Low Solar Elevation Research Articles

Visualization of UV exposure of the human body based on data from a scanning UV-measuring system.

In general, measurements of UV radition are related to horizontal surfaces, as in the case of the internationally standardized and applied UV index, for example. In order to obtain more relevant information on UV exposure of humans the new measuring system ASCARATIS (Angle SCAnning RAdiometer for determination of erythemally weighted irradiance on TIlted Surfaces) was developed and built. Three systems of ASCARATIS have been in operation at different locations in Bavaria for 3 years, providing erythemally weighted UV irradiation data for 27 differently inclined surfaces every 2 min. On the basis of these data virtual three-dimensional models of the human body surface consisting of about 20,000 triangles could be created and each of these triangles coloured according to its UV irradiation. This allowed the UV exposure of the human body to be visualized for any kind of body posture and spatial orientation on the basis of real measuring data. The results of the UV measurements on inclined surfaces have shown that measuring UV radiation on horizontal surfaces, as done routinely worldwide, often underestimates the UV exposure of the human skin. Especially at times of the day or year with low solar elevations the UV exposure of parts of the human skin can be many times higher than that of the horizontal surface. Examples of three-dimensional modelling of the human UV irradiation are shown for different times of the day and year, altitudes above sea level, body postures and genders. In these examples the UV "hotspots" can be detected and, among other things, used to inform and educate the public about UV radiation.

Observations of reflectance distribution around sunglint from a coastal ocean platform.

A scanning spectral photometer is deployed on a rigid coastal ocean platform to measure upwelling solar radiances from the sea surface at nine elevation angles spanning 150 degrees of azimuth. Measured radiance distributions at 500 nm wavelength have been compared with traditional model simulations employing the Cox and Munk distribution of wave slopes. The model captures the general features of the observed angular reflectance distributions, but: (a) the observed peak value of sunglint near the specular direction is larger than simulated, except for a very calm sea; the model-measurement differences increase with wind speed and are largest for low solar elevation; (b) the observed sunglint is wider than simulated. In contrast to some previous studies, our results do not show a clear dependence of the mean square sea-surface slope on stability (air-sea temperature difference).

Two‐stream approximations revisited: A new improvement and tests with GCM data

AbstractTwo‐stream approximations (TSAs) are employed for multiple‐scattering calculations in most solar radiation schemes used in atmospheric general‐circulation models (GCMs). Two issues related to TSAs are addressed. First, a simple extension of the two‐stream formalism is presented that allows different parametrizations for different factors, such as water clouds, ice clouds and aerosols. A ‘tuned two‐stream approximation’ that uses this extension is then developed. Second, the accuracy of the tuned TSA and selected other TSAs is quantified using a global GCM‐generated dataset (nearly 25000 atmospheric columns with clouds, aerosols and absorbing gases).In the tests with the GCM dataset, the Practical Improved Flux Method (PIFM) provided slightly better results than the delta‐Eddington approximation. Two recent modifications of delta‐Eddington yielded no, or only slight, improvement. For PIFM, the root‐mean‐square (r.m.s.) errors in top‐of‐the‐atmosphere (TOA) and surface net short‐wave fluxes and in total column absorption were 2.22, 3.09 and 2.98 W m−2, as compared with delta‐16‐stream discrete‐ordinate method results. For the tuned TSA, which was developed using the present dataset, the TOA and surface r.m.s. errors were 33% and 48% smaller than for PIFM, with substantially larger improvements at low solar elevations. The robustness of these results is tested in several sensitivity experiments in which the properties of the GCM dataset are modified systematically.While typical delta‐TSAs performed, overall, fairly well, they all had a negative bias in atmospheric absorption and a positive bias in surface net short‐wave flux of ≈1 W m−2 in the global mean (including, also, night grid points). Furthermore, the absorption errors depended substantially on solar elevation. The reasons for these problems are discussed.The tests with GCM data also confirmed that the delta‐four‐stream method is considerably more accurate than the TSAs, especially so for atmospheric absorption. Copyright © 2002 Royal Meteorological Society

Spatial and temporal variations of Venus haze properties obtained from Pioneer Venus Orbiter polarimetry

