Introduction

Outdoor workers are at special risk of heat exposure in the United States1 and globally2, particularly in agriculture, construction, and landscaping. Heat exposure is getting worse as the frequency and severity of heat waves are increasing in the US in recent decades3. Deaths due to extreme heat are higher than all other weather-related causes in the US4. Extreme heat events are also rising in California in recent periods5.

The Imperial and Coachella Valleys (ICV) in southern California (Fig. 1) have a population of nearly a half-million, with 25% living below the poverty line6. ICV is one of the largest and most productive agricultural regions of California and lies in a semi-arid region. Generally speaking, work-related death rates are high in the agriculture sector1,7 because agricultural labor is often characterized by long hours, strenuous physical activity, and exposure to extreme environmental conditions such as extreme heat. The ICV region has both the highest summer temperatures and the highest heat-related illness rate in California8. The majority of farm workers in the ICV are migrants and seasonal workers of Hispanic origin9. According to monthly surveys conducted by the Current Employment Statistics (CES) program10, there were 142,000 farm employees in Riverside County and 124,300 in Imperial County, with a total of 266,300 in the ICV region in 2020. The common crops grown in the ICV region include lettuce, cauliflower, kale, and broccoli in winter; onion, citrus, oranges, and carrots in spring; melons, corn, and wheat in spring and early summer; and dates, lemons, and olives in summer and fall11. Certain crops like sugarcane, alfalfa, and fodders are grown throughout the year. Due to the seasonal nature of employment, many migrant workers experience job insecurity and often work in physically demanding environments without consistent enforcement of labor protections7. In the US, field workers in agriculture are frequently exposed to extended work hours with minimal rest, particularly due to economic pressure12,13. Mandated rest periods and days off are often ignored during the harvest seasons, which creates additional health risks to farmworkers14. Given its desert environment with high temperatures, low rainfall (2.9 in/year), high solar radiation, added humidity from irrigation, and the presence of a large number of outdoor farm workers, the ICV region is an ideal place to examine heat exposure risks and to evaluate existing heat-related policies.

Fig. 1: Study area—the Imperial and Coachella Valleys (ICVs) in southern California.
figure 1

a WRF modeling domains; red box indicates the main WRF domain (d03), green and black boxes indicate the parent domains, and blue box indicates the IV cropped region, b satellite view of the ICV region; red line indicates the US-Mexico border.

Full size image

The risk of heat-related illness is governed by two sources of heat, environmental heat and metabolic heat15. Environmental heat refers to external heat due to the environment and weather. Metabolic heat refers to internal heat generated by the body, which depends upon the workload. In environments with high heat exposure, such as agriculture, prolonged work without sufficient hydration and rest increases the risk of heat-related illnesses16. High sweating rate in hot environments causes dehydration, which reduces evaporative and convective heat loss17. A recently proposed heat regulation by the federal Occupational Safety and Health Administration (OSHA) requires employers to provide workers access to cool drinking water, at least 1 quart per worker per hour18. Heat stress on workers also depends upon their acclimatization status, level of workload (light, medium, or heavy), and types of clothing19,20,21. An acclimatized worker is the one who has had repeated exposure to elevated heat stress that causes a series of physiologic adaptations whereby the body becomes more efficient in coping with the heat stress19. Similarly, an unacclimatized worker is the one who has not been exposed to the same working environment or level of heat stress and thus needs some time to “acclimatize”. Despite the availability of occupational guidelines that provide recommendations for hydration, protective temperature thresholds, and work-rest cycles in hot environments, compliance in agricultural settings is inconsistent in the US1.

Extreme heat stress can cause heat cramps, heat rashes, fainting, muscle rupture22, kidney dysfunction23,24, and death due to heat strokes25. Under heat stress, the cardiovascular system is called to meet the oxygen demand of the working muscles and to transfer the metabolic heat from the body core to the surface26. To reduce the risk of heat-related illnesses and deaths, various preventive measures are used which can broadly be classified into two categories: engineering controls, such as shielding from heat sources, ventilation and reducing metabolic heat with fans or air conditioning, and administrative controls, which include the adoption of work-rest regimens27, the main focus of this study. Preventive controls usually practiced in agricultural settings include making cool drinking water available to workers, providing access to shade, allowing preventive cool-down rest as needed, providing first-aid or emergency response for the workers showing heat illness symptoms, and observing employees through a “buddy” system or through effective communication28. Several studies have investigated the health impacts of environmental conditions on workers9,29,30,31, but research on work-rest schedules specifically in agricultural environments is limited. Proper work-rest cycles help the body to get rid of excess heat and improve human comfort, occupational health, and work productivity32. A rest break intervention study in southern California for stoop labor showed that alternative patterns of rest breaks 5 min every hour in addition to other regular breaks reduced workers’ musculoskeletal discomfort and fatigue without impacting productivity14. While there has been significant progress in identifying the factors affecting heat stress and in its monitoring and forecasting, the development of data-based work/rest schedules specifically designed to protect workers has not received much attention33. In many agricultural settings, work-rest schedules are determined based on the local situation, with breaks often governed by the progression of tasks rather than physiological needs or environmental conditions.

