do land characteristics affect farmers' soil fertility management?

12
Journal of Integrative Agriculture 2014, 13(11): 2546-2557 November 2014 RESEARCH ARTICLE © 2014, CAAS. All rights reserved. Published by Elsevier Ltd. doi: 10.1016/S2095-3119(14)60840-6 Do Land Characteristics Affect Farmers’ Soil Fertility Management? TAN Shu-hao School of Agricultural Economics and Rural Development, Renmin University of China, Beijing 100872, P.R.China Abstract Soil fertility management (SFM) has important implications for sustaining agricultural development and food self-sufficiency. Better understanding the determinants of farmers’ SFM can be a great help to the adoption of effective SFM practices. Based on a dataset of 315 plots collected from a typical rice growing area of South China, this study applied statistical method and econometric models to examine the impacts of land characteristics on farmers’ SFM practices at plot scale. Main results showed that in general land characteristics affected SFM behaviors. Securer land tenure arrangements facilitated effective practices of SFM through more diversified and more soil-friendly cropping pattern choices. Plot size significantly reduced the intensities of phosphorus and potassium fertilizer application. Given other factors, 1 ha increase in plot size might reduce 3.0 kg ha -1 P 2 O 5 and 1.8 kg ha -1 K 2 O. Plots far from the homestead were paid less attention in terms of both chemical fertilizers and manure applications. Besides, plots with better quality were put more efforts on management by applying more nitrogen and manure, and by planting green manure crops. Significant differences existed in SFM practices between the surveyed villages with different socio-economic conditions. The findings are expected to provide important references to the policy-making incentive for improving soil quality and crop productivity. Key words: land characteristics, soil fertility management, farm household, rice cropping, South China INTRODUCTION Soil degradation on agricultural land has been viewed as a serious environmental and economic problem in many developing countries (Koning et al. 2001; Scherr and Yadav 2001; Sanchez 2002; Heerink 2005; Rasul and Thapa 2007; Ye and van Ranst 2009). In China, a resource-poor country with pressure of securing the huge population’s food self-sufficiency, it is especially severe (Lindert 1999; Yang 2006; Guo et al. 2010). The goal of soil fertility management (SFM), which normally includes soil tillage, fertilization, irrigation, cropping systems and straw application, is to create good soil conditions for crop growing with high yield (Greenland and Nabhan 2001; Yang 2006). Since soil quality is fundamental for agricultural production, improving SFM is becoming increasingly crucial for policy-making with regard to food security, poverty reduction and environ- mental protection. Although most SFM practices can facilitate higher yield in a short-term, their net effects on soil quality and thus on potential agricultural productivity can be very different in a long-term. Inappropriate SFM practices, such as long-term mono-cropping, overused/unbalanced application of chemical fertilizers, and lack of farm yard manure application, can accelerate land degradation and threaten the agricultural sustainability and environmental health (Zhen et al. 2006). Overused chemical nitrogen Received 28 October, 2013 Accepted 18 February, 2014 Correspondence TAN Shu-hao, Fax: +86-10-62511064, E-mail: [email protected]

Upload: shu-hao

Post on 24-Mar-2017

214 views

Category:

Documents


1 download

TRANSCRIPT

Page 1: Do Land Characteristics Affect Farmers' Soil Fertility Management?

Journal of Integrative Agriculture2014, 13(11): 2546-2557 November 2014RESEARCH ARTICLE

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.doi: 10.1016/S2095-3119(14)60840-6

Do Land Characteristics Affect Farmers’ Soil Fertility Management?

Tan Shu-hao

School of Agricultural Economics and Rural Development, Renmin University of China, Beijing 100872, P.R.China

Abstract

Soil fertility management (SFM) has important implications for sustaining agricultural development and food self-sufficiency. Better understanding the determinants of farmers’ SFM can be a great help to the adoption of effective SFM practices. Based on a dataset of 315 plots collected from a typical rice growing area of South China, this study applied statistical method and econometric models to examine the impacts of land characteristics on farmers’ SFM practices at plot scale. Main results showed that in general land characteristics affected SFM behaviors. Securer land tenure arrangements facilitated effective practices of SFM through more diversified and more soil-friendly cropping pattern choices. Plot size significantly reduced the intensities of phosphorus and potassium fertilizer application. Given other factors, 1 ha increase in plot size might reduce 3.0 kg ha-1 P2O5

and 1.8 kg ha-1 K2O. Plots far from the homestead were paid less attention in terms of both chemical fertilizers and manure applications. Besides, plots with better quality were put more efforts on management by applying more nitrogen and manure, and by planting green manure crops. Significant differences existed in SFM practices between the surveyed villages with different socio-economic conditions. The findings are expected to provide important references to the policy-making incentive for improving soil quality and crop productivity.

Key words: land characteristics, soil fertility management, farm household, rice cropping, South China

INTRODUCTION

Soil degradation on agricultural land has been viewed as a serious environmental and economic problem in many developing countries (Koning et al. 2001; Scherr and Yadav 2001; Sanchez 2002; Heerink 2005; Rasul and Thapa 2007; Ye and van Ranst 2009). In China, a resource-poor country with pressure of securing the huge population’s food self-sufficiency, it is especially severe (Lindert 1999; Yang 2006; Guo et al. 2010). The goal of soil fertility management (SFM), which normally includes soil tillage, fertilization, irrigation, cropping systems and straw application, is to create good soil

conditions for crop growing with high yield (Greenland and Nabhan 2001; Yang 2006). Since soil quality is fundamental for agricultural production, improving SFM is becoming increasingly crucial for policy-making with regard to food security, poverty reduction and environ-mental protection.

although most SFM practices can facilitate higher yield in a short-term, their net effects on soil quality and thus on potential agricultural productivity can be very different in a long-term. Inappropriate SFM practices, such as long-term mono-cropping, overused/unbalanced application of chemical fertilizers, and lack of farm yard manure application, can accelerate land degradation and threaten the agricultural sustainability and environmental health (Zhen et al. 2006). Overused chemical nitrogen

Received 28 October, 2013 Accepted 18 February, 2014Correspondence TAN Shu-hao, Fax: +86-10-62511064, E-mail: [email protected]

Page 2: Do Land Characteristics Affect Farmers' Soil Fertility Management?

