Bilateral comparison of irradiance scales between PMOD/WRC and PTB for longwave downward radiation measurements

In this work, comparison measurements are presented between two independently realised and characterised blackbody cavities which serve as irradiance standards, namely the well-established Tilted Bottom Cavity BB2007 and the new Hemispherical Blackbody (HSBB). Both are used to realise the unit for thermal infrared irradiance. The BB2007 at the Physikalisch-Meteorologisches Observatorium Davos/World Radiation Center has long provided the reference for longwave downward radiation within the Baseline Surface Radiation Network. Longwave downward radiation is constantly measured at multiple stations around the world using specific broadband infrared radiometers with a hemispherical acceptance angle, for example pyrgeometers. The HSBB, developed at the Physikalisch-Technische Bundesanstalt (PTB) in recent years, was specifically designed to calibrate radiometers with a hemispherical acceptance angle which measure longwave downward radiation. The HSBB is directly traceable to the Radiation Temperature Scale of PTB and, in turn, via this scale to the SI. Comparison measurements between the BB2007 and HSBB in this work were carried out with three transfer instruments: the dedicated Temperature Stabilised Radiation Thermometer, an Infrared Integrating Sphere (IRIS) instrument and a Kipp and Zonen CG4 pyrgeometer. The results show good agreement with respect to the target irradiance uncertainty of 0.5 W m−2 provided by the HSBB. This study thus supports and validates the traceability of atmospheric longwave downward radiation to the SI linking measurements performed with the World Infrared Standard Group of pyrgeometers to the traceability of the reference blackbody BB2007 using IRIS instruments as the transfer standard.


Introduction
Longwave downward radiation refers to the infrared radiation that is emitted in and transmitted through the atmosphere and is incident on the surface of the Earth. It is an important quantity for the surface energy budget of the Earth [1]. Furthermore, it is closely linked to the Earth's greenhouse effect and is therefore particularly interesting for climate research. Longwave downward radiation measurements take place outdoors with specific broadband infrared radiometers that have a hemispherical acceptance angle and are sensitive in the relevant wavelength range from approximately 4 µm to 50 µm. Pyrgeometers are mainly used for this, and measurements take place at many weather stations and research institutes around the globe. These measurements are coordinated, for example, by national weather services or worldwide by the Baseline Surface Radiation Network (BSRN) [2].
The reference for tracing longwave downward radiation measurements within the BSRN to the SI is provided by the Tilted Bottom Cavity BB2007 [3] which was developed at the Physikalisch-Meteorologisches Observatorium Davos/World Radiation Center (PMOD/WRC) and is in operation there. The BB2007 therefore plays a central role in the comparability of longwave downward radiation to other quantities of the surface energy budget of the Earth. Furthermore, PMOD/WRC developed the Infrared Integrating Sphere (IRIS) instruments [4] -windowless transfer standard radiometers featuring a hemispherical acceptance angle and a broad spectral sensitivity -to connect pyrgeometer measurements to the irradiance provided by the BB2007 as the reference. In addition, PMOD/WRC operates the World Infrared Standard Group (WISG), a set of four pyrgeometers with longterm stability that serve as an international standard, and carries out various international comparison measurements with the radiometers in operation [5].
The Physikalisch-Technische Bundesanstalt (PTB) is the national metrology institute of Germany and has a long history of top-level radiation thermometry dating back to the end of the 19th century. Non-contact temperature measurements in the range from −170 • C to 962 • C are realised at PTB with traceability to the fixed points of the International Temperature Scale of 1990 and thus to the SI [6] and at temperatures above 962 • C as thermodynamic temperature directly traceable to the SI [7]. The Radiation Temperature Scale of PTB is verified through various international comparison measurements such as within the scope of the Traceability in Infrared Radiation Thermometry project [8] or by comparison to the Radiation Temperature Scale of the National Physical Laboratory [9]. Four heat-pipe blackbodies of different heat-pipe materials are operated at PTB and serve as national standards for the Radiation Temperature Scale at PTB, covering the temperature range from −60 • C to 962 • C. They are traceable via standard platinum resistance thermometers [6]. Monte Carlo simulations to determine the effective emissivity of these blackbody cavities are performed with high precision. This is achieved by characterising the blackbody wall materials via spectral emissivity measurements at PTB which are traceable to the SI [10].

