**1. Introduction**

Ecosystem services are critical for the sustainability of the natural environment, food security, and livelihoods of the resource-poor inhabitants of the cold winter deserts (CWDs) in Uzbekistan. CWDs are not well endowed with natural resources, and food production heavily depends on the availability of water. Even though some efforts are being made by government agencies to sustain the ecosystem services in the CWDs, there is no evidence to suggest that these efforts will continue or grow to make a difference in these deserts. On the other hand, the farming community in the CWDs generally tends to rely on government investments for improvement and management of the ecosystem services.

Fast growing urbanization and the investment required are making it increasingly difficult for the government of Uzbekistan to give as much emphasis to these important, but very fragile CWDs. Therefore, contributions by the local community will be essential to improve and sustain ecosystem services in the CWDs of Uzbekistan. It is necessary that the communities in these deserts engage more actively in the planning and implementation of sustainable land and water management activities. An important tool in managing common pool resources is payment for ecosystem services [1]. The government agency managing the deserts may enforce such a payment in the future. However, without the users' willingness to participate in the process, the governmental decision on the payment system for ecosystem services may not be sustainable [2,3]. CWDs provide a number of

**Citation:** Kassie, G.T.; Boboev, H.; Sharma, R.; Akramkhanov, A. Willingness to Pay for Irrigation Services in the Cold Winter Deserts of Uzbekistan. *Sustainability* **2022**, *14*, 94. https://doi.org/10.3390/su14010094

Academic Editors: Alban Kuriqi and Luis Garrote

Received: 26 November 2021 Accepted: 17 December 2021 Published: 22 December 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

ecosystem services, including biomass production, sand fixation, firewood, and belowand above-ground carbon storage and buffer functions in the inter-annual carbon cycle [4]. Access to and use of irrigation water is one of the most important ecosystem services in CWDs of Uzbekistan. It can be argued that irrigation schemes developed based on the needs and preferences of the target users are more likely to be owned and efficiently managed by the users. Hence, the costs and benefits of the irrigation schemes need to be estimated and compared from the perspective of the target users as well.

The key components of the cost are mainly direct and can be measured relatively easily. However, the direct and indirect benefits of the irrigation services can hardly be measured with as much simplicity. This is why non-market-based economic valuations are gaining traction in estimating the value of access to irrigation services or the different attributes of irrigation services. The value of the access to irrigation is commonly estimated using different adaptations of contingent valuation [5–10]. Direct elicitation methods would not, however, enable us to look into the relative importance of different components of the irrigation services. The services are the sum total of the different attributes characterizing them. Discrete choice experiments (DCEs) are one of the stated preference-elicitation methods that can help us assess the relative importance of the components. Compared to the other elicitation methods, DCEs would enable us to rank the different attributes and the attribute levels.

A few studies have employed DCEs to estimate implicit prices of the attributes of irrigation services [11]. This study analyzed the way irrigation water should be managed in South Africa and recommended a shift towards on-farm volumetric water pricing in the irrigation schemes. Another study that used DCE reported positive and significant willingness to pay (WTP) for cropping intensity, frequency of watering, and crop under production aspects of the irrigation system among smallholder farmers in Ethiopia who have never paid for irrigation water use [12]. A DCE study in Punjab, Pakistan reported that sample farmers had a WTP much higher than the current average rates for improved surface water reliability [13]. A study in four regions of India and Pakistan investigated preferences of farmers for different forms of irrigation fees and models of local governance [14]. The study emphasized the heterogeneity across the study areas and the need for localized approaches in determining irrigation fees and governance.

Despite the popularity of DCEs in other field of applied economics research, we could not find other peer reviewed studies that used DCEs to elicit preferences for characteristics of irrigation services. In fact, there is not any study on the valuation of irrigation services from the perspective of the target users in Uzbekistan or in Central Asia in general.

Uzbekistan's agriculture is cotton and wheat-centric and almost entirely dependent on irrigation. About 90% of the water resources in Uzbekistan are used in the agriculture sector, and it is used with low efficiency [15]. Only 11–12% of the water consumed in the country comes from within the country [16]. Most of the food and feed production happens in irrigated agriculture, which covers merely 10% of total cultivable land (4.3 million ha), demonstrating the importance of water to people's livelihoods not only in the CWDs but also in the entire country. CWDs and semi-deserts constitute about 85 percent of Uzbekistan's land mass [17]. It is almost impossible to overemphasize the importance of these deserts and the implications of the availability of water in Uzbekistan and in our study area. Agrarian livelihoods in the CWDS are becoming more and more fragile and vulnerable because of scarcity and high variability of the water supply.

A report in 2009 indicated that the welfare of the republic depended on the possibility of ensuring the water supply for almost 29 million people, for the irrigation of 4.3 million hectares, and for industry and for the environment [16]. At present, the total annual water use in the republic is 55.1 km3, of which irrigated agriculture uses 49.7 km3 and the domestic and drinking water supply for urban and rural populations uses 3.4 km<sup>3</sup> [16]. The immediate solutions revolve around increasing water use efficiency (WUE) and developing sustainable water-management systems.

