Contents lists available at ScienceDirect
Agriculture, Ecosystems and Environment
journal homepage: www.elsevier.com/locate/agee
Farmers’ perceptions of crop pest severity in Nigeria are associated with
landscape, agronomic and socio-economic factors
Wei Zhanga,⁎, Edward Katoa, Felix Bianchib, Prapti Bhandarya, Gerrit Gortc, Wopke van der Werfd
a Environment and Production Technology Division, International Food Policy Research Institute (IFPRI), 1201 Eye Street NW, Washington, DC, 20005, United States
b Farming Systems Ecology, Wageningen University, P.O. Box 430, 6700, AK, Wageningen, The Netherlands
c Biometris, Wageningen University, 6708 PB, Wageningen, The Netherlands
d Centre for Crop Systems Analysis, Wageningen University, P.O. Box 430, 6700, AK, Wageningen, The Netherlands
A R T I C L E I N F O
Keywords:
Pest regulation
Household survey
Crop management
Land use
Production situations
Africa
A B S T R A C T
Insect pests are a major cause of crop yield losses around the world and pest management plays a critical role in
providing food security and farming income. This study links Nigerian farmers’ perceptions of pest severity to the
landscape, agronomic, biophysical, and socio-economic context in which agricultural production takes place. A
farm household survey was conducted during 2012–2013, collecting data on household characteristics, cropping
systems, pest severity and pest management from 805 households in 12 states of Nigeria. Village characteristics
and land use information were collected from an accompanying semi-structured village survey. Reported pest
severity was negatively associated with the proportions of forest and unused land at the landscape scale. This
finding suggests the existence of pest suppressive effects of a diverse landscape under African smallholder
agriculture settings, confirming findings of more industrial and larger scale agroecosystems in the temperate
zone. Application of fertilizers (chemical and manure) was negatively related to reported pest severity.
Moreover, reported pest severity was lower in mixed-cropping systems than in mono-cropping systems, re-
inforcing the idea of a pest suppression benefit of diverse cropping systems. In conclusion, our findings suggest
that the presence of non-crop areas in the landscape and the diversification of agroecosystems may be a viable
strategy for smallholder farmers to manage pests with limited reliance of chemical insecticides in Nigeria, but
that actual pest management decisions are influenced by a wide range of context-specific factors. The paper adds
new evidence on the relationship between different production situation characteristics and pest severity for
Nigeria, based on which policy implications are discussed.
1. Introduction
Insect pests are a major cause of crop yield losses around the world
(Oerke, 2006) and an important cause of food insecurity in developing
countries (Zakari et al., 2014). Farmers make crop and pest manage-
ment decisions within the realm of their production situations, i.e., the
physical, biological, technical, social, and economic context in which
production takes place (Penning de Vries, 1982; Savary et al., 2006a,b),
and their decisions in turn shape their production situations. While the
interdependence between the susceptibility to pest infestation and the
production situation has been demonstrated before (Allinne et al.,
2016; Avelino et al., 2006; Savary et al., 2006a,b; Savary et al., 2017),
little is known about the relationship between production situations
and farmer reported pest severity on common crops in Nigeria. Ana-
lyzing the perceptions of farmers on pest severity within the context of
their production situation can provide important new insights in the
ways to encourage ecologically-based pest management attitudes and
practices.
The management of pests has important implications for African
agriculture where the majority of the farmer community consists of
smallholder farmers with low agricultural productivity (Bature et al.,
2013). In Nigeria, insect pests and plant diseases are major yield re-
ducing factors, threatening food security and farmers’ incomes. For
example, insect pests and diseases in yams resulted in a 25% mean
annual yield loss (Tobih et al., 2011; Amusa et al., 2003) and 25–30% of
yield loss of cocoa was attributed to the brown cocoa mirid alone
(Ndubuaku and Asogwa, 2006). While Nigerian farmers are aware of
the availability of several methods of pest control, including chemical,
biological and traditional cultural control methods, farmers commonly
do not actively control pests in their field crops (Alghali, 1991;
Bottenberg, 1995; Banjo et al., 2003; Ofor et al., 2009). Farmers who
actively manage pests rely primarily on chemical insecticides, but can
https://doi.org/10.1016/j.agee.2018.03.004
Received 13 November 2016; Received in revised form 6 March 2018; Accepted 8 March 2018
⁎ Corresponding author.
E-mail address: w.zhang@cgiar.org (W. Zhang).
Agriculture, Ecosystems and Environment 259 (2018) 159–167
Available online 21 March 2018
0167-8809/ © 2018 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/BY-NC-ND/4.0/).
T
be constrained by the cost and availability of insecticides (Banjo et al.,
2003). Traditional and cultural pest management methods include
sprinkling of wood ash on plants, manual removal of pests, beating the
crop with branches, application of kerosene/ash sprays, crop rotation,
intercropping, and leaving land fallow are cheap and readily available,
but their impact may be limited and some of these methods are labor
intensive (Bottenberg, 1995; Alghali, 1991; Amusa et al., 2003; Banjo
et al., 2003).
The ecosystem service of pest regulation provided by natural ene-
mies has been estimated to represent a worldwide value of 100–400
billion USD per year (Costanza et al., 1997; Pimentel et al., 1997). The
effectiveness of natural enemies in suppressing pest populations relies
on both agricultural management at the field scale, and the structure,
composition, and functioning of the surrounding landscape (Tscharntke
et al., 2005; Bianchi et al., 2006; Chaplin-Kramer et al., 2011; Veres
et al., 2013). However, little information is available on the effect of
landscape factors on natural pest control in developing countries, and
Africa in particular. Ironically, natural pest regulation is a critical
ecosystem service for poor smallholder farmers who have limited eco-
nomic access to external inputs and therefore rely on ecosystem services
provided by agroecosystems and their surrounding landscapes. Pro-
moting natural pest regulation may not only improve productivity and
profit, but may also reduce farmers’ dependence on the use of chemical
insecticides, which can have negative impacts on human health and the
environment (Pimentel et al., 1997; Naylor and Ehrlich, 1997; Antle
and Pingali, 1994), and negatively affect natural enemies that suppress
pest populations (Eveleens, 1983; Hansen, 1986). The extent to which
the natural enemy community is conserved and utilized to substitute or
complement chemical insecticides-based pest management has im-
portant implications for the socio-economic and environmental resi-
lience of farming systems in developing countries.
