Factors Influencing Farmers’ Adoption of Protected Agriculture: a Case of Vegetable Farmers in Dry Zone Sri Lanka

The farmers’ adoption of Protected Agriculture (PA) in dry zone vegetable cultivation is considerably low in Sri Lanka. However, a few farmers have been practiced PA technologies to minimize negative impacts of climate change on vegetable production in dry part of Sri Lanka. Rain shelters were identified as the main type of PA technology used by dry zone vegetable farmers followed by poly tunnels. Understanding the factors that influencing farmers’ adoption to PA could be more useful to popularize the technology in suitable areas while formulating policies to overcome some barriers for adoption. This study analyzed the factors influencing farmers’ adoption of PA for vegetable cultivation in dry zone of Sri Lanka. A total of 120 vegetable farmers from 12 Agrarian service centers (ASC) in Anuradhapura and 2 ASC in Matale districts were selected randomly to identify the factors that would be influencing farmers’ adoption on PA technologies for vegetable production in the dry zone of the country. Pre tested structured questionnaire were used for the survey in data collection. Binary logistic regression model was applied to identify the factors affecting on adoption to PA by vegetable farmers. Results reveal that farmers age, household size, cultivated extent, access to credit, access to new technologies, access to extension services and primary employment have a significant effect on adoption to PA. These results highlighted that importance of extension services and financial support to promote PA technologies for small scale vegetable farmers in dry zone.


Introduction
However, open field vegetable cultivations are highly depending on weather conditions mainly rainfall. In recent times vegetable production has been undergoing rigorous fluctuations due to unpredictable weather conditions in the country [2]. Compared to other climatic zones in the country, dry zone is more vulnerable to unexpected extreme climatic events such as droughts and floods [3]. As the major vegetable producing area, it is needed to identify and implement new farming technologies to minimize the impacts of extreme climatic events on vegetable production to maintain constant vegetable supply in the country. Adoption to protected agriculture on vegetable production will be one of the effective solution for minimize the impacts of these extreme climatic events.
Protected Agriculture (PA) is identified as one of the feasible adaptation strategy to minimize the impacts of climate changes on vegetable production as well as off seasons production with limited resources [4]. PA can be defined as cropping techniques where in the micro climate surrounding of the plant body is controlled partially or fully as per the requirement of the plant species grown during their period of growth [5]. PA technology involves the cultivation of vegetables especially in designed protective structures such as glass houses, poly tunnels, net houses and rain shelters where in the factors like temperature, humidity, light, soil, water and fertilizers are manipulated to attain maximum produce as well as allow a regular supply of them even during an off season [6]. Research results have shown that by adopting PA, productivity of vegetable crops can be increased by three to five times as compared to open field cultivations [7].
About 115 countries in the world are using PA technologies for vegetable production commercially. China is the major country using PA and 90 percent of its protected cultivations used for vegetable production [6]. Several studies showed that countries like India, Pakistan, Afghanistan, Israel and Iran are successfully practicing vegetable cultivation using PA under extreme climatic conditions. Capsicum, tomato, cucumber, brinjal, leafy vegetables and many other tropical vegetables are cultivated in these countries using PA and it was observed higher yield when compared to open field conditions [4,6].
PA was introduced to Sri Lanka somewhere around 1987 and it has been adopted mostly by the middle class farmers [8].
Main purposes of introduction of PA farming are to minimize the impacts of erratic weather and provide optimum environmental for horticultural crop production as well as attraction of younger generation for agricultural sector with modern technologies (2,9).
PA is currently practiced mainly in up country and low country wet zone areas and very less in dry zone areas. Most commonly cultivated vegetables are pepper, tomato, cauliflower, lettuce, broccoli and Japanese cucumber. Frequently used PA structures in the country are net houses, poly tunnels and rain shelters [9].
Niranjan [8] mentioned that PA has a high potential for expansion into different agro ecological regions in the country with suitable protected structures for cultivation of vegetables. However, due to the requirement of high initial capital and technical knowledge limited number of farmers adapted to PA technologies with their own funds.

