The Expectations and Prospects for Quantitative Volcanology in the Upcoming Surface Biology and Geology (SBG) Era

The Surface Biology and Geology (SBG) mission is part of the National Aeronautics and Space Administration's planned Earth System Observatory. The mid/thermal infrared (MIR/TIR) portion of SBG will deliver high spatial and temporal resolution data designed (in part) to answer fundamental volcano science questions, as outlined in the 2018 Decadal Survey (DS). The planned 60 m spatial resolution, high TIR saturation temperature (∼500 K), and addition of two high temperature MIR bands (∼800–1200 K) will improve thermal feature and spatial variability detection across all volcanic surfaces, as well as limit data saturation over high temperature targets such as lava flows. However, the MIR/TIR part of the mission also has design aspects that will hinder volcanic science. The ability to derive accurate spectral information of volcanic, and indeed other geologic surfaces, is a function of the number and placement of the TIR bands. The originally proposed five TIR bands do not advance these diagnostic investigations beyond what is currently available. In fact, it may hinder them. Six or more TIR bands are vital for accurate assessment of mineral and glass abundances in volcanic deposits. The SBG MIR/TIR mission temporal resolution (3–5 days) will improve repeat coverage of volcanic systems, aiding in longer‐term monitoring, but observations of highly dynamic processes will not be possible, despite their importance in the 2018 DS. Most problematic is the proposed local overpass time of 12:30, which will result in cloudier scenes (especially in tropical/lower latitude regions) and complicate atmospheric corrections compared to the traditional morning overpass times of earlier TIR missions.


The Expectations and Prospects for Quantitative Volcanology in the Upcoming Surface Biology and Geology (SBG) Era
These objectives are inherently in conflict with one another, requiring very different scales of data. However, the Earth Surface and Interior panel clearly saw these both as most important for the science and for NASA to achieve. As planned, the current SBG architecture achieves S-1a and fails S-2a, particularly for volcano science.
The proposed spatial (60 m) and temporal (3-5 days at the equator) resolutions of the MIR/TIR instrument do dramatically improve what is currently available (e.g., 90-100 m and 16 days at the equator), but does not provide the much higher data rates required to understand dynamic systems that are the focus of the S-2a objective (Figure 1). High-temporal resolution satellite-based data at optimal overpass times are key to understanding dynamic volcanic systems during an eruption, whose processes vary over timescales of seconds to hours (Ramsey & Harris, 2013). Despite decades of orbital data ( Figure 1, Table 1), we are not yet able to resolve these small (10s of meters scale) dynamic systems to quantify, for example, the flux rates of lava flows and volcanic gases, as well as constrain the contributions to atmospheric budgets over rapid (sub-daily) timescales (Carn et al., 2017). The size, volume, and flux rate of these systems are far below the spatial and temporal resolution of current sounding sensors that have higher spectral resolution, whereas higher spatial resolution imagers lack the requisite temporal and spectral resolution.
As important for surface science as the temporal cadence of the observations are both the spectral resolution and overpass time. Overpass time plays a role in the amount of expected clear-sky observations, especially over mountainous, volcanic regions. More clouds at certain times of the day will inhibit the detection of precursory thermal and gas activity (Laiolo et al., 2022). Furthermore, the spectral resolution in the 8-12 μm region determines the accuracy of surface and gas compositional retrievals, as well as improving temperature determination. For example, the development of instruments with higher spectral resolution, collecting data from diagnostic regions of the electromagnetic spectrum, allows the physical and petrologic properties of volcanic rocks to be determined, in turn improving volcanic hazard forecasting (NASEM, 2018). SBG may provide improved spatial and temporal resolution TIR data from that which is available now, but it is not currently configured to significantly reduce these knowledge gaps (Figure 1).

Results and Discussion
Multispectral TIR data from current satellite instruments (e.g., Advanced Spaceborne Thermal Emission and Reflection Radiometer (ASTER), ECOsystem Spaceborne Thermal Radiometer Experiment on Space Station (ECOSTRESS), Moderate Resolution Imaging Spectroradiometer (MODIS), Geostationary Operational Environmental Satellites Advanced Baseline Imager (GOES ABI)) as well as the airborne (MODIS/ ASTER Airborne Simulator (MASTER)) instrument, are used here to assess the potential observations SBG will make with its currently-proposed spatial, spectral, and temporal resolution. Kinetic temperature and emissivity are derived from the calibrated surface radiance data using the temperature emissivity separation (TES)  (Gillespie et al., 1998) where possible or alternatively using the emissivity normalization algorithm (Realmuto, 1990). The effect of data specifications on the accuracy of the derived analyses and conclusions for volcanic applications are discussed. Finally, the potential consequences of overpass time are also considered for evaluating hazard assessments and forecasting. Resolutions of current and planned orbital instrument data together with the trade space for the Earth Surface and Interior measurement objectives (S-1a, S-2a). Note, thermal infrared (TIR) instruments with spatial resolutions >1 km are omitted, including, for example, the Geostationary Operational Environmental Satellites Advanced Baseline Imager (GOES ABI -2 km/0.007 days).

