the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
08 Jan 2021
08 Jan 2021
Reviews and syntheses: Ongoing and emerging opportunities to improve environmental science using observations from the Advanced Baseline Imager on the Geostationary Operational Environmental Satellites
- 1Nelson Institute for Environmental Studies, University of Wisconsin – Madison, Madison, WI, USA
- 2Department of Biological Systems Engineering, University of Wisconsin – Madison, Madison, WI, USA
- 3Department of Atmospheric and Oceanic Sciences, University of Wisconsin – Madison, Madison, WI, USA
- 4Department of Land Resources and Environmental Sciences, Montana State University, Bozeman, MT, USA
- 5USDA ARS, Hydrology and Remote Sensing Laboratory, Beltsville, MD, USA
- 6Space Sciences and Engineering Center, University of Wisconsin – Madison, Madison, WI, USA
- 7Cooperative Institute for Meteorological Satellite Studies, University of Wisconsin – Madison, Madison, WI, USA
- 8Short-term Prediction Research and Transition Center, NASA Marshall Space Flight Center, Earth Science Branch, Huntsville, AL, USA
- 9Department of Electrical and Computer Engineering, Montana State University, Bozeman, MT, USA
- 10NASA Goddard Space Flight Center, Greenbelt, MD, 20771, USA
- 1Nelson Institute for Environmental Studies, University of Wisconsin – Madison, Madison, WI, USA
- 2Department of Biological Systems Engineering, University of Wisconsin – Madison, Madison, WI, USA
- 3Department of Atmospheric and Oceanic Sciences, University of Wisconsin – Madison, Madison, WI, USA
- 4Department of Land Resources and Environmental Sciences, Montana State University, Bozeman, MT, USA
- 5USDA ARS, Hydrology and Remote Sensing Laboratory, Beltsville, MD, USA
- 6Space Sciences and Engineering Center, University of Wisconsin – Madison, Madison, WI, USA
- 7Cooperative Institute for Meteorological Satellite Studies, University of Wisconsin – Madison, Madison, WI, USA
- 8Short-term Prediction Research and Transition Center, NASA Marshall Space Flight Center, Earth Science Branch, Huntsville, AL, USA
- 9Department of Electrical and Computer Engineering, Montana State University, Bozeman, MT, USA
- 10NASA Goddard Space Flight Center, Greenbelt, MD, 20771, USA
Abstract. Environmental science is increasingly reliant on remotely-sensed observations of the Earth's surface and atmosphere. Observations from polar-orbiting satellites have long supported investigations on land cover change, ecosystem productivity, hydrology, climate, the impacts of disturbance, and more, and are critical for extrapolating (upscaling
) ground-based measurements to larger areas. However, the limited temporal frequency at which polar-orbiting satellites observe the Earth limits our understanding of rapidly evolving ecosystem processes, especially in areas with frequent cloud cover. Geostationary satellites have observed the Earth's surface and atmosphere at high temporal frequency for decades, and their imagers now have spectral resolutions in the visible and near-infrared regions that are comparable to commonly-used polar-orbiting sensors like the Moderate Resolution Imaging Spectroradiometer (MODIS), Visible Infrared Imaging Radiometer Suite (VIIRS), or Landsat. These advances extend applications of geostationary Earth observations from weather monitoring to multiple disciplines in ecology and environmental science. We review a number of existing applications that use data from geostationary platforms and present upcoming opportunities for observing key ecosystem properties using high-frequency observations from the Advanced Baseline Imagers (ABI) on the Geostationary Operational Environmental Satellites (GOES), which routinely observe the Western Hemisphere every 5–15 minutes. Many of the existing applications in environmental science from ABI are focused on estimating land surface temperature, solar radiation, evapotranspiration, and biomass burning emissions along with detecting rapid drought development and wildfire. Ongoing work in estimating vegetation properties and phenology from other geostationary platforms demonstrates the potential for expanding ABI observations to estimate vegetation greenness, moisture, and productivity at high temporal frequency across the Western Hemisphere. Finally, we present emerging opportunities to address the relatively coarse resolution of ABI observations through multi-sensor fusion to resolve landscape heterogeneity and to leverage observations from ABI to study the carbon cycle and ecosystem function at unprecedented temporal frequency.
Anam M. Khan et al.
