2.1. Description of the Study Sites The Bengal Delta, also known as the Ganges-Brahmaputra-Meghna (GBM) Delta, located in the northern shores of the Bay of Bengal, is one of the populous deltas of the world. This delta has a unique ecosystem characteristic comprising the three mighty river systems (Ganges-Brahmaputra Meghna) enclosed by the terrestrial, aquatic and marine ecosystems. The city Khulna is located in the lower part of the Bengal Delta frequently stressed by tidal surge-related inundation, salinity intrusion, tropical cyclone and hydro-climatic variabilities. The city is highly dependent on peri-urban agriculture for its food supply. Khulna is also a regional food supply hub, though this region of Bangladesh is particularly vulnerable to the effects of climate change. The city’s importance in regional food supply and the vulnerability of its farming communities to climate change impacts were key reasons for its selection for this study. Indeed, Khulna is a zone of multiple vulnerabilities as well as opportunities. It is the third-largest (64.78 km2) metropolitan area in Bangladesh. Khulna district has 9 Upazilas (sub-districts) and about 2.3 million inhabitants. The climate in Khulna is subtropical warm and humid, with four distinct seasons: (i) dry winter (December to February), (ii) hot pre-monsoon summer (March to May), (iii) rainy monsoon (June to September) and (iv) post-monsoon autumn (October to November).
We obtained climatological data from the Bangladesh Meteorological Department (BMD). According to this data, over the 1948–2018 period the average annual rainfall was 1752.3 mm and the mean annual air temperature was 26.7 0C (see Supplementary Material A). The average monthly minimum and maximum temperatures were 21.9 and 31.3 0C, respectively. January was the coldest month, with a mean minimum temperature of 12.9 0C. April was the warmest month, with a mean maximum temperature of 34.9 0C, with an average annual rainfall anomaly of ±48.6% to 35.5% from 1981 to 2014. According to the literature, some 80% of precipitation occurs in the monsoon season from May to September [51–53]. However, our data for 1948–2018 indicate that the monsoon season extended from May to October, with some 90% of the total annual rainfall occurring during these months. The highest and lowest rainfall quantities were found in July and December, with monthly averages of 327.6 mm and 4.5 mm, respectively. Due to the abundance of rainfall in the region, the area offers excellent opportunities for rainfed agriculture.
For our study, we selected two sub-districts: Batiaghata (~248 km2 ) and Rupsa (~120 km2 ). Major crops grown here were paddy, jute, sesame and vegetables, with small-scale aquaculture-agriculture also observed year-round. Farmers grew various short-term crops and vegetables as well, such as beans, gourds, eggplants and tomatoes, in integrated aquaculture-agriculture farming systems. Integrated farming systems have been found to provide greater economic returns than paddy or vegetable monocrops. Additionally, [55] found that an integrated farming system consisting of paddy, vegetables and aquaculture was more resilient to recurrent hydro-climatic variability.
2.2. Site Selection, Data Collection and Analysis This study used an explorative research framework combining desk research, secondary data collection, field visits, focus groups and expert interviews. A mixed-method approach for this study allows us to collect information and triangulation through different participatory tools and approaches. We focused on six peri-urban villages within Rupsa and Batiaghata sub-districts, where we carried out 200 farmer interviews, 4 focus group discussions and 20 expert interviews from August 2017 to May 2018. Questionnaires (Supplementary Material B) and checklists (Supplementary Materials C and D) guided primary data collection through the interviews and focus groups. Secondary data were collected from documentary sources: journals, reports from the Bangladesh Bureau of Statistics (BBS) and unpublished documents from DAE district and sub-district offices regarding crops, population and livelihood characteristics of farm households in the study area.
Three of the surveyed villages—Sreefaltala, Domra and Peyara—were located in Rupsa, and the remaining three villages—Jharbhanga, Raingamari and Sanchibunia—were in Batiaghata. In Bangladesh, the lowest administrative jurisdiction of government is called a “union”. Unions are divided into agricultural blocks, each of which has its own Sub-Assistant Agriculture Officer (SAAO). The SAAO, working under the DAE, provides extension services to some 2000 farm households. All of the selected villages exhibited peri-urban characteristics; that is, they were located in close proximity to the Khulna metropolitan area, and were highly interdependent and interconnected with the city. After site selection, data on information use in agricultural decision-making were gathered through focus group discussions and farmer and expert interviews, guided by checklists and questionnaires.
Farmer interviews. Two hundred farmers were interviewed. This sample size was determined following Berenson and Levine (1992) to obtain a 95% confidence level based on the total population of households (858). In total, 62 households were selected from Batiaghata and 138 from Rupsa. With the interviews, we sought to obtain a quantitative overview of the farming communities and farming practices, including farmers’ access to and the quality of information sources and their interest in and need for expanded information services for agricultural decision-making. We designed a semi-structured questionnaire to guide the farmer interviews. These were informed by consultations with experts, field visits and random farmer interviews carried out ahead of time. Furthermore, the questionnaire was pre-tested with 10 randomly selected farmers at both sites. Based on the pre-test, we made minor changes to the questionnaire. Farmer interviews were conducted using the open-access online interview tool KoBoToolbox (www.kobotoolbox.org). Two female master’s degree students conducted the interviews. In Bangladeshi society, women have easier access to unknown households than men. It took two months to complete the interviews. Simple random sampling was used to select interviewees. Finally, SPSS Statistics 20 software was used to analyze and interpret the quantitative data obtained.
Expert interviews. We interviewed 20 experts using open-ended questionnaires. Interviews focused on the extension services currently provided, limitations of existing information and extension services, need for hydro-climatic information for agricultural decision-making and current agricultural decision-making practices of farmers in Khulna. Several more specific topics were also discussed with the experts, such as how farmers dealt with hydro-climatic information and what role available information sources and quality played in agricultural decision-making. Finally, the interview results were summarized and entered into Excel for analysis and interpretation.