Do Land Characteristics Affect Farmers’ Soil Fertility Management? 2547

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

fertilizer has caused significant acidification in major Chinese croplands since the 1980s (Guo et al. 2010), which has been a serious challenge to food self-suffi-ciency in China. Effective SFM practices, such as bal-anced application of chemical fertilizers, crop rotation, crop straw recycling and manure application, have been proved to play a vital role in sustaining agriculture by building up soil organic matter (Moreno et al. 2006; Chen et al. 2009; Sombrero and de Benito 2010), and promoting soil and water conservation (Huang et al. 2008). Effective SFM practices have also been regarded as a cost-effective way of cooling down the earth and improving agricultural sustainability and environmental health (IPCC 2007). However, the effective SFM prac-tices have not been popularly adopted by farmers world-wide till now (Li et al. 2011). Knowing which factors determine the adoption of SFM practices and how, can provide important references to the policy-making in-centive for improving soil quality and crop productivity.

Socio-economic factors play an important role in farmer’s decision-making for SFM practices. During the past decades, a lot of attention has been paid to examine the determinants of SFM practices. However, most existing studies have focused on bio-physical aspects. In fact, farmers make their decisions related to SFM not only based on the farmland natural conditions and technological availabilities, but also on socio-economic factors. Fortunately, research on socio-economic factors has drawn increasing interests from researchers during the past decades. For example, Katz (2000), Deininger and Jin (2003) argued that secure land tenure and appro-priate land use policies could encourage farmers to better manage their soil by providing incentives. Omamo et al. (2002) explored the driving factors of small farms’ SFM in Kenya, and found that lower farm-to-market transport cost, or larger quantity of family labor significantly encouraged chemical fertilizer application. Tittonell et al. (2010) examined the effects of rural livelihood strategies on SFM in Kenya and Uganda, and found that farm livelihood strategies significantly influenced the soil management practices through cropping pattern choices. The existing researches have greatly enhanced the understanding of socio-economic factors’ influence on SFM practices.

Soil fertility is more related to specific plots, however, most available researches were performed at village or farmhouse scale. Thus, it is desirable to examine the

impacts of land characteristics on farmers’ SFM prac-tices at plot scale, especially in South China where some typical SFM problems like paddy soil acidification and soil compaction broadly exist. Moreover, land is very seriously fragmented due to the prevailing system of land allocation and land reallocation in this area. According to Tan et al. (2006), cultivated land area per household was 0.61 ha in 1986, and it became 0.53 ha in 1999; farm households had on average 8.43 plots in 1986 and 6.06 plots in 1999. Land fragmentation in South China was severer. Data from the Rural Fixed Observation showed that each farm household had 8.95 plots, with plot size less than 0.05 ha in Jiangxi in 1999 (Tan et al. 2006). Furthermore, based on a survey from 17 provinc-es during 1999 to 2010, land has been reallocated in all the surveyed provinces. In Jiangxi Province, more than 90% of the sample villages had reallocated their land at least 4 times. Although land consolidation program has been launched in some provinces during the past 2 decades, reducing land fragmentation to some extent, cultivated land was still severely fragmented. In 2010, each household on average only had 0.41 ha cultivated land, spreading over 4.4 plots (Feng et al. 2011).

as an important characteristic, land transfer is becom-ing more and more popular in China (Ye et al. 2010; Gao et al. 2012; Qin and Tan 2013). In 2008, transfer took place on 8.17% of the total contracted land. It was 11% in 2009 and 20% in 2012. A study of Gao et al. (2012) showed that the cultivated land rental rate was 19% for the sample households, almost doubled that of in 2000 (10%). The share of households involved in land transfer was as high as 50% in the developed areas like Shanghai. actually, land transfer was carried out by farmers at plot level. Land rental and other socio-economic characteris-tics have significant impacts on farmers’ SMF strategies (Gao et al. 2012), and have implications for sustaining agricultural development and food security (Yu et al. 2003; Tittonell et al. 2007). However, detailed empirical research on how plot characteristics affect farmers’ SFM strategies has not been fully understood.

Large uncertainties still remain on the impacts of plot factors on farmers’ SFM strategies. Related stud-ies are rare with a few exceptions such as Ali (1996), Sah et al. (2010) and several ones on China (Li et al. 1998; Yu et al. 2003; Gao et al. 2012). For example, Li et al. (1998) applied a dataset with 160 plots from 80 households, to examine the effects of land tenure on

Page 3: Do Land Characteristics Affect Farmers' Soil Fertility Management?

2548 Tan Shu-hao

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

production input intensity. They found that longer land use right encouraged land-saving investments, such as the application of organic manure and phosphate fertil-izer. Later, Yu et al. (2003) used a two-phase dataset with 180 plots to examine the impacts of use rights security and land transfer on agricultural land resource degradation. Their main findings were that secure use rights improved long-term soil fertility, and the infor-mal land transfer among farm households gave rise to soil degradation. More recently, Gao et al. (2012) also applied a two-phase household dataset to examine how land rental markets affected agricultural investment in China. The authors argued that the rise of land rental markets partly explained the fall of the sampled house-holds’ manure use from 13 t ha-1 in 2000 to 5 t ha-1 in 2008. Although they examined the effects of tenure on land investments very well, these researches mainly focused on manure application, which is only part of the SFM practices.

In order to better understand how plot-level char-acteristics affect farmers’ SFM, this study applied an in-depth surveyed dataset to examine their effects on farmers’ cropping pattern choices, fertilization (i.e., chemical fertilizer and manure application), and green manure. Main results are expected to serve as reference for designing policies with regard to promoting SFM, alleviating resource and environmental pressure and thus sustaining agricultural production. Jiangxi was selected as the surveyed area due to its typical SFM problems and land characteristics. More explanations on the research areas can be found in the next section.

