Back to ErdblickDeutsch

What we know – and what we do not

Erdblick shows numbers for the years 2026 to 2050. Not all numbers are equally certain. Some were measured. Some were calculated by a research institute. Some we extended ourselves.

This page says which number is which kind, what is missing and what that means for you. Our rule: we do not invent numbers. When we extend something, we say so.

This text is written once and fixed. An AI does not write it anew on every visit. The years and values come from the same files as the cards.

Three kinds of numbers

How to tell on the main page: a solid line is measured. A dashed line is calculated or extended. Under “Source”, every card says which kind it is.

Putting numbers together is hard

Two sources can use the same word and mean something different. Two examples from this site:

Forest. There is no single rule for what a “forest” is. The FAO counts an area as forest if it is larger than half a hectare, the trees are taller than 5 metres and the tree crowns cover more than 10 % of the ground. The land must also be used mainly as forest. Fields and cities do not count. Satellite maps such as Global Forest Watch count differently: every pixel with trees taller than 5 metres and crowns covering at least 30 % counts, plantations included. The model we used for forests earlier assigns a land cover to each pixel. For 2020 it was about 290 million hectares away from the FAO. None of these numbers is wrong. They measure different things.

Greenhouse gases. Some sources count only CO₂. Others count all greenhouse gases and convert them into CO₂ equivalent. Over 100 years, methane has the effect of 27 tonnes of CO₂, for example. Both numbers are often called “emissions” but they are not the same. That is why “Pessimistic” carries a note, and why we do not compare it with the other two scenarios.

The rule behind this: two numbers may only be compared if they measure the same thing. When we are not sure, we say so.

Evidence: FAO, Global Forest Resources Assessment (definition of forest); World Resources Institute, Global Forest Review (satellite maps): terms and methods ↗

Why is there not a measurement everywhere for 2026?

On the main page you can choose a year. The first year is 2026. But not every topic has a measured value for 2026 yet. Some series end earlier. Then the card shows a calculation or an extension for that year.

What do “Optimistic”, “Moderate” and “Pessimistic” mean?

These are three possible paths into the future. For each topic they came about differently. That matters when you compare them.

The topics one by one

Greenhouse gas emissions

What we have: For Optimistic and Moderate: pathways from the Climate Action Tracker (as of 13 November 2025). The values are in “CO₂ equivalent”: all greenhouse gases, converted into their warming effect. The last real value is 2023, after that these are pathways. “Moderate” is the average of the upper and lower bound of the tracker. The tracker gives no value of its own for this. For Pessimistic: an older IIASA model (2018 release) with the anchor years 2020, 2030, 2040, 2050. It counts only CO₂, not the other gases.

What is missing: For “Pessimistic” there is no current pathway without climate policy in CO₂ equivalent. The tracker describes only policy that has been implemented or pledged. We have not yet added a long series from the past here.

What that means for you: You cannot compare “Pessimistic” directly with the other two. The number can be lower than for “Moderate” even though the path is worse. The reason: different gases, different model, different vintage. The card says so too.

Temperature rise

What we have: Measured annual values from 1850 to 2026 (Met Office Hadley Centre, HadCRUT5 series). The value for 2026 is an interim figure: we fetched the data on 9 September 2026, and the year was not over yet. From 2027 we use the scenarios of the institute IIASA. We take only how much it changes from year to year and add that change to the 2026 measurement (1.40 °C). That way the card has no jump.

What is missing: The model knows only a few anchor years (2030, 2040, 2050). In between it draws a straight line. In reality the temperature swings from year to year. The warmest year in our series is 2024 (1.53 °C). 2025 was 1.41 °C.

What that means for you: The line from 2027 shows a path, not a forecast for a single year. Single years deviate upwards and downwards. By 2050 the three paths range from about 1.8 to 2.4 °C above the pre-industrial level.