Status: final response (author comments only)
-
RC1: 'Comment on bg-2020-454', Anonymous Referee #1, 15 Feb 2021
The article revisits a number of applications of geostationary-based observations in environmental sciences, focusing primarily on the ABI instrument on-board the current generation of GOES satellites. Geostationary platforms provide high-frequency observations, which are widely used in Meteorology and Nowcasting. However, other fields that value more spatial resolution, even if at the expense of revisiting times, as is often the case of environmental monitoring applications, greatly underexplore those observations. Given the high quality of the instruments on the current geostationary satellites – such as the GOES-R series, but also those operated by European and Asian agencies – I find this article to be very timely. I think that aspects related with the actually added-value of high frequency observations, synergies with observations from other geostationary and/or polar-orbiting platforms should be further revised. I consider the manuscript should be subject to a moderate revision before being accepted for publication.
1) High frequency observations may be of great use for many different reasons: to follow rapidly evolving phenomena, such as weather systems; for surface monitoring purposes, high frequent observations in visible/infrared bands may be important in areas which are frequently cloud-covered, by increasing the probability of acquiring clear-sky observations; variables with strong diurnal cycles may not be properly represented by low-frequency observations; identification of rapid changes, such as phenology. Although the article touches some of these aspects, I think they should be further systematized and clearer with further examples of applications where GOES-R observations play a significant role.
2) It is acknowledged that some of the issues that one needs to deal with when using ABI or other similar instruments on geostationary satellites to monitor the Earth’s surface are well covered – parallax, atmospheric correction, increasing pixel size towards the disk’s edge. The article then lists a number of products that may be derived from ABI followed by a few examples of relevant applications of those data. Here, I suggest that the actual added-value of ABI observations/products with respect to other commonly used data should be further highlighted, namely:
2.1) the use of GOES/ABI observations and products, including LST, to derive evapotranspiration is well-known. Here I would expect to see an example of such an application including the input satellite data/products and respective evaporative fluxes. Showing the on-set of drought conditions highlighting the advantages of high frequency data would be of interest to users.
2.2) Landscape fires usually present significant diurnal cycles and therefore high frequency observations are better positioned for their full characterization. However, small agricultural fires aimed at burning agricultural waste, are known to cause a significant amount of emissions and pollution, but often these are too small and weak to be “seen” by the relatively coarse geostationary satellites. This is especially the case as we move away from the sub-satellite point, as the minimum detectable fire radiative power scales with pixel area. Please consider these problems and address the potential benefits of the combined use of ABI/GEO’s observations with higher spatial resolution.
2.3) the coarse resolution of vegetation products/indices derived from geostationary observations is a major caveat as often pointed by users, especially when compared with the same type of variables derived from MODIS or VIIRS. The phenology example presented is relevant application, but again the actual advantage of ABI with respect to other more commonly used data is far from being clear. The authors show two cases presenting post-fire disturbances and vegetation recovery in California. However, the comparison with MODIS is only shown for one of them; the time-series do not cover the full growing season. Furthermore, I would say that the most relevant contribution of geostationary/ABI observations would be to map phenological cycles in a consistent way (given the high number of observations) over large areas. So although zooming at a relatively small area affected by fires may be of interest, I think that that is not the most appropriate example to illustrate the benefits of those types of observations, especially when compared with others.
2.4) Following the above, it should be noted that only TOA vegetation indices are shown, while several atmospherically corrected and BRDF-corrected vegetation properties can be also derived from both ABI (geostationary) and MODIS or VIIRS (polar-orbiting) instruments. Why not using those types of indices – in principle less noisy – than TOA values?
3) The combined use of multiple geostationary satellites is only very briefly mentioned. Given the high quality of current sensors, there’s a great potential to build nearly global high frequency fields and derived products. The combination of geostationary data with high spatial resolution observations (from MODIS/VIIRS to Sentinel-2/Landsat) targets a different set of users and applications, with the latter usually more focused on regional-to-local applications. Both global (or large scale) and regional/local scale applications are relevant and worth addressing.
4) I would encourage the authors to improve the examples given in section 5 and to better link these to section 6. Evapotranspiration and energy fluxes are deeply linked to plant’s carbon up-take and ecosystem dynamics. A brief reference is already made to “diurnal land attributes” – I suggest the authors further explain what is meant by that and how ABI observations can address those.5) Editorial
- line 51 – please clarify what is meant by “diurnal behavior of land surface function”.
- line 280 – by “effective surface temperature” do the authors mean brightness temperature? I find it unclear what the authors are mean as a physical surface temperature. Whatever temperature you consider, it will most likely have a physical meaning, being radiometric temperature, aerodynamic temperature, brightness temperature, …
- line 314 – S. Ha et al, 2020 ?
- lines 379-380 – this is not shown here, sicne the time-series does not cover April.
-line 391 – please refer to European satellites as “Meteosat” instead of METEOSAT.
