the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 04 Aug 2026
VIIRS Global Area Coverage (VGAC): A reduced resolution VIIRS dataset for climate data record continuity
Abstract. The Visible Infrared Imaging Radiometer Suite (VIIRS) instrument represents a significant change from the legacy weather satellite observations. For example, there are several differences in scanning strategy, channels, file structure, etc. from the Advanced Very High Resolution Radiometer (AVHRR) which has flown on low earth orbit satellites since the 1970s. Several longstanding climate data records were developed expressly using the AVHRR instrument’s Global Area Coverage (GAC) data. The VIIRS Global Area Coverage (VGAC) dataset is an attempt to more closely bridge the old GAC technology with the new. The VGAC data file provides one orbit of VIIRS observations that have been degraded in resolution to approximate the AVHRR GAC resolution by defining a swath projection model. Each channel is provided using a swath that simulates a scan at ~3.9 km resolution. VGAC provides mean and standard deviation values for each moderate resolution (750 m) VIIRS channel. Higher resolution VIIRS imaging channels (375 m) contain more information, including maximum and minimum radiances, which benefit algorithms like cloud and surface retrievals. The result is that VGAC data are 50 times smaller with 1800 fewer files than the original VIIRS data. The main purpose of VGAC data is to facilitate climate processing to continue the historic climate records from AVHRR and other satellites. VGAC data described here are available from NOAA via https://doi.org/10.25921/gsef-pg81 (Knapp et al., 2024).
Competing interests: Kenneth R. Knapp operates Knapp WeatherSat Services, an independent consultancy, and performed this work under contract with NC State. Yuhan (Douglas) Rao is a topic editor for the ESSD journal.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(3056 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
-
CC1: 'Comment on essd-2026-339', Hilawe Semunegus, 19 Aug 2026
-
CC2: 'Reply on CC1', Hilawe Semunegus, 18 Sep 2026
I would like to correct the interpretation of one result in the technical annex attached to this earlier community comment. This is a follow-up community comment.
The annex reported a median difference of about 0.34 km between the reconstructed and stored coordinates, and attributed that residual to averaging of contributing observations. The subsequent test does not support that interpretation. The earlier comparison held the time-dependent longitude coefficient at a nominal 15 degrees per hour. Fitting one such coefficient on samples from the first half of the same orbit reduces the median difference to about 0.18 m on a sample from the held-out second half. Across all 8260713 valid stored locations in that orbit the median difference is also about 0.18 m, with 0.62 m at the 95th percentile and 0.95 m maximum.
I therefore withdraw the inference that this residual shows the stored positions to be means of contributing observations. What the new result establishes is close reconstruction agreement for one orbit after fitting a parameter within that same orbit. It does not establish the production algorithm, the physical interpretation of the coefficient, or transfer to other orbits, and the fitted value should not be read as a verified Earth-rotation rate. The comparison uses the same archived file as the original annex.
The index-convention and tutorial comparisons in the annex concern the version examined at that time. This reconstruction result does not establish whether later tutorial or production versions differ.
Two requests follow from this. Please read the annex Summary and its sections 2 and 5 with this correction to the residual interpretation. And I would welcome clarification of how the production code defines the time-dependent longitude term, and what locations the stored coordinates are intended to represent.
Citation: https://doi.org/10.5194/essd-2026-339-CC2
-
CC2: 'Reply on CC1', Hilawe Semunegus, 18 Sep 2026
-
RC1: 'Comment on essd-2026-339', Anonymous Referee #1, 06 Sep 2026
The authors describe a methodology and format to represent VIIRS data in a way such as to approximate legacy AVHRR GAC data. This is an important development, allowing to much more easily process the VIIRS data consistently with AVHRR, thus contributing to the construction of homogeneous climate data records from these instruments. The manuscript is fairly clear, and I have no major comments. Below are some specific comments and questions that can hopefully help to further improve the paper.
Specific comments
Abstract: ‘1800 fewer files’ sounds like the number of files is 1800 less (which is not what is meant). It may be clearer to write something like ‘the number of files is reduced by a factor 1800’.
L57-58: AVHRR/1 had 4 channels. AVHRR/3 has 6 channels.
