HS1.1.5

Soil hydraulic properties provide important information about soil behavior under unsaturated and saturated conditions. Often sampling of undisturbed soils is not possible and soil samples have to be repacked for laboratory analysis. The HYPROP® measuring system (METERgroup, Munich, Germany) is a convenient method for determination of soil water retention characteristics and unsaturated hydraulic conductivity of undisturbed soil samples. It measures the matric potential of the saturated and drying soil sample using two tensiometers placed at different depths. Although the tensiometers are based on a new design that theoretically withstands cavitation at higher tension values, they are still considered to operate in the low tension range. Since soil water retention properties in the low tension range are strongly influenced by soil structure and pore size distribution, we were interested in the changes in hydraulic properties when measured on disturbed and then repacked samples, and undisturbed soil samples. Therefore, we investigated the soil hydraulic properties of three different soil types using the evaporation method on undisturbed and repacked samples. The results provide important insights for the interpretation of the results when the collection of undisturbed samples is not possible, and for designing laboratory experiments with repacked soils.

How to cite: Pečan, U., Žvokelj, L., Ferlin, J., Zupanc, V., and Pintar, M.: Measurement of soil hydraulic properties of structured and repacked soils, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-3985, https://doi.org/10.5194/egusphere-egu21-3985, 2021.

Near surface soil hydraulic conductivity is an essential parameter for various hydrological, geotechnical, and environmental-related studies. Currently, many instruments are in practice for evaluating this parameter, both in field, and laboratory. The rainfall simulator (RS) and mini disc infiltrometer (MDI) are two instruments used for the indirect estimation of hydraulic conductivity by many researchers and engineers. However, both the devices differ in their working philosophy and evaluation methodology. While the RS works by considering large soil volumes and providing a positive soil pressure, the MDI works for small sampled volumes and supply negative boundary head. Therefore, the two devices can result in varying estimates of hydraulic conductivity. In this study, a comparative assessment is carried out between the saturated hydraulic conductivity (K_s) estimates from the two instruments using laboratory experiments for two different soil textures (loam and sand). The infiltration results from the RS are analyzed using the Green-Ampt method, and from the MDI is analyzed using the Zhang's method followed by the Kutilek and Nielson method to produce K_s values. The K_s results from both the instruments are compared with the values obtained using the laboratory falling-head permeameter test. A one-way ANOVA and the Fisher’s Least Significant Difference (LSD) test as a posthoc test are carried out to analyze the statistical significance of the differences in the estimates of K_s by the two devices. The results showed that the two devices produced varying K_s results for both the soil textures, with the MDI mean values being one order higher than the RS mean. Compared with the permeameter values, the mean values from the RS were closer to the permeameter than the MDI. However, the ANOVA test and the Fisher’s LSD test reported that the variations between the two devices with that of the permeameter were not significant for both the soil textures. On the other hand, the RS and MDI variations were reported significant by the ANOVA and post hoc test.

How to cite: Naik, A. P. and Pekkat, S.: A Comparative assessment of estimated soil hydraulic conductivity from rainfall simulator and infiltrometer using laboratory repacked soil samples, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7099, https://doi.org/10.5194/egusphere-egu21-7099, 2021.

For decades, soil erosion has been a major environmental problem as it degrades the most productive soil layers, which threatens, among other things, food production worldwide. Although these effects have been known for a long time, there are still a variety of challenges to mitigating soil erosion in different ecosystems. As climate change progresses, the risk of soil loss increases, making the preparation of effective solutions very urgent. A current research focus is on the restoration of a protective soil cover following disturbances in the vegetation layer, e.g., through the reestablishment of biological soil crust communities. These are often dominated by bryophytes in humid climates. So far, several studies examined the general protective influence of bryophytes against soil erosion, however only few of them addressed how individual species affect specific erosion processes in detail.

