Climate data and temporal aggregation

All indicators are derived from daily NEX-GDDP-CMIP6 temperature, humidity and precipitation. We aggregate results into two seasons relevant for global sport calendars:

• JJA (June–August): boreal summer / austral winter,
• DJF (December–February): boreal winter / austral summer.

These windows capture the dominant competitive periods in both hemispheres and reflect established seasonal patterns of thermal and hydrometeorological stress.

Multi-model ensemble (MME)

All results are presented as multi-model ensemble means. For each indicator and time period, we compute the ensemble mean from the five selected models. To assess robustness, inter-model agreement is quantified using a 4/5 sign-consistency criterion, whereby stippling indicates regions where at least four models agree on the direction of change. This threshold is aligned with the “medium-to-high confidence” framing used in IPCC AR6 and prevents overinterpretation of model-specific outliers.

Wet-Bulb Globe Temperature (WBGT)

WBGT is the standard biometeorological index used to assess exertional heat-stress risk in athletes and outdoor workers [6,7,24]. We compute WBGT using the classical formulation of [25]:

where is dry-bulb air temperature, the natural wet-bulb temperature and the globe temperature.

Reconstruction of time-of-day humidity.

The NEX-GDDP dataset provides daily mean relative humidity (RH_mean), but WBGT requires humidity that is physically consistent with the temperature at the time of day considered. To obtain morning and afternoon humidity, we reconstruct RH as follows:

1. We first compute daily vapour pressure:

where is the saturation vapour pressure from the Tetens formula.

2. For each time window, we recompute RH using:

where T_time is equal to:

• for morning WBGT,
• for afternoon WBGT,
• T_mean for evening WBGT.

3. Reconstructed RH values are constrained to the range 1–99%.

This procedure follows the approach used in large-scale heat–health assessments, ensuring that humidity is consistent with the thermal state of the atmosphere during the relevant sport time window.

Heat Index (HI)

The Heat Index (HI) is computed using the NOAA polynomial algorithm [27]. It should be noted that the Heat Index is known to overestimate thermal stress in hot and humid equatorial environments [8,9]. In such regions, HI values should be interpreted as a conservative upper limit of heat strain rather than an accurate physiological estimate. By contrast, WBGT — which explicitly accounts for radiation and wind — represents the most robust metric for assessing heat stress in high-performance sport contexts [4,7]. Daily maximum temperature is converted to Fahrenheit:

and daily RH is constrained to 1–99%. The raw Heat Index is:

with NOAA coefficients . The index is applied only when F and RH ≥40%; otherwise HI is approximated by the dry-bulb temperature. We convert back to Celsius:

Indicator: number of days with HI ≥ 40 °C per season, corresponding to the “danger” category in NOAA guidelines [7,27].

Heavy rainfall events

Heavy rainfall days are defined as:

This threshold represents precipitation intensities likely to produce waterlogged natural surfaces, reduced friction, degraded footing and increased probability of cancellations [15]. It is also consistent with thresholds used in hydrological impact studies to classify disruptive rainfall.

Indicator: number of 20 mm/day events per season.

Heatwaves

Heatwaves are defined as:ca

an absolute-temperature criterion commonly used in global heat-health studies [14]. We count the number of days belonging to heatwave episodes in each season, capturing cumulative thermal load not reflected in single-day metrics [24].

Composite Sport Climate Risk Index (SCRI)

Each indicator is normalised across all model×scenario combinations:

The composite index is:

Equal weights (0.25) avoid imposing an arbitrary prioritisation of hazards, reflecting the absence of any established medical or sport-scientific consensus on the relative severity of heat stress, rainfall disruption and heatwave exposure for outdoor sport practice. Any alternative weighting scheme would require empirical or normative justification that does not currently exist in the literature. The equal-weight approach therefore represents the most neutral, transparent and reproducible option, consistent with composite index construction in analogous multi-hazard frameworks [14].

