My Work at BGI | News | Publications
My work has focused on how plants use water, including investigating how carbon and water cycles interact, as well as estimating transpiration across a wide variety of ecosystems. I primarly use a data driven approach which utilizes machine learning, but is also guided by physiological understanding. Currently, I'm working with the FLUXCOM team to produce the next generation of global data driven estimates of terrestrial carbon, energy, and water fluxes.
The leaf economics spectrum 1,2 and the global spectrum of plant forms and functions 3
revealed fundamental axes of variation in plant traits, which represent different
ecological strategies that are shaped by the evolutionary development of plant
species 2 . Ecosystem functions depend on environmental conditions and the traits of
species that comprise the ecological communities 4 . However, the axes of variation of
ecosystem functions are largely unknown, which limits our understanding of how
ecosystems respond as a whole to anthropogenic drivers, climate and environmental
variability 4,5 . Here we derive a set of ecosystem functions 6 from a dataset of surface
gas exchange measurements across major terrestrial biomes. We find that most of the
variability within ecosystem functions (71.8%) is captured by three key axes. The first
axis reflects maximum ecosystem productivity and is mostly explained by vegetation
structure. The second axis reflects ecosystem water-use strategies and is jointly
explained by variation in vegetation height and climate. The third axis, which
represents ecosystem carbon-use efficiency, features a gradient related to aridity, and
is explained primarily by variation in vegetation structure. We show that two
state-of-the-art land surface models reproduce the first and most important axis of
ecosystem functions. However, the models tend to simulate more strongly correlated
functions than those observed, which limits their ability to accurately predict the full
range of responses to environmental changes in carbon, water and energy cycling in
terrestrial ecosystems 7,8 .
Check out these projects at EGU this year from two excellent PhDs I get the chance to work with:
Improved eddy covariance flux based transpiration estimates at high relative humidity and comparison to sap flux
by Weijie Zhang
Eddy covariance (EC) directly measures evapotranspiration (ET), which consists of transpiration and evaporation (E) from the soil and other surfaces. For process understanding it is pivotal to separate ET into its components. Yet, its computation is highly sensitive to the methodology used to estimate T. Among the multiple methods proposed in recent years, T has been estimated from EC via the Transpiration Estimation Algorithm (TEA, Nelson et al., 2020), and from the sap flux measurement network SAPFLUXNET (Poyatos et al., 2020). These methods are applicable to a large number of measurement sites worldwide, and can help constrain the global estimates of the ratio of T to ET, T/ET. While EC measures water and carbon fluxes across ecosystems globally, water vapor flux measurements can be underestimated at high relative humidity (Ibrom et al., 2007; Mammarella et al., 2009) causing errors in the measured ET and propagating into the predicted T.
Here we report a method to detect and correct the high relative humidity error caused by attenuation of high frequency in water vapor measurements of a closed-path EC system. Our results of the comparison between present water use efficiency (WUE) with previous TEA-based WUE show that the corrected WUE is lower at high relative humidity than that derived from previous TEA at the sub-daily scale. Besides, we compare the corrected T estimates from EC to concurrent SAPFLUXNET sites to show an improved relationship between sap flux and EC based T, T/ET, and WUE. Finally, we explore the main abiotic factors, such as vapor pressure deficit, air temperature, and precipitation, influencing WUE estimated from different T estimation methodologies. These results provide an improved data-driven approach to the ongoing research on ET partitioning and the factors influencing the WUE across ecosystems globally.
Adsorption of water vapor by soil in semi arid ecosystems: reconciling estimates from Lysimeters and Eddy Covariance
by Sinikka Paulus
Current climate change scenarios project altered rainfall frequencies which boosts scientific interest in ecosystems' responses to prolonged dry conditions. Under less rainfall, NRWI may play an increasingly important role, Yet, only sparse data are available to assess the role of non-rainfall water input (NRWI) during times of low water availability across ecoregions. Particularly, soil water vapor adsorption has received little attention at field scale. This term is used for the phase change of water from gas to liquid at highly negative matric potential. Under such conditions, water condensates already at relative humidity < 100%. The process has been broadly studied in laboratories but little is known from field experiments, which rarely cover periods longer than one month. Yet, several studies report soil water uptake from the atmosphere during soil surface cooling and in the early mornings. Lysimeters have played a strong role in quantifying these NRWI. Eddy Covariance (EC) measurements, in contrast, are known for their limited data quality under nighttime conditions when a stable boundary layer hinders the turbulent exchange of mass and energy. Therefore, EC has not been tested yet to trace soil adsorption.
In this contribution we adapt a methodology to derive NRWI from lysimeters data and compare them to EC measurements. We focus mainly on adsorption and evaluate the consistency between adsorption estimated with the lysimeters and negative (downward) latent heat (LE) fluxes from EC. We apply the method to a data set that comprises three years of observations from a semi-arid Spanish tree grass ecosystem.
