the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Irradiance and cloud optical properties from solar photovoltaic systems
James Barry
Stefanie Meilinger
Klaus Pfeilsticker
Anna Herman-Czezuch
Nicola Kimiaie
Christopher Schirrmeister
Rone Yousif
Tina Buchmann
Johannes Grabenstein
Hartwig Deneke
Jonas Witthuhn
Claudia Emde
Felix Gödde
Bernhard Mayer
Leonhard Scheck
Marion Schroedter-Homscheidt
Philipp Hofbauer
Matthias Struck
Abstract. Solar photovoltaic power output is modulated by atmospheric aerosols and clouds and thus contains valuable information on the optical properties of the atmosphere. As a ground-based data source with high spatiotemporal resolution it has great potential to complement other ground-based solar irradiance measurements as well as those of weather models and satellites, thus leading to an improved characterisation of global horizontal irradiance. In this work several algorithms are presented that can retrieve global tilted and horizontal irradiance and atmospheric optical properties from solar photovoltaic data and/or pyranometer measurements. Specifically, the aerosol (cloud) optical depth is inferred during clear sky (completely overcast) conditions. The method is tested on data from two measurement campaigns that took place in Allgäu, Germany in autumn 2018 and summer 2019, and the results are compared with local pyranometer measurements as well as satellite and weather model data. Using power data measured at 1 Hz and averaged to 1 minute resolution, the hourly global horizontal irradiance is extracted with a mean bias error compared to concurrent pyranometer measurements of 11.45 W m−2, averaged over the two campaigns, whereas for the retrieval using coarser 15 minute power data the mean bias error is 16.39 W m−2.
During completely overcast periods the cloud optical depth is extracted from photovoltaic power using a lookup table method based on a one-dimensional radiative transfer simulation, and the results are compared to both satellite retrievals as well as data from the COSMO weather model. Potential applications of this approach for extracting cloud optical properties are discussed, as well as certain limitations, such as the representation of 3D radiative effects that occur under broken cloud conditions. In principle this method could provide an unprecedented amount of ground-based data on both irradiance and optical properties of the atmosphere, as long as the required photovoltaic power data are available and are properly pre-screened to remove unwanted artefacts in the signal. Possible solutions to this problem are discussed in the context of future work.
James Barry et al.
Status: final response (author comments only)
-
RC1: 'Comment on amt-2022-335', Anonymous Referee #1, 12 May 2023
This paper presents a method to calculate the global irradiance at the surface and some atmospheric properties using photovoltaic panels. To achieve this, the inversion algorithm accommodates both a description of the hardware and one of the characteristic state of the atmosphere.
I found the article very well written, clear in exposition and scientifically correct in its premises, logical development and description of results. But at the same time it is by no means easy to digest. The text as a whole is very complex, with constant cross-references between one section and the next, and I am not sure it is easily approached by the average reader. This is due to the need to describe all those components of the experiment that determine, at the end of the chain, the final accuracy of the inversion.
In another situation, with another article, I would have requested the authors to streamline the text and make it easier for a wider audience to understand. In this case, however, it is clear to me that this is not easy, given the experimental complexity of the task the authors are facing.
So, as far as I am concerned, I am asking for minor revisions. These are mostly non-critical changes but should add some additional detail regarding the setup of the simulations and the reach and the description of the results.
General considerations after reading the full text- It is surprising that the mean bias error increases with wider temporal aggregations. Can the authors comment on this?
- I assume that the efficiency of the photovoltaic modules is highest in clear sky conditions and lowest in overcast conditions. If correct, how does the lower efficiency relate to the accuracy of the derived cloud optical light extinction? Is it a linear relationship?
- COD_PV is verified (at least in my eyes it is not a validation) with the values of COSMO and APOLLO_NG. But if I am not mistaken, COSMO and APOLLO_NG are at the same time the source of some critical inversion inputs. In this sense it is more of a verification than a validation. Since there are now established methods for determining even the optical thickness of clouds from ground-based measurements (e.g., www.atmos-meas-tech.net/8/1361/2015/), I wonder if it is not worth trying to find some stations in the area of interest that provide true validation as independent.
Specific comments
Abstract, L 7: it is stated that results on COD and AOD will be presented. While several scatterplots are presented and discussed for COD in the text, I found no mention of any retrieval of AOD. One might guess that the results with COD<2 contain, within them, also those of AOD. If true, then the text of the abstract is misleading, as one would expect a retrieval of atmospheric opacity directly due to aerosols. This would probably explain why, for the generation of LUTs with MYSTIC, there is no additional dimension referring to aerosols.
Introduction, L50 and ff: Can the authors add a sentence or two summurazing the accuracy of the cited methods?P3 L73: what is the reasoning behind the 5 Wm-2 desired threshold?