- line 449 – Fig. 2 is called after Figs 3 and 4 – so why this figure order?-
AC1: 'Reply on RC1', Anam Khan, 13 Mar 2021
The article revisits a number of applications of geostationary-based observations in environmental sciences, focusing primarily on the ABI instrument on-board the current generation of GOES satellites. Geostationary platforms provide high-frequency observations, which are widely used in Meteorology and Nowcasting. However, other fields that value more spatial resolution, even if at the expense of revisiting times, as is often the case of environmental monitoring applications, greatly under explore those observations. Given the high quality of the instruments on the current geostationary satellites – such as the GOES-R series, but also those operated by European and Asian agencies – I find this article to be very timely. I think that aspects related with the actually added-value of high frequency observations, synergies with observations from other geostationary and/or polar-orbiting platforms should be further revised. I consider the manuscript should be subject to a moderate revision before being accepted for publication.
Author response: We want to thank Referee 1 for the time they took to complete a thoughtful and insightful review of our manuscript. We are glad that our manuscript is timely for the environmental science community given the recent updates to geostationary imagers.
1) High frequency observations may be of great use for many different reasons: to follow rapidly evolving phenomena, such as weather systems; for surface monitoring purposes, high frequent observations in visible/infrared bands may be important in areas which are frequently cloud-covered, by increasing the probability of acquiring clear-sky observations; variables with strong diurnal cycles may not be properly represented by low-frequency observations; identification of rapid changes, such as phenology. Although the article touches some of these aspects, I think they should be further systematized and clearer with further examples of applications where GOES-R observations play a significant role.
Author response: Thank you for this point. Our manuscript continuously points to the advantage of the availability of clear observations throughout the year in almost every section where the point is applicable. We will include some more specific examples for the importance of capturing rapid changes during phenological transitions. There are some existing figures from various papers that we cite that illustrate the advantage that geostationary imagers provide for phenology. In the revised manuscript we will specifically note these key figures in the interest of comprehensively reviewing the existing literature.
2) It is acknowledged that some of the issues that one needs to deal with when using ABI or other similar instruments on geostationary satellites to monitor the Earth’s surface are well covered – parallax, atmospheric correction, increasing pixel size towards the disk’s edge. The article then lists a number of products that may be derived from ABI followed by a few examples of relevant applications of those data. Here, I suggest that the actual added-value of ABI observations/products with respect to other commonly used data should be further highlighted, namely:
2.1) the use of GOES/ABI observations and products, including LST, to derive evapotranspiration is well-known. Here I would expect to see an example of such an application including the input satellite data/products and respective evaporative fluxes. Showing the on-set of drought conditions highlighting the advantages of high frequency data would be of interest to users.
Author response: We will use the example of detecting flash droughts and further elaborate on this application to illustrate the advantage of thermal observations with high temporal resolution. We will also include a more detailed explanation of the process of going from the input required data to the output product and finally the product’s use in detecting drought conditions.
2.2) Landscape fires usually present significant diurnal cycles and therefore high frequency observations are better positioned for their full characterization. However, small agricultural fires aimed at burning agricultural waste, are known to cause a significant amount of emissions and pollution, but often these are too small and weak to be “seen” by the relatively coarse geostationary satellites. This is especially the case as we move away from the sub-satellite point, as the minimum detectable fire radiative power scales with pixel area. Please consider these problems and address the potential benefits of the combined use of ABI/GEO’s observations with higher spatial resolution.
Author response: Thank you for this point. The WildFire Automated Biomass Burning Algorithm for GOES is an effort to detect sub-pixel fires and estimate their sub-pixel area and temperature. We will note that GOES is more suitable for detecting larger fires and discuss the importance of detecting small fires in burned area estimates as recently discussed in Ramo et al. (2021). We will also address opportunities for spatiotemporal fusion between ABI and higher spatial resolution for potentially improving the detection of very small sub-pixel fires.
2.3) the coarse resolution of vegetation products/indices derived from geostationary observations is a major caveat as often pointed by users, especially when compared with the same type of variables derived from MODIS or VIIRS. The phenology example presented is relevant application, but again the actual advantage of ABI with respect to other more commonly used data is far from being clear. The authors show two cases presenting post-fire disturbances and vegetation recovery in California. However, the comparison with MODIS is only shown for one of them; the time-series do not cover the full growing season. Furthermore, I would say that the most relevant contribution of geostationary/ABI observations would be to map phenological cycles in a consistent way (given the high number of observations) over large areas. So although zooming at a relatively small area affected by fires may be of interest, I think that that is not the most appropriate example to illustrate the benefits of those types of observations, especially when compared with others.
