ISS Observations Offer Insights Into Plant Function E

PUBLISHED: 22 JUNE 2017 | VOLUME: 1 | ARTICLE NUMBER: 0194 comment ISS observations offer insights into plant function E. Natasha Stavros, David Schimel, Ryan Pavlick, Shawn Serbin, Abigail Swann, Laura Duncanson, Joshua B. Fisher, Fabian Fassnacht, Susan Ustin, Ralph Dubayah, Anna Schweiger and Paul Wennberg

In 2018 technologies on the International Space Station will provide ~1 year of synchronous observations of ecosystem composition, structure and function. We discuss these instruments and how they can be used to constrain global models and improve our understanding of the current state of terrestrial ecosystems.

pace-based observations are index (a metric of vegetation greenness • How does land management interact increasingly central to global ecology, often used as a proxy for ecosystem with climate to control future disturbance Sopening windows to address questions productivity or carbon storage) influences regimes? pertaining to the carbon cycle, biodiversity, estimates of absorbed photosynthetic Common to all these questions is that productivity and disturbance. Until active radiation (fPAR) and leaf area they require multiple observations to test recently, satellite observations provided a index (a metric of one-side leaf area per specific hypotheses, and reject incorrect limited range of observations, albeit critical ground surface area)1. Thus, more direct model formulations. Fortunately, there are ones, relating light absorption to support observations of functioning and structure many exciting new developments in remote plant growth, disturbance and land use/ are needed to enable distinct testing and sensing that will provide key observations cover change. These observations have improvement of modelled representations of the land (for example, Sentinel series, revolutionized our knowledge of global of ecosystem processes. BIOMASS, NISAR). ecology, but have not addressed a series Questions include: Although new technology makes of critical questions about ecosystem • How sensitive are ecosystems to many key observations feasible, it’s process that must be resolved in order temperature and water availability and uncommon to have observing platforms to reduce uncertainty in future climate how will this affect future carbon and with multiple instruments collecting a and ecosystem service projections. These energy balances? variety of parameters at synergetic spatial key questions have been addressed in • Do diverse plant communities respond to resolutions and coordinated temporal local studies, but the findings of these climate change differently from simpler acquisition, particularly at scales needed local experiments cannot be tested across communities? for management. Yet this is exactly what broad landscapes with observations • Does diversity lead to ecosystems is needed to resolve how plant function from existing satellite technologies. For responding non-linearly to change, in affects ecosystem processes at landscape example, normalized difference vegetation ways we cannot predict now? scales in different regions across the globe.

Box 1 | Description of ISS instruments.

The instruments for deployment on ISS to produce observations in 25-metre OCO-3. A pointing instrument that are: GEDI, ECOSTRESS, OCO-3 and footprints along ten parallel tracks. The will use three high-spectral-resolution HISUI. Each instrument is being designed tracks are separated by ~600 m each, and grating spectrometers to measure solar- independently, but the nature of the footprints are separated by 25 m along induced fluorescence, a measure of observations offers an unprecedented the tracks. photon emission during photosynthetic opportunity to resolve regional and global partitioning that provides a direct proxy scale understanding of ecosystem process ECOSTRESS. Because transpiration for gross primary productivity, and

and function. performs the same cooling function as atmospheric column CO2, with high sweat, when plants close stomata their precision (±1 ppm). GEDI. A geodetic-class LIDAR that leaf temperatures increase. This rise in will measure canopy heights and foliar temperature can be detected and such HISUI. Will provide contiguous visible vertical profiles to establish a baseline of observations at different hours throughout to shortwave infrared (400–2,500 nm) vegetation structure and global terrestrial the day are crucial for mapping vegetation surface reflectance at very high spectral biomass. GEDI observations leverage a water-use efficiency, represented by the resolution (400–979 nm at ~10 nm sampling scheme that is then interpolated ability of vegetation to close stomata in the average per band for 57 bands and to produce a map. The sample scheme afternoon to avoid transpiration that would 900–2,500 nm at ~12.5 average per band uses three lasers (at 242 pulses per second) exceed water uptake from the soil. for 128 bands).

