Challenges for In-Flight Calibration of Thermal Infrared Instruments for Earth Observation

Round 1

Reviewer 1 Report

As written, this paper is not ready for publication. It does not add enough new insights into radiometric calibration of IR instruments or why PCC's should be considered by others in the community.  The discussion of the requirements flowdown is common and well known to anyone who has calibrated instruments. It can be significantly trimmed. The new technologies at the end are identified as possibilities but there is insufficient linkage of the results to the requirements/needs for this to be useful.

This paper could be significantly improved by:

1. Tightening and shortening the introduction. It is too long relative to the discussions of hardware.
2. Including numbers in the flowdowns. While the full flowdown is not required, we are told that SLSTR needs <0.3K and then we are given a list of the "primary sources of uncertainty." Each of these primary sources should have a numerical estimate of its uncertainty contribution. How much does each of these items contribute to being able to meet (or not meet) the <0.3K requirement?
3. The data that is shown (Figure 4) is not linked either the flowdown or the primary sources of uncertainty. If it supersedes the flowdowns, then why are they included? If it validates them, a stronger linkage is needed.
4. The numbered issues in cross-calibration again need numbers to be useful. Otherwise, it just implies that the issue is hard instead of that the authors have solved a difficult problem. Good inter-satellite cross-calibration is a subject worthy of its own paper. If that is not a primary subject of this paper, the works needs a citation.
5. Next, the paper concludes that Phase Change Cells may improve the calibration. The numbers in this section need to be linked to the flowdowns in a clear, unambiguous way. I don't know how <20mK BB temperature relates to the 0.3K instrument reqt. from what I've read in this paper.
6. The whole paper should be logically streamlined. E.g., "Here is what we need at the top level. Here is what we need at the blackbody level. Here is how PCC's will get us there." Alternatively, "SLSTR has been extremely well calibrated. Here's the proof." This paper attempts to combine these arguments and it is hard to follow.

Author Response

General:

The reviewers main concerns have been noted by the authors. Consequentially, the document has been revised significantly to improve the flow. Furthermore, it will be submitted as a letter as it is intended to be an overview paper.

This paper could be significantly improved by:

1. Tightening and shortening the introduction. It is too long relative to the discussions of hardware.

Response:

Abstract and introduction has been shortened. However, the introduction does provide the background as to why on-board calibration sources and improvements to improve the traceability of the measurements are necessary.

2. Including numbers in the flow-downs. While the full flowdown is not required, we are told that SLSTR needs <0.3K and then we are given a list of the "primary sources of uncertainty." Each of these primary sources should have a numerical estimate of its uncertainty contribution. How much does each of these items contribute to being able to meet (or not meet) the <0.3K requirement?

Response: A clearer breakdown of the budgets has been included.

3. The data that is shown (Figure 4) is not linked either the flowdown or the primary sources of uncertainty. If it supersedes the flowdowns, then why are they included? If it validates them, a stronger linkage is needed.

Response:

The point about the pre-flight calibration is to validate the calibration model and demonstrate that the radiometric performance can be achieved, and additional uncertainty effects can be characterised and accounted for. A separate paper is in preparation that describes the results of the pre-launch calibration.

4. The numbered issues in cross-calibration again need numbers to be useful. Otherwise, it just implies that the issue is hard instead of that the authors have solved a difficult problem. Good inter-satellite cross-calibration is a subject worthy of its own paper. If that is not a primary subject of this paper, the works needs a citation.

Response:

It is not the purpose of this paper to describe the IASI intercomparisons – these will be cross referenced. However, no uncertainty estimates are provided for cross-calibration hence there is no link to SI which is why it is important to improve the on-orbit calibration sources.

5. Next, the paper concludes that Phase Change Cells may improve the calibration. The numbers in this section need to be linked to the flowdowns in a clear, unambiguous way. I don't know how <20mK BB temperature relates to the 0.3K instrument reqt. from what I've read in this paper.