Author response: Interestingly, the spatial scale of many ABI bands is similar to many MODIS bands but we all agree that finer spatial detail is beneficial for remote sensing applications. We can add a comparison on MODIS on all time series. The time series includes all seasons although they are not in the usual order from Jan - Dec. Instead the time series runs from March 2019 to March 2020. This was done to capture the time before and after the Kincade fire. We can expand on some phenology examples published in the literature to highlight the advantage of using geostationary observations in capturing phenological transitions.
2.4) Following the above, it should be noted that only TOA vegetation indices are shown, while several atmospherically corrected and BRDF-corrected vegetation properties can be also derived from both ABI (geostationary) and MODIS or VIIRS (polar-orbiting) instruments. Why not using those types of indices – in principle less noisy – than TOA values?
Author response: We have written about various efforts currently underway for atmospheric correction of TOA observations from geostationary imagers in the paper. The purpose of sticking to TOA MODIS values was to compare TOA ABI values with TOA MODIS values. We can add NDVI values from the existing MODIS NDVI product which is generated from atmospherically corrected data as an added comparison.
3) The combined use of multiple geostationary satellites is only very briefly mentioned. Given the high quality of current sensors, there’s a great potential to build nearly global high frequency fields and derived products. The combination of geostationary data with high spatial resolution observations (from MODIS/VIIRS to Sentinel-2/Landsat) targets a different set of users and applications, with the latter usually more focused on regional-to-local applications. Both global (or large scale) and regional/local scale applications are relevant and worth addressing.
Author response: Thank you for pointing this out. We will look for some more examples in the literature that combine geostationary sensors and elaborate on how this can be advantageous in other applications covered in our paper.
4) I would encourage the authors to improve the examples given in section 5 and to better link these to section 6. Evapotranspiration and energy fluxes are deeply linked to plant’s carbon up-take and ecosystem dynamics. A brief reference is already made to “diurnal land attributes” – I suggest the authors further explain what is meant by that and how ABI observations can address those.
Author response: We can link section 6 and section 5 more completely by elaborating on the application of geostationary observations for studying carbon-water-energy fluxes together, as alluded to in Anderson et al. (2000). This link can be drawn through stomatal conductance which is the linking mechanism between assimilation and transpiration and in section 6.1, we can discuss models that can estimate canopy latent heat and assimilation fluxes with stomatal, boundary layer, and aerodynamic resistances (Anderson et al. 2000)
5) Editorial
Author response: Thank you for the careful read, we will address all of these important editorial details in the revised manuscript.
- line 51 – please clarify what is meant by “diurnal behavior of land surface function”.
- line 280 – by “effective surface temperature” do the authors mean brightness temperature? I find it unclear what the authors are mean as a physical surface temperature. Whatever temperature you consider, it will most likely have a physical meaning, being radiometric temperature, aerodynamic temperature, brightness temperature, …
- line 314 – S. Ha et al, 2020?
- lines 379-380 – this is not shown here, sicne the time-series does not cover April.
-line 391 – please refer to European satellites as “Meteosat” instead of METEOSAT.
- line 449 – Fig. 2 is called after Figs 3 and 4 – so why this figure order?
Author response references
Anderson, M. C., Norman, J. M., Meyers, T. P., and Diak, G. R.: An analytical model for estimating canopy transpiration and carbon assimilation fluxes based on canopy light-use efficiency, Agricultural and Forest Meteorology, 101, 265–289, https://doi.org/10.1016/S0168-1923(99)00170-7, 2000.
Ramo, R., Roteta, E., Bistinas, I., Wees, D. van, Bastarrika, A., Chuvieco, E., and Werf, G. R. van der: African burned area and fire carbon emissions are strongly impacted by small fires undetected by coarse resolution satellite data, PNAS, 118, https://doi.org/10.1073/pnas.2011160118, 2021.
-
AC1: 'Reply on RC1', Anam Khan, 13 Mar 2021
-
RC2: 'Comment on bg-2020-454', Anonymous Referee #2, 16 Feb 2021
General comments:
This is a well written and valuable synthesis of the emerging and potential uses of new geostationary sensors. I think it will be a useful reference for a broad audience of ecologists and environmental scientists, especially those who perhaps already use remote sensing but haven’t yet added geostationary sensors to their toolbox. It certainly spurred new ideas for me.
I really only have one general comment (though see below for some more specific and technical comments): is this a review specifically of ABI, or of the new generation of geostationary sensors more broadly? The title states ABI, but there are many places throughout the article that seem much more general than the title implies, even places (especially in sections 3 and 4) that don’t seem tied to geostationary sensors at all but are more general overviews of broadly applicable issues in remote sensing (e.g., descriptions of atmospheric correction, vegetation indices, and derivation of incident PAR that could use more specifics on how these would be applied to and/or benefit from geostationary observations and what the challenges would be in their derivation/use). My main suggestion would just be to tighten up this focus a bit throughout. (I have a couple of additional specific suggestions below on a few specific places where this would be helpful.)