One limitation to this study to be mentioned is about the household and village sample size. The data were from 315 plots of 3 counties in Jiangxi Province. Due to the severe land fragmentation in the survey area, only 58 households from 3 villages were investigated for this study. However, given other factors which may also influence farmers’ SFM, such as the household char-acteristics, input and output prices, and some physical conditions like temperature and rainfall, the sample size would be big enough to allow for examining how plot-level characteristics affect farmers’ SFM. Mean-while, the discrepancy between the plots was generally greater than that between the villages. Thus, investigat-ing more plots with enough typical villages may obtain more detailed and convincing information at a rational cost. Moreover, in order to see if this sample size would

make sense or not, the results were compared with the related studies in the later part of this study.

DATA AND METHODS

Description of survey sites and sampling

Data were collected in three villages of east Jiangxi Province, China in 2011. Jiangxi is a typical rice cropping area. Rice production took up a big share in cropping systems. Normally, nearly 90% of farmers’ plots are planted with rice. For example, in the surveyed villages, the double rice and single rice were popularly planted. They accounted for 48 and 41% of the total plots, indicating that rice production dominated in this area (Table 1). According to local farmers, researchers and officials, the selected villages represented a soil-deg-radation prone area mainly caused by inappropriate SFM. Green manuring is an important traditional SFM practice in this area. It is an effective way to build up soil organic matter and to improve soil quality, and thus to reduce chemical fertilizer application (Rui and Zhang 2010). The villages were selected based on their socio-economic development levels and SFM practices. In this study, the villages were initiated as A, B and C. although they are from 3 counties, the selected villages are located close to each other, allowing for the control of some natural and socio-economic conditions such as rainfall, temperature and input/output prices, which may also influence SFM practices. From village A to village B, it is about 70 km; from village A to village C, it is 90 km; and from village B to village C, it is 80 km. It takes about 2 to 3 h driving from one village to another, the difference in time taken being mainly due to qualities of the road. Some background information of the 3 villages can be found in Table 2.

Village a is located in a hilly area with around 1 400 persons from 280 households. Field infrastructures such as irrigation condition were good and also local market outlets were available in this village. Village B is located in a remote mountain area with 2 400 persons from 580 households. It takes about 2 h driving from this village to the county capital. Its socio-economic development level was the lowest in Jiangxi Province. Village C, which is located in a plain area with a medium distance to the county capital, was the largest village among the

Page 4: Do Land Characteristics Affect Farmers' Soil Fertility Management?

Do Land Characteristics Affect Farmers’ Soil Fertility Management? 2549

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

three. It had 3 750 persons from 920 households, and its socio-economic development level was moderate. In 2011, income per capita of the three villages was 8 000, 2 800 and 6 500 CNY for villages A, B and C, respectively, and farmland area per capita was 0.13, 0.05 and 0.07 ha in A, B and C again, respectively. The cropping patterns were more diversified in C compared with that in A and B. For example, the red sprout colocasia was a traditional and popularly planted crop in C but not in A and B.

Because of change in population, the contracted land in the three villages had been adjusted at varied frequency and magnitude. It was reallocated twice since the first land allocation, namely once in about 10 years in village A. During land reallocation, all plots allocated to the individual households were taken back by the village and re-allocated to each household afresh. Small land adjustments took place with the frequency and magnitude differing across the hamlets of village B. Land was also reallocated during the past decades in village C. This conformed to the findings from Feng et al. (2011) that more than 94% villages in Jiangxi Provinces had real-located their land.

Land tenure, fragmentation and soil quality of plot,

the major characteristics to be examined in the study area could to some extent be reflected in a typical household of village C. This household cultivated about 1.07 ha land, among which 0.47 ha were contracted land, and the remaining 0.60 ha were rented in from other farmers. All the 1.07 ha land was spread over 7 plots. The contracted plots were distinguished into 4 grades in terms of soil quality and irrigation condition: basic field, with medium soil quality and irrigation condition; bad field, easy to suffer from flooding1; good field, with fertile soil and good irrigation access; and dry-land, rain-fed normally. Land had been adjusted in terms of population change, as seen in many villages of Jiangxi Province. However, the frequencies were different for fields with different qualities which took into account soil quality and irriga-tion condition. The basic fields were normally adjusted each 3 years. The bad fields, good fields and dry-land were adjusted each 9 years. All plots together with the basic fields had to be taken back and reallocated to each individual household then.

The survey for this study was based on a previous one, which was conducted in 2001 firstly. In that survey, about 23%, namely 54, 109 and 168 households were

Table 2 Background information of the sample villages1) Village a Village B Village C

Location Distance to city Close Remote Medium Road quality Poor Bad Good Topography Hilly area Mountain area Plain area

Population Persons 1 400 2 400 3 750Households 280 580 920

Income Per capita income (CNY) 8 000 2 800 6 500Land Per capita farmland (ha) 0.13 0.05 0.07

number of plots per household 5.4 6.8 4.21) Source: Based on the fieldwork.

1 During a revisit in 2013, the farmer reported that such fields suffered from a severe flooding at the end of June 2013.

Table 1 Distribution of cropping pattern choicesCropping systems number of plots1) Share (%) Crop type number of plots Share (%)Mono-cropping 108 35.18 Vegetable2) 11 10.19

Cash crop 22 20.37Single rice 75 69.44

Double-cropping 115 37.46 Single rice+Cash crop 28 24.35Double rice 77 66.95Single rice+Green manure 10 8.70

Triple cropping 84 27.36 Double rice+Cash crop 4 4.80Single rice+Cash crop+Green manure 12 14.20Double rice+Green manure 68 81.00

1) The total number of plots is 307. 2) Vegetables can be produced for one season, two seasons and even more than 3 seasons, however, in this survey, just include this as one season crop.

Page 5: Do Land Characteristics Affect Farmers' Soil Fertility Management?

