Search

One of the major undetermined problems in evaporation (ET) retrieval using thermal infrared remote sensing is the lack of a physically based ground heat flux (G) model and its integration within the surface energy balance (SEB) equation. Here, we present a novel approach based on coupling a thermal inertia (TI)-based mechanistic G model with an analytical surface energy balance model, Surface Temperature Initiated Closure (STIC, version STIC1.2). The coupled model is named STIC-TI. The model is driven by noon–night (13:30 and 01:30 local time) land surface temperature, surface albedo, and a vegetation index from MODIS Aqua in conjunction with a clear-sky net radiation sub-model and ancillary meteorological information. SEB flux estimates from STIC-TI were evaluated with respect to the in situ fluxes from eddy covariance measurements in diverse ecosystems of contrasting aridity in both the Northern Hemisphere and Southern Hemisphere. Sensitivity analysis revealed substantial sensitivity of STIC-TI-derived fluxes due to the land surface temperature uncertainty. An evaluation of noontime G (Gi) estimates showed 12 %–21 % error across six flux tower sites, and a comparison between STIC-TI versus empirical G models also revealed the substantially better performance of the former. While the instantaneous noontime net radiation (RNi) and latent heat flux (LEi) were overestimated (15 % and 25 %), sensible heat flux (Hi) was underestimated (22 %). Overestimation (underestimation) of LEi (Hi) was associated with the overestimation of net available energy (RNi−Gi) and use of unclosed surface energy balance flux measurements in LEi (Hi) validation. The mean percent deviations in Gi and Hi estimates were found to be strongly correlated with satellite day–night view angle difference in parabolic and linear pattern, and a relatively weak correlation was found between day–night view angle difference versus LEi deviation. Findings from this parameter-sparse coupled G–ET model can make a valuable contribution to mapping and monitoring the spatiotemporal variability of ecosystem water stress and evaporation using noon–night thermal infrared observations from future Earth observation satellite missions such as TRISHNA, LSTM, and SBG.

In 2020, the Australian and New Zealand flux research and monitoring network, OzFlux, celebrated its 20th anniversary by reflecting on the lessons learned through two decades of ecosystem studies on global change biology. OzFlux is a network not only for ecosystem researchers, but also for those ‘next users’ of the knowledge, information and data that such networks provide. Here, we focus on eight lessons across topics of climate change and variability, disturbance and resilience, drought and heat stress and synergies with remote sensing and modelling. In distilling the key lessons learned, we also identify where further research is needed to fill knowledge gaps and improve the utility and relevance of the outputs from OzFlux. Extreme climate variability across Australia and New Zealand (droughts and flooding rains) provides a natural laboratory for a global understanding of ecosystems in this time of accelerating climate change. As evidence of worsening global fire risk emerges, the natural ability of these ecosystems to recover from disturbances, such as fire and cyclones, provides lessons on adaptation and resilience to disturbance. Drought and heatwaves are common occurrences across large parts of the region and can tip an ecosystem's carbon budget from a net CO2 sink to a net CO2 source. Despite such responses to stress, ecosystems at OzFlux sites show their resilience to climate variability by rapidly pivoting back to a strong carbon sink upon the return of favourable conditions. Located in under-represented areas, OzFlux data have the potential for reducing uncertainties in global remote sensing products, and these data provide several opportunities to develop new theories and improve our ecosystem models. The accumulated impacts of these lessons over the last 20 years highlights the value of long-term flux observations for natural and managed systems. A future vision for OzFlux includes ongoing and newly developed synergies with ecophysiologists, ecologists, geologists, remote sensors and modellers.

The Vegetation Optimality Model (VOM, Schymanski et al., 2009, 2015) is an optimality-based, coupled water–vegetation model that predicts vegetation properties and behaviour based on optimality theory rather than calibrating vegetation properties or prescribing them based on observations, as most conventional models do. Several updates to previous applications of the VOM have been made for the study in the accompanying paper of Nijzink et al. (2022), where we assess whether optimality theory can alleviate common shortcomings of conventional models, as identified in a previous model inter-comparison study along the North Australian Tropical Transect (NATT, Whitley et al., 2016). Therefore, we assess in this technical paper how the updates to the model and input data would have affected the original results of Schymanski et al. (2015), and we implemented these changes one at a time. The model updates included extended input data, the use of variable atmospheric CO2 levels, modified soil properties, implementation of free drainage conditions, and the addition of grass rooting depths to the optimized vegetation properties. A systematic assessment of these changes was carried out by adding each individual modification to the original version of the VOM at the flux tower site of Howard Springs, Australia. The analysis revealed that the implemented changes affected the simulation of mean annual evapotranspiration (ET) and gross primary productivity (GPP) by no more than 20 %, with the largest effects caused by the newly imposed free drainage conditions and modified soil texture. Free drainage conditions led to an underestimation of ET and GPP in comparison with the results of Schymanski et al. (2015), whereas more fine-grained soil textures increased the water storage in the soil and resulted in increased GPP. Although part of the effect of free drainage was compensated for by the updated soil texture, when combining all changes, the resulting effect on the simulated fluxes was still dominated by the effect of implementing free drainage conditions. Eventually, the relative error for the mean annual ET, in comparison with flux tower observations, changed from an 8.4 % overestimation to an 10.2 % underestimation, whereas the relative errors for the mean annual GPP remained similar, with an overestimation that slightly reduced from 17.8 % to 14.7 %. The sensitivity to free drainage conditions suggests that a realistic representation of groundwater dynamics is very important for predicting ET and GPP at a tropical open-forest savanna site as investigated here. The modest changes in model outputs highlighted the robustness of the optimization approach that is central to the VOM architecture.

