On-farm evaluation of pigeonpea (Cajanus cajana L. Millsp.) -  neem (Azadirachta indica A. Juss.) agroforestry systems in the Deccan Plateau

The rainfed ecosystems are ecologically more fragile, characterized with inadequate and ill distributed rainfall, scarcity of soil moisture, low organic carbon, fluctuation of water table, higher evapotranspiration and soils highly susceptible to erosion. Therefore, the productivity of agriculture has declined or remained stagnant over years in spite of breakthroughs in crop varieties and crop husbandry. These days, tree based land use system is projected as a panacea for all problems in rainfed ecosystem. In fact, tree based land use system is an age old traditional practice followed by millions of farmers wherein farmers retain the useful trees on the farmland mainly to get fodder, fuel wood, timber, fruits and other minor forest products singly or together. That apart, trees are also extending many ecosystem services such as shade, protection and demarcation of land and congenial living habitat (Chittapur et al., 2017).
       
The traditional agroforestry land use systems are complex in structure and function with both complementary and competitive effects (Muturi et al., 2005). However, yet it is to be ascertained whether the traditional agroforestry systems are ecologically more desirable, economically viable and socially preferred. Often these land use system have significant species diversity and density variation of useful plants which are managed through planting or by selection and management of naturally regenerated plants on the farm land by the farmers (Schroth et al., 2004). However, today trees on farm land are under threat due to intensification of agriculture though trees are ecologically important to maintain favorable conditions for crop production through creation of healthy microclimatic conditions, enhanced soil fertility and reduced soil erosion besides providing food and fodder in the off season (Coulibaly et al., 2014). The irony is, the various ecological and economical benefits offered by the system are considered just as additional benefits and hence, loose the focus of the producer.
       
The farmers are more normally interested in the direct and immediate benefits viz., crop yield/fodder which would probably reduced in presence of trees though there accrue ecological and economical benefits. At times, they appear suspicious of any benefit that is not cognizable. Hence, under the circumstance understanding the nature of interaction between the tree and crop is of utmost importance to determine the management practices in agroforestry systems. Moreover characterization of component interaction in agroforestry system is important in determining the nature and extent of competition and complementarity between tree and crop.
       
Traditionally most preferred agroforestry system practiced by the farmers under rainfed ecosystem is scattered planting which is also called as parkland system and also farmers prefer the trees on the bunds and farm boundaries with varying density of 15 to 40 trees per hectare. The indigenous system with many different tree species were reported from many parts of the tropical countries of the world. Faidherbia albida in semi arid tropics in Westen Africa (Vandenbelt 1992), Vitellaria paradoxa and Parkia biglobosa in semiarid sub –Saharan Africa (Bremen and Kessler, 1995) and Prosopis cineraria with millets in Rajsthan (Tejwan, 1994), Kalaichelvi and Swaminathan (2002) reported medicinal perennial trees grown as boundary planting in agroforestry systems but the reports on the neem (Azadirachta indica) based indigenous agroforestry systems are rare though neem is predominant on farmlands in tropical India providing growers with various goods and services. With this background a study was undertaken to know the nature and extent of competition or otherwise of neem tree on associated pigeonpea [Cajanus cajan (L.) Millsp], crop in neem based agroforestry system under rainfed ecosystem. 

The study was under taken in Yadgir district of North eastern dry zone (Zone II) of Karnataka during 2016-17 on farmers’ fields on participatory mode. The climate of the region is dry semi arid with cool winters and dry hot summers. The average rainfall is around 750 mm and mean annual temperatures range from 18.6 to 32.5°C and mean elevation ranges from 350-680 m.  The soils are deep to very deep black soils and medium black soils in major areas while sandy loam and light textured soils are also found in some pockets. The major crops grown are pigeonpea, bengalgram, soybean, groundnut, greengram, pearlmillet (bajra), sorghum, sunflower and safflower under rainfed condition, and chilli, cotton, sugarcane, paddy and maize under irrigated condition.
 
       
Bund, boundary and scattered planting were the major traditional agroforestry systems practiced by the farmers in the rainfed ecosystem in the region with neem, babul, ber, tamarind and mango, however neem found to be dominant tree in terms of frequency of occurrence and density mainly due its multiple utility and adoptability.  Hence, for the study purely neem based agroforestry systems were selected and 4 farmers in each system with a plot size of one hectare were selected and as the plots were located in different locations average of 4 farmers without trees representing the locations were taken as control for comparison. The pigeonpea is the major field crop grown in the region under rainfed ecosystem. The system components, cropping pattern and other details are given in the Table 1. Further, the soil particle analysis was carried out to know the genetic difference in the soil characteristics and the soil of the study area belonged to sandy clay loam texture (sand - 66.19 to 74.19 %, silt - 5.10 to 9.04 % and clay -  20.04 to 28.72 %  at 0-15 cm depth).
 

