Rational determination of real‐life UV exposure across latitudes: An extended Diffey‐based model integrating behavioural factors

Abstract

Objectives

In this study, we quantify both the UVB and UVA doses emitted by the sun at ground level and the realistic effective UVB and UVA doses received (ERD) by an individual across latitudes from 0° to 60° North and for every month of the year. These outputs provide a quantitative basis to derive protection indices and to support informed discussion on photoprotection. Our aim is not to prescribe SPF guidance for consumers or dermatologists, but to offer a robust scientific framework that clarifies how real‐life exposure, behavioural factors, effective UVB and UVA‐weighted doses and solar geometry collectively shape the necessary UV protection.

Methods

Daily UV doses were computed for the 21st day of each month to capture seasonal extremes. These doses, expressed as Standard Erythemal Doses (UVB) and UVA‐weighted doses, were calculated for each latitude based on Diffey's spectral irradiance framework. Two corrective factors were applied: (i) effective outdoor exposure time and (ii) body orientation. These adjustments allowed conversion of ambient daily doses into effective UVB and UVA‐received doses (ERD). Finally, a third factor, representing realistic sunscreen application density (0.8 mg/cm²), was introduced to estimate the necessary level of UV protection.

Results

Across all latitudes and seasons, UV‐weighted doses requirements follow predictable patterns driven by solar geometry. At most temperate latitudes, daily SED values range from approximately 5 to 60, while extreme values of 65–70 SED occur in tropical regions (10–30° N). The greatest differences between latitudes arise from seasonality. UVA‐PF requirements exhibit similar constraints.

When converted into effective received dose and then expressed in SPF terms, values above 60 provide minimal additional UVB benefit under realistic conditions.

Conclusions

By converting ambient irradiance into effective received dose, this model provides a realistic framework for daily UV exposure. It complements Diffey's approach rather than replacing it, supports rational discussion on SPF and UVA‐PF adequacy without prescribing individual recommendations and offers a quantitative basis for realistic product design, regulatory discussion and consumer guidance.

This study models real‐life UVB and UVA exposure across latitudes and seasons using an extended Diffey framework with behavioural corrections. Findings reveal predictable dose patterns and support pragmatic SPF and UVA‐PF caps, offering a quantitative basis for realistic photoprotection and regulatory discussion.

INTRODUCTION

Today, more than one in three cancers worldwide is a skin cancer. Over 8 million cases are attributed to solar overexposure. Approximately 80% of visible signs of skin ageing are also linked to excessive UV exposure [1, 2, 3]. These figures continue to rise globally, with few regional exceptions [4, 5]. While these numbers highlight the urgency of effective photoprotection, they also underline the need for a sound scientific understanding of real‐life UV exposure and sunscreen use.

Traditional models, such as the spreadsheet developed by Diffey [6, 7], provide robust estimates of ambient spectral irradiance under standardized, cloudless‐sky conditions. These tools are invaluable reference frameworks. However, sunscreen performance and consumer protection depend not only on ambient irradiance but on the effective UV dose actually received by human skin (ERD).

Misinterpretation of theoretical irradiance often fuels misconceptions, such as the assumption that consumers are continuously exposed to extreme UV levels or that SPF values must increase indefinitely to ensure safety. In reality, behavioural, anatomical and practical factors substantially reduce effective exposure.

The purpose of this study is not to prescribe SPF values for consumers or dermatologists or to minimize the risks associated with solar exposure. It aims to complement Diffey's approach by converting environmental irradiance into realistic effective received doses (ERD), integrating behavioural and anatomical factors. Such a framework can support rational SPF and UVA‐PF ranges, inform product design, regulatory and scientific discussions.

MATERIALS AND METHODS

This study builds primarily on two foundational publications by Brian Diffey [6, 7]. The strength of Diffey's model lies in its relative simplicity: with only latitude, date and time as inputs, it provides spectral irradiance under clear‐sky conditions.

We therefore adapted the calculations generated by the Excel sheet in Appendix S1 proposed by B. Diffey [7] to compute daily UVB doses weighted by the CIE erythemal action spectrum and UVA doses weighted by the PPD action spectrum [8]. These computations were performed for the 21st day of each month and for all latitudes between 0° and 60° N, a choice that captures seasonal extremes near solstice and equinox positions without adding unnecessary complexity.

Our objective was not to increase complexity, but to translate ambient irradiance into realistic human exposure using a minimal set of scientifically justified corrections. These modifications form the basis for calculating ERD, which integrate behavioural and anatomical factors into an extended Diffey‐based framework.

