Advertisement

Plant and Soil

, Volume 424, Issue 1–2, pp 73–89 | Cite as

Quantifying root water extraction after drought recovery using sub-mm in situ empirical data

  • Indu Dhiman
  • Hassina Bilheux
  • Keito DeCarlo
  • Scott L. Painter
  • Lou Santodonato
  • Jeffrey M. Warren
Regular Article

Abstract

Aims

Root-specific responses to stress are not well-known, and have been largely based on indirect measurements of bulk soil water extraction, which limits mechanistic modeling of root function.

Methods

Here, we used neutron radiography to examine in situ root-soil water dynamics of a previously droughted black cottonwood (Populus trichocarpa) seedling, contrasting water uptake by the two major components of the root system that differed in initial recovery rate as apparent by ‘new’ (whiter, thinner), or ‘old’ (darker, thicker) parts of the fine root system.

Results

The smaller diameter ‘new’ roots had greater water uptake per unit surface area than the larger diameter ‘old’ roots, but they had less total surface area leading to less total water extraction; rates ranged from 0.0027–0.0116 g cm−2 h−1. The finest most-active roots were not visible in the radiographs, indicating the need to include destructive sampling. Analysis based on root-free bulk soil hydraulic properties indicated substantial redistribution of water via saturated/unsaturated flow and capillary wicking across the layers - suggesting water uptake dynamics following an infiltration event may be more complex than approximated by common soil hydraulic or root surface area modeling approaches.

Conclusions

Our results highlight the need for continued exploration of root-trait specific water uptake rates in situ, and impacts of roots on soil hydraulic properties – both critical components for mechanistic modeling of root function.

Keywords

Neutron radiography Modeling Populus Hydraulic redistribution Rhizosphere Water uptake 

Introduction

The ability of fine root systems to adjust to and recover from severe drying conditions is essential to maintaining plant water uptake. However, our knowledge of temporal root function is largely limited to indirect measurements of bulk soil water extraction, thereby limiting mechanistic modeling of root function. Here, we used high-resolution in situ neutron radiography (NR) measurements of soil water dynamics to test the validity of key assumptions underlying common modeling approaches to root water extraction, focusing on extreme drought conditions.

Data used to develop models of root water extraction have been largely comprised of indirect measurements, such as total plant water use (based on atmospheric demand, pot weight, or calculated with sap flow sensors, gas exchange or eddy covariance techniques), soil water sensors (volumetric content or water potential), or based on a site water balance using meteorological data and soil physical properties such as soil particle distribution, bulk density or organic matter content. Direct measurements of soil, rhizosphere and root water dynamics are difficult and have been limited, yet advances in imaging techniques such as neutron radiography are providing quantification of soil-root water dynamics at unprecedented spatial and temporal resolutions (Oswald et al. 2008; Carminati et al. 2010; Moradi et al. 2011; Zarebanadkouki et al. 2012; Warren et al. 2013a; Ahmed et al. 2016). Neutron radiography is a non-destructive technique, complimentary to several other experimental methods, such as X-ray, nuclear magnetic resonance imaging, or gamma imaging. The key advantage of NR is its sensitivity to lighter elements such as hydrogen and its stable isotope deuterium, which are particularly useful for investigation of the root water (H2O) uptake and transport through plants. Roots in unsaturated soil can be distinguished from soil due to differential attenuation of a neutron beam based on elemental composition, especially H. This allows microscopic visual quantification of actual root function (i.e., high resolution growth, distribution and water uptake dynamics) and can help link root traits to uptake. Imaging can be used to understand root function in situ, as well as for the development and validation of theoretical models of root water extraction (Roose et al. 2016) that could be scaled to the landscape level (Warren et al. 2015). Recent insights from NR studies have enhanced our understanding of water dynamics in roots and soil, and the critical rhizosphere interface between them.

Initially, different experimental studies showed that soil closer to roots had lower water content, in comparison to bulk soil (Hainsworth and Aylmore 1989; Macfall et al. 1990; Segal et al. 2008). In contrast, neutron based studies have indicated higher water content in soil in the immediate vicinity of roots (Nakanishi et al. 2005; Tumlinson et al. 2008; Esser et al. 2010). Subsequent work by Oswald et al. (2008), Carminati et al. (2010) and Ahmed et al. (2016) showed that water content in the rhizosphere was highly dynamic, with higher water content during drying and lower water content after rewetting. Their results suggest that the rhizosphere has much different properties from the bulk soil, in part due to the impact of root excreted mucilage (Carminati et al. 2010). Other NR results illustrate that soil water dynamics are not spatially uniform, but driven by evaporation, root density, soil type, soil properties and light (Moradi et al. 2008; Carminati et al. 2011; Zarebanadkouki et al. 2012; Warren et al. 2013a). NR has also illustrated the substantial variation in soil, rhizosphere and root properties and their impacts on water movement, and paired water potential driving forces. Understanding the distinct changes in the hydraulic continuum from soil to root is necessary for accurate representation of water flow and uptake in models. Yet resolving these processes and translating them into mechanistic models of water or nutrient uptake has been a challenge.

Root uptake of soil resources, including water, mobile and immobile ions, and organic molecules, depends both on the distance from the root to the resource, and the root functional traits. Younger, finer roots are much more permeable to water and nutrients, so most of the focus has been on understanding fine root structure (Guo et al. 2008), despite few studies that link that structure to function in situ. Fine root water uptake models vary in their structural approach – ranging from first-principle microscopic models, to those that are parameterized at the macroscopic level (Feddes et al. 2001; Dardanelli et al. 2004; Warren et al. 2015). The models usually bulk fine roots into a single pool despite differences in root traits such as suberization that limit water uptake, indicating a need to move towards a separation of active and inactive fine roots (McCormack et al. 2015). Relatively recent models explicitly model root water extraction through the profile linked to dynamic soil and root hydraulic properties (e.g., Lai and Katul 2000; Doussan et al. 2006). The challenge is that most of the root models proposed in literature assume similar properties in soil and rhizosphere (soil region in contact with roots). However, significant differences between chemical and physical properties of bulk soil and rhizosphere have been observed, such as the relative rehydration/dehydration kinetics described above, and the development of a physical gap between the root and the soil as both roots and soil dry (Carminati et al. 2013). In addition, soil hydraulic properties are often derived without active roots, thus modeled soil characteristics such as porosity or water retention may not represent the actual dynamics in situ. For example, fine roots and the even-finer hyphae of root-associated mycorrhizal fungi can fill soil pores, and block or enhance flow paths of water (Allen 2007; Querejeta 2017), yet these types of impacts on soil properties are rarely studied primarily due to a lack of experimental data for validation.

