HESSHydrology and Earth System SciencesHESSHydrol. Earth Syst. Sci.1607-7938Copernicus PublicationsGöttingen, Germany10.5194/hess-20-1911-2016The use of semi-structured interviews for the characterisation of farmer irrigation practicesO'KeeffeJimmyj.okeeffe12@imperial.ac.ukBuytaertWouterhttps://orcid.org/0000-0001-6994-4454MijicAnaBrozovićNicholasSinhaRajivhttps://orcid.org/0000-0002-3321-8095Department of Civil and Environmental Engineering, Imperial College London, London, UKRobert B. Daugherty Water for Food Institute, University of Nebraska, Lincoln, Nebraska, USADepartment of Earth Sciences, Indian Institute of Technology Kanpur, Uttar Pradesh, IndiaJimmy O'Keeffe (j.okeeffe12@imperial.ac.uk)12May20162051911192419June201524August201522April201626April2016This work is licensed under a Creative Commons Attribution 3.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by/3.0/This article is available from https://hess.copernicus.org/articles/20/1911/2016/hess-20-1911-2016.htmlThe full text article is available as a PDF file from https://hess.copernicus.org/articles/20/1911/2016/hess-20-1911-2016.pdf
For the development of sustainable and realistic water security, generating
information on the behaviours, characteristics, and drivers of users, as well
as on the resource itself, is essential. In this paper we present a
methodology for collecting qualitative and quantitative data on water use
practices through semi-structured interviews. This approach facilitates the
collection of detailed information on actors' decisions in a convenient and
cost-effective manner. Semi-structured interviews are organised around a
topic guide, which helps lead the conversation in a standardised way while
allowing sufficient opportunity for relevant issues to emerge. In addition,
they can be used to obtain certain types of quantitative data. While not as
accurate as direct measurements, they can provide useful information on local
practices and users' insights. We present an application of the methodology
on farmer water use in two districts in the state of Uttar Pradesh in
northern India. By means of 100 farmer interviews, information was collected on
various aspects of irrigation practices, including irrigation water volumes,
irrigation cost, water source, and their spatial variability. Statistical
analyses of the information, along with data visualisation, are also
presented, indicating a significant variation in irrigation practices both within and
between districts. Our application shows that semi-structured interviews are
an effective and efficient method of collecting both qualitative and
quantitative information for the assessment of drivers, behaviours, and their
outcomes in a data-scarce region. The collection of this type of data could
significantly improve insights on water resources, leading to more realistic
management options and increased water security in the future.
Introduction
The interactions between humans and water resources are often poorly
understood: an issue which can be reflected in the decisions behind water
resource planning. While some anthropogenic influences, such as greenhouse
gas emissions and land use change, have been incorporated in much of the
current modelling and decision-making framework, less work has been carried
out on the human–water interface . This shortfall is seen
as a major challenge in earth system modelling and
consequently decisions on water resource management. Given that human-induced
issues of water scarcity affect many parts of the world , there is a need to
understand anthropogenic–hydrological linkages in order to better manage
water resources in the future. Socio-hydrology provides a means of supporting
sustainable societal development in a changing environment
. Indeed, the significance of including so-called soft
data has been well documented (see , and
). argue the importance of
including qualitative information to improve model realism; while this may
lead to reduced model efficiency, it can help produce a more realistic
representation of catchment behaviour. Making use of this “experimental
common sense” is an important step in more accurately
representing anthropogenic water use in models. While this paper is primarily
concerned with data collection, the importance of obtaining and using soft,
qualitative data is implied. Globally, irrigation water consumption accounts
for some 70 % of total groundwater and surface water withdrawals
. This figure has increased dramatically over the last 60
years, largely as a result of population growth, market expansion, and
technological advances in water abstraction. Consequently, irrigation water
use needs to be explored in more detail than non-irrigative demand
.
Representing water use presents many challenges, many of which stem from a
lack of data . This often leads to
oversimplification, either in resolution or in user
behaviour, which can subsequently be reflected in model outputs. For example,
irrigation water requirements are often calculated based on the ideal crop
water requirement (see , and ),
giving a false representation of what is actually taking place on the ground
– as users will often over- or under-irrigate depending on prevailing
social, economic, or environmental conditions. Large-scale model outputs or
data representations also provide excellent tools for examining water use or
resource trends . While such approaches are
useful as an overview of large-scale issues, they are inadequate for
developing realistic solutions at a meaningful, implementable level. The data
collection methods described in this paper are aimed at providing information
for more local-scale models and decision making, particularly in instances
where such information is scarce. This dearth of information includes both
quantitative and qualitative data. In order to come up with suitable options
for the use of water, it is important to generate information at a realistic
spatial resolution, not only on the water resource itself but also on the
behaviours, characteristics, and drivers of its managers and users.
In social sciences and healthcare the collection of both qualitative and
quantitative information through interviews is relatively common practice
; however, such
methods are less used in the fields of earth and engineering sciences. For
the purposes of data collection for hydrological studies little guidance
exists. In a both time- and resource-constrained setting the use of
semi-structured interviews provides an efficient and effective method for
qualitative and quantitative data collection. This is particularly true of
data-scarce regions, as in our case study, where limited field information
exists. According to , using an ethnographic methodology
is useful in instances where the theory is incomplete and the phenomena are
observable and important at a local level. For the most part little room
exists for the inclusion of “non-experts” into the application of
scientific research methods . The incorporation of local
knowledge however can have many advantages, including better defining the
research questions and raising locally important, as well as unimportant,
factors. Unlike a structured interview which contains a series of set
questions asked the same way to all interviewees, a semi-structured interview
is organised around a topic guide. The topic guide ensures the main points of
interest are satisfied during the interview , while still
allowing the overall direction to be shaped by the participants' own
understanding, so-called experiential or traditional knowledge, of their
environment. This naturally highlights issues which are of most importance to
the interviewee and allows room to incorporate new themes. Semi-structured
interviews can quickly produce rich and detailed data sets
offering an accurate assessment of the characteristics of individuals and
phenomena. Importantly, it can also shed light on the drivers of these events
and the motivations behind user decisions, providing a valuable contribution
to earth systems modelling. Semi-structured interviews allow for the
collection of qualitative and quantitative information efficiently and cost
effectively, in an unobtrusive and open manner. While qualitative approaches
such as semi-structured interviews are widely recognised and regularly
applied by social scientists working on water resources, they are scarcely
used by natural scientists in the context of hydrology and modelling. In this
paper we show how the method can be used for hydrological research; however,
we see much greater scope for interdisciplinary dialogue on semi-structured
interviews and its broader relevance in addressing hydrological model
uncertainties. Aspects of the approach reported herein may differ from
traditional methods (see , and ),
for example in terms of sampling. However, we believe semi-structured
interviews provide an effective tool for data collection on water use. In
this study, we applied this approach to two districts in the northern Indian
state of Uttar Pradesh to study irrigation water use, and the results are
presented as a case study in Sect. 3, with the methodology used described in
Sect. 2.
