**3.1 Typical growth habits of tropical forage grasses**

In terms of tropical perennial forage grasses, two main morphogenetic groups with typical growth habits are usually recognized: *i*) tussock grasses, which are grasses that produce tillers, and have an erect and clumped growth form; and *ii*) creeping grasses, spreading by stolons, rhizomes or both(Skerman & Riveros 1989, Cruz & Boval, 2000,Van de Wouw et al., 2009).The first group includes tufted species as *Pennisetum purpureum, Panicum maximum*, *Andropogon gayanus, Hyparrhenia rufa* and *Brachiaria brizantha* while the examples of the second group are: *Brachiaria humidicola*, *Brachiaria mutica*, *Digitaria decumbens* and *Cynodon nlemfuensis* (Figure 2). Of course there are also species which can be defined as morphological intermediates, examples are: *Dichanthium aristatum, Dichanthium annulatum, Bothriochloa pertusa* and *Digitaria decumbens* (Cruz & Boval, 2000), and less common are: *Brachiaria* (Syn. *Urochloa*) *decumbens*. As pointed out by Cruz & Boval (2000) these morphological intermediates have the capacity to develop into stolons when growing as isolated plants, but in dense stands, most of the stems do not reach the ground but grow laterally at the top of the canopy.

#### **3.2 Measurement of solar radiation interception**

The instantaneous or daily integrated PAR absorbed at each level by a grass canopy (APAR) is the main factor determining the rate of carbon assimilation of individual leaves (Nobel et al., 1993). Therefore it is an important input for canopy photosynthesis models, once at the ecosystem level, leaf canopy is the unit of photosynthesis (Nouvellon et al., 2000, Hikosaka, 2005). Conceptually APAR, as an expression of the energy flux available, is the result of the difference between the net radiation above the canopy and the net radiation below the canopy (Norman & Arkebauer, 1991). It is calculated as:

$$\text{APAR} = (Ei\lambda \text{-} Er\alpha \text{)} \text{ - (} Et\lambda \text{ -} \text{Ers} \lambda \text{ )}\tag{13}$$

where *Erc* and *Ers* are the symbols for radiation reflected by the canopy and the soil surfaces, respectively, and λ = 400-700 nm. For most purposes APAR is approximated by a more easily estimable quantity, the intercepted PAR (IPAR) which expresses a difference between incoming PAR (*Ei*) and the radiation transmitted to the bottom of the canopy (*Et*).

$$\text{IPAR} = E\_i \text{ - } E\_t \tag{14}$$

One measure of light interception efficiency is given by the relationship between IPAR and the total incident PAR at the top of the canopy, named fractional IPAR (fIPAR).

Fig. 2. Some tropical perennial forage grasses. A) Porto Rico Stargrass (*Cynodon nlemfuensis*), showing the stoloniferous growth habit at the edge of a experimental plot (September 2011, end of dry season). B) Dairy heifers grazing on the same pasture, during the rainy season (January 2011). C) Regrowth of napier grass (*Pennisetum purpureum*) sixty days after cutting; D) with subsequent growth, napier grass shows a like-cane growth habit, with erect culms. E) A marandu palisadegrass (*Brachiaria brizantha* cv. Marandu) sward in a dairy farm in Southern Minas Gerais State. (Photographs: courtesy by Dr. Sérgio T. Camargo Filho, Researcher of PESAGRO, Rio de Janeiro).

where *Erc* and *Ers* are the symbols for radiation reflected by the canopy and the soil surfaces, respectively, and λ = 400-700 nm. For most purposes APAR is approximated by a more easily estimable quantity, the intercepted PAR (IPAR) which expresses a difference between incoming PAR (*Ei*) and the radiation transmitted to the bottom of the canopy (*Et*).

One measure of light interception efficiency is given by the relationship between IPAR and

*Cynodon nlemfuensis* (dry season) *Cynodon nlemfuensis* (rainy season)

**B**

*Pennisetum purpureum Brachiaria brizantha*

**E**

Fig. 2. Some tropical perennial forage grasses. A) Porto Rico Stargrass (*Cynodon nlemfuensis*), showing the stoloniferous growth habit at the edge of a experimental plot (September 2011, end of dry season). B) Dairy heifers grazing on the same pasture, during the rainy season (January 2011). C) Regrowth of napier grass (*Pennisetum purpureum*) sixty days after cutting; D) with subsequent growth, napier grass shows a like-cane growth habit, with erect culms. E) A marandu palisadegrass (*Brachiaria brizantha* cv. Marandu) sward in a dairy farm in Southern Minas Gerais State. (Photographs: courtesy by Dr. Sérgio T. Camargo Filho,

Researcher of PESAGRO, Rio de Janeiro).

