Disclaimer: This article is republished with permission from the author, Johan Rockström. The original articles were published on LinkedIn in three separate parts. Part 1 can be found here. Part 2 can be found here. Part 3 can be found here.  Any views expressed in this article are those of the original authors and do not necessarily reflect the views of Altiorem.

Fig. 1: Emissions pathways with same carbon budget. Solid lines show gross emissions (black) & removals via CDR & land use. Source: J. Rockström

Part 1: The risks we don’t know we’re taking

In this series of posts, I’d like to give you my take on where science stands on Earth system risk, action to avert climate change and the path forward.

1.5°C physical limit not target

Science is increasingly clear that a 1.5°C departure away from the Holocene average global mean surface temperature (GMST) on Earth of 14°C is a physical limit. It is not a goal or a target, as it is often treated in policy, media and general debates. Go beyond 1.5°C warming, a place our civilisations have never experienced and Earth has not seen in the last 100,000 years, and we will not only see major human and economic damage, but are likely to cross multiple Earth system tipping points, causing amplified warming with billions of people affected (Armstrong McKay et al. 2022).

And now we are very close to breaching the 1.5°C limit, likely even earlier than 2030-2035 as predicted in IPCC AR6 middle-of-the-road scenarios (Bevacqua, Schleussner, and Zscheischler 2025; Cannon 2025).

This calls for a probing safety check. Are our targets for staying within the 1.5°C guardrail really enough? Or are we cheating ourselves, when we think that if all countries and companies join the “best in class” by following the carbon law (i.e. cutting emissions by half each decade, Rockström et al. 2017), we will have a reasonable chance of a “safe landing”? This is the conviction that guides both the voluntary climate action efforts (e.g. the over 7000 companies with science-based targets as part of the Science Based Targets initiative SBTi) and the European Green Deal and its “Fit-for-55” policy (stipulating emissions cuts in the EU of 55% by 2030), and one that, admittedly, I was deeply involved in crafting. But today, we have several reasons to question if these science-based targets are enough.

Let me briefly walk through the reasons for concern, putting into question even our (current) best climate efforts. As I go along, keep in mind that, globally, we are depressingly far from following even the carbon law, which would require going from increasing global emissions by around 1% per year (as we do today) to immediately starting to reduce fossil-fuel emissions by some 5% per year.

Remaining carbon budget and what is tolerable risk?

The first reason for concern is, of course, that we have largely consumed our entire global carbon budget. The remaining carbon budget (RCB) is the total amount of future carbon dioxide emissions that can stay in the atmosphere without crossing a prescribed global warming guardrail (e.g. 1.5°C). The size of the budget also depends on the level of confidence you want to have that you won’t transgress the guardrail: higher confidence, lower RCB and vice versa.

The RCB was estimated by the IPCC in the most recent assessment in 2021 (AR6) at approximately 500 GtCO2 (for a 50% chance of holding the 1.5°C limit). But we have since continued to “eat up” approximately 40 GtCO2 per year, and, combined with some additional updates and adjustments, are now down to around 200 GtCO2 as of the beginning of 2025 (estimates range from 160GtCO2 up to 310GtCO2; Friedlingstein et al. 2024).

But, if we take the risks of crossing 1.5°C seriously, then the question is the following: are we really taking responsibility for the future of our children and all future generations on Earth by “tossing a coin” (50% chance of success)? My answer is no. Even a 67% chance gives an uncomfortably high likelihood of failure. But OK, let’s aim for the slightly loaded dice in our favor. What does this give us in terms of RCB? Something around 150 GtCO2. At the current rate of emissions, this will be exhausted within 4 years!

If we take these numbers seriously, which we should, it translates into a completely different speed of transition than what we are generally talking about. Staying within a 150 GtCO2 budget (for a 67% chance of success) would require removing around 6 GtCO2 per year (~15% of today’s total emissions) to reach zero emissions just after 2030. This is obviously completely unrealistic – even if we decide to shut down the global economy “à la COVID pandemic”, which temporarily, and for the wrong reasons, cut global emissions by 6-7%.

