Socio-economic conditions of farmers, especially in the coastal region in Bangladesh, have been severely affected because of climate change. This study was focused on analyzing the farmers’ perception of climate change by examining three vital issues: (1) description of the socioeconomic characteristics of farmers; (2) reporting on the perception of farmers experiences with climatic change; and (3) identification of the socio-economic factors associated with farmers’ perception of climate change. The study area encompasses three villages within the coastal region (Sathkhira district) of Bangladesh, a geographic region where climate change literature has highlighted as prone to accelerated degradation. A logit model, along with weighted indexes for ranking and descriptive statistics, was used to analyze the result of 100 farmers surveyed by questionnaire. We found that the majority of the farmers (88%) perceived changes in climatic conditions. Almost all farmers indicated increases in temperature, droughts, floods, cyclones, salinity level and decreasing rainfall over the last 20 years. The logit model explained that out of the nine factors surveyed; education, family size, farm size, family income, farming experiences and training received were significantly related and influential factors to perception of climate change. Therefore, government and non-governmental organizations are recommended to push forward with interventions, especially focusing on identified factors, in order to strengthen the farmers’ capacity to battle against climate change effects.
Bangladesh is considered as one of the most vulnerable countries in the world to climate change because of its geographical location, economic dependence on agriculture, and recurrence of natural hazards [
The predicted climate variability of Bangladesh was developed by the National Adaptation Programme for Action (NAPA) in 2008, which was shown in
The coastal region of Bangladesh has been facing numerous ongoing climatic threats which have resulted in severe damage to agricultural production, while about 70 percent of its people depend on agriculture for their livelihoods [
The socio-economic conditions of the farmers in Bangladesh are some of the most vulnerable in the world to climate change. Therefore, adaptive processes to the effects of climate change are the priority for Bangladesh and crucial to build resilience into the lives of the farmers. But before adaptation, it is necessary to understand the farmers’ perception of climate change. Previous literature exist
Year | Temperature Change (˚C) Mean | Rainfall Change (%) Mean | Sea Level Rise (cm) | ||||
---|---|---|---|---|---|---|---|
Annual | DJF | JJA | Annual | DJF | JJA | ||
2030 | 1.0 | 1.1 | 0.8 | 5 | −2 | 6 | 14 |
2050 | 1.4 | 1.6 | 1.1 | 6 | −5 | 8 | 32 |
2100 | 2.4 | 2.7 | 1.9 | 10 | −10 | 12 | 88 |
Source: GOB, 2005 [
on the perception and adaptation to climate change in different vulnerable countries similar to Bangladesh, but very limited studies are available that identify the factors affecting the farmers’ perception of climate change. Therefore, this study addressed a gap in the research on specific factors and entitled “Determinants of Farmers’ Perception of Climate Change: A Case Study from the Coastal Region of Bangladesh”. However, the specific objectives of the research included describing the farmer’s socioeconomic factors, determining the farmers’ perception of climate change, and identifying significant variables to determine the farmers’ perception of climate change.
Perception is a process of receiving information and stimuli from our surroundings and converting them into psychological responsiveness [
Haque et al. [
A research study conducted by Maddison (2007) also revealed that adaptation of climate change has involved two processes, perception of climate change and then taking a decision, whether to adopt or not [
Tiwari et al. [
The study was carried out in three villages, within the coastal region of Bangladesh. This area was selected because it is prone to frequent natural hazards such as floods and cyclones. Face to face interviews were conducted with 100 farmers in May 2012 using structured questionnaire with both open and closed questions. Logistic regression, weighted indexes, and descriptive statistics were used to analyze the farmers’ responses to the questionnaire.
Perception is measured by a dummy variable in the model which was assigned a value of 1 for farmers’ who perceived climate change and a value of 0 for farmers who did not perceive climate change events. It indicated that the probability of an individual with a given set of attributes will fall in one choice (perceive) rather than the alternative (or not) but not both. Climate change events were defined in the questionnaire as increased or decreased temperature, rainfall, drought, flood, salinity, etc.