**C D**

the total incident PAR at the top of the canopy, named fractional IPAR (fIPAR).

IPAR = *Ei* - *Et* ሺͳͶሻ

Soil nitrogen (N) has a strong effect on plant growth. Many studies, worldwide have shown that crop N uptake is co-regulated by both soil N supply and biomass accumulation processes (Fernandes & Rossiello, 1995; Hikosaka, 2005; Lemaire et al., 2007).When water supply is non-limiting, both carbon and nitrogen capture and use processes are closely linked with one another by the development of leaf area and the pattern of intercepted radiation (Lemaire et al., 2007; Giunta et al., 2009), since about half of leaf nitrogen is invested in photosynthetic proteins (Ghannoum et al., 2005; Hikosaka, 2005). Because of these interrelationships, we also consider the roles of N nutrition in the processes of interception and use of solar radiation by forage grasses.

In Figure 3, are presented the data related to fractional PAR intercepted by swards of Tifton 85 bermudagrass (*Cynodon* spp.), during field measurements to evaluate the effects of nitrogen fertilization on several physiological and morphological traits of the grass. Data were selected due to its simplicity of expression, as they show clearly that an increase in availability of N causes a temporal acceleration of the fractional IPAR by the sward, according to a logistic pattern. In practical terms we can say that N accelerates the canopy closure. Thus, at the high level of application of N, a fractional IPAR value of 0.9 was rapidly obtained, about three weeks after the beginning of the rest period. In contrast, in the same period, the control treatment did not surpass the level of 0.5 fIPAR.

Fig. 3. Temporal variations of instantaneous fractional PAR interception values by Tifton 85 bermudagrass canopies under three rates of nitrogen fertilization. Experimental plots (4m x 4 m) were established in July 2008 on a Typic Fragiudult soil in Seropédica, RJ, Brazil. Several measurements were performed between November 30th and December 28th (rainy season) using optical instrumentation. After twenty-eight days, when the 150-N plots reached a level of 0.95 fIPAR (mean of four replicates) the regrowth cycle was finished by cutting. Environmental conditions during the period were as follows: mean solar radiation: 15.9 MJ m-2 day-1, mean air temperature: 25.4° C and total rainfall: 222.5 mm (E. Barbieri Junior & R. Rossiello, unpublished data).

These results can be attributed to the effects of N on the morphogenetic traits responsible for the structural features in this type of pasture (Cruz & Boval, 2000), where axillary meristems develop as horizontal stolons (Figure 2A) under high levels of sunlight and good water supply. Particularly, in this case, nitrogen stimulated significantly (p≤0.05) canopy height growth rates, tiller population density and leaf area development (data not shown).

Some issues related to measurement procedures are pertinent. Photosynthetically active radiation was measured at the top of the sward canopy using a single quantum sensor (LI-191SA) while at the bottom the transmitted PAR was recorded with a LI-191 SA line quantum sensor connected to a LI-250A light meter (LI-COR Inc., Nebraska, USA). The sensor was inserted at the soil surface level regardless how much of dead material was present. This was possible due to young age of this hybrid bermudagrass with a small amount of dead material accumulated at the base. However, older perennial pastures may have sizeable amounts of dead phytomass accumulated on the bottom of the canopy (Le Roux et al., 1997, Guenni et al., 2005, Sbrizzia & Silva, 2008).In stoloniferous species, after 4 or 5 weeks of growth under non-N-limiting conditions, the loss of leaf biomass as a consequence of changes in allocation patterns can account for half of the leaf tissues produced (Cruz & Boval, 2000). In these situations an appropriate evaluation of IPAR may be a substantial problem. As an example, let us consider data on vertical light distribution in the pasture of Porto Rico Stargrass showed in Figure 2A. Measurements were taken under clear sky using optical equipment described above (Figure 4). Results showed that canopy light interception at 12.5 cm and at sensor level heights were 0.873 and 0.986 respectively, i.e. dead phytomass layer was responsible for about 11% of fIPAR (Figure 4A). On the other hand, in a plot adjacent, vegetated by Swazi grass (*Digitaria swazilandensis* Stent), the same variables have values of 0.930 and 0.988 respectively (profile not shown) reflecting morphological differences among the structural components of the two pastures. As noted previously by many researchers, when measuring grass light interception with optical sensors, it is nearly impossible to position the sensor under the grass canopy without disturbing it (Russell et al., 1989). One possible way of avoiding this problem, when disturbance is very apparent (Figure 4B), is to use a single sensor screwed to a transparent ruler graduated (Figure 4C), it is a solution more functional at plot than at field scale. Under field conditions, it may be more interesting to consider the bottom of the sward canopy a given "cut level" above the horizon of standing dead material, knowing however that the amounts of dead material accumulated are seasonally determined.

