Compound and cascading droughts and heatwaves decrease maize yields by nearly half in Sinaloa, Mexico

Climate change-driven heatwaves and droughts are major contributors to global hunger and rising food prices and are among the most important stresses affecting crop growth and yields. Even mild events during the vegetative stage can substantially reduce yields, with prior research reporting maize yield losses up to 40% under drought. Impacts intensify when heatwaves and droughts co-occur (compound events) compared to single extremes. Many prior studies used statistical methods, often not accounting for both single and compound/cascading occurrences during crop growth. This study aims to assess the impacts of single, compound (simultaneous), and cascading (consecutive) drought and heatwave events on maize yield using a process-based model (WOFOST), addressing a gap in the literature. Compound drought-heat (CDH) is defined as concurrent drought and heatwave in the same region/month; cascading (CaDH) refers to one following the other. The state of Sinaloa, Mexico—one of the country’s major maize-producing regions with substantial exposure to these extremes—was selected. The study integrates detailed soil, crop, and management data with ERA5 meteorology (1990–2022) to simulate yields and identify hazards (SPI-3 for agricultural drought; 90th-percentile Tmax threshold for heatwaves). Yields under normal, drought, heatwave, compound, and cascading conditions were compared, and Pearson correlation was used to relate hazards to yields. Although a case study, the approach has broader relevance.

Prior research in Mexico has often examined droughts and heatwaves in isolation, documenting significant agricultural losses in years such as 1993–1995, 2010, and 2011, despite compound and cascading (CnC) events occurring in some of these periods. Few studies systematically assess the joint occurrence and interactions of droughts and heatwaves in Mexico; some references to CnC events stem from global theoretical studies. Other work (e.g., Zhang et al.) projects increases in CnC drought-heat events in northern Mexico under extreme warming scenarios (SSPs > 5). The limited focus on CnC in Mexico may reflect limited institutional engagement with the specific high impacts triggered by CnC events, even as single-event impacts are widely recognized; drought impacts have led to national programs such as the Mexican National Drought Program.

Study region and period: Sinaloa, Mexico; maize growing season November–May annually; analysis covers 1990–2022 using ERA5 meteorological data combined with detailed soil, crop, and management information.

Yield simulation: WOFOST crop growth model was used to simulate annual maize yields under a potential production setup (high inputs/irrigation/pest control), keeping planting date, thermal time since emergence (Tsum), fertilizer, and soil properties constant each year so interannual yield variations reflect weather and hazard conditions. WOFOST simulates temperature- and water-stress effects on assimilation; optimum 20–30 C; about 3% reduction at 36 C, up to ~56% reduction by 42 C; water stress reduces assimilation via transpiration reduction.

Drought identification: Agricultural drought identified using Standardized Precipitation Index with a 3‑month accumulation period (SPI‑3). SPI computed by fitting gamma distributions to monthly precipitation and transforming to normal. Categories: moderate (−1 to −1.49), severe (−1.5 to −1.99), extreme (< −2). For each crop cycle (Nov–May), one representative SPI‑3 was selected: the lowest SPI‑3 if any drought occurred during the cycle; otherwise, the highest SPI‑3 to represent wet/normal conditions.

Heatwave identification: Heatwaves defined when daily maximum temperature exceeds a seasonally varying threshold set at the 90th percentile of Tmax, using a 31‑day centered moving average to create smoothed daily thresholds (365/366 per year). Heatwaves require at least three consecutive days above threshold. Only heatwaves occurring within the crop cycle were counted for yield comparisons.

Compound and cascading (CnC) events: CDH (compound) when drought and heatwave occur in the same month and region; CaDH (cascading) when they occur consecutively (heatwave month immediately before drought month). Drought is monthly (SPI‑3), heatwaves are daily; monthly overlap was used to determine compounding/cascading months.

Event cataloging and analysis: ERA5 time series used to catalog droughts, heatwaves, and CnC events. Events outside the crop cycle were excluded from yield-impact analyses. Pearson correlation was applied to assess relationships between SPI‑3 (minimum per cycle), number of heatwave events, and annual yields.

