Forest Ecology and Management 310 (2013) 612–622

Contents lists available at ScienceDirect

Forest Ecology and Management

journal homepage: www.elsevier .com/ locate/ foreco

Effects of rainwater harvesting on plant growth, soil water dynamicsand herbaceous biomass during rehabilitation of degraded hillsin Rajasthan, India

0378-1127/$ - see front matter � 2013 Elsevier B.V. All rights reserved.http://dx.doi.org/10.1016/j.foreco.2013.09.002

⇑ Corresponding author. Tel.: +91 291 2729150; fax: +91 291 2722764.E-mail address: singh_g_dr@yahoo.co.in (G. Singh).

G. Singh ⇑, D. Mishra, K. Singh, R. ParmarDivision of Forest Ecology, Arid Forest Research Institute, New Pali Road, Jodhpur 342005, India

a r t i c l e i n f o

Article history:Received 19 January 2013Received in revised form 31 August 2013Accepted 2 September 2013

Keywords:AfforestationDegraded hillsPlant growthHerbaceous biomassSoil and water conservation

a b s t r a c t

Degraded hills can be restored by afforestation and conserving soil and water through rainwater harvest-ing. Three slope categories viz. <10%, 10–20% and >20% of a degraded hill were treated with rainwaterharvesting (RWH) structures: Contour trench (CT), Gradonie (GD), Box trench (BT) and V-ditch (VD)including a control with a view to rehabilitate it by conserving and minimizing gradient in soil waterand nutrients. Soil water content (SWC), height and collar diameter of Acacia catechu, Azadirachta indicia,Emblica officinalis, Holoptelia integrifolia and Zyziphus mauritiana planted in August 2005, herbaceous bio-mass and photosynthetically active radiation (PAR) interceptions by tree, vegetation and tree-vegetationcombine (PARintT/V/C) were monitored for suitability of RWH devices and tree species in rehabilitation.SWC decreased from December to June and it was linearly related to rainfall and vegetation height. PARint

by tree, vegetation, and tree-vegetation combine were 30.0%, 54.6% and 84.6%, respectively and helpedconserve soil water. SWC, plant and vegetation growth and PARintT were lowest (P < 0.05), whereas her-baceous biomass and PARintV were highest in 10–20% slope. Vegetation height and SWC were linearlyrelated to biomass indicating improvement in micro-climate and herbaceous growth. Highest SWC in<10% slope promoted plant growth and mean annual increment (MAI) in height and collar diameter,which enhanced PARintT and PARintC. These variables were highest in CT/BT treatments and lowest in con-trol plots. Characteristic root distribution of Acacia catechu and A. indica promoted growth in V-ditch,whereas E. officinalis, Z. mauritiana and H. integrefolia performed best in CT treatment. RWH enhancedherbaceous biomass between 22.4% and 60.7% over control. Conclusively, VD/GD structures found bestfor growth of herbaceous vegetation as well as A. catechu and A. indica plants, whereas CT/BT structuresfavoured growth of other tree species. Rainfall influenced SWC, but RWH helped conserve soil and water,promoted plantation and herbaceous growth and facilitated restoration process, and may be promoted torestore degrading lands.

� 2013 Elsevier B.V. All rights reserved.

1. Introduction

Growing demands for fodder and fuelwood are leading causefor over-exploitation and vegetation removal and are acceleratingland degradation throughout the world (Gao et al., 2011). Severityof the degradation is relatively more in hill regions, which is animportant landscape that maintains hydrological cycle, vegetationstatus and people livelihood (Runhaare et al., 1997; Rahman,2011). The status of degradation of Indian hills is also on similarline and many hillocks of Aravalli hills-an oldest mountain systemin India, are exposed and devoid of vegetation. Afforestation iscommon approach for restoration and biodiversity conservation(Cao, 2008; Cao et al., 2009a; Cao et al., 2011), eco-environmental

improvement (Cao, 2011; Cao et al., 2011) and people livelihood(Cao et al., 2009b; Cao 2011). However, the establishment of vege-tation on these degraded hills is constrained by the inadequateavailability of soil as well as water (Li et al., 2008), whereas lowand irregular rainfall affects plant growth too (Barron et al.,2003). In such limited resource availability, the only option forincreasing biomass production is to increase the water availabilitythrough conservation of soil and water. Adoption of improvedwater conservation and harvesting technologies may contributesoil water storage, improve soil nutrients mobility, and supportsa higher number of plants and biomass production (Gowinget al., 1999). Further, water harvesting supports flourishing agricul-ture in many dry areas (Suleman et al., 1995; Faroda et al., 2007). Ahuge literature exists on rain water harvesting throughout theworld (Prinz, 2001; Mati, 2005; Waterfall, 2006), but much interestin this was developed predominantly because of migration of more

G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622 613

and more people to live and utilize the meagre resources in dryareas. Many site-specific rainwater harvesting (RWH) structureshave been designed to address the soil and water conservationissues and to improve crop yield (Li et al., 2000), plant growth(Gupta, 1995) and forage production (Jia et al., 2006).

Sustainability and eco-hydrological functioning of various RWHtechniques depend upon the timing, number of rain days and theamount of rainfall (Cohen et al., 1995). There is a need to under-stand the complex interactions between ecology and hydrologyinvolving rainwater harvesting micro-catchments and their influ-ences on availability of soil water and corresponding improvementin plantation growth and vegetation status. Further research onsoil and water conservation and its role on biological diversityimprovement have also been emphasized in a conclusive reviewby Vohland and Barry (2009). Information on ecological and hydro-logical interaction may determine the resource use and its influ-ence on plant growth (Ludwig et al., 2005; Yu et al., 2008).Further, rainwater harvesting devices differ in their effects depend-ing upon slope gradient and characteristics of the planted species-rooting pattern and soil water use from soil profile. Experimentsinvolving different RWH devices, slope categories and trees speciesmay provide suitable combinations useful in restoring degradedforestlands and help in developing and extending water-adaptiveforest management practices (Yanhui et al., 2012). The increasein water yield through micro-catchment RWH may enhance plan-tation growth and promote herbaceous vegetation, which helprehabilitate degraded hills/wastelands by conserving soil andwater and generating biomass for local benefits (Zhuang, 1997;Cao, 2011).

Therefore, objectives of this study were: (i) to study the effectsof rainfall pattern and different rainwater harvesting structures onsoil water storage; (ii) to find out suitable combination of RWH de-vice and tree species in a particular slope by observing the effectsof soil water on the growth of plantation and herbaceous vegeta-tion (Cao et al., 2010); and (iii) to observe role of herbaceousgrowth on soil water use and conservation.

2. Materials and methods

2.1. Site characteristics

The study was carried out at the junction of Aravalli Hills andMalwa Plateau covering an area of about 17 ha, which spread over23� 250 27.000 N to 23� 250 43.400 N latitudes and 74� 240 00.500 E to74� 240 23.100 E longitude. Altitude of the area ranged between 248to 320 msl. The site is located 17 km south-west of Banswara (23�320 28.200 N and 74� 260 30.300 E), Rajasthan, India. Air temperaturevaried from 4 �C in January to 42 �C in May. The mean minimum an-nual temperature during experimental period ranged from 18.7 �Cin 2008 to 20.1 �C in 2006, whereas mean maximum annual temper-ature ranged from 32.8 �C in 2008 to 34.3 �C in 2010 (Table 1).Average annual rainfall from 1993 to 2004 was 960 ± 352 mm(mean ± standard deviation) with 54 numbers of rainy days. Rainfall

Table 1Average annual temperature, number of rain days and cumulative monthly and annual ra