Modeling the root system provides an opportunity to improve our understanding about root water uptake behavior between different traits, and with different environmental conditions, independently. Several theoretical models based on Gardner’s work (Gardner 1960) have been proposed to estimate the water redistribution in the soil and roots (e.g., Roose and Fowler 2004; Doussan et al. 2006; Javaux et al. 2008). Combining root imaging in situ with paired modeling efforts should provide new insights into understanding pore-level rhizosphere dynamics and their impacts on transfer of water and ions between the bulk soil and root (Roose et al. 2016). Indeed, work done by Carminati et al. (2010); Zarebanadkouki et al. (2014) and Ahmed et al. (2016) illustrate the utility of paired neutron imaging experiments and modeling to assess soil physical properties during drying, soil water extraction and explicit radial and axial flow through roots, and new modeling approaches based on new insights are emerging, such as allowing air gaps to develop in the R-SWMS model, thereby regulating water uptake (Javaux et al. 2013). Another recent study used X-ray CT to quantify root growth over a 12-day period, and successfully paired root growth and geometry with Richard’s equation based modeling of soil water distribution and uptake (Daly et al. 2017). Such approaches that link estimation of soil and root hydraulic properties can be further extended to other types of root architecture with different soil conditions. While NR usually measures water uptake by roots based on near root reductions in soil water content in soil, redistribution of water by roots or soils can confound model application (Warren et al. 2013a, 2013b; Zarebanadkouki et al. 2014). By combining NR experimental data with model development, synergistic progress can be made in our understanding of plant-soil water dynamics, and the key interactions between co-occurring processes such as differential root water uptake, rhizosphere hydration dynamics, capillarity within the soil, and soil and root hydraulic redistribution.

In the present study, our aim was to quantify soil-root water dynamics for a droughted plant system at high spatial resolution. We hypothesized that newer roots grown after extreme drought stress would have greater uptake capacity than older roots that survived the drought stress treatment since they would be younger and presumably have less radial resistance to water flow into the root. In addition, we quantified the vertical profile in root water uptake and compared that profile to the vertical profile of specific root surface area, which is often assumed to be a proxy for the vertical variation in root water uptake in models. This addresses the current and expanding interest in the concept that root spatial distribution does not always correlate with root functional distribution, which is dynamic and affected by both external conditions such as soil water availability and internal mechanisms such as aquaporins (Schymanski et al. 2008; Johnson et al. 2014; Warren et al. 2015).

Materials and methods

Plant material and treatments

Black cottonwood (Populus trichocarpa) seeds were propagated in pure silica quartz sand (Flint #13, U.S. Silica Company, Berkeley Springs, WV) in an aluminum plate chamber (20 × 18 × 1.2 cm (OD), 19 × 17.6 × 1 cm (ID). The chamber was irrigated to field capacity every 3 days, and a balanced fertilizer solution with micronutrients was added periodically. After 1 month, water was withheld to subject the seedlings to extreme drought, as indicated by most leaves reaching the permanent wilting point. After 1 week, plants were rewatered to field capacity and maintained for two weeks. At this point the plants were excavated from the chamber, revealing both ‘older’ (presumably pre-drought) and ‘newer’ (primary post-drought) root systems based on growth form and color (Fig. 1a). The ‘new’ roots on the left side of the figure clearly indicate greater vigor; the lower portion of the main root and the attached secondary roots were visibly white with dense fine root and abundant mycorrhizal hyphae. In contrast, the ‘old’ right side of the root system was darker and less dense and may have been more damaged during the drought period. Even so, there were some visible white root tips and some hyphal masses indicating it was still viable, albeit recovering slower than the left side of the root system. This presented a unique experimental opportunity to compare water uptake capacity of these two root systems that began at different stages of drought recovery. The single plant system was replanted into another aluminum plate chamber with the ‘new’ and ‘old’ root systems displayed to each side. The roots were carefully placed between layers of sand and the plant system was maintained under hydrated conditions for 1 month in a growth chamber prior to neutron radiography. To prevent evaporation, the sides top and bottom of the chamber were sealed with tape – the stem was carefully sealed around so that the only water loss was via plant transpiration – thus mass-balance could be used to estimate total water uptake, transport and transpiration. The 11-week-old plant system was then placed into the beamline and initial baseline neutron radiographs were collected (see below). At this point, a syringe was pushed through the taped top, and 10 ml of water was slowly dripped onto the soil surface, 5 ml on each side of the plant. The syringe holes were sealed, a growth lamp suspended above the plant was turned on, and sequential neutron radiography was initiated to assess spatial dynamics of water movement and extraction through time. The plant chamber was periodically removed from the beamline and weighed to validate the estimates of total water content based on NR. Beamline conditions during NR – photosynthetically active radiation was ~500 μmol m−2 s−1, air temperature was 21–23 °C, relative humidity was 23–25%. Following NR, the plant-system was carefully harvested to assess actual root length and surface area by depth for both the older and younger root systems (Fig. 1b). Roots were scanned and their dimensions measured using WinRhizo software (Regent Instruments Inc., Quebec, Canada). Results from the scanned images were separated by depth and aligned with NR results based on identifiable root morphology in the driest radiograph (e.g., a major root branch or a bend in root could be seen in both radiograph and scanned image, Fig. 1c and d).
Fig. 1

a distinct newer (white, ‘finer’, left-side) and older (brown, ‘coarser’, right-side) root system following recovery from extreme drought and prior to replanting, b final harvested root system after radiography, (c) analysis of root structural display by depth for left (L) or (d) right (R) side of root system – layers 1–9 correspond to the layers 1–9 used in the radiography analysis. (e) initial radiograph prior to drying cycle, (f) final radiograph after 2-day drying cycle

NR conditions and experiment

Neutron radiography experiments were performed at the CG1-D beam line using cold neutrons, located at High Flux Isotope Reactor (HFIR) facility at Oak Ridge National Laboratory (Santodonato et al. 2015). Neutron attenuation by the plant samples was detected with a 25 μm LiF / ZnS scintillator linked to a charge coupled detector (CCD) camera system (iKon – L 936, Andor Technology plc. Belfast, UK). The cold neutron wavelength, λ, ranged from 0.8 to 6 Å, with a peak neutron intensity of 2.2 × 106 at 2.6 Å. In comparison to thermal neutrons, cold neutrons have the advantage with higher attenuation coefficients for most of the penetrated materials. This can lead to substantial improvement in contrast between two samples. The field of view (FOV) for this setup was ~ 7.4 × 7.4 cm, with effective pixel area of 36 μm × 36 μm, with the radiograph resolution of ~ 100 μm. Aluminum, and to a lesser extent sand (SiO2) have low attenuation coefficients for cold neutrons, therefore the chamber and dry soil transmitted most the neutron beam flux. Attenuation occurred due to the presence of soil water and water filled roots. The plate chamber was larger than our CCD detector FOV, thus to measure the whole plant system in front of the detector, multiple radiographs were recorded while moving the chamber in different directions, left, right, up and down, using the motorized stage. Thus, every 16 (4 × 4) radiographs corresponded to one complete scan of a plant chamber. Exposure time for each radiograph was 120 s. To assess soil water distribution, and root distribution, growth and water uptake, 2D radiographs of root – soil system were imaged through time.

Image reconstruction and segmentation

Neutron radiographs were normalized with respect to flat field / open beam and dark field using iMARS software (Bilheux and Bilheux 2015). Open beam radiographs are measurements carried out without any sample in the beam, considering the beam inhomogeneities, while the dark field measurements are the correction towards the electronic noise from the detector set-up. The 16 partial chamber radiographs were stitched together into a collage using ImageJ plugins ‘grid’ and ‘collection stitching’, using a linear blending method with 15% of intensity overlap (Preibisch et al. 2009). Thereafter, all the radiographs were aligned using template matching and slice alignment, based on subpixel registration (Thévenaz et al. 1998). Figure 1e and f depict fully stitched transmission radiographs, measured just after watering, then after two days of drying, respectively.

