Factors and practices: farmers' adaptation to climate change in Bangladesh

Climate change is a well-acknowledged phenomenon that is linked to frequent extreme weather occurrences and has far-reaching effects on various dimensions of the environment. Among the many far-reaching consequences of climate change, agriculture is suspected to be highly impacted because of its high dependence on the environment (Mendelsohn & Davis 2001; Stern 2006; Howden et al. 2007; IPCC 2007; Karl et al. 2007). Scientific consent also provides enough clarity on the impending highly negative consequences of climate change on the agricultural sector specifically agriculture in the least developed and developing countries (Howden et al. 2007; Parry et al. 2004; Schellnhuber et al. 2013; Wheeler & Von Braun 2013; Bandara & Cai 2014; Kahsay & Hansen 2016; Myers et al. 2017). Additionally, the global trend of climate change impact, particularly in the Asian region, is predicted to be worse more clearly with shooting high temperatures over continental interior Asia. Other robust findings through analyzing data from 1990 to 2005 indicate a significant increase in precipitation in northern and central Asia and a reverse (drought condition) in southern Asia (Mannig et al. 2017; Dimri et al. 2018). Considering this, agriculture as a climate-sensitive sector, is more likely to experience production losses. Moreover, an estimated, approximately 2.5 billion, number of rural livelihoods directly or indirectly related to agricultural production arrangements are also going to be threatened (Mendelsohn & Davis 2001; Jones & Boer 2003).

Studies also confirm that the absence of adaptation measures, which helps to curb negative effects on a farm level, brings harmful consequences due to climate change. But a great sense of optimism is still left as there are various potential adaptation choices at hand for existing agricultural systems. One thing to remember is that the extents of detrimental effects on the agricultural sector because of climate change are dependent on farming communities' adaptive capacities (Downing 1991; Rosenzweig & Parry 1994; Smith & Lenhart 1996; Mendelsohn & Dinar 1999; Smit & Skinner 2002; Howden et al. 2007; Gbetibouo 2009).

A forecast from IPCC proclaims that in facing climate change, developing countries, like Bangladesh, in upcoming decades will continue to face extreme weather events, notably water shortages, temperature rise, and exuberant rainfall causing floods and other climate events (IPCC 2001, 2007). Moreover, in Bangladesh, the agricultural sector is most likely to experience a significant reduction in crop yield on various occasions of climate variability (Islam et al. 2011). On top of that, a significant mass of research claims that in Bangladesh, due to climate change, negative environmental impacts are already taking place and temperature is predicted to rise by 1.0, 1.4, and 2.4 °C by 2030, 2050, and 2100, respectively (Agrawala et al. 2003; Ahmed 2006; Ruhul Amin et al. 2015). Within the same time period, research works also predicted a projected rise in sea-level water by 70 cm (60–80 cm) and 105 cm (85–125 cm). As a consequence, all of these incidents will culminate in tidal surges and extreme tidal events, exacerbating the risk of flooding in coastal regions or near-sea rivers (Brakenridge et al. 2013; Schellnhuber et al. 2013). Moreover, erratic changes in precipitation have also been predicted such that monsoon seasons will be expected to have more precipitation while other seasons will endure prolonged dry seasons. Due to this event, there will be severe floods during the monsoon and extreme drought throughout the other seasons (Ruhul Amin et al. 2015; Islam & Hasan 2020).

In this context, farm-level adaptation is quite critical to combat climate change. A significant body of research in Bangladesh focuses on their efforts to farm-level adaptation and overcome limitations, particularly the issue of salinity in the south-west coastal region and drought-prone regions in the north-west (FAO 2006; Kainan Ahmed et al. 2006; Habiba et al. 2012; Alauddin & Sarker 2014; Hassan et al. 2014; Vivekananda et al. 2014; Alam 2015; Khan et al. 2015). Additionally, a considerable portion of research has also identified a number of adaptation strategies including development of adverse environment-resistant crop variations, advanced crop management, enriching human resources with education, diversification of livelihood, easy access to physical resources, and financial resources with market and institutional flexibility. Moreover, research works have also acknowledged the importance of adaptation research to strategize and invocation of fruitful future policy formulation. Incorporation of technology in agriculture is also important to adapt with climate change, particularly through advancement in bioengineering, advanced irrigation systems, and nanotechnology are some of the few aspects of high-tech farming (Deressa et al. 2009; Alauddin & Sarker 2014; Alam 2015; Abdul Rajak 2022).