Year Average annual temp. (�C) Rain days Rainfall durin

Max. Min. June J

2005 33.2 19.9 42 82.6 22006 33.1 20.1 63 148.4 62007 33.5 19.0 44 102.0 22008 32.8 18.7 29 101.0 12009 33.9 19.1 30 69.0 42010 34.3 20.0 39 3.0Average – – 41.2 84.3 3

varied from 562.5 to 2266.0 mm during 2005 to 2010 with averagevalue of 1073 ± 626 mm (Table 1). Slope gradients of the hills andfoot hills in the study area varied from 3% to 53%. These slopes werecategorized into steep (>20%), medium (10–20%) and gentle (<10%)slopes. The surface of steep slope was covered with crystalline grav-els and pebbles of varying size with randomly growing Lantana ca-mara L. and sporadic Themada quadrivelvis (L.) Kuntze and Apludamutica L. grasses. Medium slope had light textured sandy loam soilsof shallow depth and was mostly covered by Prosopis juliflora (Sw.)DC. and occasional L. camara shrub and Aristida funiculata Trin. &Rupr. and Heteropogon contortus (L.) P. Beauv. ex Roem. & Schult.grasses. Lantana camara and P. juliflora are invasive to this region.Soils in gentle slope were loamy to clay loam in texture and shallowto deep in soil depth. This area was dominantly covered by P. julifloraand L. camara bushes and the grasses like Dichanthium spp. andCenchrus spp. Soil pH was slight acidic to neutral in reaction (6.34–7.02). Average SOC, available NH4–N, NO3–N and PO4–P of the sitewere 0.76%, 22.15 mg kg�1, 2.50 mg kg�1 and 4.51 mg kg�1, respec-tively (Singh, 2012).

2.2. Experimental design

The experiment was laid in a complete randomized block de-sign in five replications because of hillocks of varying height andaspects associated with different drainage lines. Because of thismajor emphasis was given to have a plot of equal size though varyin shape to adjust between hillslopes and the drainage line. Sev-enty-five plots of 700 m2 area were laid in the slope categories of<10%, 10–20% and >20% distributed in about 17 ha area coveringalmost all aspects. Each plot was separated by individual boundaryof trench (2025 cm2 cross section area, 45 cm � 45 cm) cum bundto prevent water flow into the plots from other areas or plots anddivert the flowing water toward the drainage line. Four rainwaterharvesting structure viz. contour trench (CT), gradonie (GD),Box trench (BT) and V-ditch (VD) of 30 running meter length wereprepared to harvest rainwater in the plots (Supplementary Fig. 1).In addition there were control plots without any rainwater har-vesting structures (Singh, 2009). Contour trenches were excavatedat different contour levels to conserve the run-off water and thetrenches were 45 cm � 45 cm in cross section. Box trenches were2 m length trenches excavated intermittently at different contourlevels and 15 in numbers with cross section area similar to thatin the CT. Gradonie and V-ditches were across the contour and1800 cm2 cross section area, but differences were only in verticalcut of 30 cm height. In V-ditch the vertical cut was downside ofthe slope in VD, whereas in gradonie ditch the cut was upside ofthe slope (to reduce velocity of surface run-off water). The exca-vated soil was always kept downside of the dugout. A mixed plan-tation of Acacia catechu (L.f.) Willd, Azadirachta indica A. Juss.,Emblica officinalis Gaertn., Holoptelia integrifolia (Roxb.) and Zizy-phus mauritiana Lam.were carried out in August 2005. There were35 seedlings of above-mentioned tree species planted in a45 cm � 45 cm � 45 cm pit size at the rate of 500 plants per ha.

infall (mm) near the experimental site, Banswara, Rajasthan, India.

g monsoon (June–October) and annual (mm)

uly Aug. September October Total Annual

90.0 169.1 483.5 – 1025.2 1026.748.6 630.3 489.1 44.0 1960.4 2266.032.0 565.0 254.0 – 1153 1391.092.3 129.5 74.0 – 496.8 562.557.0 333.0 – – 859 859.093.0 427.0 71.0 5.0 599 636.018.8 375.7 228.6 8.2 1015.6 1123.5

614 G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622

Three slopes, five rainwater harvesting treatments and five replica-tions (three slopes � five treatments � five replications) lead to 75experimental plots.

2.3. Soil sampling and analysis

Soil samples for texture and initial nutrient analysis were col-lected in 0–40 cm soil layer in June 2005. Collected soil sampleswere dried and passed to a 2 mm sieve for separation of graveland soil. For soil water content (SWC) determination, collected soilsamples from the vicinity of the plants were put immediately intopolyethylene bag to avoid water loss during transport. Soil watercontent was estimated by oven drying of the sample at 110 �Cfor a constant weight and calculated using following equation:

Soil water content ð%Þ ¼ ðMass of moist soils

�Mass of dry soils� 100=Mass of dry soils

An average data of 12 observations (recorded during December2005 to June 2011), six each for December and June have been usedhere to observe the effects of natural slope and RWH treatments aswell as the effects of the vegetation growing in the plots.

2.4. Observation recording

Height and collar diameter (15 cm above from the soil surface)of trees were recorded in June (before monsoon) and December(after monsoon) each year to monitor seasonal growth i.e., springand monsoon, respectively (total 12 observations). Mean annualincrements (MAI) in height and collar diameter were calculatedby dividing December 2010 data by 5 after subtracting the dataof December 2005. To observe the rooting pattern and biomassaccumulation minimum three plants (one from each slope) of aver-age growth were harvested between 2010 and 2011. Roots wereexcavated up to 0.5 cm diameter to study the root length and thelateral distribution of the roots of these species. Fresh weightwas recorded immediately after excavation, whereas dry bio-masses of the shoot and root were recorded after oven drying ofthe sample at 80 0C. For recording of herbaceous layer productivityabove ground living vascular vegetations were clipped from threeplaces (three sampling plots of 1 m2 size) in the plot and sortedto species in October each year (2005–2010). Dry mass of individ-ual species was recorded using electronic top loading balance afterdrying the sample at 80 �C in an electric oven to a constant weight.The summed dry mass of all the aboveground living vascular plantsfrom the harvest was averaged for a plot and considered as dryherbage biomass. Height of the herbaceous vegetation in a sam-pling plot was calculated as follows:

H ¼X

nihi=N

where H is height of the herbaceous vegetation, ni is the populationof ith species, hi is the average height of ith species and N is the pop-ulation of all species in the sampling plot. Photosynthetically activeradiation (PAR) was measured at above-mentioned sampling plotsin October 2010 using a portable CI-301 and CO2 gas analyzer tomonitor the effects of canopy closure of planted/regenerated treeson herbaceous growth and biomass production. PAR was measuredoutside the plantation (as control), top of herbaceous vegetationand inside the herbaceous vegetation (i.e., at soil surface) to moni-tor total PAR, reduction in PAR reaching to the vegetation (intercep-tion due to tree, PARintT) and reduction in PAR reaching to soilsurface (interception due to both tree and herbaceous vegetationcombined, PARintC), respectively. PAR interception due to vegetation(PARintV) was calculated by subtracting PARintT from PARintC.

2.5. Statistical analysis

Data were analyzed statistically using SPSS version 8.0 statisti-cal package. Since data on height and dry biomass of herbaceousvegetation were recorded six times (in October month each year),these data were analyzed using Repeated Measure ANOVA consid-ering year as the ‘Tests of within- subjects effects’ and slope andRWH treatments as the ‘Tests of between-subjects effects’. Plantsurvival, growth (height and collar diameter) and mean annualincrement (separately for each species) data were analyzed usinga two-way ANOVA. Above-mentioned parameters were the depen-dent variables and slope and RWH treatments were the fixed fac-tors. Percent soil water content (SWC) data of December andJune were also analyzed using a two-way ANOVA after square roottransformation to reduce heteroscenesdity (Sokal and Rohlf, 1981).Duncan Multiple Range Tests (DMRT) was applied to group homo-geneous subsets of slope and rainwater harvesting treatments atthe P < 0.05 levels. To obtain relations among rainfall, SWC, growthand biomass of herbaceous vegetation, and plant growth, Pearsoncorrelation coefficient were calculated. Regression relations wereobserved to related rainfall with SWC, and herbage biomass withits height as well as SWC.

