Novel scanning procedure enabling the vectorization of entire rhizotron-grown root systems
© Lobet and Draye; licensee BioMed Central Ltd. 2013
Received: 15 October 2012
Accepted: 17 December 2012
Published: 4 January 2013
This paper presents an original spit-and-combine imaging procedure that enables the complete vectorization of complex root systems grown in rhizotrons. The general principle of the method is to (1) separate the root system into a small number of large pieces to reduce root overlap, (2) scan these pieces one by one, (3) analyze separate images with a root tracing software and (4) combine all tracings into a single vectorized root system. This method generates a rich dataset containing morphological, topological and geometrical information of entire root systems grown in rhizotrons. The utility of the method is illustrated with a detailed architectural analysis of a 20-day old maize root system, coupled with a spatial analysis of water uptake patterns.
Root systems are responsible for the capture of below-ground resources such as nutrients and water. As such, they are thought to be play a central role in the yield establishment of crop plants [1–3]. The availability of a given resource for the plant can be seen as the integration of soil and roots bio-physical constraints. On the one side, the resource distribution in the soil and its mobility defines its potential availability for the plant. On the other side, the root system architecture (RSA, including morphology and topology) and the root placement in the soil domain (spatial correlation between the roots and the resource) defines the actual resource availability for the plant . However, root and soil constraints do not add independently and emerging behaviors are likely to arise from the non linearity of the soil-root system [5, 6]. Therefore, detailed datasets containing root system architecture, root placement and soil resource dynamics are required to improve our understanding of resource capture by plant roots.
Unfortunately, few techniques allow the simultaneous acquisition of precise soil and root information. Precise quantification of the root systems can be done by using, for instance, X-ray computed tomography [7, 8], magnetic resonance imaging  or transparent artificial soils . Among more classical techniques, growing plants in flat transparent culture boxes (rhizotrons) is widely used in root research. This simple technique provides an easy way to observe the growth and development of a large number of plants in a soil-like substrate . Moreover, rhizotrons allow some level of soil observation, as with the light transmission imaging  or neutron radiography . These enable a fine analysis of soil-root relations, given that sufficient information is obtained about the plant and the soil.
A general restriction of the currently available root image analysis softwares lies in the difficulty to analyze highly branched root systems with a large degree of root overlap in 2D (Figure 1B, close-up). Not surprisingly, a trade-off is usually observed between the degree of automation and the level of detail obtained. Among existing software, SmartRoot  provides, to our knowledge, the most exhaustive root architecture dataset, containing detailed morphological, topological, spatial and temporal information. In addition, the SmartRoot dataset enables more complex architectural analysis . For example, Fitter’s architectural indexes can be computed from topological information  or dynamic traits such as the growth rates of first and second order roots can be inferred from static topological and morphological traits .
Comparison of the different methods
The acetate sheet (obtained in situ) is then attached on the glass of the scanner (Figure 2D) and, one by one, each root part is laid on the scanner, aligned to its corresponding tracing and scanned (Figure 2E). The whole scanning procedure yields a set of high resolution registered images where roots (1) are positioned as in the rhizotron and (2) display a much reduced degree of root overlap. The subsequent tracing of the roots is realized with SmartRoot (Figure 2F) on the individual images and the resulting morphological datasets are combined into a unique and complete vectorized root system.
Results and discussions
The ability of the method to generate vectorized versions of entire root systems and its utility in the framework of soil-root-interaction research is demonstrated with the analysis of water uptake dynamics in maize. Plants were grown in thin rhizotrons under non-limiting conditions until emergence of the sixth leave. The nutrient solution supply was then interrupted and the evolution of the 2D soil water content was monitored during three days using light transmission imaging .
Step 1: Architectural analysis of complex root systems
These different variables are highly informative in the framework of root water uptake research. For instance, assuming that the cumulative volume (Figure 6B) is proportional to the quantity of water crossing the segments, Figure 6B highlights the large differences in terms of axial water flow between the first and second order roots (up the 12 orders of magnitudes). This assumption is supported by the fact that first order roots have much thicker roots (Figure 6D) with larger cross-sectional xylem area. On the other hand, the second and third order roots have much thinner root (Figure 6D), but represent the majority of the total root length and surface (Figure 6C and E). Moreover, these roots tend to have a lesser gravitropic behavior (Figure 6F), ensuring a better exploration of the horizontal soil layers. The root system can therefore be divided in two functional type of roots: the primary roots, responsible for (1) the vertical exploration of the soil and (2) the majority of the axial transport of water to the shoot, and the lesser order roots responsible for (1) the horizontal exploration of the soil layers and (2) the majority of the water extraction.