The magnitude of the terrestrial carbon (C) sink may be overestimated globally due to the difficulty of accounting for all C losses across heterogeneous landscapes. More complete assessments of net landscape C balances (NLCB) are needed that integrate both emissions by fire and transfer to aquatic systems, two key loss pathways of terrestrial C. These pathways can be particularly significant in the wet-dry tropics, where fire plays a fundamental part in ecosystems and where intense rainfall and seasonal flooding can result in considerable aquatic C export (sigma F-aq). Here, we determined the NLCB of a lowland catchment (similar to 140 km(2)) in tropical Australia over 2 years by evaluating net terrestrial productivity (NEP), fire-related C emissions and sigma F-aq (comprising both downstream transport and gaseous evasion) for the two main landscape components, that is, savanna woodland and seasonal wetlands. We found that the catchment was a large C sink (NLCB 334 Mg C km(-2) year(-1)), and that savanna and wetland areas contributed 84% and 16% to this sink, respectively. Annually, fire emissions (-56 Mg C km(-2) year(-1)) and sigma F-aq (-28 Mg C km(-2) year(-1)) reduced NEP by 13% and 7%, respectively. Savanna burning shifted the catchment to a net C source for several months during the dry season, while sigma F-aq significantly offset NEP during the wet season, with a disproportionate contribution by single major monsoonal events-up to 39% of annual sigma F-aq was exported in one event. We hypothesize that wetter and hotter conditions in the wet-dry tropics in the future will increase sigma F-aq and fire emissions, potentially further reducing the current C sink in the region. More long-term studies are needed to upscale this first NLCB estimate to less productive, yet hydrologically dynamic regions of the wet-dry tropics where our result indicating a significant C sink may not hold.

International audience; Global evaporation monitoring from Earth observation thermal infrared satellite missions is historically challenged due to the unavailability of any direct measurements of aerodynamic temperature. State-of-the-art one-source evaporation models use remotely sensed radiometric surface temperature as a substitute for the aerodynamic temperature and apply empirical corrections to accommodate for their inequality. This introduces substantial uncertainty in operational drought mapping over complex landscapes. By employing a non-parametric model, we show that evaporation can be directly retrieved from thermal satellite data without the need of any empirical correction. Independent evaluation of evaporation in a broad spectrum of biome and aridity yielded statistically significant results when compared with eddy covariance observations. While our simplified model provides a new perspective to advance spatio-temporal evaporation mapping from any thermal remote sensing mission, the direct retrieval of aerodynamic temperature also generates the highly required insight on the critical role of biophysical interactions in global evaporation research.

Animals, such as termites, have largely been overlooked as global-scale drivers of biogeochemical cycles1,2, despite site-specific findings3,4. Deadwood turnover, an important component of the carbon cycle, is driven by multiple decay agents. Studies have focused on temperate systems5,6, where microbes dominate decay7. Microbial decay is sensitive to temperature, typically doubling per 10°C increase (decay effective Q10 = ~2)8–10. Termites are important decayers in tropical systems3,11–13 and differ from microbes in their population dynamics, dispersal, and substrate discovery14–16, meaning their climate sensitivities also differ. Using a network of 133 sites spanning 6 continents, we report the first global field-based quantification of temperature and precipitation sensitivities for termites and microbes, providing novel understandings of their response to changing climates. Temperature sensitivity of microbial decay was within previous estimates. Termite discovery and consumption were both much more sensitive to temperature (decay effective Q10 = 6.53), leading to striking differences in deadwood turnover in areas with and without termites. Termite impacts were greatest in tropical seasonal forests and savannas and subtropical deserts. With tropicalization17 (i.e., warming shifts to a tropical climate), the termite contribution to global wood decay will increase as more of the earth’s surface becomes accessible to termites. Acknowledgements. For materials, field, and lab assistance, we thank the Australian Landscape Trust Team, Juan Martínez de Aragón, Eric Roy, Hugh Wilson, Kylie Brice, Bethanie Coleman, Brendan Delroy, Coline Deveautour, Suzanne Donn, Generci Assis Neves, Ricardo Marques Barreiros, Ricardo Marques Barreiros, Generci Assis Neves from Resineves Agroflorestal company, Alfred Lochner, Pousada Vale do Céu, Humberto Robles, María M. Rivera, and Grant Kirker. Steve Allison and Tom Crowther provided comments on the manuscript and Abbey Yatsko helped with figure and manuscript formatting. Funding for this work was provided by US National Science Foundation (DEB-1655759, 2149151, 1713502, 1713435, 1647502, 1546686, 1831952), George Washington University, USDA Forest Service, Centre College Faculty Development Funds, Australia Terrestrial Ecosystem Research Network (TERN) National Collaborative Research Infrastructure Strategy (NCRIS), Royal Society-FCDO Africa Capacity Building Initiative, New Phytologist Foundation, Fondecyt grant:1160329, Coordenação de Aperfeiçoamento de Pessoal de Nível Superior, Brasil (CAPES), Department of Ecology and Conservation of the Federal University of Lavras, CNPq, FAPEMIG, Australian Academy of Science 2017 Thomas Davies Research Grant, Australian Research Council (DP160103765), UK National Environment Research Council (NE/L000016/1), Fundação de Amparo à Pesquisa do Estado de São Paulo, Brazil (NERC - FAPESP)19/07773-1, Environment Research and Technology Development Fund (ERTDF; JPMEERF15S11420) of the Environmental Restoration and Conservation Agency (ERCA) of Japan, COLCIENCIAS (No. FP44842-046-2017), Spanish Government (grant PID2019-110521GB-I00), Catalan Government (grant SGR 2017-1005), Fundación Ramón Areces (grant ELEMENTAL-CLIMATE), Consejo Nacional de Investigaciones Científicas y Técnicas (CONICET), Agencia Nacional de Promoción Científica y Tecnológica (ANPCyT) PICT 2019- 2472, National Agency for the Promotion of Research, Technological Development and Innovation, Scientific and Technological Research Project 2018-01561 (PICT 2018- 01561), iDiv German Research Foundation (DFG– FZT 118, 202548816), European Research Council (ERC), European Union's Horizon 2020 research and innovation program (grant agreement no. 677232).