So, what are world leaders banking on?

CDR assumptions allow for higher gross emissions

The way science and policy have handled this conundrum so far is to assume/hope that carbon dioxide removal technologies (CDR) will help us to an “orderly” transition towards a safe landing around 1.5°C, rather than a global crash landing at a much higher temperature.

You can think of the RCB as a bucket being filled with greenhouse gas emissions. Once the bucket is full, the RCB is exhausted. If, however, there is some way of getting rid of the emissions through a valve at the bottom of the bucket, say using carbon dioxide removal (CDR), the bucket will fill more slowly and the total amount of emissions used to fill the bucket (gross emissions) will be higher than the volume of the bucket itself (net emissions).

Here we stumble across one of the confusing (and disturbing) points in the climate debate. The carbon budget is generally used to describe the net remaining fossil-fuel emissions of CO2 (including CO2 emissions from agriculture). On the other hand, the policy targets and decisions on the pace of emission reduction (i.e., the entire mitigation policy agenda) allow for almost twice as much gross emissions as the budget allows, because they assume that CDR will do much of the heavy lifting by drawing down massive amounts of carbon from the atmosphere. In other words, the world’s most ambitious emissions reductions targets (prescribing fast, but still far-from-sufficient reductions) correspond to gross emissions that would blow the budget, if it weren’t for the assumed removal of greenhouse gases from the atmosphere.

Take a look at the two 1.5°C-compatible gross emissions scenarios in Figure1 above: one relies on CDR (the solid line) and the other does not (dashed). Both scenarios lead to the same overall carbon budget, but the gross emissions in the scenario that relies on CDR are considerably higher than in the scenario without CDR (the difference is shown by the area between the dashed and solid lines) – a difference assumed to be compensated for by substantial removals through CDR and in land systems (see the red curve). It is immediately clear how much more rapidly emissions need to be reduced if CDR is employed.

So, if you assume a lot of CDR technologies (DAC, CCS, BECCS) and believe in a rapid agricultural transformation, you can stay within the carbon budget while tolerating relatively high gross fossil fuel emissions. This is the status quo and it means that gross emissions reductions targets alone (without CDR) do not align with the RCB. Even if we are able to achieve the current gross emissions reductions targets, failing on CDR will leave us blank.

The remaining carbon budget should instead be seen as a credit, with CDR and negative emissions from agriculture offering a second line of credit amounting to hundreds of GtCO2 according to middle-of-the-road scenarios. So, betting on CDR gives us an imaginary path towards a safe and orderly landing. The combined credit can more than double the allowed emissions for a 50% chance of holding 1.5°C, letting us believe that removing around 3 GtCO2 or 7.5% of current emissions from the global economy per year (UNEP Gap Report 2024) will be enough.

Feasibility of CDR assumptions

But is it realistic that we actually receive the second credit? Can global agriculture abruptly transition from the world’s single largest emitting economic sector in society, to become a major carbon sink, in less than 10 years? Is it really possible to scale CDR from millions of tons of pilot schemes today to billion tons of removals within years? So far, it looks like a long shot. However, the above quick walk-through of the numbers shows one thing beyond any doubt: whether we like it or not, we have burned such a large share of the RCB, and at the same time have so much evidence that we need to do everything we can to hold on to 1.5°C as a hard limit, that we have no choice but to massively scale up CDR and shift agriculture from source to sink, together with dramatically accelerating the phase out of fossil-fuel emissions (at a rate of more than 7.5% per year, following the 2024 UNEP Gap Report on required gross emission reductions, i.e., factoring in assumptions that CDR is scaled).

In my next post I’m going to look at the other very important assumptions on which the RCB hinges, and what the Earth might be telling us about the validity of these assumptions.