To fill this research gap we investigated the impact of six moss species on soil erosion, percolation and carbon relocation by means of rainfall simulations. Therefore, we used topsoil substrate from four sites in the Schönbuch Nature Park in South Germany which covers different kinds of bedrock and varying soil texture and pH. Subsequently, they were sieved by 6.3 mm and filled into metal infiltration boxes (40 x 30 cm) up to a height of 6.5 cm. The moss species differ in origin (either collected in the field or cultivated in the lab) as well as growth form (pleurocarpous or acrocarpous). Rainfall simulations were performed for bare soil substrates, as well as for moss-covered soil substrates six months later and both in dry and wet conditions. Additionally, we conducted rainfall simulations with leaf and coniferous litter on bare soil substrates. During the simulations we monitored soil moisture in two position - 3 cm depth plus soil surface - with biocrust wetness probes (BWP) and quantified surface runoff, percolation and sediment discharge. Afterwards we determined carbon contents of the sediment and dissolved organic carbon in the liquid phase of runoff and percolated water.

While surface runoff was increased by 5% due to the litter cover compared to the bare soil substrate, sediment discharge decreased to 97%. Runoff rates could also be mitigated by 90 % as a result of the moss cover. Furthermore, due to the dense moss cover sediment rates were almost reduced to zero. Preliminary results show that there are differences between the moss species in terms of sediment discharge, but not in context with runoff. The analyses of carbon contents in surface runoff and the percolated water are still in progress, as is the evaluation of the BWP measurements. These outcomes will be presented at vEGU21.

How to cite: Gall, C., Grabherr, L., Nebel, M., Scholten, T., Thielen, S. M., and Seitz, S.: On the effect of different moss species on soil erosion, percolation and carbon relocation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8389, https://doi.org/10.5194/egusphere-egu21-8389, 2021.

Carbon dioxide (CO₂) emissions from running waters represent a key component of the global carbon cycle. However, quantifying CO₂ fluxes across air-water boundaries remains challenging due to practical difficulties in the estimation of reach-scale standardized gas exchange velocities (k₆₀₀) and water equilibrium concentrations. Whereas craft-made floating chambers supplied by internal CO₂ sensors represent a promising technique to estimate CO₂ fluxes from rivers, the existing literature lacks of  rigorous  comparisons  among  differently  designed chambers and deployment techniques. Moreover, as of now the uncertainty of k₆₀₀ estimates from chamber data has not been evaluated.  Here, these issues were addressed analyzing the results of a flume experiment carried out in the Summer of 2019 in the Lunzer:::Rinnen - Experimental Facility (Austria). During the experiment, 100 runs were performed  using two different chamber designs (namely, a Standard Chamber and a Flexible Foil chamber with an external floating system and a flexible sealing) and two different deployment modes (drifting and anchored). The runs were performed using various combinations of discharge and channel slope, leading to variable turbulent kinetic energy dissipation rates (1.5 10^-3< ε < 1 10^-1 m² s^-3). Estimates of gas exchange velocities were in line with the existing literature (4 < k₆₀₀ < 32 m d^-1), with a general increase of k₆₀₀ for larger turbulent kinetic energy dissipation rates. The Flexible Foil chamber gave consistent k₆₀₀ patterns in response to changes in the slope and/or the flow rate. Moreover, Acoustic Doppler Velocimeter measurements indicated a limited increase of the turbulence induced by the Flexible Foil chamber on the flow field (22 % increase in ε, leading to a theoretical 5 % increase in k₆₀₀).
The  uncertainty  in  the  estimate  of  gas  exchange  velocities  was  then estimated  using  a  Generalized Likelihood Uncertainty Estimation (GLUE) procedure. Overall, uncertainty in k₆₀₀ was moderate to high, with enhanced uncertainty in high-energy setups. For the anchored mode, the standard deviations of k₆₀₀ were between 1.6 and 8.2 m d^-1, whereas significantly higher values were obtained in drifting mode. Interestingly, for the Standard Chamber the uncertainty was larger (+ 20 %) as compared to the Flexible Foil chamber.  Our study suggests that a Flexible Foil design and the anchored deployment might be useful techniques to enhance the robustness and the accuracy of CO₂ measurements in low-order streams. Furthermore, the study demonstrates the value of analytical and numerical tools in the identification of accurate estimations for gas exchange velocities.
These findings have important implications for improving estimates of greenhouse gas emissions and reaeration rates in running waters.

