Measurement report: Spectral actinometry at SMEAR-Estonia

Systematic spectral measurements of downwelling solar radiation, both of global and diffuse, have been collected in summertime during 8 years in the hemi-boreal zone in south east Estonia near the SMEAR-Estonia research station. The measurements provided information about the variation of spectral and total fluxes of downwelling hemispherical global and diffuse solar radiation in the wavelength range from 300 to 2160 nm with spectral resolution of 3 nm in UV to 16 nm in SWIR spectral 5 regions. Unique data have been collected and quantitative description of the variability of the measured spectra is provided. For the description of the synoptic situation during measurements, instead of cloud cover, the ratio of measured to possible total flux in the spectral range of 320–1800 nm is used. This ratio could be used as the primary meteorological parameter instead of cloud cover which is difficult to measure instrumentally.


10
Measurements of global and diffuse downwelling radiation are carried out in Estonia at Tartu-Tõravere actinometric station of Estonian Weather Service since 1955 (Russak and Kallis, 2003). Both, global and diffuse irradiance are measured with pyranometers which are spectrally integrating sensors. While in the studies of energy fluxes in the atmosphere and at Earth's surface global fluxes are needed, in ecological studies of the vegetation-atmosphere interaction the energy distribution over the radiation spectrum is important. John et al. (2013) studied the allometry of cells and tissues within leaves, and found 15 that future work is needed to consider the possible influence of the environment. Many trait allometries shift substantially due to plasticity across different growing conditions, i.e., different supplies of light, nutrients, and/or water. The study by Solomakhin and Blanke (2010) showed how the changes in the spectrum of incident light due to colored hailnets affect leaf anatomy, vegetative and reproductive growth as well as fruit coloration in apple. Ji et al. (2020) found that solar radiation components photosynthetically active radiation and ultraviolet-B have different associations with leaf nitrogen and phospho-20 rus content. The direct associations, when solar radiation is indicated by spectral components, are greater than the indirect associations. So when predicting the effects of global dimming on ecosystem nutrient fluxes, the roles of direct, diffuse, and spectral components of solar radiation must be distinguished. The study by Moon et al. (2020) concludes that accurate and spectrally resolved canopy radiative transfer models are critically necessary to realistically determine chemical reactions and gas concentrations within plant canopies and in the immediately overlying atmospheric boundary layer.

25
The spectral composition of extraterrestrial solar radiation is monitored on-board of satellites (Harder et al., 2005(Harder et al., , 2010Kopp, 2014). For designing photoelectric solar power equipment some episodic measurements of irradiance spectra at Earth's surface have been carried out (Norton et al., 2015), and model simulations with atmospheric radiative transfer models for standard situations are provided (Renewable Resource Data Center (RReDC), 2003;Gueymard, 2004). Eddy covariance (EC) sites for measuring fluxes of trace gases usually include optical sensors measuring downwelling, and sometimes also upwelling 30 shortwave radiation, however, a review by Balzarolo et al. (2011) revealed that only 5 out of 40 European EC sites involved in the study had hyperspectral radiometers while the majority used multispectral or broadband sensors.  and at least one measurement is acquired with all three spectrometer modules. Integration time is automatically adjusted independently for each spectrometer module so that the maximal value of the recorded raw signal is between 60% and 90% of the full scale value. Maximal usable integration times are limited by dark signal levels and are set to 200 s, 250 s, and 1.5 s for 50 the UV, VNIR, and SWIR modules of the spectrometer, respectively. Dark measurements are made by covering the fore-optics with a mechanical shutter. A shadow disk is used for blocking approximately 6 • in the direction of the Sun during the diffuse measurements. An azimuth-elevation tracking system is used for moving the shadow disk with complete fore-optics tracking the Sun in azimuthal direction. The detailed description of the spectrometer, calibration and metrological processing of data is provided by Kuusk and Kuusk (2018).

Data
The summary of SkySpec measurements at Järvselja is in Table 1    291, 400, 526, 649, and 721, respectively, altogether 2587 spectra. The clear sky spectra were extracted manually, observing that the global flux was stable during several minutes before and after the moment when the sun zenith angle was equal to one 65 from the prescribed set (40 • , 45 • etc).

Global radiation
Mean spectra of global radiation in case of clear sky at various sun zenith angle are plotted in Fig. 3. Standard deviations are marked by errorbars for sun zenith angles of 40 • and 60 • .
70 Kondratyev (1965) suggested an approximate solution of the radiative transfer equation for the calculation of spectral hemispherical solar radiation during cloudless sky: of 2018 was very sunny, thus the peak at Q/Q 0 = 1 is higher and the level of the distribution at Q/Q 0 in the range between 0.1 and 0.5 lower than in other years.

100
Change of cloudiness modifies not only the relative global flux Q/Q 0 but also the distribution of energy in the spectrum of total radiation. With increasing cloudiness the share of spectral flux at some wavelengths increases but at some other wavelength decreases. Fig. 7 shows how change of cloudiness (change of relative global flux Q/Q 0 ) modifies relative share of radiation at various wavelengths in the integral flux.

Diffuse radiation
105 Energetic values of spectral diffuse sky radiation vary in wide range. Here we analyze the relative spectra of diffuse radiation as the ratio to the spectral global flux. The records of global and diffuse spectra are not simultaneous, therefore the spectral global flux at the moment of recording diffuse flux is calculated as the mean value of the last global spectrum before and the first one after the diffuse irradiance measurement stage.

Discussion and conclusions
Systematic spectral measurements of downwelling solar radiation, both global and diffuse, have been collected during 8 years in 120 the hemi-boreal zone in south east Estonia near the SMEAR-Estonia research station. The measurements provided information about the variation of spectral and total fluxes of downwelling hemispherical global and diffuse solar radiation in the wavelength range from 300 to 2160 nm with spectral resolution of 3 nm in UV to 16 nm in SWIR spectral regions. Unique data have been collected and quantitative description of the variability of the measured spectra is provided. For the description of the synoptic situation during measurements the ratio Q/Q 0 in the spectral range of 320-1800 nm is used. We suggest to use this measure mainly visually, and thus may have subjective errors. At the same time the measurement of integrated global radiation with pyranometers is not very complex nor expensive. The number of meteostations where the measurements of downwelling solar radiation with pyranometers are carried out is increasing (Ohmura et al., 1998). Recording of downwelling solar flux is much more simple than the instrumental measurement of cloud cover at a meteorological station.

130
Data availability. Spectra of downwelling global and diffuse solar radiation for the period and spectral range of interest are available at Tartu Observatory, Estonia. Please contact authors for arranging details.