Outline of the measurements
The traceability of the BB2007 is provided by contact thermometry and Monte Carlo simulations [3]. The objective of this work is to validate the traceability of the BB2007 by performing comparison measurements against an independent irradiance standard that has a target irradiance uncertainty of 0.5 W m −2 . Using those comparison measurements, a second independent traceability path of the BB2007 is established. The comparison measurements may help re-evaluate and reduce the uncertainties in the calibration chain of global longwave downward radiation measurements.
More specifically, comparison measurements between the BB2007 and the Hemispherical Blackbody (HSBB), which was developed, built and brought into operation at PTB [11], are presented. In the first stage of the measurements, the HSBB was calibrated by comparison to an ammonia heat-pipe blackbody which is a national standard for the Radiation Temperature Scale at PTB. In the second stage, the HSBB was brought to PMOD/WRC for blackbody comparison measurements between the BB2007 and HSBB.
The chain for tracing longwave downward radiation measurements within the BSRN to the BB2007 is presented in detail in figure 1. Furthermore, the different traceability paths of the BB2007 and HSBB are illustrated.

Experimental setup and procedure
A detailed description of the BB2007 can be found in [3]. While the BB2007 is a well-established instrument, the HSBB was developed more recently. As the HSBB is a novel blackbody for the measurements presented in this work, a brief description of the HSBB is given in section 3.1.

The HSBB
The HSBB was specifically designed to calibrate pyrgeometers and other broadband infrared radiometers with a hemispherical acceptance angle. Details of the development, operation and Monte Carlo simulations of the HSBB are described in [11]. In short, the HSBB consists of a black coated cone in combination with a highly specular reflecting golden hemisphere. The cone of the HSBB that was employed for the measurements in this work is coated with Nextel 811-21, and this blackbody is also referred to as HSBB1 if it needs to be distinguished from the HSBB2 [11]. For simplicity, the acronym HSBB is used rather than HSBB1 in the following. Four preaged and calibrated Pt100 resistance thermometers are used to The HSBB consists of a black coated cone in combination with a highly specular reflecting golden hemisphere and is designed to specifically calibrate broadband infrared detectors with a hemispherical acceptance angle. Marked are the positions of the Pt100 thermometers at the cone vertex (1), cone edge (2) and the hemisphere (3). In the picture, an IRIS instrument (4) is placed below the opening of the HSBB. measure the temperatures of the HSBB at different positions, namely at the cone vertex, the cone edge, the hemisphere on the left and the hemisphere on the right. The Pt100 thermometers of the HSBB are read out by a resistance readout measurement device. Temperatures are calculated from the resistances according to individual calibration of each thermometer. A rendered sectional view of the HSBB is presented in figure 2. For all measurements presented in this work, the HSBB had an opening aperture diameter of 40 mm.
The HSBB was developed to provide a target irradiance uncertainty of 0.5 W m −2 and was specifically designed to show an almost constant effective emissivity for different viewing conditions, i. e., for different opening or acceptance angles. The effective emissivity results for normal incidence and the hemispherical opening angle determined by Monte Carlo simulations are the same within their uncertainties [11]. These results are supported by measurements with a radiation thermometer under angles of observation from 0 • to 30 • with respect to the optical axis of the HSBB. The measurement results show that the radiation temperature of the HSBB is independent of the angle of observation [11]. The difference in radiation temperature between normal incidence and the hemispherical opening angle is considered negligible in relation to other uncertainty contributions in measurements with the HSBB.

