ARTICLE IN PRESS Journal of Arid Environments
Journal of Arid Environments 70 (2007) 152–163
www.elsevier.com/locate/jaridenv
Effect of tree density on productivity of a Prosopis cineraria agroforestry system in North Western India G. Singh, S. Mutha, N. Bala Division of Forest Ecology, Arid Forest Research Institute, New Pali Road, Jodhpur 342004, India Received 13 September 2005; received in revised form 24 November 2006; accepted 16 December 2006 Available online 31 January 2007
Abstract An experiment was initiated in 1991 to evaluate crop productivity and define optimum tree density with advancing age of Prosopis cineraria (L.) (Khezri) in an agroforestry system. Plots of P. cineraria at densities of 417 (D1), 278 (D2) and 208 trees/ha (D3) were intercropped with Vigna radiata (L.) (mungbean) in 1995, 1997, 1999 and 2000 with Pennisetum glaucum (L.) R. Br. (pearlmillet) in 1998 and 2001. Tree height and collar diameter increased by 2.5 and 2.2-fold in D1, 2.2 and 2.4-fold in D2 and 2.2 and 2.0-fold in D3 plot, respectively in the 6-year period. The highest crop yields were found in D2 plots in 1995 and 1996, in D3 plots in 2000 and in the control plots in 2001. The lowest crop yields were found in D1 plots throughout the duration of the experiment. Trees produced utilizable biomass of 19.1, 15.8 and 10.3 tones/ha and dry leaf weight of 0.85, 0.67 and 0.50 tones/ha, respectively in the D3, D2 and D1 plots at the age of 12 years (June 2002). Low soil water content at 1 m distance from tree base compared to that at the center of four trees indicated greater utilization of soil water within the tree rooting zone. The yield of the annual crop increased when density of P. cineraria was appropriate (i.e., optimum tree density). But optimum tree density varied with tree size/age due to competition for soil resources. Yield of the annual crops was the highest at optimum tree densities of 278 trees/ha (4 m 9 m) at 6 and 7 years, 208 trees/ha (8 m 6 m) at 10 year and o208 trees/ha at 11 years of age. The study indicated greater benefits of P. cineraria tree integrated at optimum density through tree produced and synergistic effects on the annual crops. r 2007 Published by Elsevier Ltd. Keywords: Crop production; Prosopis; Soil nutrients; Soil water; Tree biomass
Corresponding author. Tel.: +91 291 2722550; fax: +91 291 2722764.
E-mail address:
[email protected] (G. Singh). 0140-1963/$ - see front matter r 2007 Published by Elsevier Ltd. doi:10.1016/j.jaridenv.2006.12.003
ARTICLE IN PRESS G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
153
1. Introduction Low and erratic rainfall, intense solar radiation, high wind velocity, recurring drought and famines are common features of the Indian arid zone. Farmers have practiced agroforestry as a strategy to ease out these problems (Malhotra, 1984). Farmers maintain and promote growth of randomly and widely spaced trees of Prosopis cineraria (L.), Tecomella undulata (L.) and Ziziphus nummularia (L.) on their cultivated fields. These trees sustain the farmers during the crop failure by producing food, fodder and timber. Current densities of these trees are very low and variable. For example, density of P. cineraria ranges from 3 trees/ha in western to more than 80 trees/ha in eastern part of Indian desert (Tejwani, 1994). A density of P. cineraria trees 4150 ha1 has also been observed in some cultivated fields (personal observation). Considering an average of only 1.27% forest cover in the arid region of western Rajasthan (Forest Survey of India, 1999), tree cover must be increased in agricultural fields to meet the local need of fodder and fuel. Growing trees at higher densities would be the best option to increase overall productivity of the farmland and to fulfill the increasing demand for fodder and fuel wood. Due to limited resources in arid and semi-arid regions, benefits from agroforestry systems largely depend on the judicious management of soil and water resources (Joshi et al., 1989; Parton et al., 1987). Improved selection of appropriate tree and crop species, growth of trees at optimum densities and adoption of pruning/lopping protocols are important management considerations to increase overall system productivity (Karim and Savill, 1991; Ong et al., 1992). P. cineraria (L.). (Khezri) is the most widely grown tree in the Indian desert because both its leaves and fruits have high fodder and human food value, respectively. Information on P. cineraria cultivation is mostly based on data from randomly growing scattered trees (Aggarwal and Kumar, 1990; Aggarwal et al., 1976; Kaushik and Kumar, 2003; Tejwani, 1994). There is no systematic study to define optimum density of P. cineraria on farmland with increasing tree size/age. There is also an urgent need to increase overall productivity of farmland. We hypothesized that sequential thinning depending upon the outputs of agricultural crop yield and tree growth and biomass would provide appropriate tree density. The objective of the study was to determine the effect of density of P. cineraria (L.) trees with advancing age on crop productivity for integrating this tree at a higher density. Tree-crop interactions were evaluated in terms of tree growth and biomass, crop yield and soil water status. 