The development of effective policies to support more sustainable
pest management requires a better understanding of the factors that
determine farmers’ pest management decision making within the
landscape, agronomic, socio-economic and biophysical context of
farming systems (Savary et al., 2017). Previous studies have examined
the effects of socio-economic factors on the likelihood of using pro-
duction inputs such as fertilizers and insecticides (e.g., Nkamleu and
Adesina, 2000; Zhou et al., 2010; Waithaka et al., 2007), but studies
that also incorporate agronomic and ecological factors in a household
analysis are scarce. The aim of this paper is to assess the ecological,
agronomic, and socio-economic factors that are associated with farmer
perceptions of the severity of pests in their field crops in Nigeria. The
study comprised three agro-ecological zones spanning a 1000 km
North-South gradient, 102 villages and 805 households. Factors asso-
ciated with reported pest severity are identified and policy implications
are discussed.
2. Materials and methods
2.1. Agro-ecological and socio-economic context
Nigeria encompasses semi-arid savanna ecosystems in the north and
tropical forest ecosystems in the south (Aregheore, 2009). Amidst these
diverse agroecological conditions there is also heterogeneity in ethni-
city and cultures (Aregheore, 2009), as well as vast economic dis-
parities between different regions of the country (Oxford Poverty and
Human Development Initiative, 2015). After a period of marginal ex-
pansion from 1997 to 2007, the area of arable land is now declining
(FAOSTAT, 2016a). Land degradation has been recognized as one of the
most important natural resource management problems in Nigeria,
constraining agricultural and rural development (FAO and ITPS, 2015;
Odemerho, 1992; Titilola and Jeje, 2008). Meanwhile, the population
has been steadily growing at an annual rate of around 2.8% (FAOSTAT,
2016a) and there has been a robust economic growth in the last decade
(African Development Bank Group, 2015). Nevertheless, the proportion
of the population that is multidimensionally poor is 53.3% nationally
and 70% in rural areas, with remarkable regional variation (Oxford
Poverty and Human Development Initiative, 2015).1
Nigeria’s agricultural sector has a relatively high insecticide use as
compared to other African countries. For instance, insecticide import by
Nigeria accounted for 11% of the total import value for the whole of
Africa in 2011 (FAOSTAT, 2016b). Despite a seven-fold increase in net
pesticide imports from US$31 million to US$221 million between 1997
and 2012, progress on increasing cereal production (which is mainly
used for domestic consumption) and per capita food supply has stag-
nated (FAOSTAT, 2016a). While the increase of pesticide inputs has
contributed to the productivity growth of agricultural workers
(FAOSTAT, 2016a), this has not been translated into significant food
security gains. These findings question the effectiveness of strategies
that are solely based on pesticides, and highlight the need for more
sustainable pest management strategies that go beyond pesticide-based
pest management.
2.2. Data collection
2.2.1. Sampling
Survey field work for this study was carried out in Nigeria during
late 2012 – early 2013. The design of the field work was linked to the
midline survey of an impact evaluation study conducted for the Nigeria
Third National Fadama Development Project (“Fadama III” project)
which covered all 37 Nigerian states (Appendix A). Using the sampling
framework established for the Fadama III project, we adopted a stra-
tified sampling approach by first selecting 12 states that covered the
three primary agro-ecological zones (AEZs) in Nigeria: Sudan
Savannah, Guinea Savannah, and Humid Forest (four states for each
AEZ). These states have relatively high poverty rates based on the 2010
Nigeria poverty profile (National Bureau of Statistics, 2012) and a high
incidence of conflicts over the use of common natural resources
(Nkonya et al., Unpublished data). Northeastern states were excluded
from consideration due to security concerns. In each of the 12 states, 6
to 10 villages were randomly selected from the midline survey sample
(Appendix A). Finally, we randomly selected households from each of
the villages, giving us a sample consisting of 851 households from 102
villages, with 34 villages in the Humid Forest zone, 36 in the Guinea
Savannah, and 32 in the Sudan Savannah (Fig. 1). Village and house-
hold surveys were conducted in each selected village. After removing
missing values, outliers and inappropriately measured responses, the
final dataset used in the regression analysis contained data from 805
households. While this sampling strategy was not fully random across
Nigeria (Appendix A), the sample provided comprehensive geographic
coverage of the country and covers all three primary AEZs.
2.2.2. Survey instruments
In the farm household survey (see Appendix B for the household
survey questionnaire), respondents were asked information on house-
holds’ social and demographic characteristics, such as ethnicity, age
and gender of household head, family size, and farm size, as well as
detailed information on pest management. Each household was asked
to report up to three main crops that were grown in the previous
growing season, and to list up to two important insect pests for each
crop. A field guide for insect pests, natural enemies, and pollinators in
15 main crops of Nigeria was developed to assist farmers identifying
insect species. Perceived pest severity was expressed at a 3-level scale
(1= significant yield reduction, 2=moderate yield reduction, and
1 The global Multidimensional Poverty Index (MPI), developed by the Oxford Poverty
& Human Development Initiative (OPHI), is an international measure of acute poverty
covering over 100 developing countries. It complements traditional income-based pov-
erty measures by capturing the severe deprivations that each person faces at the same
time with respect to education, health and living standards (Alkire et al., 2016; OPHI,
2007–2016).
W. Zhang et al. Agriculture, Ecosystems and Environment 259 (2018) 159–167
160
3= little or no yield reduction) for each reported insect pest. These
data were complemented with data on management practices, such as
irrigation, use of improved seeds, use of chemical fertilizer and manure,
and cropping system (monocropping vs. mixed cropping) for each
household from the Fadama III project survey. While the Fadama III
survey was completed six months prior to our survey, and we cannot
rule out changes in management practiced in these six months, this was
the best possible information at hand.
Village level information was elicited using semi-structured group
interviews conducted in local dialects, and included village character-
istics (e.g., road access and distance to the nearest agro-chemical store),
welfare indicators (e.g., the proportion of households in the village that
had less than two meals a day), and prevalence of insecticide use (see
Appendix C for the village survey questionnaire). A land cover and land
use assessment exercise was conducted with each village group, en-
abling us to estimate the proportion of land use area in the village for
eight main land use types, including cultivated land, unused land, re-
sidential area, forest, lowland floodplain, grazing land, woodland, and
water. Ninety-five percent of the villages (97 out of 102) were located
more than 5 kilometers apart, which ensures that villages can be con-
sidered to a large extent independent with respect to landscape effects
on pest and natural enemy communities (Bianchi et al., 2006, 2008;
Rusch et al., 2016; Thies et al., 2003).
2.3. Data analysis
Since the household survey recorded “reported” (or perceived) pest
severity, as opposed to “observed” pest severity, it is important to
consider the socio-economic factors that may influence respondents’
perceptions, along with other important dimensions of production si-
tuations (i.e., landscape, agronomic and biophysical factors) (Table 1).