Dedicated radiation thermometer
For the comparison measurements at normal incidence, a Heitronics TRT IV.82-type radiation thermometer was used which measures at normal incidence and is spectrally sensitive in the wavelength range from 8 µm to 14 µm. It was inserted into a dedicated temperature-controlled housing with good thermal isolation which had been specifically built for these measurements. The radiation thermometer is therefore called the Temperature Stabilised Radiation Thermometer (TSRT). It is shown in figure 3(a). Throughout all measurements, the housing was actively temperature-controlled at 23 • C in order to prevent possible changes in the signal of the radiation thermometer due to changes in room temperature and therefore ensure reproducible measurement conditions. It was necessary to use a tilted mirror mounted in front of the radiation thermometer, as shown in figure 3(b), because the BB2007 and HSBB were operated face down and the radiation thermometer needed to detect the radiation beam horizontally. The mirror was also used in the measurements between the HSBB and the ammonia heat-pipe blackbody. The calibration of the reflectivity of the mirror in the wavelength range in which the radiation thermometer is sensitive is implicitly included in the calibration of the HSBB against the ammonia heat-pipe blackbody.

General measurement procedure
Following the outline described in section 2, the actual measurements were carried out in two stages: • First stage, at PTB: calibration of the HSBB by comparison to the ammonia heat-pipe blackbody; performed using the TSRT • Second stage, at PMOD/WRC: comparison between the BB2007 and HSBB; performed using the TSRT, IRIS instrument and pyrgeometer During the measurements, all blackbodies -the ammonia heat-pipe blackbody, the BB2007 and HSBB -were operated between −20 • C and 20 • C, representing the relevant temperature range for atmospheric longwave downward radiation. The corresponding dominant wavelength spectrum ranges from approximately 4 µm to 50 µm. To avoid icing and water condensation from air humidity, all blackbody cavities were purged with dry air or nitrogen. The purging of the HSBB was set to 5 in the unit l min −1 for the measurements with the TSRT. During the relevant measurement periods with the pyrgeometer and IRIS instrument, the purging of the HSBB was briefly turned off to avoid cooling the optical surfaces of the radiometers. For each measurement, the two blackbodies involved -either the ammonia heat-pipe blackbody and HSBB or the BB2007 and HSBB -were operated at the same nominal temperature. Measurements with the TSRT were always performed as enclosed measurements: during the first stage, two measurements at the ammonia heat-pipe blackbody enclosed one measurement at the HSBB, and during the second stage, two measurements at the HSBB enclosed one measurement at the BB2007. This was done to account for possible drifts of the radiation thermometer which served as the transfer instrument.

Measurements at PTB
To ensure the comparability of the measurements with the TSRT between the ammonia heat-pipe blackbody and HSBB, an additional aperture resembling that of the ammonia heat-pipe blackbody was placed below the HSBB. Both apertures were black coated, actively temperature-controlled at 23 • C and had an opening diameter of 30 mm. This setup improves the comparability as it has the same influence on the measured radiation temperature of both blackbodies due to the size-of-source effect [12] or environment factor which is inherent to any radiation thermometer. Furthermore, the room temperature of the laboratory in which the measurements were carried out was actively temperature-controlled at 23 • C. Thereby, the radiation temperature of the HSBB was determined with respect to the cone vertex temperature of the HSBB. The data points denoted by 'First measurements' and 'Second measurements' refer to the measurements carried out before and after the comparison measurements performed at PMOD/WRC, respectively.
To calibrate the HSBB, comparison measurements were performed between the HSBB and the ammonia heat-pipe blackbody at nominal temperatures of −20 • C, 0 • C and 20 • C before measurements took place with the BB2007. After the measurements against the BB2007, measurements of the HSBB against the ammonia heat-pipe blackbody were performed at all nominal temperatures at which the HSBB had been operated during the measurements at PMOD/WRC, including −20 • C, 0 • C and 20 • C. This procedure was done to preclude that transportation might have affected the calibration of the Pt100 thermometers or the coatings of the HSBB. In fact, the measurement results from beforehand could be well reproduced afterwards with deviations of less than 0.012 K. The results of the calibration of the HSBB against the ammonia heat-pipe blackbody are shown in figure 4. The fact that the measurement results from beforehand could be well reproduced implicitly verifies the reproducibility of the Pt100 thermometer measurements. Separate comparison measurements of the resistance readout device against a standard for electrical resistance at PTB were carried out before it was brought to PMOD/WRC. The results could also be well reproduced afterwards with deviations of less than 0.010 K after conversion of electrical resistance to temperature for measurements with Pt100 thermometers. The deviations are small in comparison to the measurement uncertainties presented in section 6.

