Gross Primary Production Estimates Research Articles

A comparison of moderate and high spatial resolution satellite data for modeling gross primary production and transpiration of native prairie, alfalfa, and winter wheat

Although agroecosystems have a significant potential to offset carbon dioxide (CO2), the amount of CO2 captured can vary significantly depending on management practices. Accurate estimation of gross primary production (GPP) and transpiration (T) of agroecosystems at the field scale are essential for the study of food security and water resource management. To date, the carbon and water fluxes data products for commercial agroecosystems are limited, mostly at the moderate spatial resolution (MSR, hundreds of meters), which cannot be used to assess the temporal dynamics of GPP and T at the field scale. This study used the vegetation photosynthesis model (VPM) and vegetation transpiration model (VTM) to estimate field-level daily GPP (GPPVPM) and T (TVTM), respectively, in native prairie, alfalfa (Medicago sativa L.), and winter wheat (Triticum aestivum L.) in central Oklahoma, USA. We evaluated the reliability and advantages of vegetation indices (enhanced vegetation index, EVI and land surface water index, LSWI) in monitoring the land surface phenology using moderate spatial resolution data from Moderate Resolution Imaging Spectroradiometer (MODIS) and high spatial resolution (HSR, tens of meters) data from Landsat and Sentinel-2. The accuracy of GPPVPM and TVTM estimates at different spatial scales was evaluated using GPP (GPPEC) and evapotranspiration (ETEC) from the eddy flux tower sites, respectively. Results demonstrate the capacity of VPM and VTM to estimate the field-level carbon and water flux dynamics and their responses to weather conditions. The use of HSR vegetation indices helped to address certain challenges faced by MSR indices, especially in capturing the crop phenology in smaller areas with conservation measures or disturbances. The findings highlight the importance of using HSR GPP estimates to reduce uncertainty in quantifying CO2 fluxes for croplands and grasslands. The findings also demonstrate the ability of the models to track field-level vegetation phenology, carbon uptake, and water use in agroecosystems under different management practices.

Machine learning algorithms improve MODIS GPP estimates in United States croplands

Introduction: Machine learning methods combined with satellite imagery have the potential to improve estimates of carbon uptake of terrestrial ecosystems, including croplands. Studying carbon uptake patterns across the U.S. using research networks, like the Long-Term Agroecosystem Research (LTAR) network, can allow for the study of broader trends in crop productivity and sustainability.Methods: In this study, gross primary productivity (GPP) estimates from the Moderate Resolution Imaging Spectroradiometer (MODIS) for three LTAR cropland sites were integrated for use in a machine learning modeling effort. They are Kellogg Biological Station (KBS, 2 towers and 20 site-years), Upper Mississippi River Basin (UMRB - Rosemount, 1 tower and 12 site-years), and Platte River High Plains Aquifer (PRHPA, 3 towers and 52 site-years). All sites were planted to maize (Zea mays L.) and soybean (Glycine max L.). The MODIS GPP product was initially compared to in-situ measurements from Eddy Covariance (EC) instruments at each site and then to all sites combined. Next, machine learning algorithms were used to create refined GPP estimates using air temperature, precipitation, crop type (maize or soybean), agroecosystem, and the MODIS GPP product as inputs. The AutoML program in the h2o package tested a variety of individual and combined algorithms, including Gradient Boosting Machines (GBM), eXtreme Gradient Boosting Models (XGBoost), and Stacked Ensemble.Results and discussion: The coefficient of determination (r2) of the raw comparison (MODIS GPP to EC GPP) was 0.38, prior to machine learning model incorporation. The optimal model for simulating GPP across all sites was a Stacked Ensemble type with a validated r2 value of 0.87, RMSE of 2.62 units, and MAE of 1.59. The machine learning methodology was able to successfully simulate GPP across three agroecosystems and two crops.

Exploring the spatio-temporal patterns of global mangrove gross primary production and quantifying the factors affecting its estimation, 1996–2020

