Synergy between TROPOMI sun-induced chlorophyll fluorescence and MODIS spectral reflectance for understanding the dynamics of gross primary productivity at integrated carbon observatory system (ICOS) ecosystem flux sites

: An accurate estimation of vegetation Gross Primary Productivity (GPP), which is the amount of carbon taken up by vegetation through photosynthesis for a given time and area, is critical for understanding terrestrial-atmosphere CO 2 exchange processes, ecosystem functioning, and as well as ecosystem responses and adaptations to climate change. Earliest studies, based on ground, airborne and satellite Sun-Induced chlorophyll Fluorescence 15 (SIF) observations have recently revealed close relationships with GPP at different spatial and temporal scales and across different plant functional type (PFT). However, questions remain regarding whether there is a unique relationship between SIF and GPP across different sites and PFT and how can we improve GPP estimates using solely remotely sensed data. Using concurrent measurements of daily TROPOMI (TROPOspheric Monitoring Instrument) SIF (daily SIF d ), daily MODIS Terra and Aqua spectral reflectance, and vegetation indices (VIs, 20 notably NDVI (normalized difference vegetation index), NIRv (near-infrared reflectance of vegetation) and PRI (photochemical reflectance index)) and daily tower-based GPP across eight major different PFT, including mixed forests, deciduous broadleaf forests, croplands, evergreen broadleaf forests, evergreen needleleaf forests, grassland, open shrubland, and wetland, the strength of the linear relationships between tower-based GPP and SIF d at 40 ICOS (Integrated Carbon Observation Systems) flux sites was investigated, and the synergy between these 25 variables to improve GPP estimates using a data-driven modelling approach was evaluated. The results revealed that the strength of the linear relationship between GPP and SIF d was strongly site-specific and PFT-dependent. Furthermore, the GLM (Generalized Linear Model) model, fitted between SIF d , GPP, site and vegetation type as categorical variables, further supported this site-and PFT-dependent relationship between GPP and SIF d . This study also showed that the spectral reflectance bands (RF-R), SIF d plus spectral reflectance (RF-SIF-R) models explained 30 over 80% of the seasonal and interannual variations in GPP, whereas the SIF d plus VIs (RF-SIF-VI) model reproduced only 75% of the tower-based GPP variance. In addition, the relative importance results demonstrated that the spectral reflectance bands in the near-infrared, red and SIF d appeared as the most influential and dominant factors determining GPP predictions, indicating the importance of canopy structure, biochemical properties and vegetation functioning on GPP estimates. Overall, this study provides insights into understanding the strength of 35 the relationships between GPP and SIF and the use of the spectral reflectance and SIF d to improve GPP across sites and PFT. this study primarily intends to evaluate at daily timescale the strength of the linear relationships between SIF and GPP at 40 ICOS flux sites, including several vegetation functional types (mixed forests, deciduous broadleaf forests, croplands, evergreen broadleaf forests, evergreen needleleaf forests, 115 of along diminish integration time) and 3.5 to 14 km across track (based on the viewing angle) and the spectral range between 675-775 nm in the near infrared with a spectral resolution of 0.5 nm, which allows the retrieval of far-red SIF To decouple SIF emissions from the reflected incident sunlight, a statistical and data-driven approach is used, see Köhler et al. (2018) for more details. We used instantaneous and daily ungridded soundings of TROPOMI far-red SIF at 740 nm obtained from Caltech dataset 165 between February, 2018 and December, 2020 (https://data.caltech.edu/records/1347). Instantaneous SIF data were reported in (mW m -2 sr -1 nm -1 ). Daily SIF (hereafter referred as SIF d ) is computed by timing instantaneous SIF with a day length correction factor included in the dataset. the GLM model further supported this hypothesis. Exploring the newly launched satellite instruments such as OCO-3 and ECOSTRESS and upcoming FLEX and GeoCarb satellite missions which are planned to have diurnal sampling or fine-spatial resolution (for instance 300 m for FLEX), along with ongoing 420 ground-based and airborne-based SIF and GPP data altogether will improve the abilities to not only better that the relationship between GPP and SIF d on data pooled across all sites was weak but statistically significant, confirming the PFT dependence of the relationship between SIF d and GPP. The GLM model results supported this PFT-dependent relationship between GPP and SIF d as the site, year and PFT have meaningful effects on the slope of the relationship between GPP and SIF d This study also demonstrated that the spectral reflectance bands, and SIF d plus reflectance explained over 80% of the tower-based GPP variance. The RF models were able to represent the GPP seasonal and interannual variabilities across all sites. In addition, from the mean decrease in impurity results obtained from the RF models, it is inferred that the spectral reflectance bands in the near-infrared, red and SIF d appeared as the most influential and dominant factors determining GPP predictions. In summary, this study provides insights into understanding the strength of the linear relationships between GPP and SIF across different ICOS flux sites and the use of the daily MODIS surface spectral reflectance and SIF d TROPOMI on predicting GPP across different