       
The productivity of the pigeonpea - neem based agroforestry system was assessed in terms of yield and biomass. The biomass and grain yield of pigeonpea was recorded with a net plot size of 6 m x 5 m and the sample plots were laid out randomly with 3 replication at a desistance of 5 m, 10 m and 15 m from the tree line and total plots laid were 9 in each farmers field whereas in control 9 plots of 6 m x 5 m were laid out randomly in the entire field. The mature crop was harvested at ground level, later grain and haulm were separated. Further, they were dried and weighed and were computed, averaged and extrapolated to per ha basis. The harvest index of the crop was calculated with the following equation (Donald, 1962).
               

 

       
The tree biomass was calculated by non-harvesting method by taking the measurements of girth and height of the standing trees and then volume was calculated. 20 trees were randomly selected in each agroforestry system and mean girth, height and volume of tree was worked out. The tree height was measured with help of Ravi altimeter from the ground level to the tip of the tree and expressed in meter (Chaturvedi and Khanna, 1981), tree girth at breast height (1.37 m) from the ground level was measured, and crown spread was measured by taking the width of the crown in North-South and East-West direction by projecting the perimeter of the crown vertically to the ground and calculated by using the following formula (Chaturvedi and Khanna, 1981).

 
                             Crown spread = (D₁+D₂)/2,

where D₁ - crown width in North-South direction 
            D₂ -crown width in East-West direction.
 
Further, the tree volume was calculated with the following formula (Chaturvedi and Khanna, 1981).

 
 

 
where ‘g’ is girth at breast height (cm), ‘h’ is height in (m), π is 3.4125
       
The composite soil sample collected at a depth of 0 - 15 cm analysed for physic chemical properties following standard procedures (Table 2). The data was analyzed for univariate analysis and one way ANOVA at a significance level of 0.05 and further, to know the difference between the means post hoc test was performed using Duncan test at significance level of 0.05 by using SPSS (Statistical Package for Social Science) version 20.0.
 

To know the tree crop interface the crop yield and biomass were measured at different distances from the tree line. Non- significant differences were observed in grain yield, biomass yield and harvest index among the different agroforestry systems (Table 3).  However, significantly lower grain (1192 kg ha^-1) and biomass (5563 kg ha^-1) yields were recorded at 5 m distance away from the tree line compared to a distance of 10 m (1450 and 6942 kg ha^-1, respectively) and 15 m (1449 and 6720 kg ha^-1, respectively) away from the tree line. However, differences in harvest index with respect distance from the tree line were not significant.
       

 
Significantly higher grain yield (1520 kg ha^-1) was observed in control without proximity of trees compared to the agroforestry systems; the latter accounted an average of 11 per cent reduction in arable crop yield.  Whereas, the grain yield did not differ significantly among the various agroforestry systems though numerically higher yield was observed in boundary planting (1378 kg ha^-1) followed by bund planting (1370 kg ha^-1) and scattered planting (1357 kg ha^-1).  Similar was the situation with regard to biomass and harvest index (Table 4). However, numerically higher values for grain and biomass yield were observed in control in comparison to agroforestry systems.
 

       
The lower grain and biomass yield near the tree line could be ascribed to the competition offered by trees for light, moisture and nutrients with field crops through shading and root extension in to crop land. The finding highlighted the fact that the competition effect was more near the tree line compared to away from the tree line. In the present study, competition effect of trees was registered up to 5m distance from the tree line. That means while selecting the tree species for agroforestry systems light crown and deep rooted species should be selected or the timely pruning management are imposed for shoot (branches) and roots to reduce the competition effect on the crops to desirable levels (Gaddanakeri, 1991). The results are in line with Muthuri et al., (2005) who observed reduction of grain, biomass yield and harvest index of maize in agroforestry system over sole maize and reported nearly 36% of grain yield reduction close to the trees rows at a distance up to 5 m. Thus in the present investigation the lower grain yield in the agroforestry compared to the sole crop is mainly attributed to the competition effect of neem trees for light, moisture and nutrients with field crops.
       
Further, numerically higher grain yield was recorded in boundary planting followed by bund planting and scattered planting this suggests that boundary planting has minimal effect on the crop compared to the scattered planting where trees are in crop/land and, therefore, had higher competitive effect all round the tree whereas in bund planting it would be on two sides and in boundary planting it would be on only one side of the crop. The findings are in line with Chauhan et al., (2015).
 
