VZJ sign up for etocs
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published online 17 May 2007
Published in Vadose Zone J 6:269-281 (2007)
DOI: 10.2136/vzj2006.0067
© 2007 Soil Science Society of America
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (2) Citing Articles via ISI Web of Science (2)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Pierret, A.
Right arrow Articles by Pagès, L.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Pierret, A.
Right arrow Articles by Pagès, L.
GeoRef
Right arrow GeoRef Citation
Agricola
Right arrow Articles by Pierret, A.
Right arrow Articles by Pagès, L.
Related Collections
Right arrow Structure and Properties
Right arrow Root Growth/Water Uptake Models
Right arrow Root Development

SPECIAL SECTION: SOIL BIOPHYSICS

Root Functional Architecture: A Framework for Modeling the Interplay between Roots and Soil

Alain Pierreta,*, Claude Doussanb, Yvan Capowiezc, François Bastardiec and Loïc Pagèsd

a Inst. de Recherche pour le Developpement–Int. Water Management Inst.–Natl. Agric. and Forestry Res. Inst., c/o Ambassade de France BP 06, Vientiane, Lao PDR
b Inst. Natl. de la Recherche Agron.–Unité Climat, Sol et Environ., Domaine Saint Paul, Site Agroparc, 84914 Avignon Cedex 9, France
c Inst. Natl. de la Recherche Agron., Lab. de Toxicologie Environ., UMR INRA/UAPV, Domaine Saint Paul, Site Agroparc, 84914 Avignon Cedex 9, France
d Inst. Natl. de la Recherche Agron., Unité Production et Systèmes Horticoles, Domaine Saint Paul, Site Agroparc-84914 Avignon Cedex 9, France
* Corresponding author (alain.pierret{at}ird.fr).

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


Received 7 May 2006.



   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 The Complex Interplay between...
 Root Heterogeneity, Root Growth,...
 Modeling Root Functioning and...
 Conclusions
 REFERENCES
 
Soil ecosystems support a plethora of intertwined biophysical and biochemical processes. Soil structure plays a central role in the formation and maintenance of soil biological activity by providing a diversified habitat for soil organisms and determining the movement and transport of the resources on which they rely. At the same time, the formation and preservation of soil structure and fertility is also strongly linked to soil biological activity through feedback loops. In most soil ecosystems, soil biological activity and associated processes are concentrated in the soil located around living plant roots and influenced by root activity, an environment known as the rhizosphere. Consequently, among the wide array of soil life forms, plants play a dominant role in the regulation of many soil processes. In this paper, we illustrate the functional complexity of soil ecosystems using specific examples of root–soil interactions and associated processes. Through examples taken from the literature, we examine the origins and variations in soil physical, chemical, and biological properties and their impact on root growth. Next, we consider how the response of root systems to their environment affects resource acquisition by plants. Finally, we describe how the concept of root functional architecture can improve the integration of research advances from fields operating as independent disciplines and improve our understanding of soil ecosystems.

Abbreviations: EPS, extracellular polysaccharides.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
The Complex Interplay between...
 Root Heterogeneity, Root Growth,...
 Modeling Root Functioning and...
 Conclusions
 REFERENCES
 
In the current context of food production intensification, agronomists must concentrate on finding new solutions to increase crop productivity while minimizing water and nutrient losses and soil degradation (Passioura, 2006). The design of sustainable cropping systems can only be achieved if sufficient knowledge about the biophysical context(s) in which they are intended to be implemented is available. A key to this challenge is to better understand the intricacies of soil biological, chemical, and physical processes. In particular, improved knowledge about root–soil interactions may contribute to the design of practices that ensure optimized resource capture while providing leverage to minimize soil and water degradation, problems that increasingly plague most intensive cropping systems (Tilman et al., 2002).

Soil is a highly complex environment encompassing physical and chemical heterogeneity across a wide range of spatial and temporal scales. It bridges the mineral world with all the other trophic levels in the biosphere. Soil structure is central to such a fundamental linking role, as it provides the habitat for organisms and the pathway for essential resources on which they depend. In turn, soil biological activity impacts the formation and preservation of soil structure and fertility. Although the array of soil life forms is quite extensive—for example, bacteria, protozoa, fungi, nematodes, and macroinvertebrates—plants play a dominant role in the formation and maintenance of all other soil processes through root growth and functioning. Crucially, plants represent a major input of C to the soil: up to about a third of photosynthates allocated to roots can be lost to the soil as cap cells, mucilages, soluble exudates and lysates, and decaying tissues (Hawes et al., 2003; Hutsch et al., 2002; Nguyen, 2003). Because of roots' inherent nutritional value as a carbon substrate and the wide range of metabolites that they secrete into the soil (Rovira 1965), rhizosphere soil and root surfaces are also the main habitats for many soil organisms. Many aspects related to soil heterogeneity, rhizosphere processes, and root–soil interactions have been covered in three recent reviews (Doussan et al., 2003; Hinsinger et al., 2005; Gregory, 2006). As clearly outlined in these reviews, heterogeneity of soil physical processes and their variation in space and time have received much less attention than their biological and chemical counterparts.

