Maciek W. Lubczynski

Measuring soil surface water content in irrigated areas of southern Tunisia using full-waveform inversion of proximal GPR data

Full-waveform inverse modelling of proximal ground-penetrating radar was used to measure soil surface water content in irrigated areas of southern Tunisia. The ground-penetrating radar system consisted of a handheld vector network analyser combined with an off-ground monostatic horn antenna, thereby setting up an ultra wideband stepped-frequency continuous-wave radar. Inversion of the radar Green’s function was performed in the time-domain, on a time window focused on the surface reflection only. Results were compared with volumetric and time-domain reflectometry measurements, as well as with an improved version of the standard reflection coefficient method. Except for water contents close to saturation, good agreements were obtained between ground-penetrating radar, time-domain reflectometry and volumetric samples. Significant differences were observed when the soil electric conductivity was high and could not be neglected anymore in the inversion process. Accounting for electric conductivity provided better results. Remaining errors were attributed to the different scales of characterization dealt with in relation to the vertical variability of water content in the top few centimetres of the soil. The proposed method appears to be more practical and accurate than the standard reflection coefficient method and shows great promise for real-time mapping of surface soil moisture at the field scale.

Exploiting the MRS-phase information to enhance detection of masked deep aquifers: examples from the Netherlands

Several magnetic resonance soundings (MRS) in the Netherlands showed a monotonous single peak anomaly on the amplitude versus excitation moment sounding pattern, which were interpreted as a single aquifer when using an amplitude-only mode MRS data inversion. However, in all these soundings, borehole logs documented the presence of two or three aquifers separated by clay-rich aquitards in the first 100 m below ground surface. Such environments were electrically conductive so a phase excursion was noticeable on the MRS soundings. Multi-aquifer systems, in a conductive environment, may show interference among signals originating from different parts of the systems including amplitude masking or destructive interference. A new version of an off-the-shelf MRS forward modelling and inversion tool (Samovar 11.3) allowing complex amplitude and phase inversion was used to detect and parameterize deep, MRS-masked second aquifers at two selected sites in the Netherlands, one near Delft and one near Waalwijk. At the Delft site, the proposed strategy was effective in the detection and characterization of a second previously missed aquifer at 45 m below ground surface, while at the Waalwijk site, the second aquifer was not detected because of a considerably deeper aquifer at 85 m and too small excitation (6000 A ms). However, forward modelling showed that with a larger excitation moment (e.g., 13 000 A ms), detection and parameterization of the second aquifer would become possible when using both amplitude and phase.

Assessing Groundwater Storage Changes Using Remote Sensing–Based Evapotranspiration and Precipitation at a Large Semiarid Basin Scale

A method is presented that uses remote sensing (RS)-based evapotranspiration (ET) and precipitation estimates with improved accuracies under semiarid conditions to quantify a spatially distributed water balance, for analyzing groundwater storage changes due to supplementary water uses. The method is tested for the semiarid Konya basin (Turkey), one of the largest endorheic basins in the world. Based on the spatially distributed water balance estimation, the mean irrigation for croplands was 308 mm yr−1, which corresponds to a total reduction of 2270 million cubic meters per year (106 m3 yr−1, or MCM yr−1) in the groundwater storage during the study period 2005–09. The storage change estimated as the residual of the spatially distributed water balance was confirmed by the volume change calculated from groundwater table observations. To obtain an improved precipitation distribution, the monthly Tropical Rainfall Measuring Mission (TRMM) rainfall product was assessed. After a bias removal, TRMM data were combined with the snow water equivalent estimated by a multivariate analysis using snow gauge observations, the Moderate Resolution Imaging Spectroradiometer (MODIS) snow cover product, and the digital elevation model. With respect to the distribution of ET, the standard SEBS and the soil moisture integrated SEBS-SM models were compared; SEBS-SM proved to better reflect the water-limited evapotranspiration regime of semiarid regions. The RS-based distributed water balance calculation as presented in this study can be applied in other large basins, especially in semiarid and arid regions. It is capable of estimating spatially distributed water balances and storage changes, which otherwise, by ground-based point measurements, would not be feasible

Integrating MRS data with hydrologic model - Carrizal Catchment (Spain)