Specifications of rest minutes in current policies and field practices may not adequately account for the adverse effects of heat stress, long exposure to the sun, and physical exertion in agricultural settings. The rest minutes per hour (RMPH) specified in current policies are significantly lower than in some suggested guidelines34. For example, the National Institute for Occupational Safety and Health (NIOSH) guidelines on heat exposure19, originally based on the United States Environmental Protection Agency (EPA)35 and American Conference of Governmental and Industrial Hygienists (ACGIH®) specifications20, suggest equal or more rest minutes than the work period (30/30 min per hour) when the air temperature exceeds 95°F, for moderate work. However, Title 8, Section 3395 of California Code of Regulations (T8 §3395 CCR)28, which is an enforced policy on heat safety protocols in the Agriculture and Construction industry, specifies only 5 min of rest minutes for every hour of work performed when the air temperature is >95°F, and has no additional thresholds above 95 °F. Laboratory studies (treadmill walking or cycling) have shown that current ACGIH® or NIOSH guidelines on work-rest schedule do not adequately protect workers and thus may need further refinement27. Although various regulating agencies have provided guidelines on work-rest schedules, the effectiveness of work-rest regimens has not been validated experimentally27. A few studies have tested the physiological effectiveness of these specifications but only within laboratory conditions (treadmill, cycling)27,36.

We calculate average RMPH based on three key policies and guidelines that specify rest breaks for outdoor workers using three different heat stress indicators (HSIs) (Table 1). The first one is provided by the National Weather Service (NWS)37 based on Wet Bulb Globe Temperature (WBGT), which presents several thresholds with different rest-minute recommendations. The second is prescribed by CCR28 and is based on dry bulb temperature (DBT), which is simply the 2-m air temperature, with one threshold. The third one is a recently proposed rule “Heat Injury and Illness Prevention in Outdoor and Indoor Work Settings” by the federal OSHA18, which is based on the heat index (HI).

Table 1 Heat stress indicators and their criteria used to specify rest periods
Full size table

WBGT is a widely used standard measure for quantifying and monitoring heat stress and is an international standard for occupational heat exposure21. Unlike other simple heat indices that typically include the effect of air temperature and humidity only, WBGT additionally considers the effect of solar and thermal radiation and wind speed, and thus provides a more reliable measure of heat stress in outdoor environments38. WBGT was first adopted by the US Army after 600 heat casualties occurred in Marine Corps Recruit Depot, Parris Island (South Carolina) in the summer of 1952. After the adoption of WBGT to monitor heat stress during training, the heat-related casualties decreased by five to tenfold. The NWS has recently started using WBGT as an experimental product.

WBGT for outdoor conditions, which is applicable in agricultural settings, is given by the weighted average of natural wet bulb temperature (WBT), black globe temperature (BGT), and DBT:

$${WBGT}= \, 0.7 \, \times {WBT}+ \, 0.2 \, \times {BGT}+ \, 0.1 \, \times {DBT}$$
(1)

where WBT is typically measured by a “wet-bulb” thermometer exposed to natural environment under the influence of solar and thermal radiation as well as wind, BGT is measured by a black-globe thermometer (a thermometer or a temperature sensor placed at the center of a black globe which absorbs all radiation falling onto it), and DBT is air temperature commonly measured at 2 m above the ground surface at weather stations. All of these parameters can be simulated using weather models with reasonable accuracy39,40,41,42,43.

The HI, also known as the apparent temperature, is what the temperature feels like to the human body when relative humidity is combined with DBT. It was widely used in the past by NWS for heat stress forecasting, but has become less common with the emergence of WBGT. The HI is calculated as a function of DBT and relative humidity using the equation given by Rothfusz (1990)44, which was obtained by a multiple regression analysis on a look-up table originally developed by Steadman (1979)44,45, with further corrections applied at extremes of temperature and humidity46, as adopted by NWS47.

The various criteria adopted by different agencies to specify RMPH based on the above three HSIs are detailed below.

WBGT-based rest breaks recommended by NWS (WBGT-RMPH)

The NWS WBGT-RMPH prescribes rest minutes for different ranges of WBGTs, beginning at 15 rest minutes every hour when WBGT exceeds 80 °F, with additional rest minutes at 85, 88, and 90 °F (Table 1). These NWS specifications of rest minutes37 appear to be adapted from ACGIH recommendations20, which are also found in a separate report produced by US-EPA in 199235. Similar WBGT-based rest minutes can also be found in the US Army’s Technical Bulletin (TB MED 507)48.

DBT-based rest breaks specified by CCR (DBT-RMPH)

The DBT-RMPH prescribed by CCR28 requires employers to provide 5 rest minutes every hour when the air temperature exceeds 95 °F (Table 1). DBT-based specifications can also be found in all NIOSH19, ACGIH20, EPA35, and US Army48 documents. This heat-related rest break is referred to as preventative cool-down rest period in CCR which has the same meaning as “recovery period” in California Labor Code Section 226.749.