In the first part of this paper, we review the main biological factors that influence soil physical and chemical heterogeneity from the micropore to the macropore scale. We show that separating soil physical processes from their chemical and biological analogs is somewhat artificial because the processes are almost systematically associated and in interaction with each other. Because of their indisputable importance regarding soil functioning, roots represent a natural entry point for the study of the functional complexity of soil ecosystems. The second part focuses on root functional heterogeneity to examine how roots explore the soil and adapt to the soil's inherent physicochemical heterogeneity. It is widely accepted that all roots have similar, if not identical, functional characteristics (Zobel, 2003). However, recent work by Zobel et al. (2006) showed that even within roots <1mm in diameter, several functional classes can be identified on the basis of their responses to environmental conditions. The concept of root functional architecture proposed by Zobel (2003) acknowledges this inherent complexity so that descriptions of roots and root systems integrate multiple genetically and anatomically determined functional root classes. In the third section, we examine how a modeling approach based on the concepts of functional architecture (Dunbabin et al., 2002a, 2002b; Doussan et al., 2006) has the potential to provide clearer insights into processes of soil exploration and utilization by roots. We also describe how these new developments in modeling open perspectives to: (i) quantify soil exploration by roots and root functioning at scales ranging from the individual root to the entire root system and (ii) study interacting soil physical, chemical, and biological processes.


   The Complex Interplay between Soil Physical, Chemical, and Biological Processes from the Micropore to the Macropore Scale
TOP
 ABSTRACT
 INTRODUCTION
 The Complex Interplay between...
Root Heterogeneity, Root Growth,...
 Modeling Root Functioning and...
 Conclusions
 REFERENCES
 
Physical and chemical heterogeneity is a common feature of most soils. Depending on soil mineralogy, the formation of zones of highly heterogeneous soil strength can result from purely physical processes such as cracking and swelling or freezing–thawing cycles. Human activity such as agricultural practices induces structural modifications that, although generally limited to the first 10 to 30 cm of the soil profile, are more rapid than natural processes and thus tend to have "traumatic" effects on soils (Whalley et al., 1995), such as subsoil compaction due to tillage, wheel traffic or trampling by cattle, formation of plow pans, slotting, and deep ripping. Physicochemical heterogeneity also occurs as local soil properties are progressively modified by pedogenesis (weathering and accumulation processes resulting in more or less differentiated soil horizons). Finally, an important part of soil physicochemical heterogeneity results from biological activity such as perforation, ingestion, and deposition. In the following paragraphs, we discussion some of the main biological factors influencing the formation of soil heterogeneity, from the micropore to the macropore scale.

Biological Activity and Soil Heterogeneity at the Micro- and Mesoscales
Soil biota range in size from microscopic (e.g., bacteria or endomycorrhizal hyphae) to centimetric (e.g., earthworms or ants). Hence, soil biological activity impacts soil heterogeneity at all scales, ranging from the basic arrangement of soil elementary particles—textural scale—to the macroscopic arrangement of aggregates, macropores, and soil layers—structural scale. Endomycorrhizal (or arbuscular mycorrhizal) hyphae are only about 12 to 15 µm in diameter (Staddon et al., 2003) and so do not significantly modify the physical arrangement of soil particles. However, an examination of field samples by scanning electron microscopy revealed that, because of their filamentous nature, fungal hyphae tend to tightly enmesh soil particles (Gupta and Germida, 1988). In addition, as they are covered with polysaccharide-rich mucilage, fungal hyphae can temporarily join together soil microaggregates, thus fostering the formation of stable macroaggregates (>0. 25 mm) (Tisdall and Oades, 1982; Tisdall, 1991).

Although difficult to quantify accurately, indexes of spatial correlation between bacterial densities, nutrient hotspots, and different pore-size classes have been reported by a number of authors (e.g., Gaillard et al., 1999; Nunan et al., 2003). Laboratory experiments conducted under controlled moisture conditions (Dorioz et al., 1993) showed that the microstructure of clay pastes is prone to modification by bacteria that induce polysaccharide-mediated aggregation of clay particles. In a series of experiments aimed at understanding the influence of soil matrix geometry on nitrogen mineralization and nitrification, Strong et al. (1998, 1999) investigated the relationships between pore-size class, microbial activity, and physicochemical properties of an Australian red earth. They found that organic N was concentrated in micropores less than 0.6 µm and in mesopores larger than 10 to 30 µm, but not in the intermediate pore-size class. They interpreted this finding to be the result of, on the one hand, protection from microbial decomposition in micropores and, on the other hand, the fact that moisture conditions are less frequently favorable to microbial activity in the bigger mesopores than in the medium-sized ones (hence the scarcer amounts of organic N in the latter compared with the former). According to the scenario proposed by Strong et al. (1998), as microbial colonies consume organic substrates from within smaller mesopores and excrete extracellular polysaccharides (EPS), mineral particles are rearranged, leading to increased mesoporosity at the expense of microporosity. Electron microscopy observations of field samples have further confirmed the role of microbial activity in the formation of mesoporosity: increases in microbial colony size by cell multiplication or by EPS secretion were reported to be consistently associated with the rearrangement of nearby clay minerals, to form compacted layers of overlapping clay platelets impregnated with EPS (Foster and Rovira, 1978; Foster et al., 1983; Foster, 1988; Chenu, 1993). With time, such microbially generated mesopores can be reclaimed as micropores: EPS-bound clay domains can be broken down by drying–wetting cycles, leading to the release of previously adsorbed organic compounds in the soil solution (Lund and Goksoyr, 1980). Finally, Strong et al. (1999) could link the micropore/mesopore balance with local redox processes related to microbial activity. They suggested that, under the anaerobic conditions that prevail in moist, microporous soil volumes, the reduction of metallic oxides (typically Mn or Fe oxides) is enhanced and relieves the pH stress that N mineralizing and nitrifying organisms would otherwise experience (as a result of H+ release following oxidation of C substrates). This example clearly highlights how interacting physical, chemical, and biological components of the soil induce the formation of microbial microsites and diffusion gradients that are important determinants for many soil functions.

