The prevalent life cycle of agricultural flash droughts

The study addresses how agricultural flash droughts evolve and what their prevalent life cycle is across different regions and climates. Flash droughts—rapid soil drying over weeks to a couple of months—have become more common than slow-onset droughts in many regions since the 1950s and are projected to increase in frequency and severity under 21st-century warming, particularly over croplands. Their rapid onset and intensification make them difficult to predict and highly impactful on crop yields and ecosystems. Agricultural flash droughts occur when soil moisture deficits prevent plants from meeting water requirements, especially during critical crop growth periods. Precipitation deficits and positive temperature anomalies typically drive rapid soil moisture depletion. Soil moisture is a key proxy for land-atmosphere moisture state and vegetation stress, with shallow-rooted crops being especially sensitive. Understanding flash drought evolution requires examining the coevolution of meteorological anomalies (e.g., enhanced evaporative demand) and land-atmosphere coupling that can accelerate onset. Prior studies use varied indicators and definitions (soil moisture variability, evaporative stress ratio, evaporative demand, multivariate approaches), which can yield differing characterizations and drivers. Motivated by the need to couple rapid root-zone soil moisture depletion with impacts on vegetation, the authors propose a soil water availability-based indicator tailored to agricultural regions to assess the prevalent life cycle and drivers of agricultural flash droughts globally, and to evaluate their timing relative to crops’ critical growth stages.

Previous research has examined flash drought drivers and characteristics at regional to global scales using diverse indicators, including soil moisture anomalies and variability, evaporative stress ratio, evaporative demand-based metrics, and multivariate frameworks. Studies highlight the roles of precipitation deficits, high temperatures, and elevated evaporative demand in rapid soil drying and drought propagation, with land-atmosphere coupling accelerating onsets. However, varying definitions and datasets can lead to different outcomes regarding occurrence, intensity, onset timing, and key drivers. Recognizing the importance of linking rapid root-zone soil moisture declines to vegetation stress, recent work emphasizes incorporating soil hydrological thresholds (field capacity, wilting point, and a critical soil moisture separating energy- and water-limited regimes) to better represent agricultural impacts. Case studies (e.g., 2012 central-eastern U.S., 2010 Russia/Ukraine, 2013 southern China) have documented rapid drought evolution and extensive agricultural impacts, motivating physically grounded, vegetation-relevant indicators.

Data and preprocessing: The study uses ERA5 reanalysis data at 0.5° × 0.5° spatial resolution for 1960–2020. Variables include root-zone volumetric soil moisture for the top 1 m, latent and sensible heat fluxes (to compute evaporative fraction, EF = LH/(LH+SH)), total precipitation, evapotranspiration, and 2-m air temperature. Non-overlapping pentads (5-day means) were created from hourly data. Pentad anomalies were computed as departures from the 1960–2020 mean annual cycle and normalized by local standard deviations to produce standardized anomalies for global comparison. Soil hydraulic properties (field capacity, θFC, and wilting point, θWP) from ERA5 (static by grid) were used to assess soil water availability. Land cover (high/low vegetation fraction) from ERA5 and external cropland products (GFSAD: Global Cropland-extent at 30 m and 1-km Cropland Dominance) informed agricultural relevance of regions. Indicator and thresholds: The Soil Water Deficit Index (SWDI) is defined as SWDI = ((θFC − θ)/(θFC − θWP)) × 10. SWDI = 0 at field capacity (no deficit); SWDI ≤ −10 when θ ≤ θWP (no available water for plants). A critical soil moisture θcrit (between θWP and θFC) marks the transition from energy-limited (θ > θcrit) to water-limited (θ < θcrit) evapotranspiration. The agricultural flash drought definition integrates four elements: (1) rapid root-zone soil moisture decay represented by SWDI decreasing from above −3 to below −5 within 20 days (4 pentads); the upper threshold (−3) lies at the beginning of the transitional regime ensuring noticeable ET reduction; the lower threshold (−5) corresponds to onset of plant water stress for many crops; (2) an intensification period lasting at least 15 days (≥3 pentads) to exclude short synoptic events; (3) prior non-drought conditions requiring the three pentads before onset to have SWDI values higher than −4 in magnitude (|SWDI| < 4), preventing overlap of events; and (4) explicit coupling to plant water stress via the SWDI thresholds. The annual soil moisture cycle is not removed, consistent with the focus on growing-season vulnerability. Event detection and compositing: The method was applied globally to detect events and to compute frequencies (annual and seasonal). Hotspots were defined where area-averaged frequency exceeded two events per decade and at least half the area had three or more events per decade. To analyze lifecycle, composites of standardized anomalies (precipitation, temperature, evapotranspiration, soil moisture) were constructed for lags −4 to +4 pentads around onset (lag 0), and spatial composites were produced over 12° × 12° windows centered on events. Sensitivity tests varied SWDI threshold ranges (e.g., −4<SWDI<−2; −6<SWDI<−3) showing similar spatial patterns though frequencies changed. A comparison with a soil moisture percentile-based method highlighted differences, especially in extreme climates where the SWDI-based approach avoids false identification due to physical thresholds relative to θFC and θWP.