Mangrove ecosystems, as an important component of “Blue Carbon”, play a curial role on global carbon cycling; however, the lack of the global estimates of mangrove ecosystem gross primary production (GPP) and the underlying environmental controls on its estimation remain a gap in knowledge. In this study, we utilized global mangrove eddy covariance data and applied Gaussian Process Regression (GPR) to estimate GPP for global mangrove ecosystems, aiming to elucidate the factors influencing these estimates. The optimal GPR achieved favorable estimation performance through cross-validation (R2 = 0.90, RMSE = 0.92 gC/m2/day, WI = 0.86). Over the study period, the globally annual averaged GPP was 2054.53 ± 38.51 gC/m2/yr (comparable to that of evergreen broadleaf forests and exceeds the GPP of most other plant function types), amounting to a total of 304.82 ± 7.71TgC/yr, hotspots exceeding 3000 gC/m2/yr observed near the equator. The analysis revealed a decline in global mangrove GPP during 1996–2020 of −0.89 TgC/yr. Human activities (changes in mangrove cover area) played a relatively consistent role in contributing to this decrease. Conversely, variations in external environmental conditions showed distinct inter-annual differences in their impact. The spatio-temporal distribution patterns of mangrove ecosystems GPP (e.g., the bimodal annual pattern, latitudinal gradients, etc.) demonstrated the regulatory influence of external environmental conditions on GPP estimates. The model ensemble attribution analysis indicated that the fraction of absorbed photosynthetically active radiation exerted the dominant control on GPP estimations, while temperature, salinity, and humidity acted as secondary constraints. The findings of this study provide valuable insights for monitoring, modeling, and managing mangrove ecosystems GPP; and underscore the critical role of mangroves in global carbon sequestration. By quantifying the influences of environmental factors, we enhance our understanding of mangrove carbon cycling estimates, thereby helping sustain of these disproportionately productive ecosystems.

Evaluation of Ecosystem Water Use Efficiency Based on Coupled and Uncoupled Remote Sensing Products for Maize and Soybean

Accurate quantification of ecosystem water use efficiency (eWUE) over agroecosystems is crucial for managing water resources and assuring food security. Currently, the uncoupled Moderate Resolution Imaging Spectroradiometer (MODIS) product is the most widely applied dataset for simulating local, regional, and global eWUE across different plant functional types. However, it has been rarely investigated as to whether the coupled product can outperform the uncoupled product in eWUE estimations for specific C4 and C3 crop species. Here, the eWUE as well as gross primary production (GPP) and evapotranspiration (ET) from the uncoupled MODIS product and the coupled Penman–Monteith–Leuning version 2 (PMLv2) product were evaluated against the in-situ observations on eight-day and annual scales (containing 1902 eight-day and 61 annual samples) for C4 maize and C3 soybean at the five cropland sites from the FLUXNET2015 and AmeriFlux datasets. Our results show the following: (1) For GPP estimates, the PMLv2 product showed paramount improvements for C4 maize and slight improvements for C3 soybean, relative to the MODIS product. (2) For ET estimates, both products performed similarly for both crop species. (3) For eWUE estimates, the coupled PMLv2 product achieved higher-accuracy eWUE estimates than the uncoupled MODIS product at both eight-day and annual scales. Taking the result at an eight-day scale for example, compared to the MODIS product, the PMLv2 product could reduce the root mean square error (RMSE) from 2.14 g C Kg−1 H2O to 1.36 g C Kg−1 H2O and increase the coefficient of determination (R2) from 0.06 to 0.52 for C4 maize, as well as reduce the RMSE from 1.33 g C Kg−1 H2O to 0.89 g C Kg−1 H2O and increase the R2 from 0.05 to 0.49 for C3 soybean. (4) Despite the outperformance of the PMLv2 product in eWUE estimations, both two products failed to differentiate C4 and C3 crop species in their model calibration and validation processes, leading to a certain degree of uncertainties in eWUE estimates. Our study not only provides an important reference for applying remote sensing products to derive reliable eWUE estimates over cropland but also indicates the future modification of the current remote sensing models for C4 and C3 crop species.

Empirical upscaling of OzFlux eddy covariance for high-resolution monitoring of terrestrial carbon uptake in Australia