MethodologyStudy preparation and interview design
The collection of qualitative and quantitative data in the field requires an
understanding of the social nuances that exist in a study region, as well as
the relevant existing published research. This knowledge is essential in the
planning phase, including the design of the topic guide, around which the
semi-structured interview is based . The literature review
and pre-fieldwork planning, which should also take practicalities such as
logistics and cost into account, help define the main study area and the
target interview participants. In this paper we treat the semi-structured
interview purely as a tool for the collection of hydrological data in the
field. Careful and consistent phrasing of questions in the interview is
important and draws on the pre-fieldwork research as well as knowledge of the
local characteristics. Questions should be unambiguous and easily understood
by interviewees, related to their own experiences, and ethically and culturally
sensitive, and they should ensure that they assist, rather than impede, the flow of
information. In addition, the interviewer must ensure that the questions
provide data which will address the research questions appropriately
. Interviewees may not be able to give a direct answer to a
technical question; however, skilfully crafted component questions can be
combined to produce the required information (e.g. abstraction rates achieved
via depth of water applied and irrigated area).
A significant advantage of semi-structured interviews is the opportunity for
previously unknown information to emerge. Participants can be regarded as
experts by experience; therefore when sufficient opportunity to speak freely
is provided, new and novel information can emerge. This approach allows both
quantitative and qualitative data extraction, for example the volume of water
a farmer takes from a particular source and their reason for this. This
approach can yield considerable benefits in terms of cost whilst ensuring a
useful representation of parameters. Semi-structured interviews are
traditionally comprised of open-ended questions. The collection of
quantitative data, however, is best achieved through direct questions. For
this reason the topic guide used in the case study contains both open-ended
and direct questions (see Supplement). While acquiring quantitative
information in this manner is not as accurate as, for example, metered data,
we believe this approach can provide a useful representation of the important
parameters and has a place in situations where other measures could be
considered unacceptable to the sample or unfeasible in the environment.
Sampling
Sampling comprises an integral part of study design. It allows us to select
cases from a wider population, too big to be studied completely, enabling us
to generalise the final research conclusions to an entire population, not
just to the individual participants of a study . This is an
important consideration when collecting information which could be used in
policy, as any decisions arising from the data should be as applicable as
possible to the wider population. The sampling procedure traditionally
adopted with semi-structured interviews does not aim to achieve a
representative sample. However, a representative sampling was a useful
strategy for the purpose of the case study reported herein, in order to
produce more universally acceptable results. This is achieved through a
combination of sampling techniques. For example, purposive sampling provides
a useful starting point by selecting participants who are thought to be
information rich. Purposive sampling involves the random selection of
sampling units from a part of the population likely to contain the most
information on the characteristics of interest to the researcher
. Purposive sampling allows subjects to be selected based
on their characteristics; while this approach is often used to highlight and
study extreme or deviant cases, it can allow the researcher to target sample
populations which are likely to provide information of most relevance to the
research questions. Once a sample group has been identified, randomisation
should take place to ensure a representative cross section of the study group
is achieved. Prior to undertaking fieldwork it is necessary to set
participant inclusion and exclusion criteria, as it is likely that potential
interviewees who fall outside the research area interests will be approached.
Inclusion and exclusion criteria help promote the best use of available
resources.
Conducting the interview
Correct introduction of the study to potential participants is essential when
gaining informed consent. This involves a clear and concise explanation of
the purpose of the research, what the interview will involve, and how you are
going to use and store the information collected. It should also be
highlighted that the respondent is under no obligation to answer any of the
questions if they do not wish to . This component of the
research is important not only in creating the right kind of environment
where the interviewee feels they can provide the information, but also in
building good rapport with the individual . The subject
of ethics is an important consideration when entering other people's
environments and collecting data on their livelihoods. While it is outside
the scope of this paper to provide guidelines on ethics, it is strongly
recommended that they are taken into account during the planning stage of the
study.
Semi-structured interviews may need to be carried out via translator(s).
Pre-project training should be provided to translators beforehand to ensure
consistency in terms of interview style. In the field, interviews may be
conducted in the presence of family members or neighbours. While for
practical and cultural reasons it may not be possible to avoid this, care
should be taken at all times to address the question to and receive the
response from the designated participant, bearing in mind the potential
impact others' presence may have on the answers received. It is important
that the interview is recorded in as much detail as possible, ideally through
a mixture of field notes and a voice recorder. Again, consent should be
sought from the interview participant prior to the recording of any
conversation. GPS readings of where the interview takes place and any other
pertinent locations, for example wells or canal access points, should also be
taken, along with photos and samples where applicable. Data should be stored
safely and securely following all applicable institutional guidelines. It
should be made clear to the participants that their privacy and
confidentiality will be maintained to the highest degree possible.
Data processing and analyses
Following the collection of data, all interviews should be transcribed
verbatim. While time-consuming, a full transcription is paramount in avoiding
bias introduced through selective data extraction by the researcher, who may
have particular themes or research questions in mind. It also ensures that
all data remain available for further analyses, rather than what is of
interest to the researcher at that time. Reading the transcripts results in
various themes emerging from the text, from which a thematic analysis begins.
Themes are referred to as codes during the analysis. As the analysis
progresses, commonality of codes across interviews may become apparent.
However, thematic analysis allows new themes or ideas to constantly emerge.
The use of qualitative data analysis software, for example RQDA
, provides a useful platform for processing large amounts of
qualitative data. Words or sections from a discussion are coded, allowing the
frequency and relationships across topics to be analysed .
While the analysis of textual data can be a difficult process, it is made
more straightforward using the appropriate software. It is also important to
note that that such tools do not analyse the data, which is the task of the
researcher; they only make the handling of such data more straightforward
. This also allows information, both qualitative and
quantitative, on each theme to be recalled easily. Once the data have been
coded, the dominant themes can be identified. Overviews of the distributions
of variables within the database can also be produced. A significant portion
of the data collected may also be quantitative and suitable for some
statistical analyses and modelling purposes.