2. Materials and methods 2.1. Site conditions The experiment was initiated at the Arid Forest Research Institute, Jodhpur with P. cineraria seedlings established at varying spacing with different agricultural crops. The site received 340, 515, 440, 516, 296, 293, 429 and 58 mm rain in 1995, 1996, 1997, 1998, 1999, 2000, 2001 and 2002, respectively, with mean annual rainfall of 350 mm. The rain during July–September (monsoon) was 306, 312, 294, 204, 205, 264, 327, 25 mm in the respective years. In July 1997, 1998 and 1999, the rainfall was 50.2, 56.7 and 58.6 mm, respectively. The maximum temperature rises to as high as 48 1C in the summer and the
ARTICLE IN PRESS 154
G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
minimum drops to 0 1C in the winter. Wind velocity in the summer months is 20–30 km/h. The experimental farm is flat land with loamy sand soil (coarse loamy, mixed hyperthermic family of typic camborthids according to US soil taxonomy) underlain with a thick concretion of calcium carbonate at a depth of 75 cm. Soil had low content of soil organic matter (0.27%), available P (10.2 mg/kg soil), NO3-N (4.01 mg/kg soil) and NH4-N (5.92 mg/kg soil) and pH of 7.8. Soil moisture storage in the upper 75 cm layer varies from 120 mm at 0.01 MPa to 35 mm at 1.5 MPa (Gupta et al., 1998; Singh et al., 2000). 2.2. Seedling plantation and agricultural crops The experiment started in July 1991 by thinning the 1-year-old plantation to maintain tree densities of 1666 tree/ha at 2 3 m2 spacing, 833 tree/ha at 4 3 m2 spacing and 417 tree/ha at 4 6 m2 spacing. Trees were intercropped till 1994 when crop production was highest in 417 tree/ha plot (Gupta et al., 1998). The plantation was further thinned in June 1995 by removing alternate rows of trees in both the axes (X and Y axis) in the 2 3 m2 plot (1667 tree/ha). Two rows of trees alternatively from one axis (X axis) were removed in the 4 3 m2 plot (833 tree/ha), whereas tree rows alternatively from one axis (Y axis) were removed in the 4 6 m2 plot (417 tree/ha). Tree densities after thinning in June 1995 were 417 tree/ha (D1 plots), 278 tree/ha (D2 plots) and 208 tree/ha (D3 plots) in plots sizes of 16 18, 20 18 and 32 18 m2, respectively. The experiment was laid in a randomized block design with three blocks. There were 12 trees/plot in D1 and D3 and 10 trees/plot in D2. Three control plots (size 15 20 m2) without trees were established 20 m from the experimental plots to avoid any interference/competition by trees. Since the study was to monitor the effect of tree density on associated agricultural crop, no control as tree only plot was taken. Following the general practice of crop rotation in the region, Vigna radiata (L.) (mungbean) was grown in 1995, 1997, 1999 and 2000; Pennisetum glaucum (L.) R. Br. (pearlmillet) was grown in 1996 and 2001 and Sesamum indicum (L.) (til) was grown in 1998 in the plots. Crops were sown and harvested in rainy season (July–October) under rainfed conditions. Mungbean (variety S-8) crop was sown at the seed rate of 15 kg seed/ha to maintain 30 15 cm2 spacing, whereas pearlmillet (variety, MH-179) was sown at the seed rate of 5 kg/ha maintaining 45 15 cm2 spacing. 2.3. Observations Tree height and collar diameter (15 cm above the ground level) were recorded twice a year, before crop sowing (June) and after crop harvest (December). Grain and haulms/ straw yields of crops were recorded each year after the harvest (i.e., complete plot harvest in 1995 and 1996). Low rainfall during July in 1997, 1998 and 1999 affected sowing and germination of agricultural crop as well as drying of crop plantlets, hence crop yield was not recorded. In 2000 and 2001, crop yield was recorded at increasing distance from the tree trunk (base), i.e., at 1 and 2 m distances from the base in a tree row as well as at the center of four trees (center) by laying out sample plots of 1 m2 area (three sample plots per tree density plot). Height and collar diameter increment was calculated (Eq. (1)) for each year to monitor the effect of intercrop and rainfall in that
ARTICLE IN PRESS G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
155
particular year: Growth increment ð%Þ ¼ ½ðgrowth variables in a particular year2growth variables in previous yearÞ 100=growth variables in previous year:
ð1Þ