Natural enemies and their pest suppression services were largely un-
known to Nigerian farmers (Zhang et al., 2016). As information on the
prevalence of natural enemies was not available from survey re-
spondents, we did not model natural enemy presence directly, but in-
stead incorporated their role in the system through two channels. First,
the pest suppression ecosystem service of natural enemies was captured
by the landscape factors with regard to land use types and presence of
non-crop habitat. Second, the mortality effect of insecticide applica-
tions on natural enemies, which in turn may result in pest population
resurgence or increase, was accounted for by incorporating the village-
level extent of insecticide application among farmers.
The relationship between pest severity and chemical insecticide
application is complex: (i) insecticide usage may increase with per-
ceived pest severity, (ii) use of insecticides usually reduce pest levels as
an immediate effect and farmers who use insecticides may tend to ra-
tionalize that insecticide use has decreased pest severity, and (iii) use of
broad-spectrum insecticides will kill natural enemies, making the crop
more susceptible to colonization by pests once the insecticide is no
longer effective. In such reciprocal causality system, estimating the
effect of insecticide use on pest severity with observational data is
challenging. To address the second relationship, insecticide use should
be included as an explanatory variable to explain reported pest severity.
However, this is not desirable from an estimation perspective because it
Fig. 1. Surveyed villages in 12 states across three agro-ecological zones in Nigeria.
W. Zhang et al. Agriculture, Ecosystems and Environment 259 (2018) 159–167
161
violates the assumption of independence of explanatory variables and
the error term, referred to as an endogeneity problem in the econo-
metrics literature (Wooldridge, 2013). In this case, the cause of the
violation arises because of simultaneity, a loop of causality between the
explanatory (household’s insecticide use) and dependent (household’s
reported pest severity) variables of the model. To deal with this issue,
we excluded household’s insecticide use from the model of reported
pest severity and included a village-level variable “the distance to the
nearest insecticide store of the village” to capture insecticides access in
the village, which is expected to affect the probabilities of individual
households using insecticides. The simultaneity concern is mitigated
because pest severity does not affect village-level access to insecticides,
at least not in the short term. Lastly, to control for the third relationship
(impact of insecticides on natural enemies), we included village-wide
extent of insecticide application. Robustness checks were conducted to
provide additional justification for the exclusion of household’s in-
secticide use as an explanatory variable in the reported pest severity
model (Appendix D).
Based on this conceptual framework, we constructed a multiple
regression model to empirically identify the factors associated with
reported pest severity:
PestSeverityiv = β0+ β1*Lv+ β2*Aiv+ β3*Hiv+ β4*Vv+ β5*Biv+ εiv
(1)
where i and v index households and villages, respectively. PestSeverityiv
is pest severity reported by household i in village v and is measured
with a composite “pest severity index”. Lv is a vector of village-level
land use variables that proxy the ecological (landscape) factors; Aiv is a
vector of household-level management variables that capture the
agronomic factors; Hiv is a vector of household socio-economic char-
acteristics; Vv is a vector of village characteristics; Biv is a vector of
biophysical factors including altitude (meters above sea level) recorded
at the household level (Altitudeiv) and dummy variables for agro-eco-
logical zones (Table 1). Lastly, εiv is a random error term.
The pest severity index, PestSeverityiv, was computed by averaging
reported pest severity levels of different pest groups for each household
as
∑ ∑ Pestseveritylevel
Totalnumberofpestgroupsreportedbythehousehold
m
M
n
N
nm , where n and m index insect pest
groups and crop types, respectively, for each household, ∈n (1,2) (as
each household can report up to two insect pest groups per crop type)
and ∈m (1,3) (as each household listed up to 3 crop types). The com-
posite “pest severity index”, ranging between 1 and 3, is thus an ag-
gregate measurement of pest pressure perceived by households. Since
pest severity index is a censored continuous variable (i.e., bounded
between 1 and 3),2 we estimated a Censored Least Absolute Deviations
(CLAD) estimator, which corrects for censoring the dependent variable
(Powell, 1984) and is robust against departures of errors from homo-
scedasticity and normality (Wilhelm, 2008). All regression analyses
were conducted in STATA (StataCorp LP, 2013).
3. Results
3.1. Main field crops, pests and management practices
Thirty-six crops were reported in the survey, but some at low fre-
quencies (Table E-1 in Appendix E). Maize (Zea mays L), cassava
(Manihot esculenta Crantz), and sorghum (Sorghum bicolor L.) were the
most common crops, which were cultivated by 54%, 45% and 32% of
the households, respectively (Fig. 2). Other common crops were yam
(Dioscorea spp.), rice (Oryza sativa L.), cowpea (Vigna unguiculata L.),
millet (Pennisetum glaucum L.), groundnut (Arachis hypogaea L.), and
egusi melon (Colocynthis citrullus L.). Maize and rice were grown in all
three zones, while cassava, yam, and egusi melon were only reported in
the Humid Forest and Guinea Savannah (Fig. 2). Sorghum, cowpea,
millet and groundnut are crops of the drier areas, and were virtually
absent in the Humid Forest.
A total of 54 pest insect groups were reported (Table E-2 in
Appendix E). The most frequently reported pest insects include grass-
hoppers (Caelifera spp.), unspecified caterpillars (Lepidoptera: i.e.
larvae of butterflies and moths), African armyworm (Spodoptera ex-
empta Walker), aphids (Aphidoidea spp.), stemborers (larvae of specific
Lepidoptera species that bore into plant stems), great yam beetle
(Heteroligus meles Billb.), termites (Isoptera spp.), sorghum midge
(Contarinia sorghicola Coq.), and pod borers (larvae of specific
Lepidoptera species that bore into pods). While most insect groups were
reported by households from all AEZs, some pest insect groups were
Table 1
Summary statistics for variables used in the regression analysis (N=805).