Case study – data collectionStudy region: the Ganges Basin, northern India
The Green Revolution has led to enormous gains in agricultural productivity
in India, largely through the use of more reliable seeds and improved
irrigation technology . This has allowed India to become
food self-sufficient and has undoubtedly improved life for
the majority of rural poor. The Indian Green Revolution has also received
much criticism for its environmental and socio-economic impacts. This
includes a reduction in India's water resources while becoming one of the
most intensely irrigated areas of the world (;
; ). However, to correctly
investigate water security, field studies and an understanding of the often
highly localised spatial variations in water abstraction need to be
considered. While the large-scale impacts on water resources are known, the
factors influencing irrigation practices on a local level are much less
understood. In order to develop realistic and socially acceptable options for
water use in the future, this local variability needs to be taken into
account.
Uttar Pradesh (UP), located on the plains of the Ganges Basin, is the highest
producer of food grains and sugarcane in the country and
the most densely populated
. Rice,
grown during Kharif (the monsoon season from June to October), and wheat,
grown during Rabbi (November to April), are the two most dominant crops
. In the past, the dominant irrigation method in Uttar
Pradesh has been via canal, much of which is supplied by the Ganges and
Yamuna rivers. However according to , canal irrigation
has declined by approximately 40 % during the last 4 decades, with a
13-fold increase in irrigation by tube wells.
The following sections comprise a description of a case study in which data
were collected through a series of semi-structured interviews. This was
carried out in a data-scarce region, with the collected information, through
mapping and statistical analyses, used to gain a better insight into regional
irrigation practices and the motivations of users. Based on irrigation water
source information contained within the statistical abstract of Uttar Pradesh
, two
districts – Jalaun, the highest user of surface water in the state, and
Sitapur, one of the highest irrigators in UP using groundwater – were chosen
for investigation. The highest was not considered a viable option due to
logistical constraints. A map of the study area, along with the interview
locations, is presented in Fig. 1.
Map of the study region including the locations of the field
interviews carried out.
Jalaun
Jalaun is located in the south central region of Uttar Pradesh, and is
bounded by the Yamuna River to the north and the Betwa River to the east,
covering an area of 4565 km2. It is home to over 1.5 million people
. Jalaun
receives an average annual rainfall of 811 mm, about 70 % of which falls
during the monsoon season of June to August .
Approximately 139 000 ha of land is irrigated per year using canal water,
making it one of the highest users of this resource in the state. While canal
water is generally applied through gravity flow along irrigation channels,
groundwater is abstracted predominantly via diesel pumps. It was noted that
there were approximately 10 421 diesel pump sets recorded in 2012 in the
district, with electricity powering just 356 units. As there is no
restriction on the number of wells that can be drilled or on pump
specifications, it is likely that there are many more diesel pumps in use.
The main crop grown in the district is wheat, with a total cropped area of
146 307 ha. Jalaun is classed as one of Uttar Pradesh's 35 more deprived
districts and is known to be one of the
more drought-prone regions of the state .
Sitapur
Sitapur, also considered one of Uttar Pradesh's less developed districts
, is located to the north of the state
capital, Lucknow, and has a population of approximately 4.5 million
. The average
rainfall in Sitapur is 903 mm, 66 % of which falls during the monsoon
months . On a district scale it is one of the largest
irrigators in Uttar Pradesh and supplies its 374 445 ha of irrigated land
largely using groundwater, with canal water only accounting for 17 914 ha.
Using electricity for groundwater abstraction in this region is rare, and
farmers predominantly use diesel pumps. As with Jalaun, lack of regulations
and difficulty in counting wells indicate a larger number of pumps in use
across the district. The main crops grown are rice, wheat, and sugarcane,
with most farmers carrying out a rice–wheat rotation on their land.
Interview design
The main focus of this study was to investigate farmer irrigation behaviour
in the Ganges Basin of northern India and to collect relevant quantitative as
well as qualitative information, all of which may be used for informing and
driving models. Following a detailed literature review, a methodology
employing semi-structured interviews was designed and a topic guide was
organised around the following themes:
farm and crop information (farm size, soil type, crop type, crop calendar, yield),
irrigation practices (number of irrigation events, irrigation volume, irrigation methods),
water source (water source reliability, irrigation cost, irrigation method, influences on irrigation, presence of water market, power source, constraints),
other (perceptions of challenges faced, potential rationales, changes in water availability, livelihood sustainability).
The topic guide was designed to collect relevant information with as much
flexibility as possible, allowing the interview to be shaped by the
interviewees' own understandings, the interests of the researcher, and
any unexpected themes that emerge. The topic guide used during the interviews
is presented (see Supplement). While the contents of the topic
guide are presented as questions, they were treated as prompts. This allows
the conversation to progress with as much flexibility as possible while still
keeping the interviews relevant to the research questions. However, while the
aim is to highlight new data through open-ended questions and a fluid
interview structure, some direct questions are included, for example relating
to farm size or the depth of water.
Sampling
As described, fieldwork was undertaken in two districts, chosen based on
their irrigation water source, with Jalaun the highest user of canal water in
Uttar Pradesh, and Sitapur irrigating almost exclusively through groundwater.
This initial targeted approach was deemed necessary to capture a
representative sample of water users, including both conjunctive and
groundwater-only users, producing as rich a data set as possible, whilst also
considering logistics and other resource constraints, such as time and
finances. Following the identification of the fieldwork regions, a list of
villages in each district was obtained
. These were
randomised with 15 villages picked as data collection points. Between 3 and 5
interviews were conducted in each of the attended villages, with 50 farmers
interviewed in each of the two districts. After approaching a potential
interviewee, inclusion and exclusion criteria were used to determine whether
or not the participant was eligible. Interview participant inclusion criteria
were a farmer who (1) who grew wheat and/or rice, (2) irrigated their crops
rather than depended on rain only, (3) had land within approximately 5 km
from the village centre, and (4) had the authority to answer the questions.
Participants were excluded if they were (1) too young or did not have the
authority to answer the questions, or (2) if their land was too close to a
previously interviewed farmer.
Data collection – conducting the interview
The fieldwork team consisted of the researcher, a translator, and a driver.
All interviews were conducted through a translator. Potential interviewees
were approached when seen in the field. No “gatekeeper”, such as a village
head or government official, was approached in order to facilitate meetings
with participants as it was unnecessary and could have impeded the data
collection and potentially impacted on the information received. Once a
potential participant was identified, he was approached by the researcher and
translator, who made an introduction, described the project, and asked if
they would be willing to answer questions. It was made clear that the
interviewee was under no obligation to take part if they did not wish to do so and
that all information collected would be treated in the strictest confidence.
