Original Research Article

Outlook on Agriculture
1–9
Irrigation Advisory Services: Farmers ª The Author(s) 2021
Article reuse guidelines:

preferences and willingness to pay for [Link]/journals-permissions

DOI: 10.1177/00307270211002848
[Link]/home/oag
innovation

Filiberto Altobelli1 , Anna Dalla Marta2 , Marius Heinen3,

Claire Jacobs3, Elisa Giampietri4, Marco Mancini2,
Orlando Cimino1, Samuele Trestini4 , Remco Kranendonk3,
Andre Chanzy5, Marta Debolini5, Dominique Courault5,
Ewa Kanecka-Geszke6, Wiesława Kasperska-Wolowicz6 ,
Francisco José Blanco-Velázquez7, Marı́a Anaya-Romero7,
Marlene De Witt8 , Willem de Clercq8,
Antonio Diaz Espejo9, Virginia Hernandez-Santana9,
Francesco Caracciolo10 and Teresa Del Giudice10

Abstract
Irrigation Advisory Services (IAS) are powerful management instruments aiming to achieve the best efficiency in irrigation
water use. So far the literature on farmers’ preferences for a specific scheme design of IAS’ characteristics and the related
willingness to pay (WTP) is scant. This study provides evidence on farmers’ preference towards six attributes related to
the IAS configuration by using a hypothetical choice experiment. Data were collected from an original survey among 108
farmers from Spain, The Netherlands, Italy, Poland and South Africa. Moreover, we investigated the interplay between
these preferences and the individual risk attitude (elicited through a lottery task) as a novel contribution. On average, the
results suggest a clear farmers’ preference, especially for receiving weather forecasts from the service and for the feature
related to water data recording; as the opposite, on average, crop water requirement seems irrelevant. Finally, we found
that farmers’ WTP for the different IAS services varies across countries and, in some cases, also according to the individual
risk attitude.

Keywords
Decision support system, irrigation advisory service, mixed logit, risk attitude

Introduction 1
CREA Research Centre for Agricultural Policies and Bioeconomy, Rome,
Italy
The irrigated areas are major contributors to the world’s 2
Department of Agriculture, Food, Environment and Forestry, University
food supply (FAO, 2014). In fact, the differences between of Florence, Florence, Italy
3
precipitation and irrigation water requirements are so pro- Wageningen Environmental Research, Wageningen, The Netherlands
4
found that irrigation management nowadays represents a Department of Land, Environment, Agriculture and Forestry, University
of Padova, Padova, Italy
priority for the sustainable and economically profitable 5
INRAE, Avignon Université, UMR EMMAH, Avignon, France
crops’ production (IDAE, 2005; Navarro-Hellı́n et al., 6
Institute of Technology and Life Sciences, Falenty, Raszyn, Poland
2015). Accordingly, innovative irrigation practices can 7
Evenor Tech (Evenor), Seville, Spain
8
enhance water efficiency, gaining an economic advantage Stellenbosch University Water Institute (SUWI), Stellenbosch
while reducing environmental burdens; at the same time, University, Western Cape, South Africa
9
Irrigation and Ecophysiology Group, Instituto de Recursos Naturales y
farmers’ efficient use of irrigation technologies alone may
Agrobiologı́a (IRNAS), Consejo Superior de Investigaciones Cientı́ficas
not conserve water. In some cases, the necessary knowl- (CSIC), Seville, Spain
edge has been provided by extension services, helping 10
Department of Agricultural Sciences, University of Naples Federico II,
farmers to adapt and implement viable solutions, thus gain- Naples, Italy
ing more benefits from irrigation technology (Levidow
Corresponding author:
et al., 2014). Site-specific extension recommendations that Filiberto Altobelli, CREA Research Centre for Agricultural Policies and
are better adapted to the needs of individual farmers and Bioeconomy, Rome, Italy.
fields, and enabled by digital technologies included on Email: [Link]@[Link]
2 Outlook on Agriculture XX(X)