Abstract. We develop high-resolution (1 km) estimates of gross primary productivity (GPP), ecosystem respiration (ER), and net ecosystem exchange (NEE) over the Australian continent for the period January 2003 to June 2022 by empirical upscaling of flux tower measurements. We compare our estimates with nine other products that cover the three broad categories that define current methods for estimating the terrestrial carbon cycle and assess if consiliences between datasets can point to the correct dynamics of Australia's carbon cycle. Our results indicate that regional empirical upscaling greatly improves upon the existing global empirical upscaling efforts, outperforms process-based models, and agrees much better with the dynamics of CO2 flux over Australia as estimated by two regional atmospheric inversions. Our nearly 20-year estimates of terrestrial carbon fluxes revealed that Australia is a strong net carbon sink of −0.44 PgC yr−1 (interquartile range, IQR = 0.42 PgC yr−1) on average, with an inter-annual variability of 0.18 PgC yr−1 and an average seasonal amplitude of 0.85 PgC yr−1. Annual mean carbon uptake estimated from other methods ranged considerably, while carbon flux anomalies showed much better agreement between methods. NEE anomalies were predominately driven by cumulative rainfall deficits and surpluses, resulting in larger anomalous responses from GPP than ER. In contrast, we show that the long-term average seasonal cycle is dictated more by the variability in ER than GPP, resulting in peak carbon uptake typically occurring during the cooler, drier austral autumn and winter months. This new estimate of Australia's terrestrial carbon cycle provides a benchmark for assessment against land surface model simulations and a means for monitoring of Australia's terrestrial carbon cycle at an unprecedented high resolution. We call this new estimate of Australia's terrestrial carbon cycle “AusEFlux” (Australian Empirical Fluxes).

Non-Ignorable Differences in NIRv-Based Estimations of Gross Primary Productivity Considering Land Cover Change and Discrepancies in Multisource Products

The accurate estimation of gross primary productivity (GPP) plays an important role in accurately projecting the terrestrial carbon cycle and climate change. Satellite-driven near-infrared reflectance (NIRv) can be used to estimate GPP based on their nearly linear relationship. Notably, previous studies have reported that the relationship between NIRv and GPP seems to be biome-specific (or land cover) at the ecosystem scale due to both biotic and abiotic effects. Hence, the NIRv-based estimation of GPP may be influenced by land cover changes (LCC) and the discrepancies in multisource products (DMP). However, these issues have not been well understood until now. Therefore, this study took the Yellow River basin (YRB) as the study area. This area has experienced remarkable land cover changes in recent decades. We used Moderate-Resolution Imaging Spectroradiometer (MODIS) and European Space Agency (ESA) Climate Change Initiative (CCI) land cover products (termed MCD12C1 and ESACCI, respectively) during 2001–2018 to explore the impact of land cover on NIRv-estimated GPP. Paired comparisons between the static and dynamic schemes of land cover using the two products were carried out to investigate the influences of LCC and DMP on GPP estimation by NIRv. Our results showed that the dominant land cover types in the YRB were grassland, followed by cropland and forest. Meanwhile, the main transfer was characterized by the conversion from other land cover types (e.g., barren) to grassland in the northwest of the YRB and from grassland and shrubland to cropland in the southeast of the YRB during the study period. Moreover, the temporal and spatial pattern of GPP was highly consistent with that of NIRv, and the average increase in GPP was 2.14 gCm−2yr−1 across the YRB. Nevertheless, it is shown that both LCC and DMP had significant influences on the estimation of GPP by NIRv. That is, the areas with obvious differences in NIRv-based GPP closely correspond to the areas where land cover types dramatically changed. The achievements of this study indicate that considering the land cover change and discrepancies in multisource products would help to improve the accuracy of NIRv-based estimated GPP.

The roles of radiative, structural and physiological information of sun-induced chlorophyll fluorescence in predicting gross primary production of a corn crop at various temporal scales

Extensive research suggests that sun-induced chlorophyll fluorescence (SIF) and gross primary productivity (GPP) have a near-linear relationship, providing a promising avenue for estimating the carbon uptake of ecosystems. However, the factors influencing the relationship are not yet clear. This study examines the roles of SIF's radiative, structural, and physiological information in predicting GPP, based on four years of field observations of a corn canopy at various temporal scales. We quantified SIF's radiative component by measuring the intensity of incident photosynthetically active radiation (iPAR), and separated the structural and physiological components from SIF observations using the fluorescence correction vegetation index (FCVI). Our results show that the R2 values between SIF and GPP, as estimated by linear models, increased from 0.66 at a half-hour resolution to 0.86 at a one-month resolution. In comparison, the product of FCVI and iPAR, representing the non-physiological information of SIF, performed consistently well in predicting GPP with R2>0.84 at various temporal scales, suggesting a limited contribution of the physiological information of SIF for GPP estimation.. The results further reveal that SIF's radiative and structural components positively impacted the SIF-GPP linearity, while the physiological component had a negative impact on the linearity for most cases, changing from 0.6 % to -27.5 %. As for the temporal dependency, the controls of the SIF-GPP relationship moved from radiation at diurnal scales to structure at the seasonal scales. The structural contribution changed from 14.8 % at a half-hour resolution to 92.4 % at a one-month resolution, while the radiative contribution decreased from 118.0 % at a half-hour resolution to 11.7 % at a one-month resolution. This study contributes to enhancing our understanding of the physiological information conveyed by SIF and the factors influencing the temporal dependency of the SIF-GPP relationship.

