Environ Monit Assess (2014) 186:2013–2024 DOI 10.1007/s10661-013-3514-7
Land use impact on soil quality in eastern Himalayan region of India A. K. Singh & L. J. Bordoloi & Manoj Kumar & S. Hazarika & Brajendra Parmar
Received: 13 February 2013 / Accepted: 28 October 2013 / Published online: 14 November 2013 # Springer Science+Business Media Dordrecht 2013
Abstract Quantitative assessment of soil quality is required to determine the sustainability of land uses in terms of environmental quality and plant productivity. Our objective was to identify the most appropriate soil quality indicators and to evaluate the impact of six most prevalent land use types (natural forestland, cultivated lowland, cultivated upland terrace, shifting cultivation, plantation land, and grassland) on soil quality in eastern Himalayan region of India. We collected 120 soil samples (20 cm depth) and analyzed them for 29 physical, chemical, and biological soil attributes. For selection of soil quality indicators, principal component analysis (PCA) was performed on the measured attributes, which provided four principal components (PC) with eigenvalues >1 and explaining at least 5 % of the variance in dataset. The four PCs together explained 92.6 % of the total variance. Based on rotated factor loadings of soil attributes, selected indicators were: soil organic carbon
A. K. Singh Department of Agricultural Chemistry and Soil Science, SASRD, Nagaland University, Medziphema 797106 Nagaland, India L. J. Bordoloi : M. Kumar (*) : S. Hazarika Division of Natural Resource Management (Soil Science), ICAR Research Complex for NEH Region, Umiam 793103 Meghalaya, India e-mail: [email protected] B. Parmar Division of Soil Science, Directorate of Rice Research, Rajendranagar 500030 Hyderabad, India
(SOC) from PC-1, exchangeable Al from PC-2, silt content from PC-3, and available P and Mn from PC-4. Indicators were transformed into scores (linear scoring method) and soil quality index (SQI) was determined, on a scale of 0–1, using the weighting factors obtained from PCA. SQI rating was the highest for the least-disturbed sites, i.e., natural forestland (0.93) and grassland (0.87), and the lowest for the most intensively cultivated site, i.e., cultivated upland terrace (0.44). Ratings for the other land uses were shifting cultivation (0.60)>cultivated low land (0.57)>plantation land (0.54). Overall contribution (in percent) of the indicators in determination of SQI was in the order: SOC (58 %)>exch. Al (17.1 %)>available P (8.9 %)>available Mn (8.2 %)>silt content (7.8 %). Results of this study suggest SOC and exch. Al as the two most powerful indicators of soil quality in study area. Thus, organic C and soil acidity management holds the key to improve soil quality under many exploitatively cultivated land use systems in eastern Himalayan region of India. Keywords Northeastern hill regions . Soil quality indicators . Shifting cultivation . Land use changes
Introduction Soil quality is defined as “the capacity of a soil to function within ecosystem and land-use boundaries, to sustain biological productivity, maintain environmental quality, and promote plant and animal health” (Doran and Parkin 1994). This clearly establishes soil quality as
2014
a vital determinant of agricultural productivity as well as environmental sustainability, the two most important challenges facing the world today. In order to meet the needs of growing population, there have been drastic changes in land use and soil management practices in recent years, which are considered to have caused definite deterioration in soil and environmental quality. Such changes in land use systems and the associated impacts are however not uniform across the regions. Therefore, to determine the sustainability of land management systems prevalent in a region, quantitative assessment of soil quality on regional scale is necessary. Soil quality index (SQI) is an easily comprehensible way of ascertaining whether soil quality is improving, stable, or declining under different land use systems (Masto et al. 2008). It is comprised of indicators sensitive to regional scale changes in land use, and could be useful to researchers and policy makers for assessing the effects of different land use and management practices on soil quality over broad geographical areas (Brejda and Moorman 2001). Soil quality is broadly a function of soil properties but it may not be adequately represented by individual soil properties; SQI derived from integrated soil quality indicators based on a combination of soil properties could better reflect the status of soil quality (Masto et al. 2009). There are three main steps involved in soil quality indexing: first, selecting appropriate indicators; second, transforming indicators to scores; and third, combining the scores into index (Andrews et al. 2002). Indicators’ selection can be made using expert opinion (Doran and Parkin 1994) or statistical techniques like principal component analysis (PCA; Andrews et al. 2001, 2002; Sharma et al. 2005) and factor analysis (Brejda et al. 2000). Scoring of indicators can be done using linear and nonlinear scoring methods (Andrews et al. 2002) and, finally, using multiplicative (Singh et al. 1992), simple additive (Andrews and Carroll 2001), or weighted additive (Karlen et al. 1998) methods, scores can be integrated into final SQI. Although there are several studies on the effect of land uses and soil management practices on soil quality in different regions of the world including IndoGangetic Plains of India (Xu et al. 2006a, b; Masto et al. 2007, 2009; Ayoubi et al. 2011), no such information is available for the eastern Himalayan regions of India which has witnessed drastic changes in land use in recent times. This region, consisting of seven hilly states
Environ Monit Assess (2014) 186:2013–2024
of India, is characterized by diverse agroclimatic and geographical situations. About 65 % of the total geographical area is under forests, ~16 % under crops, and the rest either under nonagricultural uses or uncultivated land (Saha et al. 2012). Extreme forms of soil acidity and unacceptably high rates of soil erosion, induced by high rainfall and sloppy terrains have been the major constraints for agricultural production in majority of the areas, which lags far behind the national average crops’ productivity. Besides, negligible adoption of the soil acidity amelioration practices (such as liming), minimal use of external nutrient inputs, loss of top soil along with the nutrients and organic carbon thereof, and subsequent decline in overall soil quality further dampen the prospect of enhancing crop productivity in the region. Under the pressure of fast-growing population’s need, natural forestlands are being increasingly cleared to be put under different land uses with exhaustive cultivation practices, such as shifting cultivation (also known as slash and burn cultivation, and locally known as Jhum cultivation), terrace cultivation, etc. These land use changes are often implicated for the declining soil quality in the region, which could be a major obstacle towards increasing crop production in future. Thus, identification of the major soil quality determining parameters under the prevailing land use types of the region is crucial for evolving measures to improve soil quality and its productivity in future. Since such studies have not been undertaken in the region so far, objective of this study was to identify the appropriate soil quality indicators and to evaluate the differences in soil quality (using soil quality indexing approach) under the six most prevalent land use types in eastern Himalayan region of India.