2550 Tan Shu-hao

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

randomly selected from the villages a, B and C, respec-tively, resulting to a total 331 households2. Afterwards, several other surveys with different purposes had been undertaken within the same sample. Questionnaires for the current study were conducted in 2011, and some fol-low up visit took place in 2013. Due to time and budget constrains3, only 15, 21 and 22 households in villages a, B and C were randomly selected from the 331 sample households, resulting in a sample of 58 households and 315 plots surveyed. Of the 315 plots, 102 plots were from village A, 88 from village B and the 125 from village C. All the sample households had an average of 5.4 plots each, and specifically an average of 6.8 plots in village A, 4.2 plots in village B and 5.7 plots in village C. Land fragmentation degree represented the current situation of paddy rice areas in South China, but it was a little higher than that (4.4 plots) of whole China in 2010 (Feng et al. 2011).

Structure of database and descriptive statistics of variables

The dataset included household characteristics, plot-lev-el factors, village characteristics and input and output

prices. Farm household characteristics included age and education of the household head, farm size, family size and the family’s income and routine consumption. In this survey, over 20% of the households could not report their exact income but all the households knew their consumption. Consumption was therefore used as a proxy for wealth, as was also done in Tu et al. (2011). Plot level factors included plot-specific characteristics, namely distance of homestead to the plot, plot size, plot tenancy (i.e., whether a plot was reserved, contracted or rented-in from other farmers or from the collective), and the quality of a plot. Given soil parent materials, the quality of a plot was to a great extent resulted from long-term management (Tittonell et al. 2010), and it was seen here as one plot-level socio-economic character-istic; crop production, which included output of a plot within a production cycle, the main inputs such as labor, herbicide, pesticide, chemical fertilizers and seed, and the input and output prices; and SFM practices, which included cropping systems and fertilization (i.e., manure and chemical fertilizer application). In addition, two dummies were used as village characteristics to cover the differences excluded in the models to be estimated. Table 3 presents the descriptive statistics and the related information of the database.

Table 3 Structure of the database and descriptive statistics of variables used1)

Variable Code Unit Mean S.D. Min Maxage age Year 47.180 11.500 28 75Education Edu Year 4.800 2.720 0 13 Household size Hhsize Person 5.170 1.790 1 10Consumption Cons Thousand CnY 9.845 9.033 0.480 43.3Farm size Fsize ha 0.891 0.528 1.800 35.6Plot size Psize ha 0.113 0.101 0.020 16.1Plot tenure (dummy)2) Dt - 0.210 0.410 0 1Distance3) Dist Minute 12.550 10.360 1 60Land quality (dummy)4) Quality - 0.350 0.480 0 1Green manuring (dummy)5) Dgren - 0.310 0.460 0 1Manure application (dummy) Dman - 0.440 0.500 0 1Pure nitrogen n kg ha-1 210 203 0 334Pure phosphorus P2O5 kg ha-1 109 94 0.900 145Pure potassium K2O kg ha-1 151 194 0.900 290Manure Manure kg ha-1 4 920 11 235 0 20 0001) Source: Author’s survey.2) 1 for rented-in plots and 0 for own-contracted plots. 3) Walking distance from plot to homestead. 4) 1 for plots with good quality, 0 for plots with bad or medium quality. Land quality was decided by villages when plots were allocated to farmers. They judged land

quality in terms of soil texture, irrigation condition, soil workability and crop yields, etc. 5) 1 for plots planting green manure crops (they were mainly Chinese milk vetch), 0 for otherwise.

2 There were about 235, 474 and 730 households in the three villages in 2001, respectively. 3 In order to obtain a relatively reliable dataset, an in-depth survey was conducted with all the selected 58 households and their 315 plots visited during the fieldwork, which

was a time- and money-consuming activity.

Page 6: Do Land Characteristics Affect Farmers' Soil Fertility Management?

Do Land Characteristics affect Farmers’ Soil Fertility Management? 2551

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

Reserved land was either allocated to individual households more than a half century ago, or developed by households from waste land. The reserved land could be cultivated for a long time but only represented a very small share of the total land. In principle, around 5% of the total farmland area could be reserved. In many villag-es, however, this share was even lower. Among the 315 plots surveyed, only 8 plots were reserved, accounting for only 2.54% of total plots and were therefore excluded from the model analysis. Gender of the household head was not considered as a household characteristic, because in the surveyed samples, most production decisions were made by males in the family, although some of the males were not always present in the villages. In addition, theoretically, input prices should serve as variables in the econometric models, however, because the selected vil-lages are located close to each other, the fertilizer prices reported by farmers were almost the same. Observations from the author’s fieldworks also confirmed this. The very small variations in fertilizer prices did not allow for putting them in the models. Therefore, the input prices were also excluded from the models.

On average, a household with 5 family members cultivated about 0.9 ha of land, with average plot size of 0.113 ha. About 21% of the sampled plots were rented in, 3 percentage points higher than that found by Gao et al. (2012). Normally, farmers rented the plots once a year without signing any formal contract with the renters, even though they intended to cultivate the plots for some years. It took 12 min walking from homestead to the plots. In addition, the sample plots had quality below the medium level. This was consistent with the national level investigation, which showed that above two-thirds of cultivated lands in China are of low- and medium-yield potential. Green manure crops, which mainly included the Chinese milk vetch and the legume crops like peanut, bean and pea, were planted on 31% of the plots sampled.

Nitrogen, phosphorus and potassium fertilizers were the chemical fertilizers applied. Farmers could use each fertilizer individually, or applied the compound fertilizers with a certain content of N, P2O5 and K2O. For convenience, we calculated the content of N, P2O5

and K2O applied in each surveyed production season. The amount of chemical fertilizer applied on a plot de-pended on the number of production seasons, intensity of fertilizer use, and efficiency of fertilizer application.