Improving the ability of solar-induced chlorophyll fluorescence to track gross primary production through differentiating sunlit and shaded leaves

Recently, solar-induced chlorophyll fluorescence (SIF) is a promising tool to estimate gross primary production (GPP). Photosynthesis gradually saturates with the increasing light, but fluorescence tends to keep increasing, leading to a nonlinear SIF-GPP relationship. This nonlinearity occurs for sunlit leaves but not for shaded leaves for which photosynthesis is light-limited. However, the separation of sunlit and shaded SIF has not been systematically investigated when estimating GPP from SIF. Therefore, it is promising to develop a model for GPP estimation considering such differences. This study proposed an approach to separate the total canopy SIF emission (SIFtotal) from TROPOspheric Monitoring Instrument (TROPOMI) SIF into their sunlit and shaded components (SIFsun and SIFshade). The nonlinearity and linearity in SIF-GPP relationships for sunlit and shaded leaves were incorporated into a two-leaf hybrid model, which was fitted using flux tower data and then evaluated using leave-one-site-out crossing validation. We also elucidated the distinct SIF-GPP relationships between sunlit and shaded leaves using the Soil-Canopy-Observation of Photosynthesis and the Energy balance (SCOPE) model simulation. Compared to previously used linear (R2 = 0.68, RMSE = 2.13 gC·m−2·d−1) or hyperbolic (R2 = 0.72, RMSE = 2.01 gC·m−2·d−1) model based on the big-leaf assumption, our proposed two-leaf hybrid model has the best performance on GPP estimation (R2 = 0.77, RMSE = 1.79 gC·m−2·d−1). We also applied this two-leaf hybrid model to estimate the global GPP during the main growing season in Northern Hemisphere, which were highly correlated with several existing GPP products, with R2 ranging from 0.79 to 0.88. These results will improve our understanding of the relationship between SIF and GPP for sunlit and shaded leaves and will advance application of satellite SIF data to GPP estimation.

Estimation of photosynthetic dynamics in forests from daily measured fluorescence and PRI data with adjustment for canopy shadow fraction

We conducted year-long measurements of the photochemical reflectance index (PRI) and solar-induced fluorescence in the O2A oxygen band (SIFA) at a Norway spruce forest and a European beech forest to study relationships of these remote sensing variables to photosynthesis by trees in grown forest stands. Measured PRI and SIFA values were linked to changes in forest gross primary productivity (GPP) and light-use efficiency (LUE). Changes in the shadow fraction (αS) within tree crowns influenced PRI and fluorescence signals. In the spruce forest, the quantum yield of SIFA (FYSIFA) decreased around midday together with photosynthesis and GPP. Such decreases in FYSIFA were accompanied by an increase in the αS. In the beech forest, we detected an increase in FYSIFA together with a decrease in αS in the afternoon hours. The overall sensitivity of PRI to LUE was variable according to the season, presumably influenced by complex changes in photosynthetic pigments. PRI and FYSIFA showed weak correlations with canopy LUE; however, when considered together, the correlation was strengthened (R2 was 0.63 and 0.34 in spruce and beech forest, respectively). Our model predicting LUE dynamics includes a diurnal minimum of PRI and canopy αS to make allowances for seasonal changes in photosynthetic pigments and for diurnal variability of the shadow fraction in forests. The incorporation of these correcting factors allowed us to estimate LUE at R2 = 0.68 (spruce) and 0.53 (beech). The modeling equations appeared sensitive to the absorbed photosynthetically active radiation (APAR), but less sensitive to the GPP of these forests. Substituting pigments correction with introducing differential PRI (ΔPRI) into the model did not significantly improve the LUE estimation across the season. Our results show that the joint use of PRI and fluorescence improves LUE and GPP estimation accuracy in both daily and seasonal observations.