Model evaluation and limitations

The performance of bias-corrected CMIP6 projections, including NEX-GDDP, has been evaluated in several regional studies [28–30]. These studies generally show improved skill for temperature and mean precipitation, although regional discrepancies persist. Given these uncertainties, we focus on relative changes rather than absolute values.

Limitations include the simplified WBGT formulation, grid-scale representation of radiation and wind, absence of stadium or urban microclimates and omission of acclimatisation, cooling strategies and behavioural responses. While absolute magnitudes may differ from venue-level conditions, spatial patterns and relative trends remain robust for comparative risk assessment.

Overview of indicator workflow

Fig 1 summarises the indicator workflow and conceptual structure of the multi-hazard framework developed in this study. Daily CMIP6 projections from the NASA NEX-GDDP dataset provide the common foundation for all calculations. From these daily fields, we derive four sport-relevant indicators capturing distinct climatic constraints on outdoor sport: (i) Wet-Bulb Globe Temperature (WBGT) for morning, afternoon and evening match conditions, representing acute heat–humidity–radiation stress; (ii) Heat Index (HI ≥ 40 °C), representing humidity-amplified heat strain; (iii) heavy rainfall days (P ≥ 20 mm/day), associated with surface degradation and cancellations; and (iv) heatwave days, defined as sequences of at least three consecutive days with C, representing cumulative thermal load.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 1. Multi-hazard indicator workflow.

Daily CMIP6 NEX-GDDP projections (air temperature, humidity, precipitation and solar radiation) provide the common input for computing four sport-relevant indicators: WBGT (morning/afternoon/evening), Heat Index (HI > 40 °C), heavy rainfall days ( mm day⁻¹), and heatwaves (at least 3 consecutive days with C). Each indicator is aggregated for JJA and DJF over three 30-year periods (1991–2020, 2031–2060, 2071–2100), normalised, and combined into a composite Sport Climate Risk Index (SCRI, 0–1). The SCRI enables the production of global 0.25° maps of multi-hazard risk for outdoor sport.

https://doi.org/10.1371/journal.pclm.0000969.g001

All indicators are aggregated to two seasons relevant for global sport scheduling (JJA and DJF) and three 30-year climate periods (1991–2020, 2031–2060 and 2071–2100). Each indicator is then normalised and combined with equal weight into a composite Sport Climate Risk Index (SCRI), summarising multi-hazard exposure across regions, seasons and emission scenarios. This structure links daily-scale climate physics to seasonal-scale constraints on outdoor sport, providing a coherent and fully integrated assessment framework.

Seasonal exposure to extreme heat

Across the historical period (1991–2020), exceedances of the WBGT extreme-risk threshold already occur frequently in low-latitude regions. In JJA, afternoon exposure reaches 20–40 days per season across the Sahel, northern India, the Arabian Peninsula and parts of Central America (Fig 2). Morning conditions remain generally below 10 days except in humid equatorial regions, while evening exceedances rarely exceed 5–10 days outside West Africa and northern South America. By mid-century (2031–2060, SSP5–8.5), afternoon extreme-WBGT days intensify and expand poleward: South Asia, the Sahel and Southeast Asia frequently exceed 40–60 days, and southern Europe, southeastern China, the southeastern United States and northern Argentina begin to experience more than 10–20 days per season. By 2071–2100, afternoon extremes surpass 60–80 days across much of the tropics, morning exceedances reach 20–40 days in South Asia and West Africa, and evening exceedances reach 20–30 days in the most humid regions. These trends signal the near-disappearance of a “safe” morning window in several densely populated regions.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Seasonal exposure to extreme heat (WBGT  >  32 °C) in JJA.