HI-based rest breaks proposed in OSHA’s new rule (HI-RMPH)

The federal OSHA has recently (2024) proposed a new rule on heat injury and illness prevention18 that requires employers to use either HI or WBGT to determine access to shade and rest periods for employees. Employers are required to provide 7 ½ rest minutes every hour when the HI exceeds the high-heat trigger of 90 °F. This specification is 2 ½ min higher than the CCR but uses HI criteria instead of DBT. The specification using WBGT is not explicitly mentioned in OSHA’s new rule and is referred to NIOSH guidelines. The original specification of rest minutes in CCR and OSHA’s proposed new rule is 10 and 15 min every 2 h, respectively. We have converted them to RMPH, i.e., 5 and 7 ½ min per hour for CCR and OSHA, respectively, for making the values comparable across different HSIs.

While the existing work-rest criteria represent a compromise among policymakers, industrial representatives, labor unions, and community organizations, a scientific, data-based calculation of rest periods is desired in order to compare policies. In this study, we calculate rest minutes for farmworkers in the ICV region considering different factors that affect heat exposure including geography or crop environment, acclimatization status, clothing, work shift hours, and workload. We use the NIOSH recommended criteria based on WBGT19 to calculate rest minutes, which were originally derived using experiments involving mine workers in South Africa. These guidelines are expected to protect most healthy workers, and are assumed to be applicable to all outdoor workers including agricultural and construction workers. The WBGT limits for acclimatized and unacclimatized workers, prescribed in the NIOSH document (Fig. 8–1, 2 and pages 94–95) as the recommended heat stress alert limits (RALs) and recommended heat stress exposure limits (RELs), respectively, are derived for a male of 70 kg with 1.8 m2 body surface area. NIOSH notes that given the similar physiological ability to tolerate heat, there are no significant differences between the sexes as they both adapt well to heat exposure19. The use of different clothing factors and metabolic rates allows adaptation of these guidelines to agricultural settings considering different levels of work activity.

We calculate rest minutes based on the exceedance frequency of the three different HSIs mentioned earlier, derived from hourly WRF model outputs at 1-km spatial resolution. The WRF model for the ICV has been validated previously and includes the effect of irrigation39, which is not typically included in global and regional climate models, but is a key parameter governing heat stress in the study region. We calculate the hours above or within each HSI threshold and translate them into average rest minutes based on respective policies/guidelines (see “Methods”). Since heat exceedance hours, and consequently the rest minutes, vary by season, we also consider the seasonal variation in rest requirements by considering all warm months of a year (April–October), with an aim to better inform relevant labor and public health policies. We also consider work shift hours in our analysis, as it is possible to optimize work schedules by minimizing rest minutes required, which will also be of interest to policymakers, employers, and farmworkers. Our methodology of calculating rest minutes based on WBGT satisfies criteria adopted by NIOSH for heat stress monitoring19: it is simple to calculate, does not require physiological monitoring, and is applicable to a wide range of working conditions. We intentionally use degree Fahrenheit (°F) instead of degree Celsius (°C) as the unit of different HSIs used in this study, because the former is more frequently used in policy guidelines and also provides a more familiar scale of temperature for people and policymakers in California and across the US. We explore the following three specific research questions in this study:

  1. 1.

    Which of the heat stress indicators used for specifying rest breaks, WBGT, DBT, or HI, is most protective of farmworkers?

  2. 2.

    How do rest minute requirements vary by work shift and season?

  3. 3.

    What are the rest minute needs for unacclimatized and acclimatized workers considering occupational heat exposure criteria?

Results

Comparison of heat stress indicators (HSIs)

WBGT values exceed the lowest threshold of 80 °F (Table 1) for 24% of the grid-hours while considering all work hours (including day and night), but exceed 43% of the grid-hours during the standard work hours (6 am–2 pm) (Fig. 2) for April to October. The thresholds of DBT (95 °F) and HI (90 °F) are exceeded 39% and 45% of the time during work hours. The exceedances of this lowest threshold are comparable among the three HSIs. The histogram of HI values has a sharp vertical change around 75 °F because HI is defined only within the air temperature range of 68–122 °F45. The HSI thresholds are most frequently exceeded in June, July, August, and September with considerable exceedances in April and October as well (Fig. S1).

Fig. 2
figure 2

Histogram of WBGT, DBT, and HI in the IV crop field region (Fig. 1, blue box) The pink, red, orange, and yellow colors for WBGT represent the four WBGT categories of ≥90°F, 88–90°F, 85–88°F, and 80–85°F, respectively, as defined in Table 1. For DBT and HI, the yellow color denotes their respective threshold values of 95°F and 90°F. Histograms represent all hours (top row) and work hours (i.e., 6am-2pm, second row) during the study period (April to October 2020).