NR root segmentation from the soil was carried out using mean-based local thresholding and morphological cleaning methods, which allowed for root tapering. To segment roots from soil a mask (binarized radiograph) was created using the driest radiograph measured at the end of two-day experimental cycle. Due to low water content in this radiograph, there was good contrast between the roots and soil, making it easier to segment the roots. However, the bottom part of the chamber was highly attenuating, containing a large density of fine roots and more soil water leading to greater error in segmentation. The mask was applied to all the stitched radiographs measured over a period of two days. Further calculations and data analysis was carried out using code written in MATLAB.

Calibration to actual water content

Calibration measurements were carried out to determine the attenuation coefficient of water. Towards this, similar steps as described by Kang et al. (2013) were followed. The data measured in the present study was verified based on prior measurements, and used values for the attenuation coefficient Σ = 5.542 cm−1, and empirical correction factor for beam hardening and scattering effects β = − 2.140 cm−2 (Kang et al. 2013).

The water thickness of each pixel for the plant – soil system, for all the radiographs was calculated using the following equation:
$$ {\tau}_{\left(i,j\right)}=-\frac{\varSigma }{2\beta }-\sqrt{{\left(\frac{\varSigma }{2\beta}\right)}^2-\frac{1}{\beta}\mathit{\ln}{\left(\frac{I_{wet}}{I_{dry}}\right)}_{\left(i,j\right)}} $$
(1)
For the calculation of pixel-wise absolute water thickness, it is important to remove the attenuation contributions from soil and chamber (I dry ). To estimate the contribution from soil and the chamber together, a region of interest (ROI) in upper most section of the driest transmission radiograph with no roots and only residual water content was selected and applied as a correction to all pixel values. Eq. 1 was implemented in MATLAB code to calculate the absolute water thicknesses of all the radiographs, segmented soil and roots. Further, to evaluate the volumetric water content (V w , in cm3) of each pixel, water thickness τ (i, j) was multiplied by pixel area (36 μm × 36 μm) and beam path length (l) through the chamber (inner dimension of 1 cm). Therefore,
$$ {V}_W={\tau}_{\left(i,j\right)}\times pixel\kern.4em area\times l $$
(2)

To test root-specific water extraction patterns, the NR chamber data was segmented into 18 ROI – nine soil layers for the left side (newer roots) and the right side (older roots), and total water content was assessed within each ROI. For each ROI, the water content was calculated as the sum of water content in each pixel for that region. Each of these 18 ROI were associated with known root dimensions as described above.

Quantification of root water uptake

The changes in water content in each of the nine horizontal soil layers were caused not only by transpiration, but also by vertical redistribution of soil water (upward or downward) as the system responded to the initial pulse of water. To estimate the actual vertical profile of root water uptake, those two processes need to be separated using a modeling framework. In the closed experimental chamber, the rate of change of total water volume below a specified location (z) is a balance between the volumetric flux of soil water and the volumetric flux of water in roots.
$$ \frac{d}{dt}\left[ water\kern.3em volume\kern.3em below\kern.3em z\right]=\left[ flux\kern.3em of\kern.3em water\kern.3em in\kern.3em soil\kern.3em at\kern.3em z\right]-\left[ flux\kern.3em of\kern.3em water\kern.3em in\kern.3em roots\kern.3em at\kern.3em z\right] $$
(3)
where we define soil water flux as positive downward and root water flux as positive upward. The latter quantity is the root water uptake, cumulative from the bottom of the chamber, which can be estimated from Eq. 3 with the other two terms constrained by the measurements. The term on the left side is available directly from the measured soil water content by differencing in time. The first term on the right side (the soil water component of the total water flux) can be estimated from the measured water content because we have good constraints on the soil properties (Kang et al. 2014). Specifically,
$$ {\varGamma}_{soil}={Ak}_s{k}_{rel}\left(\theta \right)\left(\frac{\partial \psi }{\partial z}+1\right) $$
(4)
where A (cm2) is the horizontal cross-sectional area of the chamber, k s is saturated hydraulic conductivity (cm s−1), k rel is relative permeability, and ψ is suction head (cm). Here z is the depth coordinate, defined as positive downward. The water content in Eq. 4 is that of the soil, but the NR gives a composite of roots and soil. For the purposes of this analysis, we used the measured root volume fraction assuming roots are 90% water to calculate the water content in the soil.
Kang et al. (2014) examined soil water release curves for the sand used in this experiment and estimated parameters for the Brooks-Corey relationship (1964)
$$ \theta ={\theta}_r+\left({\theta}_s-{\theta}_r\right){\left(\frac{\psi_a}{\psi}\right)}^{\lambda } $$
(5)
or
$$ \psi ={\psi}_a{\left(\frac{\theta -{\theta}_r}{\theta_s-{\theta}_r}\right)}^{-1/\lambda } $$
(6)
and
$$ {k}_{rel}\left(\theta \right)={\left(\frac{\theta -{\theta}_r}{\theta_s-{\theta}_r}\right)}^{3+2/\lambda } $$
(7)

Kang et al. (2014) find that λ= 6.649, ψ a = 17.25 cm, θ r  = 0.025,and θ s  = 0.364 provide a good match with the soil water retention curve. We use θ s  = 0.41, based on the weight of the sand-filled chamber.

Saturated hydraulic conductivity for a clean flint sand is 1.66 ± 0.32 × 10−2 cm/s (Kang et al. 2014). However, bulk hydraulic conductivity for root-filled soils are likely lower than that of a root-free soil because roots preferentially occupy larger pores and shift the pore-size distribution (Zarebanadkouki et al. 2016; Moradi et al. 2011; Zarebanadkouki et al. 2013). Leung et al. (2015a) find the reduction in saturated hydraulic conductivity to be approximately a factor of 2. Data on the reduction in hydraulic conductivity are limited, and are anticipated to be soil- and species-dependent. We use k s  = 1.66 × 10−3 cm/s and judge this to be a lower bound on the rate of soil drainage. Root- or mycorrhizal-induced shifts in the pore-size distribution may also shift the soil water release curve (Leung et al. 2015a). Specifically, roots filling the pore space reduce the available porosity, resulting in larger suction for root-filled soils for the same water content (Leung et al. 2015a; Leung et al. 2015b; Ng et al. 2016; Querejeta 2017). In our case, the root volume fraction increased with increasing depth, implying an increase in suction with increasing depth, which would have the net effect of increasing the downward movement of water relative to the no-root case. We ignore this potential effect in our reference case in the interest of obtaining a lower bound on water drainage through the soil and of root water uptake, but also examine sensitivity to the parameters in soil water release curve.

A forward model based on Richards equation with specified root water uptake provides a check on the analyses method just described. Richards equation in mixed form in 1-D is
$$ \frac{\partial \theta }{\partial t}=-\frac{\partial }{\partial z}\left[{k}_{rel}\left(\theta \right)\left(\frac{\partial \psi }{\partial z}+1\right)\right]-{q}_{root}\left(t,z\right) $$
(8)
where q root (cm3/cm3-s) is the root water uptake defined as positive for flow from soil to roots, z is depth from the surface. The total transpiration is known from the change in the weight of the experimental chamber over time. We distribute the transpiration demand vertically according to root specific surface area, solve Eq. 8 numerically, then compare the resulting water content with that observed. The modeled and observed water content should be in agreement if the root water uptake is proportional to root specific surface area and the hydraulic properties are correct.