Variables Type of
variable
Mean Std. Dev. Min Max
Reported pest severity
(dependent variable
PestSeverityiv)
Censored
continuous
2.17 0.62 1 3
Household-level management
practices (Aiv):
Grew maize Binary 0.54 0.50 0 1
Grew rice Binary 0.20 0.40 0 1
Grew yams Binary 0.31 0.46 0 1
Grew cassava Binary 0.45 0.50 0 1
Grew sorghum Binary 0.32 0.47 0 1
Grew millet Binary 0.16 0.37 0 1
Grew leafy vegetable Binary 0.07 0.25 0 1
Irrigation Binary 0.25 0.44 0 1
Improved varieties Binary 0.63 0.48 0 1
Chemical fertilizer Binary 0.80 0.4 0 1
Manure Binary 0.38 0.5 0 1
Mixed cropping Binary 0.52 0.5 0 1
Crop diversity (number of crop
types)
Continuous 5.23 1.8 1 13
Household socio-economic
characteristics (Hiv):
Age of household head Continuous 51.8 13.0 15 96
Female-headed household Binary 0.07 0.2 0 1
Household size Continuous 10.01 4.7 1 42
Farm size (hectare) Continuous 4.04 8.1 0 110
Ethnicity: Hausa Binary 0.33 0.47 0 1
Ethnicity: Nupe Binary 0.09 0.29 0 1
Ethnicity: Ibo Binary 0.11 0.32 0 1
Ethnicity: Yoruba Binary 0.16 0.37 0 1
Ethnicity: Other Binary 0.30 0.46 0 1
Village-level land use (Lv):
Unused land (area%) Continuous 15.73 20.18 0 90.9
Residential land (area%) Continuous 9.69 11.01 0 57.0
Forest (area%) Continuous 10.38 15.03 0 60.0
Floodplain (area%) Continuous 8.59 8.53 0 47.4
Grazing land (area%) Continuous 0.96 2.98 0 20.0
Woodland (area%) Continuous 3.50 6.47 0 27.3
Water (area%) Continuous 3.00 5.83 0 51.5
Cultivated (area%) Continuous 26.41 23.61 0 100
Village characteristics (Vv):
Percentage of farmers spraying
(%)
Continuous 34.78 38.1 0 100
Distance of village to
insecticide store (km)
Continuous 3.8 6.1 0 30
Percentage of households
eating< 2 meals/day (%)
Continuous 18.79 28.2 0 99
Distance to all-weather road
(km)
Continuous 5.62 12.9 0 100
Biophysical factors (Biv):
Altitude (meters above sea
level)
Continuous 273.1 185.3 14 795
AEZ: Humid Forest Binary 0.35 0.48 0 1
AEZ: Guinea Savannah Binary 0.34 0.47 0 1
AEZ: Sudan Savannah Binary 0.31 0.46 0 1
2 Censoring refers to a condition in which the value of a measurement is not observable
for part of the population (Wooldridge, 2002). Censored regression models developed in
the field of econometrics may be used to handle censored data (e.g., Tobin, 1958;
Schnedler, 2005).
W. Zhang et al. Agriculture, Ecosystems and Environment 259 (2018) 159–167
162
restricted to the AEZs where the host plants are grown (e.g. great yam
beetle was only reported in the yam producing Guinea Savannah and
Sudan Savannah).
Pest insect groups were associated with certain crops. Even when
insect groups were aggregated into 8 major classes, crop-insect re-
lationships remained apparent (Fig. 3). For instance, grasshoppers were
often reported in cassava and to a lesser extent in maize, whereas
Dipteran pests (e.g. African rice gall midge, Orseolia oryzivora) were
important in rice and sorghum. Stem- and pod borers were often re-
ported infesting maize and millet, while leaf feeding Coleoptera were
mainly reported infesting yam.
Three quarters of all reported insect pest cases were considered as
serious or moderate, of which nearly 40% being serious. Chemical
insecticide use was the primary control method in 75% of the reported
cases of pests, whereas cultural control methods (pest management
based on the manipulation of crop systems, e.g., crop rotation, inter-
cropping, and early planting; Agrios, 2005; Goodell, 2009) accounted
for 8.7%. For the top three crops maize, cassava, and sorghum, about
70%, 57% and 78% of the households applied insecticides on pests
affecting the respective crops. Insecticides were applied 2.3 ± 1.9
(mean ± sd), 2.0 ± 0.9, and 2.0 ± 1.0 times for each crop in the
growing season, respectively. While insecticides were used to control a
wide variety of pests, a relatively large proportion of households re-
ported insecticide applications against leaf feeding Lepidoptera. The
most important decision factors for choosing insecticide products were
efficiency (51%) and price (18%). Seventy percent of households
Fig. 2. Frequency distribution of main crops grown in three agro-ecological zones in Nigeria (see Fig. 1 for the location of the agro-ecological zones).
Fig. 3. Frequency distribution of crop-pest groups associations.
W. Zhang et al. Agriculture, Ecosystems and Environment 259 (2018) 159–167
163
reported an increasing trend in insecticide prices over time. Never-
theless, almost half of the households reported an increase in use of
chemical insecticides in the last 5 years. Almost 18% of households
reported using application doses higher than the label recommended
dose.
The far majority of the respondents considered chemical insecticides
hazardous (96%), which they learned from own experience or from
neighbors or friends (62%), extension agents or insecticide salespersons
(20%), and product labels (13%). While the majority of the respondents
took measures for personal safety (e.g., wearing protective clothing
when spraying and washing hands after spraying), unsafe handling and
disposal of empty containers and wash water was widespread. Among
those who washed insecticide sprayers after application (90%), 16%
washed by the river or lake, 21% dumped the wash water into a crop
field, and 19% dumped it anywhere that was convenient. In addition,
37% of the households reported leaving the empty insecticide con-
tainers in the crop field, 16% left them anywhere that was convenient,
and 14% burned the containers.
3.2. Factors associated with reported pest severity
Regression analysis indicated that the proportions of forest, unused
land, and residential area in the village were significantly associated
with lower reported pest severity by households, as compared to cul-
tivated land, which was the reference land use type (Table 2). This
implies that, converting any of these land uses to cultivation land use in
the village may increase the average pest pressure experienced by in-
dividual households, holding everything else constant. The share of
lowland floodplain, relative to cultivated land, was associated with
higher pest severity reported by households.
In terms of management practices, the use of chemical fertilizers
and manure was associated with lower pest severity as compared to
households that didn’t fertilize. Households that adopted mixed-crop-
ping reported lower pest severity than those that didn’t adopt this
practice. Households that grew maize, yam, or cassava reported lower
average pest severity as compared to those that didn’t cultivate each
respective crop, whereas growing rice or leafy vegetables was asso-
ciated with higher reported pest severity. This implies that staple crop
cultivation may be less prone to pest attack than vegetables. Moreover,
the age of household head and female headship were negatively related
to reported pest severity.
In terms of village-level characteristics, the percentage of farmers in
the village that applied insecticides was negatively associated with pest
severity perception, but its squared term had a positive association. The
distance to the nearest chemical insecticide store or to an all-weather
road was not significantly correlated with reported pest severity. With
respect to the biophysical factors, altitude was positively related to
reported pest severity.
4. Discussion
Reported pest severity was associated with a suite of landscape,
agronomic, and socio-economic factors, highlighting the complexity
underlying pest management decisions and the importance of farmers’
production situations.
Reported pest severity was negatively associated with the propor-
tion of forest and unused land, as compared to cultivated land that
served as a reference. This finding is in line with the general pattern of a
positive relationships between the proportion of non-crop habitat in the
landscape, higher and more diverse natural enemy communities (not
quantified in this study), and a tendency for better natural pest sup-
pression (Bianchi et al., 2006; Chaplin-Kramer et al., 2011; Veres et al.,
2013, but see Tscharntke et al., 2016 for counter examples). Our study
indicates that the pest suppressive effects of a diverse landscape hold
also in African smallholder agriculture settings, which is of particular
importance for farmers who lack access to insecticides.
