ECOLOGICAL BASIS OF AGROFORESTRY - CHAPTER 13 potx

13 Root Competition for Water between Trees and Grass in a Silvopastoral Plot of 10 Year Old Prunus avium Philippe Balandier, François-Xavier de Montard, and Thomas Curt CONTENTS 13.1 Introduction 253 13.2 Materials and Methods 255 13.2.1 Experimental Plot 255 13.2.2 Climate and Soil 255 13.2.3 Experimental Design 255 13.2.4 Measurements 256 13.2.4.1 Tree Dimension 256 13.2.4.2 Tree and Grass Water Status 256 13.2.4.3 Grass and Tree Root Growth 256 13.2.4.4 Soil Water Content 256 13.2.4.5 Data Analysis 257 13.3 Results 257 13.3.1 Aboveground Tree Growth 257 13.3.2 Soil Water Content 257 13.3.3 Water Status of Trees 260 13.3.4 Root Growth 262 13.4 Discussion 264 13.5 Conclusion 267 Acknowledgments 267 References 267 13.1 INTRODUCTION In temperate Europe, fast-growing broad-leaved trees such as wild cherry (Prunus avium L.) supply highly valued wood with a veneer end use. The wild cherry tree has a high light requirement (Ruchaud, 1995), which makes it a species potentially well adapted for agroforestry purposes where trees are planted with very wide spacing to allow intercropping or grazing (Balandier and Dupraz, 1999). Cattle or sheep maintain grass and shrubs at low height and add an income from animal products for the owner. With the help of tree pruning (Balandier, 1997), such a silvopastoral system has proved efficient in produci ng straight knot-free quality boles (Balandier et al., 2002). Agroforestry practice requires that the biological and physical relationships between the differ- ent components of the system (for instance tree and crop or pasture) generate a favorable balance Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 253 9.10.2007 9:35am Compositor Name: VAmoudavally 253 Copyright 2008 by Taylor and Francis Group, LLC between negative and positive interactions (Anderson and Sinclair, 1993). In other words, the trees must utilize resources that the crop does not (Cannell et al., 1996) and vice versa. This is also called the niche theory: two or more species must use resources differently if they are to coexist on a site (Kelty, 1992). However, though often postulated, such a relationship has seldom been demon- strated, particularly as regards interactions at the root level (Kelty, 1992), namely competition for water, nutrients uptake, allelopathy, etc. The wild cherry tree is a species known to be sensitive to intra- and interspecies competition (i.e., other trees, shrubs, or grass; Collet et al., 1993), which adversely affect its growth and the quality of its wood (Le Goff et al., 1995). Therefore, when it is associated with a crop or pasture in agroforestry, the question arises of whet her the balance of interactions will be positive. The basic mechanisms that lead to growth impairment of wild cherry in competition with grass or shrubs are not fully known (Lucot, 1997). Most studies have been indirect: the elimination of weeds or shrubs around trees has a positive effect on growth in height and especially in diameter or biomass for young trees (Monchaux, 1979; Frochot and Lévy, 1980; Britt et al., 1991; Collet and Frochot, 1992; Campbell et al., 1994; Le Goff et al., 1995; Balandier et al., 1997; Cain, 1997; Davis et al., 1999) and a positive effect on root growth (Larson and Schubert, 1969). Interactions between trees and weeds or shrubs, although demonstrated practically, need to be more fully described in terms of specific processes to form a basis for improving tree management (Nambiar and Sands, 1993). Some functional physiological studies have been conducted on very young trees, but often in containers or not in natural conditions (Collet et al., 1996; Jäderlund et al., 1997; Johnson et al., 1998; Mohammed et al., 1998). For instance, the leaf water potential of trees in association is often more negative than that of trees in bare soil (e.g., Juglans regia L. with Trifolium subterraneum L., Pisanelli et al., 1997; Pinus strobus L. with Populus tremuloides Michx., Boucher et al., 1998; Quercus robur L. and Fagus sylvatica L. with natural herbaceous vegetation, Löf, 2000). Tree transpiration, leaf CO 2 assimilation, and leaf conductance can also be altered by herbaceous competition (Pinus radiata D. Don with Dactylis glomerata L., Miller