A logistic regression model was selected to identify the significant variables that determined whether farmers were perceptive of climate change, or not. The data could have been analyzed by different probability models where the dependent variable is a dummy. The models that were considered include the Linear Probability Model (LPM), logit model and probit model. Justification of the logit model was based on the following drawbacks of the LPM and probit model.
The LPM showed the uniformity of error terms and possibility of getting the probability function result out of 0 and 1. Due to this problem LPM is not logically attractive model for dummy responsive variables. It is better to use Cumulative Distribution Function (CDF) namely logit or probit models when analyzing this type of questionnaire [
The functional form of logit model was specified which is as follows:
For simplicity Equation (1) was expressed as
where,
Pi is the probability of perception of the farmers the ith respondent and it ranged from 0 - 1.
ezi: stands for the irrational number e raised to the power of Zi,
Zi: is a function of N-explanatory variables and expressed as:
where,
Therefore,
Prior to the estimation of the logistic regression model the explanatory variable were checked for the existence of multicolinearity. For this purpose, the presence of co-linearity was checked for categorical variables using contingency coefficient test. Therefore, age, number of plots and extension contact were omitted from the logistic model after the multicolinearity test.
The survey results for personal, economic and social characteristics of the respondents are shown in
The average farm size for the farmers surveyed was 1.57 hectare. This was nearly three times that of the national average farm sized of 0.6 hectare. The average annual income of the farmers’ in the study area was BDT 238,201 ($3043 US), which is more than the national average of BDT 112,240 ($1403 US) [
The sources of farmers’ income stated were mainly crop production, livestock rearing, fish culture and others non-farming activities (e.g. work other than agricultural sector). Non-farming activities include remittances, working at for government organizations and non-governmental organizations and off-farming activities, such as day labor, milling, and post-harvest operations. Shrimp culture proves to be a more profitable enterprise compared to other types of farming in Bangladesh. Since the farms are located in the coastal region (suitable for shrimp culture), this may have been a reason as to why the incomes of the farmers surveyed was nearly doubled the national average.
The majority of the farmers said they had received a line of credit previously and had access to the markets for selling their agricultural products. Just over
Farmers’ characteristics | Categories of the farmers | Scoring method | % respondents | Range | Mean | Std. deviation | |
---|---|---|---|---|---|---|---|
Min | Max | ||||||
Age | Young (18 - 35) | Years | 28 | 28 | 65 | 41.77 | 8.433 |
Meddle aged (36 - 50) | 58 | ||||||
Old aged (>50) | 14 | ||||||
Farming experiences | Low (Up to 13 years) | Years | 20 | 5 | 45 | 19.44 | 7.811 |
Medium (14 to 26 years) | 67 | ||||||
High (Above 26 years) | 13 | ||||||
Training received | Training received (1) | Dummy | 40 | ||||
Not training received (0) | 60 | ||||||
60 | |||||||
Education | Illiterate (0) | Years of schooling | 27 | 0 | 18 | 6.16 | 4.929 |
Primary (1 - 5) | 27 | ||||||
secondary (6 - 12) | 39 | ||||||
Higher studies (13 & above) | 7 | ||||||
Family size | Low (Up to 4) | Number | 45 | 1 | 13 | 4.93 | 1.827 |
Medium (5 to 6) | 48 | ||||||
High (Above 6) | 7 | ||||||
Farm size | Landless & Marginal (Up to 0.2 ha) | Hectare | 2 | 0.18 | 20 | 1.57 | 2.638 |
Small (0.21 - 1 ha) | 63 | ||||||
Medium (1.1 - 3.0 ha) | 25 | ||||||
Large (3.1 ha and above) | 10 | ||||||
Number of plots | Low (Up to 5) | Number | 60 | 1 | 27 | 6.35 | 5.170 |
Medium (6 - 10) | 28 | ||||||
High (above 10) | 12 | ||||||
Family income | Low Up to (85000) | 1 BDT = 0.01$ | 36 | 46,600 | 5500,000 | 238,201 | 276,737 |
Medium (85001 - 200000) | 41 | ||||||
High (above 200000) | 23 | ||||||
Credit received | Credit received (1) | Dummy | 75 | ||||
No credit received (0) | 25 | ||||||
Market access | Market access (1) | Dummy | 76 | ||||
No market access (0) | 24 | ||||||
Cooperative involvement | Involved cooperative (1) | Dummy | 51 | ||||
Not involved (0) | 49 | ||||||
Extension contact | Low (up to 11) | Scale | 21 | 6 | 38 | 19.09 | 8.269 |
Medium (12 - 22) | 42 | ||||||
High (23 and above) | 37 |
Source: Author’s field survey.