Decision Support Systems (DSS) tools, could potentially we elicit the individual risk attitude through a lottery game
bring about yield and productivity improvements and water and use it to examine whether farmers’ WTP changes
saving. According to Oyinbo et al. (2019), improving the among individuals with different risk attitudes. The results
design of extension tools to enable provision of information can inform both the agricultural research and extension
on the riskiness of expected outcomes and flexibility in programmes to consider farmers’ preferences and acceler-
switching between low-risk and high-risk recommenda- ate this technology’s adoption.
tions can help farmers to make better informed decisions,
and thereby improve the uptake of extension advices and
the efficiency of extension programmes.
Materials and methods
Against this background, Irrigation Advisory Services
(IAS) are agricultural extension services and powerful Data collection
management instruments to achieve the best efficiency in The research activity was carried on under the framework
irrigation water use (Bonfante et al., 2019). These systems of the international project OPERA – Operationalizing the
are often conceptually oriented to simulate or predict crop increase of water use efficiency and resilience in irrigation
water demand, providing a set of options. Currently, many (ERA-NET Cofund WaterWorks2015 Call)
systems exist that can optimize farmers’ economic profits A structured survey was administered through direct
by efficiently simulating/forecasting both water demand interviews with a sample of 108 farmers from Italy, The
and crop yield, and estimating the percentage of agricul- Netherlands, Spain, Poland and South Africa between 2017
tural area with specific water supply need and planted area and 2019. As highlighted by Chèze et al. (2020) and Des-
constraints (Kuo et al., 2000). Others are DSS more general potović et al. (2019) a rather small sample size is common
in managing irrigation schemes (Mateos et al., 2002). Such when dealing with farmers, due to the recognized difficulty
DSS are used in some part of the world, such as in Australia to reaching them, e.g. as opposite to consumers. Respon-
(IrriSAT) and Europe, IRRISA in France, IRRISAT and dents were informed that the survey was implemented
IRRINET in Italy (D’Urso et al., 2013; Rossi et al., within an international project aiming at identifying the
2004), BEWARE (Crete, Greece), Anglia river best IAS solution.
Basin (UK). The questionnaire, written in each country’s original
The irrigation scheduling can be transferred to the user language, was pretested on a small sample of farmers
in several forms and with different delivery systems such as (N ¼ 20) before the survey (Holmes et al., 2017). The
internet services, mobile phones (Belmonte et al., 1999; de questionnaire was divided into several sections, including
Santa Olalla Sanchez, 1999), online-bulletins. These solu- questions on the choice experiment, the measure of farm-
tions offer to the irrigation DSS the possibility of providing ers’ risk attitude through a lottery-based experiment and,
simpler and more suitable mobile decision support to farm- finally, farm and farmer’s characteristics.
ers (Mannini et al., 2013).
Recent researches (Altobelli et al., 2019; Manhoudt
AGE et al. 2002) showed that farmers have a positive pre-
The choice experiment
ference and a higher related willingness to pay (WTP) for This study uses a choice experiment that allows to eliciting
an environmental certification (EC) scheme (related to farmers’ preferences for specific IAS attributes in order to
agricultural products) that guarantees an efficient water improve these tools following farmers’ needs. Moreover,
use. Nevertheless, so far farmers’ preferences for a specific although relying on hypothetical market, CE are also useful
scheme design of IAS’ characteristics and the related WTP to elicit the value that farmers assign to each investigated
are poorly understood. Indeed, the assessment of IAS product/service characteristic, which is extremely useful
related WTP is essential to determine farmers’ likelihood for IAS since a fully functioning market does not yet exist
to use these services and therefore to prefigure the devel- and thus this can inform decisions on implementing more
opment in the near future (Small and Svendsen, 1990). tailored tools. CE are stated preference methods that simu-
To this purpose, Altobelli et al. (2018) found that farmers’ late the real choice (Hensher et al., 2005; Louviere et al.,
preferences are positively influenced by many attributes 2000): farmers are asked to choose among several multi-
as the scale (i.e. entire area of the farm instead of single attribute IAS options their favourite one: behind the func-
fields), the duration of the service delivering contract tioning of this method lays the assumption that individuals
(i.e. 3 years); as opposite, they found that the impact of derive the value of a good by summing the value of its
water saving on farmers’ choice was poor. characteristics (Lancaster, 1966). Moreover, the respon-
In this paper, we present a structured survey carried out dents’ choice is function of the probability that the utility
in five countries to understand the importance of technolo- of an option, that derives from the sum of the utility of its
gical innovation in agricultural water management and to characteristics, is higher than the utility from the other
identify farmers’ preferences and willingness to pay for the alternatives in the same choice set, thus relying on the
IAS. In particular, we investigated which are the most pre- random utility theory (McFadden, 1974). According to
ferred among six different IAS attributes. To do this, we this theory, the utility (U) that farmer i derives from a
apply a hypothetical choice experiment (CE), coherently specific IAS configuration s is function of a deterministic
with similar studies related to farmers’ adoption of new part (V) and a stochastic unobservable element (e) of the
technologies (e.g. Lambrecht et al., 2015). Furthermore, farmer choice:
Altobelli et al. 3

Table 1. List of the attributes and levels used in the choice experiment.

Attribute description Code Levels

Weather forecasts (every 1, 2/3 or 4/5 days) forec 1 day 2–3 days 4–5 days
Crop water requirement (every 1, 10 or 15 day) cwr 1 day 10 days 15 days
Registration of water data and information (1, 2 or 3 times/month) regwa 1 time 2 times 3 times
Crop monitoring (every 1/7 or 15 days) crom 1 day 7 days 15 days
Contract duration (years) cont 1 year 2 years 3 years
Price price 5 €/ha 10€/ha 15€/ha
Source: Our elaborations.