Fig. 2: Illustration of assumption of a near-linear relationship between cumulative emissions & global warming. Source: IPCC AR6 WGI SPM Fig10, adapted

Part 2: What the Earth might be telling us about resilience and climate sensitivity

In my last post I walked through the reasons for concern about whether the current set of climate targets, based on the remaining carbon budget (RCB) are enough to keep long-term global warming below 1.5°C. Now I’d like to take a look at three key assumptions underpinning the RCB:

1. All non-CO2 gases (methane, nitrous oxide as the prime ones) follow a declining path along with fossil-fuels.
2. The linear relationship between cumulative carbon and temperature (the Transient Climate Response to Cumulative Carbon Emissions, TCRE – see plot above) holds also in the coming years/decades.
3. Earth system feedbacks do not deviate beyond the stable conditions we have experienced over the past 10 000 years, i.e., climate sensitivity remains at around 3°C for a doubling of CO2 concentration (i.e., no shifts in Earth system resilience, and no tipping points are crossed).

Warming acceleration  – observations

These are the basis upon which science provides us with a RCB of approximately 200 GtCO2 for a 50% chance of holding the 1.5°C limit. Without points 1-3 above being correct, the budget disappears. These are optimistic assumptions. Especially today, when we are experiencing what may be an acceleration of warming, with indicative trends showing a jump from 0.18 °C/decade from 1970-2014 (Hansen et al. 2023) to 0.27°C/decade for the 2014-2023 decade (Forster et al. 2024). Sea-surface temperature shows the same trend, with an abrupt shift in rate of warming over the past decade (Merchant, Allan, and Embury 2025). Added to this we have the record braking 1.6°C above the pre-industrial average for the year 2024, i.e., that we for the first time breach 1.5°C over 12 consecutive months.

We know that we’ve just come out of a human-amplified (Cai et al. 2023) El-Nino, so couldn’t this all simply be a freak event on top of the steady (0.18°C/decade) warming we expect? With a high degree of certainty, the answer is no. The observations of a – off the charts – 0.3°C jump in the atmosphere and the ocean, cannot be explained only by a linearly reinforced El Nino. Something more is at play. This concern is further reinforced by the unexpected observation that January and February 2025 – which should have been a “cool” start of the year, as we have now entered a La Nina phase, were warmer than January 2024 (in the midst of the super El Nino). In fact, January 2025 is the warmest January ever recorded (1.75°C above pre-industrial temperature) and February was still anomalously high. This does not follow the standard climate script.

So, what is happening? The jury is still out, and many candidates have been brought forward.

Aerosol emissions

Helge Goessling (2024) and colleagues concluded that high solar cycle intensity, the 2022 sub-marine Hunga-Tonga volcanic eruption and new sulfur regulations on commercial shipping (reducing cooling sulfate aerosols), together only can explain 1/3 of the abrupt shift, i.e., 0.1°C. Jim Hansen et al. (2025) disagree, placing a 2/3 of the abrupt warming on the additional forcing caused by reduction in aerosol loading (air pollutants, in particular sulfates and nitrates, which act as reflective mirrors for incoming shortwave radiation, thus dimming the lower atmosphere). Here, I lean more towards the Goessling argument.

Importantly though, irrespective of the magnitude of the effect of aerosol loading, the reduction of these air pollutants would increase the overall climate forcing (removing, in Jim Hansen’s language, the Faustian Bargain of us cancelling out GHG-driven heating with cooling from air pollution). Keep in mind that the avoided air pollution shortens the lives of 7-8 million people each year (btw, a rate of lives lost that is in the same order of magnitude as during the COVID pandemic).

Earth system feedbacks

Two (main) candidates remain. Both are Earth system feedbacks, i.e., factors that change the basic Earth response to human-caused forcing, making them factors that not only push us higher up along the linear TCRE line, but also shift the slope of the line. This would very likely translate to higher climate sensitivity – very bad news indeed!

