It is not really questioned today that climate can be described in theory by an ensemble of trajectories differing in their initial conditions, which is then translated to numerical ensembles in climate models. It is also widely accepted that any evolution observed within a few decades after initialization is not yet relevant to climate. Evolution at a later stage, instead, is then used to characterize climate and its change, under the implicit assumption that slower processes do not considerably contribute to differences between ensemble members, allowing internal variability of climate to be identified with these differences. However, a careful justification for this practice is as yet lacking. In particular, a definition of climate in support of this practice is outstanding, including the identification of the kind of time scales at play through providing an argumentation for their relevance. Our study aims at filling this gap. After pointing out that the most important criterion for a definition of climate is the uniqueness of the probability measure on which the definition relies, we first recall the naive proposal to represent such a probability measure by the distribution of ensemble members that has, loosely speaking, converged to the natural probability measure of the so-called snapshot or pullback attractor of the dynamics. We then consider the time scales of convergence and refine the proposal by taking a probability measure that is conditional on the (possibly time-evolving) state of modes characterized by convergence time scales longer than the horizon of a particular study. We design an ensemble simulation initialization scheme for studying convergence time scales and uniqueness of ensembles in Earth system models.
Boreal permafrost over the Northern Hemisphere high latitudes, defined as areas where the ground temperature is below 0 °C for two or more years, stores more than twice as much carbon as the atmosphere. Therefore, thawing of the permafrost, a tipping element, due to global warming may lead to additional carbon emissions and accelerate the warming. To investigate the permafrost response to increase and decrease of CO2 emissions, we conducted a series of numerical experiments using an emission-driven Earth System Model, MIROC-ES2L, and adopting idealized overshooting scenarios in which a prescribed CO2 emission of 10 PgC yr−1 is given until the global warming level reaches different values between 2 and 8 °C followed by the negative emission until the cumulative emission becomes zero. We found that the response of permafrost area to surface warming and cooling is reversible but has hysteresis for all the emission scenarios. Furthermore, the permafrost property such as the ratio of frozen to liquid water was shown to have irreversibility in the deep soil layer; part of the frozen area in the initial condition was replaced by a mixed water-ice area in the final state despite ground temperature returning almost to the initial condition. Sensitivity experiments reveal that the hysteresis and irreversibility are attributed to the delay of the soil freezing and melting associated with the soil heat conductivity and specific heat of water phase change. This result indicates that once permafrost thaws with warming it will continue for decades after warming diminishes and the delay in the permafrost recovery is larger at global warming levels greater than 2 °C. An offline calculation shows that the additional CO2 emission during the permafrost hysteresis cycle accounts for about 0.6 %–41 % of the cumulative carbon emission released from thawed permafrost in flat10 and NEC experiments.
The Congo Basin in Central Africa remains one of the few regions globally where the Intergovernmental Panel on Climate Change has neither assessed changes in hot extremes and extreme precipitation since the 1950s nor attributed such changes to anthropogenic influences, primarily due to the sparsity of in situ observational data. Although extensive daily weather records exist, spanning from the 1900s to the early 2000s and covering numerous stations across the basin, the majority of these remain archived on paper, limiting their accessibility for climate analysis. Here we present and analyse over 1 million temperature and precipitation observations from 37 weather stations in the Democratic Republic of the Congo, which have until now been unavailable to the research community. To produce this dataset, we first digitized (imaged) 9885 paper sheets stored in local archives during two dedicated field campaigns. We subsequently apply an improved version of the MeteoSaver software (version 1.1) to transcribe the imaged sheets, yielding daily time series of maximum, minimum, and average temperatures, precipitation, as well as dry bulb and wet bulb temperatures measured at three times per day (06:00, 15:00, and 18:00 local time). After quality control, analysis of multi-decadal temperature data across the basin reveals a consistent and accelerating warming signal since the 1960s, characterized by a warming shift in the distribution of daily maximum, minimum, and average temperatures with each successive decade. Median trends across 21 of the 37 stations with sufficient data availability are 0.22, 0.10, and 0.15 °C per decade for daily maximum, minimum, and average temperatures, respectively, corresponding to approximately 0.7, 0.3, and 0.5 °C of warming during the 1961–1990 period. We further find an increasing frequency of hot extremes and a decreasing frequency of cold extremes with each successive decade, with the most recent decade exhibiting nearly twice as many hot days per year and fewer cold days compared to the earliest decade. Analysis of precipitation at three stations with sufficient data indicates an increased frequency of heavy precipitation events at two stations and no substantial change at the third; however, limited spatial coverage due to data availability restricts broader conclusions for the DRC. Overall, this analysis of rescued weather data from Central Africa highlights the urgent need to close the knowledge gap on climate trends in the Congo Basin, one of the world's most data-sparse yet climatically significant regions.