over the crop growing period and change in soil moisture

Uis ¼ Vi ðXs Þ þ eis ¼ bi Xs þ eis ð1Þ content on the other side. It varies considerably with cli-
where Xs represents a vector of observable characteristics matic conditions, seasons, crops and soil types. In this
of the alternative s, while bi is a conformable parameter study, the irrigation consumptive water use is computed
vector indicating the direction and magnitude of the statis- for each country on the basis of the irrigated crop calendar
tical association between farmers’ characteristics and the for a specific year, as the difference between the crop
utility associated to the Irrigation Advisory Services’ attri- water requirement and the water balance under natural
butes and levels. By assuming the rationality of the indi- conditions.
viduals, it follows that the famer i chooses the alternative The third attribute relates to the number of times (i.e.
configuration of IAS s compared to the other options l in one time, two times or three per month) water informa-
the choice set (namely, a finite set of alternatives, three in tion, in terms of irrigation dates and volumes applied,
our design) if its utility (Uis) is higher than the utility he have to be recorded (regwa). The CE also considered an
derives from the other alternatives (Uil) as follows: attribute describing crop monitoring (crom) in terms of
the frequency of satellite data availability (1, 7 or
Uis > Uil 8l 6¼ s ð2Þ 15 days). Furthermore, another attribute is the preferred
Parameter bi are here considered as random parameters, duration of the IAS contract (cont) between annual, crop
varying across farmers. The distribution of each parameter cycle-based or triennial option; indeed, according to the
follows across the sample a normal distribution N * literature (Altobelli et al., 2018; Biswas and Venkatacha-
(m, s2) allowing the possibility to investigate the hetero- lam, 2015), a long contract is more likely to be preferred
geneity of preferences across farmers explicitly for the dif- because farmers recognize the positive impact of the ser-
ferent IAS characteristics: the greater (in case of statistical vice on crop production, due to an optimized irrigation
significance) the s2, the wider will be the heterogeneity of management.
preferences across farmers. Finally, m and s2 of bi can be Finally, the attribute price (price) is linked to several
estimated through a mixed logit conditional model by aspects of IAS, such as the cost for training farmers and
maximum likelihood estimator. the cost of implementing the irrigation services: here, each
Through the CE, farmers’ preferences towards six attri- level of price represents an average and indicative value
butes related to the IAS configuration was evaluated. These based on evidence from a previous investigation among
derived from the focus groups conducted among different different irrigation services carried out within the consid-
stakeholders as farmers, farmers’ associations, and land ered countries (Altobelli et al., 2018); the amounts varied
reclamation consortia in each country. Each attribute is from 5 € to 10 € and 15 € per hectare.
characterized by three levels (Table 1). Since a full factorial design consisting of all combi-
The first attribute regards the time length of weather nations of experimental attributes and levels would
forecasts (forec). Weather forecasts provide key informa- require 729 possible choice sets, an orthogonal experi-
tion for strategic farming management decisions. The sec- mental design was generated, generating a total of
ond attribute regards the forecast of total water amount 24 choice sets that will be then aggregated into four
required by the crop (cwr), a relevant factor to consider blocks. Hence, each questionnaire presented six choice
from the sowing until the harvesting (Gowing and Ejieji, sets to each respondent, that is considered an acceptable
2001; Memon and Jamsa, 2018). In particular, crop water number by the literature (Bech et al., 2011). In each
requirement refers to a range of forecasting for 1, 10 or choice set the individual could choose only once among
15 days and represent the amount of water to provide by three alternative options (A–B–C) and an opt-out option
irrigation to the crops. More specifically, irrigation is nec- (no choice or alternative D); this latter (status quo)
essary to compensate for the evapotranspiration (crop tran- allows realism in choices (Adamowicz and Boxall,
spiration and soil evaporation) deficit. Accordingly to 2001). Each alternative from A to C was described by
FAO, irrigation consumptive water use is defined as the six attributes and represented a specific combination of
volume of water needed to compensate for the deficit attribute levels, as shown in the example in Table 2.
between potential evapotranspiration of the considered Each attribute was explained in detail to respondents
crop (kc*ET0) on the one side and effective precipitation in order to avoid misinterpretation bias.
4 Outlook on Agriculture XX(X)

Table 2. Example of a choice set.

Which of the following alternative options (from A to D) would you prefer?

A B C D

Weather forecasts (every 1, 2/3 or 4/5 days) 4–5 days 1 day 4–5 days nothing
Crop water requirement (every 1, 10 or 15 day) 1 day 1 day 1 day
Registration of water data and information (1, 2 or 3 times/month) 2 times per month once per month 3 times per month
Crop monitoring (every 1/7 or 15 days) per day per day 15 days
Contract duration (years) 3 years 2 years 1 year
Price 15 €/ha 5 €/ha 10 €/ha
Source: Our elaborations.