Specific comments:
-Line 112: I might choose a different heading here. This section seems less about working with ABI data (which I had initially interpreted as being from the perspective of a typical environmental science user) than about some of the issues and uncertainties in ABI data that need to be corrected before you can even start working with it.
-Section 3.3.2: This section is kind of general and could be more specifically focused on LST from ABI (or others of the newer generation of geostationary sensors).
-Lines 231-232: It might also be worth noting that EVI has a soil background correction factor built in that may also make it more suitable for open canopy systems (Huete et al., 2002).
-Lines 356-358: There’s a really good article from Zhang et al. (2009) on the impact of temporal resolution on vegetation phenology retrieval (as well as the impacts of missing data around vegetation transitions) that might be useful here. It might be worth expanding a bit on where and how much geostationary data could improve on the composite data (typically 8- to 16-day) that are often currently used for vegetation phenology retrieval. The authors specifically mention how it could be useful where cloudy conditions occur around seasonal transitions, but another thing to highlight might be the utility of geostationary observations in some dryland systems, where phenological transitions can occur very rapidly and unpredictably in response to rainfall pulses (Smith et al., 2019), which might be either missed by polar orbiting sensors or washed out in the composite. Monitoring phenology in these regions has definitely been a long-standing problem (e.g., the White et al. 2009 and Ganguly et al. 2010 articles that are already cited by the authors). Of course, there are other challenges aside from temporal resolution that make dryland phenology difficult to model/detect (Smith et al., 2019), but any potential improvement from geostationary observations in these regions could be worth noting.
-Line 416: I think this should say “linear relationship between absorbed PAR (APAR) and _gross_ primary production,” right? The MODIS algorithm first models GPP at an 8-day frequency then estimates annual NPP as the annual integral of GPP minus the annual integral of autotrophic respiration (modeled from allometric relationships and air temperature).
-Lines 417-420: There are a couple (admittedly nitpicky) details of the MODIS algorithm that I think are incorrect here and should be double-checked. First, in the operational MODIS GPP/NPP product, NDVI is not the primary way that APAR is estimated (though in principle, it certainly could be). APAR is estimated from the MODIS FPAR product (MOD15), which is itself based primarily on an inversion of a radiative transfer model, with the NDVI~FPAR relationship only being used as a backup algorithm in case the full inversion fails (Knyazhikin et al. 1999; Myneni et al. 2002). Second, and this is merely a technicality, there are five (not three) biome-specific parameters in the model since Tscale and Wscale require two parameters each: a lower Tmin/VPD threshold and an upper Tmin/VPD threshold. (If you also include the allometry and Q10 parameters needed for the NPP estimates, then there are even more than five.)
-Lines 429-432: Maybe here, or maybe elsewhere, it might be worth noting possible synergies with other satellite-based sensors. Here, for example, I’m thinking of possible synergies with soil moisture estimates from microwave sensors (e.g., SMAP or AMSR-E/AMSR-2) given the widespread importance of soil moisture for primary production and the known deficiencies in existing LUE models’ ability to represent soil moisture stress (Stocker et al. 2018, 2019).
-Lines 434-435: This is very true, and worth noting that there is at least one relatively new LUE model that does indeed include effects of diffuse irradiance (Zhang et al., 2016).
-Lines 439-441: Could the authors expand on this a bit?
-Section 6.2: Since recovery from disturbance is a relatively slow process (occurring mostly on the time scale of weeks to years, I imagine?), it’s unclear to me what geostationary brings to the table that we’re not already getting from polar-orbiting sensors like MODIS, especially given the challenges raised by the authors in lines 451-454. Could the authors expand more on why geostationary observations would be useful in this regard and how they would complement or expand on the capabilities already available from polar-orbiting sensors?
-Section 6.3: This is a really interesting section, and something that I hadn’t really considered before. I think it could be more specifically tied to the article’s focus on geostationary observations, though. Is this not something that can be done with existing polar-orbiting satellites (e.g., the morning/afternoon/nighttime overpasses of Terra/Aqua)? How specifically could you envision geostationary observations contributing to this?
Editorial suggestions/corrections:
-Line 113: I would suggest changing “certain” to “given” and deleting “on the Earth”.
-Line 131: This should be section 3.2, not 3.1.
References:
Huete, A., Didan, K., Miura, T., Rodriguez, E. P., Gao, X., & Ferreira, L. G. (2002). Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sensing of Environment, 83(1–2), 195–213.