### **3.3 Leaf area index, extinction coefficient and angular distribution of canopy elements**

Interception of PAR is modified by canopy architecture as represented by the extinction coefficient, *k* (Bréda, 2003, Zhou et al., 2003). For simulation purposes in canopy photosynthesis and radiation interception models, a fixed value for *k* is sometimes assigned (Thornley, 2002). However, research with real canopies has shown that this coefficient varies seasonally, in line with changes in traits such as leaf angle, canopy height or LAI (Bréda, 2003, Polley et al., 2011). A fixed value of *k* may be appropriate for estimating values of ceiling LAIs, i.e. when fIPAR is around 0.90 and the crop growth rate is near its maximum. However, during the previous vegetative growth, in several instances, it has been shown that *k* changes as sward architecture changes. Nouvellon et al. (2000), working with shortgrass ecosystems in northwestern Mexico found that the *k* value for diffuse and global

These results can be attributed to the effects of N on the morphogenetic traits responsible for the structural features in this type of pasture (Cruz & Boval, 2000), where axillary meristems develop as horizontal stolons (Figure 2A) under high levels of sunlight and good water supply. Particularly, in this case, nitrogen stimulated significantly (p≤0.05) canopy height

Some issues related to measurement procedures are pertinent. Photosynthetically active radiation was measured at the top of the sward canopy using a single quantum sensor (LI-191SA) while at the bottom the transmitted PAR was recorded with a LI-191 SA line quantum sensor connected to a LI-250A light meter (LI-COR Inc., Nebraska, USA). The sensor was inserted at the soil surface level regardless how much of dead material was present. This was possible due to young age of this hybrid bermudagrass with a small amount of dead material accumulated at the base. However, older perennial pastures may have sizeable amounts of dead phytomass accumulated on the bottom of the canopy (Le Roux et al., 1997, Guenni et al., 2005, Sbrizzia & Silva, 2008).In stoloniferous species, after 4 or 5 weeks of growth under non-N-limiting conditions, the loss of leaf biomass as a consequence of changes in allocation patterns can account for half of the leaf tissues produced (Cruz & Boval, 2000). In these situations an appropriate evaluation of IPAR may be a substantial problem. As an example, let us consider data on vertical light distribution in the pasture of Porto Rico Stargrass showed in Figure 2A. Measurements were taken under clear sky using optical equipment described above (Figure 4). Results showed that canopy light interception at 12.5 cm and at sensor level heights were 0.873 and 0.986 respectively, i.e. dead phytomass layer was responsible for about 11% of fIPAR (Figure 4A). On the other hand, in a plot adjacent, vegetated by Swazi grass (*Digitaria swazilandensis* Stent), the same variables have values of 0.930 and 0.988 respectively (profile not shown) reflecting morphological differences among the structural components of the two pastures. As noted previously by many researchers, when measuring grass light interception with optical sensors, it is nearly impossible to position the sensor under the grass canopy without disturbing it (Russell et al., 1989). One possible way of avoiding this problem, when disturbance is very apparent (Figure 4B), is to use a single sensor screwed to a transparent ruler graduated (Figure 4C), it is a solution more functional at plot than at field scale. Under field conditions, it may be more interesting to consider the bottom of the sward canopy a given "cut level" above the horizon of standing dead material, knowing however that the

growth rates, tiller population density and leaf area development (data not shown).

amounts of dead material accumulated are seasonally determined.