Soil water content (SWC) was determined gravimetrically in the upper 0–75 cm layer in the marked sample plots. In September 1995, March 1996 and September 2001, soil samples were collected from different depths, i.e., 0–25, 25–50 and 50–75 cm soil layers and dried at 110 1C till constant weight for SWC. Gravimetric SWC was converted to mm (Eq. (2)) to give total soil water to the soil depth of 0–75 cm (Gupta et al., 1998): Total soil waterðmmÞ ¼ SWC ð%Þ soil depth in mm bulk density of soil=100=density of water:
ð2Þ
Sixty percent of lower crown of trees was pruned in December 1996 and again in 1999 to give trees proper shape and get benefits of fuel wood and fodder (i.e., leaf) (Bhimaya et al., 1964; Sharma and Gupta, 1981). Fresh weight was recorded immediately after pruning and dry weight was recorded after oven drying (80 1C) of the representative samples. Half of the trees in each plot (6 trees per plot in D1 and D3 and 5 trees per plot in D2) were harvested in June 2002. Of the harvested trees, two trees of each plot were separated into stem, branches and leaves and fresh weight was recorded immediately. Dry biomasses were recorded after oven drying (80 1C) of the samples of stem, branches and leaves. Tree roots were excavated up to 1 m soil depth by 0.5 radius and fresh as well as the dry weight recorded as above. Tree total production per ha or utilizable biomass (stem+branches) was calculated (Eq. (3)): Tree production ðtone per haÞ ¼ Number of tree per ha tree total dry mass ðkg per treeÞ=1000:
ð3Þ
2.4. Statistical analysis The data collected were statistically analyzed using SPSS statistical package version 8.0 for ‘‘Windows 2000’’. Tree height and collar diameter, pruned biomass and harvested tree biomass were analyzed using a one-way ANOVA and the growth variables as the dependent variables. The variations in crop yield due to tree density and distance was tested using a two way ANOVA model considering distance and tree density as the main factors. Since SWC was determined in different soil layers and also at different distance from tree trunk (2001), the data were analyzed using a repeated measure ANOVA considering soil depth as test of within subject effects and tree densities as test of between subject effects. Percent soil water was square root transformed before statistical analysis to reduce heteroscenesdity (Sokal and Rolf, 1981). To obtain the relations between crop production and tree growth variables, SWC and soil nutrients, Pearson correlation coefficient were calculated. The least significant difference test was used to compare treatments at the Po0:05 levels.
ARTICLE IN PRESS 156
G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
3. Results 3.1. Tree growth The effect of density on tree growth was significant ðPo0:05Þ in June of both 1995 and 2002 (Table 1). At 12 years of age (June 2002), trees were significantly ðPo0:01Þ taller and thicker in D3 plots compared with the D1 and D2 plots. The collar diameter was positively correlated with tree height (r ¼ 0:854, Po0:01, n ¼ 9). However, effects of tree density in responses of tree height and collar diameter from 1996 to 2002 were not significant ðP40:05Þ when height and collar diameter of 1995 were considered as covariant. The increase in height over the base data of 1995 was 2.5-fold in D1 and 2.0-fold in D3 plots. The corresponding increase in collar diameter was 2.2-fold in D1 and D3 and 2.4-fold in D2 plots. Percent increments in tree height and collar diameter (1995–96 to 2001–02) were highest ðPo0:05Þ in 1995–96 and decreased in the subsequent years (Fig. 1). Growth increment was comparatively more when V. radiata was the intercrop and/or the rainfall was higher than when the pearlmillet was the intercrop.
Table 1 Growth of Prosopis cineraria as influenced by tree density and the associated crops (recorded in June) at Jodhpur, India Tree density (no./ha) Height (cm) 417 (D1) 278 (D2) 208 (D3) ANOVA P-value Collar diameter (cm) 417 (D1) 278 (D2) 208 (D3) ANOVA P-value
1995
1996
1997
1998
1999
2000
2001
2002
178a (10.1) 223b (4.0) 278c (7.5)
253a (4.0) 332b (9.9) 417c (5.0)
344a (9.3) 378b (5.5) 461c (2.5)
383a (4.5) 406ab (3.6) 485b (30.9)
406a (4.2) 422b (2.3) 500c (5.7)
421a (2.1) 435b (2.3) 514c (3.6)
439a (5.1) 469b (3.5) 534c (6.9)
450a (9.1) 497b (8.9) 542c (9.9)
o0.01
o0.01
o0.01
o0.05
o0.01
o0.01
o0.01
o0.01
5.1a (0.09) 6.0b (0.11) 8.3c (0.14)
7.1a (0.12) 8.8b (0.10) 10.7c (0.12)
8.1a (0.12) 10.4b (0.17) 11.5c (0.25)
8.8a (0.15) 11.2b (0.23) 12.6c (0.24)
10.0a (1.13) 12.3b (0.42) 13.5c (0.51)
10.2a (0.20) 12.4b (0.21) 14.1c (0.15)
10.8a (0.17) 13.7b (0.12) 14.9c (0.21)
11.1a (0.45) 14.3b (0.34) 16.5c (0.35)
o0.01
o0.01
o0.01
o0.01
o0.05
o0.01
o0.01
o0.01
P-values are result of a one-way ANOVA. Values within columns followed by different letters differ significantly at Po0.05. Tree densities 417, 208 and 208 tree/ha are designated by D1, D2 and D3, respectively. Values are means of three replicate plots with SE7 in parentheses.