Knyazikhin, Y., Glassy, J., Privette, J. L., Tian, Y., Lotsch, A., Zhang, Y., et al. (1999). MODIS Leaf Area Index (LAI) and Fraction of Photosynthetically Active Radiation Absorbed by Vegetation (FPAR) Product (MOD15) Algorithm Theoretical Basis Document, https://modis.gsfc.nasa.gov/data/atbd/atbd_mod15.pdf.
Myneni, R. B., Hoffman, S., Knyazikhin, Y., Privette, J. L., Glassy, J., Tian, Y., et al. (2002). Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data. Remote Sensing of Environment, 83, 214–231.
Smith, W. K., Dannenberg, M. P., Yan, D., Herrmann, S., Barnes, M. L., Barron-Gafford, G. A., et al. (2019). Remote sensing of dryland ecosystem structure and function: Progress, challenges, and opportunities. Remote Sensing of Environment, 233.
Stocker, B. D., Zscheischler, J., Keenan, T. F., Prentice, I. C., Peñuelas, J., & Seneviratne, S. I. (2018). Quantifying soil moisture impacts on light use efficiency across biomes. New Phytologist, 218(4), 1430–1449.
Stocker, B. D., Zscheischler, J., Keenan, T. F., Prentice, I. C., Seneviratne, S. I., & Peñuelas, J. (2019). Drought impacts on terrestrial primary production underestimated by satellite monitoring. Nature Geoscience, 12(4), 264–270.
Zhang, X., Friedl, M. A., & Schaaf, C. B. (2009). Sensitivity of vegetation phenology detection to the temporal resolution of satellite data. International Journal of Remote Sensing, 30(8), 2061–2074.
Zhang, Y., Song, C., Sun, G., Band, L. E., McNulty, S., Noormets, A., et al. (2016). Development of a coupled carbon and water model for estimating global gross primary productivity and evapotranspiration based on eddy flux and remote sensing data. Agricultural and Forest Meteorology, 223, 116–131.
-
AC2: 'Reply on RC2', Anam Khan, 13 Mar 2021
General comments:
This is a well written and valuable synthesis of the emerging and potential uses of new geostationary sensors. I think it will be a useful reference for a broad audience of ecologists and environmental scientists, especially those who perhaps already use remote sensing but haven’t yet added geostationary sensors to their toolbox. It certainly spurred new ideas for me.
I really only have one general comment (though see below for some more specific and technical comments): is this a review specifically of ABI, or of the new generation of geostationary sensors more broadly? The title states ABI, but there are many places throughout the article that seem much more general than the title implies, even places (especially in sections 3 and 4) that don’t seem tied to geostationary sensors at all but are more general overviews of broadly applicable issues in remote sensing (e.g., descriptions of atmospheric correction, vegetation indices, and derivation of incident PAR that could use more specifics on how these would be applied to and/or benefit from geostationary observations and what the challenges would be in their derivation/use). My main suggestion would just be to tighten up this focus a bit throughout. (I have a couple of additional specific suggestions below on a few specific places where this would be helpful.)
Author response: We would like to thank Referee 2 for their comments and review of our manuscript. We are glad that our manuscript could serve as a useful reference for ecologists and environmental scientists and excited that our manuscript can add inspiration for users of remote sensing. Our review is meant to focus on the ABI as we felt that it is a largely untapped resource for environmental science in the Western Hemisphere while other geostationary imagers such as SEVIRI already offer various products for environmental science. When we refer to other geostationary imagers, our intention was to reference the existing work with geostationary imagers that could be extended and/or further developed with ABI while not excluding other imagers that can be combined to create near-global coverage. The purpose of reviewing the more general issues of atmospheric corrections and vegetation indices was to explain some fundamentals of remote sensing to offer a starting point in expanding the use of ABI. We believe each section has specific references to ABI (the current efforts in each section being carried out for ABI and/or the issues that might arise with ABI) or other geostationary sensors when an example for ABI was not found, but we will expand on some more ABI specific details.
Specific comments:
-Line 112: I might choose a different heading here. This section seems less about working with ABI data (which I had initially interpreted as being from the perspective of a typical environmental science user) than about some of the issues and uncertainties in ABI data that need to be corrected before you can even start working with it.
Author response: Thank you for this suggestion. We will change the title of the section to “Preprocessing ABI data” or similar following discussions with coauthors.
-Section 3.3.2: This section is kind of general and could be more specifically focused on LST from ABI (or others of the newer generation of geostationary sensors).
Thank you, we wrote about LST and specifically the GOES LST algorithm in section 4.4. We will move the entire discussion of retrieving land surface temperature with land emissivity estimates to section 4.4. We can keep section 3.3.2 focused on the removal of atmospheric effects on thermal band data.