It was also highlighted that if participants had any questions they were free
to ask. During the interview the participant was given as much opportunity as
possible to expand on topics that were of most interest to them. All
interviews were recorded using a dictaphone and field notes, with GPS
readings of pertinent locations and photographs taken throughout.
Differences in irrigation practices between the districts of
Sitapur and Jalaun, Uttar Pradesh, northern India. The boxplots represent variability
between farmers in each district. The boxes represent the 25 to 75 percentiles;
the whiskers represent 1.5 times the interquartile range (IQR). The P values give the chance of equal mean obtained from Student's t test.
Data processing and analyses
Once data collection was completed, all interviews were transcribed verbatim
and uploaded to the qualitative data analysis package RQDA
to allow for thematic analyses of the data. During the interviews and while
reading the transcripts, a number of themes emerged as being important, for
example the cost of irrigation, the reliability of their water source, and
the importance of conjunctive surface and groundwater use. These themes were
coded to different sections from the transcribed interviews, allowing not
only a commonality of themes to emerge across interviews but also unique
perspectives to be highlighted. A significant portion of the data collected
was quantitative. This allowed for statistical analyses of variables to
assess differences in irrigation practices between and within the two
districts. These included the volume of water applied (m3 ha-1), the
volume of water required to produce 1 t of wheat (m3 t-1), the cost
of wheat irrigation during the growing season (r ha-1), the crop yield
in tonnes per hectare (t ha-1), the farm area (ha), and the cost of
irrigation water per cubic metre (rupees m-3). The cost of water in
cubic metres was calculated by taking into account the cost of irrigation and
the volume of water applied per hectare. The case study analyses focus on
wheat. While both wheat and rice are grown in Sitapur, rice is not commonly
cultivated in Jalaun, with only 1 farmer out of 50 interviewed growing the
crop. The results of the analyses can be found in Figs. 2 to 6, with a
description of results below.
Case study – discussion and resultsQuantitative results
The results presented in Fig. 2 and in the maps in Figs. 3 and 4 show there
is a significant variance in the irrigation practices of farmers in Jalaun
and Sitapur. This can be seen in the volumes of irrigation water used
(Fig. 2a), with farmers in Sitapur applying on average 1555 m3 ha-1
more than farmers in Jalaun. This is also reflected in the overall cost of
irrigation, with farmers in Sitapur paying on average over
7000 rupees ha-1 season-1 more to irrigate their wheat crop than
their counterparts in Jalaun (Fig. 2b). This is despite the basic cost of
water per cubic metre being largely the same: 3.58 r m-3 in Sitapur
and 3.84 r m-3 in Jalaun (Fig. 2f).
Sitapur is by area one of the largest irrigators in Uttar Pradesh and for
the most part relies on water from the underlying aquifers. The primary
method of abstraction is by diesel pump, which, although reliable and
versatile, is expensive, with farmers in Sitapur paying on average
12 782 r ha-1 season-1 to irrigate their wheat crop. Jalaun,
however, is one of the highest irrigators using canal water in Uttar Pradesh,
with the majority of farmers interviewed (33/50) making use of the resource,
often in conjunction with groundwater. This provides a cheap, and sometimes
free, source of irrigation water (Figs. 2c and 3). In addition, farmers in
Sitapur produce smaller yields than farmers in Jalaun, almost 2 t ha-1
less (Fig. 2d). As can be seen in Fig. 2b and in Fig. 4, farmers in Sitapur
apply 1017 m3 of irrigation water, with those in Jalaun using only
396 m3 to produce a tonne of wheat.
Spatial variations in the annual price paid for the irrigation of wheat by farmers
in Jalaun and Sitapur, Uttar Pradesh, northern India.
Spatial variations in the volume of water applied per tonne of wheat produced
in Jalaun and Sitapur, Uttar Pradesh, northern India.
When comparing tube well users only in both districts, further differences
emerged. In terms of production efficiency, farmers in Sitapur required an
average of 1017 m3 of irrigation water per tonne of wheat produced, while
their counterparts in Jalaun applied 800 m3 less (Fig. 5b). When only
tube well users were taken into account, the price paid per cubic metre of
irrigation water was found to be very different. Farmers in Sitapur paid an
average of 3.58 r m-3, whereas farmers in Jalaun pay significantly
more: an average of 8.71 r m-3 (Fig. 5d). The fact that farmers
applied less irrigation water in Jalaun, however (Fig. 5a), is reflected in
the overall cost of irrigation by both groups (Fig. 5c). Farmers in Sitapur
paid an average of 1167 r ha-1 more to irrigate their wheat crops
despite the fact that the cost per cubic metre of water is less.
Differences in irrigation practices between tube well only users in
the districts of Sitapur and Jalaun, northern India. The boxplots represent variability
between farmers in each district. The boxes represent the 25 to 75 percentiles;
the whiskers represent 1.5 times the interquartile range (IQR). The P values give the chance of equal mean obtained from Student's t test.
Differences in irrigation practices between tube well and canal users
and canal-only users in the district of Jalaun, northern India. The boxplots represent
variability between farmers in each district. The boxes represent the 25 to 75
percentiles; the whiskers represent 1.5 times the interquartile range (IQR). The P values give the chance of equal mean obtained from Student's t test.
In Jalaun many of the interview participants had access to both tube wells
and the cheaper but less reliable Irrigation Department-supplied canal
water. Conjunctive use of surface and groundwater is often promoted as a
realistic option to solving groundwater overdraft caused by irrigation
, and developing an understanding of farmer
behaviour in this type of environment is important when formulating
solutions. To investigate irrigation behaviour between farmers who have a
choice in their water source (canal and tube well) and those who do not
(tube well only), a comparison of the data collected within the district of
Jalaun was undertaken, the results of which can be seen in Fig. 6. In terms
of the volume of irrigation water applied, there was a statistically
significant difference between both groups (Fig. 6a), with farmers who had
canal access applying over 1722 m3 of water more than those who relied
on tube wells only. While more water was used by farmers who have access to
canals to produce 1 t of wheat (Fig. 6b), the difference between the
two groups was not found to be statistically significant. The cost of
irrigation water however, per cubic metre, was found to be significantly different
between both users (Fig. 6d); canal users paid an average of
2.09 r m-3, whereas farmers who use tube wells pay an average of
8.71 r m-3. As can be seen in Fig. 6c, in terms of the overall price
paid for irrigation by both groups, farmers who had access to canal water
were applying more, and also paid 7805 rupees ha-1 season-1 less to
irrigate their wheat.