Measurements at PMOD/WRC
The setup for the blackbody comparison measurements between the BB2007 and HSBB with the TSRT is similar to Figure 5. Setup for the comparison measurements between the BB2007 and HSBB. The TSRT measurements were carried out with the additional temperature-controlled aperture below the BB2007 and below the HSBB. In addition to the TSRT measuring at normal incidence, comparison measurements were also performed with an IRIS instrument and a Kipp and Zonen CG4 pyrgeometer which have a hemispherical acceptance angle. The picture is not drawn to scale. the setup described in section 3.4. The additional temperaturecontrolled aperture was used below the BB2007 as well as below the HSBB. Besides the measurements performed using the TSRT at normal incidence and in the wavelength range from 8 µm to 14 µm, measurements between the BB2007 and HSBB were also performed with an IRIS instrument and a Kipp and Zonen CG4 pyrgeometer which served as broadband transfer instruments with a hemispherical acceptance angle. The schematic setup for the measurements is depicted in figure 5.
The TSRT measurements between the BB2007 and HSBB were carried out by calibrating the TSRT on site against the HSBB which is the calibrated transfer standard. Thereby, the radiation temperature of the BB2007 was compared to the radiation temperature of the HSBB. In contrast, the comparison measurements performed using the IRIS instrument and pyrgeometer were performed in a slightly different procedure. The IRIS instrument and pyrgeometer were calibrated at the BB2007 in their usual calibration procedure [3,4], and the obtained calibration coefficients of the IRIS instrument and pyrgeometer were used for irradiance measurements at the HSBB. The irradiances of the HSBB, known from the calibration of the HSBB against the ammonia heat-pipe blackbody, were compared to the irradiances detected by the IRIS instrument and the pyrgeometer.
In accordance with the calibration procedures of the IRIS instrument and the pyrgeometer at the BB2007, measurements with the IRIS instrument at the HSBB were carried out at nominal blackbody temperatures of 0 • C, 5 • C, 10 • C, 15 • C and 20 • C as well as at −20 • C, −10 • C, 0 • C and 10 • C with the pyrgeometer at the HSBB. Measurements with the TSRT were carried out in the entire relevant temperature range at nominal blackbody temperatures of −20 • C, −10 • C, 0 • C, 10 • C and 20 • C. For redundancy, the measurements with the TSRT were repeated once in an independent measurement series.
Contrary to the IRIS instrument, the housing of the pyrgeometer was actively temperature-controlled by a temperature-controlled plate at certain nominal temperatures following the standard calibration procedure at the BB2007. The housing temperatures of the IRIS instrument and the pyrgeometer were read out by internal sensors. The data points showing for which housing temperature which HSBB irradiance was present are plotted in figure 6.