Materials and methods Description of study sites The study area is located in Dimapur district of Nagaland state (25°06′–27°04′ N, 93°21′–95°15′ E) in eastern Himalayan region of India (Fig. 1). Average altitude ranges from 260 to 690 m, with average annual temperature ranging from 18 to 20 and 23 to 25 °C in the higher and lower elevations, respectively. Average annual rainfall is 1,825 mm, 90 % of which is experienced during June to September. The soils of the study sites were acid Alfisols (rich in Fe and Al). Detailed
Environ Monit Assess (2014) 186:2013–2024
2015
receives a regular but small application of chemical fertilizers (5–10 kg ha−1 of NPK). Soil sampling and analysis
Fig. 1 Location of the study area shown on India’s map
description of these sites is shown in Table 1. We identified six adjacently located land use systems (LUS) viz., natural forestland (NFL), cultivated lowland (CLL), cultivated upland terrace (CUT), shifting cultivation (SC), plantation land (PL), and grassland (GL) for this study. Group discussion and key informants’ interviews revealed that these LUSs were in place for at least past 20 years. The NFL represents an undisturbed forest site with majority of the vegetation comprising of Terminalia bellerica, Mesua ferrea, Ficus elastica, Gmelina arborea, Alnus nepalensis, Tectona grandis, Bauhinia variegate, Panicum spp., Pongamia spp., Schima wallichii, etc. Monocropping with local paddy cultivars (during July–August) is practiced in CLL site where application of FYM at 5–7 Mg ha−1 before paddy transplanting is common. CUT are mostly under nutrient exhaustive annual crops (paddy, maize, millets, etc.), and represent the most intensively cultivated system in this study. These terraces receive no application of chemical fertilizers, with only nominal applications of organic manures at times (~0.25 Mg ha−1). The SC site has been subject to a cycle of 3–4 years of cultivation followed by 4 years of fallow periods. Nutrient supply in the first 2 years of cultivation comes entirely from the ashes of the slashed and burnt vegetation, with some external application of organic manures in third/fourth years (~2 Mg ha−1). Rice, maize, tapioca, sweet potato, banana, pine apple, etc. are the commonly grown crops in this land use. PL site is mostly under mixed plantations of citrus, pineapple, banana and mango, and
Altogether, 120 composite surface (0–20 cm depth) soil samples (each composite sample derived from three randomly collected subsamples) from six land uses were collected (October 2009), air dried, ground, and passed through 2-mm sieve (0.5 mm sieve for organic carbon estimation). Samples were analyzed for 29 soil attributes as follows: Soil pH (1:2 soil/water) was measured using glass electrode. Soil organic carbon (SOC) was determined by wet digestion method (Walkley and Black 1934). Cation exchange capacity (CEC) was estimated by sodium saturation method (Jackson 1973). Particle size distribution was determined by international pipette method (Page et al. 1982). Available N was estimated using alkaline KMnO4 distillation method (Subbiah and Asija 1956); available P by Bray-II method (Bray and Kurtz 1945). Exchangeable K and Na were determined using ammonium acetate extraction followed by emission spectrometry (Jackson 1973). Exch. Ca and Mg were determined through versene titration (Baruah and Barthakur 1999), and exch. Al and acidity by potassium chloride method (Page et al. 1982). Effective CEC (ECEC) was calculated as the sum total of exchangeable cations (Na, K, Ca, Mg, and Al). Available Fe, Mn, Cu, and Zn were determined using 0.005 M DTPA extraction followed by atomic absorption spectrometry (Lindsay and Norvell 1978). Available B and Mo were measured colorimetrically using hot water and ammonium oxalate solution as extractants, respectively (Page et al. 1982). Soil dehydrogenase activity was estimated through reduction of TTC (triphenyltetrazolium chloride) to TPF (triphenylformazan) (Page et al. 1982). Soil microbial biomass C (SMBC) was estimated by chloroform fumigation–incubation procedure given by Jenkinson and Powlson (1976) as modified by Srivastava and Singh (1988). Soil microbial N (SMBN) and P (SMBP) were extracted following chloroform fumigation extraction procedures using 0.5 M K2SO4 and Na-HCO3, respectively (Anderson and Ingram 1993). Soil respiration was measured by alkali trap method following Page et al. (1982). Percent base saturation was estimated as the proportion (in percent) of CEC contributed by exchangeable bases (Na, K, Ca, and Mg), while percent Al saturation was estimated as the proportion (in percent) of exch. Al in ECEC
– U (7), M (6), L (7)
– 12–73
17–31
73–118
14–22
13–38
Cultivated lowland (20) Natural forest (20)
Grassland (20)
Shifting cultivation (20)
Upland terraces (20)
Plantation land (20)
Adret (6), Ubac (6), semi-Adret (8) Adret (5), semi-adret (7), semi-adret, ubac (8)
Ubac (7), semi-adret (5), Adret (8)
Adret (6), ubac (6) semi-ubac (8)
– Adret (11), Semiadret (9)
Aspect
U, M, L upper, middle, and lower position
The numbers in parenthesis are the numbers of samples collected
U (6), M (8), L (6)
U(8), M (5), L (7)
U (7), M(8), L (5)
U (6), M (8), L (6)
Position
Gradient (°)
Land use types (no. of samples)
Table 1 Characteristics of the study sites
Valley (7), Terrace (7), Plateau (6) Valley (7), hill side (6), hill slope (7)
Hill slope (20)
Valley (4), hill side (6), plateau (3), hill slope (7)
Floodplain, valley Valley (5), hill side (6), plateau (4), hill slope (5)
Relief
2.00 2.00
32.00
2.00
2.00
2.00 2.00
Min.
32.00
32.00
32.00
32.00 32.00
Max.