**3.3 Leaf area index, extinction coefficient and angular distribution of canopy elements**  Interception of PAR is modified by canopy architecture as represented by the extinction coefficient, *k* (Bréda, 2003, Zhou et al., 2003). For simulation purposes in canopy photosynthesis and radiation interception models, a fixed value for *k* is sometimes assigned (Thornley, 2002). However, research with real canopies has shown that this coefficient varies seasonally, in line with changes in traits such as leaf angle, canopy height or LAI (Bréda, 2003, Polley et al., 2011). A fixed value of *k* may be appropriate for estimating values of ceiling LAIs, i.e. when fIPAR is around 0.90 and the crop growth rate is near its maximum. However, during the previous vegetative growth, in several instances, it has been shown that *k* changes as sward architecture changes. Nouvellon et al. (2000), working with shortgrass ecosystems in northwestern Mexico found that the *k* value for diffuse and global radiation decreased as LAI sward increased. We found the same trend in canopies of Tifton 85 bermudagrass modified by nitrogen fertilization. In our study, decreases in *k*PAR with sward height seemed to fit a linear pattern (Figure 5). In this case, sward height is a direct surrogate for herbage biomass or foliage density, structural properties with which it is highly correlated (Oliveira et al., 2010).

Fig. 4. Vertical light distribution in canopies of two stoloniferous species. A) Profile of a Porto Rico Stargrass sward (same as Figure 2A). At the site, mean heights of canopy and dead phytomass were 43 cm and 12.5 cm respectively. Sensor height is 2.5 cm and therefore, when facing upward, this is the location of its sensitive surface nearest ground surface. B) Plot of Swazi grass (*Digitaria swazilandensis* Stent) near the anterior. The proper deployment of the line quantum sensor is impeded by a dense layer of dead material so that its sensitive surface lies suspended at 7.5 cm from the soil surface. C) A better option may be to move a simple sensor through a vertical length of the canopy (Mean height: 29 cm). Measurements were taken on day 21stof September 2011 at PESAGRO Experimental Dairy Farm, Seropédica, RJ, Brazil, at the time corresponding to SZA between 24.8º and 28.0º.

Preliminarily we must recognize that *k*PAR data as shown in Figure 5 are, to some degree, oversized. This fact is a result of using an optical approach for the measurement of intercepted PAR. The instrumentation used in this method does not discriminate between leaves and stems, or among green, senescent or dead leaf blades (Asner et al., 2003, Bréda, 2003). Overestimation arises because k values are calculated with the transmission values of the whole standing foliage, but with LAI values that includes only green leaf blades, which is the so-called "true LAI"(He et al., 2007). However the magnitude of this overestimation is a matter of debate because it is species- specific (Bréda, 2003, Guenni et al., 2005, He et al., 2007), and in our case it is assumed that it is distributed equally among the treatments since the canopy leaf to stem ratio was almost invariant (Figure 5). Given this, the data indicate that under nitrogen influence, there is a structural change in the canopy towards a more erectophile condition.

Fig. 5. Canopy heights of Tifton 85 bermudagrass related to corresponding extinction coefficients (*k*PAR) values, after a regrowth period of 35 days (From January 25th to March 1st 2007). Variations in these traits were induced by nitrogen fertilization (0, 75,150,227 and 300 kg N-urea/ha). Concurrent values of Green Leaf Area Index ranged from 0.78 to 4.19.Mean leafiness (leaf blade: stem ratio) was 1.09 ± 0.05 and did not vary significantly (p > 0.05) among applied N doses (A.P. Oliveira & R. Rossiello, unpublished data).

Fig. 6.Maximum stolon length of clonal propagules of Tifton 85 bermudagrass grown in Hoagland solution culture, in response to N levels (0.5 or 10 mM) and days of regrowth in a controlled growth environment . Photosynthetic photon flux density: 450 µmol photon m-2 s-1, air temperature (day/night): 30/24º C, photoperiod: 12 h. (R. Rossiello, unpublished data).

2007), and in our case it is assumed that it is distributed equally among the treatments since the canopy leaf to stem ratio was almost invariant (Figure 5). Given this, the data indicate that under nitrogen influence, there is a structural change in the canopy towards a more

Fig. 5. Canopy heights of Tifton 85 bermudagrass related to corresponding extinction coefficients (*k*PAR) values, after a regrowth period of 35 days (From January 25th to March 1st 2007). Variations in these traits were induced by nitrogen fertilization (0, 75,150,227 and 300 kg N-urea/ha). Concurrent values of Green Leaf Area Index ranged from 0.78 to 4.19.Mean leafiness (leaf blade: stem ratio) was 1.09 ± 0.05 and did not vary significantly (p > 0.05) among applied N doses (A.P. Oliveira & R. Rossiello, unpublished data).