3. Results

3.1. Soil water content

Average (six observation in each) soil water content (SWC) wasgreater (P < 0.01) in December than in June (Table 2). SWC washighest (P < 0.05) in <10% slope and lowest in 10–20% slope in boththe months. As compared to SWC in 10–20% slope, the increases inSWC in >20% and <10% slopes were 24.3% and 58.9% in Decemberand 21.1% and 29.4% in June, respectively. Among the RWH treat-ments, SWC was highest (P < 0.05) in BT treatments in December,whereas it was highest in CT treatment in June. The lowest SWCwas in the control. Average increase in SWC was 17.7% in VD treat-ment to 28.6% in BT treatment in December. In June BT and CTtreatments exhibited 15.0% and 16.7% higher SWC as comparedto the control. Soil water depletion during December to June was74% (P < 0.05) in <10% slope to 67.4% in10–20% slope. It was lowestin the control (65.8%). Among RWH treatments the order of soilwater depletion was CT < BT < VD < GD.

3.2. PAR interception

Average photosynthetically active radiations (PAR) ranged from1238 to 1397 l mol m�2 s�1 (without any interception), whereas itranged from 905 to 1020 l mol m�2 s�1 at top of vegetation (PARinterceptions due to tree, PARintT), and 190–258 lmol m�2 s�1 atsoil surface (interception due to both tree and herbaceous layer,PARintC) in October 2010. The PARintT and PARintC were 30.0% and84.6%, respectively, whereas interception due to vegetation(PARintV) was 54.6% across the slope and RWH treatments. Amongslopes, PARintT and PARintC were highest (P < 0.05) in <10% slope,whereas PARintH was highest in 10–20% slope (Table 3). While con-sidering RWH treatments, PARintT was highest in BT and lowest inGD treatments. The values of PARintV and PARintC were highest inGD and CT treatment, respectively, whereas these values were low-est in the control plots.

3.3. Growth and biomass of herbaceous vegetation

Both growth (height) and dry biomass of herbaceous vegetationvaried temporally between the years (P < 0.01). Vegetations weretallest in 2010, whereas dry biomass was highest in 2008

Table 2Soil water content influenced by slope gradient, rainwater harvesting and afforestation in degraded hills in Rajasthan, India. Values are mean with ±SE of five replications.

Slope RWH treatment Average soil water content (%) Soil water depletion (%)

December June During December to June

<10% Control 4.69 ± 0.49 1.55 ± 0.11 65.07 ± 5.30Contour trench 5.75 ± 0.45 1.55 ± 0.23 73.58 ± 1.82Gradonie 4.85 ± 0.60 1.19 ± 0.08 74.69 ± 1.82Box trench 6.38 ± 0.37 1.55 ± 0.24 76.09 ± 2.58V-ditch 6.15 ± 0.38 1.19 ± 0.08 80.46 ± 1.24

10–20% Control 2.67 ± 0.54 0.89 ± 0.16 64.08 ± 5.06Contour trench 4.03 ± 0.60 1.25 ± 0.18 68.90 ± 1.41Gradonie 4.16 ± 0.69 1.08 ± 0.17 73.32 ± 1.89Box trench 3.63 ± 0.41 1.21 ± 0.13 65.75 ± 4.01V-ditch 3.02 ± 0.53 1.04 ± 0.16 64.76 ± 3.61

>20% Control 3.98 ± 0.53 1.16 ± 0.16 68.30 ± 4.11Contour trench 4.44 ± 0.62 1.41 ± 0.13 66.84 ± 3.31Gradonie 4.35 ± 0.50 1.34 ± 0.09 67.82 ± 3.57Box trench 4.57 ± 0.71 1.38 ± 0.09 68.25 ± 2.70V-ditch 4.39 ± 0.53 1.29 ± 0.10 69.49 ± 2.58Mean 4.47 ± 0.18 1.27 ± 0.04 69.83 ± 0.93

F value of two-way ANOVASlope (S) 16.67** 5.60** 6.22**

Treatment (T) 1.65NS 1.65NS 1.69NSS � T 0.81NS 0.74NS 1.39NS

NS, not significant.**, Significant at p < 0.01.

Table 3Interception in photosynthetically active radiations (PAR) due to tree (PARintT) and herbaceous vegetation (PARintV) and tree and herbaceous vegetation combine (PARintC)influenced by rainwater harvesting and afforestation during rehabilitation of degraded hills. Values are mean with ±SE of five replications.

Slope RWH treatment Light interception (%)

PARintC PARintT PARintV

<10% Control 87.49 ± 6.33 45.22 ± 8.78 42.26 ± 6.66Contour trench 94.00 ± 0.33 31.59 ± 9.50 62.41 ± 9.77Gradonie 91.43 ± 1.17 31.94 ± 10.92 59.49 ± 11.73Box trench 85.62 ± 6.63 42.88 ± 5.76 42.74 ± 6.94V-ditch 90.99 ± 1.77 42.14 ± 11.69 48.85 ± 10.35

10–20% Control 61.49 ± 12.95 18.73 ± 8.51 42.75 ± 11.18Contour trench 84.65 ± 4.24 20.37 ± 6.50 64.27 ± 5.43Gradonie 85.44 ± 3.15 10.55 ± 3.65 74.90 ± 2.76Box trench 85.43 ± 5.14 31.30 ± 6.95 54.14 ± 6.88V-ditch 75.40 ± 6.97 17.34 ± 3.78 58.05 ± 4.52

>20% Control 82.64 ± 6.71 29.94 ± 8.38 52.69 ± 4.94Contour trench 85.93 ± 3.31 34.74 ± 7.75 51.19 ± 6.92Gradonie 82.30 ± 6.20 30.15 ± 8.33 52.16 ± 5.98Box trench 90.48 ± 1.49 36.16 ± 7.37 54.33 ± 6.45V-ditch 86.35 ± 4.50 27.08 ± 9.43 59.28 ± 8.23

84.64 ± 1.58 30.01 ± 2.19 54.63 ± 2.04

F value of Two-way ANOVASlope (S) 5.23** 7.06** 1.29NSRWH treatment (T) 1.83NS 0.95NS 2.22NS*

S � T 1.14NS 0.38NS 0.94NS

RWH, rainwater harvesting.NS, not significant.*, significant at P = 0.078.**, significant at P < 0.01.

G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622 615

(458.8 g m�2) across the slopes and RWH treatments. Vegetationswere taller (P < 0.05) in >20% slope throughout the study(Fig. 1a). Shortest vegetation was either in 10–20% slope or in<10% slope (Supplementary Table 1). While considering RWHtreatments, vegetation was taller in GD treatment in 2005 to VDtreatment in 2010 (Fig. 1b). Lowest vegetation growth was in thecontrol. Significant (P < 0.05) year � slope, year � RWH treatmentand year � slope � RWH treatment interactions indicated depen-dency of herbaceous biomass on rainfall and efficiency of waterconservation by RWH structures in different slopes. DMRT showed

highest (P < 0.05) herbaceous biomass in <10% slope in 2005, in10–20% slope in 2006 and 2008, and in >20% slope in 2007, 2009and 2010 (Fig. 1c). Average herbaceous biomass was highest in10–20% slope (i.e., 383.4 g m�2) and the biomass reduced by12.0% in <10% slope and by 7.1% in >20% slope. Herbaceous bio-mass was highest (P < 0.05) in VD treatment in most of the years(Fig. 1d). As compared to the lowest value (viz., 293.8 gm�2) inthe control (Supplementary Table 2), the increases in herbaceousbiomasses were 15.9%, 21.1%, 25.2% and 70.5% in CT, GD, BT andVD treatments, respectively.

Fig. 1. Effects of slope gradient (left panels), and rainwater harvesting treatments (right panels) on herbaceous vegetation (top panels) and dry biomass (bottom panels)production in different years. Legend for both left panels and right panels are same. Error bars are ±SE of five replications. C, control; CT, contour trench; GD, gradonie; BT, boxtrench and VD, V-ditch rainwater harvesting structures.