Temperature (°C)
1,504–2,500
1,504–2,500
1,504–2,500
1,504–2,500
1,504–2,500 1,504–2,500
Mean annual rainfall (mm)
Pineapple(Ananas comosus), Banana (Musa sp.), Mango (Mangifera indica), Citrus (Citrus limon), Arecanut (Areca catechu), Cardamom (Elettaria cardamomum), Passion fruit (Passiflora edulis)
Lowland paddy (Oryza sativa) Bastard myrobalan (Terminalia bellerica), Ironwood (Mesua ferrea), Indian rubber (Ficus elastica), Beechwood (Gmelina arborea), Alder (Alnus nepalensis), White cedar (Chukrasia tabularis), Teak (Tectona grandis), East Indian Almond (Terminalia myriocarpa), Camel’s foot (Bauhinia variegata), Panic grass (Panicum spp.), Indian beech (Pongamia spp.), Needlewood (Schima wallichii) Nut grass (Cyperus rotundus), Lantana (Lantana camara), Goat weed (Ageratum conyzoides), Alligator weed (Alternanthera philoxeroides), Cobbler’s pegs (Bidens pilosa) Upland paddy (Oryza sativa), Maize (Zea mays L), Sweet potato (Ipomoea batatus) Tapioca (Manihot esculenta), Potato (Solanum tuberosum), Banana (Musa sp.), Pineapple (Ananas comosus), Citrus (Citrus limon), Ginger (Zingiber officinale) Maize (Zea mays L), Paddy (Oryza sativa), Millets
Vegetation
2016 Environ Monit Assess (2014) 186:2013–2024
Environ Monit Assess (2014) 186:2013–2024
2017
was used to segregate the significance of difference among the mean values. Pearson’s correlation coefficients were used to determine the strength of relationships among soil attributes. For determination of SQI, three steps were followed: (1) most critical soil quality indicators, i.e., minimum data set (MDS) of indicators that best represented soil function was selected; (2) the indicators were scored; and (3) the scores were integrated into final SQI. For selection of MDS, PCA was
of soil. Metabolic quotient was calculated as the ratio of BR and SMBC. Statistical analysis and soil quality indexing All data were checked for normality of distribution, and the analysis of variance (ANOVA) was performed using SPSS (version 16) to assess the effects of different land uses on soil properties. The Duncan’s multiple range test Table 2 Soil attributes as influenced by different land uses under study Soil attributes
Land use systems NFL
pH (1:2 soil/water) −1
Soil organic carbon (g kg ) −1
CLL
CUT
SC
PL
GL
5.3a
4.3e
4.6d
5.3a
4.9c
5.1b
a
d
f
e
c
26.5b
a
702.6
562.5d 148.4b
28.7
14.6
e
10.7
f
12.1
b
15.5
c
Sand (g kg )
559.3
537.5
Silt (g kg−1)
138.6c
210.5a
92.8e
110.6d
89.3f
a
d
e
c
f
−1
Clay (g kg )
302.1 −1
+
Exch. acidity [cmol(P )kg ] −1
+
−1
+
252.0 e
0.62
d
Exch. Al [cmol(P )kg ] CEC [cmol(P )kg ] % Base saturation
d
+
−1
Exch. Ca [cmol(P )kg ] Exch. Mg [cmol(P )kg ] −1
5.1b
d
28.9b
b
11.2d
c
172.6b
d
7.6e
22.5
0.16 e
90.1
32.8e
3.02
−1
b
Available B (mg kg )
0.40
−1
b
Available Mo (mg kg ) SMBC (μg g−1) −1
SMBN (μg g ) −1
Dehydrogenase (μg g −1
SR (μg CO2-C g
−1
h )
−1
h )
qCO2 (μg CO2-C μg−1 SMBC h−1)×10−3
2.14
a
0.52
a
75.1e
81.3
34.7d d
2.57
c
2.72
c
0.32
c
34.2d c
2.64
d
2.66
d
0.29
d
38.7b e
2.86a
e
2.83b
d
0.24e
e
0.14c
2.53 2.37 0.30
0.22
0.14
0.10
0.08
712.2a
257.9c
207.1e
240.3d
252.7c
a
c
f
d
e
24.6 b
SMBP (μg g )
f
1.2b
c
0.5
0.18 35.2
−1
2.48
2.8b
e
1.4
76.9
42.5a f
d
d
86.7
11.7 c
21.1 e
589.4b 31.3b
14.2 d
e
5.21
3.15
1.07
1.84
1.26
a
c
e
d
d
18.2
15.1
10.3
a
5.13
17.1b
19.9a
12.2
c
12.8
d
5.84a 16.7b
12.4
e
3.29
3.14
7.60e
13.7c
12.4d
1.57
0.17b
0.13
c
1.1
b
152.8b d
0.17 c
0.8
a
136.1d b
1.6
d
0.4
a
0.15 1.0
f
7.9
141.2c c
f
1.2
2.81
8.2
89.8f b
157.3
b
7.1
124.8e
b
127.9
f
8.1
32.7
e
113.7
c
a
36.3c
11.2
f
139.2
26.7
d
29.9
d
1.9
27.9
c
35.6
a
11.4
27.1
a
0.19
DTPA Mn (mg kg−1) DTPA Zn (mg kg )
15.6b
3.4d
71.2
−1
8.6
c
DTPA Fe (mg kg ) DTPA Cu (mg kg )
e
3.6c
f
−1
1.12
c
d
3.8
−1
+
0.57c
0.40
3.2f
a
Exch. Na [cmol(P )kg ]
0.69d
f
1.75
e
159.9a
−1
f
e
3.3e
8.5
+
0.96
0.54
a
a
Available K (mg kg−1)
b
a
6.8a
184.1
Available P (mg kg )
1.66
289.2b
208.1 f
8.3
a
−1
277.7 c
9.4
7.4
Available N (mg kg )
a
1.17
e
−1
1.72
a
34.2
% Al saturation
611.7
225.1 b
0.51 18.5
ECEC [cmol(P+)kg−1]
682.2
d
9.65b 16.4b
Values followed by different letters within the same parameter under different land uses are significantly different (P1 (Brejda et al. 2000) and those that explained at least 5 % of the variation in the data
(Sharma et al. 2005; Gui et al. 2010) were selected and subjected to varimax rotation to maximize correlation between PCs and the measured attributes (Shukla et al. 2006). Within each PC, the attribute with highest factor loading or the attribute with higher correlation sum (in case the factor loading of the second most weighted attribute was in very close proximity with the highest factor loading, and both were highly correlated too; this was done to avoid any redundancy in indicators’
Table 4 Results of principal component analysis showing principal components (PC) with their Eigenvalues and proportion of variance (in percent) explained, along with rotated factor loadings and communalities of soil attributes Soil attributes
PC-1
PC-2
PC-3
PC-4
PC-5
Communalities
pH
0.388
0.826
−0.345
0.196
0.003
0.990
SOC
0.933
0.287
0.189
0.021
−0.007
0.990
sand
−0.322
−0.278
−0.867
−0.205
0.109
0.987
silt
0.148
−0.195
0.920
0.230
−0.160
0.984
clay
0.416
0.752
0.487
0.100
−0.007
0.985
CEC
0.751
0.593
0.245
0.041
0.104
0.987
ECEC
0.822
0.457
0.158
0.025
0.293
0.996
Exch. acidity
−0.296
−0.936
−0.097
−0.113
0.085
0.994
Exch. Al
−0.192
−0.976
0.002
−0.021
0.020
0.990
PBS
0.603
0.602
−0.232
−0.076
0.412
0.955
PAlS
−0.403
−0.913
−0.023
−0.027
−0.001
0.997
Available N
0.946
0.122
0.031
0.273
−0.034
0.987
Available P
0.251
0.285
0.246
0.842
0.247
0.974
Available K
0.660
0.454
0.053
0.553
−0.180
0.982
Exch. Na
0.248
0.568
0.537
−0.082
0.402
0.841
Exch. Ca
0.822
0.510
0.153
0.041
0.183
0.995
Exch. Mg
0.568
0.706
0.083
−0.062
0.395
0.988
DTPA Fe
−0.598
−0.715
0.241
−0.150
−0.020
0.950
DTPA Mn
0.038
0.145
−0.167
−0.935
0.087
0.932
DTPA Cu
0.684
0.674
0.040
−0.209
−0.140
0.988 0.986