Step 2: Local analysis of root-soil interactions
As mentioned earlier, a detailed description of root system architecture (including root placement) and quantification of the distribution of the observed resources in the soil domain is required to analyze the interplay between root systems and their environment [4–6]. Because the vectorized root system is spatially registered on the rhizotron surface, the architectural information can be combined with any 2D description of soil resources. Here, we crossed the SmartRoot dataset with 2D maps of soil water content obtained with the light transmission imaging technique .
Depending on the size and complexity of the root system, this method can be time consuming and is not amenable even to moderate throughput. In our example, the image acquisition step (cleaning and scanning) took between 20 to 50 minutes per root system. The time required for root tracing with SmartRoot ranged from 1 to 4 hours. Nevertheless, the methodology proves to be a valuable tools to analyze complex root systems since, to the authors knowledge, no other methods today is able to extract such detailed dataset from entire root systems well beyond the seedling stage.
We have presented a new method based on a multiple scan approach to vectorize entire root systems grown in rhizotrons. This methods combines the strengths of two classical methods: in situ recording of root placement and ex situ high resolution scanning of root system fragments. The method yields ultimately a rich dataset containing detailed information on every root (position, morphology and topology) which can be easily crossed with spatial soil information data to analyze the interplay between the root system and local soil conditions. The method has been successfully used to vectorize root systems of 20-day old maize plants and has been used for the analysis of spatial root water uptake patterns.
Despite the time required by the method (both root scan and image analysis), we believe that it opens new perspectives for root-soil research. It proved an affordable way to precisely describe complex root system architecture and their interplay with their direct environment. We are currently using it for the parametrization of functional-structural plant models that simulate water and nutrient movement in the soil-root domain [21–25]. Using the digital structure created with the hybrid method, these models could be used to analyze in silico the water dynamics of the system.
The hybrid approach is neither bound to a specific acquisition device nor to a specific image analysis software. It provides a framework which should improve with future technical advances (e.g. faster scanners) and software developments (e.g. increased tracing automation). Following the root tracing protocol described by  the method could also be extended to the temporal analysis of root growth and its merging with relevant local soil conditions (e.g. soil water content and mechanical impedance).
Finally, the method presented here could be extended to generate precise estimation of topological and morphological traits of field-grown root systems, although some aspect of the root placement information would be lost.
Modified Hoagland solution
2M Ca(NO3)2 × 4H2O
1.5 × 2 = 3
2M MgSO4 × 7H2O
MnCl2 × 4H2O
ZnSO4 × 7H2O
H3MoO4 × H2O or
Na2MoO4 × 2H2O
(pH 6.0 w/ 3M KOH)
In situ tracing of the root axis
At the end of the growing period, before removal of the plant from the rhizotrons, the visible roots were manually traced on a transparent sheet (e.g. Avery 2503 Transparents JE 90 Microns, but any transparent sheet works) placed on the rhizotron surface (Figure 2A). Different roots orders were drawn using different colors (Stabilo OHPen universal, red and green) for an easier placement of the roots on the scanner (see below). We used light colors that do not appear on the final scan.
Root system preparation
After the tracing, rhizotrons were open and plants were taken out. Root systems were separated from the shoot and cleaned from the substrate (Figure 2B). Root cleaning was performed by soaking them during 5 minutes into water with a mild detergent. This procedure has the advantage of removing the majority of the soil particles without breaking the different roots. Finer soil particles still attached to the root were removed using a small painting brush. Plant were stored in a 50% ethanol solution before the scanning procedure.
Individual root scan
Every root image was analyzed using the root image analysis software SmartRoot  (Figure 2F). The software enables a semi-automated tracing of individual roots and generates morphological and topological information. In order to streamline the tracing of complete root systems, several new tools were implemented in SmartRoot which allow the user to perform actions simultaneously on multiple roots: Root list panel A new tool was implemented that displays the different roots as a hierarchical list (every root being nested in its parent node). This tool allows a fast observation of basic root statistics and an easy selection of multiple roots (even not adjacent in the image). Action on multiple roots Complementary to the multiple root selection tool in the root list panel, the possibility to perform actions on multiple roots was implemented. These actions, that were applied on single roots in previous versions, include the deletion of roots, their attachment to a common parent or the detection of lateral roots. Import multiple files A new import function was implemented to allow the import of multiple datafiles (e.g. those from the different root scans) into a single image. This tool enables the reconstruction of a complete root system.