ARTICLE IN PRESS G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
157
80 Height
Height increment (%)
70 60
a a a b
50 40
b
30 20
b a
a
10
b a
a
a a
b a a
b a
a
a
a
0 70 Collar diameter
417 tree per ha (D1)
b
a per ha (D2) 278 tree
60 Collar diameter increment (%)
208 tree per ha (D3)
c
50 b 40
a 30 c
20
b
b a
10
a a
a a
a
a
a
a
a
b a a
0 1995-96
1996-97
1997-98
1998-99 Year
1999-00
2000-01
2001-02
Fig. 1. Effect of tree density and agriculture crop on percent growth increment in height and collar diameter of Prosopis cineraria trees. Error bars are 7SE of three replicate plots. Columns within years denoted by different letters differ significantly at Po0:05. Tree density and years are same in both the panels.
3.2. Biomass Pruned biomass (60% of crown) increased with tree age and varied with tree size in both the years 1996 and 1999 (Table 2). It was greater ðPo0:01Þ in D3 compared to D1 and D2 plots. The contribution of leaves and the branches in the pruned biomass were 11% and 89% respectively. Tree total biomass at 12 years was higher ðPo0:01Þ in D3 than the trees in D1 and D2 plots (Table 2). Utilizable biomass (stem+branches) of trees was 19.1 tone/ ha in D3 as compared to only 15.75 tone/ha in D2 and 10.3 tone/ha in D1 plots. Dry leaf production was 0.85 tone/ha in D3 to 0.50 tone/ha in D1 plots.
ARTICLE IN PRESS G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
158
Table 2 Pruned biomass in 1996 and 1999 and tree total biomass of 12-year-old tree of P. cineraria influenced by tree density and associated crop at Jodhpur, India Tree density (no./ha) Pruned biomass (kg per tree) Tree total biomass (kg per tree)
417 (D1) 278 (D2) 208 (D3) ANOVA P-value
1996
1999
June 2002
Total
Total
Stem
9.72a (0.41) 15.40b (0.47) 19.64c (0.84) o0.01
13.55a (0.87) 20.56b (0.65) 22.10c (0.55) o0.01
Branch
17.66a (1.27) 43.33b (3.39) 68.16c (1.91) o0.01
7.00a (0.64) 13.31b (0.67) 23.23c (1.72) o0.01
Leaf 1.16a (0.10) 2.35b (0.09) 4.10c (0.52) o0.01
AB
Root
25.82a 8.16a (1.31) (0.42) 58.99b 17.00b (3.00) (1.60) 95.49c 29.83c (3.83) (1.41) o0.01 o0.01
Total 33.98a (1.18) 75.99b (3.47) 124.65c (7.08) o0.01
P-values are result of one-way ANOVA. Values within columns followed by different letters differ significantly at Po0.05. AB, above ground biomass (stem+branch+leaf). Values are means of three replicate plots with SE7 in parentheses.
3.3. Crop production In 1995 the highest ðPo0:01Þ grain yield for V. radiata was in the D2 plots, with a much lower yield in the D1 and slight lower yield in the D3 plots (Fig. 2). The reduction of grain yield was 22.3% ðPo0:05Þ in D1 and 3.5% in D3 plots compared to that in D2 plots. Haulms/straw yield followed a trend similar to grain yield. In 1996, pearlmillet was badly infested by birds and, therefore, grain yield could not be recorded (Fig. 2a). Straw yield recorded in 1996, indicated highest ðPo0:01Þ yield in D2 plots. The D1 plot produced 58.2% less straw compared to D2 plots. In 2000 and 2001, grain production was highest ðPo0:05Þ in the D3 plots and the production decreased with increasing tree density from D3 to D1 plots. Production of straw was also highest in D3 plots (not significant in 2000 and Po0:01 in 2001) (Fig. 2). Distance from tree trunk had a significant ðPo0:05Þ effect during 2000 on both grain and haulms yield (V. radiata), which showed highest yield at center than at 1 and 2 m distances from the trees (Table 3). Reductions in grain and straw yields of pearlmillet in 2001 at 1 and 2 m distances were significant ðPo0:05Þ in D1 and D2 plots (Table 3). In center of four trees, grain yield was similar ðP40:05Þ to that in the control plot. Crop yield (grain+straw, kg/ha) had exponential relation with tree total biomass (TB in kg/tree) in 2001 (Eq. (4)): Crop yield ðkg=haÞ ¼ 4000:546 exp0:002432TB , R2 ¼ 0:727;
RMSE ¼ 0:00405;
F ¼ 18:60;
(4) Po0:05.