DTPA Zn
0.462
0.722
−0.156
−0.406
0.249
Available B
−0.089
−0.427
0.672
0.342
0.459
0.970
Available Mo
0.152
−0.138
0.883
−0.051
0.301
0.914
SMBC
0.883
0.399
0.215
−0.035
0.102
0.995
SMBN
0.676
0.427
0.535
0.200
−0.022
0.966
SMBP
0.750
0.315
0.474
−0.037
−0.193
0.926
Dehydrogenase
0.721
0.229
0.460
0.246
−0.019
0.844
Soil respiration
0.846
0.337
0.366
0.093
0.061
0.976
Metabolic quotient
0.363
0.001
0.648
0.516
−0.192
0.854
Eigenvalues
17.27
5.70
2.18
1.71
1.07
Variance explained (%)
59.54
19.64
7.52
5.89
3.67
Cumulative variance explained (%)
59.54
79.18
86.71
92.60
96.29
PBS Percent base saturation
Environ Monit Assess (2014) 186:2013–2024
selection), or the attributes with highest negative and positive loadings were selected for further scoring. Every observation of selected indicators was transformed into score using a linear scoring method (Andrews et al. 2002; Mandal et al. 2008). Indicators were categorized based on whether a higher value was considered “good” or “bad” in terms of soil function. For “more is better” indicators, each observation was divided by the highest observed value so that the highest observed value received a score of 1 while all others received a score of PL (0.54)>CUT (0.44). Highest SQI rating of NFL, followed closely by GL, can be understood based on their highest SOC (28.7 g kg −1 in NFL and 26.5 g kg−1 in GL) and relatively much lower exch. Al content [0.62 cmol(P+)kg−1 in NFL and 0.69 cmol (P+) kg−1 in GL], which were the two most powerful indicators of soil quality in present study. SOC contributed ~67 % and exch. Al contributed ~15 % in SQI rating of both these land uses. Higher SOC content in these land uses can be associated with its gradual build-up over the years, due to least anthropogenic disturbances and maximum carbon inputs from vegetation thereunder. Similar explanation holds good for lowest SQI rating of CUT, which had the lowest SOC content (10.7 g kg−1) and relatively much higher exch. Al [1.66 cmol(P+)kg−1]. It
Table 5 Soil quality index statistics for six land use types and their classification Land use
samples
SQI (mean)
Std. deviation
Std. Error
Min.
Max.
Soil quality classification
NFL
20
0.927a
0.014
0.003
0.9
0.95
High
CLL
20
0.569d
0.010
0.002
0.55
0.59
Medium
CUT
20
f
0.435
0.015
0.003
0.41
0.46
Low
SC
20
0.596c
0.017
0.003
0.56
0.64
Medium
PL
20
0.542e
0.012
0.002
0.53
0.57
Medium
GL
20
0.866b
0.018
0.004
0.82
0.89
High
Total
120
0.654
0.177
0.016
0.41
0.95
Medium
Means followed by different letters are different at P=0.05 SQI0.75, high NFL natural forestland, CLL cultivated lowland, CUT cultivated upland terrace, SC shifting cultivation, PL plantation land, GL grassland
Environ Monit Assess (2014) 186:2013–2024
is worth noting that CUT system underwent most exhaustive cropping practices during last two decades, with practically no external inputs of chemical fertilizers or organic manures. SQI ratings of the remaining three LUSs (i.e., SC, CLL, and PL) were significantly better than that of CUT, which may be due to relatively less intensive cultivation practices and more (though not sufficient) supply of external inputs. This was also reflected in higher SOC contents of these land uses relative to CUT. Interestingly, SQI rating of SC system was better than those of other three systems including CLL, PL and CUT. This could be due to 4 years of fallow period maintained after every 3–4 years of cultivation in SC system, which might have helped restoring the soil fertility and relieving the “cultivation stress”. On the whole, percentage contribution of the selected indicators in determination of SQI was in the order: SOC (58 %) > exchangeable Al (17.1 %) > available P (8.9 %)>available Mn (8.2 %)> silt content (7.8 %). Since classification of soil quality could be useful for devising land management strategies (Xu et al. 2006b), we also classified the soils of the study sites in different quality categories based on their SQI (Table 5). The soils under natural forest and grassland were “high” in quality rating (SQI>0.75), while those under intensively cultivated upland terrace were in “low” category (SQI< 0.50), with the remaining land use types classified as “medium” (0.50
Land use impact on soil quality in eastern Himalayan region of India A. K. Singh & L. J. Bordoloi & Manoj Kumar & S. Hazarika & Brajendra Parmar
Received: 13 February 2013 / Accepted: 28 October 2013 / Published online: 14 November 2013 # Springer Science+Business Media Dordrecht 2013
Abstract Quantitative assessment of soil quality is required to determine the sustainability of land uses in terms of environmental quality and plant productivity. Our objective was to identify the most appropriate soil quality indicators and to evaluate the impact of six most prevalent land use types (natural forestland, cultivated lowland, cultivated upland terrace, shifting cultivation, plantation land, and grassland) on soil quality in eastern Himalayan region of India. We collected 120 soil samples (20 cm depth) and analyzed them for 29 physical, chemical, and biological soil attributes. For selection of soil quality indicators, principal component analysis (PCA) was performed on the measured attributes, which provided four principal components (PC) with eigenvalues >1 and explaining at least 5 % of the variance in dataset. The four PCs together explained 92.6 % of the total variance. Based on rotated factor loadings of soil attributes, selected indicators were: soil organic carbon
A. K. Singh Department of Agricultural Chemistry and Soil Science, SASRD, Nagaland University, Medziphema 797106 Nagaland, India L. J. Bordoloi : M. Kumar (*) : S. Hazarika Division of Natural Resource Management (Soil Science), ICAR Research Complex for NEH Region, Umiam 793103 Meghalaya, India e-mail: [email protected] B. Parmar Division of Soil Science, Directorate of Rice Research, Rajendranagar 500030 Hyderabad, India