-Lines 231-232: It might also be worth noting that EVI has a soil background correction factor built in that may also make it more suitable for open canopy systems (Huete et al., 2002).
Author response: Thank you for pointing this out. We will include this advantage of EVI.
-Lines 356-358: There’s a really good article from Zhang et al. (2009) on the impact of temporal resolution on vegetation phenology retrieval (as well as the impacts of missing data around vegetation transitions) that might be useful here. It might be worth expanding a bit on where and how much geostationary data could improve on the composite data (typically 8- to 16-day) that are often currently used for vegetation phenology retrieval. The authors specifically mention how it could be useful where cloudy conditions occur around seasonal transitions, but another thing to highlight might be the utility of geostationary observations in some dryland systems, where phenological transitions can occur very rapidly and unpredictably in response to rainfall pulses (Smith et al., 2019), which might be either missed by polar orbiting sensors or washed out in the composite. Monitoring phenology in these regions has definitely been a long-standing problem (e.g., the White et al. 2009 and Ganguly et al. 2010 articles that are already cited by the authors). Of course, there are other challenges aside from temporal resolution that make dryland phenology difficult to model/detect (Smith et al., 2019), but any potential improvement from geostationary observations in these regions could be worth noting.
Author response: Thank you for these suggestions, we were unaware of the Zhang and Smith references and will incorporate these ideas into the revised manuscript.
-Line 416: I think this should say “linear relationship between absorbed PAR (APAR) and _gross_ primary production,” right? The MODIS algorithm first models GPP at an 8-day frequency then estimates annual NPP as the annual integral of GPP minus the annual integral of autotrophic respiration (modeled from allometric relationships and air temperature).
Author response: In principle we agree but there are a couple of notable differences for how the algorithm works in practice. According to Monteith’s original logic, the linear relationship is specified between APAR and aboveground NPP (Medlyn, 1998). According to the MOD17 User’s Guide, NPP is first simulated with the BIOME-BGC ecosystem model and epsilon which is the conversion efficiency under ideal environmental conditions is estimated between APAR and NPP (Running and Zhao, 2015).
-Lines 417-420: There are a couple (admittedly nitpicky) details of the MODIS algorithm that I think are incorrect here and should be double-checked. First, in the operational MODIS GPP/NPP product, NDVI is not the primary way that APAR is estimated (though in principle, it certainly could be). APAR is estimated from the MODIS FPAR product (MOD15), which is itself based primarily on an inversion of a radiative transfer model, with the NDVI~FPAR relationship only being used as a backup algorithm in case the full inversion fails (Knyazhikin et al. 1999; Myneni et al. 2002). Second, and this is merely a technicality, there are five (not three) biome-specific parameters in the model since Tscale and Wscale require two parameters each: a lower Tmin/VPD threshold and an upper Tmin/VPD threshold. (If you also include the allometry and Q10 parameters needed for the NPP estimates, then there are even more than five.)
Author response: Thank you for pointing out these important details. We will add that the calculation of Wscale, Tscale and respiration require additional biome-specific parameters. We will also add the updated MODIS FPAR (MOD15) input for APAR to clearly delineate the line of reasoning that leads to the MODIS carbon cycle products.
-Lines 429-432: Maybe here, or maybe elsewhere, it might be worth noting possible synergies with other satellite-based sensors. Here, for example, I’m thinking of possible synergies with soil moisture estimates from microwave sensors (e.g., SMAP or AMSR-E/AMSR-2) given the widespread importance of soil moisture for primary production and the known deficiencies in existing LUE models’ ability to represent soil moisture stress (Stocker et al. 2018, 2019).
Author response: Thank you, we will add some details about using space-based soil moisture estimates in efforts to improve representation of soil moisture limitations in GPP estimates that use satellite data and refer to modeling techniques that have been able to capture the impacts of soil moisture on carbon assimilation (Anderson et al. 2000).
-Lines 434-435: This is very true, and worth noting that there is at least one relatively new LUE model that does indeed include effects of diffuse irradiance (Zhang et al., 2016).
Author response: Thank you for this reference. We will refer to it in our discussion on including the effects of diffuse radiation on GPP in models.
-Lines 439-441: Could the authors expand on this a bit?
-Section 6.2: Since recovery from disturbance is a relatively slow process (occurring mostly on the time scale of weeks to years, I imagine?), it’s unclear to me what geostationary brings to the table that we’re not already getting from polar-orbiting sensors like MODIS, especially given the challenges raised by the authors in lines 451-454. Could the authors expand more on why geostationary observations would be useful in this regard and how they would complement or expand on the capabilities already available from polar-orbiting sensors?
