Periodic Reporting for period 1 - ReSPEc (Towards a Remotely sensed estimation of the Photosynthetic Energy balance)
Reporting period: 2018-07-01 to 2020-06-30
During photosynthesis, light and water are used by plants to assimilate CO2 and produce organic compounds, while oxygen (O2) is released. Without photosynthesis, life on earth would be very different. While the produced biomass serves as food, fibre and energy source, the exchange of carbon, oxygen and water affects the composition of the atmosphere, as well as the climate. The carbon assimilated during photosynthesis over a certain time and space is referred to as gross primary production (GPP). The quantification of GPP and its spatio-temporal description from field to global scale is therefore of fundamental importance, not only in terms of climate change research, but also in respect to food security.
Space- and airborne-based spectroscopy can be considered the only technology that continuously observes vegetation status and functioning at field to global scale, allowing to derive GPP. Because it is directly related to the photosynthetic process, SIF (sun-induced chlorophyll fluorescence) is a promising signal that might allow overcoming the limitations of traditional reflectance-based methods (e.g. surface greenness) for estimating GPP. Advances in optical sensor technology and methods allow reliable measurements of SIF using ground, airborne, and satellite sensors and have shown good relationships between SIF GPP. However, it is still unclear if this relationship is primarily driven by their common dependence on the absorbed incoming light (APAR), rather than being directly related. Recent studies have shown that the SIF fractior per incoming light (SIF/APAR = Fyield) might be related to light-use efficiency (LUE), a measure of the fraction of photons used in the photosynthesis reactions.
Nevertheless, there are two major factors which complicate the predictability of LUE by Fyield: i) SIF and Fyield are related to the photosynthetic light reactions, while the LUE concept includes the dark reactions, and the stomatal conductance; ii) only 1-2 % of APAR is re-emitted as SIF, while heat dissipation (non-photochemical quenching, or NPQ) can contribute by 17.5 to 98 %. Therefore, the relation between Fyield and LUE will always be indirect and poor. To overcome this problem, the SIF signal has to be linked to the light reactions and the signal of NPQ needs to be accessed. It is currently being discussed if the photochemical reflectance index (PRI), which is sensitive to changes in xanthophyll activity (a pigment used for protection from photodamage), could be used to estimate NPQ on plant and ecosystem level. However, the application of the PRI at plant and ecosystem level is complicated by factors like canopy structure, leaf angle distribution and changes in leaf pigment pools. Furthermore, plant or ecosystem-derived PRI can only be compared to leaf level NPQ measurements, further complicating the use of PRI to estimate NPQ.
The main objective of ReSPEc “Towards a Remotely Sensed estimation of the Photosynthetic Energy balance” is to overcome the named problems by developing a semi-mechanistic model that allows an improved estimation of plant and ecosystem photosynthesis by traditional and novel remote-sensing applications.
Space- and airborne-based spectroscopy can be considered the only technology that continuously observes vegetation status and functioning at field to global scale, allowing to derive GPP. Because it is directly related to the photosynthetic process, SIF (sun-induced chlorophyll fluorescence) is a promising signal that might allow overcoming the limitations of traditional reflectance-based methods (e.g. surface greenness) for estimating GPP. Advances in optical sensor technology and methods allow reliable measurements of SIF using ground, airborne, and satellite sensors and have shown good relationships between SIF GPP. However, it is still unclear if this relationship is primarily driven by their common dependence on the absorbed incoming light (APAR), rather than being directly related. Recent studies have shown that the SIF fractior per incoming light (SIF/APAR = Fyield) might be related to light-use efficiency (LUE), a measure of the fraction of photons used in the photosynthesis reactions.
Nevertheless, there are two major factors which complicate the predictability of LUE by Fyield: i) SIF and Fyield are related to the photosynthetic light reactions, while the LUE concept includes the dark reactions, and the stomatal conductance; ii) only 1-2 % of APAR is re-emitted as SIF, while heat dissipation (non-photochemical quenching, or NPQ) can contribute by 17.5 to 98 %. Therefore, the relation between Fyield and LUE will always be indirect and poor. To overcome this problem, the SIF signal has to be linked to the light reactions and the signal of NPQ needs to be accessed. It is currently being discussed if the photochemical reflectance index (PRI), which is sensitive to changes in xanthophyll activity (a pigment used for protection from photodamage), could be used to estimate NPQ on plant and ecosystem level. However, the application of the PRI at plant and ecosystem level is complicated by factors like canopy structure, leaf angle distribution and changes in leaf pigment pools. Furthermore, plant or ecosystem-derived PRI can only be compared to leaf level NPQ measurements, further complicating the use of PRI to estimate NPQ.
The main objective of ReSPEc “Towards a Remotely Sensed estimation of the Photosynthetic Energy balance” is to overcome the named problems by developing a semi-mechanistic model that allows an improved estimation of plant and ecosystem photosynthesis by traditional and novel remote-sensing applications.
Two field experiments with potato plants were setup from April to September 2019. The main experiment focused on obtaining all necessary parameters needed for a successful completion of the project. During the experiment we focused ok two mesocosm, one control and the other under increasing drought conditions. The measurements were performed over eight consecutive days from 8 am to 7 pm during the heatwave that affected Antwerp in August 2019. In addition to the main experiment, an additional experiment was performed to study the structural effects that might affect the SIF signal under stress conditions. An experiment with 15 mesocosm at three different treatments (control, drought, non-fertilized) and five repetitions were set up. From July to August 2019 SIF measurements at leaf and canopy were conducted under clear-sky conditions. The collected data set is unique in its extensive parameter list, temporal resolution and environmental conditions (highest ever recorded temperature in Belgium, 40.2° and a major drought at continental scale).