(SOC) from PC-1, exchangeable Al from PC-2, silt content from PC-3, and available P and Mn from PC-4. Indicators were transformed into scores (linear scoring method) and soil quality index (SQI) was determined, on a scale of 0–1, using the weighting factors obtained from PCA. SQI rating was the highest for the least-disturbed sites, i.e., natural forestland (0.93) and grassland (0.87), and the lowest for the most intensively cultivated site, i.e., cultivated upland terrace (0.44). Ratings for the other land uses were shifting cultivation (0.60)>cultivated low land (0.57)>plantation land (0.54). Overall contribution (in percent) of the indicators in determination of SQI was in the order: SOC (58 %)>exch. Al (17.1 %)>available P (8.9 %)>available Mn (8.2 %)>silt content (7.8 %). Results of this study suggest SOC and exch. Al as the two most powerful indicators of soil quality in study area. Thus, organic C and soil acidity management holds the key to improve soil quality under many exploitatively cultivated land use systems in eastern Himalayan region of India. Keywords Northeastern hill regions . Soil quality indicators . Shifting cultivation . Land use changes
Introduction Soil quality is defined as “the capacity of a soil to function within ecosystem and land-use boundaries, to sustain biological productivity, maintain environmental quality, and promote plant and animal health” (Doran and Parkin 1994). This clearly establishes soil quality as
2014
a vital determinant of agricultural productivity as well as environmental sustainability, the two most important challenges facing the world today. In order to meet the needs of growing population, there have been drastic changes in land use and soil management practices in recent years, which are considered to have caused definite deterioration in soil and environmental quality. Such changes in land use systems and the associated impacts are however not uniform across the regions. Therefore, to determine the sustainability of land management systems prevalent in a region, quantitative assessment of soil quality on regional scale is necessary. Soil quality index (SQI) is an easily comprehensible way of ascertaining whether soil quality is improving, stable, or declining under different land use systems (Masto et al. 2008). It is comprised of indicators sensitive to regional scale changes in land use, and could be useful to researchers and policy makers for assessing the effects of different land use and management practices on soil quality over broad geographical areas (Brejda and Moorman 2001). Soil quality is broadly a function of soil properties but it may not be adequately represented by individual soil properties; SQI derived from integrated soil quality indicators based on a combination of soil properties could better reflect the status of soil quality (Masto et al. 2009). There are three main steps involved in soil quality indexing: first, selecting appropriate indicators; second, transforming indicators to scores; and third, combining the scores into index (Andrews et al. 2002). Indicators’ selection can be made using expert opinion (Doran and Parkin 1994) or statistical techniques like principal component analysis (PCA; Andrews et al. 2001, 2002; Sharma et al. 2005) and factor analysis (Brejda et al. 2000). Scoring of indicators can be done using linear and nonlinear scoring methods (Andrews et al. 2002) and, finally, using multiplicative (Singh et al. 1992), simple additive (Andrews and Carroll 2001), or weighted additive (Karlen et al. 1998) methods, scores can be integrated into final SQI. Although there are several studies on the effect of land uses and soil management practices on soil quality in different regions of the world including IndoGangetic Plains of India (Xu et al. 2006a, b; Masto et al. 2007, 2009; Ayoubi et al. 2011), no such information is available for the eastern Himalayan regions of India which has witnessed drastic changes in land use in recent times. This region, consisting of seven hilly states
Environ Monit Assess (2014) 186:2013–2024
of India, is characterized by diverse agroclimatic and geographical situations. About 65 % of the total geographical area is under forests, ~16 % under crops, and the rest either under nonagricultural uses or uncultivated land (Saha et al. 2012). Extreme forms of soil acidity and unacceptably high rates of soil erosion, induced by high rainfall and sloppy terrains have been the major constraints for agricultural production in majority of the areas, which lags far behind the national average crops’ productivity. Besides, negligible adoption of the soil acidity amelioration practices (such as liming), minimal use of external nutrient inputs, loss of top soil along with the nutrients and organic carbon thereof, and subsequent decline in overall soil quality further dampen the prospect of enhancing crop productivity in the region. Under the pressure of fast-growing population’s need, natural forestlands are being increasingly cleared to be put under different land uses with exhaustive cultivation practices, such as shifting cultivation (also known as slash and burn cultivation, and locally known as Jhum cultivation), terrace cultivation, etc. These land use changes are often implicated for the declining soil quality in the region, which could be a major obstacle towards increasing crop production in future. Thus, identification of the major soil quality determining parameters under the prevailing land use types of the region is crucial for evolving measures to improve soil quality and its productivity in future. Since such studies have not been undertaken in the region so far, objective of this study was to identify the appropriate soil quality indicators and to evaluate the differences in soil quality (using soil quality indexing approach) under the six most prevalent land use types in eastern Himalayan region of India.