Table 1 | A list of high-level data products that will provided by each instrument expected for deployment to the ISS starting in 2017. Data ECOSTRESS OCO-3 GEDI HISUI product (40–60 m pixels, 20–30 samples (5 km2 footprint, capable of (~500 m2 footprint spaced (30 m pixels with 20 km swath and 10 nm spectral level per hour of the day collected mapping up to one hundred 60 m along a track with no resolution over 0.4–2.5 μm spectral range) throughout the year) 1,000 m2 areas per day) temporal repeat)

2 Surface temperature Atmospheric column CO2 Height metrics Atmospherically corrected surface reflectance with Emissivity Solar-induced fluorescence Canopy metrics quality assurance (not validated) 3 Evapotranspiration Gridded level 2 Gridded level 2 4 Water-use efficiency Aboveground biomass Evaporative stress index (footprint and gridded)

Fortunately, with a coordinated strategy, dynamics questions that cannot be answered on investment. Coordinated observations such an opportunity will arise for ecologists from any one instrument and that have can include using aircraft versions of these in the next few years. the ability to substantially enhance our spaceborne instruments (for example, understanding of ecosystem responses to LIDAR for GEDI, PhyTIR for ECOSTRESS, Novel Earth observations global change. For example, an estimation CFIS for OCO-3, and AVIRIS for HISUI) The International Space Station (ISS) will of carbon sink potential (Fig. 1) can be both before and after launch. Airborne host four instruments from JAXA (Japan estimated using observations of carbon campaigns with instruments analogous to Aerospace Exploration Agency) and NASA flux (OCO-3) and carbon storage (LIDAR- those deployed for space are frequently used (Box 1) in 2018 that will advance our ability derived biomass from GEDI), both of which for calibration and validation activities of to monitor and model terrestrial ecosystems can be affected by the efficiency of individual delivered data products both leading up between latitudes ~50° North and South. plant species (HISUI) to sequester carbon to and after launch. Furthermore, these Three of the instruments are from NASA and under variable access to water (ECOSTRESS). instruments may help to scale understanding include: the Global Ecosystem Dynamics Despite added information value from of ecological processes between ground-based Investigation (GEDI), the Ecosystem synergistic observations, the protocols and and spaceborne observations. Additional Spaceborne Thermal Radiometer Experiment measurement strategies for any one of these consideration for integration with airborne on Space Station (ECOSTRESS), and the instruments do not currently consider the (high spatial resolution) and ISS (unique Orbiting Carbon Observatory 3 (OCO-3). other instruments. Thus, there is an argument temporal acquisition) observations is the All three are being developed for deployment and need for collaboration during mission use of other spaceborne observations (for to the ISS in 2018 and will distribute the free development to coordinate observation example, Sentinel series and Landsat), as data products listed in Table 1. The fourth strategies and maximize scientific returns these can provide additional information for instrument, also for deployment in 2018, is the Hysperspectral Imager Suite (HISUI) ECOSTRESS GEDI from JAXA1,2. Complementing HISUI is the German Aerospace Agency (DLR) EnMAP3, OCO-3 HISUI Land surface Height Volume for which the observational plan has already temperature Leaf area index been established, thus it is not the focus Emissivity Surface of the imaging spectroscopy discussion Surface reﬂectance Spectral Water-use eciency henceforth. These four instruments will absorption roughness make observations of three-dimensional Biomass Species structure (GEDI) that can be used to scale biochemical signals4–6 to the canopy level Solar-induced ﬂuorescence Canopy Live vs dead (HISUI) and derive estimates of ecosystem Root depth functional composition, productivity (solar-induced traits Canopy water fluorescence from OCO-3), and water-use Atmospheric Carbon sink content efficiency (a product of ECOSTRESS) at column CO2 potential landscape scales. This instrument suite offers Nutrients a unique opportunity because the drifting Disturbance ISS platform and pointing capabilities of Carbon ecology Evapo- some of the instruments enable co-location residence transpiration time of high-spatial-resolution measurements Plant in space and acquisition times covering the Carbon-use temperature eciency and water diurnal cycle throughout the year (that is, stress more than 20 observations per hour of the Light- and nutrient-use day throughout the year). eciency Call for coincident observations Coordinating the spatial and temporal Figure 1 | Spatial and temporal synergy of observations and their applications. A pretzel diagram of coincidence of measurements from GEDI, observations (red text) from each instrument (coloured shapes) and the synergistic physical parameters ECOSTRESS, OCO-3 and HISUI would that can be derived (black text) when observations are taken at synchronous and complementary spatial be an opportunity to address ecosystem and temporal resolutions.