On average, N, P2O5 and K2O were applied at levels of 210, 109 and 151 kg ha-1, respectively in one production cycle on the sampled plots, with the ratio of N, P2O5 and K2O being 1:0.5:0.7. According to a field experiment on 262 pilots in rice cropping areas of South China, at the maximum economic yield, the ratio of N, P2O5 and K2O was 1:0.28:0.32 (MOA 2011), whereas at the maximum physical yield, the ratio was 1:0.55:0.7. At this point, the average maximum yield was 5 860 kg ha-1. The actual ratio of our sampled plots was very close to the one that yielded maximum physical level. However, the yield was only 4 292 kg ha-1, about 30% lower than that at recommended ratio. This means that the yield could be greatly improved and the ratio of chemical fertilizer application could be more economically efficient by adopting some effective SFM practices.

applying manure and green manuring were two im-portant and traditional ways to maintain or to improve soil fertility in the surveyed area. Manure was mainly a mixture of animal dung, urine and crop residue. In the surveyed areas, the main sources of manure were cattle, pig and chicken dung. Manure was normally applied during land preparation as basal fertilizer. Of the sam-pled plots, manure was applied on 44% of the plots at the rate of 5 000 kg ha-1 per production cycle with a range of 0-20 000 kg ha-1. Information about green manuring is presented in Tables 1 and 4.

Methods of data analyses

Statistical method was used to analyze the distribution of cropping pattern choices, and econometric models were applied to estimate the equations of the other SFM deci-sions4. As discussed before, SFM decisions depended on socio-economic factors (Omamo et al. 2002; Zingore et al. 2007). In this specification, they included village, household and plot-specific characteristics. The linear relationships between SFM practices and the explanatory

4 The estimation household model was built upon Sadoulet and De Janvry (1995).

Page 7: Do Land Characteristics Affect Farmers' Soil Fertility Management?

2552 Tan Shu-hao

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

variables were assumed5.SMi

*=P1iα+H1i β+V1iφ+υ1 A second equation, of the binary choice type, is spec-

ified whether the farm household planted green manure, or applied manure or not. That is:

Zi*=P2iα+H2i β+V2iφ+υ2

Where we have the observation rule:SMi=SMi

*, Zi=1 if Zi*>0

SMi=0, Zi=0 if Zi* 0

Where SMi denotes a farmer’s actual SFM practices on plot i, namely how many nitrogen (n), phosphorus (P2O5), potassium (K2O) and manure (manure) were applied on the plot, while the binary variable Zi simply indicates if the farmer planted green manure crop or applied manure or not, which is a standard Probit model. Pi is a vector of characteristics of plot i, Hi is a vector of characteristics of household to which plot i belonged; Vi is a vector of village level difference to which the plot i belonged, α, β and φ are vectors of unknown co-efficients to be estimated; υ is a vector of error terms of the equations.

The expected directions of each factor on different SFM practices mentioned above could be mixed and explained in the results. Normally, seemingly unrelated regressions (SUR) are regarded to be a consistent and efficient method for estimation in such case, however, each equation here contains exactly the same set of regressors on the right-hand-side, and it is in fact equiv-alent to ordinary least square (OLS), and therefore OLS was selected to estimate the equations with continuous dependent variables, and the Probit models were used to estimate the two binary equations.

RESULTS AND DISCUSSION

Distribution of cropping pattern choices

Tables 1 and 4 show cropping pattern choices on the plots sampled. Table 1 indicates that plots cultivated for mono-cropping, double cropping and triple cropping systems were almost evenly distributed, accounting for

about one-third of the total sample plots for each one, but with the double-cropping system a little higher and the triple cropping system a bit lower. Within each cropping system, the detailed crop types were distin-guished. Within mono-cropping system, for instance, 10% of the plots were used for vegetable production, 20% for cash crops such as water melon, rapeseed, sugar cane and peanuts, and a lion share of 70% for rice production.

as an important way of managing soil in the surveyed area, green manure cultivation was a major cropping pattern. Out of the 307 plots, 90 plots were planted with green manure, accounting for 29% of the total plots. Of which, 76% with double rice, 14% with single rice and cash crop, and the remaining one-tenth with single rice, either early rice, late rice or middle-season rice (Table 1). Compared to 31% indicated in Table 3, the green manure in Tables 1 and 4 means only the Chinese milk vetch, excluding the legume crops which were counted as cash crops.

Table 4 indicates the relationship between land ten-ancy and the distribution of cropping pattern choices in the study area. As mentioned before, the land tenure arrangements could be distinguished into rented in, contracted and reserved ones. It is interesting to note that the triple cropping system on contracted plots and on rented plots was significant at 5% level (30% vs. 15%). The green manure and legume crops were also much more planted on contracted plots compared with the rented ones (33% vs. 15% for green manure). The ANOVA test also showed the significance at 5% level. Although the ANOVA test did not show significance between cash crops planted on contracted plots and rented plots, simply seen from the data, there existed

Table 4 Plot tenure and cropping pattern choicesContracted plot

(%)Rented plot

(%)analysis of varianceF-value P-value

Mono cropping system 33.87 40.68 0.750 0.435Double cropping system 35.89 44.07 1.113 0.351Triple cropping system 30.24 15.25 7.401 0.053**

Green manure 32.66 15.25 17.011 0.015**

Cash crop 27.02 20.34 0.329 0.597**, significant at 5% level. The same as below.

5 Theoretically, the age and education of household head might have non-linear relationship with SFM, i.e., farmers at certain age or with certain education level might differ from their counterparts at younger/elder or with higher/lower education level in SFM. The linear relationship was assumed because that this study intended to focus on land characteristics, among which land quality and tenure were dummy variables, while farm size, plot size and distance were anticipated to have linear relationship with SFM.

Page 8: Do Land Characteristics Affect Farmers' Soil Fertility Management?

Do Land Characteristics affect Farmers’ Soil Fertility Management? 2553

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

some difference (27% vs. 20%). It is similar for vege-table planted (5.24% vs. 0%), which was not shown in Table 4. In addition, although the 8 reserved plots were excluded in the analysis, it was surprising to note that among the 8 reserve plots, only one was used to plant rice after harvesting peanut, the other 7 plots were used for vegetable production for the whole year, which were managed better by applying more fertilizers, manure and irrigation etc. Cropping pattern choices were not consid-ered in the available related studies on China (Li et al. 1998; Yu et al. 2003; Gao et al. 2012), therefore, it is pity that we cannot compare the results of the current study with these ones. The cropping pattern choices based on this study suggested that securer tenure enabled more diversified and more soil-friendly cropping patterns, and was thus more beneficial to agricultural system. This to some extent conformed to the findings from de Jager et al. (1998) and Salasya (2005) on Kenya that cash crops and vegetables were managed more sustainably than food crops.

