Causes of Climate Model Differences

Climate models apply the same greenhouse gas forcing coefficient to every ecosystem on Earth. A rainforest and a desert get identical treatment. Using 254 flux tower sites across 7 biomes and 4 continents, I measured what actually happens. The effective forcing varies by a factor of four. Forests attenuate it by up to 50%. Arid shrublands amplify it by up to 24%. The uniform assumption introduces 53.7% error. One takeaway: when a forest is cleared and replaced by degraded land, the effective radiative forcing at that location triples. This is independent of carbon emissions. It happens the day the trees fall. Code: https://lnkd.in/drwSsxuN Preprint: https://lnkd.in/dZshdF8k #ClimateScience #RadiativeForcing #FLUXNET #Ecosystems #LandAtmosphere

Can climate models reproduce observed trends? The answer can be challenging. Our new review paper in Science Advances led by Isla Simpson and Tiffany Shaw discusses challenges and ways forward in confronting climate models and observations. It's tricky. Climate models and observations may disagree (1) by chance, due to unforced internal variability, (2) due to error in the model response, (3) due to inaccurate prescribed external forcings, (4) due to incomplete or uncertain observations or (5) due to inappropriate comparison methods. The paper discusses ways forward in disentangling the reasons for potential mismatches between observed and simulated trends. It provides a long catalogue of examples of success, discrepancies and unclear situations that require further attention. https://lnkd.in/dHrEJfDh Let by Isla Simpson and Tiffany Shaw with Paulo Ceppi, Amy Clement, Erich Fischer, Kevin Grise, Angeline Pendergrass, James Screen, Robert Jinglin Wills, Tim Woollings, Russell Blackport, Joonsuk Kang, and Stephen Po-Chedley supported by US CLIVAR

🌿 𝗡𝗚𝗙𝗦 𝗷𝘂𝘀𝘁 𝗽𝘂𝗯𝗹𝗶𝘀𝗵𝗲𝗱 𝘀𝗼𝗺𝗲𝘁𝗵𝗶𝗻𝗴 𝘂𝘀𝗲𝗳𝘂𝗹 𝗳𝗼𝗿 𝗮𝗻𝘆𝗼𝗻𝗲 𝘄𝗼𝗿𝗸𝗶𝗻𝗴 𝗼𝗻 𝗰𝗹𝗶𝗺𝗮𝘁𝗲 𝗿𝗶𝘀𝗸. The Network for Greening the Financial System (NGFS) released a paper showing how two different economic models react to the same climate shocks. This matters because companies struggle to turn climate scenarios into actual numbers that show up on balance sheets. 𝗪𝗵𝗮𝘁 𝘁𝗵𝗲 𝗽𝗮𝗽𝗲𝗿 𝗰𝗼𝘃𝗲𝗿𝘀: • Three scenarios over five years: smooth carbon pricing, sudden carbon pricing, extreme weather events • How different assumptions (how fast people react, how detailed the industry breakdown, how central banks respond) change the results • Why the same carbon tax can produce different effects on GDP, inflation, and interest rates 𝗞𝗲𝘆 𝗳𝗶𝗻𝗱𝗶𝗻𝗴𝘀: The model you choose shapes the answer. Two credible models: same inputs > different outputs. Sometime around 2030 they show similar total damage but completely different timelines. How you assume people react matters. If companies and investors see the carbon tax coming, they adjust early. If they do not, the adjustment happens later and differently. Physical risk is harder to model. Extreme weather shocks show up weaker and slower in these models, which probably reflects how hard it is to measure rather than low actual impact. 𝗪𝗵𝗼 𝘀𝗵𝗼𝘂𝗹𝗱 𝗿𝗲𝗮𝗱 𝘁𝗵𝗶𝘀: • Finance and risk teams running climate scenarios • Anyone preparing climate disclosures under ESRS or TCFD • Anyone trying to explain to senior leadership why different models give different answers Climate risk modelling is uncertain. This paper makes that uncertainty visible and shows how your methodological choices affect what you report. This is a valuable read if you are working through climate scenario analysis or moving climate risk into your enterprise risk framework. 🔗 𝗟𝗶𝗻𝗸 𝘁𝗼 𝘄𝗲𝗯𝘀𝗶𝘁𝗲: https://lnkd.in/eWvsxt8v