Renewable energy

What we have: The share of hydropower, wind, solar, bioenergy and other renewable sources in worldwide electricity generation. Measured from 1900 to 2025 (Ember, via Our World in Data), most recently 33.8 %. For the future we use three scenarios of the International Energy Agency (IEA, World Energy Outlook 2025): Pessimistic is “Current Policies”, Moderate is “Stated Policies”, Optimistic is “Net Zero 2050”. The IEA gives values for 2035, 2040, 2050.

What is missing: The free IEA dataset has no value for 2030. So the years between 2025 and 2035 are a straight line of ours. The card shows only electricity. Heating and transport are not included. The share of all energy is a different value, and we do not show it.

What that means for you: The number 33.8 % means here: 33.8 out of 100 parts of electricity. Not: 33.8 out of 100 parts of energy. The value for 1900 (41.2 %) looks high. Back then electricity generation was small and came mostly from hydropower.

Sea temperature

What we have: Measured annual values from 1950 to 2026 (Met Office Hadley Centre, HadSST series). 2026 is also an interim figure (data fetched on 8 September 2026). All values are the difference from the average of the years 1861 to 1890.

What is missing: We have no values before 1950. For the future there is no institute calculation we can use. We took the trend of the years 1997 to 2026: 0.0185 °C more per year. We continue this straight line.

What that means for you: The line comes from us, not from experts, and is the same for all three scenarios. The number for 2050 only means: this is how much it would be if the trend of recent years simply carries on.

Species extinction

What we have: The Red List Index from BirdLife International and the IUCN, taken from the UN database. Annual values from 1993 to 2025: from 0.806 to 0.745. A value of 1 means no risk of extinction. The smaller the value, the greater the risk.

What is missing: The index covers only five groups: mammals, birds, amphibians, corals and cycads. Other groups, for example insects, are not included. For 2050 there is no official forecast. We take the change of the years 2015 to 2025 (minus 0.0023 per year) and continue it as a straight line. Optimistic: half as fast. Pessimistic: one and a half times as fast. These are our assumptions.

What that means for you: The index shows a direction for these five groups. It does not say how many species disappear in total. The numbers for 2050 show what happens if the trend carries on, not what will happen.

Water demand

What we have: The amount of fresh water withdrawn worldwide: 3,853 billion cubic metres in 2022 (World Bank, data from the FAO). This is the latest world value there is.

What is missing: Only this one measured year. There is no annual series, so the card has no look back. For 2050 the UN water report 2024 gives a range: 20 to 45 % more. We take 20 % for Optimistic, 30 % for Moderate and 45 % for Pessimistic and draw a straight line from 2022 to 2050.

What that means for you: If you choose 2035, for example, the card shows a point on this line. It was not calculated for the year 2035. The percentages are a range from a report, not a forecast for every year.

Forest area

What we have: The worldwide forest area according to the FAO (Food and Agriculture Organization of the UN), annual values from 1990 to 2025: from 4,343.5 to 4,140.2 million hectares. Where the FAO did not survey a year itself, Our World in Data filled the gap with a straight line.

What is missing: For 2050 there is no official forecast that fits the FAO numbers. We take the loss of the years 2015 to 2025 (4.12 million hectares per year) and continue it as a line. Optimistic: half the loss. Pessimistic: one and a half times the loss. These are our assumptions.

What that means for you: How large the forest area is depends on how “forest” is recorded. We used to use an IIASA model. For 2020 it was about 290 million hectares away from the FAO. That is why we switched to the FAO. Compare forest numbers only if they come from the same source.

Plastic production

What we have: The card shows how much plastic is used worldwide (the card calls this “production”). The numbers are OECD model estimates (Global Plastics Outlook 2022), annual values from 1990 (130 million tonnes) to 2019 (460 million tonnes). For the future we use the three OECD scenarios: “Baseline” (our Pessimistic), “Regional Action” (Moderate) and “Global Ambition” (Optimistic). The card also shows plastic waste from the same source: in 2019 around 353 million tonnes arose, around 79 million tonnes of it mismanaged or littered.