Determinants of chemical fertilizer application

Eviews 6.0 was used to estimate all the equations. The F-statistics of the equations were high enough to reject the null hypotheses that the listed variables could not

explain the differences in chemical fertilizer and ma-nure applications between plots. Table 5 shows the full effects of socio-economic factors on N, P2O5 and K2O application6. Considering the interests of this paper, only effects of the concerned plot-level land characteristics on chemical fertilizer application, which normally might cover urea, ammonium bicarbonate, compound fertilizer with different nitrogen, phosphorus and potassium com-bination, calcium magnesium phosphate and potassium chloride in the surveyed areas, were discussed. As mentioned before, in this study, these fertilizers were discounted as N, P2O5 and K2O.

In the study area, plot size was found to have signifi-cantly negative effects on fertilizer use. Keeping other factors constant, larger plots were cultivated less inten-sively in terms of chemical fertilizer application. Given yield level, this might also indicate that crop productions on such plots were more efficient in using the chemical fertilizers. The results from Nguyen et al. (1996) con-firmed that controlling for total farm size, there was a statistically significantly positive relationship between plot size and productivity of rice, maize and wheat. This further suggested that larger plots were helpful to efficient application of chemical fertilizers, because larger plots enabled size effect which might be caused by farmers’ behavior on fertilization. In other words,

Table 5 Regression results for chemical fertilizer applicationn (OLS) P2O5 (OLS) K2O (OLS)

Coeff t-value Coeff t-value Coeff t-value Constant 59.950 4.006*** 38.02 4.881*** 67.770 4.830***

age -0.275 -1.365 -0.145 -1.378 -0.237 -1.254 Education 0.384 0.485 -0.325 -0.788 -1.627 -2.192** Household size 3.567 3.316*** 0.911 1.628* -2.573 -2.551***

Consumption -1.719 -1.725* -0.254 -0.489 -0.431 -0.461 Farm size -0.873 -3.006*** -0.336 -2.221** 0.239 0.878 Plot size -1.508 -1.290 -1.817 -2.988*** -1.970 -1.798* Distance -0.175 -0.943 -0.187 -1.937** -0.297 -1.705* Tenure (dummy) -4.859 -1.044 -0.746 -0.308 -3.835 -0.879 Plot quality (dummy) 6.830 1.856* -0.101 -0.053 3.547 1.028 B -39.900 -8.932*** -18.780 -8.078*** -20.870 -4.984***

C -19.820 -4.338*** -6.117 -2.573*** -7.050 -1.646* R-squared 0.27 0.24 0.14 Adjusted R-squared 0.25 0.21 0.11 No. of observations 307*, significant at 10% leve; ***, significant at 1% level. The same as below.

6 Consumption (the proxy variable for income) was only found to have a weak negatively significant effect on nitrogen use, but no significant effects on phosphorus, potassium and manure application. This suggested that in general income might not be a constraint for farmers to buy fertilizers. However, if farmers spent too much money to consume, they might have no enough money to buy fertilizers sometimes.

Page 9: Do Land Characteristics Affect Farmers' Soil Fertility Management?

2554 Tan Shu-hao

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

farmers might be more casual in applying fertilizers when the plots or farm sizes were small (they might not mind wasting some fertilizer). Larger plots or farms needed more fertilizers and made farmers be more precise when applying so as to avoid big waste.

Distance of plot to the homestead showed signifi-cantly negative effects only on P2O5 and K2O. Two possibilities might explain this: either farmers paid less attention to P2O5 and K2O application in farther plots, or farther plots were cultivated more efficiently. according to Tan et al. (2010), keeping other factors constant, plots with farther distance were less efficient, implying that the second possibility did not hold. It confirmed that plots with larger distance normally re-ceived less attention from farmers, especially for P2O5 and K2O, which were more useful for long-term soil improvement compared with N.

Out of expectation, plot tenure arrangements did not show any significant effects on intensity of fertilizer application. This meant that given other factors, there was no difference between farmers’ chemical fertilizer applications on rented-in plots and on own contracted plots. However, plot quality was found to have signifi-cantly positive effects on N application. This indicated that keeping other variables constant, the plots with higher quality were applied more N compared with the plots with poor or medium quality, suggesting that farmers paid more attention to the plots with better quality. Observations from field surveys confirmed that plots with poor quality were more vulnerable to being abandoned.

Determinants of green manuring and manure application

The F-statistics and LR statistics were also high enough to reject the null hypotheses. The effects of land char-acteristics on green manuring and manure application in the research area are shown in Table 6. Similar to the earlier sections, the study focused on discussing the plot-level land characteristics. As expected, all the plot-level land characteristics had statistically significant effects on manure application. Given other factors, plot size negatively affected manure application. There was a tendency to either apply no manure or to apply less manure on larger plots. This was consistent with Zhao (2006) and Gao et al. (2012) that larger plots had lower intensity in manure application. The distance of plot to the homestead negatively influenced manure application. This also confirmed the findings of Gao et al. (2012). The reasons might be because that manure application was a labor-intensive activity and manure was limited, so farmers tended to apply on plots close to the homestead. Plot tenancy was found to have a significantly negative effect on green manure, i.e., keeping other factors con-stant, farmers tended to plant less green manure crop on rented-in plots compared with own-contracted plots. This was because green manuring is a long-term invest-ment of improving the soil, and own-contracted plots have securer tenure to guarantee the returns to such soil investment. This was consistent with the findings from Katz (2000) and Deininger and Jin (2003) that secure

Table 6 Green manure and manure applicationGreen manuring in dummy (Probit) Manure application in dummy (Probit) Manure application in quantity (OLS)