Productivity of neem in agroforestry systems
 
Growth performance of neem trees in different agroforestry systems was measured to know the competition/complementary effect of crop on neem. Significantly higher tree height was recorded in bund planting (7.61 m) followed by boundary planting (7.38 m) and scattered planting (6.46 m) whereas the crown spread recorded was significantly lower in bund planting (6.27 m) followed by scattered planting (6.38 m) and boundary planting (6.68 m) (Table 5). However, there existed non-significant differences in girth, volume and biomass of neem trees grown in different agroforestry systems.
 
  
 
The higher tree height and lower crown spread was recorded in bund planting. This suggested that the trees grown on the bund are better managed through pruning that made the trees to grow straight but tall and cause less shading effect, in turn this also reduced the overall crown spread of the trees. On the other hand, scattered planting recorded lower tree height which may be attributed to absence of competition from perennial component particularly in the off season (winter/summer) as found in bund or boundary system and consequently the trees in scattered planting need not have to outgrow any fellow tree except the under-grown vegetation hence it grew less tall compared to trees in other systems. Besides, high frequent intensity of pruning of trees inside the field might have further suppressed the growth of the trees. However, overall biomass did not differ significantly mainly because the management techniques applied in bund planting and scattered planting were the same. Whereas, thick planting of trees in the boundary might have reduced the biomass on individual plant basis, however, at the age at which the observation made, there occurred non-significant differences in the biomass of the trees grown in different agroforestry systems, which probably would not be the case as the trees grow and years pass. The findings are in line with Avinash and Ansari (2013).
       
Significantly lower soil bulk density was recorded in boundary planting (1.50 Mgm^-3) followed by bund planting (1.59 Mgm^-3) and scattered planting (1.60 Mgm^-3) in comparison to control (1.64 Mgm^-3) whereas non-significant differences occurred with regard to soil moisture holding capacity, while agroforestry system had  higher  soil moisture holding capacity compared to the control (without trees) (Table 6). Significantly higher porosity was observed in boundary planting (43.49%) followed by bund planting (39.91 %) and scattered planting (39.81%) than under control (37.45%), whereas, non-significant differences occurred in dispersion ratio, and the trend revealed lower values in agroforestry systems compared to control.
 
  
 
Improvement in soil bulk density, higher soil moisture holding capacity and porosity in agroforestry systems could be attributed to the addition of organic matter through leaf by perennial component mainly composed of trees. Among the agroforestry systems significantly lower soil bulk density and higher porosity in boundary planting was due to higher density of trees in the boundary planting which in turn added higher organic matter. The lower dispersion ratio in agroforestry systems could again be ascribed to the organic matter added by tree component in these systems as against  control which acts as cementing agent for developing stable aggregates which in turn improved the soil structure.
       
Soil pH, EC and organic carbon and CEC did not reveal significant differences (Table 7).  Numerically higher organic carbon and CEC were recorded in boundary planting (0.74%, 14.33 cmol (p+) kg^-1) followed by bund planting (0.65%, 14.61 cmol(p+)kg^-1) and scattered planting (0.66, 12.26 cmol(p+)kg^-1) compared to control (0.48, 10.73 cmol(p+)kg^-1). Further, significantly higher soil available N was recorded in boundary planting (319.56 kg ha^-1) followed by scattered planting (287.59 kg ha^-1) and bund planting (276.50 kg ha^-1) and the lowest soil available N was observed in control (208.90 kg ha^-1). Whereas, P and K varied non-significantly, with numerically values with agroforestry systems compared to the control.
 
 
 
Numerically lower pH and EC values with agroforestry systems could be attributed again to the organic matter in these systems which upon oxidation produced various organic acids which dissolved the salts that lead to loss of basic cations through leaching. Similarly, higher available N, P and K in agroforestry systems could be attributed to contribution of these elements on decomposition of organic matter which were already high, besides to management practices. Organic matter is the important source of both N and P. Trees on the farm land enhance the soil fertility through constant addition of organic matter (Kohil et al., 2007). The improvement of soil fertility depends on the amount and pattern of litter fall which varies with the species, age, growth, tree density, phenology, intercrops and season (Bhardwaj et al., 2001). In addition to this, diammonium phosphate fertilizer was the most commonly used fertilizer in crop production by farmers and more so to pigeonpea. Long time application of this form of P has led to gradual build in the soil. Application of P often exceeds P requirement of the crops cultivated. The findings are in line with Adewole and Adeoye (2014). 
               
Thus, the extent of competition with pigeonpea was of higher order near the neem tree line suggesting necessary management of tree and crop to overcome the negative effect of trees on associated crop. Pruning management especially during cropping season and crop management like selective fertilization would be helpful. The investigation also enlightened that the tree growth grown in different patterns and density in the farm land also get affected. However, merely on yield based productivity assessment of the system is incomplete and therefore economical and ecological benefits of neem trees must be considered while evaluating the productivity of the systems which would be meaningful for promotion of tree land use system.