half of the farmers of the study area (51%) stated involvement in cooperatives. The highest proportion of the respondents (79%) stated medium to high extension services contact, while only 21% of the respondents reported low contact. These questions were important because farmers require diverse information ranging from soil and water conservations techniques, production procedures, marketing system, sustainable agriculture, environmental sustainability issues that affect the farmers’ perception of climate change.
The majority of the farmers surveyed reported being within the ages of 18 to 50 (86%). Similar findings were found regarding age by Alam; Khan [
On average the farmers reported about 20 years farming experiences of the farmers. This was important to understand that the farmers surveyed have sufficient experience with the changes in climate and weather patterns and the subsequent effects on their decision making and output. It was found that experienced farmers will respond better regarding the climate change events specifically the changing pattern of temperature, rainfall, occurrences of natural hazards such as floods, cyclones, droughts, salinity etc. [
The respondents of the study were asked a dichotomous (“yes/no” response) question about whether or not they had experienced changes in the climate of the region within the past 20 years. After their initial response, the farmers were asked about their perceived experience in relation to a series of climatic events commonly associated with global climate change effects in Bangladesh (according to the literature reviewed). They could respond selecting the following; experienced decreases, increases, no change, or they did not know, in the occurrence of the event.
Climatic event | % of Respondents | |||
---|---|---|---|---|
Increased | No change | Decreased | Don’t know | |
Temperature | 100 | |||
Rainfall | 3.4 | 96.6 | ||
Occurrence of drought | 100 | |||
Occurrence of flood | 100 | |||
Occurrence of cyclones | 100 | |||
Salinity level | 100 | |||
Short winter season | 85.2 | 4.6 | 10.2 | |
Long summer season | 92 | 2.3 | 5.7 | |
Unpredictable rainfall | 90.9 | 1.1 | 8 | |
Changes of monsoon season | 80.7 | 2.3 | 17 |
Source: Author’s field survey data.
flooding, cyclones, and soil salinity. Across all events, at least 80% or more reported having experienced climatic shifts which are likely to have a negative impact on agricultural activity. While it is clear that these are perceptions of the farmers surveyed, not calculated events, such information provided important input from the farmers.
