Wind solar complementarity in Hungary, 2020 to 2025

Wind and solar complementarity in Hungary is mostly seasonal. From 2020 to 2025, with each year's output divided by that year's average, the monthly correlation was -0.41, and only -0.15 once the seasonal cycle was removed. At hourly resolution it was -0.14. Solar's best month is July and wind's is February.

Do wind and solar complement each other in Hungary?

Partly, and mostly through the seasons. Hungarian solar and wind output are negatively correlated at every resolution tested, and the relationship strengthens as the time window grows. The Pearson correlation is -0.14 for quarter-hours and for hours, -0.23 for days, -0.31 for weeks and -0.41 for months. Pearson correlation runs from -1 (the series move in opposite directions) through 0 (no linear relationship) to 1 (they move together).

Most of the -0.41 monthly correlation is the seasonal pattern: wind peaks in winter and solar in summer. With each calendar month's 2020 to 2025 average removed, the monthly correlation is -0.15. Month-to-month co-movement beyond the seasons is therefore weak. That matters for firmness and storage sizing, because a month that is poor for one technology is not reliably a good month for the other.

European work reports a similar pattern. A study in the Journal of Applied Meteorology and Climatology by Miglietta, Huld and Monforti-Ferrario used weather-model wind and satellite solar data. It reports a weak anticorrelation at hourly resolution across its European domain. The absolute value increases as the averaging window lengthens.

Correlation between solar and wind output

Pearson correlation of Hungarian public solar and wind generation at different time resolutions, 2020 to 2025, with each year's output divided by that year's average to remove solar's growth. Below zero means they tend to move in opposite directions.

Resolution	Correlation
Quarter-hour	-0.14
Hour	-0.14
Day	-0.23
Week	-0.31
Month	-0.41
Month, seasonal cycle removed	-0.15

Source: Actual generation, load and scheduled cross-border trade for Hungary, ENTSO-E Transparency Platform via the Energy-Charts API

How do wind and solar output differ across the seasons?

Solar peaks in summer and wind in winter, and this offset drives most of the monthly correlation above. Solar's best month is July, with 13.2% of its annual output, and wind's best month is February, with 11.2% of its annual output. In December solar delivers 2.6% of its year and wind 8.6%. In July the figures are 13.2% for solar and 6.6% for wind.

Grouping four months at each end of the year shows the split. From November to February wind delivers 38% of its annual output and solar 14.7%. From May to August solar delivers 49.7% and wind 26%.

Solar and wind output through the year

Each month's share of the year's output, averaged over 2020 to 2025. A flat line at 8.3% would mean an even year.

SolarWind

Dec

Solar2.6 %

Wind8.6 %

0.05.010.015.0

Share of annual output, Jan to Dec. Solar, latest 2.6 %. Wind, latest 8.6 %.

Share of annual output

Month	Solar	Wind
Jan	3.1 %	10.7 %
Feb	5.3 %	11.2 %
Mar	8.3 %	10.8 %
Apr	10.1 %	10.3 %
May	11.4 %	8.2 %
Jun	12.7 %	6.1 %
Jul	13.2 %	6.6 %
Aug	12.4 %	5.2 %
Sep	9.7 %	6.1 %
Oct	7.5 %	8.7 %
Nov	3.8 %	7.6 %
Dec	2.6 %	8.6 %

Month	Solar share of annual output (%)	Wind share of annual output (%)
Jan	3.1	10.7
Feb	5.3	11.2
Mar	8.3	10.8
Apr	10.1	10.3
May	11.4	8.2
Jun	12.7	6.1
Jul	13.2	6.6
Aug	12.4	5.2
Sep	9.7	6.1
Oct	7.5	8.7
Nov	3.8	7.6
Dec	2.6	8.6

Source: Actual generation, load and scheduled cross-border trade for Hungary, ENTSO-E Transparency Platform via the Energy-Charts API

In the figure, a flat line at 8.3% would mean an even year. Neither technology is close to it. Solar swings widely, and wind falls away through the summer months.

How much does each technology vary across the year?

The coefficient of variation (CV) of the twelve monthly shares measures it. It is the standard deviation divided by the mean, so zero would be a perfectly even year. Solar's CV is 0.44 and wind's is 0.24. An equal-energy mix, where each technology supplies half of the same annual energy, has a CV of 0.18.

The lower figure for the mix means its monthly output is more even than either technology's alone. Wind's winter surplus fills part of solar's winter gap, and solar's summer surplus fills part of wind's summer gap. A CV of 0.18 is not zero, so the mix would still have a seasonal shape.

The mix is a calculation, not Hungary's actual system. It would need wind output equal to solar's, about 12 times Hungary's 2025 wind output. In 2025 solar generated 6.81 terawatt-hours (TWh) and wind 0.56 TWh, as the year-by-year table below shows.

The same seasonal logic is reported for Europe as a whole. Heide and co-authors modelled a European system with a very high renewable share. They report that wind power is much stronger in winter and solar power in summer. The two counterbalance each other to a certain extent, and the best mix minimises the energy that has to be stored.