Materials and methods Description of study sites The study area is located in Dimapur district of Nagaland state (25°06′–27°04′ N, 93°21′–95°15′ E) in eastern Himalayan region of India (Fig. 1). Average altitude ranges from 260 to 690 m, with average annual temperature ranging from 18 to 20 and 23 to 25 °C in the higher and lower elevations, respectively. Average annual rainfall is 1,825 mm, 90 % of which is experienced during June to September. The soils of the study sites were acid Alfisols (rich in Fe and Al). Detailed
Environ Monit Assess (2014) 186:2013–2024
2015
receives a regular but small application of chemical fertilizers (5–10 kg ha−1 of NPK). Soil sampling and analysis
Fig. 1 Location of the study area shown on India’s map
description of these sites is shown in Table 1. We identified six adjacently located land use systems (LUS) viz., natural forestland (NFL), cultivated lowland (CLL), cultivated upland terrace (CUT), shifting cultivation (SC), plantation land (PL), and grassland (GL) for this study. Group discussion and key informants’ interviews revealed that these LUSs were in place for at least past 20 years. The NFL represents an undisturbed forest site with majority of the vegetation comprising of Terminalia bellerica, Mesua ferrea, Ficus elastica, Gmelina arborea, Alnus nepalensis, Tectona grandis, Bauhinia variegate, Panicum spp., Pongamia spp., Schima wallichii, etc. Monocropping with local paddy cultivars (during July–August) is practiced in CLL site where application of FYM at 5–7 Mg ha−1 before paddy transplanting is common. CUT are mostly under nutrient exhaustive annual crops (paddy, maize, millets, etc.), and represent the most intensively cultivated system in this study. These terraces receive no application of chemical fertilizers, with only nominal applications of organic manures at times (~0.25 Mg ha−1). The SC site has been subject to a cycle of 3–4 years of cultivation followed by 4 years of fallow periods. Nutrient supply in the first 2 years of cultivation comes entirely from the ashes of the slashed and burnt vegetation, with some external application of organic manures in third/fourth years (~2 Mg ha−1). Rice, maize, tapioca, sweet potato, banana, pine apple, etc. are the commonly grown crops in this land use. PL site is mostly under mixed plantations of citrus, pineapple, banana and mango, and
Altogether, 120 composite surface (0–20 cm depth) soil samples (each composite sample derived from three randomly collected subsamples) from six land uses were collected (October 2009), air dried, ground, and passed through 2-mm sieve (0.5 mm sieve for organic carbon estimation). Samples were analyzed for 29 soil attributes as follows: Soil pH (1:2 soil/water) was measured using glass electrode. Soil organic carbon (SOC) was determined by wet digestion method (Walkley and Black 1934). Cation exchange capacity (CEC) was estimated by sodium saturation method (Jackson 1973). Particle size distribution was determined by international pipette method (Page et al. 1982). Available N was estimated using alkaline KMnO4 distillation method (Subbiah and Asija 1956); available P by Bray-II method (Bray and Kurtz 1945). Exchangeable K and Na were determined using ammonium acetate extraction followed by emission spectrometry (Jackson 1973). Exch. Ca and Mg were determined through versene titration (Baruah and Barthakur 1999), and exch. Al and acidity by potassium chloride method (Page et al. 1982). Effective CEC (ECEC) was calculated as the sum total of exchangeable cations (Na, K, Ca, Mg, and Al). Available Fe, Mn, Cu, and Zn were determined using 0.005 M DTPA extraction followed by atomic absorption spectrometry (Lindsay and Norvell 1978). Available B and Mo were measured colorimetrically using hot water and ammonium oxalate solution as extractants, respectively (Page et al. 1982). Soil dehydrogenase activity was estimated through reduction of TTC (triphenyltetrazolium chloride) to TPF (triphenylformazan) (Page et al. 1982). Soil microbial biomass C (SMBC) was estimated by chloroform fumigation–incubation procedure given by Jenkinson and Powlson (1976) as modified by Srivastava and Singh (1988). Soil microbial N (SMBN) and P (SMBP) were extracted following chloroform fumigation extraction procedures using 0.5 M K2SO4 and Na-HCO3, respectively (Anderson and Ingram 1993). Soil respiration was measured by alkali trap method following Page et al. (1982). Percent base saturation was estimated as the proportion (in percent) of CEC contributed by exchangeable bases (Na, K, Ca, and Mg), while percent Al saturation was estimated as the proportion (in percent) of exch. Al in ECEC
– U (7), M (6), L (7)
– 12–73
17–31
73–118
14–22
13–38
Cultivated lowland (20) Natural forest (20)
Grassland (20)
Shifting cultivation (20)
Upland terraces (20)
Plantation land (20)
Adret (6), Ubac (6), semi-Adret (8) Adret (5), semi-adret (7), semi-adret, ubac (8)
Ubac (7), semi-adret (5), Adret (8)
Adret (6), ubac (6) semi-ubac (8)
– Adret (11), Semiadret (9)
Aspect
U, M, L upper, middle, and lower position
The numbers in parenthesis are the numbers of samples collected
U (6), M (8), L (6)
U(8), M (5), L (7)
U (7), M(8), L (5)
U (6), M (8), L (6)
Position
Gradient (°)
Land use types (no. of samples)
Table 1 Characteristics of the study sites
Valley (7), Terrace (7), Plateau (6) Valley (7), hill side (6), hill slope (7)
Hill slope (20)
Valley (4), hill side (6), plateau (3), hill slope (7)
Floodplain, valley Valley (5), hill side (6), plateau (4), hill slope (5)
Relief
2.00 2.00
32.00
2.00
2.00
2.00 2.00
Min.
32.00
32.00
32.00
32.00 32.00
Max.