Spatial Resolution
Larger-scale volcanic eruption products (e.g., ash clouds) can be tracked relatively easily through the atmosphere using lower spatial but higher temporal resolution instruments such as the GOES ABI (e.g., Pavolonis et al., 2019) or MODIS (e.g., Watson et al., 2004). In contrast, smaller-scale volcanic feature discrimination (e.g., lava flows, gas fumaroles, small eruptive plumes) is limited by the spatial resolution of those data. The ability to correctly identify different volcanic phenomena and their change over time is particularly important to understand each volcano's behavior. For example, the appearance of new fumaroles or slight increases in the thermal and/or gas flux at existing fumaroles can signal changes in the underlying hydrothermal system, which is responding to even deeper magma migration (Madonia, 2020). These initial changes are commonly too small spatially and thermally to be detected by most orbital TIR instruments. However, this is vital information as is understanding the spatial distribution of temperature and emissivity variations across lava flow surfaces, which can assist in modeling of potential hazards (e.g., FLOWGO; Chevrel et al., 2018;Harris & Rowland, 2001). Figure 2 shows the impact of spatial resolution on constraining the path and temperature variability across the lava flow surface. Kinetic temperature and emissivity are derived from the calibrated surface radiance data of ASTER and MASTER using the TES algorithm (Gillespie et al., 1998). The spatial resolution of the MASTER data is degraded using nearest neighbor resampling to evaluate the effect on subsequent analyses.
The proposed 60 m spatial resolution of the SBG MIR/TIR instrument, provides a 33% improvement in feature detection and spatial variability for lava flows ( Figure 2b) and other thermal features compared to the 90 m ASTER resolution (Abrams et al., 2015). ASTER currently provides the highest spatial resolution global multispectral TIR data from polar orbit ( Figure 2a). The ECOSTRESS instrument delivers a higher spatial resolution than ASTER but is limited to the inclined equatorial orbit of the International Space Station (Fisher et al., 2020). The improved spatial resolution of SBG will also resolve higher surface temperatures without saturating than ASTER and therefore produce more accurate emissivity data that enable the analysis of both subtle (<1 K; Figure 2 blue dashed circle) and extremely hot (>1000 K; Figure 2 red dotted circle) temperature anomalies (Corradino et al., 2020;Girona et al., 2021;Ramsey et al., 2022;Thompson & Ramsey, 2020). Additionally, more detailed mapping of volcanic deposits and volcanic ash/gas plumes will be possible, enhancing our understanding of eruption processes and hazard potential. The higher spatial resolution and saturation temperatures (MIR: <1200 K; TIR: <500 K) may also minimize "thermal bleeding" from very hot surfaces to pixels of cooler surfaces in close proximity. This effect is a function of the spatial resolution of the data, the radiometric response of the detectors, and the surface temperature. Therefore, it is more pronounced in hot anomalies (e.g., at lava lakes and flows; Figure 2 red dotted circles) (Thompson & Ramsey, 2020). Overall, the 60 m resolution could enable smaller volcanic features to be detected and analyzed more accurately than currently possible, thus improving our ability to interpret the condition of actively erupting volcanic systems in their entirety.