Fraction of daylight spent outdoors

Even in extreme situations, it is highly unlikely that an individual would spend more than 8 h per day in full sun exposure [9, 10, 11, 12]. Therefore, we selected, as an extreme yet realistic condition, an effective exposure of 70% of the daylight duration, capped at a maximum of 8 h and centered around solar noon (e.g., an 8‐h exposure would correspond to 8:00 AM to 4:00 PM).

Anatomical orientation factor

Human skin is not a horizontal sensor. Over the course of real‐life exposure [13], the orientation of the surface continuously varies due to posture and movement. When these varying orientations are integrated over time, the cumulative projection of solar radiation corresponds to about 70% of the maximum theoretical dose, which would only occur if the surface remained permanently perpendicular to the sun.

In other words, 70% represents the time integral of variable incidence angles, not a fraction of time spent at optimal orientation.

The value of 0.7 should be understood as an integrated, conservative estimate reflecting the time‐averaged projection of a moving, curved human surface, rather than a fixed geometric correction. Similar order of magnitude reductions in effective dose have been discussed in exposure assessment studies [7, 14].

Combined exposure coefficient

The ERD integrates exposure duration and orientation:

$${\mathrm{ERD}}_{\mathrm{UVB}}=\mathrm{SED}\mathrm{per}\mathrm{day}\times \text{orientation factor}\times \text{exposure fraction}$$

$${\mathrm{ERD}}_{\mathrm{UVA}}=\mathrm{UVA}\mathrm{per}\mathrm{day}\times \text{orientation factor}\times \text{exposure fraction}$$

This maintains the elegance of Diffey's model while converting irradiance into dose.

Note that the fraction of daylight effectively spent outdoors (0.7) and the anatomical orientation factor (0.7) act multiplicatively. The combined effect (0.7 × 0.7 = 0.49) reduces the ambient horizontal dose by about one‐half, which lies well within the 20–60% range of personal‐to‐ambient exposure ratios reported in literature [15].

Sunscreen application amount and necessary UV protection

Although ISO and FDA standards specify an application density of 2 mg/cm² for in vivo SPF determination, this quantity does not reflect consumer behaviour [16, 17]. It represents the minimum amount required to ensure repeatability and reproducibility of in vivo testing [18]. Numerous studies show that consumers typically apply between 0.5 and 1.5 mg/cm². In this study, an application density of 0.8 mg/cm² was considered realistic.

Over the range of sunscreen application amounts typically observed in consumers (approximately 0.5–1.5 mg/cm²), the decrease in SPF can be reasonably approximated as quasilinear [18, 19].

Therefore, a corrective factor of 2.5 was applied:

$$\text{Application factor}=\frac{2}{0.8}=2.5$$

In other words, if the application amount had been 1 mg/cm², the correction factor would have been 2.

The necessary UV protections are then calculated as:

$$\text{Necessary}\mathrm{UVB}{{{\text{protection}}_{(\mathrm{J}/\mathrm{m}}}^2}_{)}={\mathrm{ERD}}_{\mathrm{UVB}}\times 2.5$$

$$\text{Necessary}\mathrm{UVA}{{{\text{protection}}_{(\mathrm{J}/\mathrm{cm}}}^2}_{)}={\mathrm{ERD}}_{\mathrm{UVA}}\times 2.5$$

Since these protection values expressed in joules are not commonly used, we converted them into more conventional units such as SPF and PPD, using the example of a fair‐skinned individual (Fitzpatrick I–II) having a MED of 210 J/m² and a PPD of 15 J/cm².

As 1 SED = 100 J/m², the necessary UVB protection expressed as SPF is therefore:

$$\text{Necessary}\mathrm{UVB}{\text{protection}}_{\left(\mathrm{SPF}\right)}={\mathrm{ERD}}_{\mathrm{UVB}}\times \left(2.5/2.1\right)$$

And the necessary UVA protection expressed as PPD is:

$$\text{Necessary}\mathrm{UVA}{\text{protection}}_{\left(\mathrm{PPD}\right)}={\mathrm{ERD}}_{\mathrm{UVA}}\times \left(2.5/15\right)$$

METHODOLOGY

Ambient spectral irradiance

Although a cosine‐shaped dependence on solar zenith angle would more closely reflect the physical behaviour of instantaneous irradiance, the Gaussian diurnal distribution implemented in the original Diffey spreadsheet was retained for consistency with the parent model. Accordingly, ambient spectral irradiance was calculated using the same formalism as in Diffey's approach, including the Gaussian representation of the daylight distribution.