Temperature (°C)
1,504–2,500
1,504–2,500
1,504–2,500
1,504–2,500
1,504–2,500 1,504–2,500
Mean annual rainfall (mm)
Pineapple(Ananas comosus), Banana (Musa sp.), Mango (Mangifera indica), Citrus (Citrus limon), Arecanut (Areca catechu), Cardamom (Elettaria cardamomum), Passion fruit (Passiflora edulis)
Lowland paddy (Oryza sativa) Bastard myrobalan (Terminalia bellerica), Ironwood (Mesua ferrea), Indian rubber (Ficus elastica), Beechwood (Gmelina arborea), Alder (Alnus nepalensis), White cedar (Chukrasia tabularis), Teak (Tectona grandis), East Indian Almond (Terminalia myriocarpa), Camel’s foot (Bauhinia variegata), Panic grass (Panicum spp.), Indian beech (Pongamia spp.), Needlewood (Schima wallichii) Nut grass (Cyperus rotundus), Lantana (Lantana camara), Goat weed (Ageratum conyzoides), Alligator weed (Alternanthera philoxeroides), Cobbler’s pegs (Bidens pilosa) Upland paddy (Oryza sativa), Maize (Zea mays L), Sweet potato (Ipomoea batatus) Tapioca (Manihot esculenta), Potato (Solanum tuberosum), Banana (Musa sp.), Pineapple (Ananas comosus), Citrus (Citrus limon), Ginger (Zingiber officinale) Maize (Zea mays L), Paddy (Oryza sativa), Millets
Vegetation
2016 Environ Monit Assess (2014) 186:2013–2024
Environ Monit Assess (2014) 186:2013–2024
2017
was used to segregate the significance of difference among the mean values. Pearson’s correlation coefficients were used to determine the strength of relationships among soil attributes. For determination of SQI, three steps were followed: (1) most critical soil quality indicators, i.e., minimum data set (MDS) of indicators that best represented soil function was selected; (2) the indicators were scored; and (3) the scores were integrated into final SQI. For selection of MDS, PCA was
of soil. Metabolic quotient was calculated as the ratio of BR and SMBC. Statistical analysis and soil quality indexing All data were checked for normality of distribution, and the analysis of variance (ANOVA) was performed using SPSS (version 16) to assess the effects of different land uses on soil properties. The Duncan’s multiple range test Table 2 Soil attributes as influenced by different land uses under study Soil attributes
Land use systems NFL
pH (1:2 soil/water) −1
Soil organic carbon (g kg ) −1
CLL
CUT
SC
PL
GL
5.3a
4.3e
4.6d
5.3a
4.9c
5.1b
a
d
f
e
c
26.5b
a
702.6
562.5d 148.4b
28.7
14.6
e
10.7
f
12.1
b
15.5
c
Sand (g kg )
559.3
537.5
Silt (g kg−1)
138.6c
210.5a
92.8e
110.6d
89.3f
a
d
e
c
f
−1
Clay (g kg )
302.1 −1
+
Exch. acidity [cmol(P )kg ] −1
+
−1
+
252.0 e
0.62
d
Exch. Al [cmol(P )kg ] CEC [cmol(P )kg ] % Base saturation
d
+
−1
Exch. Ca [cmol(P )kg ] Exch. Mg [cmol(P )kg ] −1
5.1b
d
28.9b
b
11.2d
c
172.6b
d
7.6e
22.5
0.16 e
90.1
32.8e
3.02
−1
b
Available B (mg kg )
0.40
−1
b
Available Mo (mg kg ) SMBC (μg g−1) −1
SMBN (μg g ) −1
Dehydrogenase (μg g −1
SR (μg CO2-C g
−1
h )
−1
h )
qCO2 (μg CO2-C μg−1 SMBC h−1)×10−3
2.14
a
0.52
a
75.1e
81.3
34.7d d
2.57
c
2.72
c
0.32
c
34.2d c
2.64
d
2.66
d
0.29
d
38.7b e
2.86a
e
2.83b
d
0.24e
e
0.14c
2.53 2.37 0.30
0.22
0.14
0.10
0.08
712.2a
257.9c
207.1e
240.3d
252.7c
a
c
f
d
e
24.6 b
SMBP (μg g )
f
1.2b
c
0.5
0.18 35.2
−1
2.48
2.8b
e
1.4
76.9
42.5a f
d
d
86.7
11.7 c
21.1 e
589.4b 31.3b
14.2 d
e
5.21
3.15
1.07
1.84
1.26
a
c
e
d
d
18.2
15.1
10.3
a
5.13
17.1b
19.9a
12.2
c
12.8
d
5.84a 16.7b
12.4
e
3.29
3.14
7.60e
13.7c
12.4d
1.57
0.17b
0.13
c
1.1
b
152.8b d
0.17 c
0.8
a
136.1d b
1.6
d
0.4
a
0.15 1.0
f
7.9
141.2c c
f
1.2
2.81
8.2
89.8f b
157.3
b
7.1
124.8e
b
127.9
f
8.1
32.7
e
113.7
c
a
36.3c
11.2
f
139.2
26.7
d
29.9
d
1.9
27.9
c
35.6
a
11.4
27.1
a
0.19
DTPA Mn (mg kg−1) DTPA Zn (mg kg )
15.6b
3.4d
71.2
−1
8.6
c
DTPA Fe (mg kg ) DTPA Cu (mg kg )
e
3.6c
f
−1
1.12
c
d
3.8
−1
+
0.57c
0.40
3.2f
a
Exch. Na [cmol(P )kg ]
0.69d
f
1.75
e
159.9a
−1
f
e
3.3e
8.5
+
0.96
0.54
a
a
Available K (mg kg−1)
b
a
6.8a
184.1
Available P (mg kg )
1.66
289.2b
208.1 f
8.3
a
−1
277.7 c
9.4
7.4
Available N (mg kg )
a
1.17
e
−1
1.72
a
34.2
% Al saturation
611.7
225.1 b
0.51 18.5
ECEC [cmol(P+)kg−1]
682.2
d
9.65b 16.4b
Values followed by different letters within the same parameter under different land uses are significantly different (P1 (Brejda et al. 2000) and those that explained at least 5 % of the variation in the data
(Sharma et al. 2005; Gui et al. 2010) were selected and subjected to varimax rotation to maximize correlation between PCs and the measured attributes (Shukla et al. 2006). Within each PC, the attribute with highest factor loading or the attribute with higher correlation sum (in case the factor loading of the second most weighted attribute was in very close proximity with the highest factor loading, and both were highly correlated too; this was done to avoid any redundancy in indicators’
Table 4 Results of principal component analysis showing principal components (PC) with their Eigenvalues and proportion of variance (in percent) explained, along with rotated factor loadings and communalities of soil attributes Soil attributes
PC-1
PC-2
PC-3
PC-4
PC-5
Communalities
pH
0.388
0.826
−0.345
0.196
0.003
0.990
SOC
0.933