Introduction
In the context of climate change, understanding the role of terrestrial ecosystems in terms of exchanges of carbon, water and energy is crucial in order to fill-in the knowledge gap on climatic interactions between the biosphere 40 and the atmosphere. Terrestrial ecosystems are one of the main components of the carbon cycle and are highly sensitive to abiotic stresses. Therefore, an accurate estimation of vegetation Gross Primary Productivity (GPP), which is the amount of flux carbon taken up by vegetation through photosynthesis, is critical for understanding terrestrial-atmosphere CO2 exchange processes, ecosystem functioning, as well as ecosystem responses and adaptations to climate change (Gamon et al., 2019). Eddy Covariance (EC) techniques allow the estimation of In recent years, SIF has emerged as a promising remotely sensed tool for monitoring canopy GPP, which is functionally and fundamentally different from the aforementioned VIs (Damm et al., 2010;X. Yang et al., 2015;Köhler et al., 2018;N. Wang et al., 2021;Guanter et al., 2021). In fact, SIF does not rely on vegetation reflectance, instead it is a faint signal directly emitted by chlorophyll from the absorbed sunlight just before the occurrence of photochemical reaction (Porcar-Castell et al., 2014;Gu et al., 2019;Y. Zhang, Migliavacca, et al., 2021). SIF has 80 a physical and physiological meanings, and hence SIF offers new opportunities for global assessment of canopy GPP (Mohammed et al., 2019;Wieneke et al., 2018;Kimm et al., 2021;Dechant et al., 2022). Early studies relied on ground-based, airborne and satellite SIF data measurements at different temporal and spatial scales have indicated a strong linear site-specific and vegetation types dependent relationship between GPP and SIF (Frankenberg et al., 2011;Guanter et al., 2014;H. Yang et al., 2017;Wood et al., 2017;X. Li, Xiao, 85 He, et al., 2018;Paul-Limoges et al., 2018;J. Zhang et al., 2022). In contrast, at finer temporal scales such as diurnal and hourly, the relationship between GPP and SIF is not as strong as at longer timescales. Instead, it appears to be non-linear due to rapid changes in instantaneous variations in PAR and environmental conditions (Damm et al., 2015;Marrs et al., 2020;Kim et al., 2021). How and at which extent driving factors such as canopy structure, spatial heterogeneity and abiotic stress conditions mediate the GPP and 90 SIF relationship remains a challenge and needs to be investigated (Smith et al., 2018;N. Wang et al., 2021;. The main drawback relates to the use of SIF to predict GPP at regional and global scales lies on the difficulty in the weak SIF signals retrieval requiring averaging over large time and spatial scales, and thus hampers detecting fine-scale dynamics needed to explain underlying processes (Gamon et al., 2019;Köhler et al., 2021).Yet, the TROPOspheric Monitoring Instrument (TROPOMI) sensor, which is on board Sentinel 5-Precursor, 95 represents a novel for understanding SIF variations as well as an opportunity to fully evaluate the potential of SIF to improve GPP estimates at the ecosystem scale as it provides a quiet high temporal resolution at daily scale . In addition, the future satellite mission FLEX (Fluorescence Explorer) will provide on a single platform SIF at an unprecedented spatial resolution (300m) together with visible reflectance in the green, red and far red spectral windows (Drusch et al., 2017).

100
The surface spectral reflectance, VIs and SIF can be used altogether to better characterize highly spatiotemporal dynamics in vegetation canopy structure, canopy biochemical properties and vegetation functioning as a response to frequent changes in abiotic conditions at the site and ecosystem scales. However, to the best of our knowledge, an attempt to study the synergy between those variables have not been comprehensively addressed. Owing to most likely that the relationships between structural and functional components are not linear, and have complex 105 interactions over time and space (Hilker et al., 2007;Sippel et al., 2018;Yazbeck et al., 2021;Pabon-Moreno et al., 2022;Kong et al., 2022). Therefore, a series of observations of SIF, surface spectral reflectance and VIs at the site et ecosystem scales could give insights about how SIF is related to GPP, and whether SIF and spectral reflectance, and VIs would make better model parameters, and provide additional information on understanding the dynamics of GPP at the ecosystem scale and beyond.