Shortages of water and deterioration of water and land resources are observed throughout Uzbekistan [15,17,18]. Most of the irrigated area is subjected to salinization [19], waterlogging, water erosion, agro-biodiversity losses, and other very hazardous processes [20,21]. This hampers the development of the economic system—including the agriculture sector and aggravates the challenges faced by the poor rural communities. Almost one fourth of Uzbekistan's population (more than 6 million) suffers from the negative effects of polluted water [22,23]. Research has also shown that, in Uzbekistan, the low income of the rural population is linked to the irregular supplies of irrigation water and the deterioration of land due to, among others, salinization and waterlogging [17].

Water shortage has become a key factor limiting the sustainable development of Central Asia, especially for the downstream agricultural countries like Uzbekistan [15,24]. It is not only the availability of water that is an issue but also the level of efficiency of its use. A recent study [25] argues that low water-use efficiency is a main factor contributing to water shortages in Central Asia. Irrigation facilities in the region are relatively backward, and the cropland relies on furrow irrigation, leading to low crop yields and a low utilization efficiency of water resources [25]. Improvement of these facilities is among the mostimportant political-economic priorities of Uzbekistan. Improvement will, however, happen only if it is based on careful and well-informed planning.

Irrigation is a technology with different characteristics or components. Farmers' interest in each of the components of the irrigation scheme determines their level of engagement and efficiency in use. This is the basis of our study, where we estimated the willingness to pay for the different attributes of irrigation schemes in this vulnerable ecosystem. By investigating farmers' choice strategies, we also looked into the relative importance of the different components of the irrigation schemes.

Planning sustainable development without proper valuation of ecosystem services can hardly be meaningful, as the focus will be predominantly on direct, local, and immediate benefits [25]. There is no better evidence than the Aral Sea crisis to show the failure associated with focusing on direct benefits from ecosystem services [26,27]. Currently, the most-degraded pastures are located in the Central Kyzylkum deserts, covering the Bukhara, Navoi, Khorezm, and Karakalpakstan, regions, where the misuse emanates from inadequate access to land, inappropriate land-management systems, and a lack of knowledge on sustainable use and management of these resources [18,21]. Weak institutional structures and procedures, as well as a lack of law enforcement, also contribute enormously to the challenges of sustainable development [24,28].

We identified five key attributes of irrigation schemes in the Bukhara region's Karakul District, one of the most-fragile ecosystems in Uzbekistan, and elicited preferences and estimated implicit prices. The attributes considered were water availability in the dry season (May to October), crop water frequency, irrigation water quality, water sharing with downstream users, and the fee for irrigation. We elicited preferences with discrete choice experiments and estimated willingness to pay for these attributes using a randomparameters logit. We also estimated a series of latent class models to investigate the relative focus given by farmers to the different attributes while choosing among the hypothetical irrigation schemes.

The results of this study will serve at least two purposes. First, the evidence will inform policymakers on what the focus of the irrigation development effort should be. Not all components of the irrigation schemes are equally important to farmers and the community. Second, we strongly believe that scientific evidence-based designing and implementation of irrigation schemes helps farmers cope with the unforgiving environment better, as they will have a scheme that addresses their priorities and the implied challenges thereof.

#### **2. Description of the Study Area**

The study was conducted in Durmon village in Karakul District, which is a central south region of Uzbekistan (Figure 1). The Karakul District Forestry Department, established in 1925, includes forest, pastures, and non-used land resources. The entire land

resources in the district cover 73,542 ha. These land resources are geographically located at 39.582991◦ N latitude, 63.905707◦ E longitude, and an altitude of 242 m above sea level. The territory of the forest department encompasses part of the Kimmerikum desert, the West Kyzyl Kum plains, and the ancient valley of the Zerafshan River. Durmon village has an area of 517 ha.

**Figure 1.** Map of Bukhara and its land cover. Note: Map prepared by ICARDA's geo-informatics unit.

Water availability is the most-important determinant of the land-management practices of farmers. In terms of land use and management, the study area has two distinct populations—the smallholder farm households who are always trying to eke out a living in this harsh environment and the employees of the Forestry department whose focus is controlling the natural resources both for political and conservation reasons.

Durmon village was selected as a representative site for cold winter deserts that cover different provinces in Uzbekistan and across other countries around the Aral Sea region in Central Asia. Livelihoods of the farming communities in Durmon depend entirely on the ecosystem services provided by the cold winter desert, and the CWDs are a major policy agenda in Uzbekistan. It is within the national strategy to improve management of the CWDs to enhance their contribution to rural food security on a sustainable basis.