It is well documented that energy balance and other remote sensing-based evapotranspiration (ET) models face greater uncertainty over water-limited tree-grass ecosystems (TGEs), representing nearly 1/6th of the global land surface. Their dual vegetation strata, the grass-dominated understory and tree-dominated overstory, make for distinct structural, physiological and phenological characteristics, which challenge models compared to more homogeneous and energy-limited ecosystems. Along with this, the contribution of grasses and trees to total transpiration (T), along with their different climatic drivers, is still largely unknown nor quantified in TGEs. This study proposes a thermal-based three-source energy balance (3SEB) model, accommodating an additional vegetation source within the well-known two-source energy balance (TSEB) model. The model was implemented at both tower and continental scales using eddy-covariance (EC) TGE sites, with variable tree canopy cover and rainfall (P) regimes and Meteosat Second Generation (MSG) images. 3SEB robustly simulated latent heat (LE) and related energy fluxes in all sites (Tower: LE RMSD ~60 W/m2; MSG: LE RMSD ~90 W/m2), improving over both TSEB and seasonally changing TSEB (TSEB-2S) models. In addition, 3SEB inherently partitions water fluxes between the tree, grass and soil sources. The modelled T correlated well with EC T estimates (r > .76), derived from a machine learning ET partitioning method. The T/ET was found positively related to both P and leaf area index, especially compared to the decomposed grass understory T/ET. However, trees and grasses had contrasting relations with respect to monthly P. These results demonstrate the importance in decomposing total ET into the different vegetation sources, as they have distinct climatic drivers, and hence, different relations to seasonal water availability. These promising results improved ET and energy flux estimations over complex TGEs, which may contribute to enhance global drought monitoring and understanding, and their responses to climate change feedbacks., The research received funding from the European Union’s Horizon 2020 research and innovation programme under the Marie Sklodowska-Curie TRuStEE project (grant agreement No 721995). It was also funded by Ministerio de Economía y Competitividad through SynerTGE CGL2015-G9095-R funded by MCIN/ AEI /10.13039/501100011033/ FEDER ‘a way of making Europe’. The study also benefitted from the DIVERSPEC-TGA project, funded by the Ministerio de Ciencia e Innovación of Spain MCIN/ AEI /10.13039/501100011033. The infrastructure at ES-LM1 was partly funded through the Alexander von Humboldt Foundation, ELEMENTAL (CGL 2017-83538-C3-3-R, MINECO-FEDER) and IMAGINA (PROMETEU 2019; Generalitat Valenciana). Funding for the US-Ton AmeriFlux site was provided by the U.S. Department of Energy's Office of Science. This research was also supported by the NASA Ecostress project. We thank Siyan Ma for contributing to the collection and processing of US-Ton’s in situ data. USDA is an equal opportunity provider and employer.

The FLUXNET2015 dataset provides ecosystem-scale data on CO2, water, and energy exchange between the biosphere and the atmosphere, and other meteorological and biological measurements, from 212 sites around the globe (over 1500 site-years, up to and including year 2014). These sites, independently managed and operated, voluntarily contributed their data to create global datasets. Data were quality controlled and processed using uniform methods, to improve consistency and intercomparability across sites. The dataset is already being used in a number of applications, including ecophysiology studies, remote sensing studies, and development of ecosystem and Earth system models. FLUXNET2015 includes derived-data products, such as gap-flled time series, ecosystem respiration and photosynthetic uptake estimates, estimation of uncertainties, and metadata about the measurements, presented for the frst time in this paper. In addition, 206 of these sites are for the frst time distributed under a Creative Commons (CC-BY 4.0) license. This paper details this enhanced dataset and the processing methods, now made available as open-source codes, making the dataset more accessible, transparent, and reproducible., European Union (EU), United States Department of Energy (DOE)

Despite their size and contribution to the global carbon cycle, we have limited understanding of tropical savannas and their current trajectory with climate change and anthropogenic pressures. Here we examined interannual variability and externally forced long-term changes in carbon and water exchange from a high rainfall savanna site in the seasonal tropics of north Australia. We used an 18-year flux data time series (2001-2019) to detect trends and drivers of fluxes of carbon and water. Significant positive trends in gross primary productivity (GPP, 15.4 g C m2 y-2 ), ecosystem respiration (Reco , 8.0 g C m2 y-2 ), net ecosystem productivity (NEE, 7.4 g C m2 y-2 ) and ecosystem water use efficiency (WUE, 0.0077 g C mm-1 y-1 ) were computed. There was a weaker, non-significant trend in latent energy exchange (LE, 0.34 W m-2 y-1 ). Rainfall from a nearby site increased statistically over a 45-year period during the observation period. To examine dominant drivers of changes in GPP and WUE, we used a random forest approach and a terrestrial biosphere model to conduct an attribution experiment. Radiant energy was the dominant driver of wet season fluxes, whereas soil water content dominated dry season fluxes. The model attribution suggested that [CO2 ], growing season length and dry season Tair accounting for 90% of the modelled trend in GPP and WUE. Positive trends in fluxes were largest in the dry season implying tree components were a larger contributor than the grassy understory. Fluxes and environmental drivers were not significant during the wet season, the period when grasses are active. The site is potentially still recovering from a cyclone 45 years ago and regrowth from this event may also be contributing to the observed trends in sequestration, highlighting the need to understand fluxes and their drivers from sub-diurnal to decadal scales.

Water storage plays an important role in mitigating heat and flooding in urban areas. Assessment of the capacity of cities to store water remains challenging due to the extreme heterogeneity of the...

The Vegetation Optimality Model (VOM, Schymanski et al., 2009, 2015) is an optimality-based, coupled water-vegetation model that predicts vegetation properties and behaviour based on optimality theory, rather than calibrating vegetation properties or prescribing them based on observations, as most conventional models do. In order to determine wheter optimality theory can alleviate common shortcomings of conventional models, as identified in a previous model inter-comparison study along the North Australian Tropical Transect (NATT) (Whitley et al., 2016), a range of updates to previous applications of the VOM have been made for increased generality and improved comparability with conventional models. To assess in how far the updates to the model and input data would have affected the original results, we implemented them one-by-one while reproducing the analysis of Schymanski et al. (2015). The model updates included extended input data, the use of variable atmospheric CO2-levels, modified soil properties, implementation of free drainage conditions, and the addition of grass rooting depths to the optimized vegetation properties. A systematic assessment of these changes was carried out by adding each individual modification to the original version of the VOM at the flux tower site of Howard Springs, Australia. The analysis revealed that the implemented changes affected the simulation of mean annual evapo-transpiration (ET) and gross primary productivity (GPP) by no more than 20 %, with the largest effects caused by the newly imposed free drainage conditions and modified soil texture. Free drainage conditions led to an underestimation of ET and GPP, whereas more fine-grained soil textures increased the water storage in the soil and resulted in increased GPP. Although part of the effect of free drainage was compensated for by the updated soil texture, when combining all changes, the resulting effect on the simulated fluxes was still dominated by the effect of implementing free drainage conditions. Eventually, the relative error for the mean annual ET, in comparison with flux tower observations, changed from an 8.4 % overestimation to an 10.2 % underestimation, whereas the relative errors for the mean annual GPP stayed similar with a change from 17.8 % to 14.7 %. The sensitivity to free drainage conditions suggests that a realistic representation of groundwater dynamics is very important for predicting ET and GPP at a tropical open-forest savanna site as investigated here. The modest changes in model outputs highlighted the robustness of the optimization approach that is central to the VOM architecture.