110
The overarching objective of this work is to study the potential of SIF, spectral reflectance and VIs (namely NDVI, NIRv, and PRI) to estimate canopy GPP at the site and the ecosystem scales, and the synergy between these predictive variables. Specifically, this study primarily intends to evaluate at daily timescale the strength of the grassland, open shrubland, and wetland), and ultimately to examine the synergy between SIF, spectral reflectance and VIs to improve canopy GPP estimates based on data-driven modelling approach.

Materials and Methods
In this current section, the site characteristics and Eddy Covariance (EC) flux data are presented. Then, the remote sensing data (TROPOMI, MODIS Aqua and Terra, and Copernicus Land Cover classification) used in the study 120 are described. At last, data analysis methods used in this study are presented. and Google Earth. Detailed information and references of these sites are provided in Supplementary Materials in Tab S1. Figure 1 presents the location of these study sites, except for GF-Guy site. In the analyses, we used daily GPP values computed as the sum of the half-hourly values estimated from each site. GPP data previously gap filled by ICOS PI representing for a full year, which was the case for instance at CH-Dav, FR-Bil, 135 are filtered out and were not used in the analyses.  and MYDOCGA), centered at the location of each site, were downloaded from Google Earth Engine database. The quality assurance (QA) flag (ideal quality, QA = 0) and the cloud mask (clear, cloud state = 0) criteria were used.

145
Both MODIS Terra and Aqua contain 16 spectral bands of which, the spatial resolution from band 1 to band 7 is 500 m, and 1 km for the remaining bands (8-16) (Vermote et al., 2015). A detailed information about the MODIS data products is given in Supplementary Materials in Tab S2. We used daily MODIS surface reflectance, NDVI, NIRv, and PRI. These VIs are computed according the equation given in Table 1. For PRI computation, we used B13 as a reference band following (Hilker et al., 2009).

155
TROPOMI has a near sun-synchronous orbit with a repeat cycle of 16 days and an equatorial crossing time at around 13:30 local time , which is comparable to those of OCO-2 (Orbiting Carbon Observatory-2) and GOSAT (Greenhouse Gases Observing Satellite). However, the wide swath of TROPOMI (2600 km) is larger than that of OCO-2 (10 km), which enables TROPOMI to provide almost daily spatially continuous global coverage . TROPOMI has a spatial resolution of 7 km along track (5 km 160 since August 2019 owing to diminish integration time) and 3.5 to 14 km across track (based on the viewing angle) and covers the spectral range between 675-775 nm in the near infrared with a spectral resolution of 0.5 nm, which allows the retrieval of far-red SIF . To decouple SIF emissions from the reflected incident sunlight, a statistical and data-driven approach is used, see Köhler et al. (2018)  The TROPOMI SIF observations corresponding to each site were determined relying on the following criteria.
Firstly, we extracted all pixels which center locations are less than 5 km away from the flux tower sites for analyses.

170
The latter choice was motivated due to the fact that the relationship between TROPOMI SIF and tower-based GPP gradually weakened as the distance between sites to the center of pixels increased (data not shown). Secondly, to reduce the cloud effects on SIF data, SIFd observations with cloud fraction over 15% were excluded, even though, some findings reveal that TROPOMI SIF is less sensitive to cloud than surface reflectance values (Guanter et al., 2012;Doughty et al., 2021). The 100 m spatial resolution of the Copernicus Global Land Cover Classification map

175
for the year 2019 (Buchhorn et al., 2020) was used as a based map of the study sites. This land cover classification map was obtained from the Copernicus Global Land Service website (https://lcviewer.vito.be/download).

Data Analysis
In this study, the GPP and SIFd relationship was evaluated at the daily timescale at different spatial scales. Before investigating the link between GPP and SIFd, it was necessary to figure out a way to process outliers which were 180 mostly associated with negative SIFd values. It has been shown that excluding directly negative SIF values could have effects on studying the relationships between satellite SIF data and GPP Köhler et al., 2021). Thus, to handle the outliers, an exponential model was used to account for the structural relationship between the instantaneous SIF and the SIF error included in the dataset. A threshold of ±0.15 mW m -2 sr -1 nm -1 was then applied to the residual random error of the exponential model.