What is missing: These too are estimates from a model, not measurements. Waste and mismanaged waste exist only up to 2019. Waste projections exist in the dataset only for the baseline scenario, so the card does not show waste for the future. The number many people are really looking for is missing: how much of it ends up in the environment. There is no worldwide measurement for that, only estimates, which sometimes differ by a multiple. The OECD itself writes that data to check its models are lacking. Earlier the card showed a series from 1950 (Geyer et al. 2017 via Our World in Data) in which the years 2016 to 2018 were only extended, and 2050 values we could not substantiate. We replaced both with the continuous OECD series. That is why we no longer have values before 1990.

What that means for you: Use, waste and waste in the environment are three different numbers. Made is not the same as thrown away, and thrown away is not the same as landed in the environment. The OECD scenarios show policy paths: in 2050 use is 674 million tonnes in the “Global Ambition” scenario and 976 million tonnes in the baseline.

World population

What we have: Numbers from the United Nations (World Population Prospects 2024) for the years 1900 to 2023. For the future we use the UN “medium variant”. It gives values for 2027, 2030, 2040, 2050. In between we draw straight lines.

What is missing: Even the years from 2024 onwards are a projection. The UN also calculates higher and lower variants. We show only the medium one because only it was freely available as a continuous series. So we show no range of uncertainty. Population is the same in all three scenarios, we do not tie it to the climate.

What that means for you: The number for 2050 (9.7 billion) is the UN medium expectation, not a certain value.

Water stress

What we have: How many people live in countries under high or very high water stress. High water stress means: a country withdraws at least 40 % of the renewable water available to it. The country rating comes from the World Resources Institute (WRI, Aqueduct 4.0, 2023), the number of people from the World Bank (2024). Today (2024) that is 2.93 billion people, about 36 % of the world population.

What is missing: This is a model, not a measurement at every place. The WRI itself says Aqueduct is a tool for assessing risk. We have no annual series, only three points: 2024, 2030, 2050. In between we draw straight lines. We calculate per country: a country counts in full, even if only part of it suffers water stress. Population stays frozen at the 2024 level. So the card only shows how the rating of countries shifts, not that there are more people.

What that means for you: In 2050 the three scenarios are almost the same (2.95 to 2.98 billion people). “Optimistic” is even slightly higher than “Pessimistic”. That is not a calculation error. Water stress depends strongly on irrigation and industry, not only on the climate.

Displacement by natural disasters

What we have: Annual values from 2008 to 2025 from the Internal Displacement Monitoring Centre (IDMC). In 2025 there were 29.9 million new displacements. Cases are counted, not people. Someone displaced twice in the same year counts twice. All natural disasters count, earthquakes too. In 2025, 27.2 of 29.9 million were weather-related.

What is missing: There is no forecast, and we cannot make a serious one of our own. The values swing strongly depending on major events: 2024 had 45.8 million, 2025 only 29.9 million. So we continue the last value horizontally.

What that means for you: The flat dashed line does not say “it stays like this”. It only says: we do not know. A future year can be much higher or much lower.

Words we use

Scenario:
a possible path into the future under certain assumptions. Not a forecast.
Anchor year:
a year for which a source gives a real value. We connect the years in between with a straight line.
Interim figure:
a year that was not over yet when the data was fetched. The value can still change.
Pre-industrial:
before industrialisation, roughly before 1850. For temperature and sea temperature the zero point is the average of the years 1861 to 1890. The IIASA scenarios refer to roughly 1850 to 1900. The difference is only one to two hundredths of a degree. For the other topics we write “oldest measured value” and the year, because no such zero point applies there.
CO₂ equivalent (CO₂e):
all greenhouse gases, converted into their warming effect.
Red List Index:
measures how strongly species are threatened with extinction. 1 means: no risk.

Found a mistake?

If you find a mistake or know a better source, write to us:

Data as of: fetched between 8 September 2026 and 17 October 2026.