the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 Mar 2025
HAPI2LIBIS (v1.0): A new tool for flexible high resolution radiative transfer computations with libRadtran (version 2.0.5)
Abstract. Atmospheric radiative transfer (RT) models are useful tools to increase understanding of the physical interactions and processes occurring in the atmosphere and surface. In the category of free and open-source models, libRadtran is a widely used and versatile package. However, running high-resolution calculations with libRadtran is often tedious since libRadtran does not include all the required information to run line-by-line executions (i.e., resolving individual spectral lines of gases) in a flexible way under specific atmospheric conditions, meaning that external software is required. This poses a problem for a user since generating necessary files for libRadtran requires familiarity with the topic of molecular spectroscopy in addition to knowing how to use the external software which may not be tailored for producing libRadtran compatible files. In this paper, we present HAPI2LIBIS, a compact software tool intended to be used in close connection with libRadtran to enable easy high-resolution calculations.
- Preprint
(1284 KB) - Metadata XML
- BibTeX
- EndNote
Status: closed
-
RC1: 'Comment on egusphere-2025-220', Anonymous Referee #1, 13 Apr 2025
This manuscript introduces a software tool HAPI2LIBIS Â to compute absorption cross section of atmospheric gases based on HITRAN dataset, and can provide input file of absorption cross section for libRadtran radiative transfer computations. One of the features is the irregular interpolation method to interpolate absorption cross section at different temperature, pressure and gas concentration grids. Some comments are as follows.
- Line-by-line Radiative Transfer Model (LBLRTM, http://rtweb.aer.com/lblrtm.html) can also compute absorption cross section of gases based on HITRAN dataset, and has been widely used in the atmospheric radiation community. The authors should discuss the differences of the HAPI2LIBIS and LBLRTM in terms of gas absorption cross section calculation, and point out what are the specific strengths of the new-developed HAPI2LIBIS.
- The authors should also discuss more about the computational efficiency aspect of the HAPI2LIBIS if used in radiative transfer computation and remote sensing. Will HAPI2LIBIS be used in any remote sensing retrieval algorithms?
Citation: https://doi.org/10.5194/egusphere-2025-220-RC1 -
AC1: 'Reply on RC1', Antti Kukkurainen, 07 May 2025
Thank you for the perceptive comments! Responses to them are as follows:
- While LBLRTM also computes the gas absorption cross-sections from the HITRAN spectral line data, HAPI2LIBIS has several strengths and differences compared to it. HAPI2LIBIS is simpler and easier to use, with only two Python code files (one of which is the HITRAN API) and a configuration file, to create customized absorption molecular absorption file for libRadtran. In addition, HAPI2LIBIS automatically downloads the latest available molecular line data from the HITRAN database, so manual absorption line updating and configuration is unnecessary. In HAPI2LIBIS, it is possible to select which line broadening functions are utilized, and the computed absorption cross-sections are stored in a dynamic interpolation tables for reuse.
- Almost always retrieval algorithms are based on pre-computed tables of spectral radiances or gas absorption cross-sections, so very rarely are line broadening calculations done on-the-fly, so to say. However, the interpolation from the HAPI2LIBIS tables is rapid, as long as sufficient amount of parameter points are added, so in principle it could be used to replace spectral absorption tables, such as ABSCO [1], within a retrieval algorithm. On the other hand, in some retrievals, such as CH4 profile retrieval algorithm SWIRLAB [2], the gas absorption cross-sections are computed as a part of the retrieval algorithm, because high spectral and temperature-pressure dependency accuracies are needed. In such retrieval algorithms, HAPI2LIBIS could prove to be practical.
We will revise the manuscript by adding a brief introduction to LBLRTM along with a discussion about the differences between it and HAPI2LIBIS. We will also add a paragraph with the focus on discussion about the computational efficiency aspect as per the suggestions.