Data evaluation scheme
In the following, the main evaluation procedure is presented with a description of the evaluation equations. For clarity, smaller corrections such as drift corrections of the TSRT between the measurements are not explicitly mentioned in the formulas.
The radiation temperature of the HSBB is given by T rad,HSBB and is calculated in (1) from the comparison against the ammonia heat-pipe blackbody with the TSRT. Thereby, the calibration of the radiation temperature of the HSBB was performed with respect to the cone vertex temperature of the HSBB. The SI-traceable radiation temperature of the ammonia heat-pipe blackbody is denoted by T 90,heat-pipe . The signals of the TSRT at the HSBB and at the ammonia heat-pipe blackbody are denoted by T HSBB TSRT and T heat-pipe TSRT , respectively.
The radiation temperature of the HSBB transferred to the BB2007 is calculated in (2) in analogy to (1). The signals of the TSRT at the BB2007 and at the HSBB during these measurements are denoted by T BB2007 TSRT and T HSBB TSRT , respectively.
The data in figure 7, exemplarily shown for a nominal blackbody temperature of 0 • C, indicate the good stability of the temperatures of the Pt100 thermometers of the HSBB over all measurements at PTB and PMOD/WRC within their absolute standard measurement uncertainty of 0.041 K. The cone vertex temperature serves as the reference temperature of the HSBB and is denoted by T cone-vertex in the following. In the evalu-ation, the small changes in the cone vertex temperature during the measurements of the BB2007 against the HSBB with the TSRT, IRIS instrument and pyrgeometer compared to the calibration of the HSBB against the ammonia heat-pipe blackbody are accounted for. Thereby, only statistical uncertainties of the cone vertex temperature are used.
Typical values corresponding to (2) are, exemplary for a nominal blackbody temperature of 0 • C: Here, the uncertainties of T BB2007 TSRT and T HSBB TSRT are statistical (Type A) uncertainties only, while the uncertainty of T rad,HSBB is an absolute (Type B) uncertainty.
Due to the HSBB not having a perfect effective emissivity of 1.0, the radiation emitted by the environment is partly reflected by the HSBB. In fact, the laboratory room temperature at PMOD/WRC showed large variations. However, owing to the additional black coated aperture being consistently temperature-controlled at 23 • C and thus the temperature of most of the area around the blackbodies radiating towards them being well defined, a correction for the TSRT measurements is not necessary.
In contrast, it is necessary to correct for the radiation that is emitted by the housings of the IRIS instrument and the pyrgeometer. This radiation is then reflected from the HSBB back to the radiometers. The correction is calculated for the different housing temperatures shown in figure 6. The irradiances onto the IRIS instrument and pyrgeometer are corrected downward by up to 0.74 W m −2 apart from one data point that is negligibly corrected upward. To perform the correction in the evaluation, the radiation temperature of the HSBB that is obtained from the comparison against the ammonia heat-pipe blackbody is corrected by T IRIS corr,HSBB and T Pyrg. corr,HSBB for the IRIS instrument and pyrgeometer, respectively. The calculation of the correction terms is given in (3). The conversion from the integrated radiances in (3) to radiation temperatures is done numerically with accuracy of better than 0.0003 K.
The integrated radiances in (3) are defined in (4) and (5). The radiance of the HSBB during the comparison against the ammonia heat-pipe blackbody is given in (4), while the radiances of the HSBB during the measurements of the IRIS instrument and the pyrgeometer are denoted by L IRIS Planck,HSBB and L Pyrg. Planck,HSBB , respectively, and are given in (5). The integrated radiances correspond to the wavelength range in which the TSRT is sensitive because the temperature correction needs to be applied to the results of the comparison measurements against the ammonia heat-pipe blackbody with the TSRT. From these results, the irradiances for the IRIS instrument and pyrgeometers are subsequently calculated. The effective emissivity ε eff of the HSBB corresponds to normal incidence for TSRT measurements and is obtained from Monte Carlo simulations [11]. Its value amounts to ε eff = 0.99529 with u(k = 1) = 0.00099 and refers to the TSRT measurements with the additional aperture in the focus of the TSRT. The housing temperatures of the IRIS instrument and the pyrgeometer are denoted by T housing,IRIS and T housing,Pyrg. , respectively.
After correction of the radiation temperature of the HSBB given in (6), the irradiances of the HSBB are obtained according to (7). The irradiances for the IRIS instrument and the pyrgeometer are denoted by E IRIS HSBB and E Pyrg. HSBB , respectively, and the Stefan-Boltzmann constant is denoted by σ. The use of (7) is appropriate due to the HSBB showing the same effective emissivity for normal incidence as well as the hemispherical opening angle and the coatings having spectrally flat optical properties [11].
The Stefan-Boltzmann law is used for the calculation of the irradiance in accordance with the common calibration procedure for IRIS instruments with the BB2007 [3]. While the IRIS instrument is a windowless radiometer with spectrally flat responsivity and therefore very suitable for blackbody calibration, pyrgeometers often have spectral transmission features due to their silicon domes and are typically calibrated outdoors [4]. The correction values of the radiation temperature and of the irradiance of the HSBB can be found in the appendix listed in tables A1 and A2 for the IRIS instrument and pyrgeometer, respectively. The corresponding uncertainties can be found in section 6.