How do wind and solar output differ across the day?

Solar follows the sun, while wind is flatter and slightly stronger at night. The index in the figure is the average output in each Budapest clock hour as a percentage of the all-day average, so 100 means average. Solar's index peaks at 314 between 11:00 and 12:00 and is 0 from 21:00 to 05:00.

Wind's index stays between 84 and 114. It is highest from 22:00 to 24:00 (114) and lowest from 09:00 to 10:00 (84). When solar peaks at 314, wind's index is 90.

Solar and wind output through the day

Average output in each Budapest clock hour, as a percentage of the all-day average (100 = average), 2020 to 2025.

SolarWind

23:00 to 00:00

Solar0

Wind114

0100200300400

Output index, 00:00 to 01:00 to 23:00 to 00:00. Solar, latest 0. Wind, latest 114.

Output index

Hour	Solar	Wind
00:00 to 01:00	0	112
01:00 to 02:00	0	109
02:00 to 03:00	0	106
03:00 to 04:00	0	105
04:00 to 05:00	0	103
05:00 to 06:00	3	101
06:00 to 07:00	20	98
07:00 to 08:00	73	91
08:00 to 09:00	157	85
09:00 to 10:00	237	84
10:00 to 11:00	291	87
11:00 to 12:00	314	90
12:00 to 13:00	312	94
13:00 to 14:00	291	96
14:00 to 15:00	252	97
15:00 to 16:00	197	97
16:00 to 17:00	136	98
17:00 to 18:00	78	99
18:00 to 19:00	31	100
19:00 to 20:00	7	101
20:00 to 21:00	1	107
21:00 to 22:00	0	112
22:00 to 23:00	0	114
23:00 to 00:00	0	114

Show the data table

Hour (Budapest time)	Solar index	Wind index
00:00	0	112
01:00	0	109
02:00	0	106
03:00	0	105
04:00	0	103
05:00	3	101
06:00	20	98
07:00	73	91
08:00	157	85
09:00	237	84
10:00	291	87
11:00	314	90
12:00	312	94
13:00	291	96
14:00	252	97
15:00	197	97
16:00	136	98
17:00	78	99
18:00	31	100
19:00	7	101
20:00	1	107
21:00	0	112
22:00	0	114
23:00	0	114

Source: Actual generation, load and scheduled cross-border trade for Hungary, ENTSO-E Transparency Platform via the Energy-Charts API

Wind's daily swing is shallow next to solar's, which fits the weak hourly correlation of -0.14. In practice wind helps most in the hours when solar produces nothing, and it does little to flatten the midday solar peak. What that peak does to the rest of the system is covered in the duck-curve study, which follows residual load (demand left after wind and solar) by hour.

What is Hungary's wind capacity factor, and when were wind and solar both low?

Capacity factor is a year's mean output divided by installed capacity, here Energy-Charts installed net capacity at the end of the same year. Hungary's wind fleet was 0.32 GW in 2020 to 2022 and 0.33 GW in 2023 to 2025. Its capacity factor ranged from 22.9% in 2021 to 19.4% in 2025, and its output stayed between 0.56 TWh and 0.642 TWh a year. Solar grew from 1.59 TWh in 2020 to 6.81 TWh in 2025 over the same period.

The last column counts days when both technologies were very low. A day counts when solar's daily average output and wind's daily average output are each under 25% of that technology's average daily output for the year. On the definition used, the count was 22 in 2020, 10 in 2021, 8 in 2022, 4 in 2023, 5 in 2024 and 22 in 2025.

Year by year

Public solar and wind generation, the wind fleet's capacity factor, and days when both were very low (under 25% of that year's daily average).

Year	Solar (TWh)	Wind (TWh)	Wind capacity (GW)	Wind capacity factor (%)	Days with both very low
2,020	1.59	0.64	0.32	22.6	22
2,021	2.4	0.64	0.32	22.9	10
2,022	3.08	0.59	0.32	20.9	8
2,023	4.4	0.62	0.33	21.5	4
2,024	5.69	0.63	0.33	21.9	5
2,025	6.81	0.56	0.33	19.4	22

Source: Actual generation, load and scheduled cross-border trade for Hungary, ENTSO-E Transparency Platform via the Energy-Charts API; installed capacity from Energy-Charts

The 2020 and 2025 counts stand out against 4 to 10 days in the years from 2021 to 2024. Wind alone was under the same threshold on 78 days in 2020 and 86 days in 2025, against 61 to 71 days in 2021 to 2024. The count measures days, not runs, so it does not show whether low days came in a row.

What do the profiles imply for adding wind to a solar site?

Wind fills hours and months that solar leaves empty, but it does not lower solar's midday peak. Wind supplies 38% of its annual output from November to February, against 14.7% for solar. From 21:00 to 05:00 solar's index is 0 and wind's is between 103 and 114.