References:
[1]: https://www.sciencedirect.com/science/article/pii/S0022407320302016
[2]: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2015JD024657
Citation: https://doi.org/10.5194/egusphere-2025-220-AC1
-
RC2: 'Comment on egusphere-2025-220', Anonymous Referee #2, 22 Jun 2025
Atmospheric gas absorption has been a long-lasting task for radiative transfer simulations due to the complex absorption lines. Meanwhile, as a widely applied radiative transfer (RM) model, the libRadtran is lack of a high-resolution LBL-based gas absorption model, which clearly limits its applications. This manuscript introduces the new tool for libRadtran, i.e., HAPI2LIBIS (v1.0), for high resolution gas absorption and RT simulations. The new tool is clearly presented and introduced, and fits the GMD well. Considering the wide applicability of libRadtran, the development of such an LBL-based model/tool becomes highly necessary and meaningful, so I would suggest the paper to be accepted for publication after addressing my following concerns.
1. The Abstract briefly introduces the motivation of the new tool, while missed key techniques and contributions of the HAPI2LIBIS. Thus, the Abstract should be substantially improved to indicate the novelty and techniques of the new tool for better understanding.
2. A traditional LBL model is useful and necessary, while it is computational expensive. Thus, some fast/efficient models have been developed for high spectral resolution simulations, and show great potential for further applications. Gas absorption has been a key part for those fast models. For example, the principle-component method and machine learning techniques have been used for high-resolution gas absorption simulations (see https://doi.org/10.1007/ s00376-023-2293-5 and https://doi.org/10.1007/ s00376-021-0366-x). Would the authors think that those models be possible to be included in LibRadtran in the future to replace LBL models, and some introductions on the fast models are suggested in the manuscript.
3. Due to the oscillation features of the gas absorption, the comparisons in Figure 2 are not that clear. Thus, the differences between HAPI2LIBIS and REPTRAN are suggested to be shown as well.
4. Figure 4 compares LBL, HAPI2LIBIS and REPTRAN results, while O3 absorption is not considered in LBL simulations. Such a comparison becomes unfair, and could be possible to confuse readers. A fair comparison without O3 is suggested to be added or to replace the current results.
5. Again, Figure 5 is not clearly shown due to the absorption feature.
6. How are the continuum absorptions considered in the model? It is not mentioned that much in the manuscript.
7. The manuscript mainly shows results over the near-infrared and solar bands. However, gas absorption and high-resolutions are also widely used for infrared bands. Would some examples over the longwave-infrared bands be presented and discussed as well?
Citation: https://doi.org/10.5194/egusphere-2025-220-RC2 -
AC2: 'Reply on RC2', Antti Kukkurainen, 29 Jun 2025
Thank you for the comments. Since the open discussion has closed, we will provide a brief initial response without going into detail here and address the concrete changes to improve the manuscript at a later time. Our responses are as follows.
- We initially wrote the abstract to be concise and include only the main purpose/functionality of HAPI2LIBIS. However, we now see that the abstract is lacking necessary information about the techniques presented in the manuscript. We will extend the abstract to include these techniques and contributions and rephrase the abstract accordingly.
- Thank you for the paper suggestions. While these models could be of use for libRadtran, it is up to the libRadtran authors to include them there. Some spectral RT speedup methods are already included in libRadtran, such as correlated-k and pseudo-spectral calculation. It must be noted that HAPI2LIBIS only creates the line-by-line input for libRadtran, whereas these methods are distinct approaches for handling spectral RT. Including fast, efficient, and accurate models for gas absorption in HAPI2LIBIS is something that could be added in the future. We will add a mention of fast models and their potential in the manuscript.
- The difference plot between HAPI2LIBIS and REPTRAN will be added to the manuscript.
- Fair comparison is presented in Figure 3. The purpose of Figure 4 is to demonstrate the effect of O3 in the wavelength range 630--635 nm and highlight the capability of HAPI2LIBIS to include new gas information, as in this particular case where the O3 absorption data is missing in HITRAN for this wavelength range. We will rephrase and expand the section where Figure 4 is shown to improve clarity and motivation for this figure.
- Figure 5 will be clarified to make the altitude-dependence of the absorption lines more visible.
- Despite the water vapour continuum absorption being available on HITRAN, we have decided to focus here on the distinct gas absorption lines. Continuum absorption could be included later on as part of HAPI2LIBIS, but for example, the MT_CKD_H2O continuum absorption model is released under a non-free license and would thus not be straightforward to include in HAPI2LIBIS.