Coeff t-value Coeff t-value Coeff t-valueConstant -1.021 -1.312 -1.369 -1.934** 755 0.623 age 0.009 0.836 0.011 1.187 -3.908 -0.239 Education -0.083 -2.078** 0.042 1.125 -13.65 -0.213 Household size -0.101 -1.895** -0.038 -0.713 60.98 0.700 Consumption 0.048 0.940 0.004 0.081 76 0.939 Farm size -0.020 -1.179 -0.006 -0.443 7.519 0.320 Plot size 0.072 1.266 -0.124 -1.754* -268 -2.829*** Distance 0.015 1.523 -0.018 -2.028** -35.46 -2.357*** Tenure (dummy) 0.520 2.044** 0.296 1.322 19.55 0.052 Plot quality (dummy) 0.987 5.333*** 0.668 3.880*** 693 2.325** B -0.195 -0.880 0.849 4.004*** 1 282 3.543*** C -0.059 -0.266 1.063 4.745*** 259 0.699 R-squared 0.14 Adjusted R-squared 0.10 McFadden R-squared 0.14 0.17No. of observations 307

Page 10: Do Land Characteristics Affect Farmers' Soil Fertility Management?

Do Land Characteristics affect Farmers’ Soil Fertility Management? 2555

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

tenure encouraged farmers to better manage their soil.Plot quality was found to have significantly positive

effects on green manuring and manure application. Farmers tended to plant green manure crop and to apply manure on plots with higher quality. Moreover, manure was applied more intensively in plots with high soil qual-ity. This is consistent with Gao et al. (2012) that high soil quality plots tended to attract more organic manure investment. This implies that farmers paid more atten-tion to the plots with higher quality. Normally, such plots were easier to manage, more efficient to work on, and they had higher returns to investment in soil compared with their counterparts with low or medium quality. This also implies that farmers might neglect the plots with poor quality in practice. Unfortunately, two-thirds of cultivated land in China is of low or medium quality. The current land tenure arrangements with fragmented holdings may discourage farmers who are the main and direct land users, to manage the soil fertility inappropri-ately. This deserves a high attention in future.

CONCLUSION

In order to promote adoption of appropriate SFM strat-egies so as to maintain or improve the soil quality of cultivated land, this study applied statistical method and econometric models to examine how plot-level land characteristics affected farmers’ SFM practices. although data used for this study were based on a data-set with 315 plots from 58 households of 3 villages in 3 counties of a rice growing area in South China, main findings were interesting. In general, the results gave a positive answer to the question raised by the title of this paper. Main findings from the normally studied 3 variables, namely land tenure, plot size and the distance of the plot to homestead, were basically consistent with that from the existing studies, i.e., securer land tenure encouraged while plot distance discouraged land-saving investments, and plot size reduced the fertilization in-tensities, including both chemical fertilizers as well as organic manure.

Differing from most existing studies, the current study went one step further to examine the impacts of land characteristics on cropping pattern choices, an important aspect of SFM practice, and took into account the soil

quality of plot. Main findings showed that securer land tenure arrangement facilitated effective practices of SFM through more diversified and more soil-friendly cropping pattern choices. Notably, plots with better soil quality were found to put more effort on management by applying more nitrogen and manure, and more tended to plant green manure crop.

The results suggested that the current tenure arrange-ments with fragmented land holdings may discourage farmers to manage soil fertility in an inappropriate way. Polices for more secure tenure and less fragmented land management pattern would be desirable. Moreover, upgrading the medium- and low-yield fields is neces-sary for promoting long-term land investment in China. Since the present study was conducted at a plot scale, the household and village sampling was limited by a relatively small size due to the severe land fragmentation at the household-level. The current study might not well represent the whole rice cropping areas in South China. Thus, the implications derived from the study may need further confirmations by studies with larger samples from larger scale.

AcknowledgementsThis work was partly funded by the National Natural Science Foundation of China (71273268). The author thanks the surveyed households for their great supports, and the anonymous reviewers for their critical but constructive comments.

ReferencesAli M 1996. Quantifying the socio-economic determinants

of sustainable crop production: an application to wheat cultivation in the Tarai of Nepal. Agricultural Economics, 14, 45-60.

Chen H Q, Hou R X, Gong Y S, Li H W, Fan M S, Kuzyakov Y. 2009. Effects of 11 years of conservation tillage on soil organic matter fractions in wheat monoculture in Loess Plateau of China. Soil and Tillage Research, 106, 85-94.

Deininger K, Jin S. 2003. The impact of property rights on households’ investment, risk coping, and policy preferences: Evidence from China. Economic Development and Cultural Change, 51, 851-882.

Feng L, Ye J P, Jiang Y, Zhu K L. 2011. The sequential variation and regional difference of land reallocation in rural China: An empirical study based on a survey in 17 provinces from 1999-2010. China Land Science, 25, 14-22. (in Chinese)

Gao L L, Huang J K, Rozelle S. 2012. Rental market for

Page 11: Do Land Characteristics Affect Farmers' Soil Fertility Management?

2556 Tan Shu-hao

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

cultivated land and agricultural investments in China. Agricultural Economics, 43, 391-403.

Guo J H, Liu X J, Zhang Y, Shen J L, Han W X, Zhang W F, Christie P, Goulding K W T, Vitousek P M, Zhang F S. 2010. Significant acidification in major Chinese croplands. Science, 327, 1008-1010.

Greenland D J, Nabhan H. 2001. Soil Fertility Management in Support of Food Security in Sub-Saharan Africa. Food and Agriculture Organization of the United Nations, Rome.

Heerink N. 2005. Soil fertility decline and economic policy reform in Sub-Saharan Africa. Land Use Policy, 22, 67-74.

Huang G B, Zhang R Z, Li G D, Li L L, Chan K Y, Heenan D P, Chen W, Unkovich M J, Robertson M J, Cullis B R. 2008. Productivity and sustainability of a spring wheat-field pea rotation in a semi-arid environment under conventional and conservation tillage systems. Field Crops Research, 107, 43-55.

IPCC. 2007. Climate Change: The Physical Science Basis: Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge.

De Jager, Kariuki A, Mariri I, Odendo M, Wanyama J M. 1998. Linking nutrient balances and economic performance in three districts in Kenya. Agriculture, Ecosystems and the Environment, 71, 81-92.