185
The performance of SIFd to predict GPP using linear regression model at each site was examined. Afterward, a year and GPP are the explanatory variables. These aforementioned variables and their interaction effects may affect the changes or variations either in SIFd or GPP and consequently influence the slope and intercept of their relationships.
In order to study the synergy between SIFd , spectral reflectance and VIs to improve GPP estimates, a Random

195
Forest (RF) regressor ensemble decision tree model was used (Brieman, 2001). Briefly, a RF is a machine learning algorithm, which combines the results of several randomly ensemble decision trees to reach a final accurate output.
Before setting up the RF model, the correlation matrix between all variables was computed. It has been shown that features importance can be affected by the high correlation between feature predictors (Toloşi & Lengauer, 2011), suggesting that a decrease in importance values is observed when the level of correlation and the number of Ultimately, the coefficient of determination (R 2 ), Root Mean Squared Error (RMSE), and the p-value metrics were used to evaluate the power of the linear agreement between GPP and SIFd for the site-specific and PFT-specific 220 relationships. The aforementioned metrics plus the adjusted coefficient of determination (adj. R 2 ) were also used to evaluate the performance of the different RF models between the observed and predicted GPP. At last, but not least, a paired t-test is used to compare the performance of the RF models based on the method proposed by (Nadeau et al., 2003). A 5% significance level was used for all statistical inference.

Site-specific relationships
The first aim was to evaluate the strength of the linear relationship between tower-based GPP and SIFd encompassing different vegetation types at site level. Figure 2 shows the relationship between GPP and SIFd at each site. Overall, SIFd was significantly related with tower-based GPP at the site level and at the daily timescale 230 (as p<0.0001 was statistically highly significant), except for IT-Cp2 site of which GPP and SIFd relationship was insignificant and weak (R 2 = 0.001, p≤0.60). Furthermore, the Figure 2 indicates that the slopes and intercepts of the linear regression between GPP and SIFd are site-independent, suggesting that the difference in plant functional types and spatial heterogeneity across sites may significantly affect the relationship between GPP and SIFd.

Plant functional type-specific and overall sites relationships
To test the effects of the PFT on the relationship between GPP and SIFd at the daily timescale, data were pooled 250 across sites of the same PFT (MF, CRO, ENF, DBF, EBF, GRA, OSH, and WET) and a linear regression model was applied on each PFT. Figure 3 depicts the scatterplots of the relationships between GPP and SIFd. The linear relationship between GPP and SIFd was statistically significant for all PFT (R 2 = 0.07-0.54, p<0.0001), taken individually. Furthermore, the slope of the linear regression between GPP and SIFd was strongest for DBF and MF (10.75±0.33 and 10.53±0.87 gC m -2 d -1 / (mW m -2 sr -1 nm -1 )), respectively) and the lowest for EBF (3.08±0.72 255 gC m -2 d -1 / (mW m -2 sr -1 nm -1 )). It can also be seen from the figures that the slopes and the intercepts of their linear relationships were clearly PFT-specific, as shown in Table 2.

260
GRA, OSH, and WET at daily timescale. The R 2 represents the coefficient of determination of the relationship between GPP and SIFd. p is the probability value of the linear model. The shaded area depicted in each line is the 95% confidence interval of the linear relationships between GPP and SIFd.    Figure S1) revealed a strong dependency between predictive variables (notably B9 vs B10, B11 vs B12 and B13 vs B14), indicating that using a RF model built in these variables could be 285 affected by those high correlations. Based on these observations, the spectral reflectance of B10, B12 and B14 were excluded from the explanatory variables of RF regression models.

Performance of GPP estimates using Random Forest regression
In Figure    : Site-specific scatterplots between observed GPP and RF-SIF-R predicted GPP at daily timescale. The R 2 represents 320 the coefficient of determination of the linear relationship between observed GPP and predicted GPP. All pairwise linear relationships between observed GPP vs predicted GPP were statistically significant at all sites (with p<0.0001). The color code represents the eight different vegetation types encountered in the study sites: Red color stands for CRO, green for DBF, yellow for EBF, magenta for ENF, blue for GRA, Cyan for MF, lime for OSH, and dimgrey for WET.