Results

NR and patterns of water extraction

Following recovery from extreme drought and before neutron radiography measurements the root system was comprised of a younger, whitish and hyphal-covered root system on the left side of the chamber, largely developed within one-week post-drought, and an older, brownish, presumably more suberized (or with less absorptive capacity) root system on the right side, primarily developed pre-drought (Fig. 1a). After replanting and another month of growth, the two root systems were less visibly distinguishable, both having grown substantially in size and with most roots displayed in the lower ½ of the chamber (Fig. 1b). At harvest, total fine root length was within 5% for the younger and older root systems, but mean root diameter was 16% greater for the older root system (excluding stem section), root surface area was 13% greater and total root biomass was 25% greater for the older root system, as compared with the younger root system (Table 1, Fig. 1c and d). The finest root classes (<200 μm in diameter) represented 50% (younger) to 43% (older) of total fine root length, with ~15% of all root length < 100 μm. While 90% of total fine root length was in roots <0.07 cm, the bulk of root surface area was in the larger fine root classes (0.07–0.20 cm). The root systems were clearly revealed by neutron radiography measurements, particularly the larger root classes, and image contrast improved after several days of soil drying as water was extracted by roots (Fig. 1e and f). The radiographs also revealed some non-uniformity in the soil structure, including several voids towards the bottom of the chamber. While much of the water drained to the bottom of the chamber quickly, there was some evidence of high residual water content vertically through the profile under the injection points (Fig. 1e – notice the tape covering the injection points at the very top of the chamber).
Table 1

Plant dimensions for newer post-drought (left) and older pre-drought (right) root systems based on automated WinRhizo analysis of the excavated root systems

 

Total Root Length (cm)

Length of Fine Roots <200 μm (cm)

% Root Length with Diameter < 0.5 mm

Root Surface Area (cm2)

Mean Root Diameter (mm)

#

Projected Area (cm2)

Fresh weight (g)

Dry weight (g)

Roots – left side

952.8

478.8

82

106.8

0.357

35.11

0.120

Roots – right side

928.2

405.5

76

120.8

0.414

39.75

0.164

Leaves

  

10.94

16

55.71

1.086

0.316

Stem

6.75

  

1.89

1.27

0.189

0.049

Total volumetric water content in the soil declined through time due to root water extraction and loss through transpiration (Fig. 2a). The loss in water content over the two-day cycle was also verified by monitoring the weight of the chamber with time – total water loss was 28.65 g as soil water was progressively extracted by roots. NR based water content differed from mass-based water content due to the NR background (chamber and soil) normalization that included residual water content. Assuming θ r was 1.54% (close to the Kang et al. 2014 estimate of 2.5%) the NR and mass-based measurements of water content through time were the same. The mass-based and NR-based decline in water content through time was the same, with a slope of −0.0134 g min−1, confirming the viability of using NR to assess actual water content across a range of concentrations in agreement with our prior calibrations (Kang et al. 2013).
Fig. 2

a Total volumetric water content (cm3) in different regions of interest from top to bottom in plant – soil system as a function of time based on neutron radiography. Inset shows the different ROI’s used for the plot. b Water content for left side (finer roots) and (c) right side (coarser roots) by layer through time. The gap in data at ~23 h occurred when the chamber was removed and weighed. The numbers shown in the figures are the number of layers depicted in the inset to the respective figure

To understand the vertical patterns through the soil profile, different ROIs within the plant chamber were selected, as highlighted in the inset to Fig. 2a. Analysis of the neutron radiographs revealed that water content declined in all layers though time, with most of the water being extracted from the deeper layers that had the highest water content (total extraction of 6.87 g or 19.89 g for the four upper or four lower layers, respectively). The upper layers were relatively dry at the beginning of experiment, and approached the residual water content as water was extracted from the system. There were similar temporal extraction patterns for the newer (Fig. 2b) and older (Fig. 2c) root systems. As the upper soil layers dried there was a stronger reliance on water extraction from the deepest layer, as indicated by the increased slope of water extraction, which becomes strongly apparent after about 16 h (Fig. 2). Looking at the relative water use, more water was extracted by the plant from the right side (old roots) of the chamber, than the left (new roots) side (Fig. 3a), and the differential water extraction increased through time for the deeper soils. In some of the deeper layers it is also likely that there was some horizontal redistribution of water from the left to the right sides, based on slightly lower water content for the left side that would create a driving force for movement (e.g., layers 1–3 in Fig. 2b and c). This would suggest our measurements may have slightly underestimated the differences in water uptake from the two sides. As the upper soil dried and water became less available, water extraction became more similar for both sides of the chamber (increase in ratio for top layer at 30 h in Fig. 3a). In contrast to total water uptake, the water extraction per unit root surface area was initially 25–30% greater for the left side (new roots) of the chamber compared with the right (old roots), but this declined through time as the soil dried indicating similar rates of water extraction as water availability diminished (Fig. 3b).
Fig. 3

a The ratio of soil water content for the left (finer roots): right (coarser roots) sides of the chamber through time, separated into top, middle and bottom layers, and (b) the relative root uptake rate per unit root area per unit time for the left (finer roots): right (coarser roots) sides of the chamber

Application of root mask to NR

The root mask was used to assess water dynamics near the major roots within the newer (left) and older (right) sides of the chamber (Fig. 4). The mask could identify and segment only the larger roots and primarily roots in the upper soil, as wetter conditions and more roots towards the bottom of the chamber limited resolution. Water content was initially greater for the right side of the chamber than the left side, but this declined through time (Fig. 4c and d). For both sides of the chamber the water content was greatest near the root surface, declined sharply until ~0.03 cm into the rhizosphere, gradually increased in water content until peaking ~0.15 cm from the root, then slowly declined further into the soil (Fig. 4c and d). This wet-drier-wetter-drier pattern was most apparent in the initial radiographs, and dampened with increasing time elapsed. Using the root mask, the near root water content remained substantially greater than the bulk soil during the drying period. Overall, the water content in the rhizosphere (θrhizo) was initially 28% greater in water content than the bulk soil (θbulk), and that increased to ~89% greater by the end of the dry period as soil water became less available (Fig. 5). The left side (new roots) θrhizo increased from 31% to 83%, while right side (old roots) θrhizo increased from 24% to 94% greater than θbulk during drying.
Fig. 4

a Results of a mask applied to a radiograph to segment soil from major roots, b radiograph indicating two regions of interest, left side (finer roots) and right side (coarser roots) for automated analysis of soil water content using the mask; c Mean water content (cm3/cm3) of soil with distance from roots in upper left side of chamber (newer roots) or (d) upper right side of chamber (older roots) for six time points over the 37-h experiment

Fig. 5

Total water content in the bulk soil (0.5–0.51 cm from root) versus the near root rhizosphere soil (0–0.01 cm from root) over time, for the left and right sides of the upper portion of the chamber in Fig. 4

There were substantial differences in root distribution between digitally scanned images and segmented mask images. Both images indicated threshold root diameters of 0.010 cm, but the peak frequency of root diameters was greater for the mask images, indicating larger root diameters overall (Fig. 6a and b). The scanned images also show much larger counts of the smallest roots (>50% of total roots) as compared to the smallest roots in segmented radiographs (only ~20% of total roots) (Fig. 6c-f). The base of the stem where large lateral roots emerge (Fig. 1d) was evident as a peak at ~0.2 cm root diameter in both the scanned and masked image analyses (Fig. 6e and f) indicating good correlation between techniques for the largest root classes. The much larger number of small roots present in the frequency distribution of the scanned image reduced the prominence of this peak as illustrated by the smaller peak in the scanned image (Fig. 6e and f).
Fig. 6

