on the query string.
2. Allow end_time through 2025 where your client validates the calendar.
3. Point TMY integrations at /meteorological/typical-meteorological-year.
4. Use polar ERA5 / ERA5-Land irradiance with caution.
5. Refresh OpenAPI-based clients from the schema.

Effective date: 19 October 2026 (19/10/2026). Questions: JRC-PVGIS@ec.europa.eu.

23 July 2026 · Breaking

Response keys and CSV headers

On 23 July 2026 the web API changes JSON keys and CSV column headers. Values and shapes stay the same; only key and header names do.

Old key names will not remain available. From that date, previous JSON keys and CSV headers are dropped with no backwards-compatibility window. Clients that still expect the old names will break.

JSON keys move to standardised consistent lowercase snake_case keys across the API responses. CSV headers keep readable names but drop symbols and emoji (for instance ⌁ Power → Power).

1. Top-level / envelope JSON keys

And more…

2. Nested payload fields (JSON)

Calculation responses no longer use display-name aliases with symbols (for example "Power ⌁", "Latitude ϕ", "Fingerprint 🆔"). Nested fields use the schema field names in snake_case. Representative renames:

The same rule applies across verbose levels: if a JSON key used to be an inconsistent label or carried a symbol, expect the snake_case field name from the OpenAPI schema instead.

3. Statistics object (JSON)

4. CSV column headers (symbols removed)

CSV exports still use human-readable headers, but leading symbols and emoji are removed. Do not match on glyphs such as ⌁, 🔌, 🆔, ϕ, ∡, ☉, ⬇/⭳, ％, 🮞, or 🌡.

This affects power, performance, solar-position, surface-optimisation, and related CSV downloads. OpenAPI CSV examples follow the new headers.

For a column-by-column reference (positions, verbosity levels, old and new parameter names, and descriptions), see the explanatory notes to the PVGIS CSV output parameters.

What does not change

• URL paths and query parameter names
• HTTP status codes and error shapes
• Numeric values, physical units, and series content

What you should do

1. Update JSON clients to use snake_case keys.
2. Update CSV parsers to match headers without symbols.

Effective date: 23 July 2026 (23/07/2026). Questions: JRC-PVGIS@ec.europa.eu.