Gross primary productivity (GPP) of wooded ecosystems (forests and savannas) is central to the global carbon cycle, comprising 67%-75% of total global terrestrial GPP. Climate change may alter this flux by increasing the frequency of temperatures beyond the thermal optimum of GPP (Topt ). We examined the relationship between GPP and air temperature (Ta) in 17 wooded ecosystems dominated by a single plant functional type (broadleaf evergreen trees) occurring over a broad climatic gradient encompassing five ecoregions across Australia ranging from tropical in the north to Mediterranean and temperate in the south. We applied a novel boundary-line analysis to eddy covariance flux observations to (a) derive ecosystem GPP-Ta relationships and Topt (including seasonal analyses for five tropical savannas); (b) quantitatively and qualitatively assess GPP-Ta relationships within and among ecoregions; (c) examine the relationship between Topt and mean daytime air temperature (MDTa) across all ecosystems; and (d) examine how down-welling short-wave radiation (Fsd) and vapour pressure deficit (VPD) influence the GPP-Ta relationship. GPP-Ta relationships were convex parabolas with narrow curves in tropical forests, tropical savannas (wet season), and temperate forests, and wider curves in temperate woodlands, Mediterranean woodlands, and tropical savannas (dry season). Ecosystem Topt ranged from 15℃ (temperate forest) to 32℃ (tropical savanna-wet and dry seasons). The shape of GPP-Ta curves was largely determined by daytime Ta range, MDTa, and maximum GPP with the upslope influenced by Fsd and the downslope influenced by VPD. Across all ecosystems, there was a strong positive linear relationship between Topt and MDTa (Adjusted R2 : 0.81; Slope: 1.08) with Topt exceeding MDTa by >1℃ at all but two sites. We conclude that ecosystem GPP has adjusted to local MDTa within Australian broadleaf evergreen forests and that GPP is buffered against small Ta increases in the majority of these ecosystems.

‘Aerodynamic resistance’ (hereafter ra) is a preeminent variable in the modelling of evapotranspiration (ET), and its accurate quantification plays a critical role in determining the performance and consistency of thermal remote sensing-based surface energy balance (SEB) models for estimating ET at local to regional scales. Atmospheric stability links ra with land surface temperature (LST) and the representation of their interactions in the SEB models determines the accuracy of ET estimates.The present study investigates the influence of ra and its relation to LST uncertainties on the performance of three structurally different SEB models by combining nine OzFlux eddy covariance datasets from 2011 to 2019 from sites of different aridity in Australia with MODIS Terra and Aqua LST and leaf area index (LAI) products. Simulations of the latent heat flux (LE, energy equivalent of ET in W/m2) from the SPARSE (Soil Plant Atmosphere and Remote Sensing Evapotranspiration), SEBS (Surface Energy Balance System) and STIC (Surface Temperature Initiated Closure) models forced with MODIS LST, LAI, and in-situ meteorological datasets were evaluated using observed flux data across water-limited (semi-arid and arid) and radiation-limited (mesic) ecosystems.Our results revealed that the three models tend to overestimate instantaneous LE in the water-limited shrubland, woodland and grassland ecosystems by up to 60% on average, which was caused by an underestimation of the sensible heat flux (H). LE overestimation was associated with discrepancies in ra retrievals under conditions of high atmospheric instability, during which errors in LST (expressed as the difference between MODIS LST and in-situ LST) apparently played a minor role. On the other hand, a positive bias in LST coincides with low ra and causes slight underestimation of LE at the water-limited sites. The impact of ra on the LE residual error was found to be of the same magnitude as the influence of errors in LST in the semi-arid ecosystems as indicated by variable importance in projection (VIP) coefficients from partial least squares regression above unity. In contrast, our results for mesic forest ecosystems indicated minor dependency on ra for modelling LE (VIPa.

The errors and uncertainties associated with gap-filling algorithms of water, carbon and energy fluxes data, have always been one of the prominent challenges of the global network of microclimatological tower sites that use eddy covariance (EC) technique. To address this concern, and find more efficient gap-filling algorithms, we reviewed eight algorithms to estimate missing values of environmental drivers, and separately three major fluxes in EC time series. We then examined the performance of mentioned algorithms for different gap-filling scenarios utilising data from five OzFlux Network towers during 2013. The objectives of this research were (a) to evaluate the impact of training and testing window lengths on the performance of each algorithm; (b) to compare the performance of traditional and new gap-filling techniques for the EC data, for fluxes and their corresponding meteorological drivers. The performance of algorithms was evaluated by generating nine different training-testing window lengths, ranging from a day to 365 days. In each scenario, the gaps covered the data for the entirety of 2013 by consecutively repeating them, where, in each step, values were modelled by using earlier window data. After running each scenario, a variety of statistical metrics was used to evaluate the performance of the algorithms. The algorithms showed different levels of sensitivity to training-testing windows; The Prophet Forecast Model (FBP) revealed the most sensitivity, whilst the performance of artificial neural networks (ANNs), for instance, did not vary considerably by changing the window length. The performance of the algorithms generally decreased with increasing training-testing window length, yet the differences were not considerable for the windows smaller than 60 days. Gap-filling of the environmental drivers showed there was not a significant difference amongst the algorithms, the linear algorithms showed slight superiority over those of machine learning (ML), except the random forest algorithm estimating the ground heat flux (RMSEs of 30.17 and 34.93 for RF and CLR respectively). For the major fluxes, though, ML algorithms showed superiority (9 % less RMSE on average), except the Support Vector Regression (SVR), which provided significant bias in its estimations. Even though ANNs, random forest (RF) and extreme gradient boost (XGB) showed close performance in gap-filling of the major fluxes, RF provided more consistent results with less bias, relatively. The results indicated that there is no single algorithm which outperforms in all situations and therefore, but RF is a potential alternative for the ANNs as regards flux gap-filling.

Terrestrial evapotranspiration (ET) is thermodynamically expected to increase with increasing atmospheric temperature; however, the actual constraints on the intensification of ET remain uncertain due to a lack of direct observations. Based on the FLUXNET2015 Dataset, we found that relative humidity (RH) is a more important driver of ET than temperature. While actual ET decrease at reduced RH, potential ET increases, consistently with the complementary relationship (CR) framework stating that the fraction of energy not used for actual ET is dissipated as increased sensible heat flux that in turn increases potential ET. In this study, we proposed an improved CR formulation requiring no parameter calibration and assessed its reliability in estimating ET both at site-level with the FLUXNET2015 Dataset and at basin-level. Using the ERA-Interim meteorological dataset for 1979–2017 to calculate ET, we found that the global terrestrial ET showed an increasing trend until 1998, while the trend started to decline afterwards. Such decline was largely associated with a reduced RH, inducing water stress conditions that triggered stomatal closure to conserve water. For the first time, this study quantified the global-scale implications of changes in RH on terrestrial ET, indicating that the temperature-driven acceleration of the terrestrial water cycle will be likely constrained by terrestrial vegetation feedbacks.