The majority of farmers perceived an increased trend of repeated short winter seasons, long summer seasons, unpredicted rainfall and changes of the monsoon season. Increasing temperature along with decreasing precipitation may enhance the water scarcity resulting drought, which, in turn, may affect crop production output. These results may also prove vulnerable conditions of the coastal area in Bangladesh due to the climate change effects. The studies conducted by Dhaka et al. [
The contingency coefficient test was applied before the data analysis to diagnose colinearity and omit independent variables that were highly dependent and strongly correlated to each other, see
Generally, it is predicted that there should be a positive relationship between family income and farm size. Therefore, both were considered in the logit model reported here, instead of excluding them from the analysis. The model was run with these items omitted and the econometric estimates in those simulations were found to not have significantly changed from the model which maintains family income and farm size. Only age, extension contact, and the number of plots are omitted from the logistic regression model in determining factors affecting the farmers’ perception of climate change and shown in
The logistic regression model results (
Variables | AG | EDU | FAMSZ | FARSZ | NUMP | FAREX | FAMIN | CRRE | TRRE | COPIN | MARAC | EXCONT |
---|---|---|---|---|---|---|---|---|---|---|---|---|
AG | 1 | |||||||||||
EDU | −0.046 | 1 | ||||||||||
FAMSZ | 0.368 | 0.114 | 1 | |||||||||
FARSZ | 0.213 | 0.393* | 0.354* | 1 | ||||||||
NUMP | 0.181 | 0.35* | 0.382* | 0.853** | 1 | |||||||
FAREX | 0.887** | −0.114 | 0.360* | 0.163 | 0.162 | 1 | ||||||
FAMIN | 0.195 | 0.344* | 0.229 | 0.893** | 0.639** | 0.131 | 1 | |||||
CRRE | 0.223 | 0.113 | 0.117 | 0.202 | 0.254 | 0.178 | 0.139 | 1 | ||||
TRRE | 0.095 | 0.331 | 0.143 | 0.344* | 0.341* | 0.009 | 0.245 | 0.363* | 1 | |||
COPIN | 0.048 | 0.506* | 0.158 | 0.347* | 0.36* | −0.040 | 0.237 | 0.242 | 0.343* | 1 | ||
MARAC | 0.054 | 0.495* | 0.171 | 0.236 | 0.288 | 0.034 | 0.166 | 0.270 | 0.188 | 0.410* 1 | ||
EXCONT | 0.032 | 0.756** | 0.164 | 0.457* | 0.446* | 0.018 | 0.391* | 0.155 | 0.355* | 0.526** | 0.398* | 1 |
*Weak co-linearity between the two variables; **High co-linearity between the two variables.
Variables | Perception | ||
---|---|---|---|
Coefficient | Robust Std. Error | P Value | |
CONS | 0.1051 | 4.1100 | 0.693 |
EDU | 0.6520** | 0.30585 | 0.033 |
FAMSZ | −0.6830* | 0.3794 | 0.072 |
FARSZ | −3.1210** | 1.6065 | 0.052 |
FAMIN | 0.0001** | 0.0000 | 0.031 |
FAREX | 2.3668* | 1.3947 | 0.120 |
CRRE | −1.354 | 2.1237 | 0.512 |
TRRE | 6.861* | 4.3280 | 0.091 |
COIN | 6.929 | 3.9734 | 0.112 |
MARACC | 0.0913 | 2.092 | 0.789 |
R2 | 0.786 |
**, *indicate significant level at 5% and 10% respectively.
implies that the probability of perception of climate change is greater for those who have higher educational attainment compared to less-educated or illiterate farmers. It is apparent that educated farmers have more knowledge, ability to understand and respond to expected changes, able to forecast future scenarios and have greater access to information and opportunities than others. These issues lead to the farmers who perceive more about climate change. Education as an influencing factor of farmers perception of climate change was also found in studies conducted by other researchers [
Family size was negative and significant (at the 10% level) when related to farmers’ perception of climate change. However, the negative sign on this relationship was contradictory to what the researchers would have thought. These findings indicated that with increasing size of the family, the probability of farmers’ perception of climate change decreased. Prior to this study, it was expected that the sign of the variable family size would have a positive, the logic being that large family size makes more interaction among the family members, which increases the perception of climate change. Moreover, the results indicated that larger family size had less probability of perceiving of climate change than smaller family size. It is may be the larger family numbers interact less with each other, have less access to extension contact, are unable to attend training programs, and/or rather act as labor forces. Similar findings have found in other studies [
There was a negative and significant (at the 5% level) relationship between farm size and perception of climate change. Specifically, results show that the larger the size of a farm operation decreased the probability of farmers’ perceiving climate change. Larger farms require greater levels of investment and production inputs such as seeds, fertilizer, pesticides, irrigation facilities, which are stressors on farm budgets. To utilize these inputs require more education, experience and managerial capacity which may influence the farmers’ perception of climate change. The potential explanation may be that all farmers have the potentiality but may be lack of proper education, training, poor communication exposures and fail to perceive more. The similar results revealed studies conducted by the others [