The proportion of residential area was negatively related to reported
pest severity. There are two possible explanations for this finding. First,
residential land in rural Africa often contains home gardens that have a
high plant species diversity and a high structural vegetation complexity
(e.g., Zhang et al., 2016). These home gardens may have supported
natural enemy populations. Second, the area of residential land within a
village domain may be positively associated with income levels and
access to information. Wealthier and better-informed farmers may be
more likely to use insecticides, resulting in lower pest severity. Fur-
thermore, the proportion of lowland floodplain, relative to the pro-
portion cultivated area, was positively correlated with reported pest
severity. This may be explained by the expansion of the area of in-
tensive vegetable farming on floodplains in Nigeria in recent years
(Dam, 2012), which are associated with a high pest severity (Table 2).
In contrast to most ecological studies where land use information is
collected via GIS analyses, we adopted a semi-quantitative approach by
Table 2
Parameter estimates, standard errors and significance levels for variables explaining re-
ported pest severity. Reference variables are “Hausa” for ethnicities, “Cultivated land
(area%)” for land use, and “Humid Forest” for AEZ.
Explanatory variables Parameter s.e.
Village-level land use:
Unused land (area%) −0.004*** 0.001
Residential land (area%) −0.006*** 0.002
Forest (area%) −0.003** 0.001
Floodplain (area%) 0.005** 0.002
Grazing land (area%) −0.005 0.006
Woodland (area%) −0.001 0.003
Water (area%) −0.002 0.003
Household-level management practices:
Irrigation −0.076 0.046
Improved varieties 0.045 0.038
Chemical fertilizer −0.121*** 0.046
Manure −0.144*** 0.047
Mixed cropping −0.090** 0.039
Crop diversity (crop count) −0.004 0.010
Grew maize −0.107*** 0.038
Grew rice 0.094* 0.048
Grew yams −0.172*** 0.044
Grew cassava −0.159*** 0.054
Grew sorghum −0.017 0.047
Grew millet 0.084 0.062
Grew leafy vegetables 0.259*** 0.072
Household socioeconomiccharacteristics:
Age of household head −0.004*** 0.001
Female-headed household −0.140** 0.068
Household size −0.005 0.004
Farm size (hectare) −0.001 0.002
Ethnicity: Nupe 0.032 0.105
Ethnicity: Ibo −0.165 0.120
Ethnicity: Yoruba −0.131 0.106
Ethnicity: Other −0.095 0.095
Village characteristics:
Percentage of farmers spraying (%) −0.006** 0.003
Percentage of farmers spraying squared 6.4e−05** 0.000
Distance of village to insecticide store (km)a 0.004 0.003
Percentage of households eating<2 meals/day (%) 0.001 0.001
Distance to all weather road (km) 0.002 0.001
Biophysical factors:
Altitude (meters above sea level) 0.0003* 0.000
AEZ: Guinea Savannah 0.102 0.069
AEZ: Sudan Savannah −0.204 0.127
Constant 2.981*** 0.146
Observations 805
Pseudo R-squared 0.072
*** p < 0.01.
** p < 0.05.
* p < 0.1.
a The distance is zero for those villages that have an insecticide store in the village.
W. Zhang et al. Agriculture, Ecosystems and Environment 259 (2018) 159–167
164
asking respondents about the major land uses in the village and their
area allocations. Most landscape studies are based on ecological study
designs, with usually a maximum number of landscapes in the order of
20, beyond which the workload for GIS mapping, survey, and travel
time for taking measurements makes the workload insurmountable.
While the survey-based land use assessment approach provides less
detailed data than the GIS approach and has a drawback of potential
reporting bias, its strength is that due to the reduced work load a higher
number of landscape replicates may be obtained when combined with
socioeconomic survey field work (102 villages in this study). This level
of replication is unique for empirical studies that assess landscape ef-
fects on pest pressure. In addition, the survey-based approach may be a
low-tech alternative to GIS mapping when a GIS analysis is not feasible,
for instance in developing countries and regions dominated by small-
holder farms (Zhou et al., 2014).
Reported pest severity was associated with crop management,
which is in line with previous studies (e.g., Allinne et al., 2016; Avelino
et al., 2006; Savary et al., 2017). The use of manure and chemical
fertilizers was negatively associated with pest severity, possibly because
the enhanced soil fertility and soil biota may allow plants to better
compensate for herbivory (Rosenheim et al., 1997; Wilson et al., 2003).
Manure application has earlier been associated with reduced pest
densities (Alyokhin and Atlihan, 2005; Eigenbrode and Pimentel,
1988), while the effects of artificial nitrogen fertilizer on pest dynamics
are mixed and may differ for sucking and chewing insect herbivores
(Nicholls and Altieri, 2004). Reported pest severity was lower in mixed-
cropping systems than in mono-cropping systems. Indeed, the pest
suppression benefit of diverse cropping systems has been recognized for
a long time (Andow, 1991; Root, 1973; Altieri, 2004; Thrupp, 2000;
Tsafack et al., 2016). Our findings support the premise that diversified
agroecosystems have a higher capacity to suppress pest (Altieri, 2004;
Thrupp, 2000), and that cultural control methods offer complementary
suppression without or with a reduced dependency on chemical in-
secticides.
Our data indicated that household characteristics were associated
with pest severity perception, which has important implications for the
development of policy tools and targeting strategies. We found that
female-headed households reported lower pest severity, which corro-
borates the findings of Nkamleu and Adesina (2000) who found that
female household headship was negatively correlated with insecticide
use in Cameroon. More generally, men tend to use higher input levels
than women, and that this input gap is responsible for observed pro-
ductivity differences between men and women (Peterman et al., 2014).
Our analysis offers a new perspective on the relationship between fe-
male headship and pest severity perception in Nigeria, which merits
further investigation.
Households with older heads tended to report lower pest severity
than households with younger heads. A possible explanation is that
older farmers may have more farming experience and may be able to
apply better pest management through alternative control methods.
Furthermore, older farmers may perceive pest severity lower as they
may be more aware of the potential of crops to tolerate or compensate
for pest attack, and are therefore more likely to refrain from using in-
secticides. Evidence on the effect of farmer age on technology adoption
is mixed in the literature. While some studies show that younger
farmers are generally more likely to adopt new technologies (e.g.,
Alavalapati et al., 1995), others have argued that older farmers may
have preferential access to new information or technologies or are more
likely to invest in innovations because of greater accumulated personal
capital (Nkamleu et al., 1998; Nkamleu and Adesina, 2000). In Ca-
meroon, fertilizer use or insecticide use was not related to age of
farmers (Nkamleu and Adesina, 2000). Our data from Nigeria suggest
that using knowledge of more experienced farmers may help to reduce
the reliance on insecticides.