 **= 0,842**

 **= 0,876**

**p (two-tailed) < 0,0001**

**r = - 0,830**

**0,5 mM R2**

**10 mM R2**

**0.4**

**0**

**20**

**40**

**60**

**Stolon lenght (cm)**

**80**

**100**

**0.6**

**0.8**

**1.0**

**Canopy extinction coefficient**

**1.2**

**1.4**

**20 30 40 50 60**

**canopy height (cm)**

Fig. 6.Maximum stolon length of clonal propagules of Tifton 85 bermudagrass grown in Hoagland solution culture, in response to N levels (0.5 or 10 mM) and days of regrowth in a controlled growth environment . Photosynthetic photon flux density: 450 µmol photon m-2 s-1, air temperature (day/night): 30/24º C, photoperiod: 12 h. (R. Rossiello, unpublished data).

**0 7 14 21 28 35**

**Days of regrowth**

erectophile condition.

As the theory predicts we can suppose that this result is due to changes in LAD, but how the changes take place is unclear. A possible interpretation is that by influencing canopy development, nitrogen modifies the light spectrum within the canopy with consequences for the stolon differentiation (Cruz & Boval, 2000). In *Cynodon* species or cultivars, large variation in length and number of stolons might be due to the very plastic response of stolons to light intensity and nutrient availability (Van de Wouw et al., 2009). Willemoes et al.(1987) observed that bermudagrass stolons irradiated with red light showed an upward curvature and an increase in leaf and internode lengths in comparison with those grown in darkness or under red plus far red radiant flux. Thus even in the low levels of photosynthetic irradiance existing in the middle of Tifton 85 canopies, high N availability in the growth medium could increase stolon elongation process, as can be inferred from results obtained in controlled environmental conditions (Figure 6).

In fact, light fluxes fluctuating deeper in the canopy, with variables red/far red ratios, in the presence of a growth substrate rich in N, could form the basis of the canopy response showed in Figure 5. Of course, we can also suppose that this type of response could be a consequence of absence of grazing pressure on shoot morphology. However, in a very different context, Pinto et al. (1999) working with Tifton-85 swards being continuously grazed by sheep found that taller swards (more lenient grazing) had the lowest senescence rates and suggested that changes in sward structure with increasing sward height could be promoting changes in the canopy light environment. Clearly this is an area that deserves more ecophysiological research.

Season has a strong influence on canopy structural properties due to the seasonal course of solar elevation and the associated changes in ratios of diffuse to direct solar beam. Kubota et al. (1994) observed a large structural change of napier grass canopy with growth. Young shoots of the cultivar Merkeron were transplanted in a field and grown for 102 days. During this growth period LAI increased from 0.7 to 15.4 while *k* decreased gradually from 1.1 to 0.38 due to elongation and erection of stems (large increase in the frequency of stalks with angles of 80-90º relative to the soil surface, see Figure 1B). These results indicate that in this grass, changes from a planophile to an erectophile growth pattern (see Figures 2C-D) are accomplished by correlative variations in SAD. This type of modification protects lower leaves from heavy shading, allowing the canopy to approximate an optimum LAI throughout the growth period (Kubota et al., 1994).Besides this, a long duration of vegetative growth are regarded as the main causes of high productivity of this species, with aboveground dry mass yields of 60 tons/ha/year (Morais et al., 2009). Zhou et al. (2003) working with sugarcane (*Saccharum* spp.) cultivars in Zimbabwe, Africa, obtained different results. They found that the *k*PAR values of four cultivars (calculated by solving for *k* in the light extinction expression, as in Figure 5) increased (although not significantly, p>0.05) with increasing crop age, with a mean from 0.47 at 87 days to 0.64 at 116 days after planting. Additionally, it was observed that high stalk population cultivars intercepted more PAR than low stalk population cultivars because they had more intercepting leaf surfaces, but leaf size seemed less important than tiller population to explain differential patterns of PAR interception among cultivars. However, no information about possible differences in stalk angular distribution was given. In still other situations, there may be compensations between LAD and SAD during the growing season, so that the net effect of shifts in canopy angular distribution on light interception is decreased. This was the case in the above work of Nouvellon et al. (2000) who observed that early in the season, LAD was highly erectophile and shifted towards a less erectophile condition during the seasonal growth. However, this trend was compensated by a higher contribution of the highly erectophile stems (LAD↓ and SAD↑) to the total plant area index in the later stages of plant development. These findings suggest that, although in most cases the leaf angular distribution is the predominant factor in solar radiation interception; in some situations the role of steam angular distribution cannot be ignored.