Response:

The crux of the matter is why the PCC is actually needed. 0.1K Decade is the ideal signal that we need to measure for climate studies, do existing technologies provide this, and do the PCC help? To have even the slightest ability to measure this signal at k = 3 requires better than 0.033K/decade instrument stability AND traceability between instruments. A key component of the uncertainty budget is the degradation of the blackbody thermometers after the original calibration. Currently this is based on estimates from manufacturers data and there is no direct method to verify that the drift is within these limits. The conclusion is that the PCCs are the only practical method available to maintain the traceability of the temperature measurement to ITS-90 on-orbit. 20mK is part of the overall calibration budget of the instrument.

6. The whole paper should be logically streamlined. E.g., "Here is what we need at the top level. Here is what we need at the blackbody level. Here is how PCC's will get us there." Alternatively, "SLSTR has been extremely well calibrated. Here's the proof." This paper attempts to combine these arguments and it is hard to follow.

Response:

Accepted that the paper as submitted was not clear in its scope. The intention of the paper was to describe what is currently done to calibrate an IR instrument, the underlying limitations and how the PCCs allow a mechanism to recover some of the calibration budget. Hopefully the revised manuscript should be much clearer.

Reviewer 2 Report

This paper presents a novel technique by embedding the triple phase melt miniature temperature sensor on the blackbody target to detect the block body physical temperature so that SI traceability can be achieved under 20 mk.  The author did elaborate the different error sources when using blackbody for satellite thermal observation calibration, including the non-perfect blackbody design with less than 1 emission coefficient, the error in spectral response function, the non-linearity error of the calibration curve.  While the blackbody temperature is important to be  traceabile to SI standard, there are other ways to achieve or reduce the measurement error. Such as that on NOAA-20 VIIRS , 6 PRT temperature sensors are used to detect the blackbody temperature, while the mean values are used. Does this design achieve similar 20mk accuracy?  The author should compare if other sensors, such as VIIRS, MODIS,  IASI, what’s their blackbody temperature accuracy are.  However, this design itself is significant as a new method to improve and provide a sound absolute temperature calibration.  The authors provide very detailed instruction of the background and necessity of the design and how the traceability can be achieved under laboratory tests.  The research topic fits on the scope of  the ”engineering of remote sensing”.  I recommended it to be published on this journal.

Here are some minor comments
1.) Line 49, this table gives the application of the hyper spectra IR measurements. The authors should mention these  IR measurements are indeed used in the weather forecasting NWP data assimilation to improve weather forecasting skills

2.) Line 79-81, refraphse this sentence, not very clear.

2.) Line 176, please check the time.

3.) Please spell out some abbreviations, such as STEEP3,PAL,RAL,NPL  etc.

4.)120, there are some studies to make possible SRF correction after launch. Such as Chen et al., 2013

Chen, R., Cao, C., and Menzel, W. P. ( 2013), Intersatellite calibration of NOAA HIRS CO₂ channels for climate studies, J. Geophys. Res. Atmos., 118, 5190– 5203, doi:10.1002/jgrd.50447.

Author Response

1.) Sentence has been updated to:

Thermal InfRared (TIR) measurements from satellite instruments have a range of applications ranging from global climate change monitoring, weather forecasting Numerical Weather Prediction (NWP) data assimilation to improve weather forecasting skills and monitoring urban pollution. Table 1, presents a summary of applications from hyperspectral TIR imagers.

2.) Sentence rephrased as:

For climate change monitoring, the satellite instrument must be capable of observing atmospheric and surface temperature trends as small as 0.1 K decade–1 [3].

2.) Time corrected

SLSTR-B between October 2016 to February 2017.

3.) Changes have been made

4.) We have expanded the sentence:

From

The field of view and wavelength selection are relatively stable over time by design and can be calibrated pre-flight,

To

The field of view and wavelength selection are relatively stable over time by design and can be calibrated pre-flight and may be adjusted on-orbit [4],

Round 2

Reviewer 1 Report

The caption for Table 4 is incomplete.

The streamlined text is much easier to follow.

Author Response

Caption for table 4 has been updated

Table 4: SLSTR-A black-body temperature measurement uncertainty estimates due to degradation effects over the 7.5-year mission lifetime in addition to the terms identified in Table 3.  The combined End of Life (EOL) uncertainty estimate includes the BOL estimates and the predicted degradation effects.

also caption for Table 3 has been revised slightly for clarity.

Table 3: SLSTR-A blackbody uncertainties before launch for the 10.8um channel