Spectral Resolution
Although dominantly used for surface temperature, TIR data also provide a measure of the TIR emissivity, which is governed by the surface composition, along with other surface properties. The 8-12 μm atmospheric window of the TIR region is particularly sensitive to the fundamental vibrational frequencies of the major rock-forming minerals' dominant structures (e.g., Si-O, Al-OH, Ca-O). This makes the TIR ideal for geological mapping, exploration, and a compliment to the VSWIR, which lacks sensitivity to many of these minerals. Increasing the number of TIR wavelength bands improves the spectral resolution of a sensor and makes it more useful for mineral discrimination. For example, the ability to derive accurate composition of volcanic eruptive material is limited by the number and placement of the TIR bands in current orbital sensors like ASTER. To derive the most accurate interpretation of geologic composition from space, it is vital that the number of bands is maximized, and as important, those bands are placed in positions that allow the most accurate discrimination of the important mineral groups (e.g., those found in extrusive volcanic rocks, such as plagioclase and potassium feldspars, pyroxenes, quartz, and olivine). Volcanic glass is also important to quantify, as it can dominate the surfaces of rapidly cooled lava flows, as well as being a major component of volcanic ash and other pyroclastic deposits (up to 80% in volcanic ash deposits; Horwell, 2007). The ASTER TIR band number and placement were chosen to detect and map surface mineralogy. The five bands are placed to allow the discrimination of multiple mineral species, particularly quartz and carbonates (Watanabe & Matsuo, 2003). It is less well-suited, however, to detect feldspar minerals, although ASTER band 13, located at 10.56 μm, does provide a limited capability.
Preparatory work conducted during studies for the past (and never developed) Hyperspectral Infrared Imager (HyspIRI; Abrams & Hook, 2013) mission, proposed that this band be moved shorter to 10.11 μm (Ramsey et al., 2012) to further improve the retrieval of this important geological mineral class. During the initial development of the SBG MIR/TIR instrument, a maximum of eight bands was specified. From that set, three were set aside for: (a) cirrus cloud detection, (b) CO 2 retrieval, and (c) high temperature mapping, leaving only five bands in the TIR-the same number that we have had with ASTER since 2000. Furthermore, a specified requirement of having two bands between 11 and 12 μm was made to retrieve surface temperatures more accurately, thus leaving only three bands from 8 to 11 μm. This configuration will negatively impact the ability to map surface mineralogy and geology. The current thinking appears to be that mineral mapping will be accomplished by the VSWIR system, but this is not possible for many rock-forming mineral classes. Despite having the same number of TIR bands as ASTER, and therefore not representing an improvement 20 years later, the SBG TIR band placement is arguably worse. To understand the effect of band placement for the SBG TIR instrument, we used data acquired by the MASTER instrument over the Kelso Dunes in the Mojave Desert, California ( Figure 3). This region was chosen because past field, laboratory, and airborne instrument studies of the dunes have well constrained the composition of the material (e.g., Muhs et al., 2017;Ramsey et al., 1999), making it a valuable test and validation site.
MASTER L1B data were converted from radiance to emissivity using the emissivity normalization method (Realmuto, 1990) and then underwent spectral deconvolution modeling. TIR pixel spectra are assumed to be the sum of the areal percentage of each spectral end-member found within each pixel, meaning that a linear model can be applied using the approach of Ramsey and Christensen (1998). The Arizona State University Thermal Emission Spectrometer mineral spectral library (Christensen et al., 2000) were resampled to each image spectral resolution and used as end-members for the model. Figures 3b-3d demonstrate the change in the areal abundance of three important components of the Kelso Dunes depending on the number of spectral channels used. The fiveband results (Figure 3b) show greater detail and variability across green and cyan color end-members, indicating the presence of quartz and/or plagioclase feldspar. However, with the inclusion of a sixth and seventh spectral band, found between 10 and 11 μm, the images show increased detection and mixing of potassium feldspar, which is accurate (Ramsey et al., 1999). The placement of the sixth band at 10.11 μm was chosen based on the HyspIRI band study (Ramsey et al., 2012), whereas the seventh band at 10.63 μm was selected owing to the strong absorption of more Ca-rich plagioclase around this wavelength, relative to Na-rich plagioclase and K-feldspar. Fifty equidistant points over the dunes (Figure 3a) were used to retrieve the end-member composition ( Table 2).
The results demonstrate that the lack of a band between 10 and 11 μm negatively impacts the mapping of feldspathic minerals within the dune system. The values obtained using the six-and seven-band configurations are much more representative of the mineral abundances recorded in the literature (see Ramsey et al., 1999 andMuhs et al., 2017 for more information). The five-band data deconvolution result severely underestimated the quantity of potassium feldspar present. Additionally, mineral abundances for each data point were exclusively bimodal. None of the pixels returned abundances for both plagioclase and potassium feldspar and were instead either one or the other plus quartz. This relationship was not found with the addition of a sixth and seventh band.
Because of the relatively simple nature of the linear deconvolution approach, poor end-member selection or limited spectral resolution can commonly result in a perceived good model fit. For example, the results from  Table 2 and full results see Table S1.
the SBG data with five bands have equally low root-mean-square errors, which indicate the model's "goodness of fit," as the six and seven band datasets. The results here demonstrate that if the SBG MIR/TIR mission does not include at least one band between 10 and 11 μm, future mineral mapping modeling may yield low errors but would provide false results relative to the true mineralogy found at the site.