Full size image

Rest minutes obtained using WBGT (WBGT-RMPH) show stronger spatial variation compared to those using DBT (DBT-RMPH) and HI (HI-RMPH) (Fig. 3). The DBT-RMPH and HI-RMPH both show a lower and more uniform rest minute pattern across the study domain compared to WBGT-RMPH. There is little difference in rest minutes obtained while considering all hours (day and night) and work hours, for both DBT-RMPH and HI-RMPH. However, WBGT-RMPH shows much higher spatial variability when comparing rest minutes obtained using work hours vs. all hours (Fig. 3a, b). The WBGT-based RMPH are also much higher compared to those based on DBT or HI during the work hours. The higher spatial variability using WBGT compared to HI and DBT can be explained by the fact that WBGT includes multiple physical terms in its equation (Eq. 1) and has multiple rest minute specifications for different ranges of WBGT. The WBGT-RMPH during work hours is highest around the Salton Sea area, and in cities and desert, and fallow areas. The crop fields show lower rest minute requirements than the surrounding desert, reflecting the evaporative cooling effect of irrigation in the daytime. The higher rest minutes in urban centers reflect the urban heat island effect. In summary, the WBGT-RMPH shows stronger sensitivity to location and work hours compared to DBT-RMPH and HI-RMPH and thus better resolves the heat exposure in space and time. Therefore, we perform further calculations of rest minutes using WBGT only.

Fig. 3
figure 3

Required average heat-related rest minutes per work-hour for farmworkers in the Imperial and Coachella Valleys. Calculated based on the exceedance of (a, b) WBGT, c, d DBT, and (e, f) HI thresholds specified in three existing policy guidelines being compared (Table 1), considering all hours (left) and standard work hours (right) during the study period (April to October, 2020).

Full size image

Rest minutes for different work shift hours and seasons

The rest minute requirement using WBGT, calculated while keeping the months constant by averaging over all warm months, varies greatly by work shift (Fig. 4a). The maximum and minimum rest requirements are observed for 9 am–5 pm and 9 pm–5 am shifts, respectively. Although the 9 pm–5 am shift provides the fewest rest minutes, it may not be practical in farm work settings. The 6 am–2 pm work shift is more commonly practiced in the IV region, but a 3 am–11 am work shift is also practiced during summer, which has low RMPH (90th percentile = 4.2 min). These rest minutes reflect the average rest minutes considering all warm months; individual months have much higher or lower RMPH, as the results show below.

Fig. 4: Variation in rest minute requirements in the IV region calculated based on the exceedance of WBGT.
figure 4

a Work shift, with fixed months (Apr–Oct) b Months, with fixed work shift (6 am–12 pm). The box plot parameters show variability within the grid cells of the IV crop fields (blue box, Fig. 1), with 90th percentile values shown on the top of the whisker.

Full size image

RMPH also varies greatly with season (Fig. 4b), calculated while keeping the work shift constant to 6 am–2 pm. August requires the most RMPH (90th percentile 22.8 min) while April and October require the least. RMPH is remarkably higher when considering the mid-day work hours only (12 pm–2 pm), with ~1.5 times more RMPH than for the standard 6 am–2 pm work shift for August and double the RMPH for October (Fig. S2).

In the ICV crop fields, different crops are harvested in different months and the work shifts vary by season. We therefore use the information from Fig. 4 to adjust the rest minutes obtained from NIOSH criteria to calculate the variable rest break needs in different months and work shifts.

Rest minutes based on NIOSH occupational heat exposure criteria

The mean RMPH obtained with NIOSH criteria using WBGT (see “Methods”), considering all warm months and a standard work shift (6 am–2 pm), also differs for acclimatized vs. unacclimatized workers and at various levels of work activity (Fig. 5). Unacclimatized workers performing heavy works (metabolic rate ~415 W) need the most rest (90th percentile 39.8 min per hour). Acclimatized workers performing light work activity (metabolic rate ~180 W) require the least rest (90th percentile 14.4 min per hour). These results show that WBGT-based RMPH is sensitive to acclimatization status and work load variation.

Fig. 5: Required rest minutes per hour as per the NIOSH occupational heat exposure criteria.
figure 5

Calculated using (a, c, e) recommended heat stress alert limits (RALs) for unacclimatized workers, and (b, d, f) recommended heat stress exposure limits (RELs) for acclimatized workers, for different metabolic rates (light, medium, heavy). Includes standard work hours (6 am–2 pm) of April–October, 2020. 90th IV means 90th percentile within the IV crop fields (Fig. 1).

Full size image

Recommended RMPH for farmworkers, calculated based on NIOSH criteria, and corrected for the variation with seasons and work shifts (see “Methods”), is presented in Table 2. For the most common work shift in the study region (6 am–2 pm), recommended RMPH is minimum in April (2 min) and maximum in August (32 min). Early morning (3 am) and late afternoon shifts (3 pm) have much lower RMPH than the 9 am and 12 pm shifts. The 9 am and 12 pm shifts should preferably be avoided in July, August, and September as the rest breaks needed are very high (>25 RMPH). Even for the commonly adopted work shift (6 am–2 pm), RMPH in August is very high (32 RMPH).