Biological Activity and Soil Heterogeneity at the Macroscale
Mesofauna and macrofauna are present at a coarser scale. Mesofauna are mainly microarthropods (100 µm–2 mm) such as mites (Acari) and Collembola, which have no known impact on soil particle arrangement. They are confined to preexisting voids in the litter or soil and have negligible effects on soil heterogeneity (Lee and Foster, 1991). In contrast, soil macrofauna have major interactions with the soil. A few groups of larger soil invertebrates that are widely distributed and generally present in large numbers, namely, earthworms, termites, and ants (Lee and Foster, 1991), have the most significant effects on soil structure (provided that soil moisture is sufficient for these invertebrates to be active). Soil macrofauna have body sizes large enough to disrupt the physical makeup of most soils; for example, by burrowing, earthworms affect the transfer of water, air, and nutrients through the soil (Edwards et al., 1989; Bouma, 1991; McCoy et al., 1994; Li and Ghodrati, 1995). In general, the effect of termite and ant activity on soil structure is less extensive than that from earthworm activity: nest walls are consolidated by sticking together soil particles with excreta or salivary secretions, frequently forming massive cemented layers that locally reduce water infiltration (Lee and Foster, 1991).

The physical disruption induced by earthworm burrowing is accompanied by many biochemical modifications (Brown, 1995; Parkin and Berry, 1999; Tiunov and Scheu, 1999). Earthworms, for example, have a significant impact on the incorporation and distribution of organic matter in the soil (Shuster et al., 2001). They selectively activate mineralization and humification processes, hence promoting short and rapid cycling of nutrients and assimilable carbohydrates (Lavelle, 1988); it was reported that in some soils, 40% of all aerobic N2–fixing bacteria, 13% of anaerobic N2 fixers, and 16% of denitrifying bacteria were located in a thin layer lining earthworm burrows (Bhatnagar, 1975, cited in Anderson, 1988). As a result of their feeding activities, they produce casts that have higher cation exchange capacity, soluble carbohydrates, organic and mineral N, phosphatase and urease activity, and available P than soils from which they are derived (Satchell, 1983). Schrader et al. (1995) also noted a positive correlation between the organic C content of worm casts and their tensile strength and observed that worm casts have a higher structural stability than artificially constructed aggregates. As a consequence of the many local modifications they induce, therefore, earthworms play a central role in soil ecosystems and influence both directly and indirectly root distribution and growth (Volkmar, 1996; Lavelle, 1997).

Plant Roots and Soil Heterogeneity
Roots obviously alter soil physicochemical properties at the macroscopic scale in many ways. Growing root apices induce a reorientation of the soil particles and secrete EPS that locally bind soil particles (Cheshire, 1979; Tisdall and Oades, 1982; Dorioz et al., 1993). These two processes result in a general packing effect and the formation of macropores (Bruand et al., 1992, 1996; Jaillard and Callot, 1987). Field observations have confirmed that some deep-rooted perennial plant species can significantly alter soil macroporosity (Cresswell and Kirkegaard, 1995; Stirzaker et al., 1996; Lesturgez et al., 2004). Root water uptake induces gradients in soil water content, which, depending on soil texture and mineralogy, can lead to cracking (Lafolie et al., 1991; Bruckler et al., 1991) and also contribute to soil aggregation (Tri and Monnier, 1973). In association with the local rearrangement of soil particles that growing roots create in their immediate vicinity, there is also the development of a chemically and microbiologically differentiated environment, generally known as the rhizosphere (Darrah, 1993; Hinsinger, 1998). In some plants species, particularly graminates, this root-affected soil can take the form of rhizosheaths, which are physically bound to parts of the root system (McCully, 1995; Watt et al., 1994).


   Root Heterogeneity, Root Growth, and Resource Capture by Roots
TOP
 ABSTRACT
 INTRODUCTION
 The Complex Interplay between...
 Root Heterogeneity, Root Growth,...
Modeling Root Functioning and...
 Conclusions
 REFERENCES
 
As recently outlined by Hutchings and John (2004), most studies on root growth follow the premise that soil conditions are homogeneous, leaving serious gaps in our understanding of how plants function under natural and managed conditions and how they take advantage of patchy soil conditions. As illustrated above, soils are heterogeneous environments constantly reorganized by soil organisms and growing plants, at all scales from the micropore to the macropore. In this section, we discuss our current understanding of the mechanisms used by roots to grow and assimilate resources in such a heterogeneous environment.

