How long will the four seasons be by 2100? Increasing evidence suggests that the length of a single season or in regional scales has changed under global warming, but a hemispherical‐scale response of the four seasons in the past and future remains unknown. We find that summer in the Northern Hemisphere midlatitudes has lengthened, whereas winter has shortened, owing to shifts in their onsets and withdrawals, accompanied by shorter spring and autumn. Such changes in lengths and onsets can be mainly attributed to greenhouse‐warming. Even if the current warming rate does not accelerate, changes in seasons will still be exacerbated in the future. Under the business‐as‐usual scenario, summer is projected to last nearly half a year, but winter less than 2 months by 2100. The changing seasonal clock signifies disturbed agriculture seasons and rhythm of species activities, more frequent heat waves, storms and wildfires, amounting to increased risks to humanity.

The classical two-box model of Stommel is extended in two directions: replacing the buoyancy constraint with an energy constraint and including the wind-driven gyre. Stommel postulated a buoyancy constraint for the thermohaline circulation, and his basic idea has evolved into the dominating theory of thermohaline circulation; however, recently, it is argued that the thermohaline circulation is maintained by mechanical energy from wind stress and tides. The major difference between these two types of models is the bifurcation structure: the Stommel-like model has two thermal modes (one stable and another one unstable) and one stable haline mode, whereas the energy-constraint model has one stable thermal mode and two saline modes (one stable and another one unstable). Adding the wind-driven gyre changes the threshold value of thermohaline bifurcation greatly; thus, the inclusion of the wind-driven gyre is a vital step in completely modeling the physical processes related to thermohaline circulation.

The Pacific Walker circulation (PWC) is one of the most important components of large-scale tropical atmospheric circulations. The PWC and its influences have been studied extensively by numerical models and reanalysis. The newly released ERA5 and NCEP2 are the most widely used reanalysis datasets and serve as benchmarks for evaluation of model simulations. If the results of these datasets differ significantly, this could lead to a bias in projected long-term climate knowledge. For better understanding of future climate change, it is necessary to evaluate PWC reanalysis productions. As a result, we compared the PWC structures between the ERA5 and NCEP2 datasets from month to seasonal time scales. We used the zonal mass streamfunction (ZMS) over the equatorial Pacific to indicate the strength of the PWC. The PWC’s average monthly or seasonal cycle peaks around July. From February to June, the NCEP2 shows a higher PWC intensity, whereas the ERA5 shows greater intensity from July to December. The circulation center in the NCEP2 is generally stronger and wider than in the ERA5. The ERA5, however, revealed that the PWC’s west edge (zero line of ZMS over the western Pacific) had moved 10 degrees westward in comparison to the NCEP2. In addition, we compared the PWC mean state in the reanalysis and CMIP6 models; the mean state vertical structures of the tropical PWC in the CMIP6 multi-model ensemble (MME) are similar to those of the reanalyses in structure but weaker and wider than in the two reanalysis datasets. The PWC is broader in CMIP6, and the western boundary is 7 and 17 degrees farther west than in the ERA5 and NCEP2, respectively. This study suggests that, when using reanalysis datasets to evaluate PWC structural changes in intensity and western edge, extreme caution should be exercised.

The main seasonal characteristics in the tropics include both spatial patterns and temporal parameters of onset, cessation, duration, and the number of wet and dry seasons. Previous studies showed that wet seasons shortened and dry seasons extended with global warming, but the changes in spatial distribution and the number of wet and dry seasons are still unclear. Here, we analyze the climatic characteristics of once wet and dry season a year (annual regime) and twice wet and dry seasons a year (biannual regime), and find that regimes of wet and dry seasons have changed from 1935 to 2014. Across the equator and the Tropic of Cancer and Capricorn, some regions where there used to be an annual regime have become a biannual regime; instead, other regions have shifted from a biannual regime into an annual regime. With seasonal regimes shifting, areas of the biannual regime have expanded at a rate of 31 000 km²/decade. Meanwhile, in annual regime regions, wet seasons have been shortened in 60.3% of regions, with an average of 7 d; the onset dates of wet seasons have been delayed in 64.8%, with an average of 6 d. Besides, wet seasons have become wetter in 51.1% of regions, and dry seasons have become drier in 59.9%. In biannual regime regions, the shortened wet seasons have occurred in 83.7% of regions, with an average shortening of 8 d, and precipitation has decreased in both wet and dry seasons. Moreover, the shorter wet seasons will amplify further by the end of the 21st century. The continuous seasonal changes will threaten agricultural, ecological security, and even human well-being.