L138-150: This is presented as a list of requirements to continue the ISCCP record but is rather a list of VGAC properties. It looks like only the first two points are relevant for the ISCCP record, the other points describe other characteristics that improve the utility of the data in general. For example, the extra channels in VGAC (3rd point) do not seem to be used in ISCCP.
Section 3.3: From the first paragraph of this section, I understand that the actual quantities reported in the VGAC files are radiances, but it is not crystal clear. For example, line 216 mentions ‘a mean DNB value’. It would be good to make this more explicit, in particular also in Table 2.2.
Section 4: Here more information about the file content is provided. For example, LUTs are included to directly get brightness temperature from radiance. I would suggest to also add this in Table 2.2, so that this table more completely reflects the file content. Is information to convert radiance to reflectance (earth-sun distance and in-band solar constant) also included?
Section 5, L318: I assume the pointing accuracy is actually much higher than the VGAC spatial resolution?
Figure 7: I suggest shrinking the y-axis for clarity and adding trend lines. A slight positive trend is mentioned for SNPP. Is it significant? Suggestion for y-axis label: ‘DCC reflectance anomaly (%)’.
Section 6: The text suggests that VGAC is continuously extended into the present. What is the latency?
Section 6: Perhaps I missed it, but it would be good to specify exactly which VIIRS Level-1 data have been used for the respective instruments, since there have been reprocessings and there are NOAA as well as NASA flavours.
Technical comments
Table 1: Use comma consistently: ’10,313’.
L70: add channel after IR
L79: ‘due to .... the cosine of the satellite zenith angle’ does not make sense. Please rephrase.
L122: are -> is
L150: ‘This is a first for ...’: incomplete sentence. ‘first requirement’?
L170-171: ‘the cells clearly show that the three cells ...’ -> rephrase
Caption Fig. 3: Also, bold black lines are omitted.
L189: Approximately should be denoted by two tildes (one means same order of magnitude)
L232-233: e.g. should be i.e.? And what is ‘it’?
L247: ‘but only varies ..’. What is the subject here?
L249: ‘.. differences ... is ...’
L255: algorithm -> algorithms
L270: ‘VGAC (pixel sizes?) do not.’
L297: I believe either ‘s’ or ‘seconds’ should be used, rather than ‘sec’.
L316: ‘The data quality ... rely ...’
L317: ‘... analyzes ... analyzing ...’
L360-361: ‘Climate Monitoring Satellite Application Facility CLARA-A3 cloud data record’
L389: It seems the abbreviation STAR has not yet been introduced.
Citation: https://doi.org/10.5194/essd-2026-339-RC1 -
RC2: 'Comment on essd-2026-339', Anonymous Referee #2, 21 Sep 2026
This manuscript presents the VIIRS Global Area Coverage (VGAC) dataset, a reduced resolution of VIIRS observations intended to support the extension of climate records based on AVHRR GAC. Given the substantial differences between the two instruments, the authors introduce a continuous swath coordinate system, apply spatial aggregation while retaining selected statistics (mean, standard deviation, etc.), and organize the observations into compact orbital files. This approach significantly reduces the data volume and simplifies the organization of VIIRS observations, helping data users adapt existing AVHRR-based algorithms for use with VIIRS data.
The manuscript is very well written; the organization is very straightforward. It acknowledges the important differences between VGAC and AVHRR GAC and also refers to other application studies for evidence of VGAC’s usefulness. However, the manuscript provides relatively little quantitative information about the effects of the VGAC procedures itself. A focused assessment of these impacts, along with clarification of the extent to which VGAC supports climate-record continuity would help strengthen the paper. I recommend publication after minor revision, addressing the following comments:
- Validation of the VGAC processing
In Section 5, the authors discuss the quality of the underlying VIIRS SDRs through references to previous calibration studies. They also summarize studies that successfully used VGAC in downstream applications. These provide useful evidence supporting the dataset, but they do not directly quantify the effects of the remapping and aggregation performed during VGAC production. Could the authors report any checks undertaken to verify this processing, or provide a representative comparison between VGAC and statistics independently derived from the contributing VIIRS observations? An orbit-level comparison covering different scan positions and scene conditions would perhaps be useful. If existing assessments cover different processing periods, a brief summary would also help establish the consistency of the VGAC product over time. Please discuss whether the processing introduces any additional error or representativeness uncertainty beyond those associated with the source observations and explain its relevance to the intended climate applications.