Table 2 Recommended rest breaks (minutes per hour) for the farmworkers in California by month and work shift
Full size table

Discussion

In this work, we calculated recommended RMPH for farmworkers based on exceedance of HSIs in the agricultural regions of southern California, the ICVs, using the WRF model. We considered factors that affect heat exposure including location, work shift, season, clothing, acclimatization, and workload conditions. We calculated RMPH based on NIOSH criteria19, correcting for variable heat exposure in different work seasons and hours.

With the rising number of heat waves in California, in the United States, and globally, research on heat stress monitoring and forecasting has gained significant attention. The NWS has recently launched HeatRisk50 and the California Office of Environmental Health Hazard Assessment (OEHHA) has proposed a similar heat information tool called CalHeatScore51 to provide more localized heat forecasts while accounting for health effects52. Heat-related regulations are also rapidly changing both at federal and state levels to protect workers exposed to high-heat environments. Although the newly introduced rule on heat injury and illness prevention by OSHA requires employers to use either HI or WBGT to determine access to shade and rest periods for employees, the requirements for WBGT are not specified; instead referred to NIOSH guidelines, which have complex specifications. California’s OSHA has already introduced standards for preventing heat illnesses for indoor workers (T8 §3396 CCR)53, and is expected to change its outdoor heat illness guideline in CCR (T8 §3395 CCR)28, which currently uses air temperature (or DBT) alone to specify heat safety protocols in the Agriculture and Construction industry28. In this context, our results can provide a basis to make more specific policy guidelines on rest breaks.

Despite its accuracy in predicting heat-related illness, WBGT has not been commonly used in practice and policies. One key reason is the need for the continuous measurement or estimation of WBT, BGT, and DBT, which is challenging in large agricultural areas such as the ICV region. Secondly, WBGT-based rest minutes are also usually higher than what is currently practiced. Therefore, employers might perceive that taking longer rest breaks could reduce productivity, although taking required rest breaks may not impact productivity14. Thirdly, installing and monitoring WBGT on-site might be difficult, especially in large agricultural fields that require multiple sites. Hand-held WBGT instruments are becoming more common recently, even though their accuracy might be lower than standard WBGT instruments54. Fourthly, farmworkers are sometimes paid by piecework, so they may end up ignoring the work-rest cycle. Finally, the use of WBT would require the peer-manager to monitor WBGT (even if on an app) and adapt the work schedule accordingly. Following WBGT guidelines may increase the workload for peer managers and they may be reluctant to follow the suggested guidelines.

Our approach addresses some of the above issues by using WBGT climatology to determine rest break requirements. Our approach provides a more convenient way of using WBGT for protecting farmworkers. The employers and policy makers can refer to the recommended WBGT-based rest minutes provided in this study without having to measure WBGT continuously in the field.

Our results demonstrate that the rest minutes based on WBGT yields much higher rest break requirements than both the DBT, which is currently in use in California28, and the HI, which is proposed in OSHA’s new rule18. Our analysis suggests that WBGT is a better indicator of heat exposure for outdoor workers than DBT or HI, in large part because it specifies multiple thresholds, all of which are commonly exceeded in our study area and likely in other locations in California. We conclude that existing air temperature-based heat illness prevention regulations are likely inadequate and not protective of farmworkers in hot regions and seasons. Therefore, we recommend adopting WBGT-based rest breaks, which, as our results show, is sensitive to the spatiotemporal variation in heat exposure and thus most protective.

The recommended WBGT-based rest minute specifications (Table 2) can be directly incorporated in regulations such as T8 §3395 CCR28. We developed these recommended rest minutes basing for unacclimatized workers, which will ensure protection for both acclimatized and unacclimatized workers, and corresponding to a metabolic rate of 300 W, which should represent average activity level for most farmworkers. While these recommended rest minutes provide generalized guidelines, Tables S1 and S2 can be used to obtain a full set of rest minutes required for both acclimatized and unacclimatized workers, for all work shift hours and seasons, and considering all three different levels of work (light, medium, heavy). To get specific rest minute estimates, one only needs to multiply maximum rest minutes listed in Table S1 with the correction factor listed in Table S2 for a desired work shift and month. The rest minutes required for various crop environments can be calculated by considering the level of activity involved using Tables S1 and S2. For example, rest minutes for workers involved in picking dates and cutting sugarcane fall under “heavy” work and the rest minutes should be calculated using the rest minute values for “heavy” work provided in Table S1.