Table 3. Gamble task. particularly chosen by more risk averse subjects while the
last by risk seekers (Charness et al., 2013). Once elicited,
Please, mark only one gamble with a X
farmers risk attitude is implemented to describe farmers’
Gamble Roll Payoff Chances Choice WTP related to the IAS attributes.
The coefficient of risk attitude is measured assuming
1 low 560 € 50% constant relative risk aversion (CRRA) utility function for
high 560 € 50%
which the utility is defined as U(x) ¼ x(1  r)/(1  r), with r
2 low 480 € 50%
high 720 € 50% representing the relative risk aversion coefficient and x
3 low 400 € 50% representing wealth.
high 880 € 50%
4 low 320 € 50%
high 1,040 € 50%
5 low 240 € 50% Results
high 1,200 € 50% The highest percentage of interviewees were from Poland
6 low 40 € 50%
high 1,400 € 50%
(43%), followed from Italy (33%), Spain (16%), the
Netherlands (6%) and South Africa (2%). The sample com-
Source: Own elaboration. prised 33 farmers over 50 years old, followed by 17 farmers
aged 40–50 years, 17 aged 30–40 years, and 8 farmers that
Risk attitude elicitation were less than 30 years old. From the total, 16 farmers did
not answer. The youngest respondent was a 23-year-old
In order to elicit the respondents’ individual risk attitude, Dutch man and the eldest a 71 years old Polish man. Most
we proposed a lottery task inspired by Eckel and Grossman of the sample was made up of males (85%), while female
(2008) this elicitation technique has been already used by respondents were the 15%: these latter were between 26
the literature investigating farmers’ decision (see for and 61 years old and a quarter of them had University
instance Menapace et al., 2013; Giampietri et al., 2020) degree, as opposite to men (16%). Family farms were the
as it involves only few gambles and a single choice, thus majority (63%), followed by professional-industrial farms
resulting less complex than other techniques, e.g. the (24%) and a mixed typology (13%). The farms were mainly
widely used lottery task by Holt and Laury (2002) (Dave specialized in the production of vegetables (68%), followed
et al., 2010). More specifically, respondents were asked to by livestock (14%). The utilized agricultural area (UAA)
imagine having an amount of 560€ to play a lottery by ranged between 0.5 ha (Polish farm) to 220 ha (South
choosing only one preferred gamble (only one) among six Africa farm), with 27% of farms having a UAA below
different alternatives (from 1 to 6) (Table 3). Besides, in 5 ha, 33% between 5 and 10 ha, 30% between 10 and
addition, we added a contextual frame to the lottery by 20 ha and 0.1% over 20 ha. In recent years, 50% of farmers
asking farmers to imagine that the payoff represented the invested in innovation at farm level (both product, techno-
net income (€/ha) they as farmers would gain from a hec- logical innovation and on farm organization) and, due to
tare of cash crop (e.g., wheat). It is noteworthy that the use this, 44% of respondents stated they had a 33% increase in
of this context setting that is very intuitive to farmers (i.e., production capacity. Nonetheless, investments in innova-
the gambles are intended in terms of farmers’ net income tion have not resulted in increased exports, market share or
per hectare) made the lottery more easily comprehendible, overall employment.
thus minimizing errors in decision making (Menapace Regarding the measure of risk attitude, Table 4 shows
et al., 2016). Moreover, farmers were told to consider that the range of CRRA (r) values for farmers choosing a spe-
each gamble had two possible outcomes (low and high roll) cific gamble; we used the lower bound for the estimation
in terms of payoff (€) with 50% chance of occurring each, (Menapace et al., 2013). Consistently with the literature
exception made for gamble 1 (no loss) that entails high-risk (Iyer et al., 2020), most farmers show a significant level
aversion. The choice in the gamble task represents the indi- of risk aversion, while only a minority (6%) represent risk
vidual risk attitude, with the first gambles that are seekers.
Altobelli et al. 5

Table 4. Intervals for CRRA measure of risk aversion and share of farmers choosing each gamble.

Gamble Payoff (€) Chances (%) Expected payoff (€) Standard deviation CRRA Farmers (%)

1 560 50 560 0 r > 2.6 27.5%

560 50
2 480 50 600 120 1.2 < r < 2.6 12.7%
720 50
3 400 50 640 240 0.8 < r < 1.2 30.4%
880 50
4 320 50 680 360 0.5 < r < 0.8 15.7%
1,040 50
5 240 50 720 480 0.1 < r < 0.5 7.8%
1,200 50
6 40 50 720 680 r < 0.1 5.9%
1,400 50
Source: Own elaboration.

Table 5. Mixed logit results.

Attribute b Coef (m) [Link] t-stat p-value

Price (euro/ha) 0.089 0.022 4.04 0.000

Weather forecasts (every 1, 2/3 or 4/5 days) 0.360 0.054 6.72 0.000
Contract duration (years) 0.247 0.086 2.86 0.004
Crop water requirement (every 1, 10 or 15 day) 0.055 0.231 0.24 0.813
Crop monitoring (every 1/7 or 15 days) 0.034 0.013 2.59 0.010
Registration of water data and information (1, 2 or 3 times/month) 0.314 0.128 2.44 0.015

No Choice 0.541 0.411 1.32 0.188

b Coef (s) [Link] t-stat p-value

Weather forecasts (every 1, 2/3 or 4/5 days) 0.090 0.099 0.92 0.359
Contract duration (years) 0.333 0.097 3.44 0.001
Crop water requirement (every 1, 10 or 15 day) 1.740 0.273 6.39 0.000
Crop monitoring (every 1/7 or 15 days) 0.036 0.019 1.91 0.056
Registration of water data and information (1, 2 or 3 times/month) 0.494 0.139 3.55 0.000

WTP (€/ha) [Link]

Weather forecasts (every 1, 2/3 or 4/5 days) 4.11 0.28

Contract duration (years) 2.86 1.86
Crop water requirement (every 1, 10 or 15 day) 0.27 16.59
Crop monitoring (every 1/7 or 15 days) 0.38 0.16
Registration of water data and information (1, 2 or 3 times/month) 3.58 3.00
Source: Our elaborations.