et al., 1998; J. regia with Lolium perenne L., Picon-Cochard et al., 2001). Girardin (1994) concluded from a study on 4 year old wild cherry trees that as this species has a very shallow root system, it suffers badly from competition by grass. However, the study was indirect and the true depth of the tree root system was not measured directly. Even so, all the studies conducted suggest that trees do suffer from such competition, to different extents depending on the competing species (Nambiar and Sands, 1993; Mil ler et al., 1998; Dupraz et al., 1999; Coll et al., 2003) and that this competition can reduce their growth and sometimes prevent their establishment. Allelopathy, the release of toxic chemicals in the environment by a plant or a tree is other possible negative interference, which can reduce either tree growth or grass production. In agroforestry systems, some trees were characterized as probably having an allelopathic inhibitory effect (e.g., Juglans sp., Eucalyptus sp., Gallet and Pellissier, 2002). Many grasses were also reported to have such similar effects (Qasem and Foy, 2001). However, nothing is mentioned on a potential allelo- pathic effect of the wild cherry tree or the main herbaceous species composing the pasture (see Section 13.2) in the study reported here (Qasem and Foy, 2001), except perhaps for Holcus lanatus. Much work has been done on competition between trees and grass in agroforestry systems with pine (e.g., Nambiar and Sands, 1993; Yunu sa et al., 1995, for P. radiata) and Eucalyptus (Eastham and Rose, 1990 for Eucalyptus grandis Maiden) and for warm climates (Scholes and Archer, 1997; Balandier, 2002). However, the literature is much more scant for temperate climates and broad- leaved species such as wild cherry. Here we report on interactions at the root level between trees and grass in a temperate silvopastoral system with 10 year old broad-leaved wild cherry trees in natural conditions. Compe- tition for light and for nitrogen in such a system has already been reported (De Montard et al., 1999; Méloni, 1999). Nutrients other than nitrogen are present in the soil in supraoptimal values and competition for them was unlikely. Therefore, this chapte r focuses on interactions for water. We studied not only the aerial growth of the tree but also its water status, its root growth through direct Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 254 9.10.2007 9:35am Compositor Name: VAmoudavally 254 Ecological Basis of Agroforestry Copyright 2008 by Taylor and Francis Group, LLC measurement, time course of volume soil water content, and the interactions between these different parameters to understand more fully and so better manage the water competition relationships between tree and grass. 13.2 MATERIALS AND METHODS 13.2.1 E XPERIMENTAL PLOT The experiments took place in a silvopastoral plot of 2.9 ha in Auvergne, Central France (approxi- mate latitude 468N and longitude 38E), in hilly country, at an elevation of 810 m a.s.l. The plot slopes moderately (from about 8% – 15%). Two year old wild cherry trees (Prunus avium L.) were planted directly with minimum tillage of the soil in March 1989 at 200 stems ha À1 (6 3 8m)ona permanent pasture grazed by sheep. For practical reasons during the experiment—from 1997 to 1999—the sheeps were kept out of the experiment al plot (about 1000 m 2 ) and the pasture was regularly cut by hand to simulate sheep browsing. The main species of the pasture were orchard grass (D. glomerata L.), hairy oat grass (Avena pubescens Huds.), yellow oat-grass (Trisetum flavescens [L.] P. Beauv.), velvet grass (H. lanatus L.), Erect Brome (Bromus erectus Huds.), red fescue (Festuca rubra L.), white clover (Trifolium repens L.), Bush vetch (Vicia sepium L.), common yarrow (Achillea millefolium L.), and Germander speedwell (Veronica chamaedrys L.). Trees were weeded with glyphosate (3.6 g L À1 ) during the first 4 years after planting (i.e., from 1989 to 1992) within a radius of 0.6 m around their trunk to ensure firm rooting. None of the trees in this study were pruned. 