Seasonal number of days above the WBGT extreme-risk threshold for morning, afternoon and evening match conditions (rows), and for 1991–2020, 2031–2060 and 2071–2100 under SSP5–8.5 (columns). Base map: Natural Earth (https://www.naturalearthdata.com), public domain.

https://doi.org/10.1371/journal.pclm.0000969.g002

In DJF (Fig 3), afternoon WBGT exceedances historically reach 15–30 days in Brazil, 10–20 days in Mozambique and eastern South Africa, and 20–30 days in northern Australia. By mid-century, these regions exceed 30–50 days, with morning values approaching 10–20 days. By 2100, afternoon exposure surpasses 50–70 days across most of Brazil, southern Africa and northern Australia, while morning exceedances commonly reach 20–30 days. Evening conditions exceed 10–20 days in humid tropical regions, indicating persistent 24-hour thermal load. DJF thus becomes as hazardous in the Southern Hemisphere as JJA in the Northern Hemisphere.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Seasonal exposure to extreme heat (WBGT  >  32 °C) in DJF.

Same layout as Fig 2, highlighting the southward shift of extreme-heat hotspots in austral summer. Base map: Natural Earth (https://www.naturalearthdata.com), public domain.

https://doi.org/10.1371/journal.pclm.0000969.g003

Taken together, the JJA and DJF patterns show that by the end of the century, extreme-WBGT conditions become quasi-permanent features of the warm season in many low-latitude regions. Tropical land areas frequently exceed 60–80 dangerous days per season, while subtropical regions such as the Mediterranean Basin, the southern United States, southern Africa and eastern China transition from rare exceedances in the historical climate to 20–40 days per season. The hemispheric complementarity between JJA and DJF further implies that, on a global scale, at least one major sport region is almost always experiencing heightened heat risk, compressing the opportunity for safe scheduling across the annual calendar.

Adaptation potential and limits of time-of-day scheduling

Seasonal differences between afternoon and morning exposure illuminate the scope—and limits—of scheduling as an adaptation strategy (Fig 4). Historically, most tropical regions show a positive WBGT of 10–20 days, meaning afternoon matches are substantially riskier. Contrasts are strongest across South Asia, the Sahel, northern South America and northern Australia, while dry subtropical regions show near-zero differences.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Difference between afternoon and morning exposure (WBGT-days).

Positive values indicate locations where afternoon matches face greater heat stress; negative values indicate regions where mornings are riskier. Base map: Natural Earth (https://www.naturalearthdata.com), public domain.

https://doi.org/10.1371/journal.pclm.0000969.g004

By mid-century (2031–2060), WBGT increases by 5–15 days across most tropical belts. By 2100, afternoon amplification exceeds 20–35 days in the Indo-Gangetic Plain, West Africa, Central America, Brazil and southeastern Africa. In some equatorial regions, however, WBGT declines because mornings also exceed 32 °C, indicating that scheduling loses effectiveness as both morning and afternoon conditions surpass physiological thresholds.

Dangerous Heat Index conditions

The Heat Index expands faster than WBGT as humidity amplifies thermal stress (Fig 5). Historically, JJA HI  >  40 °C occurs 20–40 days per season in West Africa, Amazonia, South Asia and the Arabian Peninsula, and 10–20 days in Central America and Southeast Asia. DJF values reach 10–30 days across Brazil, northern Australia and southern Africa. By mid-century, South Asia, West Africa, Southeast Asia and northern Brazil exceed 40–50 days, while Central America, the Caribbean and the Gulf Coast accumulate 20–40 days. By 2100, Amazonia, West Africa, the Bay of Bengal and Southeast Asia record more than 60 dangerous-HI days per season—two months during which outdoor sport becomes physiologically unsafe regardless of time of day.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Seasonal exposure to dangerous Heat Index (HI  >  40 °C).

Seasonal number of dangerous-HI days in JJA and DJF across 1991–2020, 2031–2060 and 2071–2100. Base map: Natural Earth (https://www.naturalearthdata.com), public domain.

https://doi.org/10.1371/journal.pclm.0000969.g005

Heavy rainfall days

Heavy rainfall days (P  >  20 mm/day) show more moderate but spatially coherent increases (Fig 6). Historically, tropical monsoon regions record 10–25 days per season, while subtropical regions remain below five. By 2031–2060, West Africa, the Gulf of Guinea, Southeast Asia and Amazonia increase by 2–5 days per season, with similar changes in DJF across eastern Australia and southeastern Africa. By 2100, tropical regions frequently exceed 20–30 days, Southeast Asia reaches 25–35 days and East Africa surpasses 15–25 days. Even modest increases can disrupt multiple matchdays due to waterlogged surfaces and degraded footing.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Seasonal number of heavy rainfall days (P  >  20 mm/day).