However, most extrusive volcanic rocks are not made of pure minerals that are commonly found in spectral libraries. In many cases, lava flows and pyroclastic deposits contain a mixture of mostly glass and some mineral fragments in varying proportions. This complex composition means there are further spectral considerations needed where trying to understand the petrology of erupted volcanic rocks, which is critical to understand the volcanic processes occurring during an eruption or period of unrest. One example that has been previously studied using TIR emission spectroscopy from both the laboratory and satellite remote sensing is that of volcanic ash. Changes in the particle size and petrology of the ash, including the variation in the proportion of glass to mineral fragments, can be determined from satellite TIR data, provided the instrument bands are placed where these features emerge. A laboratory spectral library of volcanic ash samples was first developed for use with ASTER TIR data by Williams and Ramsey (2019). For this work, specific end-members representing lower and higher wt.% SiO 2 and glass content are resampled to show the effects of band placement on fine-grained volcanic material (Figure 4). Here we show the spectral resolution of the ASTER TIR instrument and compare it to the planned five-band SBG TIR instrument as well as a proposed six-band configuration.
The varying mixtures of both mineral and glass percentages are shown (Figure 4), with some spectra exhibiting the features of glass more strongly (Santiaguito between 8 and 9 μm), whereas others are more mineral-rich (Fuego). As the glass percentage increases, different spectral features appear. With increasing SiO 2 content of the glass, a subtle shoulder feature between 8.5 and 9 μm appears (Figure 4 green arrow), which can be used to assess the eruptive composition. This is particularly evident in the samples with a higher glass content, particularly where the glass is enriched in SiO 2 relative to the bulk composition. The addition of a sixth band both resolves the silicate glass spectral shape at 10 μm as well as improves discrimination between the different particle size fractions of the same composition.
In particle size ranges that are <63 μm diameter, the development of "transparency features" caused by diffraction is seen between 10.5 and 12 μm in the laboratory spectra. These features are detected at ASTER resolution (seen as a decrease in emissivity between 10.5 and 11.5 μm in the finer particle size ranges). At the planned SBG TIR spectral resolution, these same features are somewhat better resolved because of the two bands between 11 and 12 μm. However, the lack of a 10-11 μm band in the SBG planned configuration means that important spectral features caused by feldspar minerals found in the Fuego sample are not resolved, which could lead to inaccurate identifications. In our proposed six-band solution, the additional band yields significantly more spectral information, resulting in spectral shapes similar to the laboratory spectra, and an improvement over the ASTER TIR spectral resolution. In summary, the planned SBG TIR wavelength band placement should improve particle size retrievals in ash-rich plumes and deposits because of the bands between 11 and 12 μm; however, the lack of a band(s) in the 10-11 μm region will negatively impact accurate compositional retrievals for all compositions. At a minimum, a sixth band at ∼10.1 μm would solve this issue and allow more accurate spectral mapping in the TIR.
8 of 13 quent analysis of the 10-min temporal resolution data from the ABI sensor on the GOES satellites confirmed the detection of precursory and variable activity earlier (and with higher confidence) than other higher spatial but lower temporal resolution orbital instruments ( Figure 5) (Thompson et al., 2022). For example, ASTER has a   Table 2), and the proposed six-band configuration that includes an additional band at 10.1 μm (note the much better spectral agreement of this configuration to the laboratory spectra). The green arrow denotes the location of the "shoulder" feature that develops in samples containing high SiO 2 (i.e., >60 wt.%) glass. This feature is not observed in the Fuego sample but is evident in the Santiaguito sample.
nominal temporal resolution of 16 days, but with the aid of sensor webs and the instrument's pointing capability, revisit intervals can be reduced significantly. This can include day/night pairs at lower-latitude targets, and image triplets (three images within 48 hr) at higher latitude volcanoes (Ramsey & Flynn, 2020). However, the cadence of ABI data is closer to that of seismic data and other ground-based monitoring data. This increases the opportunity for data synergy and combined analysis that can aid in the identification and interpretation of magmatic and eruption processes, not possible with current low-Earth orbit instruments or with a single geophysical dataset (e.g., single station seismometer) (Thompson et al., 2022). However, this is only possible for large volcanic eruptions due to the 2 km spatial resolution of ABI data.
