Unconventional cold vortex as precursor to historic early summer heatwaves in North China 2023

Heatwaves—prolonged periods of exceptionally high temperatures—are increasing in frequency and intensity globally, posing risks to ecosystems, public health, and societal systems. Their occurrence reflects multiscale interactions: long-term trends are driven by greenhouse gas forcing and regional anthropogenic effects (including urbanization), interdecadal variability links to PDO/AMO and cryosphere change, interannual variability to ENSO, Indo-Pacific SST anomalies, extratropical SST, soil moisture, and vegetation, and subseasonal variability to ISO, NAO, and MJO. In boreal summer mid-latitudes, heatwaves are often associated with quasi-stationary anticyclones or “heat-dome” patterns linked to blocking, Rossby wave activity, and diabatic heating anomalies, and can be modulated by ENSO phase. In summer 2023, North China experienced record heat with June–July anomalies exceeding 1.5 °C, a one-month advance of the seasonal SAT peak, and multi-day temperatures exceeding 40 °C in Beijing. Unusually, an upper-tropospheric cyclonic PV anomaly (“cold vortex”) accompanied and preceded the heatwaves—contrary to canonical heat-dome situations and despite a developing El Niño that typically inhibits such extremes in this region. The study’s objective is to elucidate the mechanisms responsible for this unconventional circulation precursor and its effects on North China’s early-summer 2023 heatwaves, to identify a subtropical heatwave pattern preceded by a cold vortex, and to assess the subseasonal air–land–sea processes and their implications for S2S predictability of climate extremes.

The paper reviews drivers of heatwaves across timescales: anthropogenic warming and urbanization (urban heat island) enhance heatwave characteristics; interdecadal modes (PDO, AMO) and Arctic sea-ice loss modulate regional extremes; interannual variability involves ENSO, Indian Ocean and extratropical SST anomalies, soil moisture, and vegetation changes; subseasonal drivers include ISO, NAO, and MJO influencing Eurasia and North America. Mechanistically, extratropical summer heatwaves are commonly linked to quasi-stationary anticyclones and heat-dome patterns associated with blocking, meandering jets, Rossby wave breaking, and westerly disruptions, often arising from circumglobal wave trains or diabatic heating anomalies. Over China, heat domes tend to strengthen after mature El Niño but weaken during El Niño development. Prior studies highlight teleconnections (e.g., MJO, wave-7 patterns) and land feedbacks in notable heatwave events over Eurasia and East Asia. This context underscores the novelty of a cold-vortex precursor producing a bottom-up warming pathway, diverging from the classical top-down heat-dome mechanism.

Datasets: In situ observations from 334 stations in North China (36°–43°N, 113°–122°E) for 2-m SAT, rainfall, surface (skin) temperature, and 10-m wind; ERA5 reanalysis (1.0° × 1.0°, daily; 37 levels) for atmospheric circulation and thermal structure; NASA GPM IMERG V7 precipitation (0.1° × 0.1° daily); NOAA DOISST v2.1 daily SST (0.25° × 0.25°). Period: 1981–2023 (GPM from 2001; SST from 1982). Heatwave definition: At each station, a heatwave day occurs when daily mean SAT exceeds the 90th percentile threshold computed over 1991–2020 using a 5-day centered moving window for each calendar day. A regional heatwave day is when >30% of stations meet the threshold. Diagnostics: Surface sensible heat flux (SH) via bulk aerodynamic formula SH = (ρ cp) CD U (Ts − Ta) with ρ=1.184 kg m−3, cp=1005 J kg−1 K−1, CD=1.5×10−3; potential vorticity PV = g(f+ζ)/δθ; wave activity flux following Takaya and Nakamura. Anomalies are relative to 1991–2020 climatology. Anomalous forms: SH′ = (ρ cp) CD [U(Ts′ − Ta) + U′(Ts − Ta) + U′(Ts′ − Ta′)]; PVA′ = −V·∇PV − V′·∇PV − V·∇PV′. Indices and stages: North China Vortex Index (NCVI) defined as 250-hPa PV anomaly averaged over 35°–50°N, 115°–130°E. Two stages in early summer 2023: Cold Vortex (CV, 1–14 June) and Heatwave (HW, 15 June–3 July). ISM rainfall averaged over 10°–25°N, 70°–90°E. Zonal SSTA gradient in the Indo-Pacific warm pool defined as SSTA(Arabian Sea, 5°–25°N, 50°–90°E) minus SSTA(WNP, 5°–25°N, 120°–150°E). Niño3.4 index per standard box. S2S forecasts: ECMWF real-time S2S (CY47R2) ensemble (51 members), initiated twice weekly to day 46; horizontal 1.5° × 1.5°, 10 pressure levels (1000–100 hPa); on-the-fly reforecasts (2003–2022). Initiations used: 29 May, 1, 5, and 8 June 2023. Numerical experiments: Linear Baroclinic Model (LBM) with June mean basic state; imposed diabatic heating/cooling profiles based on observed ISM anomalies. Integrated 31 days; stationary responses from last 15-day mean. Statistical analysis: Pearson correlations, t-tests for significance; linear regressions for scatter plots; sliding 21-year correlations to assess stability over 1981–2023.