• Event occurrence: 27 distinct drought events identified (varied duration); 15 occurred within crop cycles (10 moderate: 1994, 1995, 1996, 1999, 2006, 2008, 2011, 2017, 2019; 4 severe: 2000, 2010, 2013, 2020; 1 extreme: 1999). A total of 75 heatwave events occurred (1990–2022), with only 9 during crop cycles. Compound/cascading catalog: 14 CnC events overall (10 CDH, 4 CaDH), but only 1 CDH and 2 CaDH fell within crop cycles.
• Yield impacts by condition: Droughts reduced maize yield by about 25% (e.g., from ~6 t/ha to ~3.6 t/ha). Heatwaves alone had limited impact; average yields during heatwave years often remained ≥3.9–4.3 t/ha and could match/exceed normal when occurring under wet/normal conditions. Compound events led to the strongest reductions, with yields around 2.7 t/ha; cascading events averaged ~2.8 t/ha. The study reports up to 44% yield loss under compound/cascading conditions relative to normal.
• Correlations: Strong positive correlation between SPI‑3 and yield (Pearson r = 0.77), indicating wetter conditions align with higher yields. Weak correlation between heatwave counts and yield.
• Notable years: Lowest yields (≈2.4–2.8 t/ha) simulated during the extreme drought in 2010 and CnC events in 1998–1999. Severe drought years (e.g., 2005, 2010) drove yields below 3.5 t/ha. Heatwave-only crop-cycle years (2001, 2002, 2013) showed little yield penalty with average yields above ~3.9 t/ha.
• Phenological sensitivity: Timing matters. Drought during vegetative stages tended to cause stronger losses; droughts confined to vegetative stages could sometimes have lesser impact than those spanning both vegetative and reproductive stages. The 2010 five-month drought within the crop cycle produced very low yields, comparable to CnC years.

The study demonstrates that maize yields in Sinaloa are highly sensitive to water availability, with SPI‑3 explaining a substantial portion of interannual yield variability (r = 0.77). Heatwaves in isolation had weak and inconsistent impacts, particularly when occurring during otherwise normal or wet conditions; this aligns with maize’s relative heat tolerance up to mid‑30s Celsius. However, when droughts and heatwaves co-occur or occur sequentially (compound/cascading), impacts intensify well beyond single-hazard effects, yielding losses up to approximately 44% relative to normal conditions. The findings address the research question by quantifying the distinct and joint effects of drought and heat on yields using a process-based model over three decades, highlighting that concurrent moisture deficits and high temperatures exacerbate stress through coupled water and thermal limitations on assimilation and transpiration. The results underscore the importance of monitoring and managing compound risks rather than focusing solely on single extremes, with direct relevance for maize production planning and drought risk management in Sinaloa and similar regions.

This work integrates hazard identification (SPI‑3 droughts; percentile-based heatwaves) with process-based crop modeling (WOFOST) to disentangle single versus compound/cascading impacts on maize yields in Sinaloa (1990–2022). Key contributions include (i) a catalog of events during crop cycles, (ii) quantification of yield penalties under drought (~25%) and more severe penalties under compound/cascading conditions (to ~2.7–2.8 t/ha; up to ~44% relative loss), and (iii) strong yield–moisture coupling (r = 0.77). Policy and practice implications emphasize prioritizing drought risk reduction, irrigation reliability, and targeted adaptation for CnC conditions. The approach is transferable to other regions and crops. Future research should: incorporate realistic management variability and farmer adaptation; extend analyses to heat-sensitive crops (e.g., rice, vegetables); explore sub-seasonal timing and duration effects in more detail; assess projections under climate scenarios to anticipate changing CnC risks; and integrate soil moisture and irrigation dynamics more explicitly.

• Modeling framework assumes potential production (high-input conditions), which can overestimate absolute yields relative to actual farmer conditions.
• ERA5 reanalysis and derived SPI‑3/heatwave metrics may include biases; drought is characterized monthly while heatwaves are daily, requiring monthly overlap simplifications for CnC identification.
• Constant management parameters (planting date, Tsum, fertilizer, soil properties) across years do not capture changes in practices or adaptive responses.
• Event counts within the crop cycle are limited (especially CnC events), which may constrain statistical power and generalizability.
• Case-study focus on Sinaloa limits direct extrapolation; local irrigation infrastructure and water allocation dynamics (e.g., reservoir levels) were not explicitly modeled.