Preliminary survey data of farmworkers (unpublished) shows that workers are taking rest breaks throughout the day. These breaks vary by length of workday and across workers. In addition to a 30-min lunch break, they are typically allowed 10–15 min every 2–3 h during an 8–10 h shift, though some report no additional breaks during a 6-h shift between 6 am and 12 pm (unpublished data). Employers may allow workers to leave early, typically at 12 pm for a 6 am–2 pm shift, during extreme heat conditions, while still being paid for all hours in the work shift. While these practices are commendable, the heat-related rest breaks specified by CCR above a threshold does not seem to have been followed. More frequent breaks can significantly lower the risk of heat-related illnesses14,32, so heat-related rest breaks should be adhered to for worker safety. Employers do provide access to shade required by CCR but often under a small, mobile shade structure with only overhead coverage (unpublished data). Some report lavatories connected to or in close proximity to the shade structure. While access to restrooms is an important component of heat illness prevention because it supports water consumption and elimination23,55, proximity to the shaded break area may also deter workers from spending breaks in the shade and limit their ability to recover from heat stress. Our rest minute calculations are heat-related rest minutes since they are based on the exceedance of WBGT over certain thresholds. These heat-related rest minutes are to be provided in addition to other rest breaks required by applicable labor laws. For example, the California Labor Code 226.749 has a provision of a 10 min rest break every 4 h of work regardless of heat conditions, and another 30 min meal break when the work period is more than 5 h56.

To assess the impact of year-to-year climate variability on our results, we compared rest minutes from the three HSIs derived from WRF model output from 2020 with those from 2024. The comparison of rest minutes in the two years (2020 and 2024) obtained using WBGT, HI, and DBT shows consistent spatial patterns (Fig. S3), showing the robustness of our methodology. This is reasonable because the occurrence of a few more or less heat events in a year would not significantly change the average number of exceeding hours.

Our calculations are done for the ICV region, which has the highest air temperatures and heat-related illnesses in California, so the recommended rest breaks should be protective to most farmworkers in California. However, the protective WBGT thresholds are different in different climate regimes due to geographical variation in heat exposure and acclimatization, which could influence exceedances and resulting rest minutes. Work-rest regimens depend upon environmental conditions and so are site-specific27. For other climate regions of the US, and for other global regions, these recommended values might be overprotective or underprotective depending upon the regional climate. Therefore, we recommend calculating rest minutes using our approach, but using regional WBGT climatology to derive region-specific rest break requirements.

Although the WBGT parameter used in our analysis is an internationally recognized HSI21 that accounts for key parameters affecting human heat stress including air temperature, humidity, radiation, and wind speed, it has some limitations57. WBGT does not fully consider the individual’s physical condition, age, or personal tolerance to heat. Certain workers with higher heat illness risk such as pregnant women, older adults or seniors, and those with pre-existing health conditions, may have more specific needs for rest breaks and our recommended rest minutes may not be sufficient to protect them. Our results are based on monthly averages for 2020 and thus may not reflect instantaneous rest break needs during weather events such as heat waves that involve rapid changes in temperature, humidity, winds, and radiation.

The maps of HI exceedances calculated in this study are based on the HSIs calculated using gridded WRF model outputs. Although the model calculated WBGTs used for calculating the rest minutes in this study were previously validated using station data39, they have inherent biases compared to station-derived WBGTs due to their coarse spatial resolution (~1 km). The root mean squared error of model-simulated WBT and DBT terms of the WBGT equation (Eq. 1) together range from 0.7 to 1.5 °C and that of the BGT term ranges from 0.5 to 0.7 °C, as compared to multiple stations lying in the study region39.

Methods

WRF modeling

We use high-resolution outputs from the WRF v4.4 model adapted for the ICV region to calculate the exceedance of HSIs in this study, which has been described in Parajuli et al 39. The key input parameters of HSIs including air temperature, relative humidity, dew point temperature, shortwave radiation, and wind speed, as well as WBGT parameters have been previously validated39. Briefly, the model setup consists of three nested domains with spatial resolution of 9 km, 3 km, and 1 km for d01, d02, and d03, respectively (Fig. 1a). The rest minute calculations are carried out in the innermost domain d03, which covers the entire cultivated lands of the Imperial Valley in the south and Coachella Valley in the north. Our model includes irrigation, which is heavily applied in the ICV crop fields. More than half of the farm workers in the IV area come from the cross-border city of Mexicali9, which is also included within the study domain. In California, heat surveillances (heat-related illness and deaths) are carried out during the warm season between April and October58, and heat waves are also normally defined within this period59. Therefore, we also base our analysis during this time window. WRF model simulations were carried out between April and October of 2020 at 1-km spatial resolution, then used to derive HSI exceedances and subsequently rest minutes.

Calculation of WBGT, DBT, and HI

The WBGT in this study is calculated using the formulation developed by Liljegren et al 60, which is considered a gold standard for calculating WBGT using meteorological or climate model outputs. Originally written in the C-language program, the method was recently implemented in python by Kong and Huber 40,41. Two methods are available for calculating WBGT, one using Liljegren’s original formulation that only uses shortwave radiation (WBGT_Liljegren), and a modified one to include additional radiation parameters from General Circulation Model (GCM) outputs including longwave radiation (WBGT_GCM). Weather stations typically only have shortwave radiation and many weather and climate models also do not typically save additional radiation parameters; in such circumstances, WBGT_Liljegren is preferred. However, in our case all radiation parameters required by WBGT_GCM are available from our customized WRF model so we are able to calculate WBGT incorporating the impact of both shortwave and longwave radiation on WBGT.