Given the upcoming climate change scenarios, adaptation in agriculture, particularly crop agriculture, is ubiquitous in order to function in a hostile climatic environment. In Bangladesh, several studies have been conducted in drought-prone regions to understand adaptation mechanisms (Alauddin & Sarker 2014; Alam 2015), and one particular study of Saha et al. (2016) explored adaptation practices in coastal regions. From this understanding, this research seeks to identify the primary socio-economic factors influencing farmers' crop adaptation decisions. Furthermore, there is no clear evidence on how willing farmers are to accept new agricultural adaptation practices. Taking all of this into account, this study seeks to examine the factors that influence the selection of existing agricultural adaptation practices in facing the climate change crisis.

In order to outline this research, the following sections provide a detailed explanation and structure of the study. Section 2 describes the rationale behind choosing the study area, sampling procedure, variables selection, and model specification in detail. The subsequent Section 3 elucidates the result based on the field data and the following discussion in Section 4 compares and corroborates similar findings around the study objectives. Sections 5 and 6 describe the policy implications and the limitations of this study respectively. Finally, the last section summarizes the study findings and policy implications while outlining the shortcomings succinctly.

From Equation (1) sample size (n) was calculated at 96 from each union culminating in the final sample size at 192 from two unions (study area). After determining the sample size, data have been gathered using a structured questionnaire. Each participant in this study was exposed to a set of structured questionnaires regarding their demographic information and information related to adaptation practices and procedures in facing climate change. The questionnaire was split up into two sections. The first section had a total of 15 questions concerning respondent's demographics, farming income, experience, land size, number of cultivated plots, family labor in agriculture, credit access, damaged sector and damaged amount due to climate change, and training on climate change adaptation, membership in a local organization, and finally, received assistance from the local NGO/government. The second section had 12 questions on five different adaptation options, namely crop management, water management, farm management, physical infrastructure management, and social practices. Various sub-categories of these five broad adaption measures were disclosed to understand the most practiced adaptation options. In terms of assessing the validity of the initially prepared questionnaire first, independent consultation with an expert was arranged. Secondly, from the initial consultation, various important points emerged and a second review of the questionnaire was evaluated with a group of three experts in connection with the first expert. Finally, the questionnaire was adjusted a little bit from the field experience.

The result from this index was considered as the dependent variable which reflects the range of strategies adapted. Some strategies, such as soil conservation and use of farm related infrastructure (like, temperature and moisture sensors, aerial images, and GPS technology) were observed missing in adaptation practices. Nevertheless, the composite index aids understanding of the variation in strategy adaptation and its causes, contributing to the formulation of policies for effective implementation of climate change adaptation options (Hassan & Nhemachena 2008; Deressa et al. 2009).

In order to establish a logical relationship with the various adaptation measures, disclosed in this study, initially 14 variables were considered as independent variables for stepwise regression analysis. Moreover, to maintain the explanatory power and sensitivity of the constructed model only the significantly correlated variables with the dependent variable were included in the final model. All 14 independent variables were entered step by step into the model. Finally ten variables were dropped because they had non-explanatory power, and four variables, with significant explanatory power were retained.