The negative and positive coefficients for the percentage of farmers
at the village level that applied insecticides and its squared term,
respectively (Table 2), indicate a nonlinear relationship with low per-
ceived pest severity at low levels of insecticide use and increasing
perceived pest severity at higher pesticide use at the village level. One
can conceive that as insecticide application becomes more widespread
across local systems, the non-target mortality effects of insecticide use
on natural enemies becomes an increasingly important driving factor in
pest densities, resulting in higher reported pest severity, whereas at low
pesticide use frequency with sufficient refuge for natural enemies, pest
severity might still decrease with greater use. As compared to many
emerging economies in Asia, the current rate of insecticide application
in Nigeria is still moderate (Bell et al., 2016; Huang et al., 2010; Zhou
et al., 2014), but great attention needs to be given to the possibility of
secondary pest outbreaks as farmers continue to expand the use of
broad-spectrum insecticides.
In this study, we used reported or perceived pest severity but did not
measure actual pest levels in the field. The quantification of pest levels
was not feasible given that the study comprised more than 800
households, representing a high diversity of crops and associated pest
species, and a high temporal variation in pest densities requiring mul-
tiple assessments in the growing season. Even if the reported pest se-
verities would be biased by personal attitudes and perceptions (beyond
what we have controlled for in the model), they are still relevant for
decision making because the reported or perceived pest severity in-
forms decision making. Therefore, it could be argued that from a so-
ciological and decision-making perspective, the perceived pest severity
is more relevant than the actual pest severity. Evidently, further re-
search on the relationship between actual and perceived pest severity is
needed to obtain deeper insight in drivers of pesticide use by farmers.
5. Conclusions
In conclusion, our findings indicate that the presence of non-crop
areas in the landscape and the diversification of agroecosystems may be
a viable strategy for smallholder farmers to manage pests with limited
reliance of chemical insecticides in Nigeria, but that actual pest man-
agement decisions could be influenced by a wide range of context-
specific factors. Looking at the broader implications of the study, re-
ducing yield loss to pests while reducing the reliance on chemical in-
secticides is a major challenge in Nigeria, but it is also an important
component in achieving sustainable food security and development.
Closing yield gaps via pest management requires addressing both the
constraints around (i) the sustainable use of ecosystem services asso-
ciated with biodiverse landscapes and agro-ecosystems that reduce the
need for chemical insecticides, and (ii) access to selective and en-
vironmentally benign insecticide products, their informed use, and the
affordability to farmers.
Acknowledgements
The research was carried out under the CGIAR Research Programs
on Policies, Institutions and Markets (PIM) and Water, Land and
Ecosystems (WLE) with support from CGIAR Fund Donors (http://
www.cgiar.org/about-us/our-funders/). We thank Ephraim Nkonya,
Hassan Ishaq Ibrahim, Mure Agbonlahor, and Hussaini Yusuf Ibrahim
for contributing to the data collection, Adebayo A. Omoloye for pro-
viding inputs on field crop pests in Nigeria, and Samson K. Foli,
Adebayo A. Omoloye, and James Ojo for preparing a field guide on
insects to facilitate the farmer interviews. We are grateful for farmers
who participated in our survey, the enumerators who carried out the
interviews, and the extension agents for facilitating the field work.
Figure 1 is used with permission from Zhang, W., E. Kato, P. Bhandary,
E. Nkonya, H.I. Ibrahim, M. Agbonlahor, H.Y. Ibrahim, and C. Cox,
“Awareness and Perceptions of Ecosystem Services in Relation to Land
Use Types: Evidence from Rural Communities in Nigeria” Ecosystem
Services 22(A): 150-160. Elsevier, copyright 2016.
W. Zhang et al. Agriculture, Ecosystems and Environment 259 (2018) 159–167
165
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the
online version, at https://doi.org/10.1016/j.agee.2018.03.004.
References
African Development Bank Group, 2015. Economic Report on Nigeria: 2015 Special
Edition. African Development Bank, Belvedere, Tunisia.
Agrios, G.N., 2005. Plant Pathology, fifth edition. Academic Press, New York, pp.
272–273 (952 pp.).
Alavalapati, J.R., Luckert, M.K., Gill, D.S., 1995. Adoption of Agroforestry Practices: A
Case Study from Andhra Pradesh, India .Agroforesty Systems. Kluwer Academic
Publishers.
Alghali, A.M., 1991. Studies on cowpea farming practices in Nigeria, with emphasis on
insect pest control. Trop. Pest Manage. 37 (1), 71–74.
Alkire, S., Jindra, C., Robles, G., Vaz, A., 2016. Multidimensional poverty index – 2016:
brief methodological note and results. OPHI Briefing 42, University of Oxford. .
http://www.ophi.org.uk/wp-content/uploads/OPHIBrief_42_MPI_meth_note_2016.
pdf.
Allinne, C., Savary, S., Avelino, J., 2016. Delicate balance between pest and disease in-
juries, yield performance, and other ecosystem services in the complex coffee-based
systems of Costa Rica. Agric. Ecosyst. Environ. 222, 1–12.
Altieri, M.A., 2004. Linking ecologists and traditional farmers in the search for sustain-
able agriculture. Front. Ecol. Environ. 2 (1), 35–42.
Alyokhin, A., Atlihan, R., 2005. Reduced fitness of the colorado potato beetle (Coleoptera:
chrysomelidae) on potato plants grown in manure-amended soil. Environ. Entomol. 34
(4), 963–968.
Amusa, N.A., Adegbite, A.A., Muhammed, S., Baiyewu, R.A., 2003. Yam diseases and its
management in Nigeria. Afr. J. Biotechnol. 2 (12), 497–502.
Andow, D.A., 1991. Vegetational diversity and arthropod population response. Annu.
Rev. Entomol. 36 (1), 561–586.
Antle, J.M., Pingali, P., 1994. Insecticides, productivity, and farmer health: a Philippine
case study. Am. J. Agric. Econ. 76, 418–430.
Aregheore, E.M., 2009. Country Pasture/Forage Resources Profiles: Nigeria. Food and
Agriculture Organisation of the United Nations, Italy.
Avelino, J., Zelaya, H., Merlo, A., Pineda, A., Ordoñez, M., Savary, S., 2006. The intensity
of a coffee rust epidemic is dependent on production situations. Ecol. Model. 197,
431–447.
Banjo, A.D., Lawal, O.A., Fapojuwo, O.E., Songonuga, E.A., 2003. Farmers’ knowledge
and perception of horticultural insect pest problems in southwestern Nigeria. Afr. J.
Biotechnol. 2 (11), 434–437.
Bature, Y.M., Sanni, A.A., Adebayo, F.O., 2013. Analysis of impact of national fadama
development projects on beneficiaries income and wealth in FCT, Nigeria. J. Econ.