Extracting subpixel vegetation NDVI time series for evaluating the mixed pixel effect on GPP estimation in urban areas

ABSTRACT The normalized difference vegetation index (NDVI) is the most widely used vegetation index for monitoring vegetation vigor and cover. As NDVI time series are usually derived at coarse or medium spatial resolutions, pixel size often represents a mixture of vegetated and non-vegetated surfaces. In heterogeneous urban areas, mixed pixels impede the accurate estimation of gross primary productivity (GPP). To address the mixed pixel effect on NDVI time series and GPP estimation, we proposed a framework to extract subpixel vegetation NDVI (NDVIvege) from Landsat OLI images in urban areas, using endmember fractions, mixed NDVI (NDVImix), and NDVI of non-vegetation endmembers. Results demonstrated that the NDVIvege extracted by this framework agreed well with the true NDVIvege cross seasons and vegetation fractions, with R2 ranging from 0.74 to 0.82 and RMSE ranging from 0.03 to 0.04. The NDVIvege time series was applied to evaluate vegetation GPP in Wuhan, China. The total annual GPP estimated with NDVIvege was 28-35% higher than the total annual GPP estimated with NDVImix, implying uncertainty in the GPP estimations caused by mixed pixels. This study showed the potential of the proposed framework to resolve NDVIvege for characterizing vegetation dynamics in heterogeneous areas.

Quantifying effects of cold acclimation and delayed springtime photosynthesis resumption in northern ecosystems.

Land carbon dynamics in temperate and boreal ecosystems are sensitive to environmental change. Accurately simulating gross primary productivity (GPP) and its seasonality is key for reliable carbon cycle projections. However, significant biases have been found in early spring GPP simulations of northern forests, where observations often suggest a later resumption of photosynthetic activity than predicted by models. Here, we used eddy covariance-based GPP estimates from 39 forest sites that differ by their climate and dominant plant functional types. We used a mechanistic and an empirical light use efficiency (LUE) model to investigate the magnitude and environmental controls of delayed springtime photosynthesis resumption (DSPR) across sites. We found DSPR reduced ecosystem LUE by 30-70% at many, but not all site-years during spring. A significant depression of LUE was found not only in coniferous but also at deciduous forests and was related to combined high radiation and low minimum temperatures. By embedding cold-acclimation effects on LUE that considers the delayed effects of minimum temperatures, initial model bias in simulated springtime GPP was effectively resolved. This provides an approach to improve GPP estimates by considering physiological acclimation and enables more reliable simulations of photosynthesis in northern forests and projections in a warming climate.

Phenology-Based Maximum Light Use Efficiency for Modeling Gross Primary Production across Typical Terrestrial Ecosystems

The maximum light use efficiency (LUE) (ε0) is a key essential parameter of the LUE model, and its accurate estimation is crucial for quantifying gross primary production (GPP) and better understanding the global carbon budget. Currently, a comprehensive understanding of the potential of seasonal variations of ε0 in GPP estimation across different plant functional types (PFTs) is still lacking. In this study, we used a phenology-based strategy for the estimation of ε0 to find the optimal photosynthetic responses of the parameter in different phenological stages. The start and end of growing season (SOS and EOS) from time series vegetation indices and the camera-derived greenness index were extracted across seven PFT flux sites using the methods of the hybrid generalized additive model (HGAM) and double logistic function (DLF). Optimal extractions of SOS and EOS were evaluated, and the ε0 was estimated from flux site observations during the optimal phenological stages with the light response equation. Coupled with other obligatory parameters of the LUE model, phenology-based GPP (GPPphe-based) was estimated over 21 site-years and compared with vegetation photosynthesis model (VPM)-based GPP (GPPVPM) and eddy covariance-measured GPP (GPPEC). Generally, GPPphe-based basically tracked both the seasonal dynamics and inter-annual variation of GPPEC well, especially at forest, cropland, and wetland flux sites. The R2 between GPPphe-based and GPPEC was stable between 0.85 and 0.95 in forest ecosystems, between 0.75 and 0.85 in cropland ecosystems, and around 0.9 in wetland ecosystems. Furthermore, we found that GPPphe-based was significantly improved compared to GPPVPM in cropland, grassland, and wetland ecosystems, implying that phenology-based ε0 is more appropriate in the GPP estimation of herbaceous plants. In addition, we found that GPPphe-based was significantly improved over GPPVPM in cropland, grassland, and wetland ecosystems, and the R2 between GPPphe-based and GPPEC was improved by up to 0.11 in cropland ecosystems and 0.05 in wetland ecosystems compared to GPPVPM, and RMSE was reduced by up to 5.90 and 2.11 g C m−2 8 day−1, respectively, implying that phenology-based ε0 in herbaceous plants is more appropriate for GPP estimation. This work highlights the potential of phenology-based ε0 in understanding the seasonal variation of vegetation photosynthesis and production.