Based on the data set collected and an additional data set from an experiment in 2017 two semi mechanistic models were successfully developed to retrieve the electron transport rate (ETR, a parameter describing the activity of the photosynthetic light reactions) from optical measurements under changing stress conditions.
1) Semi mechanistic model for estimating ETR by Fyield and the photochemical reflectance index
2) Semi mechanistic model for estimating ETR by the ratio of red and far-red SIF
Even though a combination of Fyield and PRI can be utilized to predict ETR, a uncertainty analysis showed that the model is highly sensitive to changes in PRI. The SIFratio (SIF687/SIF760) was found to be a more precise, accurate and robust predictor of ETR, which is of particular interest since the SIFratio, in contrast to Fyield can be obtained directly from measurements in both peaks. This can overcome the difficulty related to obtaining green APAR remotely, making the SIFratio better suited for proximal and satellite remote sensing than Fyield .
From the optical leaf measurements the PRI, the red edge chlorophyll index and the red-edge carotenoid index were calculated. It was found that the PRI showed a strong decreasing linear relationship with the epoxidation state (a proxy for NPQ) (R² = 0.64 rRMSE = 19.8%), where PRI normalized by chlorophyll and carotene pools greatly improved the prediction of EPS with an R² of 0.80 and a reduced the relative error (rRMSE of 14.8%). Our results confirm that PRI could be used to estimate stress effects on foliar EPS because of its link to the energy- (or pH-) dependent mechanism of non-photochemical quenching. The foliar pigment content, however, exerts a dominant influence on the PRI, and therefore correction methods are needed to compensate for this effect. Normalization of PRI by the chlorophyll and carotene pools further improved the estimation of leaf-level EPS and its reduction by stress.
When comparing leaf with simultaneously measured canopy measurements of SIF, a moderate positive correlation was found between them. It is known that the absorbed photosynthetic active radiation (APAR, sun-light between 400 and 700 nm) is the main driver of SIF and that the positive relationship between leaf and canopy SIF could be driven by their first order relationship with APAR. Therefore normalized SIF by APAR was used to calculate red and far-red fluorescence yield at leaf and canopy. Results show that the relationship between leaf and canopy red Fyield improved compared to SIF. The relationship between leaf and canopy far-red Fyield however showed to be non-significant.
Based on the data set collected and an additional data set from an experiment in 2017 two semi mechanistic models were successfully developed to retrieve the electron transport rate (ETR, a parameter describing the activity of the photosynthetic light reactions) from optical measurements under changing stress conditions.
1) Semi mechanistic model for estimating ETR by Fyield and the photochemical reflectance index
2) Semi mechanistic model for estimating ETR by the ratio of red and far-red SIF
Even though a combination of Fyield and PRI can be utilized to predict ETR, a uncertainty analysis showed that the model is highly sensitive to changes in PRI. The SIFratio (SIF687/SIF760) was found to be a more precise, accurate and robust predictor of ETR, which is of particular interest since the SIFratio, in contrast to Fyield can be obtained directly from measurements in both peaks. This can overcome the difficulty related to obtaining green APAR remotely, making the SIFratio better suited for proximal and satellite remote sensing than Fyield .
From the optical leaf measurements the PRI, the red edge chlorophyll index and the red-edge carotenoid index were calculated. It was found that the PRI showed a strong decreasing linear relationship with the epoxidation state (a proxy for NPQ) (R² = 0.64 rRMSE = 19.8%), where PRI normalized by chlorophyll and carotene pools greatly improved the prediction of EPS with an R² of 0.80 and a reduced the relative error (rRMSE of 14.8%). Our results confirm that PRI could be used to estimate stress effects on foliar EPS because of its link to the energy- (or pH-) dependent mechanism of non-photochemical quenching. The foliar pigment content, however, exerts a dominant influence on the PRI, and therefore correction methods are needed to compensate for this effect. Normalization of PRI by the chlorophyll and carotene pools further improved the estimation of leaf-level EPS and its reduction by stress.
When comparing leaf with simultaneously measured canopy measurements of SIF, a moderate positive correlation was found between them. It is known that the absorbed photosynthetic active radiation (APAR, sun-light between 400 and 700 nm) is the main driver of SIF and that the positive relationship between leaf and canopy SIF could be driven by their first order relationship with APAR. Therefore normalized SIF by APAR was used to calculate red and far-red fluorescence yield at leaf and canopy. Results show that the relationship between leaf and canopy red Fyield improved compared to SIF. The relationship between leaf and canopy far-red Fyield however showed to be non-significant.
A further analysis showed that red Fyield at leaf and canopy and far-red Fyield at leaf decreases with increasing heat and drought stress. A behaviour which can be expected since FY decreases with increasing NPQ under stress conditions. far-red Fyield at canopy level however showed an increase which demonstrates the risk of misinterpreting the plant physiological response to stress when using far-red SIF and Fyield. Since most estimations of the photosynthetic capacity nowadays are based on far-red SIF, our results are of utmost importance for the fluorescence community.