The spatial and temporal variations of the polarization of light scattered by Venus, as observed by the Pioneer Venus Orbiter between 1978 and 1990, is analyzed in terms of spatial and temporal variations of Venus upper haze properties. Special attention is given to choosing maps with sufficiently accurate geometrical information. This selection involves multiple‐scattering calculations for a number of reasonable models for the Venus atmosphere. For a set of observations in four wavelength bands of the Venus disk as a whole or a particular region thereof, observed values of Stokes parameters I and Q were used to obtain the average observed relative Stokes parameter qobs. Subsequently, the haze particle column density and the cloud top pressure were iteratively adjusted until values of qobs at two wavelengths were reproduced. For the planet as a whole, it was found that the haze particle column density decreased gradually during the Pioneer Venus Orbiter mission, whereas little significant temporal variability was found for the cloud top pressure. Similar long‐term behavior was derived for selected regions of Venus. Regions at higher latitudes exhibited higher values of both the haze particle column density and the cloud top pressure than regions at lower latitudes. Something similar holds for regions with low solar elevations as compared to regions in which the Sun was close to the zenith.

Characteristics of UV-B Irradiance at Syowa Station, Antarctica: Analyses of the Measurements and Comparison with Numerical Simulations

The characteristics of UV-B (λ= 290-315 nm) irradiance at Syowa Station (69°00’S, 39°35’E) in Antarctica were investigated by the numerical simulations with the multiple scattering radiative transfer model for the atmosphere-snow system, and the analyses of measured UV-B irradiance in 1996. By the numerical simulations a study has been made on the sensitivity of UV-B irradiance to the snow grain size, surface condition, cloud cover, Antarctic aerosol, solar zenith angle and total ozone amount. On the snow surface it was found that the effect of multiple reflection between atmosphere (or cloud) and snow surface of high albedo is important for the characteristics of UV-B irradiance. By the analyses of measured UV-B irradiance it was found that the occurrence of the maximum UV-B irradiance was not in December at maximum solar elevation, but in November when the low total ozone amount was observed. This shows that the time in a year reaching the minimum total ozone amount is important for determining the annual trend of UV-B irradiance. It was also found that the values of UV-B irradiance would be enhanced by the occurrence of cloud for the cloud amount of 9/10-10¯/10 due to the effect of multiple reflection between cloud layers and snow surface. As a whole, theoretically calculated values of spectral and spectrally integrated UV-B irradiance agree well with the measured ones, however, at the low solar elevation or low energy level of UV-B irradiance, the measured values of UV-B irradiance were larger than the calculated ones due to the less sensitivity of the instrument to the weak irradiance.

Aircraft observations of surface–atmosphere exchange during and after snow melt for different arctic environments: MAGS 1999

AbstractThe arctic environment, and in particular the Mackenzie Basin, displays a very dynamic interrelationship between the atmosphere and the surface for the different ecosystems represented. The Canadian Twin Otter research aircraft flew a total of 24 grid and long regional transects, over tundra, forest and delta ecosystems, during the period of snow melt (late May–early June) and early summer (early July) as part of the 1999 Mackenzie Area GEWEX (Global Energy and Water Cycle Experiment) Study (MAGS) field campaign. Observations over tundra showed a sharp rise in the sensible heat flux at the onset of melt, reaching a maximum at the end of the melting period similar to those observed in early summer. The latent heat flux showed a more gradual rise from snowmelt to early summer with a Bowen ratio (sensible heat/latent heat) of two during melt. The forested system demonstrated a similar gradual rise in the latent heat flux, whereas the sensible heat flux was already high with Bowen ratios reaching three at the start of the observation period in late May. The gradual rise in latent heat flux can be tied to gradual thawing of the root zone and the onset of photosynthesis activity. The relatively low solar elevation angle and earlier start of snow melt along the regional transect may account for the much larger sensible heat flux. An analysis of the turbulent coherent structures indicates that the spatial density of structures for both latent heat and sensible heat increases strongly with season, from snow melt into the early summer conditions. This has implications for sampling criteria and optimum flux averaging period.There are distinct differences in energy partitioning between the various arctic ecosystems. At the beginning of the observation period, almost all the net radiation over the delta and tundra regions is utilized in non‐turbulent form, whereas the forested areas use less than 50%. Model simulations of the ground heat flux showed observed diurnal imbalances and suggest that the magnitudes depend on the position of the permafrost table and may partially account for the large non‐turbulent energy. Preliminary results from the Canadian MC2 model using the MAGS aircraft data for diagnosis and validation have identified sensitive model components that may merit further investigation. The findings from this study will help to fill gaps in our knowledge about surface–atmosphere interactions in arctic environments, particularly during snow melt, and broadens our contemporary view of evapotranspiration dynamics of wet surfaces. Copyright © 2001 John Wiley & Sons, Ltd.