Variations in Root Properties among Root Types and Along Individual Roots
Roots can be classified into several categories according to their ontogenesis and functions. Details about nomenclature used to describe root types and root system architectures can be found in Harper et al. (1991), Klepper (1992), Pagès et al. (1989, 2000), and Zobel (2005a, 2005b). Many reports clearly indicate that different root types play different functional roles. For example, in wheat (Triticum aestivum L.), leaf expansion is more severely reduced when drought affects seminal rather than nodal roots (Volkmar, 1997). Similarly, the contribution (in terms of resource acquisition) of the seminal root system to the whole plant exceeds what could be expected from its fractional mass (Waisel and Eshel, 2002). Navara (1987) showed that the radicle and seminal roots of maize (Zea mays L.) play a dominant role in supplying water during a significant part of the plant's lifespan, while the nodal root system seems to be more heavily involved in the uptake of resources such as phosphate (Mistrik and Mistrikova, 1995). In barley (Hordeum vulgare L.) root systems, although nitrate uptake rates decrease overall between the vegetative and reproductive stages, they tend to remain constant in the nodal root system (Mattson et al., 1993). Lazof et al. (1992) showed that nitrate uptake rates (per unit dry weight) of the primary axis of young maize plants was up to 68% of that of the lateral roots. Waisel and Eshel (1992) demonstrated variations in Cl and K uptake between the taproot and laterals in pea (Pisum sativum L.). Mature lateral roots of maize lowered the pH at the soil–root interface, whereas the parent root made it more alkaline (Marschner, 1990).

Important changes in physiological properties also occur along individual roots. Some of these changes relate to root ontogenesis. As root tissues get older, mature, and differentiate, their physiological status evolves. As a consequence, different uptake rates and root functions are observed, at increasing distances from the root tip (Clarkson, 1996). For example, high variations in root respiration were found along primary roots of peach (Prunus persica) (Bidel et al., 2000), not only in the vicinity of the apex but up to about 20 cm from root tips. Depending on nitrogen availability, parts of some roots can release protons and participate in the acidification of the immediate root environment, while others release hydroxyl ions (Jaillard et al., 2000). Nitrate and ammonium uptake were observed to vary along roots, with zones of active (generally in the apical region) and passive uptake (Cruz et al., 1995; Lazof et al., 1992). Variations in the uptake and translocation of other ions (e.g., P, K, Ca) along roots were also reported (Clarkson, 1996). In the field, cortical senescence in older root parts seems relatively common in cereals and other grasses (Robinson, 1991). Cortical senescence may weaken ion uptake because of physiological decay but also through disruption of soil-to-root transport pathways.

Root System Development and Architecture in situ
Studies throughout the twentieth century established that the overall architecture of root systems in situ (e.g., dominance of the main axis, branching pattern) is generally more complex and subject to great inter- and intraspecific variability (Cannon, 1949; Kutschera, 1960; Weaver, 1919) than that of roots grown under standard laboratory conditions (e.g., in agar). The respective importance of the primary and adventitious root systems—that is, the relative growth rates of main axes and laterals or the number of branching orders—varies across plant species. Different families of plant species that make up the vegetation in a given ecosystem are genetically programmed to occupy different niches and thus often use different soil exploration strategies. Two different soil exploration strategies, reflected by different root system architectures are illustrated in Fig. 1, which shows the typical rooting patterns of a perennial monocotyledon (Lolium multiflorum) and a perennial dicotyledon species (Achillea millefolium) (Kutschera 1960). L. multiflorum (Fig. 1A) develops a centralized adventice root system, often referred to as fasciculated system. Such a root system, in which main roots are continuously emitted from the plant base according to a species specific emission rate, is typical of grasses and other monocotyledons. On the other hand, A. millefolium (Fig. 1B) grows a noncentralized adventice root system, in which a network of rhizomes simultaneously emits branches and main roots. Such genetically controlled growth patterns are often modulated depending on the environmental conditions experienced by plants, leading Harper et al. (1991) to define the development and functioning of a given root system as an evolutionary response to the spatiotemporal variability of resource availability and the corresponding constraints to growth. Possible effects of such responses on root system architecture are illustrated in Fig. 2. Variable soil conditions experienced locally by plant roots trigger, within species-specific limits, a range of physiological responses that help the plant minimize the potential stress arising from soil heterogeneity and enable it to take advantage of "better-than-average" conditions (Drew, 1975; Robinson et al., 1999). Such plastic root responses to heterogeneous supplies of nutrients have been extensively reviewed by Hodge (2004, 2006). Plants have developed a range of complex strategies to exploit the soil's inherent patchiness, such as proliferation, segregation, aggregative root placement (Bartelheimer et al., 2006), or preemption of nutrient supply (Craine et al., 2005). Hence, root system development or expansion can be conceptualized as the allocation of assimilates to a population of individual root apices capable of independent, though coordinated, morphological and physiological responses to their immediate environment. Thus, to a large extent, the overall functioning of a root system actually corresponds to independent physiological activities coordinated at the whole root system level and varying axially along single roots in relation to their age.


Figure 1
View larger version (43K):
[in this window]
[in a new window]

  		FIG. 1. Comparison of the rooting patterns of (A) a perennial monocotyledon (Lolium multiflorum and (B) a perennial dicotyledon species (Achillea millefolium) (from Kutschera, 1960).

 

Figure 2
View larger version (57K):
[in this window]
[in a new window]

  		FIG. 2. Effects of localized (i) nutrient supply and (ii) physical constraint on the root system architecture of a monocotyledon and a dicotyledon. Root system architecture of a barley plant (Hordeum vulgare cv. Proctor) (A) uniformly supplied with nitrate or (B) supplied with nitrate through a banded treatment. The banded treatment triggered root proliferation in the zone of nitrate supply (Drew, 1975). Comparison of the architectures of two Lupinus angustifolius root systems: (C) physically unconstrained growth conditions, and (D) taproot growth stopped by a physical obstacle at an early developmental stage.