616 G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622

3.4. Plants survival

Average survival of the plants ranged from 59.2% for H. integri-folia to 84.2% for Z. mauritiana. Survival did not differ (P > 0.05) be-tween slopes and TWH treatments both for A. indica, A. catechu, E.officinalis and Z. mauritiana. Survival of H. integrifolia plants differed(P < 0.05) between slopes. Slope � RWH treatment interaction wassignificant (P < 0.05) only for A. catechu, which showed highestplant survival in VD treatment of <10% slope. Across the treatmentsfor slopes, survival was highest in >20% slope for A. catechu, in 10–20% slope for E. officinalis and Z. mauritiana plants, and in <10%slope for A. indica and H. integrifolia plants (Table 4). While consid-ering the RWH treatments, the plant survival was lowest in thecontrol plots for all species. Highest survival of E. officinalis and Z.mauritiana was in BT, A. catechu in V-ditch and A. indica and H.integrifolia in CT treatments (see Table 4).

3.5. Plant growth

Seedling height and collar diameter of all species did not differ(P > 0.05) between slopes as well as RWH treatments in December2005 (Supplementary Table 2). In June 2011, plants of H. integrifoliawere shortest in height, whereas the plants of Z. mauritiana wereleast in diameter. Plant of A. catechu showed greater height andcollar diameter as compared to the plants of other species (Table 5).

Plants height and collar diameter of all species were highest in<10% slope (across the RWH treatments). However, collar diameterof A. catechu and height of H. integrifolia were greater in >20% slopethan in other slopes. Both height and collar diameter of A. indicaand collar diameter of other species showed a decreasing trendwith increase in slope gradient (<10% to >20%). Height of A. catechu,E. officinalis, H. integrifolia and Z. mauritiana were greater in >20%slope than in 10–20% slope. Among RWH treatments, plants ofall species were shorter and thinner in control plots than in theother treatments. Height and collar diameter of A. catechu and A.indica plants were greater (P < 0.05) in VD treatments as comparedto the other RWH treatments. Emblica officinalis, H. integrifolia andZ. mauritiana performed the best in CT treatments for both heightand collar diameter. Significant (P < 0.05) slope � RWH treatmentinteractions showed highest height growth of A. catechu and H.integrifolia VD and CT treatments of <10% slope, respectively.

3.6. Mean annual increments

Mean annual increment (MAI) in height and collar diameterwere 35.0 cm and 0.50 cm, respectively for the site. MAI in heightranged from18.2 cm for H. integrifolia to 52.4 cm for A. catechu,whereas MAI for collar diameter ranged from 0.54 cm for Z. mauri-tiana to 0.98 cm for A. catechu. MAI of both height and collar diam-eter was highest (P < 0.05) in <10% slope (Supplementary Table 4).

Table 4Percent survival of 71 months old plants of different tree species influenced by natural slope gradient and rainwater harvesting devices in degraded hills in Rajasthan, India.Values are mean with ±SE of five replications.

Slope RWH treatment A. catechu A. indica E. officinalis H. integrifolia Z. mauritiana

<10% Control 91.7 ± 8.33 86.0 ± 4.00 74.9 ± 10.52 83.3 ± 16.67 75.8 ± 15.83Contour trench 86.7 ± 6.67 90.0 ± 10.00 84.9 ± 8.13 100.0 ± 0.00 73.6 ± 3.57Gradonie 85.7 ± 14.28 68.8 ± 13.77 91.1 ± 8.89 69.1 ± 2.38 75.6 ± 11.93Box trench 75.0 ± 25.00 78.7 ± 8.79 84.4 ± 6.46 40.0 ± 0.00 88.9 ± 5.56V-ditch 100.0 ± 0.00 78.0 ± 13.57 87.0 ± 8.09 75.0 ± 25.00 88.2 ± 3.56

10–20% Control 83.3 ± 16.67 61.5 ± 14.17 73.3 ± 13.73 65.5 ± 12.74 80.0 ± 6.39Contour trench 79.2 ± 12.50 70.8 ± 17.18 78.3 ± 7.55 64.6 ± 17.80 87.8 ± 9.68Gradonie 72.1 ± 9.87 73.3 ± 13.33 98.5 ± 1.54 61.1 ± 20.69 48.3 ± 1.67Box trench 81.7 ± 9.28 76.8 ± 13.48 88.1 ± 6.14 47.7 ± 5.34 90.9 ± 3.97V-ditch 90.6 ± 9.38 80.6 ± 12.53 82.1 ± 9.57 53.2 ± 9.54 89.3 ± 6.86

>20% Control 60.5 ± 2.63 45.8 ± 4.17 82.6 ± 11.14 75.0 ± 25.00 71.7 ± 10.87Contour trench 88.9 ± 4.96 87.0 ± 12.50 69.6 ± 9.76 68.3 ± 4.41 82.6 ± 5.84Gradonie 91.5 ± 5.89 75.0 ± 25.00 71.6 ± 10.37 56.4 ± 3.46 92.9 ± 7.14Box trench 79.5 ± 8.95 83.3 ± 16.67 92.1 ± 5.10 63.3 ± 18.56 79.8 ± 18.31V-ditch 91.8 ± 3.45 83.3 ± 16.67 92.3 ± 4.49 45.2 ± 11.90 82.5 ± 11.81Average 81.1 ± 2.64 77.1 ± 3.27 83.34 ± 2.21 59.2 ± 3.55 84.2 ± 3.54

Two way ANOVA F value

Slope (S) 0.89NS 0.87 NS 0.17N 3.59* 0.38NSTreatment (T) 2.12NS 0.59NS 1.33NS 0.58NS 0.99NSS � T 2.36* 1.35NS 1.04NS 0.13NS 1.18NS

RWH, rainwater harvesting.NS, not significant.⁄⁄, Significant at P < 0.01.*, Significant at P < 0.05.

Table 5Average height of 71 months (June 2011) old plants of different species influenced by natural slope gradient and rainwater harvesting devices in degraded hills in Rajasthan.Values are mean with ±SE of five replications.

Slope RWH treatment A. catechu A. indica E. officinalis H. integrefolia Z. mauritiana

Height Collar dia. Height Collar dia. Height Collar dia. Height Collar dia. Height Collar dia.

<10% Control 317.2 ± 42.0 4.97 ± 0.34 311.7 ± 17.9 4.4 ± 0.3 264.6 ± 10.1 4.7 ± 0.3 137.5 ± 12.5 4.0 ± 0.2 241.0 ± 30.4 3.4 ± 0.4CTrench 348.0 ± 33.9 5.52 ± 0.76 343.2 ± 54.3 5.4 ± 0.7 288.9 ± 30.6 5.0 ± 0.4 177.5 ± 7.5 3.8 ± 0.1 254.9 ± 10.1 4.3 ± 0.1Gradonie 279.7 ± 40.8 5.03 ± 0.71 325.5 ± 46.8 4.7 ± 0.5 264.2 ± 29.5 4.8 ± 0.4 104.0 ± 14.0 3.6 ± 0.4 219.0 ± 9.0 2.9 ± 0.1BTrench 280.7 ± 11.2 5.22 ± 0.58 304.8 ± 43.2 4.9 ± 0.6 282.2 ± 27.8 5.3 ± 0.4 130.0 ± 0.0 3.7 ± 0.3 245.7 ± 19.7 3.1 ± 0.2V-ditch 420.8 ± 15.5 7.28 ± 0.59 390.0 ± 44.6 6.0 ± 0.7 287.7 ± 24.7 5.4 ± 0.4 60.0 ± 10.0 3.5 ± 0.2 301.0 ± 7.0 3.7 ± 0.1