JJA and DJF across 1991–2020, 2031–2060 and 2071–2100 under SSP5–8.5. Base map: Natural Earth (https://www.naturalearthdata.com), public domain.

https://doi.org/10.1371/journal.pclm.0000969.g006

Heatwave days

Heatwaves (≥3 consecutive days with ) emphasise cumulative thermal stress (Fig 7). Historically, South Asia, the Middle East, northern Australia, interior Brazil and southern Africa record 10–20 heatwave days per season. By mid-century, the Mediterranean Basin, southwestern United States, South Africa and eastern Australia experience increases of 10–20 days, reaching 20–30 days. By 2100, South Asia and the Gulf reach 40–60 days, inland Australia and central Brazil reach 30–40 days, and the Mediterranean, southern United States and eastern China reach 20–40 days—multi-week sequences with major implications for athlete safety and competition scheduling.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. Seasonal number of heatwave days.

JJA and DJF across 1991–2020, 2031–2060 and 2071–2100. Base map: Natural Earth (https://www.naturalearthdata.com), public domain.

https://doi.org/10.1371/journal.pclm.0000969.g007

Composite Sport Climate Risk Index (SCRI)

The SCRI combines all four hazard dimensions into a 0–1 metric reflecting simultaneous exposure to extreme heat, humidity, heavy rainfall and heatwaves. In 2031–2060, Fig 8 shows SSP1–2.6 yields SCRI values below 0.4 in most regions, with tropical Africa and South Asia reaching 0.5. Under SSP2–4.5, large regions shift into the 0.5–0.6 range, including northern Brazil, West Africa, Southeast Asia, northern India and parts of the southern United States and Mediterranean Basin. Under SSP5–8.5, most low-latitude regions exceed 0.6, with hotspots reaching 0.7–0.8.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 8. Composite Sport Climate Risk Index (SCRI), 2031–2060.

JJA (a–c) and DJF (d–f) under SSP1–2.6, SSP2–4.5 and SSP5–8.5. Base map: Natural Earth (https://www.naturalearthdata.com), public domain.

https://doi.org/10.1371/journal.pclm.0000969.g008

By 2071–2100, scenario divergence becomes marked (Fig 9). SSP1–2.6 stabilises with values below 0.4. SSP2–4.5 yields widespread values of 0.6–0.7 across the tropics and subtropics. Under SSP5–8.5, continental-scale regions reach SCRI values of 0.9–1.0—particularly Amazonia, West Africa, the Bay of Bengal, Southeast Asia and northern Australia—while the Mediterranean, southern United States and East Asia reach 0.6–0.8. Values above 0.8 indicate that multiple hazards are simultaneously severe for more than half the season, implying that outdoor sport becomes barely viable without major adaptation measures.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 9. Composite Sport Climate Risk Index (SCRI), 2071–2100.

JJA (a–c) and DJF (d–f) under SSP1–2.6, SSP2–4.5 and SSP5–8.5. Base map: Natural Earth (https://www.naturalearthdata.com), public domain.

https://doi.org/10.1371/journal.pclm.0000969.g009

In practical terms, high-end SCRI values under SSP5–8.5 indicate that large parts of the tropical belt enter a regime where traditional outdoor sport calendars become structurally incompatible with the emerging climate, unless competition periods, training formats and infrastructure design are fundamentally rethought.

Global and regional hotspots for outdoor sport

A consistent feature across indicators is the emergence of strong hotspots in low-latitude and warm–humid regions, where WBGT and Heat Index frequently exceed safety thresholds already in the historical baseline and rise sharply under future scenarios. These patterns are fully consistent with epidemiological and physiological evidence showing disproportionately high heat-health risks in tropical and subtropical settings [14,24,34]. For sport, this means that regions hosting some of the world’s densest populations of participants—football, running, school sport and mass-participation events—are also those facing the fastest-growing constraints.