Vegetation properties such as rooting depths and vegetation cover play a key role in coupling ecological and hydrological processes. These properties are however highly variable in space and/or time and their parametrization generally poses challenges for terrestrial biosphere models (Whitley et al., 2016). Models often use static values for dynamic vegetation properties or prescribe values based on observations, such as remotely sensed leaf area index. Here, vegetation optimality provides a way forward in order to predict such vegetation properties and their response to environmental change (Schymanski et al., 2015).In this study, we explore the utility of a combined water-vegetation model, the Vegetation Optimality Model (VOM, Schymanski et al., 2009), to predict vegetation properties such as rooting depths, foliage cover, photosynthetic capacity and water use strategies. The VOM schematizes perennial trees and seasonal grasses each as a single big leaf with an associated root system and optimizes leaf and root system properties in order to maximize the Net Carbon Profit, i.e. the difference between the total carbon taken up by photosynthesis and all the carbon costs related to the construction and maintenance of the plant organs involved. The VOM was applied along the North-Australian Tropical Transect, which consists of six savanna sites equipped with flux towers along a strong rainfall gradient between 500 and 1700 mm per year. The multi-annual half-hourly measurements of evaporation and CO2-assimilation at the different sites were used here to evaluate the model.The VOM produced similar or better results than more traditional models even though it requires much less information about site-specific vegetation properties. However, we found a persistent bias in the predicted vegetation cover. More detailed numerical experiments revealed a likely misrepresentation of the foliage costs in the model, which are based on a linear relation between leaf area and fractional vegetation cover. This finding, and the already favourable comparison with traditional models, implies that optimization of vegetation properties for Net Carbon Profit is a very promising approach for predicting the soil-vegetation-atmosphere exchange of water and carbon in complex ecosystems such as savannas.ReferencesSchymanski, S.J., Roderick, M.L., Sivapalan, M., 2015. Using an optimality model to understand medium and long-term responses of vegetation water use to elevated atmospheric CO2 concentrations. AoB PLANTS 7, plv060. https://doi.org/10.1093/aobpla/plv060Schymanski, S.J., Sivapalan, M., Roderick, M.L., Hutley, L.B., Beringer, J., 2009. An optimality‐based model of the dynamic feedbacks between natural vegetation and the water balance. Water Resources Research 45. https://doi.org/10.1029/2008WR006841Whitley, R., Beringer, J., Hutley, L.B., Abramowitz, G., De Kauwe, M.G., Duursma, R., Evans, B., Haverd, V., Li, L., Ryu, Y., Smith, B., Wang, Y.-P., Williams, M., Yu, Q., 2016. A model inter-comparison study to examine limiting factors in modelling Australian tropical savannas. Biogeosciences 13, 3245–3265. https://doi.org/10.5194/bg-13-3245-2016

Rising atmospheric CO2 concentration ([CO2]) enhances photosynthesis and reduces transpiration at the leaf, ecosystem, and global scale via the CO2 fertilization effect. The CO2 fertilization effect is among the most important processes for predicting the terrestrial carbon budget and future climate, yet it has been elusive to quantify. For evaluating the CO2 fertilization effect on land photosynthesis and transpiration, we developed a technique that isolated this effect from other confounding effects, such as changes in climate, using a noisy time series of observed land-atmosphere CO2 and water vapor exchange. Here, we evaluate the magnitude of this effect from 2000 to 2014 globally based on constraint optimization of gross primary productivity (GPP) and evapotranspiration in a canopy photosynthesis model over 104 global eddy-covariance stations. We found a consistent increase of GPP (0.138 ± 0.007% ppm−1; percentile per rising ppm of [CO2]) and a concomitant decrease in transpiration (−0.073% ± 0.006% ppm−1) due to rising [CO2]. Enhanced GPP from CO2 fertilization after the baseline year 2000 is, on average, 1.2% of global GPP, 12.4 g C m−2 yr−1 or 1.8 Pg C yr−1 at the years from 2001 to 2014. Our result demonstrates that the current increase in [CO2] could potentially explain the recent land CO2 sink at the global scale.

International audience; Leaf chlorophyll is central to the exchange of carbon, water and energy between the biosphere and the atmosphere, and to the functioning of terrestrial ecosystems. This paper presents the first spatially-continuous view of terrestrial leaf chlorophyll content (Chl(Leaf)) at the global scale. Weekly maps of Chl(Leaf) were produced from ENVISAT MERIS full resolution (300 m) satellite data using a two-stage physically-based radiative transfer modelling approach. Firstly, leaf-level reflectance was derived from top-of-canopy satellite reflectance observations using 4-Scale and SAIL canopy radiative transfer models for woody and non-woody vegetation, respectively. Secondly, the modelled leaf-level reflectance was input into the PROSPECT leaf-level radiative transfer model to derive Chl(Leaf). The Chl(Leaf) retrieval algorithm was validated using measured Chi(Leaf) data from 248 sample measurements at 28 field locations, and covering six plant functional types (PFTs). Modelled results show strong relationships with field measurements, particularly for deciduous broadleaf forests (R-2 = 0.67; RMSE = 9.25 mu g cm(-2); p < 0.001), croplands (R-2 = 0.41; RMSE = 13.18 mu g cm(-2); p < 0.001) and evergreen needleleaf forests (R-2 = 0.47; RMSE = 10.63 mu g cm(-2); p < 0.001). When the modelled results from all PFTs were considered together, the overall relationship with measured Chl(Leaf )remained good (R-2 = 0.47, RMSE = 10.79 mu g cm(-2); p < 0.001). This result is an improvement on the relationship between measured Chl(Leaf) and a commonly used chlorophyll-sensitive spectral vegetation index; the MERIS Terrestrial Chlorophyll Index (MTCI; R-2 = 0.27, p < 0.001). The global maps show large temporal and spatial variability in Chl(Leaf), with evergreen broadleaf forests presenting the highest leaf chlorophyll values, with global annual median values of 54.4 mu g cm(-2). Distinct seasonal Chl(Leaf) phenologies are also visible, particularly in deciduous plant forms, associated with budburst and crop growth, and leaf senescence. It is anticipated that this global Chl(Leaf) product will make an important step towards the explicit consideration of leaf-level biochemistry in terrestrial water, energy and carbon cycle modelling.