ARTICLE IN PRESS G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
400
159
Grain
Grain yield (kg per ha)
350 c b
300 250
b
a
d
200
a
b
c a b
a
b
150 100 50 0
Haulm/ straw yield (kg per ha)
7000
Haulms/ straw 417 tree per ha (D1) 278 tree per ha (D2) 208 tree per ha (D3) control
6000 5000
b b a
a
4000 3000 2000
a
b
a
a
d
c
a
a b b
b a
1000 0 1995
1996
2000
2001
Year Fig. 2. Effect of Prosopis cineraria tree density on yield of grain and haulms/straw for Vigna radiata in 1995 and 2000 and Pennisetum glaucum in 1996 and 2001. Error bars are 7SE of three replicate plots. Columns within years denoted by different letters differ significantly at Po0:05. Tree density and years are same in both the panels.
3.4. Soil water content Repeated-measures ANOVA indicated difference ðPo0:01Þ in SWC between soil layers but SWC did not differ ðP40:05Þ in the tree densities plots except at center of four trees in 2001(Table 4). However, LSD showed highest SWC in D3 plot in the three soil layers in 1995 and 1996 (Po0:05 in 0–25 cm soil layer) except in 50–75 cm soil layer in 1995, when SWC was highest in the D1 plots. D2 plots indicated lowest SWC in 1995. But in March 1996, SWC was lowest in D2, D3 and D1 plots in 0–25, 25–50 and 50–75 cm soil layers, respectively. In September 2001, lowest SWC (average of densities in 0–75 cm soil layer) was at 1 m (18.3 mm) as compared to 20.9 mm at 2 m and 21.6 mm at the center of four trees. SWC was in the order of 1 mo2 mocenter in 25–50 ðPo0:05Þ and 50–75 cm soil layers ðPo0:01Þ, and 1 mocentero2 m in 0–25 cm soil layer ðP40:05Þ. Among the tree density plots, lowest SWC was in D1 plots in 0–25 cm (Po0:05 at center) and in D2 plots in deeper soil layers at all the three distances (Table 4). However, highest SWC was in D3 plot in most of the observations. Considering the means for distance from tree trunk, SWC was greater at the center than at 1 and 2 m distances in 0–25 cm soil layer in D2 plots. Whereas,
ARTICLE IN PRESS 160
G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
Table 3 Average yield of Vigna radiata and Pennisetum glaucum as influenced by tree density and distance from tree trunk at Jodhpur, India Tree density
Vigna radiata (2000)
(no./ha)
1m
Grain yield (kg/ha) Control (no tree) — 417 (D1) 278 (D2) 208 (D3)
137 (3.3) 150 (5.8) 170 (5.8)
2m
Centre
Meana
1m
2m
Centre
Meana
—
—
—
—
—
153 (3.3) 160 (5.8) 183 (8.8)
170 (17.3) 177 (23.3) 227 (9.92) P-value o0.05 o0.05 ns
180c (9.01) 153a (2.91) 162a (6.67) 193b (6.97)
127 (8.3) 137 (3.0) 152 (3.29)
136 (23.5) 142 (9.8) 157 (5.8)
137 (4.56) 145 (2.89) 164 (5.5) P-value o0.01 ns ns
165b (5.01) 133a (7.77) 141a (3.76) 158b (1.58)
—
—
—
—
—
1676 (50.4) 1733 (83.3) 1786 (59.3)
1706 (93.3) 1700 (59.2) 1873 (12.0) P-value ns o0.05 ns
1720a (45.9) 1666a (42.6) 1699a (51.4) 1803a (41.8)
4180 (44.1) 4330 (57.1) 5190 (50.7)
4330 (45.4) 4390 (100.2) 5480 (85.0)
4470 (40.9) 4560.0 (62.5) 5580 (45.6) P-value o0.01 ns ns
ANOVA result Density Distance D Distance Haulm/straw yield (kg/ha) Control (no tree) — 417 (D1) 278 (D2) 208 (D3) ANOVA Density Distance D Distance
1617 (72.6) 1663 (45.0) 1750 (28.9)
Pennisetum glaucum (2001)
5470b (49.8) 4330a (53.8) 4430a (57.6) 5420b (48.6)
P-values are result of a two-way ANOVA for mean yields. Values within columns followed by different letters differ significantly (Po0.05). Values are means of three replicate plots with SE7 in parentheses. Ns, non-significant at P ¼ 0.05 level. a Average values of three sample plots (i.e., situated at 1 and 2 m distance from tree in a row, and center of four trees) in the tree density plot.
