Development of a Tool (Numerical Model) for Estimating and Forecasting Ultraviolet Surface Solar Radiation ^†

Abstract

Monitoring and accurately forecasting ultraviolet (UV) radiation is of great importance especially due to its adverse effects on human health. In this study, we develop a numerical model to estimate the UV surface solar radiation with the overarching goal of providing a fully automated UV forecasting tool in the region of Epirus, Greece, and especially at the city of Ioannina. The UV surface solar radiation (SSR) is estimated based on detailed radiative transfer (RT) calculations. To ensure their accuracy, we employ the well-established UVSPEC model included in the libRadtran RT routines. LibRadtran provides a variety of options to set up and modify an atmosphere with molecules, aerosol particles, water and ice clouds and a surface as the lower boundary. As a first step, we performed a sensitivity study of the surface solar UV radiation with respect to ozone, precipitable water, aerosol optical properties and surface albedo. Our calculations are performed initially under clear-sky conditions to eliminate the uncertainties induced by clouds. All our calculations are performed spectrally within the UV spectral range, for a specific date and time at Ioannina, Epirus.

1. Introduction

The ultraviolet solar radiation (UV) covers the wavelength range of 180–400 nm and is divided into three bands: UV-A (320–400 nm), UV-B (280–320 nm) and UV-C (180–280 nm). While it presents a small fraction of the total solar radiation that reaches the earth’s surface, it has significant effects, either beneficial or adverse, on both the environment and human health. As the UV radiation travels through the earth’s atmosphere, which acts like a filter by absorbing and scattering, it interacts with several atmospheric components and significant changes occur. The most harmful wavelengths, UV-B and UV-C are absorbed by the atmosphere, especially the UV-C that does not reach the surface. The UV-B solar radiation is partially absorbed by the ozone layer in the stratosphere and, as it is biologically active, is responsible for DNA damage and sunburn. The UV-A solar radiation is not significantly absorbed by the atmosphere and reaches the earth’s atmosphere almost entirely and has effects on aging and indirect DNA damage. Both UV-A and UV-B bands reach the earth’s surface, although they interact differently with the components of the atmosphere. Therefore, a thorough study of these interactions is essential, along with a quantitative evaluation of the contribution and impact of each atmospheric factor.

2. Data, Model and Methodology

The libRadtran model [1,2] is a comprehensive suite of tools designed for radiative transfer calculations in the Earth’s atmosphere. Its main tool, uvspec, enables the computation of radiances, irradiances and actinic fluxes across both the solar and thermal spectral regions. The model calculates both downwelling and upwelling components of shortwave (solar) radiation fluxes at the Earth’s surface, at 121 discrete atmospheric levels, as well as the top of the atmosphere. Uvspec ensures a high spectral resolution, computing fluxes at 351 wavelengths ranging from 180 to 410 nm. One of the key strengths of uvspec lies in its flexible input structure, which allows for both simple configurations and complex user-defined set ups tailored to advanced research applications. In this study, the input data were taken from the Modern-Era Retrospective Analysis for Research and Applications, version 2 (MERRA-2) reanalysis [3,4,5]. The methodological approach begins with the execution of a uvspec model run (base run), under clear-sky conditions, for an indicative specific date (20 June 2023) and hour 12:00:00 (UTC) in the region of Ioannina, Epirus. The aerosol optical depth (AOD), Single Scattering Albedo (SSA) and asymmetry parameter (g) were computed spectrally based on MERRA-2 as in [3,4,5,6] and have been used as input to the model. In addition, the model uses other input data, namely surface albedo, altitude (in kilometers), ozone concentration (in Dobson Units) and precipitable water (in millimeters). Following the baseline model run using realistic values of the input data for the specific day, each of the parameters (excluding altitude) is subsequently perturbed quantitatively by 10%, to comparatively assess the single impact of each factor on the ultraviolet (UV) radiation spectrum. All the above procedures are performed as a function of wavelength (λ), within the spectral range of 180 nm to 400 nm, to extract information specifically relevant to the region of interest, which is the ultraviolet (UV) solar radiation.

3. Results

The computed base run (not presented here) was performed using realistic data corresponding to the examined parameters. More specifically, the following initial input data values were used: surface albedo = 0.05, ozone columnar amount = 322 DU, precipitable water = 21.18 mm, altitude = 500 m and moderate aerosol load consisting mostly of sulfate and organic carbon particles. The total AOD varies between 0.306 (at 250 nm) and 0.225 (at 400 nm), while SSA varies, respectively, between 0.744 and 0.668. Following the base run, we proceeded to the next step, examining the sensitivity of UV surface solar radiation (SSR) to each parameter individually. We performed additional model runs by perturbing each parameter by 10% relative to its initial value while keeping all other parameters constant for clear-sky conditions. More specifically, we increased the aerosol load (by increasing AOD), aerosol absorptivity (by decreasing SSA), precipitable water, total ozone column and surface albedo. It should be noted that the changes in aerosol parameters (AOD, SSA) and surface albedo were applied equally in all wavelengths. All model runs were carried out spectrally, focusing on the region of interest, i.e., UV, ranging from 180 nm to 400 nm. As shown in Figure 1, in the UVB spectral region (280–320 nm), the most important factor is ozone (orange curve) due to the existence of strong absorption Hartley and fewer Huggins bands at these wavelengths. Indeed, an increase in ozone by 10% leads to less UVB at the surface by up to 15 mW/m² nm. On the other hand, in the UVA (320–400 nm), the aerosol SSA has the strongest impact among all considered factors (light blue curve), inducing decreased UVA fluxes by up to 25–30 mW/m² nm. Aerosol SSA also plays an important role in affecting the extreme wavelengths of UVB. AOD and surface albedo are less important for both UVA and UVB, while precipitable water has a negligible effect on UV solar radiation.

Also, the overall effect of drivers of UV radiation was examined by estimating the spectrally averaged percent modifications of UVA (blue bars) and UVB (orange bars) due to the increases in drivers by 10%, as in the previous analysis (Figure 2). Precipitable water is not included in Figure 2 since, as it was already shown, its effect on both UVA and UVB is negligible. Again, ozone is found to be the primary UVB driver, reducing it by up to about 9%, followed by aerosol SSA (about 3% decreases), while SSA is the most important driver of UVA, inducing decreases by up to about 3%, while changes due to ozone are much smaller. As shown in Figure 2, the overall effect of AOD and surface albedo changes is small, not exceeding 0.5% (AOD).

4. Conclusions

In this current study, a detailed sensitivity analysis is conducted for UV solar radiation at a specific midday time and summer date (near solstice) in the region of Ioannina, NW Greece, under clear-sky conditions. The analysis shows that both UV spectral and broadband UVA and UVB fluxes are significantly dependent on ozone and aerosols. Specifically, UVB is primarily driven by changes in ozone amounts, causing changes as high as 15 mW/m² nm or 9%, whereas UVA is mostly dependent on aerosol SSA, i.e., absorptivity, inducing changes up to 30 mW/m² nm or 3%. The remaining factors have a much smaller impact on UV surface solar radiation. Given the stronger erythemal action spectral values of UVB than UVA, the presented results in this study presume a stronger impact of ozone than aerosols on UVI.