Concerning the CE, Table 5 shows the results obtained positively from: a) time length of forecasts (forec): the
from the mixed logit conditional regression. The results more days available in the future, the more the service is
concerning the estimated coefficients (random coefficients chosen; b) need to record water information (regwa) (i.e.
b) can be interpreted as follows: the economic intuition of the longer the time interval farmers need to record infor-
the parameters m is strictly based on the associated marginal mation, the higher their willingness to accept the service).
utility of the corresponding attribute, while it can be statis- As opposite, a negative effect was found for: a) price: as the
tically interpreted as proportional odds ratios like a usual price increases, the probability for a farmer to adopt the
logit regression. Estimated values of m show that, excluding service decreases b) duration of the contract (cont) (i.e.
price, the IAS characteristics that generate greater utility or shorter contracts are preferred); c) frequency of crop mon-
preference to farmers are (in order of relative importance): itoring availability (crom) (i.e. when the availability of
i) weather forecasts, ii) registration of water data and infor- monitoring changes from 1 day to 7 or 15 days the prob-
mation, iii) duration of the contract, and iv) crop monitor- ability of choice decreases). The same conclusions can be
ing, while on average crop water requirement would seem obtained by calculating the monetary assessment of the
irrelevant. preferences, namely, the farmers’ WTP per hectare.
More in detail, by investigating the sign of the coeffi- Table 6 shows the average values of the WTPs for the
cients, the choices of the type of service seems depending different attributes of IAS, expressed in euros per hectares
6 Outlook on Agriculture XX(X)

Table 6. Average WTP of IAS attributes for the different countries (€/ha).

The South
Italy Poland Spain Netherlands Africa Total

Weather forecasts (every 1, 2/3 or 4/5 days) 4.10 4.09 4.05 4.32 4.27 4.11
Contract duration (years) 2.67 2.63 3.50 3.48 4.01 2.86
Crop water requirement (every 1, 10 or 15 day) 8.31 4.04 10.50 11.98 10.20 0.27
Crop monitoring (every 1/7 or 15 days) 0.37 0.36 0.45 0.40 0.38 0.38
Registration of water data and information (1, 2 or 3 times/ 3.94 3.41 4.25 0.41 5.28 3.58
month)
Source: Our elaborations.

Table 7. Average WTP of IAS attributes according the farmers risk aversion (€/ha).

Weather forecasts Contract duration

Risk seeker Risk Averse diff. Risk seeker Risk Averse diff.

Mean 4.138 4.062 0.077 2.509 3.333 0.825**

[Link] 0.039 0.036 0.271 0.202

Crop water requirement Crop monitoring

Risk seeker Risk Averse diff Risk seeker Risk Averse diff

Mean 6.214 8.865 15.1*** 0.375 0.391 0.016

[Link] 1.681 2.518 0.023 0.020

Registration of water data and information

Risk seeker Risk Averse diff

Mean 3.160 4.128 0.967*

[Link] 0.369 0.462
Note: *** significant at 1% level, ** significant at 5% level, * significant at 10% level.
Source: Own elaboration.

for the different countries involved in the study. Farmers’ longer time interval, although this appears slightly lower
WTP for the different IAS services largely varies across than risk seekers.
countries. For instance, as concerns, the weather forecast’s
length, Dutch farmers (4.32 €/ha) and those from South
Africa (4.27 €/ha) are willing to pay a higher price per Discussion and conclusion
hectare than farmers coming from other countries. Spanish Nowadays, an efficient and sustainable use of water for
and Polish farmers show a negative WTP in relation to the irrigation is essential. Accordingly, there is a plenty of
need to receive information about crop’s specific water studies showing that farmers are worried about irrigation
requirements. water scarcity, especially during the summer, and the
Finally, Table 7 shows WTP differences of for the con- conflicts arising from the competitive water consumption
sidered IAS attributes according to farmers’ risk attitude. (Alcon et al., 2014; Assefa, 2012; Aydogdu, 2019;
More in detail, average values for risk seeker farmers’ vs Aydogdu and Bilgic, 2016; Aydogdu and Yenigun, 2016;
risk averse ones have been compared. In particular, risk Khan and Zhao, 2019). This makes farmers willing to sup-
attitude is able to explain positive WTP from those nega- port and adopt any lasting solution to irrigation water scar-
tive in terms of the attribute related to crop water require- city (Oremo et al., 2020). Water managers often rely on
ment: indeed, while risk seekers are willing to pay for it (on decision support tools, including hydrological models;
average around 6 euros), risk averse farmers perceive this however, water management also requires improving the
service as a disutility, thus associating a negative WTP efficiency of water use. In this context, the introduction of
(8.9 euros). Generally speaking, risk averse farmers pre- Irrigation Advisory Services (IAS) could advance irrigation
fer both shorter contracts and a shorter crop monitoring practices and water efficiency in the near future, while
period, and perceive a higher disutility in registering water providing an economic advantage for farmers: indeed, the
data and information with a lower time frequency, com- adoption of this new irrigation management system can
pared to risk seeking farmers. As concerns the weather both increase farmers’ income and reduce energy costs
forecasts, the WTP of risk averse subjects is higher for a associated with water management consultancy. So far, the
Altobelli et al. 7