Using these tools, the root tracing with SmartRoot is performed using the following procedure: (1) trace the roots on the different images, (2) import the different tracings into a single image, (3) connect the roots from the different images and (4) export the tracing to a single database. A screencast detailing the different steps in the root tracing procedure is available at the address: http://www.uclouvain.be/smartroot
Local analysis of root-soil interactions
The obtained time-series were analyzed by crossing the data contained in the light transmission images with the vectorized root system (Figure 9B and C). This was achieved by (1) reducing the image to 50 x 50 pixels so that every pixels had a size of 1 cm2, (2) finding, for every pixel, the closest root segment (euclidian distance) and (3) creating a new database merging the soil water content and root information. Root growth was assumed to be negligible during the three days of measurement.
This research was funded by a grant from the Fonds pour la formation à la Recherche dans l’Industrie et dans l’Agriculture to GL, the grant IAP7/29 from the Belgian Science Policy Office, the grant ARC-1116-036 from the Communauté française de Belgique to XD and the European Community’s Seventh Framework Programme under the grand agreement n°FP7-244374.
The authors would like to thanks the three anonymous reviewers whose comments greatly improved the original manuscript.
- Lynch JP: Root Architecture and Plant Productivity. Plant Physiol. 1995, 109: 7-PubMed CentralPubMedGoogle Scholar
- Hammer G, Dong Z, Mclean G, Doherty A, Messina C, Schussler J, Zinselmeier C, Paszkiewicz S, Cooper M: Can Changes in Canopy and/or Root System Architecture Explain Historical Maize Yield Trends in the U.S. Corn Belt?. Crop Sci. 2009, 49: 299-312. 10.2135/cropsci2008.03.0152.View ArticleGoogle Scholar
- Hodge A, Berta G, Doussan C, Merchan F, Crespi M: Plant root growth, architecture and function. Plant and Soil. 2009, 321: 153-187. 10.1007/s11104-009-9929-9.View ArticleGoogle Scholar
- Ge Z, Rubio G, Lynch JP: The importance of root gravitropism for inter-root competition and phosphorus acquisition efficiency: results from a geometric simulation model. Plant and Soil. 2000, 218: 159-171. 10.1023/A:1014987710937.View ArticlePubMedGoogle Scholar
- Draye X, Kim Y, Lobet G, Javaux M: Model-assisted integration of physiological and environmental constraints affecting the dynamic and spatial patterns of root water uptake from soils. J Exp Bot. 2010, 61 (8): 2145-2155. 10.1093/jxb/erq077.View ArticlePubMedGoogle Scholar
- Lobet G, Hachez C, Chaumont F, Javaux M, Draye X: Root water uptake and water flow in the soil-root domain. Plant Roots: The Hidden Half. Edited by: Eshel A, Beeckman T. 2013, New York: CRC PressGoogle Scholar
- Mooney SJ, Pridmore TP, Helliwell J, Bennett MJ: Developing X-ray Computed Tomography to non-invasively image 3-D root systems architecture in soil. Plant and Soil. 2011, 352 (1-2): 1-22.View ArticleGoogle Scholar
- Carminati A, Vetterlein D, Weller U, Vogel HJ, Oswald SE: When Roots Lose Contact. Vadose Zone J. 2009, 8 (3): 805-809. 10.2136/vzj2008.0147.View ArticleGoogle Scholar
- Jahnke S, Menzel MI, van Dusschoten D, Roeb GW, Bühler J, Minwuyelet S, Blümler P, Temperton VM, Hombach T, Streun M, Beer S, Khodaverdi M, Ziemons K, Coenen HH, Schurr U: Combined MRI-PET dissects dynamic changes in plant structures and functions. The Plant J. 2009, 59 (4): 634-644. 10.1111/j.1365-313X.2009.03888.x.View ArticlePubMedGoogle Scholar
- Downie H, Holden N, Otten W, Spiers AJ, Valentine TA, Dupuy L: Transparent Soil for Imaging the Rhizosphere. PLoS ONE. 2012, 7 (9): e44276-10.1371/journal.pone.0044276.PubMed CentralView ArticlePubMedGoogle Scholar
- Neumann G, George TS, Plassard C: Strategies and methods for studying the rhizosphere - the plant science toolbox. Plant and Soil. 2009, 321 (0): 431-456.View ArticleGoogle Scholar
- Garrigues E, Doussan C, Pierret A: Water uptake by plant roots: I - Formation and propagation of a water extraction front in mature root systems as evidenced by 2D light transmission imaging. Plant and Soil. 2006, 283: 83-98. 10.1007/s11104-004-7903-0.View ArticleGoogle Scholar