Sustain. Dev. 4 (17), 23 (11).
Bell, A.R., Zhang, W., Nou, K., 2016. Pesticide use and cooperative management of nat-
ural enemy habitat in a framed field experiment. Agric. Syst. 143, 1–13.
Bianchi, F.J.J., Booij, C.J.H., Tscharntke, T., 2006. Sustainable pest regulation in agri-
cultural landscapes: a review on landscape composition, biodiversity and natural pest
control. Proc. R. Soc. Biol. Sci. 273 (1595), 1715–1727.
Bianchi, F.J.J., Goedhart, P.W., Baveco, J.M., 2008. Enhanced pest control in cabbage
crops near forest in the Netherlands. Landsc. Ecol. 23, 595–602.
Bottenberg, Harry, 1995. Farmers perception of crop pests and pest control practices in
rainfed cowpea cropping systems in Kano, Nigeria. Int. J. Pest Manage. 41 (4),
195–200.
Chaplin-Kramer, R., O’Rourke, M.E., Blitzer, E.J., Kremen, C., 2011. A meta-analysis of
crop pest and natural enemy response to landscape complexity. Ecol. Lett. 14 (9),
922–932.
Costanza, R., d'Arge, R., de Groot, R., Farber, S., Grasso, M., Hannon, B., Limburg, K.,
Naeem, S., O'Neill, R.V., Paruelo, J., Raskin, R.G., Sutton, P., van den Belt, M., 1997.
The value of the world's ecosystem services and natural capital. Nature 387 (15),
253–260.
Dam, P.D., 2012. Dry season vegetable farming in the floodplains of river Katsina-Ala in
Katsina-Ala town of Benue state, Nigeria. J. Environ. Issues Agric. Dev. Countries 4
(1), 18–23.
Eigenbrode, S.D., Pimentel, D., 1988. Effects of manure and chemical fertilizers on insect
pest populations on collards. Agric. Ecosyst. Environ. 20, 109–125.
Eveleens, K.G., 1983. Cotton insect control in the Sudan Gezira: analysis of a crisis. Crop
Prot. 2, 273–287.
FAOSTAT, 2016a. Country Profile: Nigeria. (Accessed on February 26, 2016. http://
faostat3.fao.org/home/E).
FAOSTAT, 2016b. Inputs –Insecticides Trade. (Accessed on February 1, 2016. http://
faostat.fao.org/site/423/default.aspx#ancor).
FAO, ITPS, 2015. Status of the world’s soil resources – main report. Food and Agriculture
Organization of the United Nations and Intergovernmental Technical Panel on Soils.
Rome, Italy.
Goodell, P., 2009. Fifty years of the integrated control concept – the role of landscape
ecology in IPM in San Joaquin valley cotton. Pest Manage. Sci. 65, 1293–1297.
Hansen, M., 1986. Escape from the Insecticide Treadmill, Alternatives to Insecticides in
Developing Countries. Institute for Consumer Policy Research, New York.
Huang, J.K., Mi, J.W., Lin, H., Wang, Z.J., Chen, R.J., Hu, R.F., Rozelle, S., Pray, C., 2010.
A decade of Bt cotton in Chinese fields: assessing the direct effects and indirect ex-
ternalities of Bt cotton adoption in China. Sci. China Life Sci. 53, 981–991.
National Bureau of Statistics, 2012. Nigeria Poverty Profile 2010. FCT, Abuja Nigeria.
Naylor, R.L., Ehrlich, P.R., 1997. Natural pest control services and agriculture. Nature's
Services: Societal Dependence on Natural Ecosystems 151–174.
Ndubuaku, T.C.N., Asogwa, E.U., 2006. Strategies for the control of pests and diseases for
sustainable cocoa production in Nigeria. Afr. Sci. 7 (4).
Nicholls, C.I., Altieri, M.A., 2004. Agroecological bases of ecological engineering for pest
management. In: Gurr, G.M., Wratten, S.D., Altieri, M.A. (Eds.), Ecological
Engineering for Pest Management: Advances in Habitat Manipulation for Arthropods.
CSIRO publishing (225 pp).
Nkamleu, G.B., Adesina, A.A., 2000. Determinants of chemical input use in peri-urban
lowland systems-bivariate probit analysis in Cameroon. Agric. Syst. 63, 111–121.
Nkamleu, G.B., Coulibaly, O., Tamo, M., Ngeve, J.M., 1998. Adoption of Storage Pest
Control Technologies by Cowpeas’ Traders in Western Cameroun: Probit Model
Application Monograph. International Institute of Tropical Agriculture.
Nkonya E., Phillip D., Kato E., Ahmed B., Daramola A., Ingawa S.B., Luby I., Lufadeju E.
A., Madukwe M., Shettima A.G., 2013. unpublished data.
Odemerho, F.O., 1992. Land degradation impacts on tropical agricultural basins and their
management: the Nigerian example. GEOFORUM 23 (4), 499–506.
Oerke, E.C., 2006. Crop losses to pests. J. Agric. Sci. 144 (1), 31–43.
Ofor, M.O., Ibeawuchi, I.I., M.Oparaeke, A., 2009. Crop protection problems in produc-
tion of maize and guinea corn in northern Guinea Savanna of Nigeria and control
measures. Nat. Sci. 7 (12), 8–14.
Oxford Poverty and Human Development Initiative (OPHI), 2015. Nigeria Country
Briefing, Multidimensional Poverty Index Data Bank. OPHI, University of Oxford
(December. Available at: http://www.ophi.org.uk/multidimensional-poverty-index/
mpi-2015/mpi-country-briefings/).
Penning de Vries, F.W.T., 1982. Systems analysis and models of crop growth. In: Penning
de Vries, F.W.T., Van Laar, H.H. (Eds.), Simulation of Plant Growth and Crop
Production. Pudoc, Wageningen, pp. 9–19.
Peterman, A., Behrman, J.A., Quisumbing, A.R., 2014. A review of empirical evidence on
gender differences in nonland agricultural inputs, technology, and services in de-
veloping countries. In: Quisumbing, A.R., Meinzen-Dick, R., Raney, T.L.,
Croppenstedt, A., Behrman, J.A., Peterman, A. (Eds.), Gender in Agriculture: Closing
the Knowledge Gap. The Food and Agriculture Organization (FAO) of the United
Nations and Springer Science + Business Media B.V., Dordrecht.
Pimentel, D., Wilson, C., McCullum, C., Huang, R., Dwen, P., Flack, J., Tran, Q., Saltman,
T., Cliff, B., 1997. Economic and environmental benefits of biodiversity. Bioscience
47 (11), 747–757.
Powell, J.L., 1984. Least absolute deviations estimation for the censored regression
model. J. Econometr. 25 (3), 303–325.