There are few works dealing with values of LAI and light interception of tropical and subtropical grasses under conditions of cutting or grazing. In grazed pastures, leaf tissues are subjected to discrete defoliation events, the frequency and intensity of which greatly affect the physiology of plants and therefore the rate at which new leaf tissues are produced (Lemaire & Agnusdei, 2000). As a general rule, recommendations for grazing management are made in order to preserve a residual LAI suitable for the plants to continue growing thus maintaining the persistence of the herbage resource. In this context, the height of postgrazing residue is one of the determining factors of the regrowth rates in tropical pasture grasses, especially for tussock grasses as *Pennisetum purpureum* (Zewdu et al., 2003). In this species the dynamics of tillering in terms of tiller classes also influences growth rates and herbage accumulation (Skerman & Riveros, 1989). Carvalho et al. (2007) studied the effect of these two variables on the seasonal patterns of leaf area development and light interception, in an experiment performed at an experimental field of Embrapa in Coronel Pacheco, MG, Brazil (21º 33´ S, 43º06´ W, 410 m).Two post-grazing residues (50 and100 cm) and two tiller classes (basal and aerial) were combined in a split-plot arrangement, from October 2002 to April 2003. Selected results of this work are shown in Figure 7.

Fig. 7. Effects of post-grazing residue (50 and 100 cm) and tiller class on leaf area index (LAI) and PAR interception of napier grass swards during two grazing cycles. A) Leaf area index of basal (b), aerial (a) and total (t: basal + aerial) tillers affected by sward height post-grazing residue and grazing cycle. B) Canopy PAR interception values, evaluated in pre-grazing conditions, in response to the same variables. Number at the top indicates the value of extinction coefficient (*k).* LAI was determined destructively, according to: LAI = leaf area.tiller-1 x tiller number.m-2. IPAR was measured using LI-COR optical sensors. Grazing cycles: second, from November 3th to December 6th 2002; sixth, from March 15th to April 17th 2003. (Adapted from Carvalho et al., 2007).

During the spring there was a larger appearance of basal tillers in swards managed at 50 cm post-grazing residue. Conversely, population densities of aerial tillers were predominant in the summer months. Interestingly, within each residue height, LAIs values were practically the same in the second and sixth grazing cycles, with different contributions of both tiller types (Figure 7A). However, LAI variations and PAR interception were not strongly related throughout the experimental period. The progressive increase in *k* values from 2nd to 6th cycle indicates that over the grazing cycles the foliage acquired a more planophile arrangement linked to a seasonal change in plant architecture. This shift was mediated by the afore-mentioned proliferation of aerial tillers which have a lower insertion angle than basal tillers and make up a flatter canopy (Carvalho et al., 2007). The marked dominance of aerial tillers in the last grazing cycle was apparently responsible for a greater PAR interception in the pre-grazing condition especially in pastures managed with100cm of residue (Figure 7B). However the authors do not exclude the contribution of dead material in this response. Qualitatively similar results were obtained by Giacomini et al.(2009) working with marandu palisadegrass (*Brachiaria brizantha* cv. Marandu, Figure 4E) subjected to intermittent stocking.

### **4. Radiation use efficiency**

236 Solar Radiation

of Nouvellon et al. (2000) who observed that early in the season, LAD was highly erectophile and shifted towards a less erectophile condition during the seasonal growth. However, this trend was compensated by a higher contribution of the highly erectophile stems (LAD↓ and SAD↑) to the total plant area index in the later stages of plant development. These findings suggest that, although in most cases the leaf angular distribution is the predominant factor in solar radiation interception; in some situations the