The current proposed cadence of the SBG MIR/TIR system (3 days at the equator) will be a significant improvement over what is currently available from polar-orbiting TIR instruments. The data will represent an advance in our ability to monitor the world's active volcanoes, thus achieving much of what was designated in the Earth Surface and Interior Science and Application Measurement Objective S-1a (NASEM, 2018). This resolution could be further improved through the synergy of MIR/TIR data from other proposed MIR/TIR orbital instruments, including the Thermal infraRed Imaging Satellite for High-resolution Natural resource Assessment (TRISHNA) and the Land Surface Temperature Monitoring (LSTM) instruments . For this synergy to be of use, however, it will be important to have consistent instrument performance metrics and data quality between missions. Recent discussions of changing SBG from a 3-day to 5-day revisit, as a contingency (threshold mission), will negatively impact this assessment. A 5-day revisit time only achieves what is currently possible using ASTER's pointing coupled with sensor-web based triggering (Ramsey, 2016). That together with the planned spectral resolution is arguably just a continuation of ASTER with a slight improvement in spatial resolution plus bands in the MIR.
Even at the 3-day temporal revisit time, observations of highly dynamic volcanic processes that occur during eruptions, including ash plume ascent, volcanic gas-atmospheric interactions, and lava fountaining will still not be possible with SBG. This is the fundamental step-change in volcano science required to understand how eruptions work, and the focus of Science and Application Measurement Objective S-2a and S-2b (NASEM, 2018). This higher temporal (and importantly, higher spatial) resolution data will improve understanding of transient eruption dynamics that are only possible with ground-based TIR cameras and in a limited way from geostationary data for very large eruption ( Figure 6).

Overpass Time
Active volcanoes located at sub-tropical and tropical latitudes commonly develop daily cloud cover that hinders routine observations from orbit. This is especially noteworthy after ∼10:00 local time. Geostationary data indicates that cloud cover is greatest over lower latitude volcanoes from 12:00 to 23:00 local time (Figure 7b). The greatest chance of clear-sky observations is therefore between 08:00 and 12:00 local time. Additionally, the increase in solar radiation later in the day complicates data corrections and anomaly detection by warming nonvolcanic slopes to temperatures at/higher than actual thermal anomalies caused by volcanic activity (Figure 7a). Increasing the probability of acquiring routine clear-sky observations of all active volcanoes will improve longterm analyses, precursory activity and eruption detection, as well as the accuracy of derived products (e.g., emissivity, fluxes). Unfortunately, it is likely that the SGB MIR/TIR instrument will have a nominal overpass time of midday to early afternoon driven by ecosystem requirements, plant stress, and synergy with overpass times of other future TIR imaging mission (LSTM and TRISHMA). This will reduce the opportunity to observe volcanic activity in the daytime especially those at sub-tropical and tropical latitude. The higher cloud cover effectively lowers the data's actual temporal resolution to an apparent cadence for volcanoes that is approximately 20%-60% less than that of what is planned. This range is based on previous analysis of ASTER data where ∼50% of scenes had cloud cover (Ramsey & Flynn, 2020) and cloud cover analysis using GOES (20%) that likely represents the minimum loss because small clouds over summits are undetectable at low spatial resolutions (∼2 km). In contrast, the nighttime overpass at approximately 00:30 would occur during the general low period of cloud coverage over these volcanoes (the minimum for some) between approximately 20:00 and 03:00 ( Figure 7b). Furthermore, the Figure 7. The yearly average number of (a) thermal anomalies and (b) cloud cover detected each hour using data from the Advanced Baseline Imager (ABI) on GOES-16 and -17 between 2019 and 2022. The hourly averages are reported at six of the current most active volcanoes that span latitudes from 0° to 64° and climatic environments from arid to tropical. The gray and green boxes indicate the currently proposed and our recommended Surface Biology and Geology (SBG) mid/thermal infrared (MIR/TIR) instrument daytime overpass time, respectively. The earlier overpass time will detect far more preeruptive activity based on analysis of the ABI data.
offset of overpass time with ASTER (the most comparable TIR instrument currently in orbit) and the Landsat TIR sensors reduces and complicates the ability to conduct long baseline comparative analysis. On the other hand, the synergy and harmonization with future TIR imaging missions (LSTM and TRISHNA) will reduce if an earlier SBR MIR/TIR overpass time is chosen. This should not have a major influence on volcanic analyses but will affect ecosystem science. However, this could provide an opportunity to analyze variations at different times during the diurnal cycle, potentially resulting in more insights into ecosystem and volcanic changes over several hours.

