Record-shattering 2023 Spring heatwave in western Mediterranean amplified by long-term drought

The study addresses why the April 2023 spring heatwave in the western Mediterranean—particularly the Iberian Peninsula and northwestern Africa—was so extreme and unprecedented. The authors hypothesize that antecedent severe drought conditions amplified and prolonged the heatwave by reducing evaporation and increasing surface sensible heat flux, thereby intensifying air temperatures under a favorable synoptic setup (a pronounced subtropical ridge). The event broke April records (e.g., Córdoba reached 38.8 °C, 4.8 °C above the previous April record; Portugal’s Mora reached 36.9 °C; parts of Morocco and Algeria exceeded 40 °C) amid severe multi-year drought (e.g., Spanish reservoirs ~50% below average; Morocco dam storage ~33%; Tunisia’s Sidi Salem at 16%). The objectives are to: (1) identify surface and synoptic characteristics of the heatwave; (2) quantify the influence of soil drought on surface energy partitioning and air temperature; and (3) infer the likelihood of such an event under dry versus wet soil preconditions.

Prior research has documented the role of drought and soil moisture anomalies in exacerbating heatwaves, particularly in temperate regions during summer (e.g., 2003, 2018, 2019 European heatwaves; 2010 Russian heatwave). In such settings, drought-driven soil moisture deficits can substantially alter energy partitioning and amplify temperatures through land-atmosphere feedbacks. In Mediterranean climates, summer soils are typically dry, limiting further drying effects; summer heatwaves there are often linked to heat advection from arid regions. However, the role of land-atmosphere feedbacks during spring—when climatological soil moisture is higher and anomalies can be larger—remains less understood. The study builds on metrics of soil moisture-temperature coupling and flow analogues approaches used to disentangle circulation versus land-surface feedbacks, and leverages extreme value analysis used in heatwave attribution studies.

Data: (1) In-situ daily maximum temperature (Tx) for Córdoba from AEMET Open Data via the climaemet R package. (2) Drought conditions characterized using 1-, 3-, 12-, and 24-month SPEI for the 4th week of April from the SPEI Spain Drought Monitor (1 km resolution, 1960–present). (3) ERA5-Land daily fields (0.25°): Tx, volumetric soil moisture (0–7 cm), evaporation, potential evaporation, surface net solar and thermal radiation; anomalies relative to 1950–2022 seasonal cycle. (4) ERA5 reanalysis for atmospheric fields (daily averages from hourly, resampled to 1°): 850-hPa temperature (T850) and 500-hPa geopotential height (Z500), used for synoptic characterization and the analogues experiment; percentiles computed vs 1950–2022. Soil moisture-temperature coupling: The Miralles et al. π diagnostic quantifies coupling as π = (H − Hp) T', where T' is standardized daily near-surface air temperature anomaly, and (H − Hp)' is the standardized anomaly of the sensible heat flux excess due to soil moisture constraints, derived from (Rn − λE)' − (Rn − λEp)' with λE computed as a function of T and Rn. High π arises when positive T' co-occurs with elevated (H' − Hp'), indicating strong influence of soil moisture deficits on temperature via energy partitioning. Computed daily for March–May. Flow analogues experiment: For each day of the heatwave peak (26–28 April 2023), days in 1950–2022 with similar Z500 anomalies over 20W–15E, 25N–50N (based on RMSD) were identified within a 30-day window centered on the event (excluding 2023). Analogues were split by antecedent soil moisture over south Iberia (7.5W–2.5W, 37.3N–38.6N), using the 15-day mean SM: dry if below the 33rd percentile, wet if above the 66th percentile. For each peak day, the closest 1% of days (~15) were selected for each SM class. The heatwave peak Tx field was reconstructed 5000 times by randomly sampling one analogue per day for dry and wet sets to obtain flow-conditioned Tx distributions. The linear trend was removed from SM prior to classification. Sensitivity tests varied time window size (20–40 days), analogue domain size (10 alternatives), and assessed seasonal bias; differences were negligible (<0.1 °C over south Iberia). Extreme value analysis: A Generalised Extreme Value (GEV) distribution was fit to the maxima of each reconstructed peak (5000 samples per SM class), yielding probability estimates for the observed regional maximum Tx anomaly (13.5 °C) under dry and wet preconditions, the probability ratio, and the expected maximum Tx difference between dry and wet scenarios.