In order to fit a linear model to the test, it is necessary to check the predictors for collinearity, which determines whether or not the predictors estimated in the model are dependent on each other. In this regard, the variance inflation factors (VIFs) for the coefficients are a better diagnostic for determining whether coefficients are poorly estimated due to collinearity (Tomaschek et al. 2018). In the case for our selected independent variables, the VIF value was <1.05 confirming the non-existence of multicollinearity. According to Sheather (2009) and Chatterjee & Hadi (2006), the boundary value for the collinearity problem ranges between 5 and 10. Additionally, we have also tested the condition index representing value <8, where multicollinearity exists when condition indices are higher than 10 to 30.

The model was constructed using the stepwise probability criteria of F to enter p ≤ 0.050, and probability of F to remove is ≥0.100.

This study has used a binary logit model to analyze the various factors affecting farmers' decision in accepting new adaptation strategies in facing severe weather episodes in agricultural production. Farmers' decision in accepting new adaptation strategies was explored in a discrete choice form (where 1 = yes, 0 = no). More lucidly, one (1) denotes farmers who want to adapt new strategies to climate change in their agricultural production. By contrast, zero (0) denotes farmers who do not want to adapt new strategies to climate change. This research anticipated that a variety of factors influence farmers' decisions to adapt a new strategy to climate change in agricultural production.

Most of the data analysis has been performed utilizing R version 4.1.0 and the graphical presentation of maps was done using ArcGIS, version 10.3.

One of the goals of this research was to better comprehend the factors that influence farmers' climate change adaptation strategies, as reflected in the climate change adaptation index (Table 1). Results from correlation (Table 2) show that few factors are responsible for farmers' adaptation practices. Positive significant factors like income (p = 0.034), damaged sector due to climate change impacts (p = 0.015), family labor involved in agriculture (p = 0.000), credit access (p = 0.002), and assistance from an NGO/government (p = 0.047) influence farmers' adaptation choices. But gender (−0.243, p = 0.000) and farm size (−0.138, p = 0.028) have significantly negatively correlated and had no impact in choosing adaptation strategies.

As explained previously in methodology, among 14 independent variables only 4 variables (gender, family labor in agriculture, farming experience, and damaged sector of agriculture) significantly explained the variation in deciding multiple climate change adaptation strategies (Table 3). Analysis unfolds that both multiple R and R² values have increased with the addition of each independent variables from X₁ to X₄, and have reasonable explanatory power on the models. The final model, with four independent variables, has a significantly high level of explanatory power, compared to the other three variables, as reflected in the adjusted R² = 15%. The model has a high statistical significance with a minimum error of the estimate, and it can partly explain the study area's environment in terms of the application of adaptation of climate change strategies in a complex socio-economic and socio-cultural perspectives.

Table 3

Inclusion of the significant variables in the model and the summary of the stepwise regression models

Model .	.	.	Adjusted .	Std. error of the estimate .
1	0.243a	0.059	0.054	0.339
2	0.328b	0.107	0.098	0.331
3	0.378c	0.143	0.129	0.325
4	0.410d	0.168	0.15	0.321

Model .	.	.	Adjusted .	Std. error of the estimate .
1	0.243a	0.059	0.054	0.339
2	0.328b	0.107	0.098	0.331
3	0.378c	0.143	0.129	0.325
4	0.410d	0.168	0.15	0.321

Source: Field survey, 2021.

^aPredictors: Gender.

^bPredictors: Gender, family men involved in farming.

^cPredictors: Gender, family men involved in farming, farming experience.

^dPredictors: Gender, family men involved in farming, farming experience, damaged sector.

View Large

From Table 4, the F ratio of explanatory variables in the final model is statistically significant at 0.001. The significant value of the F ratio explains that the variables included in the model are appropriate. Table 5 describes that the four variables significantly explaining the variations and among all the variables only gender, family labor in agriculture, farming experience, and damaged sector of agriculture appear to be the most influential factors, explaining nearly 60% of total variation. The result further demonstrates that the included factors in the model play influential roles in deciding adaptation strategies to climate change in a smallholder agricultural system. Regression analysis showed that all variables, except gender and farming experience, have positive effects on adaptation choices. An increase of agricultural labor and increase of damage lead to a positive move toward implementing climate change adaptation strategies by (0.343) and (0.16) times, respectively. But changes in gender and farming experience negatively influence adaptation of various climate change strategies (−0.203) and (−0.227).