The result of the logic model shows that positive and significant (at the 5% level) relationship between family income and farmers’ perception of climate change effects. This implies that farmers with high income are more likely to have the more access to resources than farmers with lower incomes, which include trainings about the effects of climate change. The Government Organizations (GOs) and Non-Government Organizations (NGOs) have programs designed to create off-farm livelihoods activities which result in increased income and continued agricultural operations in the face of climatic uncertainty. Other sources of off-farm income, such as remittances and off-farm jobs might influence to farmers’ perception. Semenza et al. (2008) [
Farming experience was found positive and significant (at the 5% level) relationship with farmers’ perception of climate change, as confirmed by the logistic regression model. Experienced farmers were more aware in changing temperature, rainfall, and other disaster events. These experiences might be helpful to understand the prediction of future changes of these events and have been identified in other research [
The results show that positive and significant (at the 10% level) relationship between training received and farmers’ perception of climate change. Any inclusion of training reported by the interviewed farmers helped them to identify climate change events, and can help the farmers more adept at handling tasks to prepare for climate change event. Similar results found by the conduction of another researcher with regards to cocoa farming and farmers’ perception on climate variability [
The findings from this study revealed that the majority (88%) of the farmers in the study area perceived changes in climatic conditions, whereas only 12% did not. Almost all farmers reported increases in temperature, droughts, floods, cyclones, salinity level and decreasing the rainfall over the last 20 years. Increasing temperature along with decreasing precipitation may enhance the water scarcity from resulting droughts which will affect crop production. The logit model explained that out of the nine factors surveyed; education, family size, farm size, family income, farming experiences, and training received, were found to be significantly related to the farmers’ perception of climate change and indentified as influential factors of farmers’ perception of climate change. Respective authorities, especially government and non-government organizations, should create policy measures that consider these influential factors of farmers’ perception of climate change. This, in turn, may have a significant contribution to farmers’ reducing the risks that they lose against climate change effects. The policy measures may be focused on capacity building of the farmers, institutional supports, easy way-out of receiving support from the concern authorities, ensuring the accountability of the supportive staff who have been working with farmers intensively.
The authors are acknowledged to the farmers from the study area for their contributions, Professor Dr. Md. Abdul Momen Miah, Professor Dr. M. Asaduzzaman Sarker and Mr. Maruf Billah (BAU, Mymensingh, Bangladesh) for their kind cooperation about data collection of this study. The paper was based on the first author’s graduate study at Humboldt University of Berlin (HUB), Germany under Erasmus Mundus Scholarship funded by European Commission (EC).
Mohammed Nasir Uddin is the primary author. He designed the field study, implemented data collection, and selected and implemented the analytical methodology. The work presented here is an extension of that completed for his master thesis, successfully defended for the International Master of Rural Development with its Secretariat at Universiteit Ghent, Belgium.
Wolfgang Bokelmann was the faculty advisor and thesis promoter of Mohammed Nasir Uddin. He contributed to the project inception, selection of the methodological framework, and writing of the thesis document upon which this article was based in an advisory capacity.
Emily S. Dunn edited and contributed to the style, contents and overall flow of the paper. She is an Instructor in the Department of Agribusiness within the College of Food and Agriculture in the United Arab Emirates University.
The authors declare no conflict of interest.
Uddin, M.N., Bokelmann, W. and Dunn, E.S. (2017) Determinants of Farmers’ Perception of Cli- mate Change: A Case Study from the Coastal Region of Bangladesh. American Journal of Climate Change, 6, 151-165. https://doi.org/10.4236/ajcc.2017.61009