The inputs required for WBGT calculation using Liljegren’s method include DBT, relative humidity, surface pressure, 2- or 10-m wind speed, surface downward and upward solar radiation, surface downward and upward longwave radiation, solar zenith angle, and the fraction of direct horizontal solar irradiance to the total horizontal solar irradiance. An example code to calculate WBGT using the publicly available Python script developed by Kong and Huber 41 is provided below for reference:

$${wbgt}{{\_}}{gcm}{{\_}}k = {xr}.{apply}{{\_}}{ufunc} \\ \quad \left(\right.{WBGT}{{\_}}{GCM},\,t2{{\_}}k,{rh},{psfc},{va},{swdnb},{swupb},{lwdnb},{lwupb},\,f,{czda},\,$$
$${False},{output}{{\_}}{dtypes}=[{float}]\left)\right.$$
(2)

The description of the above input parameters and the corresponding WRF output variables are provided in Table 3.

Table 3 Input parameters required to calculate WBGT41,61 and their equivalent WRF outputs
Full size table

HI is calculated using MetPy61, with the equation given by Rothfusz (1990), which was obtained by a multiple regression analysis on the data table originally developed by Steadman (1979)44,45, with further corrections applied at extremes of temperature and humidity46, as adopted by NWS47:

$${{\rm{HI}}}= -42.379+2.04901523\times {{\rm{T}}}+10.14333127\times {{\rm{R}}}-0.22475541 \\ \times {{\rm{T}}} \times {{\rm{R}}}-0.00683783\times {{{\rm{T}}}}^{2}-0.05481717\times {{{\rm{R}}}}^{2}+0.00122874\times {{{\rm{T}}}}^{2} \\ \times {{\rm{R}}}+0.00085282 \times {{\rm{T}}}\times {{{\rm{R}}}}^{2}-0.00000199\times {{{\rm{T}}}}^{2}{{{\rm{R}}}}^{2}$$
(3)

where T is air temperature (°F) and R is relative humidity (%).

Note that the Steadman’s original table only specifies apparent temperature within the air temperature range of 68–122 °F and dew point temperature range of 32–86 °F, so the range of HI is limited (see “Results”).

Comparison of heat stress indicators (HSIs)

For the purpose of comparing different HSIs, we calculate the average number of rest minutes required per hour in each model grid cell in terms of the exceedance of respective HSIs. Since the recommended rest breaks (Table 1) are to be provided only during the work hours exceeding the thresholds, we use the fraction of exceeding hours to calculate the average rest minutes needed as follows:

$${{RMPH}}_{{WBGT}}=\frac{{EH}(80\le {WBGT} < 85)\,\times \,15\,+{EH}(85\le {WBGT} < 88)\,\times \,30\,+{EH}(88\le {WBGT} < 90)\,\times \,40\,+{EH}({WBGT}\ge 90)\,\times \,45\,}{{Total\; number\; of\; hours}}$$
(4)
$${{RMPH}}_{{DBT}}=\frac{{EH}({DBT} > \,95)}{{Total\; number\; of\; hours}}\times \,5$$
(5)
$${{RMPH}}_{{HI}}=\frac{{EH}({HI} > \,80)}{{Total\; number\; of\; hours}}\times \,7.5$$
(6)

where EH is the exceedance hours during the month and work hours considered; 15, 30, 40, and 45 are NWS recommended RMPH based on WBGT; and 5 and 7.5 are those recommended by CCR and OSHA’s new rule based on DBT, and HI, respectively.

Rest minutes considering work shift hours and seasons

As the WBGT clearly shows a more reasonable heat exposure pattern, we use WBGT-based criteria in the rest of the calculations to derive recommended rest break requirements. In the study region, farmworkers are involved in planting and harvesting activities in all months of the year, among which the warm months (April–October) are most critical from heat perspective. The actual work shift hours adopted by employers also vary by seasons but a 6 am–2 pm shift is more common in the warm months. Therefore, one of the objectives of this work is to determine rest break requirements accounting for the variation in heat exposure during different seasons and work hours.

We calculate rest minutes across the entire study domain but focus on the IV crop fields (Fig. 1) for policy recommendations, since most of the farmworkers are employed within this area. As the rest minutes vary spatially across the crop fields (Fig. 2), we use the “90th” percentile value among all grid cells lying within the IV cropped region (Fig. 1) to ensure protection in most of the agricultural field. We calculate rest break requirements in different work shift hours by keeping the months constant, i.e., using all months from April to October. We use eight work shifts consisting of 8 h each, with a 3-h increment (viz. 12 am–8 am, 3 am–11 am, 6 am–2 pm, 9 am–5 pm, 12 pm–8 pm, 3 pm–11 pm, 6 pm–2 am, and 9 pm–5 am). Similarly, we calculate the variation of rest break requirement in different months by keeping the work shift constant, using the most common work shift (6 am–2 pm). We do these work-shift and monthly calculations using the NWS WBGT guidelines (Table 1). We perform these calculations for standard conditions without any clothing corrections applied to WBGT values because these calculations are only used to derive rest minute correction factors for different months and work shift hours. The month and work shift correction factors so derived are then used to adjust rest break requirements calculated from NIOSH criteria that incorporate clothing correction, acclimatization status, and level of work activity, as described in the following section.