Thermal infrared sensing of evapotranspiration (E) through surface energy balance (SEB) models is challenging due to uncertainties in determining the aerodynamic conductance (g(A)) and due to inequalities between radiometric (T-R) and aerodynamic temperatures (T-0). We evaluated a novel analytical model, the Surface Temperature Initiated Closure (STIC1.2), that physically integrates T-R observations into a combined Penman-Monteith Shuttleworth-Wallace (PM-SW) framework for directly estimating E, and overcoming the uncertainties associated with T0 and gA determination. An evaluation of STIC1.2 against high temporal frequency SEB flux measurements across an aridity gradient in Australia revealed a systematic error of 10-52% in E from mesic to arid ecosystem, and low systematic error in sensible heat fluxes (H) (12-25%) in all ecosystems. Uncertainty in TR versus moisture availability relationship, stationarity assumption in surface emissivity, and SEB closure corrections in E were predominantly responsible for systematic E errors in arid and semi-arid ecosystems. A discrete correlation (r) of the model errors with observed soil moisture variance (r = 0.33-0.43), evaporative index (r = 0.77-0.90), and climatological dryness (r = 0.60-0.77) explained a strong association between ecohydrological extremes and T-R in determining the error structure of STIC1.2 predicted fluxes. Being independent of any leaf-scale biophysical parameterization, the model might be an important value addition in working group (WG2) of the Australian Energy and Water Exchange (OzEWEX) research initiative which focuses on observations to evaluate and compare biophysical models of energy and water cycle components. Plain Language Summary Evapotranspiration modeling and mapping in arid and semi-arid ecosystems are uncertain due to empirical approximation of surface and atmospheric conductances. Here we demonstrate the performance of a fully analytical model which is independent of any leaf-scale empirical parameterization of the conductances and can be potentially used for continental scale mapping of ecosystem water use as well as water stress using thermal remote sensing satellite data.d

© 2017 Elsevier B.V. Deforestation and forest degradation cause the deterioration of resources and ecosystem services. However, there are still no operational indicators to measure forest status, especially for forest degradation. In the present study, we analysed the thermal response number (TRN, calculated by daily total net radiation divided by daily temperature range) of 163 sites including mature forest, disturbed forest, planted forest, shrubland, grassland, savanna vegetation and cropland. TRN generally increased with latitude, however the regression of TRN against latitude differed among vegetation types. Mature forests are superior as thermal buffers, and had significantly higher TRN than disturbed and planted forests. There was a clear boundary between TRN of forest and non-forest vegetation (i.e. grassland and savanna) with the exception of shrubland, whose TRN overlapped with that of forest vegetation. We propose to use the TRN of local mature forest as the optimal TRN (TRNopt). A forest with lower than 75% of TRNopt was identified as subjected to significant disturbance, and forests with 66% of TRNopt was the threshold for deforestation within the absolute latitude from 30° to 55°. Our results emphasized the irreplaceable thermal buffer capacity of mature forest. TRN can be used for early warning of deforestation and degradation risk. It is therefore a valuable tool in the effort to protect forests and prevent deforestation.

Savanna regions are increasingly developed for agriculture to support population growth, food demand and export economies. This is driving interest in the conversion of natural savanna to cattle grazed pastures or horticultural crops in Northern Australia. Savanna clearing leads to aboveground carbon (C) loss but impacts below-ground are less certain as studies have focused on high rainfall regions, low reactive clay soils and shallow depths of 0.3 m or less. To examine impact of land use change (LUC) in Northern Australia, we sampled the upper 1.0 m soil profile of: 1) savanna woodlands, 2) old and young C4 perennial cattle grazed pastures and 3) old and young, melon-sorghum rotation fields. Soil C concentrations and soil bulk density in the upper 0.3 m were significantly greater in cattle grazed pastures than savanna woodland, such that savanna to pasture LUC significantly increased soil C density from 30.2 ± 1.5–43.4 ± 3.0 Mg C ha-1 over 28 years. Soil bulk density in the upper 0.7 m of savanna woodland was significantly less than that in the melon-sorghum rotations. In contrast, soil C concentrations below 0.5 m were significantly greater in the savanna woodland than in the 3 y.o. melon-sorghum rotation fields. There was no significant difference in soil C density between savanna woodland and 12 y.o. melon-sorghum rotations. This may relate to the beneficial effects of ploughing in sorghum as a green manure, or simply that the decrease in soil C concentration was offset by increased soil C density in the upper layers of soil where most C is located. In contrast, conversion of savanna to cattle grazed pasture sites is equivalent to a soil sequestration rate of 0.34 Mg C ha-1 y-1, approximately 50% of the observed net ecosystem productivity estimated using eddy covariance methods. Changing the depth to which soil C density was estimated changed significant differences amongst land-use systems. Whereas, estimating soil C density using a ‘fixed depth’ or ‘equivalent mass’ method did not change the statistical significance. There is an increasing focus on the potential of soil carbon sequestration especially in north Australia and understanding potential change and impact of accounting methods is essential for meaningful policy development. Policy settings to maintain, or improve, soil C storage and soil health may contradict agricultural development goals.

The following authors were omitted from the original version of this Data Descriptor: Markus Reichstein and Nicolas Vuichard. Both contributed to the code development and N. Vuichard contributed to the processing of the ERA-Interim data downscaling. Furthermore, the contribution of the co-author Frank Tiedemann was re-evaluated relative to the colleague Corinna Rebmann, both working at the same sites, and based on this re-evaluation a substitution in the co-author list is implemented (with Rebmann replacing Tiedemann). Finally, two affiliations were listed incorrectly and are corrected here (entries 190 and 193). The author list and affiliations have been amended to address these omissions in both the HTML and PDF versions. © 2021, This is a U.S. government work and not under copyright protection in the U.S.; foreign copyright protection may apply.