• Timing and structure: A pronounced upper-tropospheric cold vortex (positive PV anomaly) formed over/east of North China on 1 June 2023 and persisted to 14 June. Following its dissipation, heatwaves rapidly developed from 15 June to 3 July, advancing the seasonal SAT peak by about one month. Beijing’s diurnal SAT exceeded 40 °C on June 22–24 for over 40 hours.
• Circulation evolution: CV stage featured paired anomalies—cyclone over North China and anticyclone over western Central Asia—and two wave trains (a North Atlantic–Europe path associated with negative NAO and a subtropical wave train from the northern Indian subcontinent). HW stage saw replacement by a weak anticyclone and warmer troposphere over North China, with a British–Baikal corridor teleconnection along the polar front jet.
• ISM teleconnection mechanism: Suppressed ISM convection in early June (deficient ISM rainfall) generated a subtropical wave train that produced positive PV (cold vortex) over North China and southwestern India and negative PV over western Central Asia. LBM forced by ISM diabatic cooling reproduced the observed PV pattern; NAO forcing produced little impact on North China in the CV stage.
• Air–land feedbacks (bottom-up warming): During CV, the cold vortex cooled/stabilized the troposphere and, via enhanced upper-level negative PV advection, promoted subsidence and reduced cloud cover, increasing net solar radiation and skin temperature over land. This amplified air–land thermal contrast, increasing sensible heating (SH) and reducing soil moisture (SM). Persistently drier soils maintained larger SH into HW, warming SAT and helping terminate the upper-level cold vortex as the column destabilized.
• Subseasonal predictability: ECMWF real-time S2S forecasts captured the sequence—stronger cold vortex → stronger subsidence/less cloud → larger SH and drier SM in CV → warmer SAT in HW. Forecast biases included underestimation of SAT and upper-level anticyclone in HW due to underestimated SM anomalies and air–land feedback intensity.
• Quantitative relationships: • In S2S and historical data, NCVI (CV) negatively correlates with ISM rainfall; both reached extreme values in 2023. • Historical correlations (CV stage): ISM rainfall vs zonal Indo-Pacific SSTA gradient r=+0.594 (p<0.001). ISM rainfall vs Niño3.4 r=−0.294 (ns at 95%). Zonal gradient vs Niño3.4 r=−0.44 (p=0.004), indicating El Niño likely contributed to the negative gradient. • Composite SAT anomalies over North China (HW stage): +1.22 °C when preceded by a cold vortex (PV>+0.2 PVU) vs +0.73 °C when preceded by an upper-level heat dome (PV<−0.2 PVU), indicating stronger heatwaves with the cold-vortex precursor.
• Extremity of 2023: Seasonal mean SAT anomalies exceeded 1.5 °C across North China with record intensity/duration at many stations; historical records show 2023 minima in ISM rainfall and zonal SSTA gradient and maxima in NCVI and subsequent SAT, underscoring the event’s severity.

The study addresses why North China’s 2023 early-summer heatwaves were preceded by a cold vortex rather than a canonical heat dome. It demonstrates a tropical–extratropical teleconnection triggered by an extreme zonal SSTA gradient across the Indo-Pacific warm pool that suppressed ISM convection, exciting a subtropical wave train and generating a cold vortex over North China. This circulation anomaly initiated a bottom-up air–land feedback: subsidence and clearer skies increased surface heating, reduced soil moisture, and subsequently amplified sensible heating, producing extraordinary heatwaves and destabilizing the column to weaken the cold vortex. These mechanisms reconcile the apparent paradox of cold upper-level anomalies coexisting with developing heatwaves and explain the one-month advance of the SAT seasonal peak. The findings highlight practical relevance for S2S predictability: initial cold-vortex conditions and monitored air–sea gradients over the Indo-Pacific warm pool provide forecast skill for imminent heat extremes, while biases in land-surface processes (soil moisture feedback) can systematically damp heatwave intensity in models. The identification of a robust cold-vortex precursor pathway complements conventional heat-dome paradigms and refines understanding of subseasonal heatwave genesis in East Asia.

This work identifies and mechanistically explains an unconventional pathway to extreme early-summer heatwaves in North China in 2023: a cold-vortex precursor generated by Indo-Pacific warm-pool SSTA gradients that suppressed ISM convection, followed by strong air–land feedbacks that drove bottom-up surface warming. It shows that: (1) the cold vortex is a statistically significant precursor of subsequent heat over North China; (2) the zonal SSTA gradient across the Arabian Sea and western North Pacific is a key driver of ISM anomalies leading to the cold vortex; and (3) ECMWF S2S forecasts can capture these coupled processes but underestimate heat intensity due to soil moisture feedback biases. The study advances heatwave theory beyond top-down heat-dome mechanisms and offers actionable S2S predictors. Future directions include: improving land-surface and soil-moisture–atmosphere coupling in S2S systems; investigating the emergence and regional generality of cold-vortex–preceded heatwaves across other extratropical regions; clarifying the nonstationary role of ENSO in shaping Indo-Pacific gradients affecting ISM; and assessing how ongoing climate change may alter the frequency and intensity of this bottom-up pathway.

• Process nonstationarity: Sliding 21-year correlations show that not all links are stable over 1981–2023. ENSO’s influence on the zonal Indo-Pacific SSTA gradient and ISM rainfall has weakened over the past two decades, introducing uncertainty in attributing El Niño’s role in 2023.
• Model biases: ECMWF S2S forecasts underestimate SAT and the HW-stage anticyclone due to biases in soil moisture anomalies and air–land feedback intensity; they may overestimate cold-vortex impacts on SH and underestimate SH–SM coupling.
• Attribution scope: The NAO pathway showed limited local influence in this case, but broader multi-teleconnection interactions may vary by event. Reanalysis/observational uncertainties (e.g., precipitation products) and resolution constraints may affect diagnostics.
• Generalizability: While historical statistics support the cold-vortex precursor over North China, further evaluation is needed to establish occurrence and robustness across other regions and climate backgrounds.