All state-of-the-art global climate models that contributed to the latest Coupled Model Intercomparison Project (CMIP6) show a robust increase in Indian summer monsoon rainfall that is even stronger than in the previous intercomparison (CMIP5). Furthermore, they show an increase in the year-to-year variability of this seasonal rainfall that crucially influences the livelihood of more than 1 billion people in India.

Higher temperatures influence both evaporation and snow processes. These two processes have a large effect on discharge but have distinct roles during different seasons. In this study, we study how higher temperatures affect the discharge via changed evaporation and snow dynamics. Higher temperatures lead to enhanced evaporation but increased melt from glaciers, overall lowering the discharge. During the snowmelt season, discharge was reduced further due to the earlier depletion of snow.

Deep Labrador Sea densities are receiving increasing attention because of their link to many of the processes that govern decadal climate oscillations in the North Atlantic and their potential use as a precursor of those changes. This article explores those links and how they are represented in global climate models, documenting the main differences across models. Models are finally compared with observational products to identify the ones that reproduce the links more realistically.

Small lakes are ecosystems highly impacted by climate change. Here, the thermal regime of a small, shallow lake over the past six decades was reconstructed via 3D modelling. Significant changes were found: strong water warming in spring and summer (0.7 °C/decade) as well as increased stratification and thermal energy for cyanobacteria growth, especially in spring. The strong spatial patterns detected for stratification might create local conditions particularly favourable to cyanobacteria bloom.

This paper presents new estimates of future extreme weather in Europe, including extreme heat, extreme rainfall and meteorological drought. These new estimates were achieved by repeating model calculations many times, thereby reducing uncertainties of these rare events at low levels of global warming at 1.5 and 2 °C above pre-industrial temperature levels. These results are important, as they help to assess which weather extremes could increase at moderate warming levels and where.

Monthly scale, seasonal-scale, and decadal-scale modeling of the atmosphere is possible using the principle of energy balance. Yet the scope of classical approaches is limited because they do not adequately deal with energy storage in the Earth system. We show that the introduction of a vertical coordinate implies that the storage has a huge memory. This memory can be used for macroweather (long-range) forecasts and climate projections.

Radiant energy is exchanged between the Earth's surface and outer space. Some of the local imbalances are stored in the subsurface, and some are transported horizontally. In Part 1 I showed how – in a horizontally homogeneous Earth – these classical approaches imply long-memory storage useful for seasonal forecasting and multidecadal projections. In this Part 2, I show how to apply these results to the heterogeneous real Earth.

We model greenhouse gas emission scenarios consistent with limiting global warming to either 1.5 or 2 °C above pre-industrial levels. We quantify the effectiveness of methane emission control and land-based mitigation options regionally. Our results highlight the importance of reducing methane emissions for realistic emission pathways that meet the global warming targets. For land-based mitigation, growing bioenergy crops on existing agricultural land is preferable to replacing forests.

We use a reduced-complexity climate model trained by observations to show that at the current rate of human release of CO₂, total cumulative emissions will pass the 66 % likelihood of limiting warming to 1.5° or 2°C in about 10 and 35 years, respectively. We also show that complex climate models often used to guide policy tend to warm faster than observed over the past few decades. To achieve the Paris Climate Agreement, CO₂ and CH₄ emissions must be severely curtailed in the next decade.

The current radiative imbalance at the top of the atmosphere is increasing the heat stored in the oceans, atmosphere, continental subsurface and cryosphere, with consequences for societies and ecosystems (e.g. sea level rise). We performed the first assessment of the ability of global climate models to represent such heat storage in the climate subsystems. Models are able to reproduce the observed atmosphere heat content, with biases in the simulation of heat content in the rest of components.

In the Earth system, climate tipping elements exist that can undergo qualitative changes in response to environmental perturbations. If triggered, this would result in severe consequences for the biosphere and human societies. We quantify the risk of tipping cascades using a conceptual but fully dynamic network approach. We uncover that the risk of tipping cascades under global warming scenarios is enormous and find that the continental ice sheets are most likely to initiate these failures.

Compound hot and dry events can lead to severe impacts whose severity may depend on their timescale and spatial extent. Here, we show that the spatial extent and timescale of compound hot–dry events are strongly related, spatial compound event extents are largest at sub-seasonal timescales, and short events are driven more by high temperatures, while longer events are more driven by low precipitation. Future climate impact studies should therefore be performed at different timescales.

We provide the first spatio-temporally explicit comparison of different model-derived fluxes from land use and land cover changes (f_LULCCs) by using the TRENDY v8 dynamic global vegetation models used in the 2019 global carbon budget. We find huge regional f_LULCC differences resulting from environmental assumptions, simulated periods, and the timing of land use and land cover changes, and we argue for a method consistent across time and space and for carefully choosing the accounting period.

Scientific tradition has left a gap between the study of humans and the rest of the Earth system. Here, a holistic approach to the global human system is proposed, intended to provide seamless integration with natural sciences. At the core, this focuses on what humans are doing with their time, what the bio-physical outcomes of those activities are, and what the lived experience is. The quantitative approach can facilitate data analysis across scales and integrated human–Earth system modeling.

We show that modern concepts of non-equilibrium statistical mechanics can be applied to large-scale environmental fluid dynamics, where fluctuations are not thermal but come from turbulence. The work theorems developed by Jarzynski and Crooks are applied to air–sea interaction. Rather than looking at the average values of thermodynamic variables, their probability density functions are considered, which allows us to replace the inequalities of equilibrium statistical mechanics with equalities.

Climate sensitivity is a key measure of how sensitive Earth's climate is to human release of greenhouse gasses, such as from fossil fuels. However, there is still uncertainty as to the value of climate sensitivity, in part because different climate feedbacks operate over multiple timescales. This study assesses hundreds of millions of climate simulations against historical observations to reduce uncertainty in climate sensitivity and future climate warming.

The roots of vegetation largely control the Earth's water cycle by transporting water from the subsurface to the atmosphere but are not adequately represented in land surface models, causing uncertainties in modeled water fluxes. We replaced the root parameters in an existing model with more realistic ones that account for a climate control on root development and found improved timing of modeled river discharge. Further extension of our approach could improve modeled water fluxes globally.

Fluxes from land-use change and management (F_LUC) are a large source of uncertainty in global and regional carbon budgets. Here, we evaluate the impact of different model parameterisations on F_LUC. We show that carbon stock densities and allocation of carbon following transitions contribute more to uncertainty in F_LUC than response-curve time constants. Uncertainty in F_LUC could thus, in principle, be reduced by available Earth-observation data on carbon densities at a global scale.

In this study, we model the relative importance of several contributors to the land-use and land-cover change (LULCC) flux based on a LULCC dataset including uncertainty estimates. The uncertainty of LULCC is as relevant as applying wood harvest and gross transitions for the cumulative LULCC flux over the industrial period. However, LULCC uncertainty matters less than the other two factors for the LULCC flux in 2014; historical LULCC uncertainty is negligible for estimates of future scenarios.