In this work we have developed a random forest regressor to predict daily evapotranspiration in three eddy-covariance sites in Northern Australia from in-situ meteorological data and fluxes, and satellite leaf area index and land surface temperature data. The variable analysis for the random forest regressor suggests that leaf area index is the most important one at this temporal scale. A development sample corresponding to the period 2010–2013 was used, while the year 2014 has been separated for testing. Using this approach, we have obtained satisfactory performances in the three sites, with RMSE errors around 1 mm/day (and relative RMSEs ~ 0.3 ) in comparison to the measured values. With the final aim of testing the predictive capability of a model that uses remote sensing data as input, regressors that predict evapotranspiration exclusively from leaf area index and land surface temperature, have been trained resulting in reasonable performances. The RMSEs over the test set are ~ 1.1 − 1.2 mm/day. In all cases, the errors are comparable to those obtained in previous works that use similar locations and different methods. When compared to the measured values, the random forest predictions of evapotranspiration are more accurate than the global MODIS ET 8-day 1 km product, which was used as benchmark for the performance evaluation of our models, in the three selected locations.

The savanna complex is a highly diverse global biome that occurs within the seasonally dry tropical to sub-tropical equatorial latitudes and are structurally and functionally distinct from grasslands and forests. Savannas are open-canopy environments that encompass a broad demographic continuum, often characterised by a changing dominance between C3-tree and C4-grass vegetation, where frequent environmental disturbances such as fire modulates the balance between ephemeral and perennial life forms. Climate change is projected to result in significant changes to the savanna floristic structure, with increases to woody biomass expected through CO2 fertilisation in mesic savannas and increased tree mortality expected through increased rainfall interannual variability in xeric savannas. The complex interaction between vegetation and climate that occurs in savannas has traditionally challenged terrestrial biosphere models (TBMs), which aim to simulate the interaction between the atmosphere and the land surface to predict responses of vegetation to changing in environmental forcing. In this review, we examine whether TBMs are able to adequately represent savanna fluxes and what implications potential deficiencies may have for climate change projection scenarios that rely on these models. We start by highlighting the defining characteristic traits and behaviours of savannas, how these differ across continents and how this information is (or is not) represented in the structural framework of many TBMs. We highlight three dynamic processes that we believe directly affect the water use and productivity of the savanna system: phenology, root-water access and fire dynamics. Following this, we discuss how these processes are represented in many current-generation TBMs and whether they are suitable for simulating savanna fluxes.Finally, we give an overview of how eddy-covariance observations in combination with other data sources can be used in model benchmarking and intercomparison frameworks to diagnose the performance of TBMs in this environment and formulate road maps for future development. Our investigation reveals that many TBMs systematically misrepresent phenology, the effects of fire and root-water access (if they are considered at all) and that these should be critical areas for future development. Furthermore, such processes must not be static (i.e. prescribed behaviour) but be capable of responding to the changing environmental conditions in order to emulate the dynamic behaviour of savannas. Without such developments, however, TBMs will have limited predictive capability in making the critical projections needed to understand how savannas will respond to future global change.

Prolonged drought has threatened traditional potable urban water supplies in Australian cities, reducing capability to adapt to climate change and mitigate against extreme. Integrated urban water management (IUWM) approaches, such as water sensitive urban design (WSUD), reduce the reliance on centralised potable water supply systems and provide a means for retaining water in the urban environment through stormwater harvesting and reuse. This study examines the potential for WSUD to provide cooling benefits and reduce human exposure and heat stress and thermal discomfort. A high-resolution observational field campaign, measuring surface level microclimate variables and remotely sensed land surface characteristics, was conducted in a mixed residential suburb containing WSUD in Adelaide, South Australia. Clear evidence was found that WSUD features and irrigation can reduce surface temperature (T s) and air temperature (T a) and improve human thermal comfort (HTC) in urban environments. The average 3 pm T a near water bodies was found to be up to 1.8 °C cooler than the domain maximum. Cooling was broadly observed in the area 50 m downwind of lakes and wetlands. Design and placement of water bodies were found to affect their cooling effectiveness. HTC was improved by proximity to WSUD features, but shading and ventilation were also effective at improving thermal comfort. This study demonstrates that WSUD can be used to cool urban microclimates, while simultaneously achieving other environmental benefits, such as improved stream ecology and flood mitigation.

This work is distributed under the Creative Commons Attribution 3.0 License., Forest ecosystems play a crucial role in the global carbon cycle by sequestering a considerable fraction of anthropogenic CO2, thereby contributing to climate change mitigation. However, there is a gap in our understanding about the carbon dynamics of eucalypt (broadleaf evergreen) forests in temperate climates, which might differ from temperate evergreen coniferous or deciduous broadleaved forests given their fundamental differences in physiology, phenology and growth dynamics. To address this gap we undertook a 3-year study (2010–2012) of eddy covariance measurements in a dry temperate eucalypt forest in southeastern Australia. We determined the annual net carbon balance and investigated the temporal (seasonal and inter-annual) variability in and environmental controls of net ecosystem carbon exchange (NEE), gross primary productivity (GPP) and ecosystem respiration (ER). The forest was a large and constant carbon sink throughout the study period, even in winter, with an overall mean NEE of −1234 ± 109 (SE) g C m−2 yr−1. Estimated annual ER was similar for 2010 and 2011 but decreased in 2012 ranging from 1603 to 1346 g C m−2 yr−1, whereas GPP showed no significant inter-annual variability, with a mean annual estimate of 2728 ± 39 g C m−2 yr−1. All ecosystem carbon fluxes had a pronounced seasonality, with GPP being greatest during spring and summer and ER being highest during summer, whereas peaks in NEE occurred in early spring and again in summer. High NEE in spring was likely caused by a delayed increase in ER due to low temperatures. A strong seasonal pattern in environmental controls of daytime and night-time NEE was revealed. Daytime NEE was equally explained by incoming solar radiation and air temperature, whereas air temperature was the main environmental driver of night-time NEE. The forest experienced unusual above-average annual rainfall during the first 2 years of this 3-year period so that soil water content remained relatively high and the forest was not water limited. Our results show the potential of temperate eucalypt forests to sequester large amounts of carbon when not water limited. However, further studies using bottom-up approaches are needed to validate measurements from the eddy covariance flux tower and to account for a possible underestimation in ER due to advection fluxes.