There are few works dealing with values of LAI and light interception of tropical and subtropical grasses under conditions of cutting or grazing. In grazed pastures, leaf tissues are subjected to discrete defoliation events, the frequency and intensity of which greatly affect the physiology of plants and therefore the rate at which new leaf tissues are produced (Lemaire & Agnusdei, 2000). As a general rule, recommendations for grazing management are made in order to preserve a residual LAI suitable for the plants to continue growing thus maintaining the persistence of the herbage resource. In this context, the height of postgrazing residue is one of the determining factors of the regrowth rates in tropical pasture grasses, especially for tussock grasses as *Pennisetum purpureum* (Zewdu et al., 2003). In this species the dynamics of tillering in terms of tiller classes also influences growth rates and herbage accumulation (Skerman & Riveros, 1989). Carvalho et al. (2007) studied the effect of these two variables on the seasonal patterns of leaf area development and light interception, in an experiment performed at an experimental field of Embrapa in Coronel Pacheco, MG, Brazil (21º 33´ S, 43º06´ W, 410 m).Two post-grazing residues (50 and100 cm) and two tiller classes (basal and aerial) were combined in a split-plot arrangement, from

October 2002 to April 2003. Selected results of this work are shown in Figure 7.

**6th- 50 6th-100**

Fig. 7. Effects of post-grazing residue (50 and 100 cm) and tiller class on leaf area index (LAI) and PAR interception of napier grass swards during two grazing cycles. A) Leaf area index of basal (b), aerial (a) and total (t: basal + aerial) tillers affected by sward height post-grazing residue and grazing cycle. B) Canopy PAR interception values, evaluated in pre-grazing conditions, in response to the same variables. Number at the top indicates the value of extinction coefficient (*k).* LAI was determined destructively, according to: LAI = leaf area.tiller-1 x tiller number.m-2. IPAR was measured using LI-COR optical sensors. Grazing cycles: second, from November 3th to December 6th 2002; sixth, from March 15th to April 17th

**Intercepted PAR (%)**

**2nd. 6th. <sup>50</sup>**

**Grazing cycle**

**k = 0,36**

**k = 0,85**

**k = 0,99 <sup>B</sup>**

**PG 50 PG 100**

**k = 0,55**

role of steam angular distribution cannot be ignored.

2003. (Adapted from Carvalho et al., 2007).

**b a t b a t b a t b a t**

**Tiller class**

**2nd- 50 2nd-100**

**LAI (m2 m-2)**

**A**

Monteith (1972) showed that phytomass production under tropical climate conditions is correlated with the amount of photosynthetically active radiation (PAR) absorbed by plants. This finding provides the basis for deriving the concept of ecosystem gross primary productivity (GPP). Ecosystem GPP may by calculated using algorithms that employ the light-use efficiency (LUE) concept (Polley et al., 2011). LUE (ε) is a conversion factor or the ratio of GPP to APAR (Equation 13). Following this concept, we have:

$$\text{GPP} = \text{APAR} \times \text{εPAR} = \text{PAR} \times \text{fAPAR} \times \text{εPAR} \tag{15}$$

where fAPAR is the fraction of PAR that is absorbed by the grass canopy. From this identity, we can infer that green or dry biomass could be increased when radiation absorption or use efficiency, or both, are maximized. However, Norman & Arkebauer (1991) considered two meanings for the term "use efficiency": *i*) mass of CO2 fixed per unit of absorbed photosynthetically active radiation, or *photosynthetic light-use efficiency* and *ii*) mass of dry matter (DM) produced per unit of absorbed photosynthetically active radiation or *dry matter light efficiency* which is the same as Equation 15. As noted by these authors, the second definition is more problematic as it involves both maintenance and growth respiration terms, which may not depend on light directly. Despite this objection, this agronomic definition is the most frequently cited in radiation use research, where IPAR can replace APAR (Norman & Arkebauer, 1991). According to Albrizio & Steduto (2005) for a given species and environment, RUE is approximately a constant value during the growth season, provided that: *i*) respiration is proportional to photosynthesis*; ii*) photosynthesis response to irradiance is essentially linear at the canopy scale and *iii*) no substantial change in the chemical composition of biomass occurs during the growth cycle considered. Under nonlimiting water and nutrient conditions, all of these conditions can be met for tropical forage grasses, however, to date there are few available data. Guenni et al. (2005) reported RUE values for five *Brachiaria* species that ranged (not significantly) from 1.3 to 1.7 g DM (MJ IPAR)-1 for *B. brizanta* and *B. humidicola,* respectively.