Coupling a light use efficiency model with a machine learning-based water constraint for predicting grassland gross primary production

Light use efficiency (LUE) model was established to estimate gross primary production (GPP) for understanding the carbon-climate feedbacks of the terrestrial ecosystems. However, water constraints in LUE models can cause large uncertainties in GPP estimates, especially in semiarid grasslands where water is a key forcing factor for multiple ecosystem processes. Here, we proposed a novel LUE-Gradient Boosting Regression Trees (GBRT) model framework where water scalar is derived from five different water constraints for improving the estimates of grassland GPP over the conterminous United States (CONUS). The performance of LUE-GBRT and ten other GPP models [i.e., LUE-RF, LUE-ERT, GBRT, RF, ERT, LUE-fEF, LUE-fVPD, LUE-fLSWI, LUE-fSM, and LUE-fLST] was evaluated against data from eddy covariance (EC) observations at 25 measurement sites over the CONUS domain from 2000 to 2021. We found that LUE-GBRT improved grassland GPP estimates at all EC sites and yielded the highest Kling-Gupta efficiency (0.85) and the lowest root-mean-square error (1.4 g C m−2 d−1) when compared with the five individual GPP models. LUE-GBRT also showed a superior performance compared to LUE-RF and LUE-ERT. Compared with GBRT, the improvements were particularly from the responses to extreme surface conditions that were better characterized and estimated. An innovation of this method is that LUE-GBRT takes machine learning complementary to the physical-based LUE framework for an optimal junction between GPP physical process and model accuracy.

Satellite-based assessment of water use and leaf area efficiencies of dryland conifer forests along an aridity gradient

Dryland forests worldwide are increasingly threatened by drought stress due to climate change. Understanding the relationships between forest structure and function is essential for managing dryland forests to adapt to these changes. We investigated the structure-function relationships in four dryland conifer forests distributed along a semiarid to subhumid climatic aridity gradient. Forest structure was represented by leaf area index (LAI) and function by gross primary productivity (GPP), evapotranspiration (ET), and the derived efficiencies of water use (WUE = GPP/ET) and leaf area (LAE = GPP/LAI). Estimates of GPP and ET were based on the observed relationships between high-resolution vegetation indices from VENμS and Sentinel-2A satellites and flux data from three eddy covariance towers in the study regions between November 2015 to October 2018. The red-edge-based MERIS Terrestrial Chlorophyll Index (MTCI) from VENμS and Sentinel-2A showed strong correlations to flux tower GPP and ET measurements for the three sites (R2cal > 0.91, R2val > 0.84). Using our approach, we showed that as LAI decreased with decreasing aridity index (AI) (i.e., dryer conditions), estimated GPP and ET decreased (R2 > 0.8 to LAI), while WUE (R2 = 0.68 to LAI) and LAE increased. The observed global-scale patterns are associated with a variety of forest vegetation characteristics, at the local scale, such as tree species composition and density. However, our results point towards a canopy-level mechanism, where the ecosystem-LAI and resultant proportion of sun-exposed vs. shaded leaves are primary determinants of WUE and LAE along the studied climatic aridity gradient. This work demonstrates the importance of high-resolution (spatially and spectrally) remote sensing data conjugated with flux tower data for monitoring dryland forests and understanding the intricate structure-function interactions in their response to drying conditions.

NIRvP as a remote sensing proxy for measuring gross primary production across different biomes and climate zones: Performance and limitations

The product of near-infrared radiation reflected by vegetation (NIRv) and PAR (NIRvP) is a promising proxy for the remote estimation of gross primary production (GPP). However, the efficiency of NIRvP in estimating the GPP and its limitations across multiple biomes and climate zones remain unclear. In this study, we aimed to evaluate the performance and limitations of NIRvP in estimating the GPP in comparison to absorbed photosynthetically active radiation (APAR), solar-induced chlorophyll fluorescence (SIF), and the MOD17A2H GPP product. Overall, the correlation between NIRvP and eddy covariance (EC) GPP was stronger than that of APAR, SIF, and MOD17A2H GPP across most biomes with usually similar seasonal variations in radiation, air temperature (TA), and precipitation. The near-​infrared (NIR) reflectance (ρNIR) and light use efficiency (LUE) exhibited a covarying relationship under these environmental conditions, which suggested that the ρNIR contributed positively to the NIRvP-GPP relationship under such climatic conditions. However, the performance of NIRvP was poor in some biomes and climate zones, which exhibited different variations in the seasonal patterns of radiation, TA, and precipitation. The resulting inconsistencies between ρNIR and LUE implied that the ρNIR contributed negatively to the NIRvP-GPP relationship in these regions. Altogether, the findings demonstrated that the NIRvP-GPP relationship was robust but attained a moderate overall relationship across ecosystems (R2 < 0.50) in the majority of biomes and climate zones. In addition, this study also elucidated the limitations of NIRvP as a GPP proxy in certain climate zones, which was attributed to the synergistic contributions of APAR and ρNIR in the NIRvP-GPP relationship.