The study site is characterized by quick climatic fluctuations, constant wind activity, the prevalence of sand, extremely low precipitation and humidity, aridity, and extensive degradation of the natural resources. The long-term average precipitation in the pilot area is only 108 mm—which is very low for rain-fed crop production. An important part of this precipitation during the growth period of the vegetation (mostly in spring) is around 30 mm only. During summer periods, the precipitation is totally absent, and the relative humidity sharply decreases with a long-term average of 36%. Although rare, timing of the first snow in the project pilot area varies from year to year, and usually starts in December and continues falling until January, sometimes lingering until mid-March.

The annual average air temperature is 14.8 ◦C, ranging from −22 ◦C (in January) to 47.1 ◦C (in July). The last days of cold weather happen in mid-spring (18.04 ◦C), while the first frost comes in mid-autumn (13.10 ◦C). The hot dry weather lasts 40–50 days during June–August and causes extreme heat and drying of vegetation. There is constant wind activity in this region, with dry and hot winds (locally called garmsel). Garmsel wind can happen for 45 to 50 days a year, blowing in June through August. This wind speed is a reason for slight-to-moderate soil erosion and movement of sand in the direction of wind. The common soil types in the study area are desert sands, takyr-like soils, grey-brown desert soils, meadow solonchak soils, and meadow irrigated solonchak soils. Soil salinity is common in this area and happens due to shallow mineralized groundwater.

The operations and management of the study area represent what is happening in the rest of the CWDs where irrigated agriculture is practiced. Depending on water availability, which is pumped from the source and delivered through canals, large parcels of land can be irrigated; however, due to water scarcity, only 13% of the total arable land is cropped, demonstrating the importance of water overall for the study site. Major crops grown in the area are vegetables, legumes, and wheat, typically requiring 5–10 thousand m<sup>3</sup> per hectare.

#### **3. Materials and Methods**

#### *3.1. The Choice Experiment*

Valuation of ecosystem services—including irrigation water—is best done with stated choice methods, as almost no services are directly marketable. Discrete choice experiments (DCEs) are a widely used data-generation method in stated choice analysis. Lancaster's characteristic theory of value (ToV) [29] and McFadden's random utility theory (RUT) [30] form the basis for estimating the relative importance of the attributes characterizing the service at hand, in this case, irrigation. The implicit prices of the characteristics show their relative importance and the structure of the latent satisfaction from the consumption of the service [29].

RUT assumes that the choice behavior of individual decision makers is probabilistic conditional on the characteristics of the services available to them and other factors that affect their choice decision. It is, therefore, expected that the choice behaviors of the individual decision makers vary because of variability in the underlying factors. The underlying factors include unobserved attributes, unobserved individual characteristics (or taste variations), and/or measurement errors [31]. The RUT also enables us to model unobserved heterogeneity in choice behavior among the sample households.

Irrigation water is a quality-differentiated service that can be described by its attributes. Transaction of such attributes does not happen in actual revealed markets, hence the need for stated choice methods such as discrete choice experiments [32]. Sample households are currently accessing irrigation water based on arrangements made by the government or water-user associations with support and guidance from the government. Sustainable use of the irrigation water resources, however, depends on farmers' actual demand for irrigation services. The demand for irrigation water services is a consolidation of the demand for the different components of the service. We derived the demand for the attributes of irrigation water services by eliciting sample households' preferences for the experimental designed irrigation schemes presented in the form of pair comparison with the option of opting out included.

The identification, definition, and prioritization of the characteristics of the hypothetical irrigation schemes in the choice experiment involved iterative focus-group discussions and a reconnaissance survey. A structured questionnaire survey was undertaken involving a sample of 200 farmers in 2020 to generate socioeconomic data and the attributes of irrigation services. The discussions with farmers resulted in the following attribute and levels for the choice experiment (Table 1).

Therefore, our design had five attributes of irrigation schemes. We used Ngene [33] to generate experimental designs that combine the attributes and create hypothetical irrigation schemes. Using main effects only, there could be 2 × 34 or 162 combinations of irrigation schemes with different levels of the five attributes. We, however, used fractional factorial design to limit the number of alternatives to a reasonable level. Our final D-optimal design had a D error of 1.32 and generated 36 alternatives paired in two to create nine choice sets. Each choice set therefore included two hypothetical irrigation schemes and an opt-out option—added to avoid forced choices.


**Table 1.** Attributes of irrigation schemes in the choice experiment.

† UZS stands for Uzbekistan Soms. In May 2020, 1 US Dollar was equivalent to 10,138.19 UZS.

To simplify farmers' choice decision-making process, we used pictorial representations for each level of the attributes in preparing the choice cards with which we elicited the choices. The DCEs were conducted at the residential homes of the respondents, and they were presented with nine choice sets to choose one among three alternatives in each of the sets. Before the interview, each respondent was briefed about the research and the mechanics of the irrigation scheme choice experiment. To ensure that that farmers have understood the experiment, one or two randomly selected choice sets were presented to them without recording the responses. Then, the nine sets were presented in random order for each of the sample respondents.