Differences between wheat irrigation volumes reported by farmers (boxplots)
and modelled irrigation water requirements (time series). The mean modelled irrigation
requirements from 1948 to 2012 are also shown (stars on boxplots) to aid comparison with
2013 reported information. The boxes represent the 25 to 75 percentiles; the whiskers
represent 1.5 times the interquartile range (IQR). Circles represent outliers: values
which exceed 1.5 times the IQR.
The data reported in this section provide an example of the type of
information that can be collected using this methodology. While it reveals a
considerable amount of detail on the irrigation behaviour of farmers in the
region, it is envisaged that this information can be further utilised,
particularly in the set-up and driving of hydro-economic and groundwater
models of the region.
Qualitative results
The most commonly reported theme during the interviews was poor water
availability. While no exact measurements were taken by respondents, a
significant proportion in both districts reported that they had noticed water
levels were falling. Among groundwater users, this was predominantly
perceived to be as a result of overuse by other farmers and poor rainfall:
Farmer 1, Dafrapur, Sitapur. Translator: “Maybe because of many people
extracting the water”.
In some cases farmers reported that while they usually got water eventually,
it was often too late to meet their crop water needs:
Farmer 1, Kishun Kheara, Sitapur. Translator: “Always get but not always on
time”.
This problem is exacerbated as water levels decrease during the post-monsoon
season, causing farmers to rely on deeper tube wells, which are fewer in
number, leading to a delay in access. A proportion of farmers in Sitapur
highlighted that they had no issues with water supply when they had reliable
access to a deep well or had land in an area with a high water table.
Farmers in Sitapur are dependent on groundwater; however, many in Jalaun have
access to both canal and surface water. Canals, while beneficial,
particularly in terms of affordability, were perceived as unreliable, with
the Irrigation Department-supplied water often arriving late or early for
irrigation. This sometimes forced farmers to turn to the more expensive
groundwater where available, to ensure their crops were irrigated. Indeed in
Jalaun, the lack of access to a reliable water source was deemed to be the
main reason for farmers' not growing rice, despite many saying that their
soil was suitable:
Farmer 6, Barha Jalaun. Translator: “So he is telling me that he generally
grows wheat, ...and don't grow rice because of lack of water, so soil is good
for rice, but because of lack of water, they generally don't grow”.
According to farmers in both districts, the lack of a dependable electricity
supply was perceived as a significant barrier to accessing a sustainable
source of water for irrigation. Electric submersible pumps allow for deeper
water abstraction and are generally considered to be a cheaper option for
farmers than the common diesel pump. Indeed the introduction or improving of
electricity in an area was seen as an obvious solution to water issues in
both districts. The fact that this could lead to further reductions in water
levels was not mentioned by participants, highlighting the often myopic
nature of farmers. Interviewees also singled out the Government for
criticism, blaming them for poor infrastructure such as badly maintained
wells, the poor electricity supply, and corruption:
Farmer 2, Gulriha Sitapur. Translator: “He is saying government is not
doing what they want here”.
Farmer 1, Mania Sitapur. Interviewer: “And what do you think are the
biggest problems you face at the moment?” Translator: “Lack of fertiliser,
lack of water, lack of electricity everything what they say first is
everything! Major thing is [government] corruption”.
In Jalaun however, a number of farmers highlighted the benefits of some
government policies, particularly that of free, or cheap, canal water. While
welcome, farmers saw this practice as a means for local politicians to secure
votes. As the participants were given freedom to elaborate where they saw
fit, additional information emerged in a number of interviews. In Sitapur
this included a system whereby access to water was shared between farmers who
owned their own wells. This was outside the water market, a common method of
irrigation water access in both districts, and was more prevalent in parts of
Sitapur where farms were fragmented. The system allowed farmers to use tube
wells owned by farmers neighbouring more distant pieces of their land, in
return allowing their own well(s) to be used by others. Farmers would move
their own pumps around to different wells as needed. A lack of labour was
also highlighted as an issue for farmers. This emerged as an important reason
why farmers did not use sprinklers for much of their wheat crop; while most
were aware of the potential benefits, particularly in Jalaun, implementation
was curtailed by the lack of available labour. Climate, particularly the lack
of rainfall, emerged as a challenge for farmers; however a number of
interviewees in Sitapur spoke of the onset of “Westerlies”: a drying wind
which had a dramatic effect on crop water requirements:
Farmer 3, Lilsi, Sitapur. Translator: “Because of Westerlies, the wind can
carry more and more moisture from the soil”.
Poor neighbour relations were highlighted as a potential problem in accessing
water when needed in both districts but was more prevalent in Jalaun,
particularly in terms of access to canal supply, with farmers further down
the canal receiving less water. Interviewees also spoke of the damming of
canals by farmers upstream as a problem in receiving water on time:
Farmer 4, Kusmra Bavani Jalaun. Translator: “...there is a conflict between
the villages because the water distribution and what happens is that the
upstream villagers they dam the canal as we have seen, and they stop the water
for 2 or 3 days”.
The perception among farmers in both districts was that irrigation water was
not cheap. However, this did not appear to change their attitude to
irrigation as a reduction in water could lead to a reduction in crop yield.
It appeared that farmers were being as efficient as they could be, given the
available resources.
Comparison with modelled irrigation requirement results
Crop water requirements can be estimated through various algorithms, for
example Hargreaves–Samani or Penman–Monteith
. These approaches are extremely useful as they can
provide results without the need for field level measurements. It is
important, however, to compare the modelled outputs to field data where
possible as results can vary considerably. The reported volume of irrigation
water applied by farmers to their wheat crop is compared to values obtained
through modelling of requirements using Hargreaves–Samani'
potential evapotranspiration method and the
Terrestrial Hydrology Research Group at Princeton University's global
meteorological forcing data set , the best data set for
the region available for this study. The crop coefficients used in the
calculation are provided by , which are estimated through
field experiments in northern India. These data allowed for the modelling of
wheat irrigation requirements from 1948 to 2012. The results are then
compared with irrigation volumes reported by farmers during fieldwork
undertaken in 2013. All results are presented in Fig. 7. The model used the
best available data set for the region; while the results do not overlap with
reported values, the difference between modelled and information obtained in
the field is clear.