Magnetic resonance sounding (MRS) provides quantitative hydrogeological information on hydrostratigraphy and hydraulic parameters of subsurface (e.g. flow and storage property of aquifers) that can be integrated in distributed hydrologic models. The hydraulic parameters are typically obtained by pumping tests. In this study, we propose an MRS integration method based on optimizing MRS estimates of aquifer hydraulic parameters through hydrologic model calibration. The proposed MRS integration method was applied in the 73 km2 Carrizal Catchment in Spain, characterized by a shallow unconfined aquifer with an unknown aquifer bottom. 12 MRS survey results were inverted with Samovar 11.3, schematized and integrated in the transient, distributed, coupled, hydrologic, MARMITES-MODFLOW model. As the aquifer bottom was unknown, the aquifer was schematized into one unconfined layer of uniform thickness. For that layer, MRS estimators of specific yield and transmissivity/hydraulic conductivity were calculated as weighted averages of the inverted MRS layers. The MRS integration with hydrologic model was carried out by introducing multipliers of specific yield and transmissivity/hydraulic conductivity that were optimized during transient model calibration using 11 time-series piezometric observation points. The optimized multipliers were 1.0 for specific yield and 3.5*10-9 for hydraulic conductivity. These multipliers were used, and can be used in future MRS investigations in the Carrizal Catchment (and/or adjacent area with similar hydrogeological conditions), to convert MRS survey results into aquifer hydraulic parameters. The proposed method of MRS data integration in the hydrologic model of Carrizal Catchment not only allowed us to calibrate the model but also to confirm the functional capability of MRS in quantitative groundwater assessment. Most importantly however, it demonstrated that if pumping tests are not available, the use of MRS integrated in distributed coupled hydrological models, or even in standalone groundwater models, provides a valuable aquifer parameterization alternative.

Groundwater evapotranspiration : underestimated role of tree transpiration and bare soil evaporation in groundwater balances of dry lands

This paper analyzes and emphasizes the importance of groundwater evapotranspiration (ETg) in groundwater balances. The ETg diminishes the net groundwater recharge that constrains groundwater flow and replenishment of groundwater resources. The ETg consists of two different components, groundwater transpiration (Tg) and groundwater evaporation (Eg), both not yet well identified in hydrogeology. The ETg values are the largest in dry locations with shallow groundwater table. The significance of the ETg however is the largest when its relative contribution to groundwater balance is high i.e. when its rate is comparable with groundwater recharge.

A framework for sourcing of evaporation between saturated and unsaturated zone in bare soil condition

ABSTRACT Sourcing subsurface evaporation (Ess) into groundwater (Eg) and unsaturated zone (Eu) components has received little scientific attention so far, despite its importance in water management and agriculture. We propose a novel sourcing framework, with its implementation in dedicated post-processing software called SOURCE (used along with the HYDRUS1D model), to study evaporation sourcing dynamics, define quantitatively “shallow” and “deep” water table conditions and test the applicability of water table fluctuation (WTF) and “bucket” methods for estimation of Eg and Eu separately. For the “shallow” and “deep” water table we propose Eg > 0.95Ess and Eg = 0 criteria, respectively. Assessment of the WTF method allowed sourcing of very small fluxes otherwise neglected by standard hydrological methods. Sourcing with SOURCE software was more accurate than the standard “bucket” method mainly because of greater flexibility in spatio-temporal discretization. This study emphasized the dry condition relevance of groundwater evaporation which should be analysed by applying coupled flow of heat, vapour and liquid water. Editor D. Koutsoyiannis; Associate editor S. Kanae

Transpiration and stomatal conductance in a young secondary tropical montane forest: contrasts between native trees and invasive understorey shrubs

It has been suggested that vigorous secondary tropical forests can have very high transpiration rates, but sap flow and stomatal conductance dynamics of trees and shrubs in these forests are understudied. In an effort to address this knowledge gap, sap flow (thermal dissipation method, 12 trees) and stomatal conductance (porometry, six trees) were measured for young (5-7 years) Psiadia altissima (DC.) Drake trees, a widely occurring species dominating young regrowth following abandonment of swidden agriculture in upland eastern Madagascar. In addition, stomatal conductance (gs) was determined for three individuals of two locally common invasive shrubs (Lantana camara L. and Rubus moluccanus L.) during three periods with contrasting soil moisture conditions. Values of gs for the three investigated species were significantly higher and more sensitive to climatic conditions during the wet period compared with the dry period. Further, gs of the understorey shrubs was much more sensitive to soil moisture content than that of the trees. Tree transpiration rates (Ec) were relatively stable during the dry season and were only affected somewhat by soil water content at the end of the dry season, suggesting the trees had continued access to soil water despite drying out of the topsoil. The Ec exhibited a plateau-shaped relation with vapour pressure deficit (VPD), which was attributed to stomatal closure at high VPD. Vapour pressure deficit was the major driver of variation in Ec, during both the wet and the dry season. Overall water use of the trees was modest, possibly reflecting low site fertility after three swidden cultivation cycles. The observed contrast in gs response to soil water and climatic conditions for the trees and shrubs underscores the need to take root distributions into account when modelling transpiration from regenerating tropical forests.