13.2.2 CLIMATE AND SOIL Average annual rainfall was 835 mm, fairly evenly distributed throughout the year but sometimes with pronounced drought periods (e.g., about 15 March–07 May, 4–11 June, 18–25 June, 15–22 July, and 30 July–20 August in 1997, 10 May–09 June and 09–30 July in 1998, and 30 May–09 July in 1999). The mean annual temperature was about 98C. The soil was a slightly acid granitic brown soil (brunisolic order—Orthic B, Canadian soil classification 1998; pH water ¼ 5.8, the organic matter ranged from about 65 g kg À1 in the upper soil layer to 6 g kg À1 in depth which corresponds to a moderately fertile soil) topped by a thin basaltic colluvium, and soil depth reached up to 180 cm. On average, the first layer (about 0–15 cm) of the soil displayed a sandy-silt texture with a micro- lumpy structure. The proportion of coarse elements (i.e., >2 mm) was about 10%. The compactness was low. The second layer (15–40 cm) had the same texture (sandy-silt) but was more compact with a high density and coarser elements (40%); the structure was heavier. The next two layers had a silty-sand texture with a heavy structure and a high proportion of coarse elements (60%–70%). Taking into account the proportion of the coarse elements, the calculated total available water content of the soil (Baize and Jabiol, 1995) to a depth of 120 cm deep was about 85 mm. Among the different trees, there were some small differences in soil layer depth and compactness. Wherever possible, we tried to take into account these small variations when analyzing growth data. For each layer of soil, the soil wat er content corresponding to the wilting point (pF of 4.2 or 16 atm., i.e., by convention, the soil potential over which plant roots cannot extract water) was assessed after establishing curves of ‘‘soil potential–soil water content’’ (Lucot, 1997); on average, for a 20 cm thick layer, the soil water content at the wilting point is about 12 mm. Apparent density was also calculated from soil samples at different depths ( d ¼ total soil sample dry weight=soil sample volume, g cm À3 ). 13.2.3 EXPERIMENTAL DESIGN Observations and measurements were made on eight trees selected among the most vigorous ones (i.e., trees that had heights and trunk diameters in the upper quartile). In this way, we avoided puny Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 255 9.10.2007 9:35am Compositor Name: VAmoudavally Root Competition for Water between Trees and Grass in a Silvopastoral Plot 255 Copyright 2008 by Taylor and Francis Group, LLC trees, for which poor growth may be due to disease and not due to competition with grass. At the beginning of the experiment (spring 1997), the average height of the trees was 6.5 m and the average trunk diameter at 1.3 m was 8 cm. Three trees were weeded (grass suppression, T À G treatment) in March 1997 with glyphosate (3.6 g L À1 ) in a 4 m radius around the tree trunk to form a control with no grass competition. Their growth was compared with that of five trees maintained in grass (T þ G treatment). Two control plots (or subplots, 100 m 2 each) were installed about 30 m away from the trees; a plot with only grass and no tree (G treatment) and a plot with bare soil (BS treatmen t). For the T À G trees, regular treatments with glyphosate (3.6 g L À1 , one treatment every year at the beginning of the growing season) and manual harrowing (several times in the year) were carried out for 3 years to keep the soil grassfree. All the trees were regularly treated against aphids and Blumeriella jaapii (with, respectively, deltamethrine 0.00075 g L À1 and doguadine 0.72 g L À1 ). 13.2.4 MEASUREMENTS 13.2.4.1 Tree Dimension Tree trunk girth at breast height (1.3 m) and total height of each tree were measured manually every week from 1997 (when trees were 10 years old) to 1999. In addition, for trunk diameter increment, an automatic electric sensor (LVDT type, Solarton DF 2.5) was fitted to the trunk of each tree at about 1.3 m height to record daily variations in trunk diameter: contraction in the day was due to water loss through transpiration flow, and increase during the night was due to water uptake and growth (Améglio and Cruiziat, 1992). The sensor was accurate to less than 2.3 mm. 13.2.4.2 Tree and Grass Water Status Predawn (c p ) and midday (c m ) leaf water potentials of tree and grass were measured each week with a pressure chamber (Scholander et al., 1965). The grass cover was made up of several species. As we were unable to make water potential measurements on all the species present, we chose the most representative species based on abundance for these measurements, that is, Avena pubescens in 1997 and D. glomerata in 1998 and 1999. 