Polarization patterns of the summer sky and its neutral points measured by full–sky imaging polarimetry in Finnish Lapland north of the Arctic Circle

Using 180° field–of–view (full–sky) imaging polarimetry, the patterns of the degree and angle of polarization of the entire summer sky were measured on 25 June 1999 at a location north of the Arctic Circle in Finnish Lapland as a function of the angular solar zenith distance. A detailed description of the used full–sky imaging polarimeter and its calibration is given. A series of the degree and angle of polarization pattern of the full sky is presented in the form of high–resolution circular maps measured in the blue (450 nm) spectral range as a function of the solar zenith distance. Graphs of the spectral dependence of the degree and angle of polarization of skylight at 90° from the Sun along the antisolar meridian are shown. The celestial regions of negative polarization and the consequence of the existence of this anomalous polarization, the neutral points, are visualized. The measured values of the angular zenith distance of the Arago and Babinet neutral points are presented as a function of the zenith distance of the Sun for the red (650 nm), green (550 nm) and blue (450 nm) ranges of the spectrum. The major aim of this work is to give a clear and comprehensive picture, with the help of full–sky imaging polarimetry, of what is going on in the entire polarized skydome. We demonstrate how variable the degree of polarization of skylight and the position of the neutral points can be within 24 h on a sunny, almost cloudless, visually clear day.

Energy Balance Of A Sparse Coniferous High-Latitude Forest Under Winter Conditions

Measurements carried out in Northern Finland on radiation and turbulent fluxes over a sparse, sub-arctic boreal forest with snow covered ground were analysed. The measurements represent late winter conditions characterised by low solar elevation angles. During the experiment (12–24 March 1997) day and night were about equally long. At low solar elevation angles the forest shades most of the snow surface. Therefore an important part of the radiation never reaches the snow surface but is absorbed by the forest. The sensible heat flux above the forest was fairly large, reaching more than 100 W m-2. The measurements of sensible heat flux within and above the forest revealed that the sensible heat flux from the snow surface is negligible and the sensible heat flux above the forest stems from warming of the trees. A simple model for the surface energy balance of a sparse forest is presented. The model treats the diffuse and direct shortwave (solar) radiation separately. It introduces a factor that accounts for the shading of the ground at low solar elevation angles, and a parameter that deals with the partial transparency of the forest.

Meteorological pre-processing of incoming solar radiation and heat flux over a sparse boreal forest at a northern site during winter conditions

Measurements from Northern Finland on radiation and turbulent fluxes over a sparse boreal forest with snow-covered ground were analysed. The measurements represent harsh winter conditions characterised by low sun angles. The absorption of incoming solar radiation in clear skies (turbidity) was found to be a strong function of the solar elevation. At low solar elevation angles, commonly used expressions for turbidity did not fit the measurements well. A simple energy balance type met-processor performed well during daytime, but it was not satisfactory during night time. Simplifications and possible improvements in the energy balance model are discussed.

Moisture and Energy Balance Components as Influenced by Row Orientations of Potato During Low Solar Elevation Angles

Moisture and Energy Balance Components as Influenced by Row Orientations of Potato During Low Solar Elevation Angles

Derivation of Cloud Index from Geostationary Satellites and Application to the Production of Solar Irradiance and Daylight Illuminance Data

We investigate in the present paper the relationship between satellite count, global irradiance and other solar and illumination resource components, bringing a particular attention to low solar elevation situations (below 20 °) which are very important in northern latitudes. Our investigation is based on data from two geostationary satellites, METEOSAT and GOES, backed by ground measurements in Switzerland and the northeastern USA. The study of different clear sky normalizations lead to the conclusion that a linear correlation between the global clearness index and the irradiance (like the heliosat method) would be inaccurate for low solar elevations, and therefore for high latitude regions. We developed a model that directly relates an elevation dependent clearness index to the could index. This methodology presents a definite advantage because it can be generalized to address the clearness index of other solar radiation components, besides global irradiance, such as direct irradiance, diffuse illuminance, etc. The correlations described in this paper were developed on the data from Geneva (in the frame of the EC program “Satellight”) and evaluated on two other independent data sets (Albany, USA and Lausanne, Switzerland). Their precisions, on a hourly basis, are respectively 30%, 40% and 60% for the global, diffuse and beam components) (90,55 and 95 W/m2). The use of independent data for thederivation and the validation of the models shows thatthose can be used in a wide range of locations, even if the applicability has to be assessed for very different climates.