This recent report from Global Association of Risk Professionals (GARP) is making the rounds and creating a lot of buzz. In it, outputs from 13 established climate risk vendors are compared (#Iris from Class 3 Technologies is too new to be included but we'll be game for the next one!) for a portfolio of global properties. The dispersion in the results, for an individual property, are stark but unsurprising. For the same location, some models predict no flooding while others predict several meters, cyclonic wind speeds can vary from Cat 1 to Cat 5, and damage and loss predictions are all over the map! This report validates what we already know, that model outcomes will vary amongst providers, causing confusion for consumers. But it doesn't really get to the root cause. I think there are a few primary drivers for the divergence that are important to be aware of: 1. Spatial resolution of the hazard - especially for hyperlocal hazards like flooding - varies widely amongst providers. And none are really meant to be site-specific, you would need very specific details on drainage and grading for that. 2. The treatment of properties as locations with buffer zones rather than as actual buildings with physical footprints can lead to inaccuracies. For instance, some providers mistakenly placed properties in the middle of the North Sea due to reliance on geolocations without verifying address/property. 3. Almost all providers use historical claims-based damage functions that were developed to serve the insurance industry where average loss estimates across portfolios of thousands of buildings are sufficient to price insurance exposure but a very blunt instrument to characterize risk at an individual building. 4. Some of the hazard models are just plain wrong - I'm not sure how you can explain the enormous variance in wind speed for example, since wind is not a hyperlocal phenomenon like flood or wildfire. What we need is clarity to cut through the noise. All risk assessments are not created equal. That's why we developed the #RiskClass Taxonomy when I was at Arup, which outlines criteria to qualify for a given Risk Class. The truth is, a couple providers on this list are much more sophisticated than others, while the vast majority were purpose built for minimum reporting requirements (Class 0), which frankly tolerate low accuracy. It's like comparing apples and oranges. Wouldn't it be nice to categorize them based on what Risk Class they can provide? Because where most of these vendors stop is where the hard questions start. If you can't be confident in your risk in the first place, how can you be confident in your resilience strategy? For that, the market needs to move towards Class 2 and 3 level assessments, which the current solutions just don't support. #Iris #resilience #confidence #accuracy https://lnkd.in/gud35BYR

The Arctic is one of the coldest places on Earth, but in recent decades, region has been rapidly warming, at a rate 3-4 times faster than global average. However, current climate models have been unable to account for this increased pace. Now, researchers reported in a study, that clouds may be to blame. Most common clouds found in the Arctic are mixed-phase clouds, which contain both ice crystals and supercooled liquid water droplets. In Arctic summer, when the sun shines arnd the clock, these clouds act like a parasol, reflecting sunlight back into space and providing a cooling effect (PIC 3). But in a long, dark Arctic winter, when there's no sunlight to reflect, these clouds act more like a blanket, trapping heat radiated from Earth's surface and sending it back down to Arctic's surface. How well these mixed-phase clouds trap heat depends on their ratio of ice to liquid. The more liquid water clouds contain, the better they're at trapping heat. But many climate models have a large bias in representing this ratio, causing incorrect predictions. In this study, they analyzed 30 climate models and compared them to satellite observations of clouds in the Arctic during winter over last decade. They found that 21 of 30 models significantly overestimated fraction of ice to liquid in wintertime Arctic clouds. These ice-dominant models are not properly accounting for present-day warming potential of clouds during winter. That's why they can't account for rapid warming we are currently seeing. However, every cloud has a silver lining. While climate models are underestimating rate of global warming in present day, they're overestimating rate of global warming in the future. Errors in future projections are due to a process called "cloud emissivity feedback." In a nutshell, as Arctic warms, clouds shift from containing mostly ice to more liquid, which increases their ability to trap heat, further warming the Arctic and creating a pos feedback loop. But importantly, this feedback loop has a time limit. Once clouds become so rich in liquid they behave like blackbodies—fully absorbing and re-emitting heat—further warming has less effect. However, because many climate models underestimate how much liquid is already present in today's clouds, they assume a larger shift still lies ahead. As a result, they overestimate how much extra heat-trapping will occur in the future, and predict feedback effect will last longer than reality suggests. Moving forward, study's findings could be used to refine climate models. Since Arctic's climate also plays a key role in shaping weather patterns further S, these findings could also lead to more accurate forecasts of extreme weather in mid-latitude regions. The biggest uncertainty in our forecasts is due to clouds! Fixing these models is essential not just for the Arctic, but for understanding its impact on weather and climate change across the globe!

Clouding the forecast! 🫠 Why so many climate models are wrong about rate of Arctic warming! The Arctic is one of the coldest places on Earth, but in recent decades, the region has been rapidly warming, at a rate three to four times faster than the global average. However, current climate models have been unable to account for this increased pace. Now, two researchers from Kyushu University—graduate student Momoka Nakanishi, from the Interdisciplinary Graduate School of Engineering Sciences, and her advisor, Associate Professor Takuro Michibata, from the Research Institute for Applied Mechanics have reported in a study, published in Ocean-Land-Atmosphere Research, that clouds may be to blame. How Does Cloud Emissivity Feedback Affect Present and Future Arctic Warming? https://lnkd.in/dz2Up_ZQ The most common clouds found in the Arctic are mixed-phase clouds, which contain both ice crystals and supercooled liquid water droplets. In the Arctic summer, when the sun shines around the clock, these clouds act like a parasol, reflecting sunlight back into space and providing a cooling effect. But in the long, dark Arctic winter, when there's no sunlight to reflect, these clouds act more like a blanket, trapping heat radiated from Earth's surface and sending it back down to the Arctic's surface. "However, how well these mixed-phase clouds trap heat depends on their ratio of ice to liquid," explains Nakanishi. "The more liquid water the clouds contain, the better they are at trapping heat. But many climate models have a large bias in representing this ratio, causing incorrect predictions." In this study, Nakanishi and Michibata analyzed 30 climate models and compared them to satellite observations of clouds in the Arctic during winter over the last decade. They found that 21 of the 30 models significantly overestimated the fraction of ice to liquid in wintertime Arctic clouds. "These ice-dominant models are not properly accounting for the present-day warming potential of the clouds during the winter," says Nakanishi. "That's why they cannot account for the rapid warming we are currently seeing." https://lnkd.in/d5r8Nt4F