D1 plots indicated greater SWC in the deeper soil layers at 1 m distance as compared to other distances (Table 4). 4. Discussion Variation in growth of P. cineraria trees in 1996–2002 was a result of variability in tree sizes at the start of the experiment in June 1995 (Table 1) as well as the tree density and resource availability (Table 4). Lower tree density and the consequent greater availability of soil water resulted in greater height and thickness of trees in the D3 plot compared with the other plots. Increased spacing between the trees as a result of thinning and the concomitant increase in water availability, compared with the previous density of 1666 tree/ha in June 1995 resulted in greater growth in trees in the D1 plot compared with
ARTICLE IN PRESS G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
161
Table 4 Soil water content (% w/w and total in mm up to 0–75 cm soil depth) as influenced by tree density of Prosopis cinerarias, associated crops and distance from tree trunk at Jodhpur, India Tree density Soil depth
September
March
September 2001
(no./ha)
(cm)
1995
1996
1m
417 (D1)
0–25 25–50 50–75 0–75 0–25 25–50 50–75 0–75 0–25 25–50 50–75 0–75
Total 278 (D2)
Total 208 (D3)
Total
0.64 (0.08) 4.27 (0.10) 4.84 (0.18) 36.6a 0.62 (0.05) 4.16 (0.32) 4.03 (0.44) 33.0b 0.91 (0.10) 4.35 (0.28) 4.44 (0.34) 36.4a
Test of within subject effects (P-value) Soil layer o0.01 Soil layer Density ns Densityb o0.05
0.48 3.00 4.03 28.2a 0.47 3.05 4.70 30.8a 0.74 2.27 5.36 31.4a o0.01 o0.05 ns
(0.08) (0.26) (0.26) (0.04) (0.08) (0.56) (0.07) (0.04) (0.32)
1.16 1.80 2.22 19.4 1.23 1.42 1.86 17.0 1.24 1.49 2.16 18.3 o0.01 ns ns
2m (0.06) (0.20) (0.18) (0.02) (0.28) (0.15) (0.13) (0.12) (0.10)
1.21 1.97 2.44 21.1 1.36 1.60 2.41 20.1 1.39 1.91 2.45 21.6 o0.01 ns ns
Meana
Centre (0.10) (0.15) (0.19) (0.05) (0.20) (0.25) (0.03) (0.16) (0.24)
1.13 2.06 2.47 21.2 1.31 1.66 2.42 20.2 1.25 2.14 2.83 23.3
(0.06) (0.08) (0.26) (0.05) (0.15) (0.07) (0.13) (0.04) (0.17)
1.18 (0.38) 1.94 (0.02) 2.37 (0.01) 20.6b 1.30 (0.01) 1.56 (0.13) 2.23 (0.15) 19.1a 1.29 (0.07) 1.85 (0.07) 2.48 (0.07) 21.1b
o0.01 o0.05 o0.05
P-values are results of a repeated-measures ANOVA (soil layers) in 1995 and 1996 and separately for each distance in 2001. Values within columns followed by different letters differ significantly at Po0.05. Values are means of three replicate plots with SE7 in parentheses. ns, non-significant at P ¼ 0.05 level. a Average values of three sample plots (i.e., situated at 1 and 2 m distance from tree in a row, and center of four trees) in the tree density plot. b Test of between-subject effects (tree density).