information on these devices is scarce among farmers and found interesting. The service should be reliable and
this represents one major obstacle to the adoption: it fol- directly delivered through Internet and/or mobile phone.
lows that the spreading of technical assistance is necessary. However, there is not a panacea that will satisfy all farmers
Moreover, a better understanding of the economic value in all geographic areas (Burton et al., 2020).
that farmers attribute to IAS among policy makers is advo-
cated. In line with this, this study provides an interesting Acknowledgement
framework for correctly understanding the potential of IAS. The authors would like to thank the EU and The Ministry of
To do this, farmers’ preferences were analysed using a Economic Affairs and The Ministry of Agriculture, Nature and
choice experiment, which seems to be an effective method Food Quality (The Netherlands), CDTI (Spain), MINECO
to estimate the non-market value of these innovative water (Spain), ANR (France), MIUR (Italy), NCBR (Poland) and WRC
use technologies. In particular, based on the trade-offs that (South Africa) for funding, in the frame of the collaborative inter-
respondents make between attributes, we were able to esti- national consortium OPERA financed under the ERA-NET
Cofund WaterWorks2015 Call. This ERA-NET is an integral part
mate the average WTP value (implicit prices for IAS) of the
of the 2016 Joint Activities developed by the Water Challenges
proposed IAS features for the whole sample. Further stud- for a Changing World Joint Programme Initiative (Water JPI).
ies may use the WTP values estimated in this analysis in a
future cost-benefit analysis of the IAS. Furthermore, in Declaration of conflicting interests
order to understand the level of importance of technologi-
The author(s) declared no potential conflicts of interest with
cal innovation in the agricultural water management, an respect to the research, authorships, and/or publication of this
analysis of the risk attitude of the farmers interviewed dur- article.
ing the survey was carried out. Our results show that farm-
ers are actually willing to pay to introduce an irrigation Funding
support system that translates into an economic advantage The author(s) received no financial support for the research,
over their current situation. On the other hand, it has been authorship, and/or publication of this article.
shown that farmers’ willingness to pay changes according
to the attributes that characterize the service offered and ORCID iDs
according to the country of origin. Filiberto Altobelli [Link]
In line to the recent overall trends weather forecasting, Anna Dalla Marta [Link]
tools for identifying pests and diseases and climate risks are Samuele Trestini [Link]
perceived as useful by most European farmers, including Wiesława Kasperska-Wolowicz [Link]
those come from Italy and Germany (Altobelli et al., 2018; 8265-8786
Bonke et al., 2018). Our results prove that the IAS func- Marlene De Witt [Link]
tions that generate greater preference to farmers are mainly
the weather forecasts, following by registration of water References
data and information, duration of the contract, while crop Adamowicz V and Boxall P (2001) Future directions of stated
water requirement seems less relevant. Appreciation for choice methods for environment valuation. In: Adamowicz
wheatear forecasts confirm that farmers can benefit sub- V and Boxall P (eds) Choice Experiments: A New Approach
stantially from short to medium weather forecasts, which to Environmental Valuation. London, 10 April 2001, pp. 1–6.
can help to optimize farming operations and deal more Alcon F, Tapsuwan S, Brouwer R, et al. (2014) Adoption of
effectively with the adverse impacts of climate variability, irrigation water policies to guarantee water supply: a choice
including extreme weather events as water scarcity experiment. Environmental Science & Policy 44: 226–236.
(Calanca et al., 2011). This brings out the need for addi- Altobelli F, Lall U, Dalla Marta A, et al. (2018) Willingness of
tional efforts to increase the quality of the forecasts farmers to pay for satellite-based irrigation advisory services:
(decrease the associated uncertainty), as well as design a southern Italy experience. The Journal of Agricultural Sci-
appropriate operational tools and promote the dissemina- ence 156: 723–730.
tion of the outcome within the farmers community. Ques- Altobelli F, Monteleone A, Cimino O, et al. (2019) Farmers’
tions that need to be addressed are, whether the forecasts willingness to pay for an environmental certification scheme:
are in an appropriate form, predict the proper variables and promising evidence for water saving. Outlook on Agriculture
refer to the relevant time scales (Garbrecht et al., 2005; 48(2): 136–142.
Wilks, 1997). As reported in this work, by investigating Assefa N (2012) Valuing the economic benefit of irrigation water:
the sign of the coefficients, the choices of the type of ser- application of choice experiment and contingent valuation
vice seems to be mainly related to the time length of fore- methods to Ribb irrigation and drainage project in South Gon-
cating (forec). der, Ethiopia. Dissertation, Addis Ababa University, Ethiopia.
The possibility of monitoring the field variability Aydogdu M (2019) Farmers’ attitudes to the pricing of natural
through remote sensing raises a strong interest, but more resources for sustainability: GAP Sanlıurfa Sampling of Tur-
at the level of the curiosity than as decision tool, since key. Water 11: 1772.
decision to be taken from such information is not straight- Aydogdu M and Bilgic A (2016) An evaluation of farmers’ will-
forward. However, a control of the irrigation system and its ingness to pay for efficient irrigation for sustainable usage of
management might be appreciated. Weather forecast, eva- resources: the GAP-Harran Plain case, Turkey. Journal of
luation of the crop water need and water stress status are Integrative Environmental Sciences 13(2–4): 175–186.
8 Outlook on Agriculture XX(X)