How to cite: Vingiani, F., Durighetto, N., Klaus, M., Schelker, J., Labasque, T., and Botter, G.: Evaluating stream CO2 outgassing via Drifting and Anchored flux chambers in a controlled flume experiment, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9177, https://doi.org/10.5194/egusphere-egu21-9177, 2021.

The continuous interaction between riparian vegetation and water has important effects on the hydraulics of a river, mainly onto the flood events propagation. Vegetation is a fundamental part of the river ecosystem, but its stage and growth need to be monitored and controlled, especially when the river passes through a densely urbanized area. In fact, vegetation obstructs the streamflow by reducing the hydraulic cross-section area and increasing the roughness of the floodplains and the relative flood risk.

In this study, experiments have been performed at the Fantoli Hydraulic Laboratory at Politecnico di Milano, to validate the methodologies that estimate the hydraulic roughness of vegetated river floodplains, starting from the vegetation properties such as size, density and elastic modulus of a case study. A model based on the mechanical properties of vegetation was used to identify the most suitable material to reproduce the dynamic behaviour of real vegetation on a laboratory scale. The tests were carried out for different spatial configurations of trees, densities and submerged conditions.

The analysis, in addition to relying on experimental work, involves the installation of six piezoresistive pressure sensors located both in the floodplains and in the main channel, to monitor head losses in a representative reach of the river under study. The field measurements allow validation of the approach used in laboratory tests.

How to cite: Herrera Gomez, L. V., Ravazzani, G., Ferri, M., and Mancini, M.: Hydraulic roughness estimation in vegetated floodplains, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14426, https://doi.org/10.5194/egusphere-egu21-14426, 2021.

The water storage in soil is a dynamic process that changes with soil, vegetation and climate properties. Water retention curves, that describe the relationship between the soil water content (θ) and the soil water potential (ψ), are used to model soil water flow and root water uptake by the plants. The overall objective of this study is to derive the retention curves of soils at two forested (Agia Marina, Platania) and two irrigated (Galata, Strakka) sites in Cyprus from in-situ soil moisture and soil water potential observations. 
The long-term (1980 – 2010) average annual rainfall at Strakka olive grove (255 m elevation), Agia Marina P. brutia forest (640 m), Galata peach orchard (784 m) and Platania P. brutia forest (1160 m) is 298, 425, 502 and 839 mm, respectively.  The average soil depth at Agia Marina is 14 cm, while at other sites it is around 1 m. We installed a total of 18 TEROS21 soil water potential sensors, 37 5TM and 19 SMT100 soil moisture sensors, at different soil depths at the four sites. 
Results from January 2019 to January 2021 show differences in the water retention curves of the four sites due to different soil textures. At the forested sites, θ reached wilting point at the summer period, indicating that trees extend their roots beyond the soil profile, to the bedrock in order to survive. At the irrigated sites, θ exceeds field capacity during irrigation, indicating over-irrigation. We found different water retention relations after rainfall and after irrigation, indicating that irrigation has an uneven spatial distribution. These findings suggest that the irrigation in these fields is not optimal and farmers may need to increase the number of irrigation drippers, while reducing the irrigation amount per dripper. From a monitoring perspective, increasing the number of sensors may give a better representation of the soil moisture conditions. 
The research has received financial support from the ERANETMED3 program, as part of the ISOMED project (Environmental Isotope Techniques for Water Flow Accounting), funded through the Cyprus Research and Innovation Foundation.

How to cite: Eliades, M., Bruggeman, A., Djuma, H., Siakou, M., Venetsanou, P., Zoumides, C., and Huebner, C.: Soil water dynamics in forested and irrigated sites in Cyprus, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14170, https://doi.org/10.5194/egusphere-egu21-14170, 2021.