those in D2 and D3 plots. Tree thinning in June 1995 provided 75% more space in D1 compared with 66.6% in D2 and 50% in D3 plots, due to reduction in tree density (Gupta et al., 1998). Greater dry biomass for both pruned and tree total biomass in the trees of D3 plot (Table 2) was due to greater availability of soil water (Table 4), which might have been utilized by other trees at higher density. An increase in crop yield with tree total biomass (Eq. (4)), which increased with tree spacing as well as tree age, suggested a complementarity effect of P. cineraria on associated agricultural crop (Ong and Leakey, 1999). Higher crop yield in D2 plots in 1995 and 1996 (Fig. 2) was a result of greater and/or efficient use of soil water resources than in the other plots (Table 4). Further, the crop production in D2 and control plot was equivalent ðP40:05Þ in the years 1995 and 1996 (i.e., at 6–7 years of age) suggesting that tree density of 278 trees/ha was advantageous for crop yield (optimum tree density). Despite relatively greater SWC in D1 than in D2 plot in September 1995 (Table 4), lower ðPo0:01Þ crop yield in D1 plot (Fig. 2) was due to soil water use in upper soil layer by both P. cineraria trees and the agriculture crop indicating that a density of 417 tree/ha showed adverse effects on crop production. Low SWC in D1 as compared to those in D2 and D3 plots in March 1996 (no crop period) also showed greater use of soil water by the densely populated trees in D1 plot (Table 4). Reduced crop yield in D3 plots in the years 1995 and 1996 was due to more open tree canopy and reduced SWC in 0–25 cm soil layer (as in 2001, center, Table 4) in the D3 plot as compared to the D2 plot.
ARTICLE IN PRESS 162
G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
Therefore, a greater soil water deficit (Gibson and Bachelard, 1986; Singh et al., 2003) as a result of probable increase in light intensity (Saunders et al., 1991) affected crop production in D3 plots compared with D2 plots. Reduced crop yield in the control plots (without tree) than those in the D2 and D3 plots in 1995 and 1996 suggested inference (Fig. 2). It is similar to the synergistic effects of P. cineraria on the productivity of Cicer arietinum (chickpea) (Puri et al., 1994). Boffa et al. (2000) recorded higher grain yield of Sorghum bicolor in Vitellaria paradoxa parkland trees of mean crown radii of 225–275 cm and an average tree densities of 12 and 31 tree/ha than in areas without trees. Higher yield of V. radiata and Pennisetum glaucum in D3 and control plots, respectively in 2000 and 2001 suggested that a density of 208 tree/ha (D3) was the optimum density for highest crop production in 2000. Low crop yield in 2000 and 2001 at 278 tree/ha suggested that tree size and densities play important role in enhancing production of intercrop (Tables 2 and 3). The increased herbaceous productivity under canopy of savanna trees is the most common pattern in tropical tree communities with low density, low rainfall and moderate soil fertility (Belsky, 1994; Chesson et al., 2004). Comparatively low yield of P. glaucum in the D1, D2 and D3 plots than with the control plot in 2001 suggests that even a tree density of 208 tree/ha (D3) became competitive in the year 2001 and optimum tree density lies below 208 tree/ha at this age and/or size of the tree. Low yield of Brassica compestris has been recorded when intercropped with 15-year-old H. binata tree as compared to control plots (sole crop) in the semi-arid region of India (Shanker et al., 2005). However, low growth in the trees with pearlmillet (1996–97 and 2001–02) intercrop as compared to mungbean (1995–96 and 2000–01) intercrop suggested that crop types and resource availability affect tree growth. 5. Conclusions The study suggests that P. cineraria enhances productivity of agricultural crops when maintained at the appropriate density. Optimum tree density declined with tree age/size, i.e., 278 tree/ha (4 9 m2) at 6 and 7 years, 208 tree/ha (8 6 m2) at 10 year and less than 208 tree/ha at 11 years of age and above. P. cineraria showed a complementarity effect on annual crops observed by an increase in crop yield with tree total biomass. Tree total biomass increased with tree spacing and tree age. Thus tree size, competition for resources at high densities and availability of soil resources at low tree density were the probable factors affecting crop production (Holmgren et al., 1997). A significant finding is the synergistic interactions between annual crops and P. cineraria trees grown and maintained at optimum tree density. These results indicate biomass and crop benefits of optimum tree density over traditional. By maintaining P. cineraria trees at optimum tree density, the complementarity effect of the trees could be utilized to obtain maximum produce from the system. Acknowledgements The authors are thankful to the Director, Arid Forest Research Institute for providing necessary facilities to carry out the experiment.