Estimating Solar Irradiance for Crop Modeling Using Daily Air Temperature Data

Crop growth models require solar irradiance as input data, yet there are few places where such data are routinely measured. For locations where measured values are not available, solar irradiance can be estimated using empirical models such as the Bristow–Campbell (B–C) model. This study was conducted to assess the spatial and seasonal accuracy of the B–C model for midcontinental locations in Kansas. A 30‐year data set from Manhattan, KS, was used to calibrate and evaluate unmodified and modified forms of the B–C model. The effect of seasonality was investigated by subdividing the yearly data into two subsets, a high noontime solar elevation angle period, ranging from DOY 121 to 273, and a low noontime elevation angle period comprising the remainder of the year. The B–C model was also evaluated without seasonal division of the year. The calibrated models were then tested against measured solar irradiance values for 10 sites distributed across the state of Kansas. Results indicate that, for the calibration site at Manhattan, irradiance was more accurately estimated using a modified form of the B–C model. For the yearly data, root mean square error (RMSE) was 3.9 MJ m−2 d−1 (25% error), compared with 5.2 MJ m−2 d−1 (24% error) for the high solar elevation angle period and 3.6 MJ m−2 d−1 (32% error) for the low solar elevation angle period. The RMSE for the 10 test sites ranged from 2.0 to 6.2 MJ m−2 d−1; percentage error ranged from 26 to 47%. Neither latitude nor distance from the calibration site significantly affected the accuracy of irradiance estimates at the evaluation sites. Results suggest that the modified B–C model provides reasonably accurate estimates of irradiance at noninstrumented sites and that the model can successfully be used at sites away from the calibration site. Seasonal subdivision of the data adds little to the accuracy of estimates.

An hourly diffuse fraction model with correction for variability and surface albedo

The paper presents an improved version of a previously published model for the diffuse fraction of hourly global irradiance. In addition to hourly solar elevation and clearness index, an hour-to-hour variability index and regional surface albedo are included among the input parameters. Moreover, to prevent excessively high normal incidence beam irradiances at very low solar elevations, the model does not allow a solar elevation dependent maximum beam transmittance to be exceeded. This new model is tuned to 32 years of data from Bergen, Norway. Moreover, a test against independent data from four European stations showed that the model performs better than the models of Erbs et al. (1982), Maxwell (1987)and Perez et al. (1992).

Biases in Shortwave Column Absorption in the Presence of Fractal Clouds

In this paper, the effect of cloud structure on column absorption by water vapor is investigated. Radiative fluxes above and below horizontally inhomogeneous liquid water clouds are computed using an efficient Monte Carlo technique, the independent pixel approximation, and plane-parallel theory. Cloud inhomogeneity is simulated by two related fractal models that use bounded cascades for the horizontal distribution of optical depth. The first (‘‘clumpy’’) model has constant cloud top and base, hence a constant geometrical thickness but varying extinction; the second (‘‘bumpy’’) model has constant extinction and cloud base, hence variable cloud-top and geometrical thickness. The spectral range between 0.9 and 1.0 mm (with strong water vapor absorption and negligible cloud liquid water absorption) is selected for a detailed study, not only of domain-averaged quantities, but also radiation fields. Column-absorption fields are calculated as the difference between the two net fluxes above and below clouds. It is shown that 1) redistribution of cloud liquid water decreases column absorption, that is, plane-parallel absorption is larger than the independent pixel approximation one by 1%‐3%; 2) 3D radiative effects enhance column absorption by about 0.6% for the clumpy model and 2% for the bumpy model, that is, Monte Carlo absorption is larger than independent pixel approximation absorption—this effect is most pronounced for the bumpy cloud model at solar zenith angle 458; and 3) plane-parallel absorption is larger than 3D Monte Carlo absorption for high solar elevations and nearly equal to it for low solar elevations. Thus, for extended clouds of thickness 1‐2 km or less, in an important water vapor absorption band (0.94 mm), the authors do not find a significant enhancement of cloud absorption due to horizontal inhomogeneity.