Katz E G. 2000. Social capital and natural capital: A comparative analysis of land tenure and natural resource management in Guatemala. Land Economics, 76, 114-132.

Koning N, Heerink N, Kauffman S. 2001. Food insecurity, soil degradation and agricultural markets in West Africa: Why current policy approaches fail. Oxford Development Studies, 29, 189-207.

Li G, Rozelle S, Brandt L. 1998. Tenure, land rights, and farmer investment incentives in China. Agricultural Economics, 19, 63-71.

Li L L, Huang G B, Zhang R Z, Bellotti B, Li G D, Kwong Y C. 2011. Benefits of conservation agriculture on soil and water conservation and its progress in China. Agricultural Sciences in China, 10, 850-859.

Lindert P. 1999. The bad earth? China’s soil and agricultural soils and development since the 1930s. Economic Development and Cultural Change, 47, 701-736.

MOA (Ministry of Agriculture of Peoples Republic of China). 2011. The Suggestions on the Guideline for Scientific Fertilization of Major Crops in 2010. MOA Office, No. [2010], China. p. 36. (in Chinese)

Moreno F, Murillo J M, Pelegrin F, Giron I F. 2006. Long-term impact of conservation tillage on stratification ratio of soil organic carbon and loss of total and active CaCO3. Soil and Tillage Research, 85, 86-93.

Nguyen T, Cheng E, Findlay C. 1996. Land fragmentation and farm productivity in China in the 1990s. China Economic Review, 7, 169-180.

Omamo S W, Williams J C, Obare G A, Ndiwa N N. 2002. Soil fertility management on small farms in Africa: Evidence

from Nakuru District, Kenya. Food Policy, 27, 159-170.Qin G Y, Tan S H. 2013. Analysis of the effects of farmers’ risk

perception on their willingness of land transfer. Journal of Northwest A&F University (Social Science Edition), 13, 61-67. (in Chinese)

Rasul G, Thapa G B. 2007. The impact of policy and institutional environment on costs and benefits of sustainable agricultural land uses: The case of the Chittagong Hill Tracts, Bangladesh. Environmental Management, 40, 272-283.

Rui W Y, Zhang W J. 2010. Effect size and duration of recommended management practices on carbon sequestration in paddy field in Yangtze Delta Plain of China: A meta-analysis. Agriculture Ecosystems & Environment, 135, 199-205.

Sadoulet E, de Janvry A. 1995. Quantitative Development Policy Analysis. The Johns Hopkins University Press, Baltimore, MD.

Sah K D, Sarkari D, Seal A, Singh S. 2010. Micro-level planning for optimum land use in a coastal area of West Bengal - A case study. Agropedology, 20, 19-29.

Salasya B D S. 2005. Crop production and soil nutrient management: an economic analysis of households in western and central Kenya. PhD thesis, Wageningen University, The Netherlands.

Sanchez P A. 2002. Soil fertility and hunger. Science, 295, 2019-2012.

Scherr S, Yadav S. 2001. Land degradation in the developing world: issues and policy options. In: Pinstrup-Anderson P, Pandya-Lorch R, eds., The Unfinished Agenda: Perspectives on Overcoming Hunger, Poverty, and Environmental Degradation. International Food Policy Research Institute, Washington, D.C. pp. 133-138.

Sombrero A, de Benito A. 2010. Carbon accumulation in soil. Ten-year study of conservation tillage and crop rotation in a semi-area of Castile-Leon, Spain. Soil and Tillage Research, 107, 64-70.

Tan S H, Heerink N, Kuyvenhoven A, Qu F T. 2010. Impact of land fragmentation on rice producers’ technical efficiency in South-East China. NJAS-Wageningen Journal of Life Sciences, 57, 117-123.

Tan S H, Heerink N, Qu F T. 2006. Land fragmentation and its driving forces in China. Land Use Policy, 23, 272-285.

Tittonell P, Muriuki A, Shepherd K D, Mugendi D, Kaizzi K C, Okeyo J, Verchot L, Coe R, Vanlauwe B. 2010. The diversity of rural livelihoods and their influence on soil fertility in agricultural systems of East Africa - a typology of smallholder farms. Agricultural Systems, 103, 83-97.

Tittonell P, Vanlauwe B, de Ridder N, Giller K E. 2007. Heterogeneity of crop productivity and resource use efficiency within smallholder Kenyan farms: Soil fertility gradients or management intensity gradients? Agricultural Systems, 94, 376-390.

Tu Q, Bulte E, Tan S H. 2011. Religiosity and economic performance: Micro-econometric evidence from Tibetan

Page 12: Do Land Characteristics Affect Farmers' Soil Fertility Management?

Do Land Characteristics Affect Farmers’ Soil Fertility Management? 2557

© 2014, CAAS. All rights reserved. Published by Elsevier Ltd.

area. China Economic Review, 22, 55-63.Yang H S. 2006. Resource management, soil fertility and

sustainable crop production: Experiences of China. Agriculture, Ecosystems and Environment, 116, 27-33.

Ye L, Ranst V E. 2009. Production scenarios and the effect of soil degradation on long-term food security in China. Global Environmental Change, 19, 464-481.

Yu H, Huang J K, Rozelle S, Brandt L, Zhang L X. 2003. Economic Research, 9, 89-91. (in Chinese)

Zhao Y. 2006. Land with Public Ownership and Private User Rights: Economic Analysis on Rural Land Institutions in

China. San Lian Book Store Publishing House, Shanghai. (in Chinese)

Zhen L, Michael A Z, Chen G B, Feng Z M. 2006. Sustainability of farmers’ soil fertility management practices: a case study in the North China Plain. Journal of Environmental Management, 79, 409-419.

Zingore S, Murwira H K, Delve R J, Giller K E. 2007. Influence of nutrient management strategies on variability of soil fertility, crop yields and nutrient balances on smallholder farms in Zimbabwe. Agriculture, Ecosystems and Environment, 119, 112-126.

(Managing editor WENG Ling-yun)