Table 6 reports and explains farmers' intention to accommodate new adaptation practices and whether or not adaptation decisions are influenced by socio-economic and demographic characteristics. If new adaptation measures were introduced to them to combat the climate crisis, their inclination to adapt those measures are also critical to assess. In this regard, respondents were asked the opinions of whether they would implement new adaptation practices or not. All the significant variables previously explored in the multivariate linear regression has also been used in this stage too. Results from the binary logistic regression shows that every increase in the unit of age (p = ≤ 0.01), income (p = ≤ 0.01), and farm size (p = ≤ 0.005) influence to accept new adaptation practices. Every unit of income increased shows that farmers are 2.65 times more likely to go for new adaptation options. Similarly, an increased unit of age of the farmers also showed that they are 2.2 times more likely to adapt new adaptation measures. Accesses to credit and received assistance from an NGO/government do not have any significant contribution in choosing new adaptation measures.

In this study, farmers' agricultural adaptation measures, as well as related factors in determining adaptation measures were examined. According to the study findings, farmers use a wide range of adaptation mechanisms, both knowingly and unknowingly. Findings suggest that various demographic and socio-economic characteristics contribute significantly to the various adaptation measures of the farmers. In the face of climate change, gender, household income, and damaged sectors all have significant associations with adaptation practices in order to maintain some adaptation measures such as water, crop, farm, and infrastructure management. Literature also confirms the significance of demographic and socio-economic characteristics in the adaptation decisions (Deressa et al. 2009; Apata 2011; Ndambiri et al. 2012). Moreover, studies by Hassan & Nhemachena (2008) and Ishaya & Abaje (2008) also showed a significant relationship with adaptation practices with gender, credit access, farming experience, farm size, and other socio-economic characteristics. According to the findings, family labor is significantly correlated with adaptation decisions. Because labor is an essential component of any practical endeavor, household labor availability appears to be an important variable. Moreover, linking this to agriculture the scenario is that households with more labor are better prepared to take on various adaptation strategies when compared to those with limited labor (Ishaya & Abaje 2008; Yila & Resurreccion 2013). A study by Deressa et al. (2009) also finds the importance of family labor in agricultural adaptation practices. Adaptation to climate change largely relies on the availability of the labor force. Other variables, such as membership in local organizations, credit accessibility, adaptation training, and damage amounts, were not significantly associated with climate change adaptation practices. Similar findings were also discovered in Nepal and Myanmar (Piya et al. 2013; Ahmad & SeinnSeinn 2015; Thoai et al. 2018).

The regression analysis results show that a household head would be able to manage all adaptation measures more effectively if she had more family labor, more farming experience, and full consideration of climate change impacts on specific sectors. Additionally, adequate household labor is effective for farm management because necessary adaptation strategies, despite their willingness, are hard to implement and beyond capability when households do not have necessary labor. In some cases, inadequate family labor forces farmers to opt for employing wage laborers. However, smallholders are beyond the affordability to manage extra hands with little surplus which does not allow them to practice climate change adaptation strategies (Yila & Resurreccion 2013). Additionally, higher age of the household head alternatively more experienced in agricultural activities increases the probability of accepting adaptation options (Deressa et al. 2009). Another study by Jin et al. (2016) showed that farmers' farm size, farming experience, education level play significant role in choosing adaptation measures mostly in crop diversification.