ARTICLE IN PRESS G. Singh et al. / Journal of Arid Environments 70 (2007) 152–163
163
References Aggarwal, R.K., Kumar, P., 1990. Nitrogen response to pearl millet grown on soil underneath P. cineraria and adjacent open site in an arid environment. Annals of Arid Zone 29, 289–293. Aggarwal, R.K., Gupta, J.P., Saxena, S.K., Muthana, K.D., 1976. Studies on soil physico-chemical and ecological changes under twelve year old five desert tree species of western Rajasthan. Indian Forester 102, 863–872. Belsky, A.J., 1994. Influence of trees on savanna productivity: tests of shade, nutrients, and tree-grass competition. Ecology 75, 922–932. Bhimaya, C.P., Kaul, R.N., Ganguly, B.N., 1964. Studies on lopping intensities of Prosopis spicigera. Indian Forester 90, 19–23. Boffa, J.-M., Taonda, S.J.-B., Dickey, J.B., Knudson, D.M., 2000. Field-scale influence of karite´ (Vitellaria paradoxa) on sorghum production in the savanna zone of Burkina Faso. Agroforestry Systems 49, 153–175. Chesson, P., Gebauer, R.L.E., Schwinning, S., Huntly, N., Wiegand, K., Ernest, M.S.K., Sher, A., Novoplansky, A., Weltzin, J.F., 2004. Resource pulses, species interactions, and diversity maintenance in arid and semi-arid environments. Oecologia 14, 236–253. Forest Survey of India, 1999. The States of Forest Report. Forest Survey of India, Dehradun, Ministry of Environment and Forest, Government of India, pp. 86–89. Gibson, A., Bachelard, E.P., 1986. Germination of Eucalyptus sieberi, L. Johnson seeds. I. Response to substrate and atmospheric moisture. Tree Physiology 1, 57–65. Gupta, G.N., Singh, G., Kachwaha, G.R., 1998. Performance of Prosopis cineraria and associated crops under varying spacing regimes. Agroforestry Systems 40, 149–157. Holmgren, M., Scheffer, M., Huston, M., 1997. The interplay of facilitation and competition in plant communities. Ecology 78, 1966–1975. Joshi, D.C., Arora, B.R., Aggarwal, R.K., Ruhal, D.C., Sharma, B.K., 1989. Forms and content of nutrient elements. In: Review on Research on Sandy Soils in India, International Symposium on ‘Managing Sandy Soils, Central Arid Zone Research Institute, Jodhpur, India, pp. 85–112. Karim, A.B., Savill, P.S., 1991. Effect of spacing on growth and biomass production of Gliricidia sepium (Jacq) Walp in an alley cropping system in Sierra Leone. Agroforestry Systems 18, 213–222. Kaushik, N., Kumar, V., 2003. Khejri Prosopis cineraria based agroforestry system for arid Haryana, India. Journal of Arid Environments 55, 433–440. Malhotra, S.P., 1984. Traditional agroforestry practices in arid zone of Rajasthan. In: Shankarnarayana, K.A. (Ed.), Agroforestry in Arid and Semi-Arid Zones. CAZRI, Jodhpur, India, pp. 263–266. Ong, C.K., Leakey, R.R.B., 1999. Why tree-crop interactions in agroforestry appear at odds with tree-grass interactions in tropical savannahs. Agroforestry Systems 45, 109–129. Ong, C.K., Odongo, J.C.W., Marshall, F.M., Black, C.R., 1992. Water use of agroforestry systems in semi arid India. In: Calder, I.R., Hall, R.H., Adlard, P.G. (Eds.), Growth and Water Use of Forest Plantations. John Wiley and Sons, Chichester, UK. Parton, W.J., Schimel, D.S., Cole, C.D., Ojima, D.S., 1987. Analysis of factors controlling soil organic matter levels in Great Plains grasslands. Soil Science Society of America Journal 51, 1173–1179. Puri, S., Kumar, A., Singh, S., 1994. Productivity of Cicer arietinum (chickpea) under a Prosopis cineraria agroforestry system in the arid regions of India. Journal of Arid Environment 27, 85–98. Saunders, D.A., Hobbs, R.J., Margules, C.R., 1991. Biological consequences of ecosystem fragmentation: a review. Conservation Biology 5, 18–32. Shanker, A., Newaj, R., Rai, P., Solanki, K.R., Kareemulla, K., Tiwari, K., Ajit, R., 2005. Micro-climate modifications, growth and yield of intercrops under Hardwickia binata Roxb. based agroforestry system. Archive of Agronomy and Soil Science 51, 281–291. Sharma, S.K., Gupta, R.K., 1981. Effect of seasonal lopping on the top feed (loong) production and growth of Prosopis cineraria (Linn.) Druce. Indian Journal of Forestry 4, 253–255. Singh, G., Gupta, G.N., Kupusamy, V., 2000. Seasonal variations in organic carbon and nutrient availability in arid zone agroforestry systems. Tropical Ecology 40 (2), 149–155. Singh, G., Bala, N., Rathod, T.R., Chouhan, S., 2003. Effect of adult neighbours on regeneration and performance of surface vegetation in shifting dune of Indian desert for the control of sand drift. Environmental Conservation 30, 353–363. Sokal, R.R., Rolf, P.J., 1981. Biometery, second ed. W.H. Freeman, New York. Tejwani, K.G., 1994. Agroforestry in India. Oxford & IBH Publishing Co. Private Ltd., New Delhi, pp. 10–21.