Radial size distribution of (a) the digitally scanned roots and (b) the segmented mask roots, including those on the left side of the chamber (finer roots) (c, d) and those on the right side of the chamber (coarser roots) (e, f). Color indicates depth, and note the segmented images (d, f) only have the upper three layers based on the analyzed mask (Fig. 4b)

Quantifying root water uptake dynamics

We calculated the upward flux of water in roots through time using measured water contents by first-order differencing in space and time, as described above. Results of that model-based analyses expressed as root water uptake per layer are shown in Fig. 7. Initially, the root water uptake in layer 1 exceeds the total transpiration by a large factor and is compensated by water loss from roots in layers 2 and 3 (shown as negative root water uptake in Fig. 7). After about 17 h, the cumulative flow increased toward the surface, indicating net flow of water from soils to roots for all layers. These results are plotted another way in Fig. 8, which shows vertical profiles of root water uptake or apparent release after 7 h, 21.75 h, and at the end of the experiment after 37.5 h. At the end of the experiment 37%, 20% and 12.5% of the total transpiration was extracted from layers 1, 2 and 3 respectively. With knowledge of root surface area in each one of the soil layers, we could test if root water uptake within the vertical profile was equivalent to the vertical profile of fine-root surface area, as models often assume. These two quantities are cross-plotted in Fig. 9 at the same three time points. There is no apparent relationship between root water uptake and specific root surface area in the early stages of the experiment as the soil column responded to the initial pulse of water (Fig. 9a). At 21.7 h the root water uptake has a monotonic but nonlinear dependence on specific surface area. The relationship between root water uptake and specific surface area approaches linear in the late stages of the experiment (Fig. 9c).
Fig. 7

Root water uptake inferred from the measured soil water contents and independently measured soil properties for each of the nine soil layers. Layer 1 is the bottom layer. Negative flow rates in Layers 2 and 3 suggest substantial root hydraulic redistribution and root release of water, or that the measured soil hydraulic parameters based on root-free soil are not applicable to this system

Fig. 8

Vertical profile of root water uptake (RWU) using model-based analysis at three times (7, 21.75, 37.5 h, left to right). Model-based estimates of root water uptake in the deepest layer was greater than used for transpiration at early times indicating substantial loss of water to the soil in overlying layers, or that modeled assumptions of drainage were incorrect

Fig. 9

Cross-plot of water uptake and root surface area at 7, 21.75 and 37.5 h

As noted above, the parameters in the Brooks-Corey model for the soil moisture characteristic curve are based on clean sand and there is uncertainty in the hydraulic properties of the soil system in the presence of fine roots. To test the robustness of the inferred root water uptake to the soil moisture characteristic curve, we repeated the analyses using different values of the ψ a and λ parameters. The inferred root water uptake versus time assuming a 50% increase in ψ a is shown in Fig. 10. As in Fig. 7, the inferred root water uptake in layer 1 exceeds the total transpiration at early times, and is compensated by a net transfer of water from roots to soil in layers 2 and 3. After about 8 h, the inferred root water uptake in layer 2 becomes larger than the total transpiration and is compensated by a net flow of water from roots to soil in layer 1. As in Fig. 7, it is only at late times that the inferred flow is consistent with water movement from soil to roots in all layers. Further increases in ψ a increased the inferred flow of water from roots to soil in layer 1 (results not shown). The same trends were obtained when the λ parameter was decreased (results not shown).
Fig. 10

Inferred root water uptake through time assuming a 50% increase in modeled air-entry suction (ψ a ) for the Brooks and Corey water release curve

Results consistent with those of Figs. 7, 8 and 9 were obtained when we assessed the data using forward modeling. Water content versus time obtained by solving Richards equation with the assumption that root water uptake is proportional to root surface area using three different values for saturated hydraulic conductivity (10−2 cm/s, 10−3 cm/s, and 10−4 cm/s) and the reference soil moisture retention parameters are shown for layers 1, 3 and 9 in Fig. 11. Recall that the value 10−2 cm/s was measured for this medium without roots and 10−3 cm/s was used in Fig. 9 and judged to be a lower bound on the rate of soil water drainage. In the deepest layer, the modeled water content increases rapidly at early times for the 10−2 cm/s and 10−3 cm/s cases, then decreases. The initial rise is inconsistent with the observations. To eliminate that initial rise, we had to decrease the saturated hydraulic conductivity by an unrealistic factor of 100 to 10−4 cm/s. The general inconsistency with the observations indicates that either root water uptake is significantly greater than modeled in the lowermost layer or the soil is draining significantly slower than would be expected. At layer 3, the modeled water content decreases faster than the observation, indicating that the root water uptake is less than the modeled result. The opposite is true for layer 1.
Fig. 11

Soil water content based on measured neutron radiographs (circles) or based on forward modeling that assumed root water uptake was proportional to root surface area (lines) for (a) the bottom layer 1, (b) layer 3 and (c) the uppermost layer 9 (from Fig. 2a). The black, blue, and orange curves in each plot correspond to soil saturated hydraulic conductivities of 1.6 × 10−2 cm/s, 1.6 ×10−3 cm/s, and 1.6 × 10−4 cm/s, respectively. Note that the measured value for the sand used in this experiment is 1.6 × 10−2 cm/s

Taken together these results are indicators that vertical profile of root water uptake is not as simple as the common assumption of making it proportional to the root surface area.

Discussion

Neutron radiography allows microscopic visual quantification of actual root function (i.e., high resolution growth and water uptake dynamics) and can provide understanding of how key root traits function in situ; e.g., how do root diameter, length, age or degree of suberization (white versus brown roots) affect water uptake? Understanding how these types of root traits are linked to function is a key component to scaling relevant data into larger models that describe biochemical, hydrological and energy cycling between the soil and the atmosphere (McCormack et al. 2015; Warren et al. 2015). Such models based on mechanistic understanding can be useful in understanding the impacts of extreme weather events such as heat or drought on ecosystem function. Further, modeling the root system provides an opportunity to improve our understanding about root water uptake by plants with different root traits, and under different environmental conditions.

Analysis of water content in the composite radiographs using our earlier calibrations and soil water release curve (Kang et al. 2013, 2014) indicated whole chamber NR-based water content differed by only 6% from actual weight-based measurements, 48.55 g and 46.37 g, respectively. The minor discrepancy could be due to small differences in soil structure due to the experimental setup (a flat plate 1 cm thick in this study, as compared with the 2.56 cm diameter cylinder used in our earlier soil water release curve calibrations – which have different degrees of sand to chamber edges), the presence of voids or differences in soil bulk density, estimated versus actual residual soil water content (due to the need to harvest fresh roots) and the extensive presence of roots that could have compacted rhizosphere soil (Aravena et al. 2011). But NR analysis did track weight based changes in water content very well, and could also delineate many of the roots (but not the finest roots), their growth and root, rhizosphere and bulk soil relative hydration.