Rest minutes based on NIOSH occupational heat exposure criteria

Heat stress experienced by a worker depends upon clothing and level of physical activity or metabolic rate. For example, a person involved in heavy work activity will experience heat stress at lower WBGT compared to the worker involved in light work activity. Similarly, heat stress experienced by the worker increases significantly if the worker is wearing double or more layers of clothing. ACGIH20/NIOSH19 recommends using adjusted WBGT (WBGTeff) to account for the effect of metabolic heat and clothing. Workers typically wear double-layered clothing in the study area, therefore, we use NIOSH recommended clothing adjustment factor of 3 °C (NIOSH, Tables 3–2) for double-layered clothing and increase our model-calculated WBGT values uniformly by 3 °C (5.4 °F) to account for clothing effect.

NIOSH recommends using different thresholds for unacclimatized and acclimatized workers to account for the effect of metabolic heat, given by Recommended Action Limit (RAL) and Recommended Exposure Limit (REL), respectively, as per the following equations:

$${RAL}\left({WBGT}\right)=59.9-14.1{\log }_{10}M$$
(7)
$${REL}({WBGT})\,=56.7-11.5\,{\log }_{10}M$$
(8)

The total heat exposure for unacclimatized and acclimatized workers should be controlled so that workers are not exposed to combinations of metabolic and environmental heat greater than the applicable RAL and REL19. NIOSH prescribes RAL and REL curves as a logarithmic relationship between WBGT and metabolic heat for four different work/rest schedules with 0, 15, 30, and 45 RMPH, or 60, 45, 30, and 15 min of work minutes per hour, respectively. However, equations for all four of these curves are not available in the NIOSH document, only the standard equations for the no-rest scenario are provided (Eqs. 7 and 8). Therefore, we determined the RAL and REL thresholds approximately by reading the values off the curves provided in the NIOSH document19 (Figs. 8–1 and 8–2). We extracted four RAL and REL threshold ranges from these curves corresponding to three metabolic heat of 180, 300, and 415 W, representing three typical work categories light, medium, and heavy, respectively15, using a grabbing software that extracts the x and y coordinates from the given curves. These thresholds are then compared to the WBGTeff values, which are obtained by adding the clothing correction factor of 3 °C (5.4 °F) to our model-derived WBGT values, to calculate exceedances. We then assign their corresponding rest minutes of 15, 30, and 45 min for the exceeding hours as in Eq. 2 to calculate the average rest minutes for all three work categories, and for both acclimatized and unacclimatized workers (Table 4). For the extreme heat conditions above the RAL/REL limit, NIOSH recommends that no work should be performed to protect workers. Therefore, we assigned 60 minutes of rest time, which effectively means no work, for the hours that have WBGTeff above the last range of RAL/REL limits (Table 4).

Table 4 NIOSHa RAL and REL ranges for light, medium, and heavy work categories
Full size table

Calculation of recommended rest breaks

To adjust the NIOSH-based rest minutes to reflect the variable rest minute requirements for different work shifts and different months, we combine the calculations in the preceding two sections as follows. Results show that the maximum rest break is needed in the month of August, and for a work shift of 9 am–5 pm (Fig. 4). Therefore, we first calculate the maximum rest break requirements considering NIOSH criteria as in Fig. 5 but using the month of August and 9 am–5 pm work shift (Fig. S4). These maximum rest minutes are summarized as 90th percentile values within the IV region (Table S1). We then apply correction factors to these maximum rest minute values to calculate variable rest requirements for different months and work-shift hours based on Fig. 4a, b. The work-shift correction factor is obtained by dividing the average rest minutes for the IV for a given work-shift by the maximum rest minute corresponding to the 9 am–5 pm work-shift shown in Fig. 4a. Similarly, the month correction factor is obtained by dividing the average rest minutes for the IV for a given month by the maximum rest minute corresponding to August presented in Fig. 4b. For example, the work shift correction factor for 6 am–2 pm is 8.8/16.0 (0.550) and the month correction factor for July is 16.5/22.8 (0.724), which are grayed in Table S2. The combined correction factors are obtained by multiplying corresponding work-shift and month correction factors from grayed cells for a given work shift and month (Table S2). Finally, these combined correction factors are applied (multiplied) to the maximum rest minute requirements obtained using NIOSH criteria (Table S1) to get variable rest minutes required for different work shifts and months. Tables S1 and S2 should be used together to determine rest minute requirements for different acclimatization levels, months, and work shifts. We calculated the recommended RMPH (Table 2) basing for unacclimatized workers, to provide protection for both unacclimatized and acclimatized workers but using an average metabolic rate of 300 W. Recommended RMPH are presented for six likely work shifts, rounded to the nearest whole number.