10–20%

Control 103.7 ± 5.1 2.78 ± 0.06 174.0 ± 52.9 3.3 ± 0.5 202.5 ± 21.4 4.1 ± 0.3 71.1 ± 8.6 2.8 ± 0.1 157.5 ± 20.9 2.6 ± 0.3CTrench 313.7 ± 30.7 5.82 ± 0.55 307.2 ± 7.8 4.7 ± 0.2 290.8 ± 24.5 5.7 ± 0.6 163.0 ± 9.7 3.7 ± 0.2 243.1 ± 18.5 3.3 ± 0.1Gradonie 279.2 ± 35.0 4.73 ± 0.65 270.0 ± 15.3 4.4 ± 0.1 202.2 ± 8.4 3.8 ± 0.1 118.3 ± 9.6 3.2 ± 0.2 187.2 ± 7.2 2.6 ± 0.5BTrench 325.7 ± 27.6 5.76 ± 0.51 272.2 ± 10.3 5.2 ± 0.3 211.0 ± 14.5 4.5 ± 0.3 130.3 ± 14.3 3.3 ± 0.2 240.9 ± 17.6 3.1 ± 0.3V-ditch 320.9 ± 17.9 5.29 ± 0.42 353.0 ± 16.6 5.7 ± 0.5 218.2 ± 18.7 4.2 ± 0.2 112.9 ± 8.5 3.0 ± 0.1 214.0 ± 21.1 3.0 ± 0.3

>20% Control 310.6 ± 20.7 5.35 ± 0.32 224.5 ± 12.5 3.3 ± 0.1 187.6 ± 11.2 3.8 ± 0.2 132.9 ± 20.6 3.1 ± 0.4 216.7 ± 27.8 2.7 ± 0.5CTrench 319.0 ± 48.0 5.29 ± 0.74 256.3 ± 4.3 3.8 ± 0.2 278.3 ± 27.2 5.1 ± 0.4 122.6 ± 11.4 2.9 ± 0.1 250.9 ± 17.5 2.7 ± 0.1Gradonie 290.3 ± 23.1 5.58 ± 0.41 262.5 ± 37.2 5.6 ± 0.9 234.9 ± 30.3 4.5 ± 0.2 113.4 ± 13.2 3.7 ± 0.1 205.0 ± 7.38 2.8 ± 0.2BTrench 333.4 ± 59.5 5.52 ± 0.86 296.7 ± 27.4 4.2 ± 0.3 207.5 ± 25.2 4.6 ± 0.4 117.0 ± 9.2 3.1 ± 0.2 198.5 ± 10.2 2.6 ± 0.1V-ditch 344.1 ± 41.4 5.72 ± 0.40 300.1 ± 12.6 4.2 ± 0.1 248.6 ± 14.6 4.0 ± 0.2 122.1 ± 5.4 2.7 ± 0.1 226.7 ± 5.5 2.8 ± 0.2

Two way ANOVA F values

Slope (S) 3.08* 2.26NS 3.79* 5.87** 2.67* 5.79** 0.07NS 6.36** 5.15** 7.64**

Treatment (T) 4.60** 3.37* 3.24* 2.91* 4.26** 4.64** 3.31* 0.94NS 3.41* 1.51NSS � T 2.27* 1.73NS 1.11NS 0.70NS 1.25NS 1.65NS 5.07** 2.07NS 0.55NS 0.94NS

RWH, rainwater harvesting.CTrench, contour trench.BTrench, box trench.NS, not significant.*, Significant at P < 0.05.**, Significant at P < 0.01.

G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622 617

Among the RWH treatments, MAI in height was greater in VD,whereas MAI in collar diameter was greater (P < 0.05) in CTtreatments.

3.7. Plant biomass and rooting pattern

Both aboveground and root dry biomass was highest from E. offi-cinalis and lowest for A. indica (Table 6). Root: shoot (above ground)

biomass ratio was highest for E. officinalis and the lowest was for H.integrifolia. Average rooting depth ranged from 235.6 cm for Z. mau-ritiana to 131.3 cm for A. catechu. Root spread was larger for Z. mau-ritiana and smaller for A. indica, but exceeded crown diameter. Totalnumber of roots ranged from 9.0 for E. officinalis to 5.0 for Z. mauri-tiana. Acacia catechu and A. indica showed both laterals as well asdeep penetrating roots indicating a fan shaped structure. Emblicaofficinalis showed multifarious umbrella shaped root system,

ble

6er

age

biom

ass

and

root

grow

thva

riab

les

ofpl

ants

ofdi

ffer

ent

tree

spec

ies

and

dist

ribu

tion

ofro

ots

indi

ffer

ent

soil

laye

rsin

degr

aded

hills

inRa

jast

han,

Indi

a.V

alue

sar

em

ean

wit

h±S

Eof

3to

12re

plic

atio

ns.

Tree

spec

ies

Hei

ght

(cm

)C

olla

rdi

amet

er(c

m)

Dry

biom

ass

(kg

per

plan

t)N

um

ber

ofro

ots

Roo

tdi

amet

er(c

m)

Roo

tdi

stri

buti

on(c

m)

Roo

tsp

read

/dep

thra

tio

Shoo

tR

oot

R/S

rati

o0–

40cm

Bel

ow40

cmR

oot

spre

adR

oot

dept

h

A.c

atec

hu

357.

7±

33.7

6.81

±0.

626.

27±

1.35

2.74

±0.

470.

506.

0±

0.7

2.3

±0.

44.

60±

0.60

315.

2±

31.2

131.

2±

18.2

3.0

±0.

4A

.in

dica

297.

5±

23.2

5.98

±0.

603.

80±

1.64

0.94

±0.

200.

325.

8±

1.5

2.0

±1.

06.

20±

0.4

264.

3±

44.4

166.

3±

22.9

1.6

±0.

1E.

offi

cin

alis

386.

0±

38.1

9.18

±1.

516.

35±

2.32

4.04

±1.

260.

695.

7±

1.2

3.3

±1.

38.

44±

1.51

318.

3±

35.0

419

7.0

±20

.51.

6±

0.1

H.i

nte

grif

olia

380.

0±

69.3

9.23

±1.

656.

33±

2.02

1.57

±0.

300.

283.

7±

0.3

1.3

±0.

39.

55±

1.10

280.

0±

55.4

219

3.0

±17

.31.

4±

0.2

Z.m

auri

tian

a32

2.0

±18

.05.

13±

0.27

4.47

±3.

031.

59±

0.42

0.35

3.0

±1.

52.

0±

0.0

6.59

±.77

357.

0±

51.4

323

5.6

±28

.21.

1±

0.1

,not

-sig

nifi

can

t.R

/Sra

tio,

root

:sh

oot

rati

o.Si

gnifi

can

tat

P<

0.05

.,Si

gnifi

can

tat

P<

0.01

.

618 G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622

Ta Av

NS⁄, ⁄⁄

whereas H. integrifolia and Z. mauritiana exhibited well-developeddeep taproot system with limited number of spreading roots. Max-imum number of roots was in 0–40 cm layer as compared to deepsoil layer. Highest density of roots was around the rootstock associ-ated with fine laterals from the roots in A. catechu and E. officinalis.Lateral spreading (0–40 cm soil layer) of the roots was greatest forA. catechu followed by E. officinalis. The spreading of lateral rootsfor Z. mauritiana was below 40 cm soil layer. E. officinalis, H. integri-folia and Z. mauritiana exhibited deep penetrating roots and lowerroot spread to root depth ratio (Table 6).

3.8. Correlations and regressions

Both annual and seasonal rainfall showed a decreasing(r = �0.591, P < 0.01, n = 450) trend during experimental periodand were positively correlated to SWC in December (r = 0.152,P < 0.01) and June (r = 0.228, P < 0.01). The correlations of slopegradient were positive with vegetation height (r = 0.356, P < 0.01)and negative with SWC in December (r = �0.138, P < 0.01) andplant growth (r = �0.145 to �0.215, P < 0.5). Height of herbaceousvegetation was positively correlated to herbaceous biomass(r = 0.380, P < 0.01), SWC in December (r = 0.291, P < 0.01) and June(r = 0.218, P < 0.01), and PARintC (r = 0.483, P < 0.01, n = 75). Positiverelationship was also observed between herbaceous biomass andgrowth variables of Acacia catechu (r = 0.131–0.171, P < 0.05), H.integrifolia (r = 0.125–0.188, P < 0.05) and Z. mauritiana(r = 0.185–0.198, P < 0.01). SWC of December indicated positivecorrelations to both height and collar diameter growth of A. catechuand Z. mauritiana (r = 0.113–0.159, P < 0.05), height of A. indica(r = 0.142–0.170, P < 0.01) and E. officinalis (r = 0.131–0.168,P < 0.01) both in December and June and PARintC (r = 0.298,P < 0.01, n = 75). PARintC showed positive relationship with SWC(r = 0.436, P < 0.01) and plant growth (r = 0.316–0.541, P < 0.05,n = 75). PARintT and PARintV were negatively related (r = �0.725,P < 0.05). Regression analysis indicated linear relationships be-tween rainfall and SWC (December as well as June), and dry bio-mass and SWC/vegetation height, whereas vegetation heightshowed curvilinear (S) relationship with SWC both in Decemberand June (Fig. 2).