At mid-latitudes, our results suggest substantial increases in the number of days above WBGT or HI safety thresholds during summer, especially under SSP2-4.5 and SSP5-8.5. This is consistent with observed and projected trends in extreme heat across Europe, North America and parts of Asia [14]. These regions currently host many major global events (World Cups, Olympic Games, World Championships, city marathons) and have long been considered “climatically safe” relative to hotter regions. Our findings show that this assumption is becoming progressively less valid, particularly in a high-emission world, and that mid-latitude regions will increasingly face non-negligible constraints on training, competition and scheduling.

Heavy rainfall and heatwaves exhibit different but overlapping patterns. In temperate regions, increases in heavy rainfall days appear modest, but even small changes can exacerbate pitch degradation, reduced ball roll, poor footing and injury risk [13]. In monsoonal and equatorial regions, heavy-rainfall days remain frequent or increase further, implying that waterlogged surfaces and cancellations will remain recurrent constraints even where heat is not the primary limiting factor. Heatwaves, defined as multi-day runs of very hot conditions, represent cumulative stress episodes that can degrade athlete performance, recovery and safety independently of single-day exceedances [24,34].

Overall, the multi-indicator perspective confirms recent qualitative assessments that extreme heat is emerging as a structural climate risk to sport [31,33], while highlighting the additional and interacting roles of humidity, rainfall and multi-day clustering that are rarely examined together.

Implications for sports scheduling and the geography of play

Our findings have direct implications for sports scheduling at all organisational levels. The projected increases in days with WBGT or HI above recommended safety thresholds suggest that the traditional summer-centred calendar will be increasingly misaligned with safe environmental windows.

Table 1 summarises the WBGT-based heat thresholds currently adopted by major international sports federations. The WBGT > 32°C threshold used in this study closely corresponds to the suspension or extreme-risk level defined by FIFA (32.0°C), ITF (32.2°C), World Triathlon (32.2°C) and World Rowing (32.0°C), enabling direct comparison between our projections and operational decision criteria. Under SSP5–8.5, regions such as South Asia, West Africa and Amazonia are projected to exceed this threshold for 60–80 days per season by 2071–2100 — equivalent to two to three months during which competitions would systematically trigger suspension protocols under current federation guidelines. This echoes calls from governance-focused studies to formally recognise heat as a core climate risk [33] and complements empirical work showing growing athlete, coach and organiser concerns regarding extreme heat [31,32].

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. WBGT-based heat policy thresholds adopted by major international sports federations and the American College of Sports Medicine (ACSM). All thresholds are expressed in °C. The WBGT > 32°C threshold used in the present study corresponds to the suspension or extreme-risk level defined by most federations.

https://doi.org/10.1371/journal.pclm.0000969.t001

Table 2 contextualises these thresholds against projected afternoon WBGT exceedance days in key hotspot regions under SSP5–8.5, illustrating the growing mismatch between current federation safety criteria and projected climatic conditions across both hemispheres.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Projected number of days per season exceeding the WBGT > 32°C extreme-risk threshold in key hotspot regions under SSP5–8.5, for afternoon conditions. Values are multi-model ensemble means expressed as day ranges consistent with the text. Thresholds from Table 1 apply across all regions listed.

https://doi.org/10.1371/journal.pclm.0000969.t002

At the elite level, organisers of major competitions may need to consider long-term relocation of events to cooler seasons or cooler venues, particularly under SSP5-8.5. Such changes are already emerging sporadically (e.g., football tournaments shifted to winter, endurance events moved to dawn or dusk), but our results suggest that this may evolve from exception to structural necessity. At community level, local clubs, schools and amateur organisers may experience a significant compression of safe outdoor hours, with consequences for training quality, participation rates and equity of access—particularly in low-income regions where indoor alternatives and cooling infrastructure are limited.