Following the irrigation event, water moved quickly through the sandy soil profile then over time redistributed throughout the rest of the soil. As such, the water content was greatest at the bottom of the chamber, where there was also the greatest density of roots. Initially, soil was expected to become dryer (reduction in water content) as a function of time, and roots and the root rhizosphere would become wetter (increase in water content) by extracting water from the soil. However, we observed that both for the root rhizosphere and soil, water content declined as a function of time. This may be related to an initial rapid water absorption by roots and rehydration over a short time span, within the first 40 mins (time difference between two 16-image data points measured).

While there are limited studies that measure root water uptake per unit root surface area, those that do report wide variation in uptake rates due to root order, root diameter, root age, suberization, species, depth in soil, soil type (or solution), or matric/osmotic potential driving force. In our study, average root water uptake rates per unit surface area during the first 24 h of the experiment were 0.0027–0.0116 g cm−2 h−1, greater than rates measured in situ for beech, oak and spruce (0.0002–0.0026 g cm−2 h−1; Leuschner et al. 2004), less than rates measured in pine seedling roots (~0.015 g cm−2 h−1; Chung and Kramer 1975) or lupine (~0.018–0.043 g cm−2 h−1; Zarebanadkouki et al. 2014), and much less than the maximum rates in solution for Citrus sp. first-third order roots (~0.1–1.0 g cm−2 h−1; Rewald et al. 2011). Our results were also in the range reported for barley, sunflower and maize root water uptake in solution (~0.0005–0.0200 g cm−2 h−1; Aston and Lawlor 1979). Further research is needed to understand the contributions of different root ages, root sizes and root order to total root water uptake (Wells and Eissenstat 2003; Rewald et al. 2011).

Our results support other observations that whiter, finer roots have greater water uptake per unit root surface area than larger, more developed roots. The root system on the left side of the chamber was initially whiter, with more visible fungal hyphae than the root system on the right, which was initially more mature and visibly browner (Fig. 1a and b), although after a month the two root systems looked much more similar, albeit displayed different diameter distributions. Despite greater length and greater water uptake rates per unit surface area in the smaller diameter root system, the larger diameter root system had more total surface area and was thus more important for total plant water uptake. Part of this difference may be attributable to root morphology; the finest root orders have high uptake rates per unit surface area, but small cell diameter and limited xylem cross-sectional area greatly restricts bulk transport axially towards the plant – a capacity that develops through time as the larger transport roots develop (Valenzuela-Estrada et al. 2008). Since fine root uptake rates also vary with age, the relative uptake rates between two root classes may change through time – initially increasing as tissues develop, then declining as tissues mature or senesce (Wells and Eissenstat 2003). Our study of the functionality of the two sides of the root system represents just one point in time of root system development following the extreme drought period – a development that is highly dynamic as new roots initiate, explore, mature or senesce. Relative uptake rates were also highly dependent on soil water availability. Uptake rates from the two sides of the root system became more similar as water availability declined, indicating either a limitation of soil water movement to the root surface or increased internal resistance due to hydraulic failure. As simple as our model system was, these dynamics illustrate the incredible complexity in resolving mechanistic whole root system water uptake in situ.

Most models of root-water uptake presume that total transpiration demand is distributed vertically to the soil according to root specific surface area. We found vertical profiles of root water uptake to be dramatically different from that assumption initially as the soil was responding to the initial water pulse. Indeed, our interpretation of the soil water dynamics suggests significant hydraulic redistribution initially, with water flowing from roots to soil in one vertical interval. Approximately 24 h after the initial pulse of water was introduced, the vertical profile of root water uptake is reasonably approximated by the profile of specific surface area. Those results suggest that root water uptake following an infiltration event may be considerably more complex than represented in typical root-water uptake modeling even if the dynamics are well approximated in the absence of an infiltration event.

It is important to note that our approach for partitioning vertical water movement to roots and soil presumes that the hydraulic properties of the soil when roots are present are not significantly different from root-free soil, which introduces uncertainty in our interpretation. Regardless, the results offer a cautionary note regarding models of the coupled soil-root system. If the hydraulic properties of the soils are well approximated by that of root-free soil, then the transient dynamics of root water uptake following infiltration events may be considerably different from that represented by common root water uptake modeling approaches. If the soil properties are not well approximated by those of root-free soil, then we cannot say anything definitive about root water uptake but that the standard approach of using pedotransfer functions to estimate bulk soil properties is inadequate for describing the system dynamics. A potential implication of either situation is that the amount of water that drains past the root zone and becomes deep infiltration may not be well represented in current ecosystem-scale models. The experimental and modeling results in our study may indicate a decline in drainage due to root presence, which both fills soil pores, shifts the capillary fringe and impacts the water retention curve.

Root and soil rehydration in the upper soils creates a driving force for water movement up through capillarity and hydraulic redistribution. The main root diameter class was 0.1–0.2 mm and the main sand grain size was 0.15–0.2 mm – so most of the pores were 0.1 mm or less, which could be readily explored by the finest roots. This network of roots (~19 m) created a hydraulic pathway through the pore spaces spanning multiple layers. As such, capillary wicking along the root exterior likely provided much of the interlayer water movement, independent of root water uptake into the xylem and internal hydraulic redistribution. After about 16 h, the cumulative root water uptake was increasing toward the surface, indicating little change in root water storage, a net transfer of water from soil to roots, and greatly reduced soil or root capillarity or hydraulic redistribution of water between layers.

While we continue to learn more about fine root anatomy and morphology (e.g., Pregitzer et al. 2002; Wells and Eissenstat 2003; Guo et al. 2008), and fine root trait data are becoming increasingly assessable (Iversen et al. 2017), linking specific fine root traits to functional uptake of water (or nutrients) remains a challenge for the research community. NR has been a valuable tool to delineate larger fine roots (e.g., 0.5–2.0 mm) and their dynamics from the background soil environment, yet current detector resolution limits identification of the finest, most active roots (<200 um), root hairs, or mycorrhizal hyphae. Newer detectors are being developed but thus far have a small field of view limiting investigation of entire root systems. Thus, a challenge remains for non-invasive techniques such as NR is resolving fine root structure and function in situ indicating a need for improved detectors or novel NR techniques, or continued development of alternate methods such as X-ray imaging. Our results also conclude that soil with roots behaves quite differently than soil without roots. As obvious as this is, there is limited literature describing how soil hydraulic properties change with presence or absence of roots and mycorrhizal hyphae, suggesting a need for additional research.

Notes

Acknowledgements

We thank Deanne Brice for plant propagation and root analysis. Research sponsored by the Laboratory Directed Research and Development Program of Oak Ridge National Laboratory (ORNL), managed by UT-Battelle, LLC, for the U. S. Department of Energy (DOE), by the DOE Office of Science, Office of Biological and Environmental Research, and by the DOE, Office of Science, Office of Workforce Development for Teachers and Scientists, Office of Science Graduate Student Research (SCGSR) program. The SCGSR program is administered by the Oak Ridge Institute for Science and Education (ORISE) for the DOE. ORISE is managed by ORAU under contract number DE-AC05-06OR23100. ORNL is managed by UT-Battelle, LLC, for the DOE under contract DE-AC05-1008 00OR22725. This research used resources at the High Flux Isotope Reactor, a DOE Office of Science User Facility operated by ORNL.

This manuscript has been authored by UT-Battelle, LLC under Contract No. DE-AC05-00OR22725 with the US Department of Energy. The United States Government retains and the publisher, by accepting the article for publication, acknowledges that the United States Government retains a non-exclusive, paid-up, irrevocable, world-wide license to publish or reproduce the published form of this manuscript, or allow others to do so, for United States Government purposes. The Department of Energy will provide public access to these results of federally sponsored research in accordance with the DOE Public Access Plan (http://energy.gov/downloads/doe-public-access-plan).