Headwater streams are important for their hydrological function and for their significant contribution to the riverine ecosystems. Nevertheless their study has always been challenging because of the ephemeral and intermittent nature of those streams. Maps representing the active part of the river network are usually drawn after field surveys performed under different hydrologic conditions, which enable an objective evaluation of the temporal changes in the length of the active network. This method is useful to describe seasonal variations of the stream length, but has significant limitations when it comes to the description of event-based changes of the flowing network, provided that visual inspections of entire catchments are highly time-consuming. In this work, electrical resistance (ER) sensors were used to analyze event-based active network dynamics along some of the tributaries of an Alpine creek in northern Italy. Current intensity values were collected every 5 minutes by the sensors and a threshold electrical signal was identified to distinguish between wet and dry status of the reaches where the probes were placed. A statistical analysis revealed a good correlation among the mean current intensity recorded, the exceedance probability of the threshold and the persistency of the nodes. Data collected by the sensors were also interpolated in space along the network to obtain a sequence of maps of the active and dry parts of the stream network. From each map the wet length (L) of the watercourse was derived and linked to the corresponding discharge (Q) at the outlet of the catchment. Small and intense precipitation events had different effects on the variations of Q and L: the network length was found to be more sensitive than discharge to small precipitation inputs; relevant stream flow variations were instead observed only during significant events that originated the largest changes in the active network length.  This heterogeneous behaviour negatively affected the quality of the fitting of empirical discharges vs. wet length data through a power law model. Water presence sensors provide an opportunity to study in depth the spatiotemporal dynamics of the active length of intermittent streams and link such dynamics to the relevant hydrological drivers.

How to cite: Zanetti, F., Durighetto, N., Vingiani, F., and Botter, G.: Analysis of the active length dynamics on intermittent streams using water presence sensors, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8526, https://doi.org/10.5194/egusphere-egu21-8526, 2021.

To improve knowledge of hydrological and hydrogeological flow processes and their dependency on climate conditions it is becoming increasingly important to integrate sensors technology, independent observation methods, and new modeling techniques. Established isotope methods are usually regarded as a supplement and extension to classical hydrological investigation methods but are rarely included in soil water balance models. However, the combination could close knowledge gaps and thus lead to more precise and realistic predictions and therefore to better water management. Within the Wasserpfad project, a project of the Department of Civil Engineering at the TH Lübeck, soil moisture has been measured since May 2018. SMT100 soil moisture sensors from TRUEBNER GmbH are used at depths of 20, 40, 60, and 80 cm. Next to the station a 2m deep soil profile was taken in 2020, to estimate groundwater recharge using stable isotope equilibration methods and cryogenic extraction combined with soil water balance modeling. Vertical profiles of stable isotopes have been determined with a 10-cm resolution and measured with Tunable Diode Laser spectrometry. Percolation through the soil profile has been estimated based on the convolution of a seasonal input function using advection-dispersion transport models. Percolation rate estimate based on environmental isotope profiles results in 230 mm per year. Fitting of the advection-dispersion equation using a sinusoidal isotope input fitted to available time series provides an estimate of 255 mm per year. This difference is due to the dispersion effect on the isotope minima and maxima. The result of modeling the soil moisture data with a soil water balance model integrating the Richards equation for water transport and Penmen-Monteith based calculation of actual evaporation is used to verify the percolation rates. The analysis of soil moisture and isotope data by modeling provides a direct and efficient way to estimate the percolation rate. The combination of isotope methods with classical hydrological measuring techniques offers the possibility to verify results, to calibrate models, or to investigate the limits of isotope methods. Thus, flow processes can be predicted more reliably in the future.

How to cite: Krüger, N., Külls, C., and Kock, M.: Groundwater recharge estimates combining soil isotope profiles and classical soil water monitoring techniques, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9061, https://doi.org/10.5194/egusphere-egu21-9061, 2021.