13.2.4.3 Grass and Tree Root Growth Grass and tree root densities and elongations were calculated using rhizotrons. Three rhizotrons were installed in April 1997 in three directions at 1.1, 2.2, and 3.3 m from the trunk of a T À G treatment tree and from the trunk of a T þ G treatment tree. One rhizotron was set up in the G treatment. In 1998, two additional rhizotrons were installed 2.2 m from a T À G tree and a T þ G tree. Each rhizotron was 1.25 m deep and 1.0 m wide. Such a dimension was necessary to assess 10 year old tree root systems. The number of rhizotrons was voluntarily limited, given their dimension, to avoid disrupting too much tree growth. In spite of some disadvantages such as modified microclimatic conditions (Taylor et al., 1990; Vogt et al., 1998), rhizotrons allow sequential measurements to be made of the same roots without any destruction (Lopez et al., 1996). Minirhizotrons were not used because they are much more expensive and require numerous long tubes to estimate such large root systems accurately (Franco and Abrisqueta, 1997). 13.2.4.4 Soil Water Content Volume soil water content was measured every week in 20 cm thick layers to a depth of 80 cm with a TDR probe (Time Domain Reflectometry IMKO device). The TDR probe used was a tube type adapted for measurements in permanent thin-walled plastic tubes . Thin-walled tubes were driven vertically into the soil with the help of an auger. Measurements were made every week by lowering the probe into the tubes with a stop measurement every 20 cm to a maximum depth of 80 cm. Three tubes were placed 1.1, 2.2, and 3.3 m (i.e., at the same distance as rhizotrons from tree trunks) from Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 256 9.10.2007 9:35am Compositor Name: VAmoudavally 256 Ecological Basis of Agroforestry Copyright 2008 by Taylor and Francis Group, LLC each T À G tree and each T þ G tree. Control tubes were driven below the G and BS treatments. The use of the TDR technique is a proven method for measuring soil water content accurately with limited disturbance of the soil and root distribution (Werkhoven, 1993; Todoroff and Langellier, 1994; Mastrorilli et al., 1998). 13.2.4.5 Data Analysis It was not p ossible to perform all the measurements in this experiment on more than eight trees, which was already a large task; a thorough statistical analysis was therefore impossible. However, as all the measurements were done at tree scale, it was nevertheless possible to link individual tree growth to each tree’s local conditions: soil characteristics, evolution of soil water content, depth and density of tree and grass roots, etc. Hence, the response of each individual tree was analyzed taking into account the ‘‘treatment’’ variable (with or without grass) as a first explanatory variable and the microsite conditions for each individual tree as a secondary, or covariate factor. Variations in soil water content, which are less sensitive to the initial conditions than absolute values, were set as a cofactor to explain tree growth. In the same way, the relative growth rate (RGR) was calculated for the different tree growth variables (height, diameter, root elongation, etc.), to take into account the initial size of the tree in its growth response (Causton and Venus, 1981; Collet et al., 1996). RGR (day À1 ), for inst ance for girth, for a given period of time t 1 to t 2 (in number of days) was calculated by: RGR ¼ (C 2 À C 1 )=(t 2 À t 1 )½ C 1 , (13:1) where C 1 is the girth at t 1 C 2 is the girth at t 2 Relationship between tree growth and causal variables (i.e., soil water content) was based on regression analysis using the general linear model (Statgraphics plus 5.1 software). Each value of water potential was the mean of three leaf measurements sampled in different parts of the crown of each tree. The value for grass was the mean of 8–10 leaves sampled on different grass clumps. Each TDR value (i.e., for a 20 cm layer from a particular tube) was the mean of three measurements made in three different directions. 