Another piece of our disaster: models miss now ~50% of the warming... Earths energy imbalance (more heat goes in then out) tells us the warming still in the pipeline and how fast the oceans accumulate energy as 90% of the total extra energy kept in the system by rising GHGs is accumulating in the oceans. Even before 2020 even the spread of models could barely keep up with the increase in Earths energy imbalance. Since that time models lost contact with reality as the EEI went off charts in 2023. Discussed reasons are a cloud feedback reinforce by aerosol emission cuts - support cloud cover reduction, and a missed SST pattern effect that also reduces cloud cover (I addressed it in a former post that SST patterns can reduce global cloud cover by decreasing tropospheric stability supporting convection). But most likely the models miss a feedback cascade of a systems that produce strengthening temperature jumps which includes a cloud feedback over warmer oceans and related SST pattern effect further reducing clouds caused e.g. by sea ice losses, atmospheric circulation shifts, stratification, mixed layer depth changes. Here what they write: Worryingly, the observed energy imbalance is rising much faster than expected, reaching 1.8 Wm 2 in 2023—or twice that predicted by climate models—after having more than doubled within just two decades (Figure 1). This strong upward trend in the imbalance is difficult to reconcile with climate models: even if the increase in anthropogenic radiative forcing and associated climate response are accounted for, state‐of‐the‐art global climate models can only barely reproduce the rate of change up to 2020 within the observational uncertainty (Raghuraman et al., 2021). The continued rise in the energy imbalance since 2020 leaves us with little doubt that the real world signal has left the envelope of model internal variability. The root cause of the discrepancy between models and observations is currently not well known, but it seems to be dominated by a decrease in Earth's solar reflectivity (Goessling et al., 2024; Stephens et al., 2022), and model experiments suggest it could be due to poorly modeled sea surface temperature patterns, the representation and emissions of polluting aerosol particles, or something else (Hodnebrog et al., 2024).     "Earth's Energy Imbalance More Than Doubled in Recent Decades"; https://lnkd.in/erUPuqt4

This chart has been widely reported over the last few days, but its importance cannot be understated. CERES ground truth data is showing that climate models are seriously underestimating the crisis. CERES is the NASA satellite mission measuring the reality of the forces driving climate change. It measures the energy coming in from the sun, the energy reflected by clouds and the Earth’s surface, the energy being radiated from the surface and even the cloud fraction and temperature. Using these observational measurements, scientists can calculate the energy being absorbed by the climate system, which is driving the accelerating warming in every aspect. In a balanced climate, the energy coming in would be matched by the energy going out. In a cooling climate, the energy going out would be more than the energy coming in. However we are observing less energy going out than is coming in leading to a positive energy imbalance and an accumulation of heat in the climate system. What’s worse still is that this imbalance is steadily increasing, more than doubling in less than 20 years. The heat is accumulating at a rate equivalent to the heat from 12 Hiroshima bombs going off every single second! What’s scary about this graphic is that it shows that all the current climate models are underestimating this energy imbalance. They are all too insensitive to changes in cloud area and brightness caused by the warming atmosphere. In so doing they are underestimating the overall climate’s sensitivity to greenhouse gas emissions, they are underestimating climate feedbacks and are underestimating the rate of warming and hence the likely changes and impacts in the pipeline. Story: https://lnkd.in/ehq3R92T Paper (paywall ) https://lnkd.in/eppXue_n #climatechange #models #ECS #globalwarming #CERES #climatecrisis 

New research published in Science suggests something disturbing: our climate models underestimate the sensitivity of the global temperature to increased greenhouse gas emissions. (Links to a story about the article and the article itself in comments.) The technical details are way above my pay grade, but the basic idea is simple: - Use satellites to measure Earth's energy imbalance (absorbed radiation versus emitted radiation back into the space) - Compare this empirical satellite measure to what 37 different climate models generate It turns out that climate models with high climate sensitivity get much closer to the empirical measurement of energy imbalance. In contrast, climate models with low climate sensitivity underestimate the energy imbalance systematically and by a wide margin. This suggests that climate models with high climate sensitivity could be better constructs for how humanity's greenhouse gas emissions are changing the planet. This is not good news. If climate sensitivity is higher than estimated, that means our current emissions trajectory will produce even more warming than previously thought. And the previous numbers have recently been bad, around 3 degrees Celsius.