Author response: We will expand on the potentially increased availability of cloud-free imagery before and after disturbances to map immediate damage to ecosystems after a disturbance. We can also expand on the complementary use of high temporal resolution with polar-orbiting sensors to capture recovery trajectories with more detail and to capture the impact of short-term resource pulses on the trajectory of recovery from disturbance.
-Section 6.3: This is a really interesting section, and something that I hadn’t really considered before. I think it could be more specifically tied to the article’s focus on geostationary observations, though. Is this not something that can be done with existing polar-orbiting satellites (e.g., the morning/afternoon/nighttime overpasses of Terra/Aqua)? How specifically could you envision geostationary observations contributing to this?
Author response: Thank you for the kind words. We will expand on the ability of diurnal measurements of land surface temperature to gain insight into ecosystem thermodynamics, which often involves integrals of an observable over a given time interval (for example when calculating the thermal response number). You are correct in that Terra/Aqua can make important LST observations given their overpass times and we will expand this section to note that geostationary satellites are not unique in this regard; rather that their hypertemporal measurement capabilities open new avenues for satellite research. We would also like to point out that ABI is expected to be around for a longer time compared to Terra/Aqua and future geostationary missions like GEO-XO will be able to provide long term and consistent data.
Editorial suggestions/corrections:
-Line 113: I would suggest changing “certain” to “given” and deleting “on the Earth”.
-Line 131: This should be section 3.2, not 3.1.
Author response references
Anderson, M. C., Norman, J. M., Meyers, T. P., and Diak, G. R.: An analytical model for estimating canopy transpiration and carbon assimilation fluxes based on canopy light-use efficiency, Agricultural and Forest Meteorology, 101, 265–289, https://doi.org/10.1016/S0168-1923(99)00170-7, 2000.
Medlyn, B. E.: Physiological basis of the light use efficiency model, Tree Physiology, 18, 167–176, https://doi.org/10.1093/treephys/18.3.167, 1998.
Running, S. W. and Zhao, M.: Daily GPP and Annual NPP (MOD17A2/A3) Products NASA Earth Observing System MODIS Land Algorithm, 28, 2015.
Referee references:
Huete, A., Didan, K., Miura, T., Rodriguez, E. P., Gao, X., & Ferreira, L. G. (2002). Overview of the radiometric and biophysical performance of the MODIS vegetation indices. Remote Sensing of Environment, 83(1–2), 195–213.
Knyazikhin, Y., Glassy, J., Privette, J. L., Tian, Y., Lotsch, A., Zhang, Y., et al. (1999). MODIS Leaf Area Index (LAI) and Fraction of Photosynthetically Active Radiation Absorbed by Vegetation (FPAR) Product (MOD15) Algorithm Theoretical Basis Document, https://modis.gsfc.nasa.gov/data/atbd/atbd_mod15.pdf.
Myneni, R. B., Hoffman, S., Knyazikhin, Y., Privette, J. L., Glassy, J., Tian, Y., et al. (2002). Global products of vegetation leaf area and fraction absorbed PAR from year one of MODIS data. Remote Sensing of Environment, 83, 214–231.
Smith, W. K., Dannenberg, M. P., Yan, D., Herrmann, S., Barnes, M. L., Barron-Gafford, G. A., et al. (2019). Remote sensing of dryland ecosystem structure and function: Progress, challenges, and opportunities. Remote Sensing of Environment, 233.
Stocker, B. D., Zscheischler, J., Keenan, T. F., Prentice, I. C., Peñuelas, J., & Seneviratne, S. I. (2018). Quantifying soil moisture impacts on light use efficiency across biomes. New Phytologist, 218(4), 1430–1449.
Stocker, B. D., Zscheischler, J., Keenan, T. F., Prentice, I. C., Seneviratne, S. I., & Peñuelas, J. (2019). Drought impacts on terrestrial primary production underestimated by satellite monitoring. Nature Geoscience, 12(4), 264–270.
Zhang, X., Friedl, M. A., & Schaaf, C. B. (2009). Sensitivity of vegetation phenology detection to the temporal resolution of satellite data. International Journal of Remote Sensing, 30(8), 2061–2074.
Zhang, Y., Song, C., Sun, G., Band, L. E., McNulty, S., Noormets, A., et al. (2016). Development of a coupled carbon and water model for estimating global gross primary productivity and evapotranspiration based on eddy flux and remote sensing data. Agricultural and Forest Meteorology, 223, 116–131.
-
AC2: 'Reply on RC2', Anam Khan, 13 Mar 2021
Anam M. Khan et al.
Anam M. Khan et al.
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 453 | 143 | 14 | 610 | 1 | 4 |
- HTML: 453
- PDF: 143
- XML: 14
- Total: 610
- BibTeX: 1
- EndNote: 4
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1