13.3 RESULTS 13.3.1 A BOVEGROUND TREE GROWTH During the whole study period, T À G trees displayed a much better height and especially girth growth than T þ G trees (Figure 13.1) and differences tended to increase with time. After 3 years, the T À G tree girth increment was about twice that of T þ G. Over the season, girth RGRs (Figure 13.2) showed some global variations according to tree phenology (i.e., in general, RGR increased at the beginning of the season and decreased at the end), and also that the girth RGRs of TÀG trees were often greater than T þ G girth RGRs, especially during the drought periods (e.g., 4–11 June, 18–25 June, and 15–22 July in 1997; similar data were found in 1998 and 1999). 13.3.2 SOIL WATER CONTENT Volume soil water content fluctuated according to rainfall events and treatments (Figure 13.3). Only data of 1997 are presented, the same soil water patterns being recorded in 1998 and 1999. Only the variations of the 0–20 cm and 40–60 cm soil layers are presented, the 20–40 cm soil layer showing results intermediate between the 0–20 and 40–60 cm soil layers, and the 60 –80 cm soil layer Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 257 9.10.2007 9:35am Compositor Name: VAmoudavally Root Competition for Water between Trees and Grass in a Silvopastoral Plot 257 Copyright 2008 by Taylor and Francis Group, LLC showing no variation. For the 0–20 cm layer, the soil water content of the T À G treatment was about the same as the BS treatment (Figure 13.3a). In contrast, the soil water contents of the G and markedly for the T þ G treatment were much lower than under the BS and T À G treatments, and fluctuated widely according to rainfall events. In the 40–60 cm deep layer, soil water content was much more stable than in the 0–20 cm deep layer (Figure 13.3b) and showed only small variations following some isolated rainfall events for the T À G treatments. In this layer, the soil water content was globally low in comparison with the 0–20 cm layer (between 20 and 30 mm for the T þ G treatment). As observed in the 0 – 20 cm deep layer, we recorded the same hierarchy among the treatments regarding soil water content in the 40–60 cm deep layer, BS > T À G > G > T þ G. 0 5 10 15 20 25 Apr. 1997 Date Girth increment (cm) T − G T + G Sept. 1997 Apr. 1999 Sept. 1999 FIGURE 13.1 Mean tree girth increment at breast height for the T À G and T þ G treatments for the period 1997–1999 (data of 1998 not shown). 0 0.001 0.002 0.003 0.004 Apr. May Jun. Jul. Aug. Sept. Month (1997) Girth RGR (day −1 ) T − G T + G FIGURE 13.2 Mean tree girth RGR time course for the T À G and T þ G treatments over the season: example for 1997. Each point corresponds to the RGR between two consecutive dates. Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 258 9.10.2007 9:35am Compositor Name: VAmoudavally 258 Ecological Basis of Agroforestry Copyright 2008 by Taylor and Francis Group, LLC Focusing on the periods corresponding to marked differences in the T À G and T þ G girth RGRs (Figure 13.2), it is clear that these periods corresponded to ranging degrees of soil water deficit according to date and treatment (Table 13.1).Total soil water contents for the 0–80 cm deep layer were always greater for the T À G treatment (about 200 mm) than for the T þ G treatment (under 140 mm), resulting in a high availability of water for trees of this T À G treatment and a corresponding high tree RGR (Table 13.1, RGRs were alw ays greater than 0.00163 day À1 ). When the amount of available water decreased severely (less than 80 mm, i.e., close to the wilting point), girth growth also decreased and even stopped in some particularly pronoun ced droughts (data not shown). Pooling all the data, a close relationship betw een girth RGR (10 À3 day À1 ) and water availability (WA) (mm) was established: RGR ¼ 0:0177 WA À 0 :8083, R 2 ¼ 0:68, n ¼ 12: (13:2) Figure 13.4 shows in detail the variations of soil water content a few days before and after 22 July 1997, a period of severe drought, according to soil layer depth, rainfall event, and treatment. (a) 0 20 40 60 80 100 Volume water content (mm) 0 20 40 60 80 100 Rainfall (mm) G BS T − G T + G Apr. May Jun. Jul. Aug. Sept. Apr. May Jun. Jul. Aug. Sept. (b) 0 20 40 60 80 100 Volume water content (mm) 0 20 40 60 80 100 Rainfall (mm) Month (1997) FIGURE 13.3 Volume soil water content dynamics (example for 1997) for the different experimental conditions. Values for the T À G and T þ G trees are those at 2.2 m from the trunk. (a) 0–20 cm deep layer and (b) 40– 60 cm deep layer. (Results of the other layers not shown.) Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 259 9.10.2007 9:35am Compositor Name: VAmoudavally Root Competition for Water between Trees and Grass in a Silvopastoral Plot 259 Copyright 2008 by Taylor and Francis Group, LLC Variations in the soil water content were much more marked for the T À G than for the T þ G treatment (Figure 13.4). For the T À G treatment (Figure 13.4a), the soil water content of the 0–20 cm deep layer varied according to the rainfall events: increase with rainfall, decrease with dry period. For the deepest layers, there was a time lag between precipitation events and increase in the soil water content. The water percolation toward the deepest layers sometimes took several days. For the T þ G treatment (Figure 13.4b), no such variations were observed, nor was any water transfer toward the deepest layers observed. Only the 0–20 cm deep layer showed some small variations. It seems that all the water coming from rainfall events was taken up in this 0–20 cm layer as there was no variation in the deepest layers. 13.3.3 WATER STATUS OF TREES Table 13.2 gives the tree and grass leaf water potentials for two consecutive dates of measurement in 1998: 23 July and 30 July, respectively, before and after a period of water deficit. The total amount of rainfall water between 11 June and 9 July was 40 mm; there was then no rainfall for 2 weeks till 23 July and a rainfall event of 15 mm between 23 and 30 July. The mean value of c m for trees was very negative and some individual values were as low as À2.5 MPa for some trees. Despite this severe stress during daytime, the much less negative values of c p indicated that the trees rehydrated themselves partially during the night (Table 13.2). However, there was a significant difference between trees of the T À G and T þ G treatments in predawn leaf water potential, whereas values for the midday water potential were insignificantly different (Table 13.2). Clearly T À G trees rehydrated themselves overnight more than T þ G trees. The recorded tree diameter microvariations between 6 and 28 July 1998 (Figure 13.5) confirmed the leaf water potential measurements: tree contraction during the day reached 0.5 mm (e.g., 20 July—day 201) indicating marked water stress. However, while T þ G tree growth was greatly reduced during this period (Figure 13.5), T À G trees continued to display an impressive growth due TABLE 13.1 Measured Total Soil Water Conten t (mm) Using the TDR Probe for the 0–80 Deep Layer at 2.2 m from Tree Trunk, Calculated Soil Water Content (mm) Corresponding to the Wilting Point of the Same Layer (See Section 13.2) and Resulting Water Content Available for Plant (Total Water Content–Wilting Point Water Content) for Three Different Dates and Associated Girth RGRs (Year 1997) Treatment T À GT1 G Date 11 June 25 June 22 July 11 June 25 June 22 July Total soil water content (mm) as measured with TDR probe for the 0–80 cm deep layer (1) 212 205 198 131 117 105 Soil water content (mm) corresponding to the wilting point for the 0–80 cm deep layer as deduced from ‘‘soil potential–water content’’ curves (2) 53 53 53 42 42 42 Resulting soil water content (mm) available for plant (1–2) 159 152 145 89 75 63 Girth RGR (10 À3 day À1 ) 2.18 1.63 1.86 1.3 0.75 0.95 Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 260 9.10.2007 9:35am Compositor Name: VAmoudavally 260 Ecological Basis of Agroforestry Copyright 2008 by Taylor and Francis Group, LLC T – G 0 10 20 30 40 Rainfall (mm) −40 −30 −20 −10 0 10 20 30 40 −40 −30 −20 −10 0 10 20 30 40 Soil water content variations (mm) 0−20 cm 20−40 cm 40−60 cm 60−80 cm 09 Jul. 17 Jul. 22 Jul. 31 Jul. 06 Aug. T + G 0 10 20 30 40 09 Jul. 17 Jul. 22 Jul. 31 Jul. 06 Aug. Rainfall (mm) Soil water content variations (mm) (a) (b) FIGURE 13.4 Relative variations (according to the initial value at the beginning of the season) of volume soil water content (mm) as measured by TDR probe at 2.2 m from tree trunk for