2 NATURE ECOLOGY & EVOLUTION 1, 0194 (2017) | DOI: 10.1038/s41559-017-0194 | www.nature.com/natecolevol ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved.

comment

ECOSTRESS 140 Santa Rita FLUXNET site OCO-3 120 Biweekly avg (1–14 Sep 2007) 0.5 100 5 GPP (gC m ) 0.4 –2 80 4 0.3 60 3 0.2 2 –2 d ET (W m 40 –2

0.1 ) 20 1 0 0 0 0 100 200 300 400 500 600 SIF Days since 1 Jun 2009 GPP 6 8 10 12 14 16 18 Hours ECOSTRESS FLUXNET Landsat-5

Terrestrial GEDI HISUI biosphere model Waveform amplitude 20 40 60 80 100 360 30

Highest reﬂecting surface height Parameterization 350 20 100 Needle−leaf Broadleaf Assimilation 340 RH75 10 80 Litter e (% ) RH50 RH100 Wood 60 Soil Elevation (m) RH25 tan c Benchmarking 330 40

0 Height above ground (m) eﬂe c

25 m R 20 Ground return 320 0 0.0 0.2 0.4 0.6 0.8 1.0 500 1,000 1,500 2,000 2,500 Model forecast Normalized cumulative return energy Wavelength (nm)

Figure 2 | A conceptual framework for how different data, observed from each ISS instrument, can be integrated into terrestrial biosphere models to improve their ability to represent and predict ecosystem processes. ECOSTRESS shows the 14-day average evapotranspiration (ET) from Santa Rita Fluxnet against the expected annual average value of ET from ECOSTRESS for each hour of the day. OCO-3 shows GOME solar-induced fluorescence data compared to the MPI-BGC gross primary productivity product. GEDI shows airborne LIDAR waveform from La Selva Biological Research Station, Costa Rica. HISUI shows an imaging spectroscopy/hyperspectral image cube and different spectra for different spectral classes. ISS image credit: Getty Images/scibak.

scaling understanding of ecosystem process ecosystems8,9. These models incorporate strategies that remain undetermined. and function. Thus, efforts to coordinate semi-empirical and mechanistic descriptions Specifically, which areas are of highest synergistic observations between the ISS of the physiological and biophysical processes priority to image given downlink constraints instruments should also consider the utility of and properties of terrestrial ecosystems that of instruments on the ISS? And, who is and coordinated efforts needed for airborne drive land–atmosphere fluxes and storage responsible for organizing this across campaigns and how their observations link of carbon, water and energy across space missions? Each instrument system and with existing and upcoming satellites. and time. In addition, terrestrial biosphere team is limited and constrained by their Organizing synchronous observations in models provide the mechanistic framework budgets and the mandate to deliver on space and time at varying scales of resolution necessary for integrating various types of their individual mission requirements will be most useful if future research uses data because they can represent ecological first. Thus, it will be necessary for the these data by integrating them with existing processes and functional responses at community to make additional efforts in information system infrastructure (that multiple spatial and temporal scales in a way order to seize this opportunity and take full is, models) in order to contextualize the that leverages our best understanding10. In advantage of this spectacular convergence information gleaned from the relatively order to improve terrestrial biosphere model of technical advances that offer a means for short period of overlapping observations parameterizations11, benchmark model constraining our mechanistic understanding (approximately one year). Specifically, future forecasts12, improve process representations, of ecological systems across scales suitable research can use the ISS observations to and evaluate alternative model structures, for management and regional analyses. establish a baseline for monitoring that can be models may need to be adjusted to integrate Importantly, although each instrument is used to attribute and improve predictions of the ISS observations at the relevant spatial largely developed and provides significant terrestrial ecosystem changes through time. and temporal scales for the processes and value independently, in aggregate, the To do so will require systematic evaluation functional responses of interest (Fig. 2). contemporaneous observations from all four of processes and ecosystem dynamics, which Although these flight projects can instruments could provide observations to can be done using models7. Terrestrial work together to coordinate observations support major advances for understanding biosphere models are the primary tool for strategies, there are many logistical issues the carbon and water cycles, and how quantifying the impacts of climate variability, (for example, explicit priority of observation ecosystem structure, function and diversity disturbance, and global change on terrestrial areas) for coordinating observational interact with them. ❐

NATURE ECOLOGY & EVOLUTION 1, 0194 (2017) | DOI: 10.1038/s41559-017-0194 | www.nature.com/natecolevol 3 ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. ©2017 Mac millan Publishers Li mited, part of Spri nger Nature. All ri ghts reserved. comment