- Carminati A, Moradi AB, Vetteriein D, Vontobel P, Lehmann EH, Weller U, Vogel H, Oswald SE: Dynamics of soil water content in the rhizosphere. Plant and Soil. 2010, 332 (1-2): 163-176. 10.1007/s11104-010-0283-8.View ArticleGoogle Scholar
- Lobet G, Pagès L, Draye X: A Novel Image Analysis Toolbox Enabling Quantitative Analysis of Root System Architecture. Plant Physiol. 2011, 157: 29-39. 10.1104/pp.111.179895.PubMed CentralView ArticlePubMedGoogle Scholar
- Barthelemy D, Caraglio Y: Plant Architecture: A Dynamic, Multilevel and Comprehensive Approach to Plant Form, Structure and Ontogeny. Ann Bot. 2007, 99 (3): 375-407. 10.1093/aob/mcl260.PubMed CentralView ArticlePubMedGoogle Scholar
- Fitter AH: An Architectural Approach to the Comparative Ecology of Plant Root Systems. New Phytologist. 1987, 106: 61-77.View ArticleGoogle Scholar
- Lecompte F, Ozier-Lafontaine H, Pagès L: The relationships between static and dynamic variables in the description of root growth. Consequences for field interpretation of rooting variability. Plant and Soil. 2001, 236: 19-31. 10.1023/A:1011924529885.View ArticleGoogle Scholar
- Galkovskyi T, Mileyko Y, Bucksch A, Moore B, Symonova O, Price CA, Topp CN, Iyer-Pascuzzi AS, Zurek PR, Fang S, Harer J, Benfey PN, Weitz JS: GiA Roots: software for the high throughput analysis of plant root system architecture. BMC Plant Biol. 2012, 12: 116-10.1186/1471-2229-12-116.PubMed CentralView ArticlePubMedGoogle Scholar
- R Core Team: R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing Vienna. 2012, [http://www.R-project.org]Google Scholar
- Gardner WR: Dynamic Aspects of Soil-Water Availability to Plants. Ann Rev Plant Physiol. 1965, 16: 323-342. 10.1146/annurev.pp.16.060165.001543.View ArticleGoogle Scholar
- Postma JA, Lynch JP: Root Cortical Aerenchyma Enhances the Growth of Maize on Soils with Suboptimal Availability of Nitrogen, Phosphorus, and Potassium. Plant Physiol. 2011, 156 (3): 1190-1201. 10.1104/pp.111.175489.PubMed CentralView ArticlePubMedGoogle Scholar
- Javaux M, Schroeder T, Vanderborght J, Vereecken H: Use of a three-dimensional detailed modeling approach for predicting root water uptake. Vadose Zone J. 2008, 7 (3): 1079-1088. 10.2136/vzj2007.0115.View ArticleGoogle Scholar
- Schroeder T, Tang L, Javaux M, Vanderborght J, Koerfgen B, Vereecken H: A grid refinement approach for a three-dimensional soil-root water transfer model. Water Resour Res. 2009, 45 (10): W10412-Google Scholar
- Couvreur V, Vandenborg J, Javaux M: A simple three-dimensional macroscopic root water uptake. Hydrol Earth Syst Sci. 2012, 16: 2957-2971. 10.5194/hess-16-2957-2012.View ArticleGoogle Scholar
- Javot H, Lauvergeat V, Santoni V, Martin-Laurent F, Guclu J, Vinh J, Heyes J, Franck KI, Schaffner AR, Bouchez D, Maurel C: Role of a Single Aquaporin Isoform in Root Water Uptake. Plant Cell. 2003, 15 (2): 509-522. 10.1105/tpc.008888.PubMed CentralView ArticlePubMedGoogle Scholar
- Le Bot J, Serra V, Fabre J, Draye X, Adamowicz S, Pagès L: DART: a software to analyse root system architecture and development from captured images. Plant and Soil. 2010, 326 (1-2): 261-273. 10.1007/s11104-009-0005-2.View ArticleGoogle Scholar
- Lobet G, Draye X: Rhizotrons blueprints. figshare.http://dx.doi.org/10.6084/m9.figshare.97204,
- Busch J, Mendelssohn IA, Lorenzen B, Brix H, Miao S: A rhizotron to study root growth under flooded conditions tested with two wetland Cyperaceae. Flora. 2006, 201: 429-439. 10.1016/j.flora.2005.08.007.View ArticleGoogle Scholar
- Berntson GM, Woodward FI: The Root System Architecture and Development of Senecio vulgaris in Elevated CO2 and Drought. Functionnal Ecol. 1992, 6 (3): 324-333. 10.2307/2389524.View ArticleGoogle Scholar
- Tidwell VC, Glass RJ: X ray and visible light transmission for laboratory measurement of two-dimensional saturation fields in thin-slab systems. Water Resour Res. 1994, 30 (11): 2873-2882. 10.1029/94WR00953.View ArticleGoogle Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.