0.287
0.189
0.021
−0.007
0.990
sand
−0.322
−0.278
−0.867
−0.205
0.109
0.987
silt
0.148
−0.195
0.920
0.230
−0.160
0.984
clay
0.416
0.752
0.487
0.100
−0.007
0.985
CEC
0.751
0.593
0.245
0.041
0.104
0.987
ECEC
0.822
0.457
0.158
0.025
0.293
0.996
Exch. acidity
−0.296
−0.936
−0.097
−0.113
0.085
0.994
Exch. Al
−0.192
−0.976
0.002
−0.021
0.020
0.990
PBS
0.603
0.602
−0.232
−0.076
0.412
0.955
PAlS
−0.403
−0.913
−0.023
−0.027
−0.001
0.997
Available N
0.946
0.122
0.031
0.273
−0.034
0.987
Available P
0.251
0.285
0.246
0.842
0.247
0.974
Available K
0.660
0.454
0.053
0.553
−0.180
0.982
Exch. Na
0.248
0.568
0.537
−0.082
0.402
0.841
Exch. Ca
0.822
0.510
0.153
0.041
0.183
0.995
Exch. Mg
0.568
0.706
0.083
−0.062
0.395
0.988
DTPA Fe
−0.598
−0.715
0.241
−0.150
−0.020
0.950
DTPA Mn
0.038
0.145
−0.167
−0.935
0.087
0.932
DTPA Cu
0.684
0.674
0.040
−0.209
−0.140
0.988 0.986
DTPA Zn
0.462
0.722
−0.156
−0.406
0.249
Available B
−0.089
−0.427
0.672
0.342
0.459
0.970
Available Mo
0.152
−0.138
0.883
−0.051
0.301
0.914
SMBC
0.883
0.399
0.215
−0.035
0.102
0.995
SMBN
0.676
0.427
0.535
0.200
−0.022
0.966
SMBP
0.750
0.315
0.474
−0.037
−0.193
0.926
Dehydrogenase
0.721
0.229
0.460
0.246
−0.019
0.844
Soil respiration
0.846
0.337
0.366
0.093
0.061
0.976
Metabolic quotient
0.363
0.001
0.648
0.516
−0.192
0.854
Eigenvalues
17.27
5.70
2.18
1.71
1.07
Variance explained (%)
59.54
19.64
7.52
5.89
3.67
Cumulative variance explained (%)
59.54
79.18
86.71
92.60
96.29
PBS Percent base saturation
Environ Monit Assess (2014) 186:2013–2024
selection), or the attributes with highest negative and positive loadings were selected for further scoring. Every observation of selected indicators was transformed into score using a linear scoring method (Andrews et al. 2002; Mandal et al. 2008). Indicators were categorized based on whether a higher value was considered “good” or “bad” in terms of soil function. For “more is better” indicators, each observation was divided by the highest observed value so that the highest observed value received a score of 1 while all others received a score of PL (0.54)>CUT (0.44). Highest SQI rating of NFL, followed closely by GL, can be understood based on their highest SOC (28.7 g kg −1 in NFL and 26.5 g kg−1 in GL) and relatively much lower exch. Al content [0.62 cmol(P+)kg−1 in NFL and 0.69 cmol (P+) kg−1 in GL], which were the two most powerful indicators of soil quality in present study. SOC contributed ~67 % and exch. Al contributed ~15 % in SQI rating of both these land uses. Higher SOC content in these land uses can be associated with its gradual build-up over the years, due to least anthropogenic disturbances and maximum carbon inputs from vegetation thereunder. Similar explanation holds good for lowest SQI rating of CUT, which had the lowest SOC content (10.7 g kg−1) and relatively much higher exch. Al [1.66 cmol(P+)kg−1]. It
Table 5 Soil quality index statistics for six land use types and their classification Land use
samples
SQI (mean)
Std. deviation
Std. Error
Min.
Max.
Soil quality classification
NFL
20
0.927a
0.014
0.003
0.9
0.95
High
CLL
20
0.569d
0.010
0.002
0.55
0.59
Medium
CUT
20
f
0.435
0.015
0.003
0.41
0.46
Low
SC
20
0.596c
0.017
0.003
0.56
0.64
Medium
PL
20
0.542e
0.012
0.002
0.53
0.57
Medium
GL
20
0.866b
0.018
0.004
0.82
0.89
High
Total
120
0.654
0.177
0.016
0.41
0.95
Medium
Means followed by different letters are different at P=0.05 SQI0.75, high NFL natural forestland, CLL cultivated lowland, CUT cultivated upland terrace, SC shifting cultivation, PL plantation land, GL grassland
Environ Monit Assess (2014) 186:2013–2024
is worth noting that CUT system underwent most exhaustive cropping practices during last two decades, with practically no external inputs of chemical fertilizers or organic manures. SQI ratings of the remaining three LUSs (i.e., SC, CLL, and PL) were significantly better than that of CUT, which may be due to relatively less intensive cultivation practices and more (though not sufficient) supply of external inputs. This was also reflected in higher SOC contents of these land uses relative to CUT. Interestingly, SQI rating of SC system was better than those of other three systems including CLL, PL and CUT. This could be due to 4 years of fallow period maintained after every 3–4 years of cultivation in SC system, which might have helped restoring the soil fertility and relieving the “cultivation stress”. On the whole, percentage contribution of the selected indicators in determination of SQI was in the order: SOC (58 %) > exchangeable Al (17.1 %) > available P (8.9 %)>available Mn (8.2 %)> silt content (7.8 %). Since classification of soil quality could be useful for devising land management strategies (Xu et al. 2006b), we also classified the soils of the study sites in different quality categories based on their SQI (Table 5). The soils under natural forest and grassland were “high” in quality rating (SQI>0.75), while those under intensively cultivated upland terrace were in “low” category (SQI< 0.50), with the remaining land use types classified as “medium” (0.50