The heterogeneous patterns in our results also imply that some currently marginal regions (e.g., high latitudes) may become more favourable for certain outdoor sports, while others may experience chronic climatic stress. This could gradually reshape the global geography of sport, potentially reinforcing existing inequalities or creating new opportunities depending on adaptation capacity, infrastructure and investment.

Adaptation via time-of-day: opportunities and limits

One of the most accessible adaptation levers is **time-of-day scheduling**. Our separation of WBGT into morning, afternoon and evening exposures clearly shows that shifting away from afternoon hours—where solar radiation, air temperature and surface heating peak—can substantially reduce the number of days above critical thresholds. This aligns with physiological evidence showing lower thermal strain during cooler parts of the day [3,5,24] and with practice-oriented recommendations to “play early or late” [1,15].

However, our projections also reveal the **limits** of this strategy. In several low-latitude regions under SSP5-8.5, mornings and evenings increasingly exceed safety thresholds, reducing the margin of benefit achievable through temporal shifting alone. Moreover, early-morning scheduling conflicts with athlete sleep, school and work constraints, while late-evening matches can create logistical and safety issues. Time-of-day adaptation is therefore useful but insufficient, particularly in regions where climate change drives WBGT exceedances across the full diurnal cycle.

Time-of-day adaptation also interacts with air quality in ways our thermal framework does not capture. Shifting activity to early morning generally lowers exposure to ozone, which tracks afternoon solar radiation, and to commuting traffic emissions; however, evening windows may coincide with elevated traffic-related PM_2.5 and NO₂, and even early-morning exposure is not negligible in dense urban cores [19,21]. The net benefit of temporal shifting therefore depends on local pollutant regimes, adding a dimension not represented in our WBGT-based analysis.

Additionally, our projections do not include microclimatic effects such as urban heat islands, shading, ventilation or stadium architecture, which may amplify or mitigate heat stress locally. Nonetheless, the broad patterns highlight a structural contraction of safe windows for outdoor activity.

Rainfall-driven cancellations and surface-related risks

While heat tends to dominate climate–sport discussions, our rainfall indicator underscores the importance of **hydrometeorological constraints**. Days with precipitation above 20 mm can lead to waterlogged natural turf, increased slip risk, reduced ball control and higher injury risk in football, rugby, cricket and other field sports. Even when athletes physiologically tolerate such conditions, surfaces may be unsafe or unplayable.

Although heavy rainfall increases appear modest in many mid-latitude regions, small increments may still lead to substantial operational impacts due to narrower recovery windows and increased pitch saturation. In monsoonal and tropical regions, heavy rainfall remains a dominant constraint, often overlapping with high thermal stress. This **compound hazard**—wet, soft surfaces combined with high temperatures—poses additional injury and scheduling challenges.

From an adaptation perspective, these patterns highlight the value of resilient surface design (drainage, hybrid turf, synthetic options) and flexible scheduling that incorporates both thermal and hydrological forecasts. However, such adaptations are costly and more accessible to professional teams than to community sport organisations, raising clear equity concerns.

Heatwave interruptions and cumulative stress

Heatwaves represent multi-day periods of elevated thermal load that challenge traditional training cycles and competition calendars. Unlike isolated hot days, heatwaves can impair recovery, elevate cardiovascular strain and increase heat-related illness risk even when individual days remain below acute danger thresholds [24].

Our projections show sharp increases in heatwave days across many regions, particularly under SSP2-4.5 and SSP5-8.5. This implies greater risk of disrupted training blocks, altered tapering, forced rest days and elevated cumulative fatigue. Recreational athletes, older adults and those with underlying conditions are likely to be disproportionately affected [14]. For sport medicine and performance staff, this underscores the need to monitor **cumulative thermal load** and integrate environmental stress into periodisation models—an emerging but underdeveloped area of practice.

Seasonal patterns (JJA vs. DJF) and hemispheric asymmetries

Our JJA/DJF analysis reveals clear hemispheric asymmetries relevant for sport scheduling. JJA shows the strongest increases in heat-related hazards in the Northern Hemisphere, while DJF becomes increasingly constrained in the Southern Hemisphere where summer sport is predominant.