Soil water content (SWC) dynamics can strongly influence catchment runoff generation processes. Knowledge about the amount and spatiotemporal distribution of SWC at the catchment scale can be useful for constraining and evaluating rainfall-runoff models.  While it is still challenging to obtain catchment scale-representative measurements of SWC, recent advances in cosmic ray neutron sensor (CRNS) technology have provided opportunities to obtain hectare scale data on SWC. Here we present a new method for obtaining spatially variable near-surface SWC by combining a high temporal resolution static CRNS sensor with ‘snapshot’ surveys using a portable CRNS. We also explored the role of these soil water storage data for catchment in rainfall-runoff generation models. We used ~4-years of near-surface SWC data from a static CRNS located in a humid mixed-agricultural catchment (~10km²) in Scotland. These data were complemented with at least three ‘snapshot’ portable CRNS surveys in each of the four main soil-land use (SLU) units in the catchment to produce SWC timeseries for each of these units. Two SLU units involved rotational crops under poorly or imperfectly draining mineral soils; one SLU unit typically supports livestock farming on freely draining mineral soils and the fourth, moorland on organic-rich soils.  While the moorland SLU unit on organic soils had the greatest difference in SWC dynamics under the static CRNS and other SLUs, we also found subtle SWC differences between mineral soil SLU units under different agricultural management. We then evaluated the additional information generated by the combined CRNS method in a rainfall-runoff model (HBV-light) calibration of dynamic catchment storage. For the purpose, we used areal weighted SLU SWC timeseries and compared the model calibration to that using the static CRNS alone. In this case, differences were marginal and model efficiencies similar, suggesting that static CRNS data from a landscape-representative location may be sufficient to inform rainfall-runoff model calibration at the catchment scale.  However, this may depend on model structure and the degree to which SWC dynamics vary within the landscape. This study demonstrated the potential of expanding the information value of permanently installed CRNS sensors using portable CRNS surveys in the context of humid mixed-agricultural environment, although testing in different environments would be required to evaluate wider applicability.

How to cite: Dimitrova Petrova, K., Rosolem, R., Soulsby, C., Wilkinson, M., Lilly, A., and Geris, J.: Combining static and portable Cosmic Ray Neutron Sensor (CRNS) data to assess catchment scale heterogeneity in soil water content and implications for runoff generation, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4751, https://doi.org/10.5194/egusphere-egu21-4751, 2021.

Monitoring and profiling the isotopic composition of soil water in combination with groundwater isotope hydrology are commonly used in studying flow and transport in soils as well as in estimating groundwater recharge. Establishing the isotopic composition of local precipitation is of essence. Towards this end and in facilitating the application of isotope hydrology in Troodos Fractured Aquifer (TFA), precipitation was monitored in 16 precipitation sampling stations, stretching from the shoreline up to 1725 m above m.s.l., from January of 2015 to December of 2017. A seasonal trend was discerned, with isotopically depleted rainfall occurring in December as opposed to the more enriched autumn and spring rainfall. Northern European air masses appear to prevail during the months of December to January during which d values tend to be on average above 25‰ whereas the more enriched rain with the lowest d values occurs in July. The averaged seasonal effect between 2015 and 2017 on δ18O, δ2H and d values are 4.53‰, 30.98‰ and 14.93‰, respectively. Cyprus’ Local Meteoric Water Line (LMWL) was found to be equal to δ2H = (6.58±0.13)*δ18O + (12.64±0.91) and a general decrease of 1.22‰ for δ2H and 0.20‰ for δ18O in precipitation was calculated per 100 m altitude.  Similar values have been found by other researchers for the region. These variations in the isotope composition of rainfall can be used to earmark seasonal input of recharge water and for deriving percolation rates from tracing their movement in the soil column.

How to cite: Christofi, C., Bruggeman, A., and Kuells, C.: The Isotopic Composition of Cyprus Precipitation. A Tool of Isotope Hydrology., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9210, https://doi.org/10.5194/egusphere-egu21-9210, 2021.

Groundwater dynamics play a crucial role in the spreading of a soil and groundwater contamination. However, there is still a big gap in the understanding of the groundwater flow dynamics. Heterogeneities and dynamics are often underestimated and therefore not taken into account. They are of crucial input for successful management and remediation measures. The bulk of the mass of mass often is transported through only a small layer or section within the aquifer and is in cases of seepage into surface water very dependent to rainfall and occurring tidal effects.

This study contains the use of novel real-time iFLUX sensors to map the groundwater flow dynamics over time. The sensors provide real-time data on groundwater flow rate and flow direction. The sensor probes consist of multiple bidirectional flow sensors that are superimposed. The probes can be installed directly in the subsoil, riverbed or monitoring well. The measurement setup is unique as it can perform measurements every second, ideal to map rapid changing flow conditions. The measurement range is between 0,5 and 500 cm per day.

We will present the measurement principles and technical aspects of the sensor, together with two case studies.

