There are a lot of new and upcoming spaceborne capabilities capturing different parts of plant stress, each with their own strengths, limitations, and unique characteristics. It could be beneficial to evaluate our ability to identify plant water stress with our range of options.
Before determining how well we can detect plant stress from space, we need to define: what is plant stress? While definitions vary widely, we can think of it as sub-optimal controls on plant growth, which may be broadly grouped into light, temperature, water, CO₂, nutrients, and disturbance. At Hydrosat we take into account what our Science Lead, Josh Fisher, coined as the ‘goldilocks zone’ for all of these, where too much or too little is not good. It should be considered that even a bit of disturbance such as wind or even fire can be good (while not necessarily for individual growth; but, good for community growth and dispersal)! Here, we focus on key players to water stress; specifically, too little water and too much heat.
Plant stress responses to temperature and water are a continuum starting at the biochemical level manifesting in stomatal closure and cessation of transpiration and CO₂ uptake for photosynthesis. The loss of evaporative cooling results in increased canopy and plant temperatures, and eventually a reduction in canopy water content and turgor pressure leading to increased leaf angles and wilting. At this point, leaves can become sun-scorched and otherwise damaged from increased susceptibility to pathogens and leaf loss. Through prolonged drought, plants will survive off their nonstructural carbohydrate carbon stores, and, ultimately, mortality.
What of all this can we see from space, and how well can we see it?
Our mission at Hydrosat is to detect plant stress responses everywhere and all the time, and particularly early so we have the best chance to do something about it. If a drought hits a region, we want to be able to see which plants, communities, or ecosystems respond first, most, and why. We want to connect that information back to water and agricultural management, which is tied to water and food security. All these stresses and responses can onset quickly with changing weather conditions, plant water and temperature thresholds and tipping points, as well as management practices.
There are many lenses through which to view ecosystems, like RGB lenses that give us different perspectives on what the world looks like and how it works, but when combined form to give us a richer true picture. For ecosystems we can think of these lenses like structure, composition, and function that combine to form the true picture of ecosystems. This may be spot lit by key controls such as CO₂, water, or temperature.
Starting with structure, two broad types of measurements get us key aspects of ecosystem structure: 1) VNIR, which gives us information about canopy cover, damage, and loss; and, 2) active remote sensing such as LiDAR and radar, which tells us about ultimate plant loss and hence mortality.
Where do these measurements lie on the space-time target zone, with no interest in low spatiotemporal resolution, and a desired target on the order of less than 100 m and daily? As with most traditional measurements, there has been a tradeoff between high spatial and high temporal resolution, evident for example with some classic VNIR capabilities. While GEDI LiDAR has good spatial resolution for biomass, its repeat cadence will be improved upon by upcoming radar missions in NISAR and BIOMASS, which both nearly hit the sweet spot target. However, hitting the target square on is Planet’s VNIR capabilities, which is illustrative of how the technological evolution of the measurements has moved from a foundation of NASA to ESA, JAXA and other governmental agencies. Now it’s making its way forward by the commercial sector which, in turn, feeds its data back to the government and academic communities through avenues such as data buys and academic ambassador programs.
So, where do these measurements lie on the Early Warning Indicator? VNIR comes in half way through, telling us about damage before mortality, though not before the stress has already been occurring. Active remote sensing structure measurements give us the final word in the story and, while important for sounding the alarm on future losses, are too late for those that have been lost already.
Turning to composition with key hyperspectral measurements from spectroscopy, gives us even more detailed information about leaf damage, delineating causes between drought stress to plant disease. Leaf pigments such as carotenoids, xanthophylls, and luteins can also give us insights from sensitivity to stresses.
Spaceborne spectroscopy has arguably plateaued for some time, but we are now entering into a hyperspectral renaissance with a plethora of sensors upcoming online recently. However, most of these sensors, including EMIT, SBG, CHIME, PRISMA, DESIS, and HISUI, have very similar characteristics with roughly 30 m spatial resolution repeated every few weeks. Only the proposed German EnMAP mission promises to increase that temporal resolution significantly. With all things considered, if these missions fly concurrently, and share spectral alignment; data processing and data sharing, then the combination of them moves them as a suite upwards on the temporal resolution axis — but that’s a big ‘if’.
Where does spectroscopy hit the Early Warning Indicator? It does a good job getting that early onset of stress, though only slightly better than VNIR. It would be remiss not to include another component in Composition, from measurements that was mentioned in Structure, which is vegetation water content (VWC) from SAR. Considered by some to be a somewhat of an afterthought of the prime SAR products, VWC has been demonstrating to be a fine indicator of plant water stress getting us even higher up on the Early Warning Indicator chart.
Finally, this brings us to Function, where critical measurements in the thermal infrared and solar induced chlorophyll fluorescence provide key insight into evapotranspiration and photosynthesis. While the history of thermal infrared measurements has been a struggle, through successful efforts of groups like the Western States Water Council for Landsat and the selection of ECOSTRESS, maturity in TIR measurements now outpaces most of the measurements covered here outside of VNIR. Building on this strong foundation there are upcoming missions in SBG, LSTM, and TRISHNA that continue to march the sector into the sweet spot target territory. And, like the hyperspectral missions, if the stars and satellites align, the technological expertise will move the group upwards along the temporal frequency axis. Moreover, like VNIR, the technological evolution of TIR has also launched the sector into a more commercialized domain, and like Hydrosat is moving towards the sweet spot right on target with a constellation of TIR (and VNIR) satellites.
With fluorescence as a plant indicator, it has had to extract itself from the atmospheric community which has had less stringent spatial resolution requirements than does the land community, although that is rapidly changing with interests in point sources. The ‘point sources’ of SIF data include, for example, OCO-2/3, TROPOMI, GOME-2, and GOSAT 1/2. Fluorescence, like VWC from radar, could be considered as a fortuitous gift of circumstance. Still, SIF and TIR top the early indicator chart for plant stress, providing the first indications of stress before the more structural responses kick in.
Putting it altogether — the structure, composition, and function measurements — that provides an incredible wealth of information for plant stress. This is an exciting data revolution for the future, and it’s the synergies among those measurements that reveal even more insights into the complexities of ecosystem functioning, responses — and lenses — that enable us to see ecosystems as a whole larger than the sum of their parts, with even further synergies to more missions especially in hydrology.
Speaking of hydrology, providing one more bonus measurement, and that’s soil moisture. It’s not a direct measure of plant characteristics, but has an enormous first-order control over them. So far, there have not been a lot of missions, after AMSR moving into SMOS and SMAP, which have steadily progressed technologically, and NISAR, which seems to be the mission that does it all. While they may not hit the target dead on, their potential for early warning indicator is outstanding. While, TIR and ET are still the best indicators for within-season stress and management, soil moisture — which is also strongly linked to TIR, especially for downscaling (see, for example, a new ECOSTRESS 70 m soil moisture product and upcoming Hydrosat data planned linkages to soil moisture), is particularly useful for advance warning.
Hydrosat strives towards high spatiotemporal resolution, which advances along a technological maturity evolution, and we need to move all measurements along that pathway. While we’re also particularly interested in early detection for management, all measurements contribute to that strength. It is through these interdisciplinary intersections and combinations that we can only reveal the true picture of ecosystems.
This article, originally written by Hydrosat’s Science Lead, Josh Fisher, can be found at the link here.
