the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Aug 2025
Total solar irradiance using a traceable solar spectroradiometer
Abstract. Accurate, precise and traceable measurements of total and spectral solar irradiance measurements are fundamental for solar energy applications, climate studies, and satellite validation. In this study, we assess the performance and the quality of the data from a commercially available, compact BTS Spectroradiometer system, by comparing its spectrally integrated total solar irradiance (TSI) values with an electric substitution cavity radiometer (PMO2), which is traceable to the World Radiometric Reference (WRR). The resulting ratio between BTS Spectroradiometer system and WRR-traceable TSI is 0.9975 with a standard deviation of 0.0050. Applying a correction factor of (-) 0.34 % to PMO2, accounting for the known offset between WRR and the International system of Units (SI) results in a relative difference between the BTS Spectroradiometer system derived TSI and PMO2 of +0.09 % with a standard deviation of 0.0050 demonstrating good consistency between BTS derived TSI and the cavity radiometer.
This comparison confirms the precision and accuracy of the BTS spectroradiometer system, and its capability to deliver SI traceable TSI from spectrally resolved solar irradiance measurements. Its spectral resolution enables accurate measurements of spectral solar irradiance, which are essential, not only for determining total solar irradiance but also for retrieving key atmospheric gases such as water vapor, ozone, and aerosols, establishing its relevance as a compact instrument for atmospheric and climate research.
- Preprint
(1102 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-4030', Joseph Michalsky, 16 Sep 2025
- AC1: 'Reply on RC1', Dhrona Jaine, 22 Oct 2025
-
RC2: 'Comment on egusphere-2025-4030', Anonymous Referee #2, 05 Oct 2025
General comments:
The manuscript presents observations of direct spectral irradiance taken with a calibrated BTS-spectrometer (covering the 280 nm to 2150 nm wavelength range) in cloud-free conditions over a one-year period at Davos, Switzerland (46.803° N, 9.836° E, 1590 m.a.s.l). The integrated spectral observations were compared to the broadband direct normal irradiance of an WRR-traceable absolute cavity radiometer by extending the spectrum beyond its observed spectral range using radiative transfer calculations (RTM) conducted with the library for radiative transfer (libRadtran) in the 2150 to 5000 nm wavelength range so that approximately 99.96 % of the Total Solar Irradiance (TSI) is considered. Modelled spectral irradiance in the 280 nm to 2150 nm wavelength range was initially validated against the respective observed spectral irradiance from the BTS and was within 0.3 % (for 3442 cloud-free spectra). A sensitivity study was then conducted to quantify the impact of the most relevant atmospheric parameters such as solar zenith angle, precipitable water vapor and Angstrom parameters on the solar spectra in the 2150 to 5000 nm wavelength range. Finally, a machine learning model based on the previously identified atmospheric parameters was used to speed up the calculations in the extended spectral range yielding excellent consistency with the spectra based on RTM calculations. BTS-system with modelled extended spectra and WRR-traceable TSI from the cavity was within 0.1% demonstrating the robustness of the method and its value for the retrieval of atmospheric gases and constituents.
Accurate observations of (spectral) direct solar irradiance is highly relevant for climate monitoring, modelling and PV-related applications and thus this study provides a valuable contribution to the community and is well suited for publication in AMT.
The manuscript is well structured and clearly written. The literature has been carefully selected and cited. Graphics and tables are clear and the captions self-explanatory. The presented methods are robust and valid. I recommend acceptance after minor/technical corrections.
Specific comments:
Section 3. 1
Could you give more details on the input for the radiative transfer model calculations and on how you conducted the calculations (note: libRadtran is not a radiative transfer model but a library for radiative transfer, i.e. a collection of different solvers and band parametrization models (correct that in line 129)). Specifically:
- what kind of wavelength grid/band parametrization model did you use to be consistent with the spectra from the BTS? What about beyond 2150 nm?
- what solver did you use?
- line 173: “synthetic atmosphere”. You may describe this a bit more detailed. You used the (US-) Standardatmosphere, normalized to the indicated observed atmospheric parameters, I guess?
Section 3.2
- Have you also studied the impact of microphysical aerosol properties (e.g. single scattering albedo (SSA)) on the fractional contribution R as the aerosol type may also change the solar spectra substantially?
Minor comments
In general: Define abbreviations once at their first occurrence and then use them throughout the manuscript (e.g.: Lines 55, 73: World Radiometric Reference is defined in line 9; line 73 World Standard Group is defined in line 56; line 77: the Bic-Tec Sensor (BTS) is defined in line 51 (or should be defined in line 7).
Line 31: Kopp et al.
Line 144: May define sza, wv, α, β here (instead in section 3.2). Instead of water content, I would use precipitable water, pw, or integrated water vapor, iwv, as you used a column measure in the sensitivity analysis. α and β is to my knowledge termed as Angstrom (or Ångström) α and Angstrom (Ångström) β, respectively
Line 189: May add a reference for the OAT method
Line 216: delete “However”
Citation: https://doi.org/10.5194/egusphere-2025-4030-RC2 - AC2: 'Reply on RC2', Dhrona Jaine, 22 Oct 2025
Data sets
Total solar irradiance using a traceable solar spectroradiometer- Datasets and python scripts Dhrona Jaine Kochuparambil and Julian Gröbner https://doi.org/10.5281/zenodo.16910121
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 144 | 40 | 18 | 202 | 6 | 8 |
- HTML: 144
- PDF: 40
- XML: 18
- Total: 202
- BibTeX: 6
- EndNote: 8
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Comments on “Total solar irradiance using a traceable solar spectroradiometer” by Jaine, Groebner, and Finsterle https://doi.org/10.5194/egusphere-2025-4030
This paper compares direct normal solar radiation measured with a WRR-traceable absolute cavity radiometer to spectrally integrated solar spectra between the wavelengths 280 to 5000 nm, which includes at least 99.5% of the spectrum for their site in Davos, Switzerland. The range from 280 to 2150 nm is measured with a Bi-Tec Sensor (BTS) spectroradiometer and beyond that (2150 – 5000 nm) is modeled using model inputs of aerosol optical depth and wavelength dependence, ozone, water vapor, carbon dioxide, and the instantaneous atmospheric pressure and solar zenith angle. After correction for scatter light in the cavity measurements the agreement between the two quantities is within 0.1%. As the authors state, this lends a great deal of validity to the spectral measurements that can be used to characterize trace constituents in the atmospheric column.
I accept the paper after a few items are clarified in the manuscript.
In Section 2.1:
What is the FOV of the BTS spectroradiometer? Does it match the 5 deg FOV of the PM02?
How often is the BTS calibrated? If only initially, how do you guarantee that it is stable?
In Section 3:
The caption for Fig. 2 is incorrect in that the grey area does not represent 90 % of the TSI.
Perhaps an insert that blows up the 4000 - 5000 nm region in Fig. 2 would clarify the points made in lines 133 and 134.
I did not understand the necessity of a machine learning approach since one needs the model inputs (eqn. 4) to estimate the 2150 – 5000 nm contribution for machine learning or the model runs; why not just run the model to calculate the contribution?
In Section 4:
Fig. 6 is difficult to examine. Perhaps a blow up of just one vertical grouping would more clearly show the degree of agreement. I think you could eliminate the left part (a) of this figure.
Other:
Line 44 “gases constituents” “gases”
Look for “could” that should be changed to “cloud” in at least two places. Lines 156 and 178.
In Fig. 2 caption “grey vertical” “vertical”