These are:

• Earth albedo change. As shown by Goessling et al., Earth is getting darker and absorbing more heat, which is explained by a trend of a pronounced decline in low-level clouds in the Northern hemisphere, which decreases the reflectivity of incoming short-wave radiation. The northward shift of tree lines as the planet warms, and accelerated melting of snow and ice, also contribute to making the planet darker.
• Relative reduction in carbon uptake in the biosphere, particularly with signs of a weakening carbon sink in boreal (Virkkala et al. 2025), temperate (Burton et al. 2024) and tropical forests. The southeastern part of Earth’s richest ecosystem on land, with a carbon stock exceeding 100 GtC, has already tipped over from sink to source (Gatti et al. 2021). Estimates show that the global land sink, which normally averages almost 10 GtCO2/yr (25% of global fossil-fuel emissions), reduced to less than 2 GtCO2 in 2023 (Ke et al. 2024). This fall off was caused primarily by droughts (over the Amazon) and fires, which are expected to increase during El Nino years (such as in 2023), but here we experienced an unexpectedly high loss of carbon uptake. The result is more carbon in the atmosphere – an Earth system feedback that increases climate forcing and accelerates warming. If this feedback persists, it would imply a higher climate sensitivity.

Earth losing resilience

The worry is this. Even though we do not know precisely which combination of these effects is ultimately responsible for the unexpectedly high global temperatures, we have received enough concerning signals from the Earth system, forcing us to seriously ask the question, are we seeing the first signs of Earth losing resilience?

That is, is its biophysical capacity to buffer the stress caused by Anthropogenic damage (i.e. excess fossil emissions) to the system eroding?

The most recent estimates already point to implications of a weaker planet showing first signs of accelerated warming. The 1.5°C limit will be breached earlier, probably already before 2030 (Cannon 2025).

Risk of crossing tipping points

And the BIG question out there is what does all this mean for the risk of crossing tipping points in the Earth system? We already have evidence that multiple tipping elements are likely to cross their thresholds when 1.5°C is breached permanently. This places us in a very delicate situation, given that these tipping elements (Tropical Coral Reef systems, the Greenland Ice Sheet, the West Antarctic Ice Sheet, abrupt thawing of permafrost, and collapse of the Barent sea ice) would not only affect billions of people, but comprise feedback systems, i.e., they can trigger permanent changes in the functioning of Earth, which would accelerate warming even further.

Again, the scientific jury is still out. But increasingly science shows (in addition to the 5 tipping elements already mentioned), that both the Amazon basin and the AMOC, are also at risk of crossing tipping earlier than previously thought, when factoring in all interacting drivers (e.g. for the Amazon: climate forcing + deforestation + biodiversity loss + shifts in hydrology; for the AMOC: slowdown of the thermohaline pump not only due to dilution from massive inflow of freshwater from melting of the Greenland Ice Sheet, but also from increased precipitation over the North Atlantic).

Climate sensitivity

These are all Earth system feedbacks, which impact on Earth’s climate sensitivity. There has been a lot of discussion around the “holy grail” in climate science, the Equilibrium Climate Sensitivity (ECS) – how much warming is to be expected at equilibrium after a doubling of CO2. Since Jules Charney and the Charney report’s first estimate in 1979 (3°C ± 1.5), the uncertainty around the ECS has remained stubbornly large, with a range of 1.5-4°C. There is some convergence around an average of 3°C (which is the ECS level that best reproduces observed global warming so far).

But 3°C is far from the end of the ECS story. Already in 2008 Jim Hansen concluded that while ECS due to fast feedbacks may be 3°C, the real ECS when factoring in slow feedbacks, like albedo shifts and changes in the cryosphere and ocean heat balance, is more likely 6°C. Several GCMs in the CMIP6 round showed ECS levels > 4°C, mostly as a result of more detailed representation of cloud dynamics (i.e., that better understanding of cloud feedbacks results in amplified warming due to changes in Earth’s albedo, confirmed by the Goessling et al. (2024) work). Importantly, none of these high-ECS simulations include tipping points, only corrections in slow-shifting changes related to, for example, cloud dynamics and thawing permafrost.