E. Natasha Stavros, David Schimel, Ryan Pavlick 4. Asner, G. P. et al. Remote Sens. Environ. 158, 15–27 (2015). Department of Energy (grant no. DE-SC0012704) to and Joshua B. Fisher are in the Jet Propulsion 5. Singh, A. et al. Ecol. Appl. 25, 2180–2197 (2015). Brookhaven National Laboratory. R.D. is funded by NASA 6. Jetz, W. et al. Nat. Plants 2, 16024 (2016). Laboratory, California Institute of Technology, Global Ecosystem Dynamics Investigation Mission (grant 7. Moorcroft, P. Trends Ecol. Evol. 9, 3857–3974 (2006). no. NNL15AA03C). P.W. acknowledges support from Pasadena, California 91109, USA. Shawn Serbin 8. Fisher, J. B. et al. Annu. Rev. Environ. Resour. 39, 91–123 (2014). the NASA Carbon Cycle Science programme (grant no. 9. Dietze, M. C. & Latimer, A. M. in Encyclopedia of Theorestical is at Brookhaven National Laboratory, Upton, NNX14AI60G). J.F. is funded by the NASA ECOSTRESS Ecology 307–316 (2012). New York 11973-5000, USA. Abigail Swann is at 10. Dietze, M. C. et al. Plant Cell Environ. 36, 1575–1585 (2013). project. A.S. was funded by the National Science University of Washington, Seattle, Washington 11. Keenan, T. F. et al. Glob. Change Biol. 18, 2555–2569 (2012). Foundation (NSF) and the National Aeronautics and 98105, USA. Laura Duncanson is at NASA’s 12. Luo, Y. et al. Biogeosciences 9, 3857–3874 (2012). Space Administration (NASA) through the Dimensions of Biodiversity programme (DEB-1342872). Goddard Space Flight Center, Greenbelt, Maryland Acknowledgements 20771, USA. Fabian Fassnacht is at Karlsruhe Many of these ideas were born and developed at the Author contributions Institute of Technology, 76021 Karlsruhe, Germany. workshop on Exploring New Multi-Instrument Approaches E.N.S. is the lead writer for the manuscript, D.S. led Susan Ustin is at University of California, Davis, to Observing Terrestrial Ecosystems and the Carbon discussions and helped articulate the key points of discussion for the manuscript, R.P. provided HISUI and California 95616, USA. Ralph Dubayah is at Cycle from Space that occurred 5–9 October 2015 and was funded by the Keck Institute for Space Studies (KISS), OCO-3 information and panels in Fig. 2 as well as with University of Maryland, College Park, Maryland organized by Michelle Judd, and hosted at the institute edits for the manuscript, S.S. helped edit the general text, 20742, USA. Susan Ustin and Anna Schweiger are facility at California Institute of Technology in Pasadena, developed Fig. 2 and provided text for terrestrial biosphere at University of Minnesota, Saint Paul, Minnesota California. Additionally, the ideas in this article would models section, A.S. helped edit the general text and 55455, USA. Paul Wennberg is at California Institute not have been possible without the contributions of each provided text for terrestrial biosphere models sections, L.D. provided the GEDI panel in Fig. 2 and text describing of Technology, Pasadena, California 91125, USA. participant. The majority of the work was internally funded and carried out at the Jet Propulsion Laboratory, GEDI, J.B.F. helped with general editing and provided e-mail: [email protected] California Institute of Technology, under a contract with the ECOSTRESS panel in Fig. 2, F.F. helped with general editing, S.U. helped develop the original manuscript References the National Aeronautics and Space Administration. Also, outline, R.D., A.S. and P.W. were key in contributing ideas 1. Matsunaga, T. et al. in Geoscience and Remote Sensing Symp. S.S. was supported during the writing of this manuscript by for the manuscript. http://doi.org/b7d7 (IEEE, 2016). the Next-Generation Ecosystem Experiments (NGEE) in 2. Iwasaki, A. et al. in Geoscience and Remote Sensing Symp. the Tropics, which is supported by the Office of Biological http://doi.org/cpxfdf (IEEE, 2011). and Environmental Research in the Department of Competing interests 3. Guanter, L. et al. Remote Sens. 7, 8830–8857 (2015). Energy, Office of Science, and through the United States The authors declare no competing financial interests.