In this study, income and farm size were important factors in defining the important elements for choosing new climate change adaptation strategies in agriculture. Confirming this consistent finding, the study of Thoai et al. (2018) also showed that households with large farm size were more likely to accept climate change adaptation strategies than smallholders. Moreover, farm size determines the farmers' adaptation decisions in maintaining climate change adaptation measures (Nhemachena & Hassan 2007; Piya et al. 2013; Ashraf et al. 2014; Ahmad & SeinnSeinn 2015; Asfaw et al. 2016; Jin et al. 2016). One of the possible reason behind this causation was explained in the study of Thoai et al. (2018), where author described that higher economic loss caused by climate change are significant in the large-scale farm size than small scale farms which works as a possible factor for adopting climate change adaptation strategies. Furthermore, in our study credit accessibility and assistance from an NGO/government organizations were not significant for considering adaption options. Similar findings also reflected in the study of Piya et al. (2013) describing access to credit as a negative impact on adaptation measures. Additionally, in accepting new adaptation measure smallholder farmers often hesitate because they are partly suspicious about the benefits and socio-economic constraints, which also works as a contributing factor. Similarly, practicing only indigenous strategies also becomes impossible for various reasons (Yila & Resurreccion 2013). Descriptive findings, from our study, regarding existing adaptation practices show that crop management is the highly practiced adaptation measures corroborating the findings of Hassan & Nhemachena (2008). In view of the fact that, high reliance on crop diversification and crop management help farmers' to diffuse risks in agriculture (Adger et al. 2003; Ashraf et al. 2014). In addition to this, infrastructure and water management were two other popular adaptation measures identified in this study.

Another important finding also shows that male-led farms are more likely to adapt water management practices. Findings from Tenge et al. (2004) also showed that female-headed households were less likely to implement soil and water conservation measures. One of the possible reasons, we observed during the study, explains the fact that water management in this study was observed in carrying water through a small water canal or carrying water with a water pot to water the plant which is troublesome. Also, this practice is more convenient for a male farmer than female. Another possible reason explained in the study of Yila & Resurreccion (2013) that, female-headed households are less likely to adapt climate change strategies due to cultural and social barriers that limit women's access to land and information. Furthermore, different geographical and cultural contexts might influence female-headed households to adopt climate change adaptation measures too. To point this out the study of Hassan & Nhemachena (2008) in South Africa showed that female-headed households are less likely to take up adaptation than male-headed households.

The final result from multinomial logistic regression, to understand the most demanded adaptation practices, water management practices stood out compared to farm management across the age and income level of the farmers. A greater emphasis on water management practices was observed by farmers to differentiate themselves from different climate change adaptation scenarios. One important thing to consider is that water management in coastal areas is important because, due to the proximity to the sea, salinity problems are highly reported in these areas. Crucial sectors like agro-economy, economy, and human health in coastal areas are in great jeopardy due to the high salinity conditions and the situation only aggravates during the cyclonic storm surge (Rakib et al. 2019a, 2019b; Uddin et al. 2019). In the study by Rahim et al. (2018) found that the maximum salinity problem occurs during the month of March to June. Due to this high salinity agricultural activities are significantly negatively impacted and cultivatable lands are left to fallow. Livelihoods related to agriculture are halted to put a stop. A study of Jin et al. (2016) outlined the fact that gender differences in agriculture showed a significant preference for technological practices in water management measures. They also found better educational qualifications of farmers' increases investments in irrigation infrastructure in adapting climate change conditions in agriculture. It is undoubtedly true that the education level of the household head and his adaptation to climate change in water management are highly associated (Maddison 2007; Deressa et al. 2009). Moreover, water management procedure in this study was highly demanded because of the unavailability of irrigation water due to drying up the water resources and a scarce number of groundwater irrigation facilities. Farmers highlighted they require proper irrigation facilities with the re-dredging of existing water canals. A study by Ashraf et al. (2014) found that due to lack of water management measures farmers' were unable to practice some of the crop management adaptation measures which ultimately resulted in low agricultural production.