Theoretical Framework of Study
Adoption may be defined as the integration of an innovation into farmers' normal farming activities over an extended period [10]. Various theoretical perspectives explained the behavior of adoption on new technologies. Innovation diffusion theoretical perspective, the economic constraints theoretical perspective and the adopter perception theoretical perspective are three different theories commonly used to explain farmers' adoption behavior and factors affecting to technology adoption in previous studies [10,11].
Information dissemination is identified as a key factor influencing adoption decision in innovation diffusion theoretical perspectives.
The adoption decision process is viewed as a series of linear stages starting by the knowledge stage followed by a persuasion, decision, implementation and finally confirmation stages [11]. The economic constraints theoretical perspective states that adoption is influenced by economic factors and economic constraints due to the asymmetric distribution of resources. Lack of ownership and access to capital constrain farmers from adopting new innovations [12]. The adopter perceptions theoretical perspective identifies farmers' perceptions as the key to the adoption of a particular farming technology [11]. This study integrated these three theories in to developing a conceptual understanding of the adoption of the PA technologies by farmers.
There is vast array of literatures on factors that determine agricultural technology adoption. Factors affecting to technology adoption is categorized in different ways in previous studies and there is no clear distinguishing feature between variables in each category. Categorization of factors is done according to the investigating technology, the study location and researchers' preference [13]. Kabir [14] categorized the determinants of technology adoption in to four broad categories of economic, social, institutional and management. Howely [15] categorized the factors affecting farmers' adoption to new technology in to a personal, social, cultural, institutional and economic factors as well as on the characteristics of the technology. In another studies factors influencing on agricultural technology adoption were categorized into technological, economic, institutional and household specific factors [13,16]. In previous studies, farmers age, educational level, gender, family size, farming experiences, farm size, primary Therefore, hereafter in this study PA term is used to denote rain shelters and polytunnels. Rain shelter is a simple form of protected structure made with Galvenized Iron (GI) pipes and only roof of the structure is covered with water resistant low density polyethylene film. It is mainly used to minimize the impact of heavy rainfall on cultivated crops and partially controlled the evapotranspiration by acting like a roof. Poly tunnels are relatively high tech protected structures and frame of structure constructed with GI pipes.
The structure is fully covered with UV stabilized transparent polyethylene films. Microclimate within the structure is partially or fully controlled in this structures. The study sample was divided into two groups of farmers, those who adopted (adopters) vegetable cultivation under PA and those who have not been adopted the PA but practicing conventional vegetable cultivation within a same locality. Seventy farmers who have been practiced vegetable cultivation under PA and fifty conventional vegetable farmers were selected randomly for data collection. Total sample size was 120 including both PA and conventional vegetable farmers.
To achieve the objective of the study primary data was collected through household questionnaire survey using pre tested semi structured questionnaire. Two separate semi structured questionnaires were used to collect data from adopters and non-

Analytical Framework: Binary Logistic Regression Model
Logistic regression model is one of the best analytical tool commonly used for analyze the farmer adoption decisions on new technologies [11,21]. Since the adoption of PA is a dichotomous or binary dependent variable, with the option of either adoption or rejection, the binary logistic regression model was applied [12]. It explores the degree and direction of relationship between independent and dependent variables in the adoption of PA at the household level. The dependent variable in this study was the farmer being an adopter or non-adopter of PA technology.
Accordingly, adoption of PA technologies in dry zone assumed to be influenced by set of independent variables and is specified as follows.
Where the subscript i means the i th observation in the sample. P is the probability of that a farmer adopts the PA and (1-P) is the probability that a farmer does not adopt PA. is the intercept term and , …. and are the coefficients to be estimated related to independent variables X 1 , X 2 ,…….,X k . Based on literature, scope of the study and the availability of data following independent variables were selected for analysis * (Table 1).

Summary of socio demographic characteristics of respondents
Socio demographic characters of sample farmers show slight variations between adopters and non-adopters (

Factors affecting to adoption of PA technologies
Results for the Tobit model was significant at the 0.05 percent level based on a model chi square statistics. Significant loglikelihood and LR Chi-square values imply that the model fitted well and the explanatory variables used in the model are collectively able to explain the determinants of adoption of PA technologies by dry zone vegetable farmers.  Household size significantly relates with propensity to adopt PA practices at 10% level of significance. The result shows that farmers with small household will have high propensity to move for protected agricultural technologies. Arene [23] and Bello [24] reported a positive and significant relationship between adoption and household size. Labour has become one of the major constraint factors in Sri Lanka hence younger farmers with less number of family labour are more likely to transfer for less labour demanding technologies. However, in another study conducted by Alam [18] identified that household size did not affect toward the level of technology adoption.
The provision of support services, such as credit, has been shown to increase the adoption of improved technologies [25].
Farmers without cash and no access to credit find it very difficult to obtain and adopt new technologies. The results of this study indicate that the variable 'access to credit' has a positive and significant influence on the likelihood of adoption of PA, which is consistent with reports from previous studies [26,27]. As discussed earlier, not like in conventional cultivation adoption of PA technology needs high initial investment which prevents farmers from adopting it.
The variable that captures the subsidy effect on adopting is not significant in the analysis by emphasizing the need of increasing the availability of credit rather than promoting subsidies that have dead weight losses. This validates the condition that farmers with high accessibility to credit are more willing to use PA technologies.  [13].
The dummy variable that explains the type of primary employment of the farmer is significantly and negatively influencing on the decision of using protected agricultural technologies among the sample farmers. Farmers whose primary employment is agriculture are poor adopters than the farmers who are not completely rely on agriculture as income source. This could be dues