the period around the 22 July 1997, which was a dry one, according to rainfall events and treatment. (a) T À G and (b) T þ G. Each curve corresponds to a soil layer. TABLE 13.2 Mean Predaw n (c p ) and Midday (c m ) Leaf Water Potential for Trees and Grass (Dactylis glomerata) for Two Dates in 1998 (see text for more details) 23 July 30 July c p in Mpa (+SD) c m c p c m T – G À0.32 (0.05) * a À1.64 (0.31) À0.28 (0.02) * a À2.08 (0.26) T þ G À0.74 (0.19) À1.77 (0.21) À0.57 (0.11) À1.92 (0.21) Grass close to the tree b À2.98 (0.72) À3.77 (0.11) À0.84 (0.42) À2.75 (0.49) Grass far from the tree À2.06 (1.41) À3.01 (0.30) À0.34 (0.08) À2.16 (0.12) a * Indicates a significant difference between TÀG and T þ G with a risk level of 5%. b Grass close to the tree is grass in a radius of 1 m around the tree trunk. Grass far from the tree is grass about 3 m from the tree. Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 261 9.10.2007 9:35am Compositor Name: VAmoudavally Root Competition for Water between Trees and Grass in a Silvopastoral Plot 261 Copyright 2008 by Taylor and Francis Group, LLC to good rehydration during the night (i.e., the water balance between night and day was strongly positive for T À G trees but near zero for T þ G trees). Grass c p and c m for 23 July were strongly negative (Table 13.2). Grass located far from the tree (3 m from the trunk) was always less stressed than grass close to the tree (1 m from the trunk) although the relationship was not statistically significant because of a wide dispersion of the water potential values for grass. After the 15 mm rainfall event betw een the two dates (23 and 30 July), grass c p for 30 July reverted to a less negative value although c m values were always very negative though increasing (Table 13.2). The trees did not benefit from this rainfall as much as the grass: their c p values were barely less negative and their c m values were more negative than the values of 23 July. 13.3.4 ROOT GROWTH Rhizotron data showed that grass roots grew mainly in the first 60 cm of soil, with a peak in the 20–40 cm layer, but there were some roots growing even at a depth of 100 cm (Figure 13.6). Tree roots grew mainly 20–80 cm deep, with a peak in the 40–60 cm layer but there were also some roots growing at a depth of 100 cm. There was practically no tree root elongation in the top layer in contrast to grass. Tree root elongation was high at 1.1 m from the trunk and decreased rapidly at 2.2 and 3.3 m from the trunk (Figure 13.6). Irrespective of the depth, the total length of the roots emitted by the grass was much higher than that of the trees (Figure 13.7). The total root length of the grass alone was higher than that of the grass under trees, and the T À G trees emitted longer roots than the T þ G trees. Therefore, it seems that in the T þ G treatment, the soil space was a limiting factor and both tree and grass root growth was limited. Although the grass root system was longer than the tree root system, the roots of the trees grew faster than those of the grass (Figure 13.8), and it was the 0 4 8 12 16 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 Day number Diameter increments (mm) −0.6 −0.5 −0.4 −0.3 −0.2 −0.1 0 0.1 0.2 0.3 0.4 0.5 T − G T + G (night) Contraction (mm) Growth (mm) (day) FIGURE 13.5 Tree diameter microvariations between 6 and 28 July 1998 for the T – G trees (thick line) and the T þ G trees (thin line), each curve represents one tree, and in insert, comparison between the T – G and T þ G tree diameter mean (and standard deviation) net growth during the night and contraction during the day for three days (199, 200, and 201). Batish et al./Ecological Basis of Agroforestry 43277_C013 Final Proof page 262 9.10.2007 9:35am Compositor Name: VAmoudavally 262 Ecological Basis of Agroforestry Copyright 2008 by Taylor and Francis Group, LLC [...]... ed T.K Karalis, 193–204 H64 of NATO ASI Springer: Berlin, Heidelberg, New York Copyright 2008 by Taylor and Francis Group, LLC Batish et al. /Ecological Basis of Agroforestry 43277_C 013 Final Proof page 268 9.10.2007 9:36am Compositor Name: VAmoudavally 268 Ecological Basis of Agroforestry Anderson, L.S and F.L Sinclair 1993 Ecological interactions in agroforestry systems Agroforestry Abstracts 6(2):57–91... 