Effect of cloud inhomogeneities on the solar zenith angle dependence of nadir reflectance

A significant discrepancy has been noted between satellite measurements of shortwave reflectance at nadir and the results of plane‐parallel model calculations: For moderate to large solar zenith angles, observed nadir reflectances increase with solar zenith angle, whereas plane‐parallel values decrease. Consequently, cloud optical depths retrieved using one‐dimensional (1‐D) theory have a bias which increases systematically with solar zenith angle. Using Monte Carlo model simulations of photon transport through stochastic, isotropic, scale‐invariant cloud fields with variable cloud top heights and volume extinction coefficients, we show that nadir reflectances from three‐dimensional cloud fields increase with solar zenith angle, consistent with the observations. The difference from the 1‐D case is shown to be explainable by cloudside illumination as well as by the presence of structured (i.e., non‐flat) cloud tops. Cloud sides enhance the amount of incident solar radiation intercepted by cloud, allowing more radiation to be scattered upward in the nadir direction. Structured cloud tops change the slope of illuminated cloud top surfaces, such that nadir reflectance at low solar elevations increases with the slope of the illuminated surface. For simple cloud geometries the two effects make equivalent contributions to the increase in nadir reflectance with solar zenith angle. While this increase is most pronounced for vertically extensive broken cloud fields, it also affects reflectances from overcast cloud fields with inhomogeneous (bumpy) cloud tops. Thus the observed solar zenith angle bias in cloud optical depth for the general cloud scene likely also occurs for extensive overcast cloud fields. Internal inhomogeneities due to small‐scale liquid water content variations within clouds are shown to cause no changes at low Sun and only slight decreases in nadir reflectance for high solar elevations.

Scattered ultraviolet radiation in the upper stratosphere 1: Observations

Observations of scattered solar near‐UV radiation in the upper stratosphere were obtained during a 1983 balloon flight at 31°N latitude. The instrument design was patterned after the UV ozone spectrometer on board the Solar Mesosphere Explorer satellite. The measurements cover the wavelength range 175 to 325 nm at a spectral resolution of 1.5 nm, spanning a range of viewing elevation angles from −3.3° to 14.6° with a field of view 1.65×10−5 sr. Observations began at midday near 46 km and continued to approximately 1700 LT near an altitude of 40 km. Calibrated radiances reported here are determined from the raw data using measured sensitivity characteristics of the spectrometer. A correction to account for the polarized nature of Rayleigh‐scattered radiation is included which employs the measured polarization response of the instrument and assumes single scattering of solar radiation. The single pass spectrometer was susceptible to a small but in some cases important contamination by stray light. This effect is quantified using observations obtained at low altitudes and low solar elevation angles, i.e., large optical paths. The uncertainty in radiances is determined by statistical and systematic errors (efficiency, polarization, and stray light determinations) and ranges typically from 15 to 30%. In contrast to previous observations, scattered radiances decrease rapidly at wavelengths shortward of 210 nm.

Luminous efficacy models and their application for calculation of photosynthetically active radiation

From a spectral radiative transfer model, an algorithm is developed for the conversion of illuminance to different measures of Photosynthetically Active Radiation (PAR) (in W m −2 or in μEm −2s −1). This illuminance to PAR conversion may even be used in combination with a luminous efficacy model and, thus, form a photosynthetic efficacy model. In this work, two luminous efficacy models are chosen, one empirical and one derived from the above radiative transfer model. Observed PAR energy flux and PAR photon flux from seven Nordic stations (56–70°N) and illuminance from one U.S. station (43°N) are, together with observed all-wave solar irradiance from all stations, used for verification. Observed and modelled luminous efficacies agree on the average within 1% at high solar elevation under cloudless sky, while it is indicated that the illuminance, PAR energy flux, and PAR photon flux radiometers are mutually inconsistent by some 6–16%. Even differences in cosine response between radiometer types are apparent at low solar elevation under cloudless sky. In the present climates, the global radiation efficacy is 10–12% higher under an average cloudless atmosphere than it is outside the atmosphere. By introducing an average cloud deck in this cloudless atmosphere, a further efficacy increase, slightly exceeding these 10–12%, is observed. However, observations indicate that the cloud transmittance algorithm used in the radiative transfer model significantly overestimates the global radiation efficacy increase caused by horizontally inhomogeneous cloud decks.