- Example image of the optical thickness of different gases in thermal infrared wavelength band will be included along with the appropriate discussion of the results.
Â
Â
Â
Citation: https://doi.org/10.5194/egusphere-2025-220-AC2
-
AC2: 'Reply on RC2', Antti Kukkurainen, 29 Jun 2025
Status: closed
-
RC1: 'Comment on egusphere-2025-220', Anonymous Referee #1, 13 Apr 2025
This manuscript introduces a software tool HAPI2LIBIS Â to compute absorption cross section of atmospheric gases based on HITRAN dataset, and can provide input file of absorption cross section for libRadtran radiative transfer computations. One of the features is the irregular interpolation method to interpolate absorption cross section at different temperature, pressure and gas concentration grids. Some comments are as follows.
- Line-by-line Radiative Transfer Model (LBLRTM, http://rtweb.aer.com/lblrtm.html) can also compute absorption cross section of gases based on HITRAN dataset, and has been widely used in the atmospheric radiation community. The authors should discuss the differences of the HAPI2LIBIS and LBLRTM in terms of gas absorption cross section calculation, and point out what are the specific strengths of the new-developed HAPI2LIBIS.
- The authors should also discuss more about the computational efficiency aspect of the HAPI2LIBIS if used in radiative transfer computation and remote sensing. Will HAPI2LIBIS be used in any remote sensing retrieval algorithms?
Citation: https://doi.org/10.5194/egusphere-2025-220-RC1 -
AC1: 'Reply on RC1', Antti Kukkurainen, 07 May 2025
Thank you for the perceptive comments! Responses to them are as follows:
- While LBLRTM also computes the gas absorption cross-sections from the HITRAN spectral line data, HAPI2LIBIS has several strengths and differences compared to it. HAPI2LIBIS is simpler and easier to use, with only two Python code files (one of which is the HITRAN API) and a configuration file, to create customized absorption molecular absorption file for libRadtran. In addition, HAPI2LIBIS automatically downloads the latest available molecular line data from the HITRAN database, so manual absorption line updating and configuration is unnecessary. In HAPI2LIBIS, it is possible to select which line broadening functions are utilized, and the computed absorption cross-sections are stored in a dynamic interpolation tables for reuse.
- Almost always retrieval algorithms are based on pre-computed tables of spectral radiances or gas absorption cross-sections, so very rarely are line broadening calculations done on-the-fly, so to say. However, the interpolation from the HAPI2LIBIS tables is rapid, as long as sufficient amount of parameter points are added, so in principle it could be used to replace spectral absorption tables, such as ABSCO [1], within a retrieval algorithm. On the other hand, in some retrievals, such as CH4 profile retrieval algorithm SWIRLAB [2], the gas absorption cross-sections are computed as a part of the retrieval algorithm, because high spectral and temperature-pressure dependency accuracies are needed. In such retrieval algorithms, HAPI2LIBIS could prove to be practical.
We will revise the manuscript by adding a brief introduction to LBLRTM along with a discussion about the differences between it and HAPI2LIBIS. We will also add a paragraph with the focus on discussion about the computational efficiency aspect as per the suggestions.
References:
[1]: https://www.sciencedirect.com/science/article/pii/S0022407320302016
[2]: https://agupubs.onlinelibrary.wiley.com/doi/full/10.1002/2015JD024657
Citation: https://doi.org/10.5194/egusphere-2025-220-AC1
-
RC2: 'Comment on egusphere-2025-220', Anonymous Referee #2, 22 Jun 2025
Atmospheric gas absorption has been a long-lasting task for radiative transfer simulations due to the complex absorption lines. Meanwhile, as a widely applied radiative transfer (RM) model, the libRadtran is lack of a high-resolution LBL-based gas absorption model, which clearly limits its applications. This manuscript introduces the new tool for libRadtran, i.e., HAPI2LIBIS (v1.0), for high resolution gas absorption and RT simulations. The new tool is clearly presented and introduced, and fits the GMD well. Considering the wide applicability of libRadtran, the development of such an LBL-based model/tool becomes highly necessary and meaningful, so I would suggest the paper to be accepted for publication after addressing my following concerns.