- Qualification of the continuity claim
The statement in lines 44-46 that VGAC allows VIIRS observations to be used directly with existing AVHRR-based algorithms appears stronger than the qualifications presented in Section 3.5. As Section 3.5 explains that VGAC retains VIIRS spectral response functions and differs from AVHRR GAC in spatial sampling and footprint geometry. Depending on the application, users may therefore need to make further adjustments before applying AVHRR-based algorithms. Please reconcile the introductory statement with the limitations already discussed and make sure that the conclusions use consistent wording. Perhaps describe VGAC as facilitating the adaption and extension of AVHRR-based processing could more accurately convey its contribution while acknowledging the remaining application-specific requirements.
Specific comments:
Line 16-17, 98-100, and 290: Please identify the specific VIIRS product used to obtain approximately 1800 files per orbit. NASA Level-1B and NOAA SDR products organize bands and granules differently.
Line 35: “While the new instruments have superior environmental products, the period of record is much shorter”. This does not seem like a strong argument for VGAC. You could emphasize the need to extend the historical AVHRR-based record while maintaining consistency in observation/retrieval method.
Line 42: “an increase by a factor of 400” is a bit confusing, please be specific as you are referring to the data volume difference.
Line 66: Table 1. The cross-scan sampled counts for AVHRR GAC is 401, but in Section 3.5 it is 409.
Line 124: “nearly impossible” seemed like an overstatement. Please clarify what you mean by “spatial processing”.
Line 199: The sine terms in R2(θ0) appear to have reversed signs relative to the stated coordinate definitions. At the swath origin, the equation returns latitude -θ0 whereas from the source code (Knapp, K. R. 2026) the equation returns the correct starting latitude (θ0). Please reconcile the listed equations with the source code.
Line 226: Figure 4: Please consider including the corresponding native-resolution VIIRS image or a subset, using the same geographic extent and color scales. This would help readers assess visual effects of the aggregation and understand the information retained by the different VGAC statistics. The figure captions should also refer to the specific sub-panels, i.e. reflectance, brightness-temperature, radiance standard deviation…
Line 236: Table 2: Please explicitly list the units and quantities associated with each variable, especially radiance for the MOD and IMG channels.
Line 330: Figure 7: Please report the fitted trend coefficients and their uncertainties to support the claim of long-term stability. Please also clarify how the reflectance anomalies were calculated (e.g., whether a common reference or satellite-specific climatology was used). This would help readers interpret the inter-satellite agreement.
Line 403: change “50x” to “50 times”
References
Knapp, K. R. (2026). knappweathersat/VGAC-tutorial: v1.0.0 - Initial release (Version v1.0.0) [Computer software]. Zenodo. https://doi.org/10.5281/zenodo.19952586Citation: https://doi.org/10.5194/essd-2026-339-RC2
Data sets
Visible Infrared Imaging Radiometer Suite (VIIRS) Global Area Coverage (VGAC) (Version 1) K. R. Knapp et al. https://doi.org/10.25921/gsef-pg81
Interactive computing environment
VGAC Tutorial Jupyter Python Notebooks, v1.0.0 K. R. Knapp https://doi.org/10.5281/zenodo.19952585
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 252 | 121 | 127 | 500 | 189 | 138 |
- HTML: 252
- PDF: 121
- XML: 127
- Total: 500
- BibTeX: 189
- EndNote: 138
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Excellent paper.
Because the projection is generic to cross-track and push-broom (along-track) instruments on near-circular orbits, it is a strong candidate for standardization in the netCDF Climate and Forecast conventions, which at present have no analytic representation for swath geometry. I am pursuing that separately and would welcome the authors' involvement. Stating the conventions precisely in this manuscript, as suggested above, would make that work considerably easier for everyone who follows."
I have also attached a PDF as a 'technical annex' feedback.