P5 L139-142: "It can be shown using the diode model [see for instance Sauer (1994); Abe et al. (2020)] that the maximum power point (MPP) current generated by a PV module is linearly dependent on the incident irradiance, and only very weakly dependent on temperature."
Surprising. But still no info on the nature of the MPP-irradiance under overcast skies.Eq.1 and 2: even if obvious, \$tau is not introduced in the text. This also leaves the room open for me to ask how \$tau infleunces Eq.1 and 2. Is \$tau role linear on non-linear?
How about a direct aerosol radiative effect at play? Is this embedded in the simulations based on the OPAC database?
P14 L 330 : AOD = 0.01 and AOD = 1. What is the typical AOD for dust events reaching Central Europe?
P14 Section 2.3.5: I have two questions with respect to the setup of MYSTIC and its range of application.
(1) why wasn't MYSTIC also equipped with an additional dimension of AOD? I would expect that, especially in intermediate situations of broken cloudiness, the presence of aerosols may not only change the intensity of irradiance but also its spectral behavior. The latter is precisely identified by the authors as one of the factors contributing to the inversion uncertainty.
(2) Later on in the text, Table 9 shows the limits of applicability of MYSTIC, but in the main text there is no dicussion on the reasons why exaclty those values are reported.P17 L 392: which overlap scheme for clouds? Can the authors be more specific on this point?
P22 onward, Tables reporting biases: (1) can the authors add to the captions the information on the asterisk purpose, so that the reader does not have to look through the text? (2) Would it be informative to report also these values expressed in %?
Spotted typos
P14 L 329: lookup table
P14 L 341: \citep instead of \cite for Crnivec and Mayer 2019Citation: https://doi.org/10.5194/amt-2022-335-RC1 -
RC2: 'Comment on amt-2022-335', Anonymous Referee #2, 17 May 2023
This is an interesting study that addresses a very relevant topic, namely the use of PV as a surrogate radiation measurement device to acquire a much larger spatial coverage than the current relatively sparse coverage of radiation measurement devices.
In my view, the work is very well done in terms of approach and analysis. My comments merely address the presentation and availability of software and data. These points should be easily addressed hence I have suggested minor revisions.
I have a couple of issues that I would like to point out. First concerning open sciences and the lack thereof in the review process. This maybe more a complaint to AMT than to the authors' although I would also encourage them to embrace open science a bit more.
1. This review process cannot be done properly if the authors provide their data in a closed github repository, for which the anonymous reviewers need to reveal themselves in order to see the software. This goes against my definition of open science in my view, and in my view should be prohibited by the journal. I do not think this level of secrecy is constructive nor necessary.
2. In line with the previous comments. It would be nice if all software is made available open source, because from the paper it is not entirely clear what are datasets and what are tools that the readers can use themselves to reproduce or slightly alter the setups. Since the respository is closed, the reviewers cannot check.
Then concerning the manuscript:
1. I would like the authors to reflect more about the feasibilty of this endeavor for operational use. They have tested this for a well designed campaign / setup, but to me it is still a bit unclear whether the authors, based on their findings, expect this to be a realistic road to operational use. They are very careful in their comments in 588, and I would like to challenge them to be a bit more outspoken.
2. Table 9. I would like the authors to comment a bit more on the restrictions presented in the _limits_ column. Are these very restrictive? How much does this influence potential operational use in the future?
3. I find it a little worrying that the authors base their weather model data on a model that according to their own comments went into retiring in 2021. Why not base the results on ICON? If this ever should be come a method for operational use it makes a lot more sense to compare it against the local state of the art.
4. Eq. 10: what is the origin of this equation and what is the physics behind it?
5. Eq. 12: how sensitive are results to the chosen thresholds?
6. Line 344. I do not understand how a PV setup sees the whole sky. Please clarify.
7. Figure 8: The 15-min averages have an almost twice as large bias compared to the 1-min data. Is this because the added measurements are so much worse?
If these points are addressed, I think the authors provide a nice basis for future work on this topic. I am glad that their results are based on physics.
Citation: https://doi.org/10.5194/amt-2022-335-RC2
James Barry et al.
Data sets
Datasets for "Irradiance and cloud optical properties from solar photovoltaic systems" James Barry, Stefanie Meilinger, Klaus Pfeilsticker, Anna Herman-Czezuch, Nicola Kimiaie, Christopher Schirrmeister, Rone Yousif, Tina Buchmann, Johannes Grabenstein, Hartwig Deneke, Jonas Witthuhn, Claudia Emde, Felix Gödde, Bernhard Mayer, Leonhard Scheck, Marion Schroedter-Homscheidt, Philipp Hofbauer, and Matthias Struck https://zenodo.org/record/7628155#.Y-yFTh_MJD8
James Barry et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 330 | 81 | 8 | 419 | 4 | 2 |
- HTML: 330
- PDF: 81
- XML: 8
- Total: 419
- BibTeX: 4
- EndNote: 2
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1