Aydogdu M and Yenigun K (2016) Willingness to pay for sus- DOI: 10.13128/bliterature investigating farmers’ decisionli-
tainable water usage in Harran Plain-Gap Region, Turkey. terature investigating farmers’ decisionae-8416.
Applied Ecology and Environmental Research 14(3): 60–147. Gowing JW and Ejieji CJ (2001) Real-time scheduling of supple-
Bech M, Kjaer T and Lauridsen J (2011) Does the number of mental irrigation for potatoes using a decision model and
choice sets matter? Results from a web survey applying short-term weather forecasts. Agricultural Water Management
a discrete choice experiment. Health Economics 20(3): 47(2): 137–153.
273–286. Hensher DA, Rose JM and Greene WH (2005) Applied Choice
Belmonte AC, Gonzalez JM, Mayorga AV, et al. (1999) GIS tools Analysis: A Primer. Cambridge: Cambridge University Press.
applied to the sustainable management of water resources: Holmes Thomas P, Adamowicz WL, and Carlsson F (2017)
application to the aquifer system 08-29. Agricultural Water “Choice experiments.” A primer on nonmarket valuation.
Management 40: 207–220. Dordrecht: Springer, pp. 133–186.
Biswas D and Venkatachalam L (2015) Farmers’ willingness to Holt C, Laury S (2002) Risk aversion and incentive effects. Amer-
pay for improved irrigation water. A case study of Malaprabha ican Economic Review 92(5): 1644–1655.
Irrigation Project in Karnataka, India. Water Economics and IDAE (2005) Ahorro y Eficiencia Energética en Agricultura de
Policy 1: 1450004. Regadı́o, Madrid. Jang, J.-S.R., 1993. ANFIS: adaptive-
Bonfante A, Monaco E, Manna P, et al. (2019) LCIS DSS – an network-based fuzzy inference system. IEEE Transactions
irrigation supporting system for water use efficiency improve- on Systems, Man, and Cybernetics 23: 665–685.
ment in precision agriculture: a maize case study. Agricultural Iyer P, Bozzola M, Hirsch S, et al. (2020) Measuring farmer risk
Systems 176: 102646. ISSN 0308-521X. 675 preferences in Europe: a systematic review. Journal of
Bonke V, Fecke W, Michels M, et al. (2018) Willingness to pay Agricultural Economics 71(1): 3–26.
for smartphone apps facilitating sustainable crop protection. Khan I and Zhao M (2019) Water resource management and
Agronomy for Sustainable Development 38(5): 51. public preferences for water ecosystem services: a choice
Burton M, Cooper B and Crase L (2020) Analysing irrigation experiment approach for inland river basin management. Sci-
farmers’ preferences for local governance using a discrete
ence of the Total Environment 646: 821–831.
choice experiment in India and Pakistan. Water 12: 1821.
Kuo Sheng-Feng, Merkley GP, and Liu Chen-Wuing (2000) Deci-
Calanca P, Bolius D, Weigel AP, et al. (2011). Application of
sion support for irrigation project planning using a genetic
long-range weather forecasts to agricultural decision problems
algorithm. Agricultural Water Management 45(3): 243–266.
in Europe. The Journal of Agricultural Science 149(1): 15–22.
Lambrecht I, Vranken L, Merckx R, et al. (2015) Ex ante appraisal
Charness G, Gneezy U and Imas A (2013) Experimental methods:
of agricultural research and extension: a choice experiment on
eliciting risk preferences. Journal of Economic Behavior and
climbing beans in Burundi. Outlook on Agriculture 44 (1):
Organization 87: 43–51.
61–67.
Chèze B, David M, and Martinet V (2020) Understanding farm-
Lancaster KJ (1966) A new approach to consumer theory. Journal
ers’ reluctance to reduce pesticide use: A choice experiment.
of Political Economy 74(2): 132–157.
Ecological Economics 167: 106349.
Levidow L, Zaccaria D, Maia R, et al. (2014) Improving water-
D’Urso G, De Michele C and Bolognesi SF (2013) IRRISAT: the
efficient irrigation: prospects and difficulties of innovative
Italian On-line Satellite Irrigation Advisory Service. In:
practices. Agricultural Water Management 146: 84–94.
EFITA-WCCA-CIGR Conference Sustainable Agriculture
Louviere JJ, Hensher DA and Swait JD (2000) Stated Choice
Through ICT Innovation, Turin, June 2013.
Dave C, Eckel CC, Johnson CA, et al. (2010) Eliciting risk pre- Methods: Analysis and Applications. Cambridge: Cambridge
ferences: When is simple better? Journal of Risk and Uncer- University Press.
tainty 41(3): 219–243. Manhoudt AGE, Van de Ven GWJ, de Haes HU, et al. (2002)
De Santa Olalla Sanchez A (1999) Una propuesta de codificacion Environmental labelling in The Netherlands: a framework for
morfosintactica para corpus de referencia en lengua espanola. integrated farming. Journal of Environmental Management
Estudios Lingüı`stica del Espan˜o 3. 65(3): 269–283.
Despotović J, Rodić V and Caracciolo F (2019) Factors affecting Mannini P, Genovesi R and Letterio T (2013) Four decades of
farmers’ adoption of integrated pest management in Serbia: an progress in monitoring and modeling of processes in the
application of the theory of planned behavior. Journal of soil-plant-atmosphere system: applications and challenges
Cleaner Production 228: 1196–1205. IRRINET: large scale DSS application for on-farm irrigation
Eckel CC and Grossman PJ (2008) Forecasting risk attitudes: an scheduling selection and/or peer-review under responsibility
experimental study using actual and forecast gamble choices. of the scientific committee of the conference. Procedia Envi-
Journal of Economic Behavior & Organization 68(1): 1–17. ronmental Sciences 19: 823–829.
FAO (2014) Available at: [Link] Mateos L, López-Cortijo I and Sagardoy JA (2002) SIMIS: the
didyouknow/[Link] (accessed 10 March 2021). FAO decision support system for irrigation scheme manage-
Garbrecht J, Meinke H, Sivakumar MV, et al. (2005) Seasonal ment. Agricultural Water Management 56(3): 193–206.
climate forecasts and adoption by agriculture. Eos, Transac- McFadden D (1974) Conditional logit analysis of qualitative
tions American Geophysical Union 86(24): 227–227. choice behavior. In: Zarembka P (ed) Frontiers in Econo-
Giampietri E, Yu X and Trestini S (2020) The role of trust and metrics. New York: Academic Press, pp. 105–142.
perceived barriers on farmer’s intention to adopt risk manage- Memon AV and Jamsa S (2018) Crop water requirement and
ment tools. Bio-Based and Applied Economics 9(1): 1–24. irrigation scheduling of Soybean and Tomato crop using
Altobelli et al. 9