Remote sensing of daily evapotranspiration and gross primary productivity of four forest ecosystems in East Asia using satellite multi-channel passive microwave measurements

Satellite multi-channel passive microwave observations provide valuable information of canopy vertical structure and vegetation water content, while their potential in the simultaneous estimation of forest evapotranspiration (ET) and gross primary productivity (GPP) has not been explored. This study combined satellite multi-channel microwave Emissivity Difference Vegetation Indices (EDVIs) from 6.9 to 36.5 GHz with the Penman-Monteith (PM), Priestley-Taylor (PT) and light use efficiency (LUE) models to estimate daily ET and GPP over four forest ecosystems in East Asia. In-situ measurements of daily ET, GPP and meteorological variables from ChinaFLUX network (2003 to 2010) were used to calculate and validate the method. An exponential relationship was found between EDVIs and canopy conductance across the forests. By inverting EDVIs-based LUE model coupled with PM and PT models when canopy carbon uptake and water content reached a maximum in growing seasons, the maximum LUE (EDVI-LUEmax) was derived and found to agree with other independent studies. Site validations showed that the EDVI-based method performed a good simultaneous estimation of daily ET and GPP at a tropical forest and three sub-tropical and temperate forests (R2 of 0.64 to 0.94 and RMSE of −11.7% to 19.6% for ET, R2 of 0.39 to 0.73 and RMSE of 31.2% to 50.4% for GPP). The modeled relationship between GPP and ET also agreed that from in-situ measurements. The results were also found to correlate with microwave frequencies. In-situ ET and GPP showed a stronger correlation with EDVI at 18.7 and 36.5 GHz than that at 10.7 and 18.7 GHz across the forests. EDVI-LUEmax of a temperate mixed forest increased as EDVI incorporated lower frequency microwave emissions from deeper canopy layers, which reduced the negative bias in GPP estimation. Overall, this study demonstrated the potential of multi-channel microwave measurements in estimating ET and GPP with the potential effects of vegetation vertical structure.

Temporal upscaling of MODIS instantaneous FAPAR improves forest gross primary productivity (GPP) simulation

Gross primary productivity (GPP) is a measure of carbon uptake by terrestrial ecosystems for carbon neutrality and serves as a key indicator for the Sustainable Development Goals. The instantaneous Fraction of Absorbed Photosynthetically Active Radiation (FAPAR) is applied in GPP estimation by light use efficiency (LUE) models. However, an obvious time scale mismatch exists between instantaneous FAPAR and daily accumulated GPP. In the study, we explored the potential of temporal upscaling instantaneous FAPAR in improving GPP simulation. The absorbed photosynthetically active radiation (APAR) derived from instantaneous and upscaled FAPARs were first compared. GPPs were estimated from instantaneous and upscaled FAPAR by MOD17 LUE model using default and optimized parameters. The optimized GPP was finally compared with three GPP products: GLASS, MOD17A2H and MYD17A2H GPP. The APARs and GPPs were evaluated with the Eddy Covariance (EC) GPP at 78 forest sites from the Fluxnet community. Results showed that when compared with EC GPP, the upscaled FAPAR-derived APAR held a higher R2 (0.26–0.73) than that from instantaneous FAPARs. The upscaled FAPAR-based GPPs by MOD17 LUE model with default or optimized parameters had a larger R2 and smaller RMSE reducing uncertainties by 6.82–7.8% (maximum value: 27.9–49%). The optimized upscaled FAPAR-derived GPP was far superior to the GPP products with larger R2 (0.69–0.85) and smaller RMSE (12.8–14.6 g C/m2/8d). It is concluded that temporal upscaling of MODIS instantaneous FAPAR can well improve the estimation of forest GPP and the parameters of MOD17 LUE model should be further optimized. Our study is an effort toward reducing the uncertainties from FAPAR in the GPP estimation for better assessing the achievement of carbon neutrality and Sustainable Development Goals.