Results
In this section, the results of the measurements are described. The agreement in radiation temperature and irradiance of the BB2007 and HSBB is explained in this section and is assessed in terms of the target irradiance uncertainty of 0.5 W m −2 . The corresponding uncertainty budget is presented in the following section. The HSBB was established as fit-for-purpose reference that provides an irradiance uncertainty of 0.5 W m −2 or better. This value was considered necessary to obtain meaningful blackbody comparison measurement results and to improve the traceability of the BB2007. In fact, the HSBB achieved even better uncertainties. The original target uncertainty of 0.5 W m −2 is considered as a reasonable benchmark value for the scale difference investigated here. The results from the TSRT measurements between the BB2007 and HSBB are shown in figure 8 and are given as the difference in the radiation temperature between the BB2007 and HSBB. Shown are the results of the two independent measurement series.
For the TSRT measurements, the differences range from −0.116 K, averaged at a nominal temperature of 20 • C, to 0.086 K, averaged at −20 • C. The differences for the temperatures from −20 • C to 10 • C are within the target temperature uncertainty corresponding to the target irradiance uncertainty of 0.5 W m −2 . The target uncertainty is indicated by the grey area in figure 8. The differences for the temperature of 20 • C are outside the target uncertainty, but the uncertainty bars of the differences extend into the target uncertainty. For each nominal temperature, the corresponding differences are in agreement with each other within their uncertainties and are therefore considered well reproducible. In terms of reproducibility, the largest deviation between two differences for the same nominal temperature was found at −20 • C with a deviation of 0.050 K.
The results from the measurements with the IRIS instrument and pyrgeometer are shown in figure 9. Here, the results are obtained by subtracting the irradiances of the HSBB calculated according to (7) from the irradiances of the HSBB which were detected by the IRIS instrument and the pyrgeometer, respectively, with their calibration coefficients obtained from measurements at the BB2007 as described in section 3.5.
The differences found in the measurements performed using the IRIS instrument are all within the target irradiance uncertainty of 0.5 W m −2 . Here, the differences range from −0.17 W m −2 with u(k = 1) = 0.55 W m −2 at a 15 • C nominal blackbody temperature to 0.37 W m −2 with u(k = 1) = 0.84 W m −2 at 0 • C. It should be noted that, due to combined uncertainties, the uncertainties here are mostly larger than 0.5 W m −2 . For the measurements performed using the pyrgeometer, the differences range from −0.63 W m −2 with u(k = 1) = 0.46 W m −2 at 10 • C to 0.54 W m −2 with u(k = 1) = 0.50 W m −2 at −20 • C. Four out of six differences found in the pyrgeometer measurements are within the target uncertainty. The remaining differences, namely one each at nominal temperatures of −20 • C and 10 • C, are outside the target uncertainty. The uncertainty bars of these differences, however, extend into the target uncertainty.
The differences corresponding to figures 8, 9(a) and (b) can be found in the appendix listed in tables A3, A4 and A5, respectively.