- The April 2023 heatwave coincided with a strong subtropical ridge (high Z500) and very warm lower-tropospheric air (T850 above the 99th percentile), alongside extreme soil desiccation (soil moisture below the 1st percentile in extensive areas). - Córdoba recorded four consecutive daily Tx records (25–28 April), peaking at 38.8 °C on 27 April, which is 4.8 °C above the prior April record (from 1960). Portugal (Mora) reached 36.9 °C, while parts of Morocco and Algeria exceeded 40 °C. - The soil moisture-temperature coupling metric (π) and its energy term (H' − Hp') were exceptionally high during the peak (26–28 April), unprecedented in southern Iberia during March–May, revealing strong land-atmosphere feedback driven by soil dryness and surplus radiation. - Flow analogues conditioned on similar circulation show that dry antecedent soils systematically yield higher Tx anomalies than wet antecedent soils: differences exceed 1 °C over most of Iberia, average about 2 °C over southern Iberia, and locally exceed 3 °C near Córdoba. - GEV-based attribution with analogue reconstructions indicates the event’s regional maximum Tx anomaly (13.5 °C) was 4.53 times (SD 0.87) more likely under dry preconditions than under wet, and that maximum Tx under dry soils was 2.19 °C (SD 0.17 °C) higher than if soils had been wet. - Despite strong drought amplification, the top reconstructed maximum Tx from the dry analogue subset was ~2.5 °C lower than the observed peak, implying additional factors (e.g., concurrent marine heatwaves and SST anomalies) likely contributed.

The results support the hypothesis that antecedent drought substantially amplified the April 2023 heatwave through enhanced sensible heat flux and reduced evaporative cooling, under a conducive synoptic driver (subtropical ridge). The strongest impacts centered on southern Iberia and parts of northern Morocco/Algeria, aligning with the deepest soil moisture deficits and highest coupling. While circulation set the stage, land-atmosphere feedbacks were crucial to the severity and persistence of the event. The discrepancy between observed peak temperatures and analogue-based dry reconstructions (~2.5 °C) suggests other mechanisms, including marine heatwave/SST anomalies, may further intensify extremes. Importantly, the study shows semiarid Mediterranean regions in spring can exhibit strong land-atmosphere coupling comparable to temperate regions in summer, implying that soil moisture can be an effective diagnostic and predictor for early-season heatwave risk. The analogue-plus-GEV framework provides an observationally grounded, computationally efficient method to quantify drought-induced amplification and probability shifts, complementing model-based attribution and forecast methods.

This study identifies and quantifies the pivotal role of antecedent soil moisture deficits in amplifying the record-breaking April 2023 western Mediterranean heatwave. Under comparable circulation, dry soils increased regional maximum temperatures by an estimated 2.19 °C and made the observed extreme 4.53 times more likely than under wet soils. The event featured unprecedented soil moisture-temperature coupling in southern Iberia, highlighting strong land-atmosphere feedback during spring in semiarid Mediterranean climates. The proposed combination of flow analogues and extreme value analysis offers a practical, observation-based approach to assess land-surface contributions to heat extremes. Future work should disentangle contributions from marine heatwaves/SST anomalies and other drivers using regional modeling, and explore integration of soil moisture metrics into subseasonal and seasonal forecasting and land management strategies to mitigate spring heatwave risks.

- The analogue method, while robust and tested for biases, cannot fully capture all concurrent drivers; the ~2.5 °C gap between observed and dry-analogue maxima indicates additional mechanisms (e.g., SST anomalies, marine heatwaves) were not explicitly quantified. - Reliance on reanalysis (ERA5/ERA5-Land) introduces uncertainties from model and data assimilation, particularly for surface fluxes and soil moisture. - The GEV analysis is conditioned on analogue reconstructions and assumes that the selected analogues adequately represent circulation variability; residual seasonal biases may persist despite sensitivity tests. - No targeted regional climate or atmospheric model experiments with prescribed soil moisture or SST perturbations were conducted, limiting process attribution and dynamical causality. - The case-study focus on the western Mediterranean (southern Iberia and nearby North Africa) may limit generalizability across different climates/seasons without further validation.