This approximation is known to slightly underestimate late‐afternoon exposure and overestimate around noon, but the 24 h integrated dose remains within ±10%, which is acceptable for consumer‐protection modelling and consistent with Diffey's original justification.

For simplicity and to capture seasonal extremes, daily UV doses were computed for the 21st day of each month (close to solstice and equinox positions). While calculations can be performed for all days of the year, this approach provides a representative monthly value without altering seasonal trends.

ERD (UVB‐ and UVA‐weighted)

Using CIE erythema and PPD action spectra [20, 21], daily doses were computed for each day of the year and for latitudes 0–60° N.

Output

For each month × latitude pair (computed for the 21st day of each month), we generated:

• daily standard erythemal dose: 1 SED = 100 J/m².

• UVA‐effective dose (J/cm²).

• Effective UVB (J/m²) and UVA (J/cm²) received dose.

• Necessary UVB and UVA protection.

RESULTS

Seasonal and latitudinal variation in daily dose

Daily standard erythemal doses (using the erythemal action spectrum) and UVA doses (using the PPD action spectrum) were first calculated for a horizontal surface under clear‐sky conditions for latitudes between 0° and 60° N. Higher latitudes were not included to avoid confounding effects related to extreme daylength, snow albedo and very low solar elevation (Figure 1).

FIGURE 1.

Daily standard erythemal dose in J/m² (SED) under clear sky conditions as a function of latitude and month. Values represent the 21st day of each month as a proxy for monthly conditions.

Please note that daily SED and daily UVA values represent the 21st day of each month as a proxy for monthly conditions.

This reproduces known climatological trends but quantifies them in units relevant to protection.

• Tropical climate: 30–70 SED/day, depending on month.

• Mediterranean climate: 4–60 SED/day, depending on month.

• Northern Europe climate: 0–40 SED/day, depending on month.

These results are consistent with solar geometry expectations and measurements [22] as they are clearly correlated with both season and latitude. The summer solstice occurs on June 21 in the Northern Hemisphere, so it is expected that the maximum values appear during this month, while the minimum values occur around the winter solstice in December. The distribution follows a Gaussian pattern, peaking near June 21 [7, 23].

It is also noteworthy that the highest doses are not observed at the equator but rather near the 22nd and 24th parallels. This can be explained by the Earth's axial tilt of approximately 23°, which results in the sun reaching a 90° zenith angle around the 23rd parallel at noon on June 21 [23, 24].

The same principle was applied for the UVA. The same conclusions regarding the maximum values and the Gaussian distribution of intensities can be drawn in this case. However, there is a notable difference compared with the previous calculation: UVA doses continue to increase between April and September, even at higher latitudes (Figure 2). This phenomenon is explained by the fact that day length at these latitudes increases much more than at lower latitudes (Figure 3) [24, 25].

FIGURE 2.

Daily UVA weighted dose in J/cm² under clear sky conditions as a function of latitude and month. Values represent the 21st day of each month as a proxy for monthly conditions.

FIGURE 3.

Day length values correspond to the 21st day of each month.

The clear‐sky spectral irradiance used in this study, based on the Diffey formalism, provides spectral shapes and UVA/UVB ratios fully aligned with reference daylight spectra. Christiaens et al. [26] reported a mean UVA/UVB ratio of 27.3 for non‐extreme conditions, with values consistently between 27.2 and 27.5 across seasons. In our model, instantaneous midday Diffey spectra at mid‐latitudes reproduce ratios in the 25–35 range, while daily‐integrated ratios are lower (around 20). This confirms that the spectral assumptions of our Extended Diffey‐based approach are fully compatible with the Standard Ultraviolet Daylight defined by Christiaens and colleagues.

As shown in Figure 3, between the equator and the tropics, day length varies very little (approximately ±2 h, with a maximum of 13–14 h), whereas at higher latitudes, the increase in daylight duration amplifies cumulative UVA exposure despite lower solar elevation [23, 27].

At higher latitudes, particularly above the 50th parallel north, these fluctuations become extreme, ranging from approximately 6 h to a maximum of 18.5 h of daylight. Since UVA radiation penetrates slightly easier in the atmosphere than UVB and considering that day length can reach up to 18 h in high latitudes during the months surrounding June, it is not surprising that the cumulative UVA dose continues to increase despite the rise in latitude [7, 25, 27].