4. Discussion

4.1. Soil water dynamics

We observed a linear relationship between SWC and both an-nual and seasonal rainfall (R2 = 0.022, F1/448 = 9.93 in Decemberand R2 = 0.051, F1/448 in June, P < 0.01). Rainfall during June to Octo-ber enhanced SWC in December, but soil water use by growingvegetation and increased evaporative demand with increase intemperature towards summer resulted in a decline in SWC by June.A decreasing trend in rainfall during 2005 to 2010 (r = -0.118,P < 0.05) and growing stock of varying rooting depth also influ-enced the spatial distribution of water inside the soil profile (Wuet al., 1985). Greater SWC in December 2010 than in December2009 was the effect of extended period of rainfall indicated by apositive relationship (r = 0.147 in December and 0.228 in June,P < 0.01) between rainfall and stored soil water (Lodge and Brenan,2009). Soil texture also played an important role in soil water stor-age as indicated by lowest (P < 0.05) SWC in 10–20% slope withgreater sand content as compared to the soils of other slopes(Singh et al., 1998; Singh, 2012). Stony soil surface restricted sur-face runoff and facilitated water infiltration during rain similar tothe observations of Van Wesemael et al. (1996) and Katra et al.(2008) and favoured soil water storage in <10% and >20% slopes.Presence of roots of trees and herbaceous vegetation and

0

0.5

1

1.5

2

2.5

3

3.5

0 50 100 150 200 250

(e) Log BM =2.342 + 0.00216* H, R2 = , SE = 0.221, F1/448 = 45.33, P < 0.001

Height (cm)

Log

Bio

mas

s(g

m-2

)

0

500

1000

1500

2000

2500

0 1 2 3 4

(b) SWC (J) = 0.864 + 0.000292 * RF, R2 =0.051, SE = 0.60, F1/448, P < 0.001

Soil water Content (%, w/w)

0

500

1000

1500

2000

2500

0 4 8 12

(a) SWC (D) = 3.50 + 0.000673*RF, R2 = 0.022, SE = 2.17, F1/448 = 9.93, P < 0.001

Rai

nfal

l (m

m)

Soil water content (%, w/w)

0

0.5

1

1.5

2

2.5

3

3.5

0 4 8 12

(f) Log BM = 2.4450 + 0.01276*SWC(D), R2

= 0.015, SE = 0.23, F1/448 = 6.62, P < 0.01

Soil water content (%, w/w)

0

0.5

1

1.5

2

0 50 100 150 200 250

(d) SWC (J) = exp(0.3859 -(24.1627/H)), R2

= 0.095, SE = 0.62, F1/448 = 47.11, P < 0.001

Vegetation height (cm)

0

3

6

9

12

15

0 50 100 150 200 250

(c) SWC (D) = exp(1.6890 - (023.9880/H)), R2 =0.120, SE = 0.52, F1/448 = 61.05, P < 0.001

Vegetation height (cm)

Soil

wat

er c

onte

nt (

%, w

/w)

Fig. 2. Relationships between annual rainfall and soil water content (a and b), vegetation height and soil water (c and d), herbaceous biomass and height (e), and herbaceousbiomass and soil water content (f) irrespective of slope gradient and rainwater harvesting treatments. Legend for both left panel and corresponding right panel are same if notdefined. D, December; J, June; SWC, soil water content; and BM, herbaceous dry biomass.

G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622 619

corresponding soil organic matter content also enhance soil infil-tration and dynamics of soil water distribution under tree canopies(Caylor et al., 2006; D’Odorico et al., 2007). RWH structures en-hanced SWC (lowest in the control), but greater storage capacityof CT/BT structures as compared to GD/VD structures facilitated

water storage to a greater extent in soil profile that remained avail-able for a longer period of time (James et al., 2003). This is reflectedby a greater SWC and plant growth in CT/BT treatments. Because ofits distribution throughout the plots BT structures were exposed toa greater evaporative soil water loss and hence greater decline in

Rooting pattern of Acacia catechu Rooting pattern of Azadirachta indica

Rooting pattern of Emblica officinalis Rooting pattern of Holoptelia integrifolia

Rooting pattern of Zizyphus mauritiana

Fig. 3. Rooting pattern of different tree species planted under different rainwater harvesting treatments at Banswara area, Rajasthan, India.

620 G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622

SWC during December to June than in the plots with CT structure.Distribution of soil water in upper soil layers and its efficient utili-zation by the growing plants/vegetation caused faster decline inSWC in VD and GD treatments as compared to CT/BT treatments(Singh et al., 2010). A curvilinear relation between vegetationheight and SWC showed beneficial effects of taller vegetation inconserving soil water by way of intercepting PAR reaching to soilsurface (positive relationship between vegetation height, PARintC

and SWC).Soil water depletion is also related to spatial distribution of

roots of growing vegetations (Wang et al., 2010), though reportsare also there that variation in precipitation history and landscapepositions are the main determinants of water-use patterns thanwould be expected on the basis of absolute rooting depth (Nippert

and Knapp, 2007). Distribution of roots of A. catechu and A. indica inboth top 0–40 cm and deeper soil layer provided enough adsorp-tive surfaces to exploit water and nutrients from the topsoil duringlow rainfall and from deeper soil layer during stress supportingtheir higher survival (Singh and Rathod, 2002).

4.2. Herbaceous biomass production

We did not find a clear relationship between growth (height)and biomass of herbaceous vegetation and annual rainfall, whereasboth variables varied between the years (P < 0.05, year � slope,year � RWH treatment and year � slope � RWH treatment interac-tions). Dominance of high altitude grasses like Themada quadrivel-vis, Apluda mutica and upcoming Heteropogon contortus – all tall

G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622 621

grasses increased average height of the herbaceous vegetation in>20% slope (positive relationship between slope gradient and veg-etation growth, r = 0.356, P < 0.01) as compared to the other slopes,enhanced herbaceous biomass. This was indicated by a weak linearrelationship between height and biomass of herbaceous vegetation(R2 = 0.092, F1/448 = 45.34, P < 0.01). Despite of low SWC and conse-quent low vegetation growth, the highest herbaceous biomass in10–20% slope as compared to the other slopes was due to low com-petition by lesser grown plantation (low PARintT) and dominance ofwater stress tolerance grasses like A. funiculata and H. contortus asreported earliar by Ludlow (1980), where H. contortus, C. cilairisand Panicum maxicum tolerated soil water stress of �120 to�130 bar. Smit (2005) observed that thinning of Colophospermummopane increased productivity of the grass layer during years ofbelow average rainfall. Highest amount of herbaceous biomass in10–20% slope was also supported by the highest value of PAR inter-ception due to herbaceous vegetation (PARintH) in this slope ascompared to the other slopes. The warm season grasses have beenfound more beneficial for soil and environment when grown inmarginal lands than grown in croplands or natural forests (Blan-co-Canqui, 2010). Highest level of water availability in <10% slopepromoted growth of both herbaceous vegetation and plantation.However, surface available soil water (because of growth duringrainfall) and greater soil water depletion in GD/VD treatments ascompared to CT/BT treatments revealed higher vegetation growthand biomass in former treatments. Increase in SWC in Decemberand June with growth (r = 0.255, P < 0.01) and biomass (curvilinear,R2 = 0.028, F = 12.82, P < 0.01) of herbaceous vegetation indicatedpositive effects of vegetation on soil water conservation. VD andGD treatments were best for growth and biomass production ofherbaceous vegetation, which were favoured by growing Acaciacatechu, H. integrifolia and Z. mauritiana as shown by positive rela-tionship between plant and vegetation growth.