Improved process representation of leaf phenology significantly shifts climate sensitivity of ecosystem carbon balance

Abstract. Terrestrial carbon cycle models are routinely used to determine the response of the land carbon sink under expected future climate change, yet these predictions remain highly uncertain. Increasing the realism of processes in these models may help with predictive skill, but any such addition should be confronted with observations and evaluated in the context of the aggregate behavior of the carbon cycle. Here, two formulations for leaf area index (LAI) phenology are coupled to the same terrestrial biosphere model: one is climate agnostic, and the other incorporates direct environmental controls on both timing and growth. Each model is calibrated simultaneously to observations of LAI, net ecosystem exchange (NEE), and biomass using the CARbon DAta-MOdel fraMework (CARDAMOM) and validated against withheld data, including eddy covariance estimates of gross primary productivity (GPP) and ecosystem respiration (Re) across six ecosystems from the tropics to high latitudes. Both model formulations show similar predictive skill for LAI and NEE. However, with the addition of direct environmental controls on LAI, the integrated model explains 22 % more variability in GPP and Re and reduces biases in these fluxes by 58 % and 77 %, respectively, while also predicting more realistic annual litterfall rates due to changes in carbon allocation and turnover. We extend this analysis to evaluate the inferred climate sensitivity of LAI and NEE with the new model and show that the added complexity shifts the sign, magnitude, and seasonality of NEE sensitivity to precipitation and temperature. This highlights the benefit of process complexity when inferring underlying processes from Earth observations and representing the climate response of the terrestrial carbon cycle.

Generating high-resolution total canopy SIF emission from TROPOMI data: Algorithm and application

Solar-induced chlorophyll fluorescence (SIF) is a rapidly advancing front in modeling global terrestrial gross primary production (GPP). Canopy total SIF emissions (SIFtotal) are mechanistically linked to the plant photosynthesis, and can be estimated from satellite observed SIF (SIFobs) through radiative transfer modeling. However, the current satellite SIFobs and thus SIFtotal are available only at coarse spatial resolutions from several kilometers to tens of kilometers, inhibiting the application at fine spatial scales. Here, we proposed an algorithm to generate both global high-resolution SIFtotal (HSIFtotal) and high-resolution SIFobs (HSIFobs) at 1 km from low-resolution SIFobs (LSIFobs) from the TROPOspheric Monitoring Instrument (TROPOMI), which has a spatial resolution at nadir of 3.5 km by 5.6–7 km. Our statistical method is based on the law of energy conservation and uses satellite derived fraction of absorbed photosynthetically active radiation, fluorescence efficiency, and the escape probability of fluorescence. We evaluated the accuracy of our HSIFtotal using the Orbiting Carbon Observatory-2 SIF (R2 = 0.78). We found that the spatial resolution had clear effects on the relationship between HSIFtotal and GPP. We also compared HSIFtotal to 8-day averaged tower GPP from 135 flux sites and found that they were better correlated when HSIFtotal was averaged over a 1-km radius around the tower than when averaged over a larger radius. Our study provided a unique high-resolution HSIFtotal product, which will advance the estimation of GPP by extrapolating site-level relationships to the global scale.

Analysing far-red SIF directional anisotropy of three structurally contrasting forest canopies towards improved GPP estimation

Solar-induced chlorophyll fluorescence (SIF) is used to estimate terrestrial gross primary production (GPP) due to its physiological link to photosynthesis. However, strong angular variations in satellite-observed SIF (SIFobs), especially for forest canopies with a high structural heterogeneity, hampers robust estimation of GPP. Here, we use directional SIFobs datasets from OCO-2 and TROPOMI satellite sensors and the 3D discrete anisotropic radiative transfer model to investigate the directional properties of far-red SIFobs and their relationships with two GPP products obtained from MODIS and FluxCom datasets for three European forest types with contrasting canopy structures, i.e., boreal, temperate and Mediterranean forests. We found bowl-like angular distribution patterns of SIFobs in both observed and simulated boreal and temperate forests, but no generic distribution pattern of SIFobs was found for the Mediterranean forest. In our GPP estimation of boreal forest, oblique SIFobs performed better than the widely used nadir SIFobs. However, we did not find an optimal view angle suitable for all three forest types. Further, the angular dependencies and forest structure impacts present in SIFobs were found to still propagate into the estimation of total emitted SIF (SIFtotal), which generally regarded as free of these confounding effects. Finally, our results demonstrate that two commonly used variables, i.e., SIFobs with a constant viewing angle and the approximated SIFtotal, both fail to produce a robust relationship with GPP for the Mediterranean forest. These findings highlight the importance of forest canopy structural heterogeneity, including forest floor, on interpretation of directional satellite SIF observations, even when corrected and expressed as a total canopy SIF emission.