 
The presence of zones of high mechanical resistance is one of the most common physical limitations to soil exploration by roots (Hoad et al., 1992). In cultivated soils, the location, lateral extension, and thickness of zones of high resistance to penetration vary during the growing season (Castrignano et al., 2002). In soil volumes of higher strength, the development of soil structure is of paramount importance to root penetration (Tardieu and Manichon, 1986; Tardieu and Katerji, 1991; Fig. 3A); increases in soil strength reduce root elongation and alter root diameters and the average number of laterals on primary axes (Bennie, 1996; Dexter, 1987; Fig. 3B). In soils that impede root growth (e.g., because of high resistance to penetration), successive generations of roots tend to reuse paths of least mechanical resistance, such as preexisting structural features like cracks, biopores, or soil casts excreted by soil macrofauna (Rasse and Smucker, 1998). This colocation of roots and macropores (McKenzie et al., 1995; Volkmar, 1996; Stewart et al., 1999) leads to the formation of a specific environment that differs significantly chemically and biologically from the bulk soil (Pierret et al., 1999; Pankhurst et al., 2002).


Figure 3
View larger version (21K):
[in this window]
[in a new window]

  		FIG. 3. (A) Root impact map illustrating soil exploration by roots in compacted soil horizons in which cracks represent paths of least resistance preferentially explored by roots (Tardieu and Katerji, 1991). (B) A model for relative root elongation rate as a function of matric potential, at different levels of soil strength Qp (MPa), measured by a penetrometer (Dexter, 1987): as soil strength increases and soil is drier, relative root elongation (R/Rmax) decreases.

 
Soil Exploration vs. Resource Acquisition
Because, as described above, root systems are not uniformly and constantly active, soil exploration by plant roots is not a reliable indicator of soil resource exploitation. It has been clearly demonstrated that, if homogeneous root behavior is assumed to model water and nitrate uptake rates, predicted values at the entire root system level are substantially overestimated. On the basis of such an exercise, Robinson (1991) inferred that, on average, only 10 and 30% of the total root length of a given root system is effectively involved in nitrate and water uptake, respectively. Thus, to understand resource acquisition by plant roots, it is essential to determine (i) the spatial pattern of root activity, that is, where root activity occurs at any given time—for example, are some root order(s) or specific region(s) of the root system preferentially involved?—(ii) how the spatial distribution of root activity varies with time, and (iii) the influence of environmental conditions on this pattern. From the point of view of resource acquisition, a root system must be regarded as a population of individual roots behaving (i) differently and independently from each other (Waisel and Eshel, 1992) (although coordinated to some degree at the root system level), (ii) as a function of tissue differentiation, and (iii) in response to changing environmental conditions (plasticity).

Root System Plasticity and Uptake Optimization
Roots probably evolved plastic responses to their environment as they differentiated as specialized tissues throughout geological times (Raven and Edwards, 2001), optimized to explore and utilize resources in heterogeneous soils (Leyser and Fitter, 1998). Root plasticity is also a response to intra- and interspecific competition. Robinson (2001), for example, showed that plastic root responses are triggered by intraspecific competition in a wheat monoculture but do not necessarily lead to greater uptake rates. Nutrient availability is known to influence many facets of root system morphology (Ford and Lorenzo, 2001): root branching, root growth (with growth of main axes generally less affected by nutritional effects than higher order axes), root diameter, root angle (e.g., low P availability decreases the angle of emission of basal roots in bean [Phaseolus vulgaris L.], soybean [Glycine max (L.) Merr.], and pea [Liao et al., 2001]), root hair length and density, and production of specific root types (cluster roots [Skene, 2000] or drought-induced roots [Vartanian, 1996]). The response of plants to variations in the location of nutrients has been well studied (see review by Robinson, 1994) compared to the influence of temporal variations in nutrient concentrations on root plasticity. Experimental observations of root responses to variations in the spatiotemporal availability of nutrients have generally been made under conditions wherein access to nutrients was artificially reduced. For example, a classic experimental design consists of providing nutrients to a small portion of the root system only, while the rest of it grows in nutrient-poor or sterile soil (Drew and Saker, 1975). Roots respond to such a heterogeneous system in two ways (Robinson, 1996): (i) the nutrient inflow rate increases but then returns to normal within hours, or (ii) roots proliferate toward and within the nutrient-rich patch over a period of several days, while root growth in the rest of the root system is inhibited. These trends vary depending on the plant species, with the induced increases in root growth and nutrient uptake varying over one order of magnitude or with a total lack of response in some species (Robinson, 1996). The stimulation in uptake rate seems to be sensitive to the nutrient considered and the duration of the starvation period. Root proliferation appears less dependent on the nutrient considered (except for K in some species). Localized responses are generally assumed to be caused by direct nutritional benefits to the roots directly exposed to nutrient patches, but there is some evidence that they can also involve indirect, sophisticated mechanisms. Zhang et al. (1999), for example, proposed a dual pathway for NO3– in Arabidopsis thaliana in which the NO3– ion acts as a signal rather than a nutrient, and root branching is modulated by opposing signals from the plant's internal N status and the external supply of NO3–.