Stratospheric aerosol injection (SAI) simulations are often short relative to climatic timescales and conducted against a background that evolves due to changes in anthropogenic greenhouse gas emissions and other forcings. This can cause challenges in assessing certain impacts of the intervention, especially for aspects of the climate that respond slowly to such changes. The early Geoengineering Model Intercomparison Project (GeoMIP) G2 experiment prescribes solar dimming to offset 1 % CO2 forcing in a preindustrial control background. Here we propose a new G2-SAI experiment, in which SAI is applied in the same scenario, to isolate SAI climate responses from transient changes other than CO2. Using the Community Earth System Model (CESM2), we present three 150-year “G2-SAI” simulations which use contemporary SAI strategies: two use the commonly-used “three degree-of-freedom” (“3DOF”) strategy, in which independent injections at 30° N, 15° N, 15° S, and 30° S are used to manage global mean temperature (T0) and large-scale meridional temperature gradients (T1, T2). Our third G2-SAI simulation uses a “1DOF” strategy that injects at 30° N and 30° S to manage global mean temperature only. Our two 3DOF simulations both maintain the same temperature targets; however, one simulation, which injects mostly at 15° S, slows but does not prevent the decline of the Atlantic Meridional Overturning Circulation (AMOC) compared to the baseline simulation, while the other, which injects mostly at 30° N and 30° S, stops the decline of AMOC entirely, similarly to the 1DOF simulation. These results demonstrate that multiple distinct Earth system states can satisfy the same temperature targets, challenging the assumption of linearity commonly used in strategy design. In addition, the results highlight that long simulations are required to identify some of the long-term impacts of SAI, such as AMOC changes. Using this knowledge, we revisit the ARISE-SAI-1.5 experiment and modify the injection strategy without changing the temperature targets, producing an “ARISE-hybrid” ensemble. We demonstrate that this results in some significant differences in the climate response to SAI, with implications for the perceived effects of the intervention.
The sustainable management of global groundwater resources is a key societal challenge and is central to the Sustainable Development Goals. The localized dynamics of groundwater abstraction, topography, and surface-water interactions, as well as the sensitivity of groundwater-dependent ecosystems, call for high-resolution information to support effective groundwater management. At the same time, groundwater observations are very limited and concentrated in a few regions, rendering large parts of groundwater resources ungauged. To address limited observations and coarse global models, we applied the global groundwater model GLOBGM (v1.1) to simulate past and future groundwater heads and water table depth at 30 arcsec (∼ 1 km) on a monthly time step. Model calibration improved mean bias in water table depth predictions from −4.8 to 3.6 m compared to GLOBGM v1.0, with depth-weighted bias reduced from 34.2 to 32.5 m across 34 800 observation wells. Groundwater dynamics are simulated for a historical reference period (1960–2019) to support model evaluation and attribution of observed impacts to climate variability and change. Baselines (1960–2014) and three combined socioeconomic-climate scenarios (2015–2100; SSP1-RCP2.6, SSP3-RCP7.0, SSP5-RCP8.5) are simulated with five global climate models, supporting detection and impact assessment of future change. Validation against monthly observations yielded skillful predictions (KGE-NPskill) in approximately 60 % of deep wells (> 60 m) and 83 % of shallow to intermediate wells (0–60 m). When validated against annual observations, 66 % of deep wells (> 60 m) and 71 % of shallow to intermediate wells (0–60 m) were skillful. Simulations are further bias corrected using a machine learning approach. Historical trend analysis (1960–2019) accurately reproduced known groundwater depletion regions such as the U.S. High Plains, Arabian Peninsula, and Indo-Gangetic Plain, while also identifying rising water tables in northern latitudes and Arctic regions, which are linked to climate-driven recharge changes. Future scenario-based simulations suggest rising water tables for most continents in the next century, with Europe being a notable exception. However, known regions of groundwater depletion are expected to persist. Regions that disagree with observations or show reduced reliability are mapped, and quality assurance flags are provided to guide the appropriate use and interpretation of the results. The resulting data set offers high-resolution information to assess groundwater dynamics for the past and future, supporting improved global water resource management and climate impact assessments.