Uncertainty budget
The main individual uncertainty components are listed in table 1. The dominant uncertainty contribution of the radiation temperature -and subsequently the irradiance of the HSBBcorresponds to the uncertainty associated with the ammonia heat-pipe blackbody against which the HSBB was calibrated.
In table 1, 'TSRT measurements between BB2007 and HSBB' corresponds to additional uncertainties for the TSRT comparison measurements between the BB2007 and HSBB, i. e., the influence of the size-of-source effect of the TSRT and the influence of the aperture's position. The items 'IRIS measurements between BB2007 and HSBB' and 'Pyrgeometer measurements between BB2007 and HSBB' correspond to the uncertainties associated with the calibration coefficients of the IRIS instrument and pyrgeometer, respectively, which were determined at the BB2007 and used for measurements at the HSBB.
To correct the radiation temperature of the HSBB for IRIS and pyrgeometer measurements according to (3), the uncertainty of the effective emissivity of the HSBB amounting to 0.00099 is used. The uncertainty of the temperature measurement with the Pt100 thermometers of the HSBB is 0.041 K, and that of the room temperature is 1.1 K. The calculation of the uncertainty of the integrated radiances in (4) and (5) is based on [13] and is presented in (8).  The correction of the reflection term was applied to all values for consistency even though it may have not been necessary for the values with housing temperatures close to 23 • C. Future measurements will reveal, depending on typical housing temperatures, whether the correction is necessary or not or if it would be useful to transform the correction into an uncertainty contribution only.

Discussion and conclusion
The objective of the presented measurements was to establish a second traceability path and at the same time to independently validate the existing traceability of the well-established BB2007 with the help of the new HSBB. To do so, comparison measurements between the BB2007 and HSBB were carried out. For the comparison measurements performed using the IRIS instrument, very good agreement was found with all differences lying within the target irradiance uncertainty of 0.5 W m −2 of the HSBB. For the comparison measurements performed using the TSRT and the pyrgeometer, the majority of the differences was within the target irradiance uncertainty of the HSBB. Small trends with decreasing values for increasing nominal temperatures of the BB2007 were obtained with the trends crossing the 0 K and 0 W m −2 line between −10 • C and 0 • C nominal temperature for both the TSRT and pyrgeometer measurements, respectively. The trends are insignificant for this work. The cause of the trends is difficult to fathom and will thus be the subject of further investigation.
To conclude, the comparison measurements performed using the IRIS instrument as the relevant instrument for the irradiance scale transfer can be regarded as highly successful. Based on these and the very good results from the measurements performed using the pyrgeometer, it can be inferred that the BB2007 and HSBB are equally well suited for the uniform irradiation of radiometers with a hemispherical acceptance angle. Overall, no systematic errors in the existing traceability of the BB2007 could be identified.
The results of this study will support the process initiated by the World Meteorological Organization to redefine the traceability chain for atmospheric longwave downward radiation that is currently based on the WISG with an estimated uncertainty of 10 W m −2 to one based on the BB2007 via IRIS instruments as the transfer standard with an uncertainty of 2 W m −2 , as proposed in [5]. As a result, measurement uncertainties associated with the WISG pyrgeometers will be reduced in the future.
Several attempts have been made so far to describe the radiation and energy transfers between the Earth, the Sun and the atmosphere by means of simulations. Relatively large discrepancies arise from the results of different simulation approaches for longwave downward radiation [14]. Reduced measurement uncertainties may increase the comparability between measurements and simulations, help identify the appropriate simulation approaches and lead to resolving the differences.
In the future, more measurements may be undertaken with the HSBB in a round robin-like system in order to improve the overall consistency of longwave downward radiation measurements within the BSRN. Pyrgeometers of different types, IRIS instruments and Absolute Cavity Pyrgeometers [15] may be employed.
In summary, comparison measurements were carried out between two independently traceable reference blackbodies, the BB2007 operated by PMOD/WRC and the HSBB operated by PTB. Good agreement was found, representing the successful bilateral comparison of the irradiance scales of PMOD/WRC and PTB for longwave downward radiation measurements. Therewith, the existing traceability of the BB2007 to the SI could be verified to a highly satisfactory overall degree. Through this comparison, a second independent traceability path for the BB2007 was established. The results can be regarded as a major achievement on the road towards reducing measurement uncertainties and improving the significance and impact of longwave downward radiation measurements.

Data availability statement
Any data that support the findings of this study are included within the article.

Acknowledgments
This project, which is part of 16ENV03 METEOC-3 and 19ENV07 METEOC-4, has received funding from the EMPIR programme co-financed by the Participating States and from the European Union's Horizon 2020 research and innovation programme.