Root, R.B., 1973. Organization of a plant-arthropod association in simple and diverse
habitats: the fauna of collards (Brassica Oleracea). Ecol. Monogr. 43 (1), 95–124.
Rosenheim, J.A., Wilhoit, L.R., Goodell, P.B., Grafton-Cardwell, E.E., Leigh, T.F., 1997.
Plant compensation, natural biological control, and herbivory by Aphis gossypii on
pre-reproductive cotton: the anatomy of a non-pest. Entomol. Exp. Appl. 85, 45–63.
Rusch, Adrien, Chaplin-Kramer, Rusch, Gardiner, Mary M., Hawro, Violetta, Holland,
John, Landis, Douglas, Thies, C., Tscharntke, T., Weisser Wolfgang, W., Winqvist, C.,
Woltz, M., Bommarco, R., 2016. Agricultural landscape simplification reduces natural
pest control: a quantitative synthesis. Agric. Ecosyst. Environ. 221, 198–204. http://
dx.doi.org/10.1016/j.agee.2016.01.039.
Savary, S., Mille, B., Rolland, B., Lucas, P., 2006a. Patterns and management of crop
multiple pathosystems. Eur. J. Plant Pathol. 115 (1), 123–138.
Savary, S., Teng, P., Willocquet, L., Nutter, F.J., 2006b. Quantification and modeling of
crop losses: a review of purposes. Ann. Rev. Phytopathol. 44, 89–112.
Savary, S., McRoberts, N., Esker, P.D., Willocquet, L., Teng, P.S., 2017. Production si-
tuations as drivers of crop health: evidence and implications. Plant Pathol. 66,
867–876. http://dx.doi.org/10.1111/ppa.12659.
Schnedler, Wendelin, 2005. Likelihood estimation for censored random vectors. Econom.
Rev. 24 (2), 195–217. http://dx.doi.org/10.1081/ETC-200067925.
StataCorp, 2013. Stata Statistical Software: Release 13. College Station, TX : StataCorp LP.
Thies, C., Steffan-Dewenter, I., Tscharntke, T., 2003. Effects of landscape context on
herbivory and parasitism at different spatial scales. Oikos 101, 18–25.
Thrupp, L.A., 2000. Linking agricultural biodiversity and food security-the valuable role
of agrobiodiversity for sustainable agriculture. Int. Aff. 76 (2), 265–281.
Titilola, S.O., Jeje, L.K., 2008. Environmental degradation and its implications for agri-
cultural and rural development: the issue of land erosion. J. Sustain. Dev. Afr. 10 (2),
116–146.
Tobih, F.O., Okonmah, L.U., Omoloye, A.A., 2011. Assessment of yield potentials and
damage of yams in uncontrolled upland yam monocrop system with varying planting
dates in Oshimili area of delta state, Nigeria. Int. J. AgriSci. 1 (3), 811–871.
Tobin, James, 1958. Estimation of relationships for limited dependent variables.
Econometrica 26 (1), 24–36. http://dx.doi.org/10.2307/1907382. (JSTOR
1907382).
Tsafack, N., Alignier, A., Head, G.P., Kim, J.H., Goulard, M., Menozzi, P., Ouin, A., 2016.
Landscape effects on the abundance and larval diet of the polyphagous pest
Helicoverpa armigera in cotton fields in North Benin. Pest Manage. Sci. 72 (8),
1613–1626. http://dx.doi.org/10.1002/ps.4197.
Tscharntke, T., Klein, A.M., Kruess, A., Steffan-Dewenter, I., Thies, C., 2005. Landscape
perspectives on agricultural intensification and biodiversity – ecosystem service
management. Ecol. Lett. 8, 857–874.
Tscharntke, T., Karp, D., Chaplin-Kramer, R., Batáry, P., deClerck, F., Gratton, C., Ives, A.,
Jonsson, M., Martin, E., Martínez-Salinas, A., Meehan, T.D., O’Rourke, M., Poveda,
K., Rosenheim, J.A., Rusch, A., Schellhorn, N., Wratten, S., Zhang, W., 2016. When
natural habitat fails to enhance biological pest control –five hypotheses. Biol.
Conserv. 204 (Part B), 449–458.
Veres, A., Petit, S., Conord, C., Lavigne, C., 2013. Does landscape composition affect pest
abundance and their control by natural enemies? A review. Agric. Ecosyst. Environ.
W. Zhang et al. Agriculture, Ecosystems and Environment 259 (2018) 159–167
166
138, 1–8.
Waithaka, M.M., Thornton, P.K., Shepherd, K.D., Ndiwa, N.N., 2007. Factors affecting the
use of fertilizers and manure by smallholders: the case of Vihiga, western Kenya.
Nutr. Cycl. Agroecosyst. 78, 211–224.
Wilhelm, Mark Ottoni, 2008. Practical considerations for choosing between Tobit and
SCLS or CLAD estimators for censored regression models with an application to
charitable giving. Oxf. Bull. Econ. Stat. 70 (4), 0305–9049.
Wilson, L.J., Sadras, V.O., Heimoana, S.C., Gibb, D., 2003. How to succeed by doing
nothing: cotton compensation after simulated early season pest damage. Crop Sci. 43,
2125–2134.
Wooldridge, Jeffrey M., 2002. Econometric Analysis of Cross Section and Panel Data. The
MIT Press, Cambridge, MA, pp. 517.
Wooldridge, Jeffrey M., 2013. Introductory Econometrics: A Modern Approach (Fifth
International Ed.). pp. 82–83 (Australia : South-Western. ISBN 978-1-111-53439-4).
Zakari, S., Ying, L., Song, B., 2014. Factors influencing household food security in west
Africa: the case of southern Niger. Sustainability 6, 1191–1202.
Zhang, W., Kato, E., Bhandary, P., Nkonya, E., Ibrahim, H.I., Agbonlahor, M., Ibrahim,
H.Y., Cox, C., 2016. Awareness and perceptions of ecosystem services in relation to
land use types: evidence from rural communities in Nigeria. Ecosyst. Serv. 22 (A),
150–160.
Zhou, Y., Yang, H., Mosler, H., Abbaspour, K.C., 2010. Factors affecting farmers’ decisions
on fertilizer use: a case study for the Chaobai watershed in Northern China.
Consilience: J. Sustain. Dev. 4 (1), 80–102.
Zhou, K., Huang, J.K., Deng, X.Z., van der Werf, W., Zhang, W., Lu, Y.H., Wu, K.M., Wu,
F., 2014. Effects of land use and insecticides on natural enemies of aphids in cotton:
first evidence from smallholder agriculture in the North China Plain. Agric. Ecosyst.
Environ. 183, 176–184.
W. Zhang et al. Agriculture, Ecosystems and Environment 259 (2018) 159–167
167