Our results show that during the dry season the gradient in water vapour established between the atmosphere (more humid) and the soil pores (more dry) leads to adsorption by the soil. The observations from both instruments suggest that during the dry season, nightly transport of humidity from the atmosphere towards the ground is driven by soil vapor adsorption. This process occurs each night typically in the second half, but begins increasingly earlier in the evening the dryer the conditions are. The amount of water adsorbed is not directly comparable between EC and the lysimeter readings. With the latter, we quantified a yearly mean uptake between 8.8 mm and 25 mm per year. With the lysimeters we measure additionally 23.1 mm of water that condenses as dew and fog in winter, when EC is impeded by stable conditions. We further analyze EC LE measurements from different sites to evaluate if adsorption can be detected from EC data collected at different locations.
We conclude that the temporal patterns of adsorption estimates from lysimeters match the nighttime negative LE data from the EC technique, although the absolute numbers are uncertain. This might open interesting perspective to fill the knowledge gap of the role of soil water vapor adsorption from the atmosphere at field scale and open the opportunity to broaden the topic across ecosystem research communities. Our results also highlight a potential shortcoming in the interpretation of EC measurements in the case that negative nighttime values, representing physically plausible adsorption, are neglected.
Exciting new research from https://www.bgc-jena.mpg.de/bgi/index.php/People/TarekEl-Madany? and the MANIP team showing how a savanna ecosystem responds to nutrient inputs. Using the TEA algorithm, they were able to show how phosphorus addition leads to an increase in ecosystem water use efficiency.
The availability of nutrients like nitrogen (N) and phosphorus (P) is important for every living organism on Earth. Due to human activities, especially combustion processes large amounts of N are transported into the atmosphere and ecosystems. Therefore, ecosystems receive additional N but no other nutrients. We are investigating if the addition of N alone will lead to deficits in other nutrients and thus impact the functioning of ecosystems. Hence, we set up a large‐scale ecosystem experiment in a Mediterranean tree‐grass ecosystem where we fertilized two plots with N (16.9 ha) and N+P (21.5 ha). A third plot served as the control treatment. While the N‐only treatment created an imbalance between the available N and P, this imbalance was relieved in the N+P treatment where both N and P were provided. Our measurements showed that both fertilized treatments increased their carbon uptake and turned the ecosystem from a carbon source to carbon neutral. One of the main differences between the fertilized treatments which is associated with the imbalance of available N and P is the loss of water through the vegetation (transpiration). This increase in transpiration was only observed in the N‐only but not in the N+P treatment. Our results show, that the N:P stoichiometric imbalance, resulting from N‐only addition, increases the water‐use efficiency (i.e. the carbon gain per water loss) less than the addition of N + P, due to the strong increase in transpiration at the N‐only treatment.
FLUXNET-ECN Spring Workshop 2021
I had great time participating in this years FLUXNET Early Career Network Spring Workshop. In case you missed my talk, you can find it here:
or watch it directly here:
Code and examples of how to estimate transpiration from eddy covariance data.
We are working toward a manuscript on evapotranspiration partitioning of eddy covariance data. The paper will utilize three partitioning methods:
‣ Perez-Priego et al (2018). Partitioning Eddy Covariance Water Flux Components Using Physiological and Micrometeorological Approaches. Journal of Geophysical Research: Biogeosciences. https://doi.org/10.1029/2018JG004637
‣ Nelson et al (2018). Coupling Water and Carbon Fluxes to Constrain Estimates of Transpiration: The TEA Algorithm. Journal of Geophysical Research: Biogeosciences. https://doi.org/10.1029/2018JG004727
‣ Zhou et al (2016). Partitioning evapotranspiration based on the concept of underlying water use efficiency: ET PARTITIONING. Water Resources Research, 52(2), 1160–1175. https://doi.org/10.1002/2015WR017766
To run a tutorial for the TEA (Nelson et al 2018), uWUE (Zhou et al 2016), and Perez-Priego methods, check ou the links below. Each can be run within a browser without any installation!
If you try out the tutorial, please feel free to give me some feedback via email ⬆⬆⬆
Data and Code
Conferences and Workshops
- Examining Transpiration from Ecosystem to Global Scales, 5-7 September 2018 at the Max Planck Institute for Biogeochemistry in Jena, Germany. http://www.bgc.mpg.de/~jnelson/transpiration2018/
- Nelson, JA, N Carvalhais, M Cuntz, N Delpierre, J Knauer, M Migliavacca, J Ogee, M Reichstein, and M Jung. 2017. “Data Driven Estimation of Transpiration from Net Water Fluxes: The TEA Algorithm.” In AGU Fall Meeting Abstracts.
- Nelson, Jacob, Martin Jung, Nuno Carvalhais, Mirco Migliavacca, and Markus Reichstein. 2016. “Understanding Ecosystems’ Sub-Daily Water and Carbon Flux Changes during Dry-down Events.” In EGU General Assembly Conference Abstracts, 18:17549.