• Validation with historical events: The SWDI-based approach accurately captured the spatiotemporal evolution of well-documented flash droughts, including the 2012 central-eastern United States event, late April–May 2010 in southwestern Russia/eastern Ukraine, late 2001 in India, and the once-a-century 2013 southern China event.
• Global hotspots and frequencies (1960–2020): Eight agricultural flash drought hotspots were identified: Southern China (SCh), Central-eastern Europe (CEEu), India (In), Southeastern South America (SESA), Southern Russia (SRus), Central-eastern USA (CEUSA), Southeastern Asia (SEAS), Northern South America (NSA), and Central-western Africa (CWAf). Area-averaged decadal frequencies and maxima include: SCh 3.3 (max 12.5), CEEu 3.2 (8.2), In 3.0 (9.7), SESA 2.9 (10.1), SRus 2.8 (8.3), CEUSA 2.3 (7.7), SEAS 2.3 (10.6), NSA 2.1 (8.5), CWAf 2.0 (8.5). Some grid points approach or exceed one event per year in SCh, In, SESA, and SEAS.
• Seasonality aligned with crop critical periods: In extratropics, peak frequencies occur in spring (MAM in the Northern Hemisphere; Nov–Dec in SESA), impacting planting and pollination of major crops (wheat, barley, corn, soybeans, rice). In subtropical/tropical regions, peaks occur in summer or SON: SCh peaks in Jul–Aug (affecting rice), India/SE Asia/CWAf peak in SON, and NSA peaks in DJF affecting mixed crops, cotton, and coffee.
• Prevalent lifecycle across climates: Composite analysis shows a consistent lifecycle worldwide. Pre-onset (lags −4 to −1): nearly steady temperature, slight precipitation decline, sufficient soil moisture supporting slight ET increase (energy-limited). Just before onset: precipitation drops rapidly, temperature rises, and ET surges due to enhanced evaporative demand, accelerating soil moisture dry-down. At and after onset (lag 0 to +3): precipitation deficits deepen then stabilize; soil moisture declines; ET decreases despite rising temperatures as water-limited conditions set in; temperature anomalies remain positive, with ET reductions implying more energy to sensible heat.
• Drivers: Precipitation deficit is the primary driver of rapid soil moisture depletion; evapotranspiration dynamics also contribute—rising pre-onset in energy-limited conditions and falling during intensification under water limitation. The transition from energy- to water-limited regimes is central to agricultural impact.
• Indicator robustness and physical realism: Threshold sensitivity tests preserve spatial occurrence patterns. Comparisons with a percentile-based method reveal that SWDI’s physical thresholds help avoid misclassification in extreme climates (e.g., deserts with persistent θ < θWP; very wet regions near θFC) while providing comparable identification in agriculturally suitable climates.

The findings demonstrate that agricultural flash droughts share a common physical evolution regardless of geography or climatic regime, clarifying their lifecycle and key drivers. By integrating soil hydraulic constraints with root-zone soil moisture, the SWDI-based indicator directly links rapid soil drying to plant water stress, which is crucial for agriculture. The consistent sequence—initial precipitation decline and enhanced evaporative demand in energy-limited conditions, followed by a shift to water-limited conditions with decreasing evapotranspiration and rising temperatures—explains the rapid intensification and agricultural impacts, including yield losses during critical crop growth stages. The approach also highlights that precipitation deficits are the dominant driver of rapid soil moisture depletion, with evapotranspiration playing a significant, regime-dependent role. Comparisons indicate the new method’s advantages in avoiding false identifications in extreme climates by anchoring thresholds to soil physical properties. These insights can improve understanding of predictability and inform early warning, as the lifecycle’s commonalities provide signals for monitoring and forecasting across diverse regions.

This study introduces a physically grounded, soil water availability-based indicator to identify agricultural flash droughts by coupling rapid root-zone soil moisture depletion with vegetation water stress using SWDI and soil hydraulic thresholds. Applied globally with ERA5 data, the method identifies consistent flash drought lifecycles across climatic regimes, quantifies hotspot regions and seasonal timing aligned with crop critical periods, and validates against notable historical events. Results underscore precipitation deficits as the primary driver, with evapotranspiration and the energy- to water-limited transition shaping intensification and temperature responses. Future work could refine estimates by incorporating improved soil property datasets and uncertainty quantification, evaluate robustness across multiple reanalyses and observations, and integrate the indicator within operational early warning systems to support agricultural risk management.

• Data uncertainties: ERA5 reanalysis has known uncertainties in soil moisture and fluxes; soil hydraulic properties (field capacity and wilting point) are derived using pedotransfer functions and sparse/heterogeneous soil profiles, introducing regional biases.
• Definition specificity: The indicator is tailored to agricultural flash droughts and may not capture all flash drought manifestations in non-agricultural or extreme climates; it intentionally excludes cases where soils/climates are unsuitable for agriculture.
• Threshold arbitrariness: Although physically motivated, selected SWDI thresholds (−3 to −5) and duration criteria have some arbitrariness; sensitivity tests show frequency changes with different ranges, though spatial patterns persist.
• Annual cycle retained: The method does not remove the annual soil moisture cycle, focusing detections during the growing season, which may influence seasonal distributions.
• Dataset/indicator dependence: Different datasets or alternative flash drought definitions can yield differences in occurrence, intensity, onset, and drivers, affecting comparability across studies.