   Modeling Root Functioning and Soil Exploration by Roots
TOP
 ABSTRACT
 INTRODUCTION
 The Complex Interplay between...
 Root Heterogeneity, Root Growth,...
 Modeling Root Functioning and...
Conclusions
 REFERENCES
 
In most crop models, water and nutrient uptake are predicted on the basis of synthetic descriptors such as the root density (e.g., length, biomass, or surface area per unit soil volume). Such descriptors are indicative of soil exploration by roots if it can be assumed that roots are regularly distributed in the soil. However, under field conditions, the assumption of a regular distribution of roots does not hold (Tardieu, 1988), and root distribution within the soil has a strong influence on resource acquisition by plants (Lynch and Nielsen, 1996; Pagès, 2002; Pagès et al., 2000). Consequently, summary parameters such as root density are not sufficient to investigate the detailed development and functioning of root systems; it is necessary to include particulars about root architecture and growth dynamics in models to gain sharper insights into soil exploration and utilization processes. Because they include explicit quantitative information about the soil volume that a given root system accesses and influences, as well as about the location and number of roots, models of root system architecture provide a unique opportunity to understand soil exploration by roots. Conceptually, models of root system architecture consist of three-dimensional sets of connected axes or segments, each characterized by properties such as diameter, water, or nutrient-uptake ability. Existing models of root system architecture include variable degrees of dynamic complexity, which have been extensively reviewed by Doussan et al. (2003). Explicit models of root system architecture are also valuable tools for including effects of heterogeneous soil conditions on root growth at the scale of the individual root segment through to the whole root system.

Using Root Architecture Models to Assess the Interactions between Roots and Their Chemical Environment
Using root architectural modeling, Ge et al. (2000) studied the effect of altered gravitropism of the basal roots of bean plants (the position of which varied from shallow to deep) to study their importance in P acquisition efficiency. The authors considered both a homogeneous P distribution and a stratified distribution with high P concentrations in the top 10 cm of the soil profile. In both cases, shallower root systems explored more soil (per unit root biomass) than deeper systems because of reduced interroot competition (i.e., the overlap of depletion zones corresponding to neighbor roots was reduced in shallow root systems; Fig. 4).


Figure 4
View larger version (25K):
[in this window]
[in a new window]

  		FIG. 4. Simulation of the influence of different degrees of basal root gravitropism on the exploitation of P by bean (Phaseolus vulgaris L.) root systems. The depletion zone of P is represented by diffusion of P to the root with time (Diffusion coefficient 10–8 cm2 s–1). (A) Bean root systems simulated with different rooting patterns (shallow; Carioca, a cultivar, and deep). (B) Volume of the overlapping exploited zones for the three root system types. (C) P uptake by the three simulated root systems at the end of simulation (320 h), in the case of a stratified soil profile of P (P concentration is higher in the first 20 cm of soil) (from Ge et al., 2000).

 
Somma et al. (1998) and Dunbabin et al. (2002b) also incorporated aspects of the effects of nutrient availability on root system development in their root architectural models. In Somma et al.'s (1998) model, the effect of nitrate on root growth was implemented via a linear impedance function, which mimics the fact that root growth remains unaffected by nutrient concentrations as long as they fall within ranges that are both plant species and nutrient specific. For each growth step, an actual elongation rate is computed for each individual root apex, based on an unimpeded elongation (function of available photosynthetic assimilates) scaled according to temperature, soil strength, and soil nutrient concentration impedance factors. Figure 5 shows an example output from this model: the simulated root system of a 25-d-old barley plant grown with water and NO3– supplied through drippers located on the soil surface. NO3– was either applied continuously (Fig. 5A) or for a finite time at the beginning of the simulation (Fig. 5B). The total amount of applied N was the same in the two cases. In the first case, simulations showed that N concentrations remained higher in the upper part of the soil and root density decreases with depth. In the second case, the NO3– plume moved downward following the application and caused a greater root density in the central part of the soil. Interestingly, peaks in root length density and NO3– concentration did not coincide, a feature linked to the relative rates of root growth and the downward percolation of NO3–.


Figure 5
View larger version (31K):
[in this window]
[in a new window]

  		FIG. 5. Simulated three-dimensional root architecture (coupled with water and nitrate transfer and uptake by the root system) with corresponding root density and nitrate concentration distribution for (A) continuous supply of nitrogen by drippers and (B) the same amount of nitrogen, but supplied at the beginning of the simulation period (from Somma et al., 1998).

 
More recently, Dunbabin et al. (2002b) encapsulated a more subtle description of root system plasticity into the root architecture ROOTMAP, initially developed by Art Diggle (1988) in Western Australia. This model's fundamental principle is to combine, at the whole plant scale, the demand for individual resources and, at the local scale, the ability of the various components of the root system to supply resources, thus driving the allocation of assimilates to the most rewarding parts of the growing root system. Depending on the soil conditions defined at the onset of numerical experiments (i.e., runs of the model aimed at testing scenarios consisting of different N and water-supply patterns), the architecture and uptake efficiency of the root systems produced by the model resulted, at least in part, from the environmental conditions roots experienced throughout growth. Hence, this model simulates both a local "sensing" response and a whole root system response. Inflow and root proliferation plasticity are features that can be modeled with this approach. The authors tested their model's performance against laboratory and field experiments with narrowleaf lupin (Lupinus angustifolius), using nitrate as an example nutrient (Dunbabin et al., 2002a, 2002b). Nitrate was supplied to the plants every second day, according (i) to a static supply pattern (same random distribution of nutrient patches along the soil profile for every successive application) or (ii) to a dynamic supply (new random distribution of nutrient patches with every successive application). Figure 6 shows the results yielded by this model for two extreme root system topologies (herringbone and dichotomous; Dunbabin et al., 2001). In the case of static nitrate supply, because of root plasticity (both morphological and functional), the dichotomous system is more efficient than the herringbone one. In the case of dynamic N supply, the herringbone system appears to be more efficient than the dichotomous one, and the latter gains almost no efficiency in uptake from plasticity.