References

  1. Ahmed MA, Zarebanadkouki M, Kaestner A, Carminati A (2016) Measurements of water uptake of maize roots: the key function of lateral roots. Plant Soil 398:59CrossRefGoogle Scholar
  2. Allen MF (2007) Mycorrhizal fungi: highways for water and nutrients in arid soils. Vadose Zone J 6:291–297CrossRefGoogle Scholar
  3. Aravena JE, Berli M, Ghezzehei TA, Tyler SW (2011) Effects of root induced compaction on rhizosphere hydraulic properties—X-ray microtomography imaging and numerical simulations. Environ Sci Technol 45(2):425–431.  https://doi.org/10.1021/es102566j CrossRefPubMedGoogle Scholar
  4. Aston MJ, Lawlor DW (1979) The relationship between transpiration, root water uptake and leaf water potential. J Exp Bot 30:169–181CrossRefGoogle Scholar
  5. Bilheux J, Bilheux HZ (2015) iMARS (iMaging analysis research software). Phys Procedia 69:343–348CrossRefGoogle Scholar
  6. Brooks RH, Corey AT (1964) Hydraulic properties of porous media hydrology paper no 3. Civil Engineering Department, Colorado State University, Fort CollinsGoogle Scholar
  7. Carminati A, Moradi AB, Vetterlein D, Vontobel P, Lehmann E, Weller U, Vogel H, Oswald SE (2010) Dynamics of soil water content in the rhizosphere. Plant Soil 332:163–176CrossRefGoogle Scholar
  8. Carminati A, Schneider CL, Moradi AB, Zarebanadkouki M, Vetterlein D, Vogel HJ, Hildebrandt A, Weller U, Schüler L, Oswald SE (2011) How the rhizosphere may favor water availability to roots. Vadose Zone J 10:988–998CrossRefGoogle Scholar
  9. Carminati A, Vetterlein D, Koebernick N, Blaser S, Weller U, Vogel H-J (2013) Do roots mind the gap? Plant Soil 367(1–2):651–661CrossRefGoogle Scholar
  10. Chung H-H, Kramer PJ (1975) Absorption of water and 32P through suberized and unsuberized roots of loblolly pine. Can J For Res 5(2):229–235.  https://doi.org/10.1139/x75-031 CrossRefGoogle Scholar
  11. Daly KR, Tracy SR, Crout NMJ, Mairhofer S, Pridmore TP, Mooney SJ, Roose T (2017) Quantification of root water uptake in soil using X-ray computed tomography and image based modelling. Plant Cell Environ.  https://doi.org/10.1111/pce.12983
  12. Dardanelli JL, Ritchie JT, Calmon M, Andriani JM, Collino DJ (2004) An empirical model for root water uptake. Field Crop Res 87:59–71CrossRefGoogle Scholar
  13. Doussan C, Pierret A, Garrigues E, Pagès L (2006) Water uptake by plant roots: II – modeling of water transfer in the soil root–system with explicit account of flow within the root system – comparison with experiments. Plant Soil 283:99–117CrossRefGoogle Scholar
  14. Esser HG, Carminati A, Vontobel P, Lehmann EH, Oswald SE (2010) Neutron radiography and tomography of water distribution in the root zone. J Plant Nutr Soil Sci 173:757–764CrossRefGoogle Scholar
  15. Feddes RA, Hoff H, Bruen M, Dawson T, de Rosnay P, Dirmeyer P, Jackson RB, Kabat P, Kleidon A, Lilly A, Pitman AJ (2001) Modeling root water uptake in hydrological and climate models. Bull Am Meteorol Soc 82:2797–2809CrossRefGoogle Scholar
  16. Gardner WR (1960) Dynamic aspects of water availability to plants. Soil Sci 89:63–73CrossRefGoogle Scholar
  17. Guo D, Xia M, Wei X, Chang W, Liu Y, Wang Z (2008) Anatomical traits associated with absorption and mycorrhizal colonization are linked to root branch order in twenty-three Chinese temperate tree species. New Phytol 180:673–683CrossRefPubMedGoogle Scholar
  18. Hainsworth JM, Aylmore LAG (1989) Non-uniform soil water extraction by plant roots. Plant Soil 113:121–124CrossRefGoogle Scholar
  19. Iversen CM, McCormack ML, Powell AS, Blackwood CB, Freschet GT, Kattge J, Roumet C, Stover DB, Soudzilovskaia NA, Valverde-Barrantes OJ, van Bodegom PM, Violle C (2017) Viewpoint: a global fine-root ecology database to address belowground challenges in plant ecology. New Phytol.  https://doi.org/10.1111/nph.14486
  20. Javaux M, Schroeder T, Vanderborght J, Vereecken H (2008) Use of a three-dimensional detailed modeling approach for predicting root water uptake. Vadose Zone J 7:1079–1088CrossRefGoogle Scholar
  21. Javaux M, Couvreur V, Vander Borght J, Vereecken H (2013) Root water uptake: from three-dimensional biophysical processes to macroscopic modeling approaches. Vadose Zone J 12.  https://doi.org/10.2136/vzj2013.02.0042
  22. Johnson DM, Sherrard ME, Domec J-C, Jackson RB (2014) Role of aquaporin activity in regulating deep and shallow root hydraulic conductance during extreme drought. Trees.  https://doi.org/10.1007/s00468-014-1036-8
  23. Kang M, Bilheux HZ, Voisin S, Cheng C, Perfect E, Horita J, Warren JM (2013) Water calibration measurements for neutron radiography: application to water content quantification in porous media. Nucl Instrum Method A.  https://doi.org/10.1016/j.nima.2012.12.11
  24. Kang M, Perfect E, Cheng CL, Bilheux HZ, Lee J, Horita J, Warren JM (2014) Multiple pixel-scale water retention curves quantified by neutron radiography. Adv Water Resour 65:18CrossRefGoogle Scholar
  25. Lai CT, Katul G (2000) The dynamic role of root-water uptake in coupling potential to actual transpiration. Adv Water Resour 23:427–439CrossRefGoogle Scholar
  26. Leung AK, Garg A, Coo JL, Ng CWW, Hau BCH (2015a) Effects of the roots of Cynodon dactylon and Schefflera heptaphylla on water infiltration rate and soil hydraulic conductivity. Hydrol Process 29:3342–3354.  https://doi.org/10.1002/hyp.10452 CrossRefGoogle Scholar
  27. Leung AK, Garg A, Ng CWW (2015b) Effects of plant roots on soil-water retention and induced suction in vegetated soil. Eng Geol 193:183–197CrossRefGoogle Scholar
  28. Leuschner C, Coners H, Icke R (2004) In situ measurement of water absorption by fine roots of three temperate trees: species differences and differential activity of superficial and deep roots. Tree Physiol 24(12):1359–1367.  https://doi.org/10.1093/treephys/24.12.1359 CrossRefPubMedGoogle Scholar
  29. Macfall JS, Johnson GA, Kramer PJ (1990) Observation of a water-depletion region surrounding loblolly-pine roots by magnetic resonance imaging. Proc Natl Acad Sci U. S. A. 87:1203–1207CrossRefPubMedPubMedCentralGoogle Scholar