Stability of polar frosts in spherical bowl-shaped craters on the Moon, Mercury, and Mars

Following Svitek ( Martian Water Frost: Control of Global Distribution by Small-Scale Processes, Ph.D Thesis, California Institute of Technology, 1992) , analytic solutions are presented for the effective albedo, the effective emissivity, and the radiative equilibrium temperature in the shadowed portions of a spherical bowl-shaped crater. The model assumes that the surface is a Lambert scatterer with visual albedo and infrared emissivity each independent of wavelength across their respective spectral ranges. Absorption, emission, and multiple scattering from the walls of the crater are treated rigorously to all orders. For airless bodies whose surfaces are in radiative equilibrium, all shadowed portions of any individual crater have the same temperature, whose value depends on four quantities: the insolation (product of the solar constant and the sine of the solar elevation angle), the depth/diameter ratio of the crater, the visual albedo, and the infrared emissivity. As long as the crater is deep enough to have shadows, the lowest temperatures are for the shallowest crater—those with the smallest depth/diameter ratio. The model is applied first to the Moon and Mercury using a depth/diameter ratio of 0.2, which is typical of the lunar highlands according to Pike ( Geophys. Res. Lett. 1, 291–294 (1974) ; in Impact and Explosion Cratering (Roddy et al., Eds.), pp. 489–509, Pergamon, New York, 1977) . For Mercury and the Moon, temperatures in shadows in polar craters are below 102 K, so the sublimation rate of water ice calculated according to the model of Watson et al. ( J. Geophys. Res. 66, 3033-3015 (1961)) is less than 1 cm per byr. The latitudinal extent of the cold zone on the Moon is greater than that on Mercury, although temperatures at the poles of the two planets are similar. The other application is to polar frosts on Mars. Illuminated water frosts in radiative equilibrium grow rougher, because the average temperature of a depression is greater than that of flat ground. Subliming CO 2 frosts, which are always at the same temperature, grow rougher at low solar elevation angles because the heat flux absorbed by a depression is greater than that for a flat surface. At high insolation rates (high Sun near perihelion) the average heat flux to a depression is less than for a flat surface. The latter evaporates faster, which makes the average surface smoother and leads to a high average albedo. This behavior helps explain the fact that the south CO 2 cap, which receives its greatest insolation near perihelion, has a higher effective albedo and therefore can survive the summer, whereas the north CO 2 cap has a lower effective albedo and disappears each year around summer solstice.

Focused sun observations using a Brewer ozone spectrophotometer

Accurate measurements of total ozone at high latitudes during winter have become increasingly important for studies of polar ozone depletion. The direct Sun measurements are the most direct and physically well defined type of measurement of total ozone with Dobson and Brewer spectrophotometers along with the focused Sun measurements. The direct Sun and focused Sun method are influenced by many sources of error at low solar elevations. A large error arises from the addition of radiation scattered from air along the path of the solar beam. A method to correct for this error using the Brewer instrument is described. The method gives reasonably accurate values down to solar elevations of 5° compared to 10°–20° without corrections. This is especially important for high‐latitude stations, where the low solar elevations are a limiting factor for accurate measurements of the total ozone particularly during the winter.

A comparison of ultraviolet radiation measured at an arctic and an alpine site

Ultraviolet radiation contributes relatively little energy to the solar spectrum; however, it is very important because it is biologicaly very active. Measurements were carried out at a high altitude station in Switzerland (Jungfraujoch 3576 m), and at a high altitude station in Alaska (Fairbanks 64.82°N) with identical instrumentation. For all season the UV flux for Jungfraujoch was larger than for Fairbanks. In summer the differences between the stations were less pronounced because the lower solar elevation is compensated by a longer day length. In winter the differences are more severe. For both stations we find an increased relative intensity of the UV (UV/Global) with increasing cloudiness, while the absolute values decreased with increasing cloudiness. This shows that the clouds absorb more in the near IR than in the UV region of the solar spectrum. For Fairbanks, the UV values in spring were substantially higher (mean value 18%) than for identical solar elevations after summer solstice. Cloudiness could not account for this, because we also observed differences for clear sky conditions. A simple model was developed, which took multiple reflections of the highly reflecting snow cover in spring into account, which correctly explained 83% of the observed differences.