Conclusions and Recommendations
As currently planned, the proposed SBG MIR/TIR orbital instrument will ultimately improve thermal volcanic analysis from space over current capabilities. This will be most noteworthy with the improved temporal and spatial resolutions. However, the planned spectral resolution and overpass time limits the volcanic applications of the data at the expense of other priorities. There is a great opportunity with the SBG mission to push the boundaries of MIR/TIR remote sensing and transform the state-of-the-art for the field of volcanology. Notably, an improved spectral resolution, coupled with the planned spatial and temporal resolution data, allows more detailed modeling of eruptive products (i.e., ash and SO 2 ), the state of a volcano, and its evolution over the course of an eruption. Recognizing that there are competing science interests for these data and that it is likely not possible to change all the planned specifications, we recommend several changes below (Table 3), which advocate for volcano science. These are organized from most important and achievable to those that are more difficult to attain (from top to bottom Table 3).
As currently planned, the SBG MIR/TIR specifications will improve volcano science, specifically monitoring over the timescale of days to months. This may ultimately improve our ability to forecast new eruptions. However, the choice of certain instrument and mission parameters at this early planning stage also makes this prospect more difficult. Some of these could be changed prior to the development schedule if the will is there from the volcanology community. However, even with all the recommended changes, SBG still will not capture data at a cadence required to understand the myriad of dynamic processes ongoing in an active eruption. These requirements for volcanology could be improved through the synergy and harmonization of TIR data from other future mission planned by other agencies (e.g., LSTM and TRISHNA), but NASA has limited influence on the specifications, parameters, and data quality of these missions and instruments. However, NASA has the opportunity to augment SBG with targeted smaller missions as part of their Earth Venture Program that could acquire these TIR data and address all the volcano-specific objectives recommended by the Earth Surface and Interior panel in the 2018 DS.

Conflict of Interest
The authors declare no conflicts of interest relevant to this study.

Table 3 Recommended Changes to the Surface Biology and Geology (SBG) Mid/Thermal Infrared (MIR/TIR) Mission to Achieve Improved Outcomes for Volcano Science
Volcano science specific recommendations Improve the spectral resolution by adding (at a minimum) a sixth TIR band in the 10-11 μm range to improve surface and volcanic compositional mapping, as well as reduce analytical errors. a Change the day and night local overpass time to mid-morning (ideally near/at the time of Terra, Landsat, and the proposed time of the SBG VSWIR instrument). This is significantly better for clear sky volcanic observations, especially in the tropics and sub-tropics where clouds are more prevalent in the afternoon and hinder observations.
Strive to achieve a higher spatial resolution. Data acquired at ≤50 m will reduce spatial mixing but will also require a higher saturation temperature to accurately resolve smaller, high temperature features.
Improve the temporal resolution to less than 1 day over all regions on Earth allows greater data volume at high-risk volcanoes and enables possible observation of more dynamic processes. Higher revisit times will also improve chance of cloud-free scenes. a Between the time of writing and editing based on the reviewer's comments, the work in this manuscript was presented to the SBG TIR Science Team. The results of the spectral study influenced them to adopt this recommendation. Now planned: the original 1.6 μm band has been eliminated in favor of adding a sixth TIR band at ∼10.2 μm. These changes are still being reviewed and are pending final approval by NASA HQ, but it is likely that SBG will now be a significant spectral improvement over all TIR instruments.

Data Availability Statement
All the data used in this study are available fully, openly, and without restrictions. MASTER data are available from NASA Jet Propulsion Laboratory (JPL) and the Oak Ridge National Laboratory Distributive Active Archive Center (ORNL DAAC) at https://asapdata.arc.nasa.gov/sensors/master/order.html. ECOSTRESS, ASTER, and MODIS data are available from NASA by selecting these instruments under the appropriate dropdown list at https://search.earthdata.nasa.gov/search/. GOES ABI data are available from the NOAA Comprehensive Large Array-data Stewardship System (CLASS) by selecting the ABI L1b Radiance Data and Clear Sky Mask product types at https://www.avl.class.noaa.gov/saa/products/search?sub_id=0&datatype_family=GRABIPRD&submit. x=25&submit.y=0. The Arizona State University Thermal Emission Spectroscopy Library is available at https:// speclib.asu.edu/. Registration is required to access the library but it is not restricted. The samples used in this study are: WAR-0024, WAR-5804, WAR-0579, BUR-3460, and BUR-4120.