Standardised, quality-controlled and robust data from flux networks underpin the understanding of ecosystem processes and tools necessary to support the management of natural resources, including water, carbon and nutrients for environmental and production benefits. The Australian regional flux network (OzFlux) currently has 23 active sites and aims to provide a continental-scale national research facility to monitor and assess Australia's terrestrial biosphere and climate for improved predictions. Given the need for standardised and effective data processing of flux data, we have developed a software suite, called the Dynamic INtegrated Gap-filling and partitioning for OzFlux (DINGO), that enables gap-filling and partitioning of the primary fluxes into ecosystem respiration (Fre) and gross primary productivity (GPP) and subsequently provides diagnostics and results. We outline the processing pathways and methodologies that are applied in DINGO (v13) to OzFlux data, including (1) gap-filling of meteorological and other drivers; (2) gap-filling of fluxes using artificial neural networks; (3) the u* threshold determination; (4) partitioning into ecosystem respiration and gross primary productivity; (5) random, model and u* uncertainties; and (6) diagnostic, footprint calculation, summary and results outputs. DINGO was developed for Australian data, but the framework is applicable to any flux data or regional network. Quality data from robust systems like DINGO ensure the utility and uptake of the flux data and facilitates synergies between flux, remote sensing and modelling.

Savanna landscapes are globally extensive and highly sensitive to climate change, yet the physical processes and climate phenomena which affect them remain poorly understood and therefore poorly represented in climate models. Both human populations and natural ecosystems are highly susceptible to precipitation variation in these regions due to the effects on water and food availability and atmosphere–biosphere energy fluxes. Here we quantify the relationship between climate phenomena and historical rainfall variability in Australian savannas and, in particular, how these relationships changed across a strong rainfall gradient, namely the North Australian Tropical Transect (NATT). Climate phenomena were described by 16 relevant climate indices and correlated against precipitation from 1900 to 2010 to determine the relative importance of each climate index on seasonal, annual and decadal timescales. Precipitation trends, climate index trends and wet season characteristics have also been investigated using linear statistical methods. In general, climate index–rainfall correlations were stronger in the north of the NATT where annual rainfall variability was lower and a high proportion of rainfall fell during the wet season. This is consistent with a decreased influence of the Indian–Australian monsoon from the north to the south. Seasonal variation was most strongly correlated with the Australian Monsoon Index, whereas yearly variability was related to a greater number of climate indices, predominately the Tasman Sea and Indonesian sea surface temperature indices (both of which experienced a linear increase over the duration of the study) and the El Niño–Southern Oscillation indices. These findings highlight the importance of understanding the climatic processes driving variability and, subsequently, the importance of understanding the relationships between rainfall and climatic phenomena in the Northern Territory in order to project future rainfall patterns in the region.

The coexistence of trees and grasses in savanna ecosystems results in marked phenological dynamics that vary spatially and temporally with climate. Australian savannas comprise a complex variety of life forms and phenologies, from evergreen trees to annual/perennial grasses, producing a boom–bust seasonal pattern of productivity that follows the wet–dry seasonal rainfall cycle. As the climate changes into the 21st century, modification to rainfall and temperature regimes in savannas is highly likely. There is a need to link phenology cycles of different species with productivity to understand how the tree–grass relationship may shift in response to climate change. This study investigated the relationship between productivity and phenology for trees and grasses in an Australian tropical savanna. Productivity, estimated from overstory (tree) and understory (grass) eddy covariance flux tower estimates of gross primary productivity (GPP), was compared against 2 years of repeat time-lapse digital photography (phenocams). We explored the phenology–productivity relationship at the ecosystem scale using Moderate Resolution Imaging Spectroradiometer (MODIS) vegetation indices and flux tower GPP. These data were obtained from the Howard Springs OzFlux/Fluxnet site (AU-How) in northern Australia. Two greenness indices were calculated from the phenocam images: the green chromatic coordinate (GCC) and excess green index (ExG). These indices captured the temporal dynamics of the understory (grass) and overstory (trees) phenology and were correlated well with tower GPP for understory (r2 = 0.65 to 0.72) but less so for the overstory (r2 = 0.14 to 0.23). The MODIS enhanced vegetation index (EVI) correlated well with GPP at the ecosystem scale (r2 = 0.70). Lastly, we used GCC and EVI to parameterise a light use efficiency (LUE) model and found it to improve the estimates of GPP for the overstory, understory and ecosystem. We conclude that phenology is an important parameter to consider in estimating GPP from LUE models in savannas and that phenocams can provide important insights into the phenological variability of trees and grasses.

This paper describes a dataset of LoRa radio channel metrics, soil moisture and temperature measurements, weather and flux tower observations, from an underground sensor network on a farm site in Western Australia. The dataset comprises half hourly readings from 10 months of continuous monitoring over 10 radio channels (5 transmitters and 2 receivers) for 24 sensors. R-notebooks are provided to to facilitate re-use of this data by documenting collection metadata and exploratory data analyses. The full dataset and metadata is available from DOI 10.26182/5d8036e5c4ca7

In our recent study in Global Change Biology (Li et al., ), we examined the relationship between solar-induced chlorophyll fluorescence (SIF) measured from the Orbiting Carbon Observatory-2 (OCO-2) and gross primary productivity (GPP) derived from eddy covariance flux towers across the globe, and we discovered that there is a nearly universal relationship between SIF and GPP across a wide variety of biomes. This finding reveals the tremendous potential of SIF for accurately mapping terrestrial photosynthesis globally.

Savannas, occupying a fifth of the global land surface, are characterized by the coexistence of trees and grasses. Accurate estimation of savanna evapotranspiration (ET) is vital for understanding the regional and global water balance and its feedback to climate. However, the overlapping phenology and different water-use patterns of trees and grasses constitute a major challenge for modeling efforts. To estimate savanna ET, we used a three-source ET model, partitioning ET among soil, trees, and grasses. To represent legacy effects of precipitation on ecosystem water use, the Normalized Ecosystem Drought Index (NEDI, i.e. a function of precipitation and potential evapotranspiration) was included to limit canopy conductances in the model and also in two other classic two-layer models (Shuttleworth-Wallace model and Penman-Monteith-Leuning model). The results of our model and the other models were tested and compared using tower-based eddy covariance flux data collected at six sites (including four savanna sites, one pasture site, and one grassland site) along a precipitation gradient in northern Australia, together with satellite-derived leaf area index, which was partitioned to represent the canopy dynamics of trees and grasses. Inclusion of NEDI significantly reduced seasonal biases in ET estimation results for all models compared with observations at savanna sites (fitted slopes were closer to unity by 0.08–0.10, R2 increased by 0.03–0.04, and RMSE decreased by 0.07–0.09 mm d−1). The three-source model provides insights into simulation of water fluxes over vegetated areas of complex composition. Our work makes a contribution to savanna research by determining a flexible indicator defining the seasonal water availability limitation on savanna ET. The inclusion of NEDI in ET models could guide future research on modeling ecosystem water and carbon fluxes in response to seasonal droughts.