Figure 7:
Scatterplots of observed GPP against RF-SIF-R predicted GPP in eight PFT at daily timescale. The R 2 represents the coefficient of determination of the linear relationship between observed GPP and predicted GPP. p denotes probability value of the linear relationships.
In Figure 8 and Table 3, it is depicted the observed and estimated GPP representing different PFT for all four RF models. The estimation for each site is given in Supplementary Materials Figure S2. Overall, all RF models' GPP 330 predictions capture very well the seasonal and interannual dynamics of the tower-based GPP. However, there are sites, years and vegetation types where observed GPP cannot be estimated with high accuracy. For instance, the RF models tend to underestimate GPP maxima in GRA, WET and EBF vegetation types. These underestimates are mostly marked by the slope of the relationships between the observed GPP and predicted GPP in Table 3.   Figure 9 shows the relative importance (or mean decrease in impurity) of the predictive variables of the RF models 345 for predicting GPP across all sites pooled together. The Figure 9 indicates that for RF-R model, the surface spectral reflectance in the near-infrared (NIR) band (B2 :841-876 nm) and the surface reflectance in the red band (B1: 620-670 nm) were found as the most important inputs variables for GPP estimates. Moreover, it can be seen that the contribution of the far-red spectral reflectance (B13) on predicting GPP is also important, whereas the contribution of the others spectral reflectance bands was on similar level.

Strength of the linear relationship between GPP and SIFd at site level and PFT scale
In this study, the first aim was to evaluate the strength of the relationship between tower-based GPP and SIFd in 370 different PFT at daily timescale and different spatial scales (at site and vegetation type levels).
At the site level, the results demonstrate that there were strong and statistically significant relationships between GPP and SIFd. However, the linear relation between tower-based GPP and SIFd across diverse sites vary significantly in terms of the slope and the intercept, which suggests a site-specific relationship. In other words, at these scales the differential variations in plant physiology and vegetation structure across sites and years and the on the relationship between GPP and SIFd across multiple sites (Dechant et al., 2020;Lu et al., 2020;X. Li, Xiao, He, et al., 2018;Sun et al., 2018;Hao et al., 2021;X. Wang et al., 2022). For instance, Wang et al. (2020) found that the relationship between OCO-2 SIF observed at 757 nm and 771 nm and tower-based GPP across eight vegetation types at 61 flux sites all over the world relies on canopy structure and Lu et al. (2020) 385 reported a better relationship between canopy GPP and SIF corrected from reabsorption and scattering effects than top of canopy SIF based on ground-based measurements, underlying the importance of canopy structure on SIF and GPP relationships.
The relationship between tower-based GPP and SIFd considering the PFT was also examined. The results revealed a strong significant PFT-specific GPP and SIFd relationships across all eight major vegetation type. Yet, the slopes studies (X. Li, Xiao, He, et al., 2018;Hayek et al., 2018;Mengistu et al., 2020;He et al., 2020;Hornero et al., 2021;. Previous researches have also reported weak relationship between GPP and SIF in EBF stands biome (X. Li, Xiao, & He, 2018;. Moreover, it is worth mentioning that the 410 biases related to cloudless sky and cloudy sky in space-based SIF retrieval, complicates the use of SIF to estimate GPP at the PFT scale because cloudless sky SIF and cloudless sky GPP are completely different from cloudy sky SIF and cloudy sky GPP and consequently, their relationship may also differ (Miao et al., 2018). Investigating understand the GPP and SIF relationship but also to completely decouple the effects of driving factors such as leaf morphology and orientation, vegetation physiology, canopy structure and abiotic stress conditions that mediate their relationships at the ecosystem scale.