The mean value reported by farmers in Sitapur is 4050 m3 ha-1 of
irrigation water applied during the wheat season. This is
368 m3 ha-1 below the modelled 2012 result of
4418 m3 ha-1. The difference in Jalaun is more significant, with a
mean reported values of 2283, 2253 m3 ha-1 less than the modelled
result of 4536 m3 ha-1. The median reported values for both
districts is also significantly lower than the modelled result (Jalaun:
1390 m3 ha-1; Sitapur: 3800 m3 ha-1), highlighting that
the majority of farmers apply less water than would be predicted through
modelling, showing the importance of using field-collected information to
address model uncertainties. The variance found between the districts is
likely due to differences in soil type with a higher proportion of sandy soil
in Sitapur ,
requiring larger amounts of irrigation to maintain soil moisture. Rainfall
rates are largely similar across both districts. The data reported in this
section provide an example of the type of information that can be collected
using this methodology. While it reveals a considerable amount of detail on
irrigation behaviours, it is envisaged that this information can be further
utilised, particularly in the set-up and driving of hydro-economic and
groundwater models of the region.
Opportunities and limitations of semi-structured interviews
The lack of reliable quantitative and qualitative information is a major
barrier in developing realistic water security options. In data-scarce
regions of the world, information is typically downscaled from larger
regional data sets; however this ignores the often significant spatial
variability that exists on a finer scale. The use of qualitative as well as
quantitative information is essential in identifying the drivers behind water
use practices; however the collection of this information is often expensive
and time-consuming. Semi-structured interviews provide a means of developing
information-rich data sets in a time- and resource-efficient manner. Direct
contact with water users and the opportunity to allow participants to expand
on the issues of most importance to them provides a unique opportunity to
develop an understanding of the human–water interface in a given location.
Despite the usefulness of semi-structured interviews, we identify some
limitations in both the data collected and the approach used. The information
collected, while useful for informing large-scale models, is most applicable
to the scale at which it was collected, which ideally should coincide with a
scale at which decisions on policy can be made and implemented: in this case
district level. The type of data collected, both quantitative and
qualitative, is useful for driving models, through numerical inputs and in
setting rules – for example who has access to which water source and when.
As can be seen in Sect. 4.3, the differences between modelled outputs and
collected field data can be significant. Incorporating field level
information where possible is an important consideration for modellers in
order to highlight bias and uncertainty. This also applies to water users and
water managers, as the approach allows for more realistic conclusions to be
drawn from model outputs. In the case studies, interviews took place from
September to November. This snapshot of the farming year in Uttar Pradesh is
during a time of peak water availability, as it is following the monsoon
season. It is possible that this influenced farmer responses. In addition,
out of 105 farmers approached, only 5 declined to be interviewed. While this
high participatory rate made fieldwork straightforward, it highlights a
potential propensity for interviewees to please the interviewers, providing
statements indicative of social desirability response bias
, which may be reflected in the collected information.
While social desirability response bias has been observed in Indian culture,
it is not culturally specific (Hebert et al., 1998) and should be considered
at all stages of data collection and analysis. In the case study reported
above, interviews required the use of translators. Shortfalls associated with
using a translator(s) are described in ; however to limit
the potential for discrepancy, training should be provided prior to
fieldwork. It is also important to remember that in their environment the
interviewee is the expert and should be treated as such. This also helps
break down some of the barriers which may exist when a researcher and
participant are from different cultures. It is important to take these
factors into consideration at all stages of the research, including
subsequent analyses. While the case study sample size (n=50 per district)
is small relative to the population (Sitapur = 623 000 farms,
Jalaun = 253 000 farms;
), we are
confident that it presents a good representation of farming practices across
the district as a whole. Verification of the objective accuracy of
self-reported data is also an important consideration. Reported information
can be triangulated with, if available, socio-economic data; outputs from
other models; or, ideally, field level monitoring of water levels,
abstraction rates, and surface water availability. While validation of
collected data through objective measures is a necessary step in data
collection, it is outside the scope of this paper. To address these
shortcomings, further fieldwork should be undertaken, focusing on different
regions of Uttar Pradesh during more water-scarce times of the year and,
importantly, gaining objective measures of the data reported herein, i.e. via
direct observation and metering of the phenomena. This would help in
quantifying the differences between modelled, reported, and collected
information, leading to more accurate hydrological model development and
outputs, allowing for more realistic predictions to changes in boundary
conditions, including those from climate change.
Conclusions
Limits in our current understanding of the human–water interface are a major
constraint in developing options for future water security. One of the major
barriers in developing this understanding is a lack of suitable qualitative
and quantitative data. In this paper we present a methodology to facilitate
the collection of information for hydrological and engineering purposes in
data-scarce regions through semi-structured interviews. We use this
methodology to investigate farmer irrigation practices in the Ganges Basin of
northern India, collecting information from 100 farmers across two districts.
Information was obtained on topics such as irrigation water volumes, the cost
of irrigation, water source, and the drivers behind these practices.
Statistical analysis of the data, along with some data visualisation, is
presented. Aspects such as a significant variability in water use practices,
as well as insights into farmer behaviours and their environment, are
highlighted. Semi-structured interviews provide a useful platform for the
collection of qualitative and quantitative information simultaneously. This
has clear benefits, including directly linking behaviours and their drivers
to reported numerical values. Semi-structured interviews facilitate the
collection of detailed information quickly, easily, and relatively cost
effectively while indicating themes which may not have been obvious
beforehand, as well as highlighting aspects of the study which may no longer
be relevant. The data collected also lend themselves to hydrological and
hydro-economic modelling, as well as providing more realistic representations
of user behaviour: an essential component in model development. While some
limitations do exist, we are confident that this approach can be employed by
natural scientists as an effective and efficient method of collecting both
qualitative and quantitative hydrological information for the assessment of
drivers, behaviours, and their outcomes in a data-scarce region.
The Supplement related to this article is available online at doi:10.5194/hess-20-1911-2016-supplement.
Acknowledgements
The authors would like to acknowledge the support of the NERC Changing Water
Cycle (South Asia) project Hydrometeorological feedbacks and changes in water
storage and fluxes in Northern India (grant number NE/I022558/1). The authors
would also like to thank all reviewers for their constructive comments, which
enabled us to significantly improve this paper. Edited by: A. D. Reeves
References
Allen, R. G., Pereira, L. S., Raes, D., and Smith, M.: Crop
evapotranspiration
– Guidelines for computing crop water requirements – FAO Irrigation and
Drainage Paper 56, Tech. rep., Food and Agriculture Organization of the
United Nations, 1998.