At midday the picture reverses. Wind's index is 90 to 94 in the hours when solar's is 312 to 314 (11:00 to 13:00), so wind is close to its daily average while solar is at its peak. The solar curve here is the national average. A single plant's curve depends on its layout, and the explainer on trackers vs fixed tilt compares measured data.

For storage, two tasks need separating. The daily solar cycle, from 0 to an index of 314 within a day, is a task for batteries that move energy within a day. The days with both technologies very low, 4 to 22 a year on the definition used, are a different task. The monthly correlation of -0.15 beyond the seasons is weak, so a battery sized for the daily cycle does not cover a run of low days.

How this analysis was done

Data and definitions

• Source. Solar and onshore wind generation for Hungary from the Energy-Charts application programming interface (API) of Fraunhofer ISE, which republishes data from the ENTSO-E Transparency Platform. ENTSO-E is the European Network of Transmission System Operators for Electricity. The series are 15-minute averages.
• Period and time. 1 January 2020 to 31 December 2025, in Budapest time (CET/CEST). The window ends in December 2025 so that every year in the seasonal shares is a full calendar year, and the study will be refreshed with 2026 in January 2027. Data refreshed: 10 October 2026.
• Correlation. Pearson correlation of solar and wind output at five resolutions: quarter-hour, hour, day, week (Monday to Sunday) and month. For all but the quarter-hour, both series are averaged over the period before the correlation is taken. Each year's values were first divided by that year's average for the technology, so solar's year-to-year growth does not count as a pattern.
• Seasonal cycle removed. The monthly correlation is repeated after removing each calendar month's 2020 to 2025 average from the monthly values, so the seasonal cycle does not count as a pattern.
• Seasonal shares. Each month's share of that year's output, averaged over the six years. Winter adds the shares for November to February, summer those for May to August.
• Daily profile. The average output in each Budapest clock hour, as a percentage of the all-hours average. Averaging in clock hours blurs the solar peak, because the clock moves by an hour with daylight saving time.
• Days with both very low. Days when solar's daily average and wind's daily average are each under 25% of that technology's average daily output for the year.

Seasonal variation and capacity factor use these formulas:

CV = standard deviation of the 12 monthly shares / mean of the 12 monthly shares
Equal-energy mix share (month) = (solar share + wind share) / 2
Capacity factor = mean output in the year (MW) / Energy-Charts installed net capacity at the end of the same year (MW)

The methods and data sources page lists every public source used on this site. The two CSV (comma-separated values) datasets behind this study, monthly generation and the average hourly profile, are listed at the top of this page.

Limits of this data

• A small, concentrated wind fleet. The fleet is about 0.33 GW. The Hungarian Wind Energy Industry Association (MSZIT) list of operating wind farms puts the large sites in the north-west and west of the country, around Mosonmagyaróvár, Levél, Bőny, Kisigmánd and Ikervár. The wind profile here therefore reflects the weather at a few sites. A 2024 study by the think tank Energiaklub argues that wind farms spread across more regions could give steadier total output. It also says modern turbines at the same sites could reach a clearly higher capacity factor, so the figures above describe the existing fleet, not new turbines.
• The public solar series. The public solar series (reported to ENTSO-E by MAVIR) covers utility-scale plants and leaves out most household and other small rooftop systems, so solar figures here are lower than total solar output. Why published Hungarian solar capacity figures differ gives the detail.
• Correlation is not firm capacity. A negative correlation describes how output moves on average. It does not say how much capacity can be relied on at the worst moment, and the days with both technologies very low show that both can be weak together.
• A short record. Six years is a short record. Wind's capacity factor alone ranged from 19.4% to 22.9%, so another six years could look different.
• No plant model. This study does not model a specific plant, a connection limit, curtailment or a battery, so its points are readings of the profiles, not a design.

What this means for developers, grid planners and investors

Developers

• Test a wind addition against the hours solar leaves empty: 21:00 to 05:00 (wind's index 103 to 114) and November to February (38% of wind's annual output, against 14.7% for solar).
• Do not count on wind to lower solar's midday peak. Wind's index is 90 to 94 when solar's is 312 to 314.
• Use site-specific wind data. This national profile reflects a few north-western sites and the existing turbines.

Grid planners

• Treat the daily solar cycle (index 0 to 314) and the days with both technologies very low (4 to 22 a year on the definition used) as two separate planning tasks.
• Plan for the seasonal offset, since wind supplies 38% of its annual output from November to February. Do not plan for month-to-month offsets: the monthly correlation beyond the seasons is -0.15.
• Read the daily shape against residual load in the duck-curve study.

Investors

• Compare wind capacity factors on the same capacity base. Here the base is Energy-Charts installed net capacity at the end of the year, and the range was 19.4% to 22.9%.
• Do not assume a wind and solar mix removes low-output days. The weak co-movement beyond the seasons (-0.15) and the both-low counts (4 to 22 a year on the definition used) point the other way.
• For prices at 15-minute resolution, see the study of negative quarter-hours in the first year of 15-minute day-ahead prices.