The first case study comprises the installation of iFLUX sensors in 4 different monitoring wells in a chlorinated solvent plume to map on the one hand the flow patterns in the plume, and on the other hand the flow dynamics that are influenced by the nearby popular trees. The foreseen remediation concept here is phytoremediation. The sensors were installed for a period of in total 4 weeks. Measurement frequency was 5 minutes. The flow profiles and time series will be presented together with the determined mass fluxes.

A second case study was performed on behalf of the remediation of a canal riverbed. Due to industrial production of tar and carbon black in the past, the soil and groundwater next to the small canal ‘De Lieve’ in Ghent, Belgium, got contaminated with aliphatic and (poly)aromatic hydrocarbons. The groundwater contaminants migrate to the canal, impact the surface water quality and cause an ecological risk. The seepage flow and mass fluxes of contaminants into the surface water were measured with the novel iFLUX streambed sensors, installed directly in the river sediment. A site conceptual model was drawn and dimensioned based on the sensor data. The remediation concept to tackle the inflowing pollution: a hydraulic conductive reactive mat on the riverbed that makes use of the natural draining function of the waterbody, the adsorption capacity of a natural or secondary adsorbent and a future habitat for micro-organisms that biodegrade contaminants. The reactive mats were successfully installed and based on the mass flux calculations a lifespan of at least 10 years is expected for the adsorption material.  

How to cite: Verreydt, G., Van Putte, N., De Kleyn, T., Cool, J., and Maiheu, B.: Innovative real-time sensing of flow dynamics in groundwater and sediments to map contaminant spreading, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-13338, https://doi.org/10.5194/egusphere-egu21-13338, 2021.

Drainage below the root zone of irrigated crops and trees is often an unknown component of the water balance. This drainage water could recharge underlying aquifers and flow to streams and is not part of water consumed by crops, as used in water productivity computations. Drainage from fields with irrigation systems that wet only part of the soil is difficult to estimate. The objective of the research was to develop a water balance model with a dynamic wetted area for analyzing soil water balance components from daily soil moisture observations. The method was applied in an olive orchard in Cyprus, with approximately 35% canopy cover. Soil moisture sensors (SMT100, Truebner and 5TM, Decagon) were installed at six trees, at 10-, 20-, 40- and 60-cm depth, approximately 90 cm from the trunk of the tree. Soil moisture was recorded hourly. The trees were irrigated weekly, with a single spaghetti tube with a discharge rate of approximately 135 L/hr. Daily reference evapotranspiration was computed with the Penman-Monteith equation from meteorological observations recorded inside the orchard (WS500, Lufft). Rainfall was measured with a tipping bucket rain gauge (15189, Lambrecht).

The model computes a daily volumetric water balance for the canopy area of the tree. During the irrigation season, soil moisture observations were assumed to represent the soil volume wetted by irrigation. Drainage below the 70-cm root zone occurred when soil moisture exceeded the field capacity, as derived from hourly observations. A canopy-area crop coefficient (Kcc-max) was estimated for all irrigation days without drainage by minimizing the sum of the daily evapotranspiration in excess of the maximum evapotranspiration (Kcc-max ETo). This one-sided error was controlled by maintaining a positive difference between Kcc-max and Kcc the day after irrigation. Wetted areas were subsequently computed for all irrigation days without drainage. For irrigation days with soil moisture above field capacity, the wetted area was adjusted manually, such that drainage was smaller on the second day than on the irrigation day, using a Kcc-max for both days. During the May to November 2019 irrigation season, drainage was 8 mm over the field area, for a field capacity of 36%, a Kcc-max of 1.3, and an error of 16 mm. Assuming a field capacity of 38%, drainage was 3 mm over the field area, with a Kcc-max of 1.4, and an error of 17 mm. Overall, the model provided a quick and robust way of estimating the irrigation water balance components.

This research has received financial support from the ERANETMED3 program, as part of the ISOMED project (Environmental Isotope Techniques for Water Flow Accounting), funded through the Cyprus Research and Innovation Foundation.

How to cite: Bruggeman, A., Siakou, M., Eliades, M., Djuma, H., and Zoumides, C.: A daily water balance model with a dynamic wetted area for estimating drainage from soil moisture observations in an irrigated orchard, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-14566, https://doi.org/10.5194/egusphere-egu21-14566, 2021.