Allen, R. G., I.Elliott, W. R. L., Howell, T. A., Itensfisu, D., Jensen,
M. E.,
and Snyder L., R.: The ASCE Standardized Reference Evapotranspiration
Equation, Tech. rep., The Task Committee on Standardization of Reference
Evapotranspiration of the Environmental and Water Resources Institute,
Reston, VA, USA, 2005.Amarasinghe, U. A., Mccornick, P., and Shah, T.: Projections of irrigation
water demand in India: What do recent trends suggest?, International Journal of River Basin Management, 7, 157–166,
10.1080/15715124.2009.9635378, 2009.Avtar, R., Kumar, P., Singh, C. K., and Mukherjee, S.: A Comparative Study
on
Hydrogeochemistry of Ken and Betwa Rivers of Bundelkhand Using Statistical
Approach, Water Quality, Exposure and Health, 2, 169–179,
10.1007/s12403-010-0035-2, 2011.Barnes, A. P., McCalman, H., Buckingham, S., and Thomson, S.: Farmer
decision-making and risk perceptions towards outwintering cattle, J.
Environ. Manage., 129, 9–17, 10.1016/j.jenvman.2013.05.026, 2013.Barriball, K. L. and While, A.: Collecting data using a semi-structured
interview: a discussion paper, J. Adv. Nurs., 19, 328–335,
10.1111/j.1365-2648.1994.tb01088.x, 1994.Burnard, P., Gill, P., Stewart, K., Treasure, E., and Chadwick, B.:
Analysing
and presenting qualitative data, Brit. Dent. J., 204, 429–432,
10.1038/sj.bdj.2008.292,
2008.Calheiros, D., Seidl, A., and Ferreira, C.: Participatory research methods
in
environmental science : local and scientific knowledge of a limnological
phenomenon in the Pantanal wetland of Brazil, J. Appl. Ecol.,
37, 684–696, 10.1046/j.1365-2664.2000.00524.x, 2000.
Chauhan, P.: Ground Water Brochure of Sultanpur District, Uttar Pradesh,
Tech. Rep. May 2007, 2007.Chowdhury, N. T.: Irrigation Institutions of Bangladesh: Some Lessons.
Problems, Perspectives and Challenges of Agricultural Water Managment,
http://www.intechopen.com/books/problems-perspectives-and-challenges-ofagricultural-water-management/irrigation-institutions-of-bangladesh-some-lessons (last access: 19 November 2015), 2012.Collins, M., Shattell, M., and Thomas, S. P.: Problematic interviewee
behaviors in qualitative research, Western J. Nurs. Res., 27,
188–99; discussion 200–9, 10.1177/0193945904268068, 2005.Creswell, J. W.: Research Design. Qualitative, Quantitative, and Mixed
Methods
Approaches, SAGE Publications Ltd, 3rd edn.,
10.1002/1521-3773(20010316)40:6<9823::AID-ANIE9823>3.3.CO;2-C, 2009.Döll, P. and Siebert, S.: Global modeling of irrigation water
requirements, Water Resour. Res., 38, 1–10, 10.1029/2001WR000355,
2002.Döll, P., Müller Schmied, H., Schuh, C., Portmann, F. T., and
Eicker, A.: Global-scale assessment of groundwater depletion and related
groundwater abstractions: Combining hydrological modeling with information
from well observations and GRACE satellites, Water Resour. Res., 50,
5698–5720, 10.1002/2014WR015595, 2014.Ellis, L. M. and Chen, E. C.: Negotiating identity development among
undocumented immigrant college students: a grounded theory study, J. Couns. Psychol.,
60, 251–64, 10.1037/a0031350, 2013.Fallon, P.: Life events; Their role in onset and relapse in psychosis,
research utilizing semi-structured interview methods, J. Psychiatr. Ment. Hlt., 15, 386–392,
10.1111/j.1365-2850.2007.01244.x, 2008.Famiglietti, J. S.: The global groundwater crisis, Nature Climate Change,
4,
945–948, 10.1038/nclimate2425, 2014.Fenicia, F., Kavetski, D., and Savenije, H. H. G.: Elements of a flexible
approach for conceptual hydrological modeling: 1. Motivation and theoretical
development, Water Resour. Res., 47,
10.1029/2010WR010174, 2011.
Flick, U.: An introduction to qualitative research, vol. 9781446267,
London,
Sage Publications, London, 5 edn., 2014.Gao, H., Birkett, C., and Lettenmaier, D. P.: Global monitoring of large
reservoir storage from satellite remote sensing, Water Resour. Res.,
48, 1–12, 10.1029/2012WR012063, 2012.GEWEX: GEWEX Plans for 2013 and Beyond – GEWEX Science Questions (Version
1),
Tech. Rep. GEWEX Document Series No. 2012-2, GEWEX,
http://www.gewex.org/gewex-content/uploads/2015/02/GEWEX_Science_Questions_final.pdf (last access: 19 November 2015), 2012.Gibson, C.: Semi-structured and unstructured interviewing – a comparison of
methodologies in research with patients following discharge from an acute
psychiatric hospital, J. Psychiatr. Ment. Hlt., 5,
469–477, 10.1046/j.1365-2850.1998.560469.x, 1998.
Government of India: Census 2011 – Provisional Population Totals, Tech.
rep., Ministry of Home Affairs, 2011.Guarte, J. M. and Barrios, E. B.: Computation Estimation Under Purposive
Sampling, Communications in Statistics – Simulation and Computation, 0918,
277–284, 10.1080/03610910600591610, 2006.
Hagirath, B., Kumar, C., Nauriyal, D. K., Nayak, N. C., Prasad, P. M.,
Rajgopalan, P., Mishra, P., Trivedi, P. L., Agrawal, A., Singh, S. P.,
Sharma, S., Mazumder, T. N., Upadhyay, V. B., Sharma, V., and Tare, V.:
Trends in Agriculture and Agricultural Practices in Ganga Basin. An
Overview, Tech. rep., Ganga River Basin Managment Plan, 2011.Hargreaves, G. H. and Samani, Z. A.: Reference Crop Evapotranspiration From
Ambient Air Temperature, Paper – American Society of Agricultural
Engineers, 96–99,
http://www.scopus.com/inward/record.url?eid=2-s2.0-0022284023&partnerID=40&md5=8dbbb2b21fcf0b47f0fb6d1e5dcefa02 (last access: 19 November 2015), 1985.Harou, J. J. and Lund, J. R.: Ending groundwater overdraft in
hydrologic-economic systems, Hydrogeol. J., 16, 1039–1055,
10.1007/s10040-008-0300-7,
2008.Hebert, J. R., Gupta, P. C., Bhonsle, R. B., Sinor, P. N., Mehta, H., and
Mehta, F. S.: Development and testing of a quantitative food frequency
questionnaire for use in Gujarat, India, Public Health Nutr., 2,
39–50, 10.1017/S1368980099000051, 1998.Huang, R.: RQDA: R-based Qualitative Data Analysis,
http://rqda.r-forge.r-project.org/ (last access: 8 September 2015), 2014.