So, in summary, while this series focused on the fact that even the most ambitious climate action is far from aligned with science (i.e., will fail on 1.5°C), this post describes how the Earth system is telling us that the remaining carbon budget is in fact even smaller. Our Earth may be losing resilience. We are seeing signs of accelerated warming over the last decade and abrupt deviations from predicted trends in the last years, which may suggest an Earth system shift in climate sensitivity. Higher climate sensitivity would mean a lower remaining carbon budget, and would erase any chance of an orderly phase out of fossil-fuels and a safe landing for humanity on a planet that can continue to provide decent life for all humans in the world.

In the next post (Part 3), I will address what all this means for our efforts of securing life support for humans on Earth.

Fig. 3: Based on Bennett et al. (2016).

Part 3: What does this all mean?

In the previous two posts in this series, I addressed two issues:

1. We are confronted with the dilemma that even our best climate efforts are insufficient to have any chance of holding the 1.5°C global warming limit.
2. There is deep scientific worry that Earth may be losing resilience, resulting in feedbacks that not only would continue to “bump up” warming, but also reduce the capacity of the planet to buffer human abuse to the system (i.e., leading to a shift in climate sensitivity, and thus reduce even further the remaining carbon budget for an orderly transition to a safe operating space on Earth).

In this final post, I’d like to share a few thoughts on what all this implies for our actions now and in the coming years. In my view, the current scientific diagnosis of the state of the planet leads to two main conclusions:

• We cannot exclude that Earth is losing resilience, which will manifest in higher Earth climate sensitivity (ECS). This does not mean that current estimates of ECS for the past are wrong, but rather that when multiple planetary boundaries are breached simultaneously – weakening the buffering capacity of the biosphere, ocean, and cryosphere – Earth may shift into a regime with a higher ECS. The fact that we are in unchartered waters – paleo-climatically and geophysically – matters. We are currently witnessing an Earth Energy Imbalance (accumulating more heat on Earth than is reflected back to space) on a planet that, already at the starting point, is in a warm, inter-glacial equilibrium (Holocene). Put another way, we are shifting from a warm (natural) state, to an even warmer trajectory, forcing Earth to move from a warm equilibrium to even warmer conditions. The glacial-interglacial cycles of the past one million years don’t offer a play book for such a shift: it is outside of the Pleistocene logic.
• We have an unfortunate convergence of facts: (1) running out of carbon budget; (2) 1.5°C is a limit, not a goal to negotiate with; and (3) a weaker planet means lower capacity to buffer human caused warming, which results in an even smaller remaining carbon budget. The only responsible action to 1+2+3 is, in my view, to take the planetary crisis seriously, and do everything we can to deviate away from the danger that is hitting us faster than we had expected.

In the UNFCCC, all countries in the world have agreed to avoid dangerous climate change. Well, we are now in the midst of danger. Already in the coming years and decades we can expect unmanageable outcomes, if we do not radically shift course, now.

The path to a solution is quite straight forward:

• Implement the Dubai agreement (signed by all countries) to accelerate the transition away from fossil-fuels this decade. Step one is to remove 3-4 GtCO2 from the global economy in 2025. This would be the necessary start on a journey to cut-global emissions by 50% by 2030. Still insufficient, but will very likely lead to self-acceleration of the transformation away from oil, coal and gas.
• Put an end-date on all fossil-fuel driven transport (as it has been done in the EU – stopping the sales of combustion engine cars by 2035).
• A global decision to forbid all investments in new oil, gas and coal extraction and utilities.
• Introduce a global price on carbon of 100-200 USD/ton (constrained and rising) for all sectors.
• Halt the expansion of agriculture into the remaining natural forests and other land-based ecosystems (approximately 50% of Earth’s land surface).
• Legally declare the ocean (its biology, sea-floor, and heat-conveyor belt) and all tipping point systems as global commons, and compensate tipping point hosting countries for their service to humanity (if they preserve them). Start at COP30 with the Amazon.