1993 Competition for water and nutrients in forests Canadian Journal of Forest Research 23:1955–1968 Copyright 2008 by Taylor and Francis Group, LLC Batish et al. /Ecological Basis of Agroforestry 43277_C 013 Final Proof page 270 9.10.2007 9:36am Compositor Name: VAmoudavally 270 Ecological Basis of Agroforestry Picon-Cochard, C., A Nsourou-Obame, C Collet, J.M Guehl and A Ferhi 2001 Competition for water... 29 May 18 Jun 8 Jul 28 Jul 1999 FIGURE 13. 7 Root cumulated length (cm mÀ2) along the 1999 season as recorded on the rhizotron windows for the different treatments Copyright 2008 by Taylor and Francis Group, LLC Batish et al. /Ecological Basis of Agroforestry 43277_C 013 Final Proof page 264 9.10.2007 9:36am Compositor Name: VAmoudavally 264 Ecological Basis of Agroforestry T−G Root RGR (day−1) 0.8 0.6... 1999 FIGURE 13. 11 Grass root relative growth rate (RGR, dayÀ1) calculated from the rhizotron windows for the G treatment in 1999 according to the soil layer depth (0–20 cm, 20–40 cm, etc.) Copyright 2008 by Taylor and Francis Group, LLC Batish et al. /Ecological Basis of Agroforestry 43277_C 013 Final Proof page 266 9.10.2007 9:36am Compositor Name: VAmoudavally 266 Ecological Basis of Agroforestry already... agroforestry plantations in France Agroforestry Systems 43:151–167 Balandier, P., H Rapey and J.L Guitton 1997 Improvement and sustainable development of medium altitude areas through agroforestry: tree-grass-animal association In Proceedings of the XI World Forestry Congress, Vol 1, 80 Italy: Food and Agricultural organisation; Antalya, Turkey: Ministry of Forestry of Turkey Balandier, P., H Rapey,... M.G Tjoelker, T Schaeffer and C Muermann 1999 Survival, growth, and photosynthesis of tree seedlings competing with herbaceous vegetation along a water-light-nitrogen gradient Plant Ecology 145:341–350 Copyright 2008 by Taylor and Francis Group, LLC Batish et al. /Ecological Basis of Agroforestry 43277_C 013 Final Proof page 269 9.10.2007 9:36am Compositor Name: VAmoudavally Root Competition for Water... depth of more than 2 m (Lucot, 1997) Such a distribution of the root systems 2500 Number of roots m−2 Grass 2000 1500 1000 500 0 28 Apr Tree 12 May 26 May 9 Jun 23 Jun 7 Jul 21 Jul 1999 FIGURE 13. 9 Cumulated number of roots for the trees and the grass along the 1999 season as recorded on the rhizotron windows Copyright 2008 by Taylor and Francis Group, LLC Batish et al. /Ecological Basis of Agroforestry. .. status of young walnuts as affected by intercropped legumes in a Mediterranean climate Agroforestry Systems 43:71–80 Eastham, J and C.W Rose 1990 Tree=pasture interactions at a range of tree densities in an agroforestry experiment I Rooting patterns Australian Journal of Agricultural Research 41:683–695 Eastham, J., C.W Rose and D.A Charles-Edwards 1990 Planting density effects on water use efficiency of. .. Clermont-Ferrand, France: Cemagref Mastrorilli, M., N Katerji, G Rana and B Ben Nouna 1998 Daily actual evapotranspiration measured with TDR technique in Mediterranean conditions Agricultural and Forest Meteorology 90:81–89 Méloni, S 1999 A simplified description of the three-dimensional structure of agroforestry trees for use with a radiative transfer model Agroforestry systems 43(1–3):121 134 Miller,... al., 1999) How nitrogen and water act together remains to be studied: a low level of water can limit nitrogen uptake by plant roots, and a high level of nitrogen can increase a tree’s resistance to drought 13. 5 CONCLUSION As stated in Section 13. 1, one of the principles of agroforestry is that the different components of the system—here trees and grass—use different resources, or get resources from . II. Agroforestry Systems 32:163–204. Batish et al. /Ecological Basis of Agroforestry 43277_C 013 Final Proof page 270 9.10.2007 9:36am Compositor Name: VAmoudavally 270 Ecological Basis of Agroforestry Copyright. 200, and 201). Batish et al. /Ecological Basis of Agroforestry 43277_C 013 Final Proof page 262 9.10.2007 9:35am Compositor Name: VAmoudavally 262 Ecological Basis of Agroforestry Copyright 2008 by. 1.3 0.75 0.95 Batish et al. /Ecological Basis of Agroforestry 43277_C 013 Final Proof page 260 9.10.2007 9:35am Compositor Name: VAmoudavally 260 Ecological Basis of Agroforestry Copyright 2008 by