ICRISAT-ICAR-IRRI: Village Dynamics in South Asia (VDSA), District Level
Database Documentation, Tech. rep., ICRISAT-ICAR-IRRI Collaborative Research
Project, 2012.Indian Council of Agricultural Research: District Profile: Uttar Pradesh,
Tech. rep., Indian Council of Agricultural Reseasrch – Zonal Project
Directorate Kanpur,
http://zpdk.org.in/sites/default/files/districtprofile(2-2-10).pdf (last access: 14 March 2016),
2010.Jewitt, S. and Baker, K.: The Green Revolution re-assessed: Insider
perspectives on agrarian change in Bulandshahr District, Western Uttar
Pradesh, India, Geoforum, 38, 73–89, 10.1016/j.geoforum.2006.06.002, 2007.Kapborga, I. and Berterö, C.: Using an interpreter in qualitative
interviews: does it threaten validity?, Nurs. Inq., 9, 52–6,
http://www.ncbi.nlm.nih.gov/pubmed/12164715, 2002.Mason, J.: Qualitative Researching, SAGE Publications Ltd, London, second
edn., 10.1016/S0143-6228(97)90005-9, 2002.McKenney, M. S. and Rosenberg, N. J.: Sensitivity of some potential
evapotranspiration estimation methods to climate change, Agr.
Forest Meteorol., 64, 81–110, 10.1016/0168-1923(95)02240-X, 1993.Ministry of Panchayati Raj: Backward Regions Grant Fund Programme,
http://www.panchayat.gov.in/details-of-brgf-districts (last access: 29 October 2015), 2014.Montanari, A.: Debates – Perspectives on sociohydrology: Introduction,
Water Resour. Res., 51, 2–31,
10.1002/2015WR017430, 2015.Mottram, A.: “They are marvellous with you whilst you are in but the
aftercare
is rubbish”: a grounded theory study of patients' and their carers'
experiences after discharge following day surgery, J. Clin.
Nurs., 20, 3143–51, 10.1111/j.1365-2702.2011.03763.x, 2011.Mueller, N. D., Gerber, J. S., Johnston, M., Ray, D. K., Ramankutty, N., and
Foley, J. A.: Closing yield gaps through nutrient and water management,
Nature, 490, 254–257, 10.1038/nature11420, 2012.Nazemi, A. and Wheater, H. S.: On inclusion of water resource management in
Earth system models – Part 1: Problem definition and representation of water
demand, Hydrol. Earth Syst. Sci., 19, 33–61, 10.5194/hess-19-33-2015,
2015a.Nazemi, A. and Wheater, H. S.: On inclusion of water resource management in
Earth system models – Part 2: Representation of water supply and allocation
and opportunities for improved modeling, Hydrol. Earth Syst. Sci., 19,
63–90, 10.5194/hess-19-63-2015, 2015b.Portmann, F. T., Siebert, S., and Döll, P.: MIRCA2000 – Global
monthly
irrigated and rainfed crop areas around the year 2000: A new high-resolution
data set for agricultural and hydrological modeling, Global Biogeochem.
Cy., 24, 1–24, 10.1029/2008GB003435, 2010.Rabionet, S. E.: How I Learned to Design and Conduct Semi-structured
Interviews : An Ongoing and Continuous Journey, The Qualitative Report, 16,
563–566, http://www.nova.edu/ssss/QR/QR16-2/rabionet.pdf (last access: 16 January 2015),
2011.Rodell, M., Velicogna, I., and Famiglietti, J. S.: Satellite-based estimates
of groundwater depletion in India., Nature, 460, 999–1002,
10.1038/nature08238, 2009.Shah, T., Bhatt, S., Shah, R., and Talati, J.: Groundwater governance
through
electricity supply management: Assessing an innovative intervention in
Gujarat, western India, Agr. Water Manage., 95, 1233–1242,
10.1016/j.agwat.2008.04.006, 2008.Sheffield, J., Goteti, G., and Wood, E. F.: Development of a 50-year
high-resolution global dataset of meteorological forcings for land surface
modeling, J. Climate, 19, 3088–3111, 10.1175/JCLI3790.1,
2006.Siebert, S. and Döll, P.: Quantifying blue and green virtual water
contents in global crop production as well as potential production losses
without irrigation, J. Hydrol., 384, 198–217,
10.1016/j.jhydrol.2009.07.031, 2010.Singh, N. J., Kudrat, M., Jain, K., and Pandey, K.: Cropping pattern of
Uttar
Pradesh using IRS-P6 (AWiFS) data, Int. J. Remote Sens.,
32, 4511–4526, 10.1080/01431161.2010.489061, 2011.Singh, R.: Environmental consequences of agricultural development: a case
study from the Green Revolution state of Haryana, India, Agr.
Ecosyst. Environ., 82, 97–103,
10.1016/S0167-8809(00)00219-X, 2000.Tiwari, V. M., Wahr, J., and Swenson, S.: Dwindling groundwater resources in
northern India, from satellite gravity observations, Geophys. Res.
Lett., 36, L18401, 10.1029/2009GL039401, 2009.Uttar Pradesh State Planning Institute: Statistical Abstract of Uttar
Pradesh 2012, http://updes.up.nic.in, Tech. rep., Economics and Statistics
Division, State Planning Institute, Uttar Pradesh, 2012.Voss, K. A., Famiglietti, J. S., Lo, M., De Linage, C., Rodell, M., and
Swenson, S. C.: Groundwater depletion in the Middle East from GRACE with
implications for transboundary water management in the
Tigris-Euphrates-Western Iran region, Water Resour. Res., 49,
904–914, 10.1002/wrcr.20078, 2013.Wada, Y., Van Beek, L. P. H., Van Kempen, C. M., Reckman, J. W. T. M.,
Vasak, S., and Bierkens, M. F. P.: Global depletion of groundwater
resources, Geophys. Res. Lett., 37, 1–5,
10.1029/2010GL044571, 2010.Winsemius, H. C., Schaefli, B., Montanari, A., and Savenije, H. H. G.: On
the
calibration of hydrological models in ungauged basins: A framework for
integrating hard and soft hydrological information, Water Resour.
Res., 45, W12422, 10.1029/2009wr007706, 2009.
Wisser, D., Frolking, S., Douglas, E. M., Fekete, B. M.,
Vörösmarty, C. J., and Schumann, A. H.: Global irrigation water
demand: Variability and uncertainties arising from agricultural and climate
data sets, Geophys. Res. Lett., 35, 1–5,
10.1029/2008GL035296, 2008.