4.3. Plant survival, growth and root distribution

Availability of soil water and its gradient in different slopes af-fected survival and growth of planted seedlings. Low availability ofsoil water in 10–20% slope and in control plots affected plant sur-vival negatively. Highest survival in BT treatment with relativelymore SWC indicates that RWH promoted plants survival (El-Attaand Aref, 2010). Greater growth during June to December was sup-ported by favorable environment i.e., enhanced soil water and con-sequently nutrient mobility (Marion and Everett, 2006). Greateravailability of soil water in <10% slope facilitated growth andMAI in height and collar diameter of most of the species (substan-tiated by greater PARintT). Tsui et al. (2004) and Yong et al. (2006)observed that spatial redistribution of surface runoff enhancednutrients and soil water availability on lower slope positions andcontributed to the greater vegetation/plant growth. Lowering ofSWC with increase in slope gradient affected plant growth partic-ularly in 10–20% slope, where competitive use of soil water af-fected the growth of both plantation and herbaceous vegetation.A significant decrease in soil moisture (but not the lights) havebeen observed under Panicum dichotomiflorum and Andropogon vir-ginicus grasses (Morris et al., 1993; Reynolds et al., 2002; Sonohatet al., 2002), but we observed highest value of PARintV in 10–20%slope and PARintC in >20% slope caused by higher population andgrowth of vegetation, respectively. Mitchell et al. (1999) observedreduction in light under Liquidambar styraciflua tree and soil mois-ture under Andropogon virginicus affecting growth of Pinus taeda.Relatively greater growth of A. catechu in >20% slope and H. integ-rifolia in 10–20% slope as compared to other slopes was due tofavorable soil conditions (i.e., relatively heavy soil in >20% slopeand sandy in 10–20% slope, respectively) similar to the observa-tions of Cao et al. (2007). Effects of different RWH structures in

facilitating soil water and nutrient availability varied among theslopes (Singh, 2009), but rooting and growth pattern of tree speciesalso influenced soil water status. Increased soil water throughRWH facilitated plants growth, but greater growth of A. catechuand A. indica in VD treatment and those of the remaining speciesin CT treatments was because of species characteristics particularlythe rooting pattern (Fig 3) and subsequent utilization of soil re-sources from different soil layers (Gupta, 1994). Fibrous roots ofplanted species were mostly allocated in top 40 cm of soil layer,but concentrations of most of the roots of A. catechu and A. indicain top soil layers provided enough adsorptive surfaces to exploitwater and nutrients from top soils in VD/GD treatments. However,relatively deep penetrating roots in E. officinalis, H. integrifolia andZ. mauritiana facilitated these species to prefer CT structures forutilization of deep soil stored water during spring growth. Alloca-tion of high proportions of root biomass in the top soil layer hasalso been reported for successional tropical trees (Singh andRathod, 2002; Das and Chaturvedi, 2012).

5. Conclusions and recommendations

Temporal changes in rainfall influenced soil water status andcorresponding growth of both planted species as well as herba-ceous vegetation. Rainwater harvesting enhanced soil water status.The effects of VD structure in improving soil water were relativelygreater in <10% slope, whereas the effect GD and CT structureswere highest in 10–20% slope and >20% slope, respectively. Highestgrowth and biomass of herbaceous vegetation in VD/GD treatmentas compared to BT/CT treatments showed the efficiency of the for-mer RWH structure in developing herbaceous vegetation cover andsubsequently their role in conserving soil water by interceptingboth water flow and PAR reaching to the soil surface. Greater soilwater status in <10% slope promoted survival and growth of plan-tation, but decreased soil water availability in higher slopes af-fected plant growth and PAR interception. However, sandy soil in10–20% slope and clay soil in >20% slope favoured growth of A. cat-echu and H. integrifolia, respectively. V-ditch and GD structure fa-voured water distribution in upper soil layers and facilitatedherbaceous growth, whereas Contour trench and Box trench facil-itated water storage in deep soil profile favorable for plant growth.Distribution of plant roots in surface and deep soil layers resultedin best growth of Acacia catechu and A. indica in V-ditch RWH struc-ture, whereas contour trench RWH structure was best for E. offici-nalis, H. integrifolia and Z. mauritiana. Conclusively, effect of rainfallon SWC was linear and RWH facilitated soil water storage andgrowth of plantation and herbaceous vegetation. RWH structurebased performance of tree species may be beneficial in further rep-lication of these combinations in restoration of degraded hills,enhancement of fodder and fuelwood and improvement in ecolog-ical benefits.

Acknowledgements

We wish to express our sincere thanks to the Director AridForest Research Institute, Jodhpur for providing necessary facilities.Financial supports from Forest Department, Government ofRajasthan and help rendered by KVK, Banswara by providingclimatic data are gratefully acknowledged.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.foreco.2013.09.002.

622 G. Singh et al. / Forest Ecology and Management 310 (2013) 612–622

References

Barron, J., Rockström, J., Gichuki, F., Hatibu, N., 2003. Dry spell analysis and maizeyields for two semi-arid locations in East Africa. Agric. For. Meteorol. 117, 23–37.

Blanco-Canqui, H., 2010. Energy crops and their implications on soil andenvironment. Agronomy J. 102, 403–419.

Cao, S., 2008. Why large-scale afforestation efforts in China have failed to solve thedesertification problem. Environ. Sci. Technol. 42, 1826–1831.

Cao, S., Chen, L., Yu, X., 2009a. Impact of China’s grain for green project on thelandscape of vulnerable arid and semiarid agricultural regions: a case study innorthern Shaanxi Province. J. Appl. Ecol. 46, 536–543.

Cao, S., Wang, G., Chen, L., 2010. Questionable value of planting thirsty trees in dryregions. Nature 465, 31–31.

Cao, S., Chen, L., Liu, Z., 2007. Impact of three soil types on afforestation in China’sLoess Plateau. Landsc. Urban Plan. 83, 208–217.

Cao, S., Xu, C., Chen, L., Wang, X., 2009b. Attitudes of farmers in China’s northernShaanxi Province towards the land-use changes required under the Grain forgreen project, and implications for the program’s success. Land Use Policy 26,1182–1194.

Cao, S., 2011. Impact of China’s large-scale ecological restoration program on theenvironment and society: achievements, problems, synthesis, and applications.Crit. Rev. Environ. Sci. Technol. 2011 (41), 317–335.

Cao, S., Sun, G., Zhang, Z., Chen, L., Feng, Q., Fu, B., McNulty, S., Shankman, D., Tang, J.,Wang, Y., Wei, X., 2011. Greening China naturally. Ambio 40, 828–831.

Cohen, I.S., Lopes, V.L., Slack, D.C., Yanez, C.H., 1995. Assessing risks of waterharvesting systems in arid environment. J. Soil Water Conserv. 50, 446–449.

Caylor, K.K., D’Odorico, P., Rodriguez-Iturbe, I., 2006. On the ecohydrology ofstructurally heterogeneous semi arid landscapes. Water Res. Res. 42, W07424,doi: 101029/2005WR004683.

Das, D.K., Chaturvedi, O.P., 2012. Root biomass and distribution of five agroforestrytree species. Agrofor. Syst. 74, 223–230.

D’Odorico, P., Caylor, K.K., Okin, G.S., Scanlon, T.M., 2007. On soil moisture–vegetation feedbacks and their possible effects on the dynamics of drylandecosystems. J. Geophys. Res. 112, G04010, doi: 101029/2006JG000379.

El-Atta, H.A., Aref, I., 2010. Effect of terracing on rainwater harvesting and growth ofJuniperus procera Hochst. ex Endlicher. Int. J. Environ. Sci. Technol. 7, 59–66.

Faroda, A.S., Joshi, N.L., Singh, R., Saxena, A., 2007. Resource management forsustainable crop production in arid zone-a review. Ind. J. Agronomy 52, 181–193.