CROPWAT 8.0. International Research Journal of Engineer- Oremo F, Mulwa R, and Oguge N (2021) Sustainable water access
ing and Technology 5(9): 669–671. and willingness of smallholder irrigators to pay for on-farm
Menapace L, Colson G and Raffaelli R (2013) Risk aversion, water storage systems in Tsavo sub-catchment, Kenya. Envi-
subjective beliefs, and farmer risk management strategies. ronment, Development and Sustainability 23(2): 1371–1391.
American Journal of Agricultural Economics 95(2): Oyinbo O, Chamberlin J, Vanlauwe B, et al. (2019) Farmers’
384–389. preferences for high-input agriculture supported by site-
Menapace L, Colson G and Raffaelli R (2016) A comparison of specific extension services: evidence from a choice experi-
hypothetical risk attitude elicitation instruments for explaining ment in Nigeria. Agricultural Systems 173: 12–26.
farmer crop insurance purchases. European Review of Agri- Rossi F, Nardino M, Mannini P, et al. (2004) IRRINET Emilia
cultural Economics 43(1): 113–135. Romagna: online decision support on irrigation. Online agro-
Navarro-Hellı́n H, Torres-Sánchez R, Soto-Valles F, et al. meteological applications with decision support on the farm
(2015) A wireless sensors architecture for efficient irrigation level. Cost Action 718: 99–102.
water management. Agricultural Water Management 151: Small LE and Svendsen M (1990) A framework for assessing
64–74. irrigation performance. Irrigation and Drainage Systems
Oremo F, Mulwa R. and Oguge N (2020) Sustainable water access 4(4): 283–312.
and willingness of smallholder irrigators to pay for on-farm Wilks D (1997) Forecast value: prescriptive decision studies. In:
water storage systems in Tsavo sub-catchment, Kenya. In: Katz RW and Murphy AH (eds) Economic Value of Weather
Environment, Development and Sustainability. Netherlands: and Climate Forecasts. New York: Cambridge University
Springer, pp. 1–21. Press, pp. 109–145.