These patterns intersect structurally with established calendars. Many global events are scheduled to optimise conditions in one hemisphere, but our results show that the concept of a “safe summer” is weakening. Federations may need to revisit long-standing assumptions about seasonal windows and consider cross-hemispheric rotations, shoulder-season scheduling or long-term calendar redesigns.

Scenario divergence between SSP2-4.5 and SSP5-8.5

The widening gap between SSP2-4.5 and SSP5-8.5 by late century has clear implications for both climate policy and sport governance. Under SSP1-2.6 and, to a lesser extent, SSP2-4.5, many regions experience increases in hazardous days that, while substantial, remain theoretically manageable through adaptation (scheduling shifts, facility upgrades, medical protocols). Under SSP5-8.5, however, several regions cross into regimes where **large fractions of the warm season** exceed thermal safety thresholds even in the coolest parts of the day.

This mirrors broader health and labour impact research [14,17], which shows that emissions trajectories determine the feasibility of safe outdoor activity. For sport, this means that global mitigation efforts are not simply an environmental imperative but a determinant of whether outdoor sport remains viable in its traditional form in major regions of the world.

Limitations and future directions

Several limitations must be acknowledged. First, although we use bias-corrected and downscaled CMIP6 projections, we do not perform explicit validation of WBGT, HI, rainfall and heatwave metrics against reanalysis products such as ERA5 or against dense station networks. While NEX-GDDP performs reasonably well for many variables, regional biases remain possible, particularly in complex terrain or coastal domains. Although station networks such as SYNOP provide temperature and humidity observations, solar radiation — required by the Liljegren WBGT formulation — is rarely available at the same locations, precluding systematic station-based validation of this indicator across hotspot regions. We therefore rely on published regional evaluations of NEX-GDDP-CMIP6 performance [28–30] as an indirect assessment of dataset skill.

Second, the WBGT formulation uses standard approximations rather than full radiative–wind models. This is common in large-scale assessments but introduces uncertainties when translating grid-cell conditions to venue-specific microclimates (e.g., shade, ventilation, stadium architecture) [6,7]. Our estimates should therefore be interpreted as **first-order indicators**, not precise diagnostics for specific facilities.

Third, we do not explicitly model acclimatisation, cooling strategies, clothing, hydration protocols or rule changes, all of which can substantially modify real-world risk [1,5]. Nor do we distinguish between sports with different metabolic loads or protective equipment. As such, our projections define a **climatic envelope of potential stress**, within which sport-specific risks may vary widely.

Fourth, our rainfall and heatwave definitions are intentionally simple and operational. More sophisticated metrics—compound indices, multi-day rainfall accumulations, sport-specific saturation thresholds—could be incorporated in future extensions. Similarly, while our multi-model mean reduces noise, we do not quantify model spread in detail, leaving some uncertainty in regional projections.

Finally, we do not model behavioural, organisational or economic responses. Recent evidence on athlete and organiser behaviour under heat stress [16,32,36] suggests that adaptive strategies are heterogeneous and context-dependent. Integrating these behavioural dimensions with physically based hazard projections remains a major research frontier.

Despite these limitations, our study provides an unprecedented global baseline for understanding how climate change may reshape the practice, scheduling and geography of outdoor sport. Future work could build on this foundation by (i) validating and refining indicators with reanalysis and observations, (ii) developing sport-specific thresholds, (iii) exploring alternative adaptation scenarios (e.g., systematic time-of-day shifts, infrastructure upgrades, calendar redesigns), and (iv) coupling climate hazard projections with participation, injury and economic data to develop integrated climate–sport risk assessments.

Future extensions could incorporate an air-quality dimension into the SCRI, given growing evidence that pollutant exposure both harms athlete health and degrades sport-specific performance [19]. Daily, bias-corrected PM_2.5 projections under CMIP6 scenarios are beginning to emerge [37], although public access remains limited; coupling such datasets with thermal hazards would allow a more complete multi-hazard assessment of outdoor sport viability.