  30. McCormack ML, Dickie IA, Eissenstat DM, Fahey TJ, Fernandez CW, Guo D, Helmisaari HS, Hobbie EA, Iversen CM, Jackson RB, Leppälammi-Kujansuu J, Norby RJ, Phillips RP, Pregitzer KS, Pritchard SG, Rewald B, Zadworny M (2015) Redefining fine roots improves understanding of below-ground contributions to terrestrial biosphere processes. New Phytol 207:505–518CrossRefPubMedGoogle Scholar
  31. Moradi AB, Conesa HM, Robinson B, Lehmann E, Kuehne G, Kaestner A, Oswald S, Schulin R (2008) Neutron radiography as a tool for revealing root development in soil: capabilities and limitations. Plant Soil 318:243–255CrossRefGoogle Scholar
  32. Moradi AB, Carminati A, Vetterlein D, Vontobel P, Lehmann E, Weller U, Hopmans JW, Vogel H-J, Oswald SE (2011) Three-dimensional visualization and quantification of water content in the rhizosphere. New Phytol 192:653–663CrossRefPubMedGoogle Scholar
  33. Nakanishi TM, Okuni Y, Hayashi Y, Nishiyama H (2005) Water gradient profiles at bean plant roots determined by neutron beam analysis. J Radioanal Nucl Chem 264:313–317CrossRefGoogle Scholar
  34. Ng CWW, Ni JJ, Leung A, Wang ZJ (2016) A new and simple water retention model for root-permeated soils. Geotech Lett 6(1):1–6CrossRefGoogle Scholar
  35. Oswald SE, Menon M, Carminati A, Vontobel P, Lehmann E, Schulin R (2008) Quantitative imaging of infiltration, root growth, and root water uptake via neutron radiography. Vadose Zone J 7:1035–1047CrossRefGoogle Scholar
  36. Pregitzer KS, DeForest JL, Burton AJ, Allen MF, Ruess RW, Hendrick RL (2002) Fine root architecture of nine north American trees. Ecol Monogr 72:293–309CrossRefGoogle Scholar
  37. Preibisch S, Saalfeld S, Tomancak P (2009) Globally optimal stitching of tiled 3D microscopic image acquisitions. Bioinformatics 25:1463–1465CrossRefPubMedPubMedCentralGoogle Scholar
  38. Querejeta JI (2017) Soil water retention and availability as influenced by mycorrhizal symbiosis: consequences for individual plants, communities, and ecosystems. In: Johnson NC, Gehring C, Jansa J (Eds) Mycorrhizal mediation of soil: fertility, structure, and carbon storage. Elsevier Inc, p 299-317Google Scholar
  39. Rewald B, Ephrath JE, Rachmilevitch S (2011) A root is a root is a root? Water uptake rates of Citrus root orders. Plant Cell Environ 34:33–42CrossRefPubMedGoogle Scholar
  40. Roose T, Fowler AC (2004) A mathematical model for water and nutrient uptake by plant root systems. J Theor Biol 228:173–184CrossRefPubMedGoogle Scholar
  41. Roose T, Keyes SD, Daly KR (2016) Challenges in imaging and predictive modeling of rhizosphere processes. Plant Soil 407:9CrossRefGoogle Scholar
  42. Santodonato L, Bilheux HZ, Bailey B, Bilheux J, Nguyen PT, Tremsin AS, Selby DL, Walker L (2015) The CG-1D neutron imaging beamline at the oak Ridge National Laboratory high flux isotope reactor. Phys Procedia 69:104CrossRefGoogle Scholar
  43. Schymanski SJ, Sivapalan M, Roderick ML, Beringer J, Hutley LB (2008) An optimality-based model of the coupled soil moisture and root dynamics. Hydrol Earth Syst Sci 12:913–932CrossRefGoogle Scholar
  44. Segal E, Kushnir T, Mualem Y, Shani U (2008) Microsensing of water dynamics and root distributions in sandy soils. Vadose Zone J 7:1018–1026CrossRefGoogle Scholar
  45. Thévenaz P, Ruttimann UE, Unser M (1998) A Pyramid Approach to Subpixel Registration Based on Intensity. IEEE Trans Image Process 7:27–41CrossRefPubMedGoogle Scholar
  46. Tumlinson LG, Liu H, Silk WK, Hopmans JW (2008) Thermal neutron computed tomography of soil water and plant roots. Soil Sci Soc Am J 72:1234–1242CrossRefGoogle Scholar
  47. Valenzuela-Estrada LR, Vera-Caraballo V, Ruth LE, Eissenstat DM (2008) Root anatomy, morphology and longevity among root orders in Vaccinium corymbosum (Ericaceae). Am J Bot 95:1506–1514CrossRefPubMedGoogle Scholar
  48. Warren JM, Bilheux H, Kang M, Voisin S, Cheng C, Horita J, Perfect E (2013a) Neutron imaging reveals internal plant water dynamics. Plant Soil 366:683–693CrossRefGoogle Scholar
  49. Warren JM, Bilheux H, Cheng C, Perfect E (2013b) Reply to: comment on ‘neutron imaging reveals internal plant water dynamics’. Plant Soil 371:15–17CrossRefGoogle Scholar
  50. Warren JM, Hanson PJ, Iversen CM, Kumar J, Walker AP, Wullschleger SD (2015) Root structural and functional dynamics in terrestrial biosphere models – evaluation and recommendations. New Phytol 205:59–78CrossRefPubMedGoogle Scholar
  51. Wells CE, Eissenstat DM (2003) Beyond the roots of young seedlings: the influence of age and order on fine root physiology. J Plant Growth Regul 21:324–334CrossRefGoogle Scholar
  52. Zarebanadkouki M, Kim YX, Moradi AB, Vogel H-J, Kaestner A, Carminati A (2012) Quantification and modeling of local root water uptake using neutron radiography and deuterated water. Vadose Zone J:11.  https://doi.org/10.2136/vzj2011.0196
  53. Zarebanadkouki M, Kim YX, Carminati A (2013) Where do roots take up water? Neutron radiography of water flow into the roots of transpiring plants growing in soil. New Phytol 199:1034–1044CrossRefPubMedGoogle Scholar
  54. Zarebanadkouki M, Kroener E, Kaestner A, Carminati A (2014) Visualization of root water uptake: quantification of deuterated water transport in roots using neutron radiography and numerical modeling. Plant Physiol 166:487–499CrossRefPubMedPubMedCentralGoogle Scholar
  55. Zarebanadkouki M, Ahmed MA, Carminati A (2016) Hydraulic conductivity of the root–soil interface of lupin in sandy soil after drying and rewetting. Plant Soil 398:267–280CrossRefGoogle Scholar

Copyright information

© US Government (outside the USA) 2017

Authors and Affiliations

  • Indu Dhiman
    • 1
  • Hassina Bilheux
    • 1
  • Keito DeCarlo
    • 1
    • 2
  • Scott L. Painter
    • 3
  • Lou Santodonato
    • 1
  • Jeffrey M. Warren
    • 3
  1. 1.Chemical and Engineering Materials DivisionOak Ridge National LaboratoryOak RidgeUSA
  2. 2.Department of Civil and Environmental EngineeringPrinceton UniversityPrincetonUSA
  3. 3.Climate Change Science Institute and Environmental Sciences DivisionOak Ridge National LaboratoryOak RidgeUSA

Personalised recommendations