In the UVB part, the seasonal variation in dose between maximum and minimum values is also strongly latitude dependent as seen in Table 1. At low latitudes (0°–15°), the difference remains minimal (ratio of 3) due to the presence of a single, prolonged summer season. Conversely, as latitude increases, the amplitude of these variations grows markedly (ratio of 14 for Mediterranean latitudes), reflecting pronounced seasonal UV exposure gradients and becomes extreme beyond 60°, where pronounced seasonal photoperiod contrasts induce significant variability in erythemal and pigmentary dose distribution.

TABLE 1.

Seasonal variation of the UVB doses depending on the latitude calculated for the 21st of June and December.

Again, even if these values correspond to the 21st day of each month, they capture seasonal extremes without altering overall trends.

Modelling effective UVB and UVA received doses

Using the calculated daily UV doses, effective UVB and UVA received doses were derived for each latitude and month by:

1. Calculate the maximum SED and UVA doses that can reach a flat surface during an effective exposure of 70% of the daylight duration, capped at a maximum of 8 h and centered around solar noon (8:00 AM to 4:00 PM) on a given sunny day at the given latitude at an incidence angle of 90°.

2. Correct the results by the anatomical orientation factor (a point on human skin is not continuously oriented at a 90° incidence angle for 8 h)

$${\mathrm{ERD}}_{\mathrm{UVB}}=\mathrm{SED}\mathrm{per}\mathrm{day}\times \text{orientation factor}\times \text{exposure fraction}$$

$${\mathrm{ERD}}_{\mathrm{UVA}}=\mathrm{UVA}\mathrm{per}\mathrm{day}\times \text{orientation factor}\times \text{exposure fraction}$$

Despite these mathematical adjustments, which account for geometric and behavioural exposure patterns, the seasonal and latitudinal trends remain unchanged. We still observe the Gaussian pattern of intensity driven by seasonality and latitude [7, 25].

Figures 4 and 5 provide a detailed representation of effective ultraviolet doses received by individuals as a function of geographic location.

FIGURE 4.

Effective SED received doses in J/m² for the 21st day of each month and latitude.

FIGURE 5.

Effective UVA received doses in J/cm² for the 21st day of each month and latitude.

These tables can serve as a basis for various calculations related to UV exposure, particularly for estimating the level of sun protection required.

For example, by slightly adjusting ERD calculations and introducing a correction factor that reflects the actual amount of sunscreen applied by consumers, it becomes possible to estimate the necessary UVB and UVA protection for different latitudes.

Consumers typically apply an average amount of 0.8 mg/cm² rather than 2 mg/cm² used in standardized SPF testing. To account for this difference, values in Figures 4 and 5 were multiplied by a factor of 2.5, yielding approximate protection requirements for UVA (in J/cm²) and UVB (in SED) as a function of latitude and month. Because these units are rarely used to express sun protection, we converted them into protection factors based on a phototype I‐II reference (Figures 6 and 7), assuming 1 MED = 210 J/m² and 1 PPD = 15 J/cm². Accordingly, UVA doses were divided by 15 and UVB doses by 2.1 (1 SED = 100 J/m²). For a phototype II (MED 210 J/m²), this corresponds to 2.1 MED.

FIGURE 6.

Necessary UVB protection values derived from ERD for the 21st day of each month.

FIGURE 7.

Necessary UVA protection values derived from ERD for the 21st day of each month.

Note that thanks to the values generated in Figures 4 and 5, anyone can calculate ERD according to phototype (by adjusting MED and PPD values) and/or the application factor.

Based on these calculations, it is noteworthy that, under ideal conditions where the consumer does not compromise the sunscreen's protection, and under the idealized assumptions of our model (clear sky, high but realistic outdoor exposure, average application amount ~ 0.8 mg/cm² and a time‐averaged protection factor), an UVB protection greater than 40 (SPF 50 in Europe) would protect the majority of fair‐skinned individuals (Fitzpatrick I–II) across all latitudes and throughout the year [23, 27].

Applying the same reasoning to UVA protection, a sunscreen with a PPD value above 10 would protect most consumers (phototype II or higher) for all latitudes and months. When comparing this model to SPF and PPD values of sunscreens available on the European market, it becomes possible to assess whether UVA protection is sufficient.