Gao, Y., Zhong, B., Yue, H., Wu, B., Cao, S., 2011. A degradation threshold forirreversible soil productivity loss in southeastern China: results of a long-termcase study in Changing County. J. Appl. Ecol. 48, 1145–1154.

Gowing, J.W., Mahoo, H.F., Mzirai, O.B., Hatibu, N., 1999. Review of rainwaterharvesting technique and their evidence for their use in semi-arid Tanzania.Tanzania J. Agri. Sci. 2, 171–180.

Gupta, G.N., 1994. Influence of rainwater harvesting and conservation practices ongrowth and biomass production of Azadirachta indica in the Indian desert. For.Ecol. Manage. 70, 329–339.

Gupta, G.N., 1995. Rain water management for tree planting in Indian desert. J. AridEnviron. 31, 219–235.

James, S.E., Pärtel, M., Wilson, S., Peltzer, D.A., 2003. Temporal heterogeneity of soilmoisture in grassland and forest. J. Ecol. 91, 234–239.

Jia, Y., Li, F.M., Wang, X.L., Yang, S.M., 2006. Soil water and alfalfa yields as affectedby alternating ridges and furrows in rainfall harvest in a semiarid environment.Field Crop Res. 97, 167–175.

Katra, I., Lavee, H., Sarah, P., 2008. The effect of rock fragment size and position ontopsoil moisture on arid and semi-arid hillslopes. Catena 72, 49–55.

Ludlow, M.M., 1980. Stress physiology of tropical pasture plants. Tropic. Grasslands14, 133–145.

Li, W.Q., Xiao-Jing, L., Khan, M.A., Gul, B., 2008. Relationship between soilcharacteristics and halophytic vegetation in coastal region of north China.Pak. J. Botany 40, 1081–1090.

Li, X.R., Ma, F.Y., Xiao, H.L., Wang, X.P., Kim, K.C., 2000. Long-term effects ofrevegetation on soil water content of sand dunes in arid region of NorthernChina. J. Arid Environ. 57, 1–16.

Ludwig, J.A., Wilcox, B.P., Breshears, D.D., Tongway, D.J., Imeson, A.C., 2005.Vegetation patches and runoff-erosion as interacting ecohydrologicalprocesses in semiarid landscapes. Ecology 86, 288–297.

Marion, G.M., Everett, K.R., 2006. The effect of nutrient and water additions onelemental mobility through small tundra watersheds. Ecography 12, 317–323.

Mati, B.M., 2005. Overview of water and soil nutrient management undersmallholder rainfed agriculture in East Africa. Working Paper 105.International Water Management Institute (IWMI), Colombo, Sri Lanka.

Mitchell, R.J., Zutter, B.R., Gjerstad, D.H., Glover, G.R., Wood, C.W., 1999.Competition among secondary-successional pine communities: a field studyof effects and responses. Ecology 80, 857–872.

Morris, L.A., Moss, S.A., Garbett, W.S., 1993. Competitive interference betweenselected herbaceous and woody plants and Pinus taeda L. during two growingseasons following planting. For. Sci. 39, 166–187.

Nippert, J.B., Knapp, A.K., 2007. Linking water uptake with rooting patterns ingrassland species. Oecologia 153, 261–272.

Prinz, D., 2001. Water harvesting for afforestation in dry areas. In: Proceedings, 10thInternational Conference on Rainwater Catchment Systems, Mannheim, 10–14September 2001, p. 195.

Rahman, M.A., 2011. Biodiversity conservation and food security of indigenouspeople hilly regions of Bangladesh. <http://feppcar.org/150/biodiversity-conservation-and-food-security-of-indigenous-people-hilly-regions-of-bangladesh>. Accessed on 3rd January 2013.

Reynolds, P.E., Curry, R.D., Bell, F.W., 2002. Effects of controlled weed densities, soiltypes, weed species and duration of weed establishment on seedlingmicroclimate and seedling growth. In: Fourth International Conference onForest Vegetation Management: Popular Summaries. H. Frochot, C. Collet and P.Balandier (compilers) Institut National de la Recherche Agronomique (INRA),Champenoux, France, pp. 258–260.

Runhaare, J., Witte, J.P.M., Verburg, P.H., 1997. Ground-water level, moisture supply,and vegetation in the Netherlands. Wetlands 17, 528–538.

Singh, G., 2009. Soil water dynamics, growth of Dendrocalamus strictus and herbageproductivity influenced by rainwater harvesting in Aravalli hills of Rajasthan.For. Ecol. Manage. 258, 2519–2528.

Singh, G., 2012. Enhancing growth and biomass production of plantation andassociated vegetation through rainwater harvesting in degraded hills insouthern Rajasthan, India. New Forest. 43, 349–364.

Singh, G., Rani, A., Bala, N., Upadhyaya, S., Baloch, S.R., Limba, N.K., 2010. Resourceavailability through rainwater harvesting influenced vegetation diversity andherbage yield in southern Aravalli hills of India. Front. Agri. China 4, 145–158.

Singh, G., Rathod, T.R., 2002. Plant growth, biomass production and soil waterdynamics in a shifting dune of Indian desert. For. Ecol. Manage. 17, 309–320.

Singh, J.S., Milchunas, D.G., Lauenroth, W.K., 1998. Soil water dynamics andvegetation patterns in a semiarid grassland. Plant Ecol. 134, 7–89.

Sokal, R.R., Rohlf, P.J., 1981. Biometery, second ed.. W.H. Freeman, New York.Smit, G.N., 2005. Tree thinning as an option to increase herbaceous yield of an

encroached semi-arid savanna in South Africa. BMC Ecol. 5, 4, 10.1186/1472-6785-5-4.

Sonohat, G., Sinoquet, H., Varlet-Grancher, C., Rakocevic, M., Jacquet, A., Simon, J.C.,Adam, B., 2002. Leaf dispersion and light partitioning in three-dimensionallydigitized tall fescue–white clover mixtures. Plant Cell Environ. 25, 529–538.

Suleman, S., Wood, M.K., Shahm, B.H., Murray, L., 1995. Rainwater harvesting forincreasing livestock forage on arid rangeland in Pakistan. J. Range Manage. 48,523–527.

Tsui, C.C., Chen, Z.S., Hsieh, C.F., 2004. Relationships between soil properties andslope position in a lowland rain forest of southern Taiwan. Geoderma 123, 131–142.

Van Wesemael, B., Poesen, J., Kosmas, C.S., Danalatos, N.G., Nachtergaele, J., 1996.Evaporation from cultivated soils containing rock fragments. J. Hydrol. 82, 62–82.

Vohland, K., Barry, B., 2009. A review of in situ rainwater harvesting (RWH) practicesmodifying landscape functions in African drylands. Agri. Ecosyst. Environ. 134,119–127.

Waterfall, P.H., 2006. Harvesting rainwater for landscape use. University of Arizona,Cooperative Extension/Low 4 Program. Second Edition, Revised 2006.

Wang, Y.M., Hudan, T., Zhang, J.Z., et al., 2010. Study on the root distributionpeculiarity of cotton in drip irrigation under mulched-film. Journal ofAgricultural University of Hebei 33, 103–106.

Wu, H.-I., Sharpe, P.J.H., Walker, J., Penridge, L.K., 1985. Ecological Field Theory: aspatial analysis of resource interference among plants. Ecological Modelling 29,215–243.

Yanhui, W., Bonell, M., Feger, K.-H., Pengtao, Y., Wei, X., Lihong, X., 2012. Changingforestry policy by integrating water aspects into forest/vegetation restoration indryland areas in China. Bullet. Chin. Acad. Sci. 26, 59–67.

Yong, L., Wang, C.W., Hongliang, T., 2006. Research advances in nutrient runoff onsloping land in watersheds. Aquat. Ecosyst. Health Manage. 9, 27–32.

Yu, M., Gao, Q., Epstein, H.E., Zhang, X., 2008. An ecohydrological analysis foroptimal use of redistributed water among vegetation patches. Ecol. Appl. 18,1679–1688.

Zhuang, X., 1997. Rehabilitation and development of forest on degraded hills ofHong Kong. For. Ecol. Manage. 99, 197–201.