Figure 6
View larger version (27K):
[in this window]
[in a new window]

  		FIG. 6. Simulation of nitrate uptake efficiency with an architecture model taking into account both inflow and morphological plasticity of the root system. Nitrate is distributed in the soil as small patches. The efficiency of uptake with plasticity is relative to the same root system with no plasticity response. Root systems are (A) herringbone system and (B) dichotomous system. In the dynamic supply case, the nutrient patches are randomly redistributed in space, which is not the case for static supply (from Dunbabin et al., 2001).

 
Using Root Architecture Models to Assess the Interactions between Roots and Their Physical Environment
Several explicit models of root architecture incorporated the influence of soil temperature on root growth or root appearance, using a thermal time scale (Diggle, 1988; Pagès et al., 1989) or a reduction coefficient that reduces root growth rates (Clausnitzer and Hopmans, 1994). The effect of soil strength has also been included, by means, generally, of indirect variables such as soil bulk density or water content (Clausnitzer and Hopmans, 1994; Pagès, 1999; Fig. 7A and 7B) combined with empirical functions that reduce optimal growth rates and alter root growth direction. To test the influence of hydrotropism on root growth in slopes, Tsutsumi et al. (2003) used an explicit model of root architecture that included a sensing mechanism of water flux gradients near root tips to modulate root bending.


Figure 7
View larger version (23K):
[in this window]
[in a new window]

  		FIG. 7. Simulation of maize (Zea mays L.) root system architecture interacting with the environment. A plow pan layer impedes root growth at 35 cm depth. (A) General morphology of the simulated maize plant. (B) Simulated (+) and observed (·) root profiles, obtained by counting the number of colonized cells (2 x 2 cm) on vertical grids. The horizontal bar represents one standard deviation (from Pagès, 1999). (C) A simple example of simulated root growth around mechanical obstacles (rocks) in a homogeneous soil (from Prusinkiewicz, 1998).

 
Currently, the modeling of the interactions between growing roots and their physical environment remains very basic. Prusinkiewicz (1998) presented a modified version of Diggle's (1988) ROOTMAP, which included root responses to mechanical obstacles (rocks) in the soil (Fig. 7C). Recently, our research group initiated a new project aimed at testing the influence of soil structure on root growth and water uptake. Technically, this modeling exercise is based on coupling a simplified model of soil structure with a modified version of the model of root hydraulic architecture developed by Doussan et al. (1998a). The simplified model of soil structure consists of a 1.5-m3 volume, with a 5-cm seed-bed, a 25-cm-thick tilled layer, a 3-cm plow layer and a ~1.2-m-deep subsoil. Soil structure in the tilled layer is described as a distribution of dense clods embedded in a looser matrix. Clod shape and size distributions were simulated on the basis of field observations made in northern France (Desbourdes-Coutadeur, 2002). For the subsoil, a soil density gradient combined with a macropore network was used to represent soil structure. Objects considered as macropores were generated using the model developed by Capowiez and Bastardie (Bastardie et al., 2002) to describe earthworm burrowing behavior. Burrows 2- to 3- and 5-mm in diameter, respectively, were simulated, corresponding to a mixed population of endogeic (Aporrectodea caliginosa) and anecic (Lumbricus terrestris) worms, with individual densities corresponding to field observations made in northeastern France. During the root growth simulation period, a local soil impedance factor was computed for each cell of the structured soil volume by combining the local soil bulk density with the local soil water content (deduced from an initial soil water profile, which was altered at each time step to mimic soil drying). This local soil impedance factor was used to modulate root elongation depending on local soil moisture and bulk density conditions (from 1, unimpeded growth, to 0, stalled apical growth). At this stage, highly simplified rules have been used regarding root response to the presence of soil structural features. First, if a macropore is present within the voxel in which a given root tip is entering, following the elongation corresponding to a simulated time step, the root tip continues its growth inside the macropore if a randomly generated number is higher than an arbitrarily set threshold. Otherwise, it is assumed that contact between root tip and macropore did not occur. Second, once inside a macropore, a root tip is forced to follow the whole extension of the macropore before it can grow back in the soil matrix. Third, root elongation remains totally unimpeded as long as the root tip remains "trapped" in the macropore. At present, rules regarding the alteration of branching patterns in response to local impedance to root growth have not yet been added to the model. Even though simplified, this model opens new avenues for understanding the effect of soil structure and soil structure manipulation on root growth and functioning. Currently, it is possible to generate maize root systems whose architecture is clearly altered by the presence of soil structural features (Fig. 8). In the near future, we hope to use this model to assess the effect of different degrees of soil structural constraints to root growth on root water uptake.


Figure 8
View larger version (33K):
[in this window]
[in a new window]

  		FIG. 8. Modeling of the interactions between roots and soil structure. (A) Comparison between two 100-d-old maize (Zea mays L.) root systems, the first one (left-hand side) grown in a homogeneous soil volume and the second (right-hand side) grown in a structured soil consisting of a 25-cm-thick tilled layer with distributed dense clods, a 3-cm plow layer and a ~1.2-m-deep subsoil with biopores (earthworm burrows). In the case of the structured soil, the interactions between growing roots and soil structure have led to reduced rooting depth and lateral expansion of the root system. This is largely due to the trapping of roots in macropores at certain soil depths (50–55 cm in particular), as shown by (B) the high occurrence of root-to-macropore distances less than the voxel size (1 cm) .