1. The Abstract briefly introduces the motivation of the new tool, while missed key techniques and contributions of the HAPI2LIBIS. Thus, the Abstract should be substantially improved to indicate the novelty and techniques of the new tool for better understanding.
2. A traditional LBL model is useful and necessary, while it is computational expensive. Thus, some fast/efficient models have been developed for high spectral resolution simulations, and show great potential for further applications. Gas absorption has been a key part for those fast models. For example, the principle-component method and machine learning techniques have been used for high-resolution gas absorption simulations (see https://doi.org/10.1007/ s00376-023-2293-5 and https://doi.org/10.1007/ s00376-021-0366-x). Would the authors think that those models be possible to be included in LibRadtran in the future to replace LBL models, and some introductions on the fast models are suggested in the manuscript.
3. Due to the oscillation features of the gas absorption, the comparisons in Figure 2 are not that clear. Thus, the differences between HAPI2LIBIS and REPTRAN are suggested to be shown as well.
4. Figure 4 compares LBL, HAPI2LIBIS and REPTRAN results, while O3 absorption is not considered in LBL simulations. Such a comparison becomes unfair, and could be possible to confuse readers. A fair comparison without O3 is suggested to be added or to replace the current results.
5. Again, Figure 5 is not clearly shown due to the absorption feature.
6. How are the continuum absorptions considered in the model? It is not mentioned that much in the manuscript.
7. The manuscript mainly shows results over the near-infrared and solar bands. However, gas absorption and high-resolutions are also widely used for infrared bands. Would some examples over the longwave-infrared bands be presented and discussed as well?
Citation: https://doi.org/10.5194/egusphere-2025-220-RC2 -
AC2: 'Reply on RC2', Antti Kukkurainen, 29 Jun 2025
Thank you for the comments. Since the open discussion has closed, we will provide a brief initial response without going into detail here and address the concrete changes to improve the manuscript at a later time. Our responses are as follows.
- We initially wrote the abstract to be concise and include only the main purpose/functionality of HAPI2LIBIS. However, we now see that the abstract is lacking necessary information about the techniques presented in the manuscript. We will extend the abstract to include these techniques and contributions and rephrase the abstract accordingly.
- Thank you for the paper suggestions. While these models could be of use for libRadtran, it is up to the libRadtran authors to include them there. Some spectral RT speedup methods are already included in libRadtran, such as correlated-k and pseudo-spectral calculation. It must be noted that HAPI2LIBIS only creates the line-by-line input for libRadtran, whereas these methods are distinct approaches for handling spectral RT. Including fast, efficient, and accurate models for gas absorption in HAPI2LIBIS is something that could be added in the future. We will add a mention of fast models and their potential in the manuscript.
- The difference plot between HAPI2LIBIS and REPTRAN will be added to the manuscript.
- Fair comparison is presented in Figure 3. The purpose of Figure 4 is to demonstrate the effect of O3 in the wavelength range 630--635 nm and highlight the capability of HAPI2LIBIS to include new gas information, as in this particular case where the O3 absorption data is missing in HITRAN for this wavelength range. We will rephrase and expand the section where Figure 4 is shown to improve clarity and motivation for this figure.
- Figure 5 will be clarified to make the altitude-dependence of the absorption lines more visible.
- Despite the water vapour continuum absorption being available on HITRAN, we have decided to focus here on the distinct gas absorption lines. Continuum absorption could be included later on as part of HAPI2LIBIS, but for example, the MT_CKD_H2O continuum absorption model is released under a non-free license and would thus not be straightforward to include in HAPI2LIBIS.
- Example image of the optical thickness of different gases in thermal infrared wavelength band will be included along with the appropriate discussion of the results.
Â
Â
Â
Citation: https://doi.org/10.5194/egusphere-2025-220-AC2
-
AC2: 'Reply on RC2', Antti Kukkurainen, 29 Jun 2025
Data sets
Initial version of HAPI2LIBIS with examples Antti Kukkurainen and Antti Mikkonen https://doi.org/10.5281/zenodo.14673990
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 702 | 118 | 21 | 841 | 21 | 41 |
- HTML: 702
- PDF: 118
- XML: 21
- Total: 841
- BibTeX: 21
- EndNote: 41
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1