The findings of this study have some important policy implications for the promotion of climate change adaptation strategies at the farm level regionally and locally. First, government should focus and extend services in improving the capability and effectiveness of different programs as a catalyst for appropriate climate change strategies. Speaking of the agricultural policy of Bangladesh, the Department of Agricultural Extension (DAE) pronounces some of the initiatives including supply and availability of quality seeds, fertilizers, irrigation, pesticides, etc. Though it is mentioned that agriculture extension service will be strengthened to encourage a self-motivated cooperative system of production, but the implementation of these services is required to be ensured (MoA 1999). Secondly, strengthening both formal and informal institutions is a vital prerequisite to effective adaptive capacity at the local and regional level. These institutions would expand knowledge and capacity about climate change in all levels of the formal governance structures and possibly promote more anticipatory capacity building. In this regard, the National Agricultural Research System (NARS) aims to strengthen and coordinate different programs through periodic evaluation (MoA 1999). Thirdly, for better water management in agriculture, excavation of the water canals is highly prioritized by the farmers. In order to make this challenge turn into a fruitful adaptation strategy Cash-For-Work (CFW) program within the Employment Generation Programme for the Poorest (EGPP) would be a great approach both for the enhanced adaption strategy as well as for the program itself too. Finally, civil societies and environmental NGOs should actively participate in crafting and advocating for robust and efficient strategies and agricultural management and adaptation. Different NGOs and banks are providing credits. Moreover, government has assigned different bodies of authorities with merging both government and private bodies to facilitate agricultural development (MoA 1999).

With the changing conditions of climate change and its negative consequences on coastal areas requires more attention into policy framework. In view of the findings of this study, government and relevant stakeholders should focus on improving the capability and effectiveness of the extension service, which is a catalyst for the promotion of appropriate climate change strategies.

In this research, we have tried to explain the possible factors responsible for choosing different adaptation decisions. Based on our field visit and pertinent literature search, we chose the factors and tested them against the existing adaptation strategies. This is sufficed, based on literature, to explain the existing process but elaborated and comprehensive planning of research works are required to understand more clearly the adaptation strategies and introducing the new adaptation options both locally and regionally. Moreover, the dominant adaptation strategies, according to this study, water management and fertilizer use were explored which also might pose some environmental impacts. Though water management in our study was explored in terms of using water from nearby existing water bodies like canals and ponds rather than exploiting ground water. Nonetheless, future researchers are also encouraged to understand the causal pathways in terms of adaptation practices against climate change are imposing further environmental threats or not. One important aspect which is also encouraged to debunk in the future study is the implication of socio-cultural factors and the mind set of farmers, if any, to understand climate change adaptation perspectives. Additionally, institutional and political factors coupled with cognitive and psychological factors also merit further deep understanding of adaptation decisions and inclinations.

The purpose of this paper was to investigate farmers' adaptation practices and to identify the factors that influence various adaptation measures. Results show that farmers' gender, income, damaged sector, family labor, farm size, access to credit, and help from NGOs/governments are significantly correlated factors in adaptation decisions. But the overall analysis shows that only gender, family labor, farming experience, and damaged sector are significant determinants in practicing adaptation measures. Furthermore, crop and infrastructure management measures were the most important of the five adaptation measures while farm management was the least practiced option observed in this study. Under the five broad adaptation measures, fertilizer and improved seed use were highly observed in crop management. The primary water management adaptation measure was irrigation from ponds and canals. Infrastructure management only was observed through the application of tractors in ploughing and processing crops. Also, finally, the social practice was maintained through the adaptation of traditional methods of agricultural process.

The findings of this study have practical implications in initiation of geography-based government interventions in boosting regional agricultural adaptation mechanisms, particularly water management, which will support poor people, farmers, and community at the same time. Additionally, agricultural policy extension to strengthen and encourage self-motivated cooperative system of production at the local level is also needs to be considered alongside.

The authors are grateful to the editor of the journal, Prof. D. Nagesh Kumar, and the anonymous reviewers for their useful, constructive comments, and suggestions. All remaining errors and shortcomings entirely befall on authors.

The authors received no financial support for the research, authorship, and/or publication of this article.

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.