 
Encapsulating Root Functional Heterogeneity into Root Models
The example of root water uptake illustrates well how root functional heterogeneity can be taken into account using root architectural models. As relatively impermeable structures differentiate away from the root tip (suberization), root water uptake is increasingly impeded along the radial pathway (which concerns water transport from the soil to xylem vessels). Symmetrically, with the presence of increasingly opened xylem vessels away from the root tip, axial water transport to the stem is facilitated. In maize main axes, late metaxylem vessels (i.e., xylem of high water-carrying capacity) are only fully open at distances up to 20 to 30 cm from the apex (Wenzel et al., 1989). On the basis of experimental measurements of the axial and radial hydraulic conductance of maize roots (Varney and Canny, 1993), Doussan et al. (1998a, 1998b) were able to model the spatial variability of root hydraulic conductance within the root system of maize. They showed that this spatial variability led to the formation of a heterogeneous water uptake pattern, even when soil water is readily and evenly available to all roots (Fig. 9). A different water uptake pattern was found for the perennial root system of Prunus (Doussan et al., 1999), indicating that genetic differences influence root water uptake heterogeneity.


Figure 9
View larger version (86K):
[in this window]
[in a new window]

  		FIG. 9. Distribution of water uptake fluxes within a simulated maize (Zea mays L.) root system. Water uptake is simulated by taking into account the variability of the root hydraulic conductance in the root system (from Doussan et al., 1999).

 
Recently, Doussan et al. (2006) developed their root architectural model further, coupling the effect of local soil and root hydraulic properties with the formation and evolution of root water uptake patterns, at the scale of the entire root system. This new model provides information about root system functional architecture and hydraulic continuity between plant and soil. Both experiments (using light and X-ray transmission imaging of root water uptake [Garrigues et al., 2006]) and modeling concurred, showing that as water is extracted from the growth medium by the plants (L. angustifolius), a water uptake front forms and moves downward along the root system (as soil dries) (Fig. 10). This uptake front's spatial extension and displacement along roots was closely related to local root and soil hydraulic properties. In particular, the water retention properties of the growth medium strongly influenced the characteristics of the front: a sharp front formed in a dominantly sandy medium, whereas, in a sandy-clay loam, the front's shape was highly attenuated. Comparisons between tap-rooted and fibrous root system architectures grown in a sandy medium showed that the tap-rooted architecture induced a more spatially concentrated uptake zone (near the soil surface) with higher flux rates, but with a xylem water potential at the base of the root system twice as low as in the fibrous architecture. Modeling provided evidence that hydraulic lift can occur when transpiration declines at night, particularly in a growth medium prone to abrupt variations in soil water potential (sand). Overall, this new way of modeling soil-to-root water transfer, demonstrates that the concept of root functional architecture is valuable for studying water uptake in relation to both plant and soil heterogeneity. We envisage extensions of this approach to analyze root uptake, the distribution of root hydraulic conductance, or the influence of heterogeneous conditions (localized irrigation, root clumping), depending on genetically selected root architectural traits.


Figure 10
View larger version (57K):
[in this window]
[in a new window]

  		FIG. 10. Simulation of the propagation of a water uptake front across the root system of a 50-d-old narrowleaf lupin (Lupinus angustifolius) with a fibrous root system, growing in a sandy rhizotron. Length scale in centimeters. Distribution of calculated water uptake within the root system 1.5, 5, 7, 9, and 11.5 h after the beginning of an uptake experiment. The rates are expressed as flux density (i.e., volumetric flow rate normalized to the root surface area [cm3 cm–2 s–1]). The red lines show the downwardly moving zone of active water uptake. The green color shows negative flux rates, that is, water exsorption by roots (from Doussan et al., 2006).

 

   Conclusions
TOP
 ABSTRACT
 INTRODUCTION
 The Complex Interplay between...
 Root Heterogeneity, Root Growth,...
 Modeling Root Functioning and...
 Conclusions
REFERENCES
 
Models based on the concept of root functional architecture provide a unifying framework for integrating root and soil heterogeneity and their complex interaction. Such a modeling tool represents a unique opportunity to unify research advances from fields that operate as independent disciplines (e.g., results obtained using novel observation techniques such as nondestructive and cryoscanning imaging of roots under field conditions [Pierret et al., 2003; McCully, 1999], microsensors [Portefield, 2002], and root-pressure probes [Steudle, 2000]). It can be applied to the analysis of root water and nutrient uptake as a function of root architectural traits (genetically selected), distribution of root properties within the root system (hydraulic conductivity, nutrient-uptake ability), or heterogeneous environmental conditions (e.g., localized water and/or nutrient availability). Models of root functional architecture could also prove useful for crop improvement as they can be used to derive robust biophysical indexes characteristic of some crop–environmental combinations, such as improved root sink terms for water uptake modeling. The many processes that can be investigated using the modeling of root functional architecture include

To date, the majority of studies on roots have been conducted based on false premises of homogeneous soil conditions, leaving serious gaps in our understanding of plant functioning under field conditions (Hutchings and John, 2004). Much remains to be understood about "how real roots work" (McCully, 1995). In the current global context of food production intensification, additional knowledge about the interplay between soil biochemical processes and soil mineral constituents, such as root–soil interactions, is urgently needed to increase crop productivity while minimizing water and nutrient losses and soil degradation. We believe that models of root functional architecture will play a key role in the design and testing of sustainable cropping systems.


   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 The Complex Interplay between...
 Root Heterogeneity, Root Growth,...
 Modeling Root Functioning and...
 Conclusions
 REFERENCES