GPP using Random Forest
The second main goal in this manuscript was to explore the synergy between SIFd from TROPOMI instrument and MODIS surface spectral reflectance and reflectance based-indices namely NDVI, NIRv and PRI for predicting GPP on data pooled across all sites. For achieving this purpose, four RF regression models were established: RF-R, RF-SIF-R, RF-SIF-R-PFT, and RF-SIF-VI. Except for RF-SIF-R-PFT model, the main advantage of using 430 solely remotely sensed data for estimating GPP is that it can be avoided using information on land cover type and land cover change, as well as meteorological data (J. Xiao et al., 2019).
The current results show that RF-R model (surface spectral reflectance alone) could explain 86% of the variance in tower-based GPP at the daily time scale, whereas RF-SIF-R (SIFd plus surface spectral reflectance), RF-SIF-R-PFT (SIFd plus surface spectral reflectance plus PFT), and RF-SIF-VI (SIFd plus reflectance based-indices) models 435 explain 82%, 83% and 75% of the interannual variabilities in GPP across all sites, respectively. These results suggest that at the seasonal scale spectral reflectance presumably capture the variations in canopy structure, while SIF is highly dependent on variabilities in changes in absorbed photosynthetically active sunlight (APAR). The seasonal variations in canopy structure, especially LAI, are strongly correlated with variations in GPP ((Dechant et al., 2022)). This could justify the strong relationship found between tower-based GPP and the predicted GPP by 440 the RF-R model. On the other hand, SIF is an integrative variable at the seasonal and interannual scales as shown in Figure 9 and on the correlation matrix results (strong contribution of SIF on GPP estimates and high correlation between GPP and SIF compared to each spectral reflectance band). This may explain why SIF, while exhibiting the highest relative importance, fails to improve the GPP estimate. Furthermore, while being limited by its spatial resolution (7 km x 3.5 km), at which SIF may lose its physiological information and most likely reflect 445 phenological, structural and illumination information (Jonard et al., 2020;Kimm et al., 2021), SIF remains a better predictor of GPP than each reflectance band individually. These results also revealed that the RF-SIF-VI have the poorest performance in predicting GPP. This lower performance could be partly due to the well-known saturation of VIs over intense canopies. In addition, the paired t-test did not show any statistically significant difference between RF-R and RF-SIF-R models, which confirms the above hypothesis, which suggests that SIF represents models outperform previous GPP products derived based on data-driven methods (Wolanin et al., 2019;Tramontana et al., 2016;Jung et al., 2019) and process-based model (Jiang & Ryu, 2016;Lin et al., 2019), which included even further inputs as predictive variables such as meteorological data, land cover type and land cover change data and were conducted mostly at longer time scales ( Furthermore, in this study, it is determined what are the main variables contributing to GPP prediction using the four RF models based on the relative importance metric of each model. Yet, it is found that SIFd, the surface spectral reflectance in the NIR band (B2), red band (B1) and far-red band (B13), as well as the vegetation type, NDVI and NIRv seem to provide useful information for the predictions of GPP as shown in Figure 9. B2 and B1 480 are well-known spectral bands for characterizing vegetation canopy structure, seasonal phenology, canopy scattering and reabsorption due to chlorophyll content within leaves, and consequently have a dominant role in estimating GPP across all sites. The high contribution of SIFd is presumably due to its integrative role at the seasonal and interannual scales as explained previously (Maguire et al., 2020;Dechant et al., 2022). PRI is known to be implied in the xanthophyll cycle, which is an important photoprotection mechanism and as a driver of GPP 485 (X. Wang et al., 2020;Hmimina et al., 2015;Soudani et al., 2014). However, in this study, the findings evidenced that the contribution of PRI on predicting GPP was weak, which could be explained by the spatial and temporal aggregation of the rapid responses in plant physiological and functional activities, observable at the finer scales (diurnal). Ultimately, the findings in this study suggest that using spectral reflectance bands and SIF for estimating GPP is an important approach for improving GPP predictions compared to GPP products that include 490 meteorological and land cover type information.

Conclusion
In this current study, the strength of the linear relationships between tower-based GPP and SIFd encompassing eight major plant functional types (PFT) at the site and interannual scales was evaluated, and the synergy between SIFd, surface spectral reflectance, and reflectance-based indices namely NDVI, NIRv and PRI to improve GPP 495 estimates using a data-driven modelling approach was examined.
At the site scale, the results showed a strong and statistically significant relationships between SIFd and GPP (p<0.0001). However, the slopes and intercepts of their relationships were site-dependent, indicating that canopy structure and environmental conditions affect the relationship between GPP and SIFd. The GPP and SIFd relationship across all sites of the same PFT was considerably significant and was PFT-specific. Furthermore, it Supplement. The supplementary materials related to this manuscript is available as a pdf document.

520
Author contributions. All authors contributed to the paper conceptualization. HB performed the data collection and preparation. HB and GH performed the data pre-processing, analyses and prepared the figures. HB led the writing of the manuscript with the contributions from all authors. KS, YG and GH supervised the project.

525
Competing interests. The authors declare that they have no conflict of interest.
Funding. This ongoing Ph.D work is jointly funded by le Centre National d'Études Spatiales (CNES) and ACRI-ST.