For example, if the modelled UVB protection based on terrestrial UV exposure and correction factors (application amount, solar incidence angle, exposure duration) yields a value of 33 for a latitude of 40° in June for a phototype II (necessary UVB Protection, Figure 6), this consumer should use a higher SPF, namely SPF 50, according to European Commission recommendations [28]. This SPF 50 product, with a UVA protection factor of 16.6, would adequately protect against the modelled UVA dose. Indeed, the corresponding UVA requirement from the model is 10, well below the 16.6 provided by the SPF 50 product, ensuring complete protection.

It is also interesting to note that, according to these calculations, the maximum dose received in the worst‐case scenario, namely in June under tropical conditions, is approximately 41 MED and 10 PPD. An SPF 50 product, with a corresponding UVA protection of one third (≤16.6), should therefore provide adequate protection for consumers. Of course, it is possible to identify scenarios that are less favourable in terms of consumer protection.

The behavioural correction applied to the labelled SPF accounts only for the average quantity of product applied (0.8 mg/cm²). Additional losses arise from incomplete and inhomogeneous spreading, film break‐up and mechanical removal, which may account for a substantial fraction of the discrepancy between labelled and effective protection, as shown by Heerfordt et al. [29], Diffey [30] and Pissavini [31]. Representing these aspects would require product‐specific parameters, and they are therefore not explicitly included here but they could be incorporated as an additional multiplicative factor when modelling a given formulation.

However, it should also be recalled that SPF 50+ products (e.g., SPF 60) exist with UVA protection values above 20.

Furthermore, sun protection does not depend solely on the amount of UV filters incorporated into the formulation, but also, and more importantly, on the amount applied and the homogeneity of spreading [31, 32, 33, 34, 35, 36]. In other words, the organoleptic properties of sunscreen formulations are just as important as their filter content. As SPF increases, sensory performance generally decreases. Consequently, increasing SPF values beyond SPF 50+ offers no real additional benefit, neither in view of the doses received nor, as previously demonstrated, from a formulation standpoint.

Finally, it should be noted that the results were generated for latitudes between 0 and 60° N under clear‐sky conditions. Snow albedo, cloud cover, aerosols and altitude can materially affect the received doses.

We did not explicitly model the time‐dependent decrease in protection due to wash‐off or photodegradation, nor the effects of re‐application or the effect of cumulative SED. Exploring these dynamics may slightly alter the precise quantitative values of ‘necessary SPF’, but is beyond the scope of the present study.

All protection factor calculations were performed using fixed MED and PPD values (210 J/m² and 15 J/cm², respectively), corresponding on average to a phototype I‐II. However, MED and pigmentary response thresholds vary with phototype. Therefore, the calculations proposed here must remain phototype‐specific. For example, phototypes IV and higher, which have different MED but with small differences in PPD values, would likely lead to quantitatively different results, although the overall conclusions would remain similar to those obtained for lighter phototypes. Future study could extend this approach to other phototypes and environmental conditions to further refine the quantitative estimates.

CONCLUSION

This framework enables visualization of UVB and UVA doses received by an individual across all latitudes from 0° to 60° North for every month of the year. These tables can therefore be used to model various exposure scenarios by applying correction factors. Readers may adjust or add such factors as needed. In this study, we have already proposed two factors to calculate realistic effective doses and a third factor to estimate UVB and UVA doses for an individual with phototype II.

This final step demonstrates the adequacy of current market offerings: products with SPF up to 60 and UVA protection up to 20 already provide strong coverage under realistic conditions. It also shows that SPF values above 60 are unlikely to provide proportionally greater benefit under the realistic exposure scenarios modelled here. Beyond a certain point, increasing SPF yields diminishing returns while compromising sensory acceptability, a factor emphasized in previous publications as essential for effective photoprotection [31, 32, 33, 34, 35, 36].

Importantly, this analysis deliberately avoids claiming universal protection: extended summer photoperiods, product reapplication, water or sweat resistance, towelling or friction can drive different cumulative UV doses, which may require refined approaches. Rather, we emphasize that context matters for everyone; latitude, season and individual characteristics collectively shape protection needs. We therefore open the door to refined UV weighting and context‐specific guidance, to be detailed in the sequel paper. At the same time, current SPF values and EU recommendations remain rational for a wide range of situations, considering both terrestrial UV doses and realistic consumer habits.

Overall, this study emphasizes that photoprotection should be interpreted through the lens of ERD, integrating solar geometry, human behaviour and product use.

Effective photoprotection is not determined by maximal nominal indices alone, but by their ability to control the effective UV dose under realistic exposure conditions.

CONFLICT OF INTEREST STATEMENT

The author declares no conflict of interest.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.