A03
Parameterization of evapotranspiration partitioning function in land-surface models using water stable isotopes
Prof. Dr. Youri Rothfuss
Forschungszentrum Jülich | +49-2461-96925 |
Prof. Dr. Nicolas Brüggemann
Forschungszentrum Jülich | +49-2461-618643 |
Daniel Schulz (PhD student)
Forschungszentrum Jülich | +49-2461-614035 |
Summary
In project A03, we analyze water component fluxes of the terrestrial water cycle by partitioning evapotranspiration (ET) into its components evaporation (E) and transpiration (T). The method is based on the non-destructive analysis of the temporal dynamics of water stable isotopologues, 1H2H16O and 1H218O, in atmospheric, soil and plant water. These analyses will be carried out in ecosystems with four specific plant functional types (PFTs) defined in CLM 5.0: managed irrigated crops, C3 grass, temperate climate needleleaf evergreen trees, and Mediterranean managed rainfed unirrigated crops. Using the gathered data, the isotope-enabled soil-vegetation-atmosphere transfer model SiSPAT-Isotope will be parameterized for field-scale simulations of T/ET for the analyzed PFTs and different climatic conditions. These modelling results will be compared to estimations of E and T using the CLM 5.0 and improve the PFT-specific modelling parameterization.
Graphical summary
Figure 1: Possibilities and complexity of using the 18O isotopic composition of water for partitioning evapotranspiration (Rothfuss et al., 2021).
Contribution to the CRC
The project contributes to the CRC’s modelling system CLM 5.0 by providing essential information for tuning sub-daily estimations of T/ET time series. Tuning data is calculated using the SiSPAT and SiSPAT-Isotope models, that directly incorporate the partitioning information collected in this study. The goal of the comparison is to statistically determine systematic errors in the CLM 5.0 estimations and to update the parameters used for T/ET calculation for the PFTs agricultural fields and grasslands. Finally, simulations with the improved CLM 5.0 will provide new values for PFT-specific characteristics used in the computation of E and T fluxes.
Approach
The isotopic partitioning of ET is based on a two-end member linear mixing model (assuming complete mixing of evaporated and transpired water in the atmosphere):
Measurements of the isotopic composition of the water in each of the three components E, T and ET therefore enable the assessment of the ratio of T/ET in environment over time. Multiple field campaigns will be carried out in the different PFT environments at different times of the year over the course of the project to account for seasonal variations of the T/ET ratio and to isotopically characterize T/ET of the different ecosystems during all seasons.
Establishment of isotopic measurement system
In 2022 a field campaign was carried out in a mediterranean, rainfed sunflower cropland (Auradé, southern France, summer season 21.06.-21.07.2022). Isotopic partitioning measurements were carried out at 2-hour resolution. was determined non-destructively from measurements of the isotopic composition of atmospheric water vapor, with contemporaneous destructive sampling of soil water (δ_E) and plant xylem (δ_T). The isotopic partitioning results were validated by additional partitioning estimates based on non-isotopic water flux measurements using eddy covariance (ET flux), sap flow sensors (T flux) and microlysimeters (E flux). From ICOS-station data T/ET was calculated additionally by using the water use efficiency WUE concept limited to the campaign period. Lastly T/ET was modelled using the SiSPAT and SiSPAT-Isotope models, as well as WUE partitioning modelled for the full year 2022. The Comparison of all time series is shown in Figure 2.
Figure 2: Partitioning results Auradé 2022
The comparison between partitioning results shows, that 2-hourly partitioning results for 18O are comparable to the daily estimates for all other presented methods, except the daily WUE T/ET, which are overestimated due to the high transpiration flux in the campaign period. Considering the full year for WUE-based partitioning shifts T/ET closer to the other models and field measurements. T/ET using 2H is overestimated in comparison due to the calculation procedure for soil water isotopic flux. The partitioning results of this field campaign show that 18O-based T/ET estimates exhibit realistic values for T/ET at a high temporal resolution compared to the other presented approaches. These results validate the applicability of the water stable isotopic ET partitioning.
Figure 3: Field campaign in summer 2022 in the PFT ‘Mediterranean managed rain-fed unirrigated crops’, Auradé, FR.
Figure 4: Field campaign in autumn 2022 in the PFT ‘Temperate grassland site’, Rollesbroich, DE.
Figure 5: Winter wheat spring field campaign Auradé 2023
High-resolution, long-term isotopic partitioning estimates
Figure 6: Isotopic partitioning field measurement setup
In 2025 a field campaign was carried out at a grassland site between 05.05.-28.09 with an optimized isotopic measurement setup and water calibration system enabling T/ET estimates at hourly resolution. The isotopic measurement system was expanded by incorporating an approach to non-destructively measure the soil water isotopic composition for determination of δ_E (soil gas probes). Isotopic partitioning measurements were correlated to ET measurements using eddy covariance and environmental parameters measured at the field site using the local ICOS station. In total 100 high accuracy 18O T/ET estimates were collected with a T/ET error of < 8 %, aligning well with the temporal developments in environmental parameters and showing close correlation with the land cover management practices. Figure 7 shows an excerpt of the campaign around a plant cut on DoY 181, with corresponding isotopic partitioning results shown in Figure 8. Absolute T/ET values decrease significantly after the cut on DoY 182, while showing sub-daily temporal developments corresponding to the solar cycle and local weather patterns. The high-resolution ET partitioning results will be confronted against SiSPAT and SISPAT-Isotope modelling like in the 2022 field campaign.
Figure 7: Soil water content, precipitation, mean VPD, plant height and cumulative daily ET across the campain period
Figure 8: 18O partitioning across the campaign excerpt
Comparing partitioning using the destructive soil samples or the non-destructive soil measurement shows good agreement between the two approaches (compare Figure 9). The comparison shows that the application of the soil gas probes presents a meaningful alternative to the labor-intensive destructive sampling of soil material for the purpose of isotopic partitioning, providing the next step towards the development of a fully non-destructive measurement system for water stable isotopic ET partitioning.
Figure 6: Isotopic partitioning based on Kelling plot, plant samples and both soil samples/soil gas probes for determination of δ_E
