Farmers must clearly perceive climate change to implement autonomous adaptations and support planned adaptation and mitigation initiatives. Based on the primary data collected from 300 farming households of the sub-Himalayan West Bengal of India, the present study compared farmers’ perceptions with meteorological trends obtained through a combination of statistical and graphical tests. Many farmers declared a change in local climate, and farmers’ perceptions mostly cognate the increasing summer temperature and decreasing monsoon precipitation from 1991 to 2020. However, a greater level of imperfect perceptions was observed for the winter temperature rise. Overall, only 23% of farmers were able to clearly perceive all the changes in climatic conditions. A binary logistic regression model was employed to identify the determining factors of farmers’ clear perceptions, and results showed that farmers who are younger, male, read newspapers, and experienced elephant crop-raiding perceived the changes more accurately. Whereas access to television and irrigation facilities decreased the probability of perceiving climate change accurately. The study recommends that bridging the knowledge gap between farmers and stakeholders is necessary for this region, which could be achieved by disseminating accurate weather information in combination with agricultural advice and targeted initiatives, especially for the older and female farmers.

  • Multiple trend analysis techniques are used to detect the trends in meteorological variables.

  • The temperature increased, and monsoon precipitation decreased over the sub-Himalayan West Bengal, India.

  • The majority of the farmers’ perceptions are consistent with the meteorological trends except for the winter temperature increase.

  • Older and female farmers need more accurate information regarding climate change.

Graphical Abstract

Graphical Abstract
Graphical Abstract

The last few decades have witnessed far-reaching impacts of climate change on various sectors, and the risks to agriculture stand out as the most crucial (Mendelsohn 2009). Studies show that as a climate-sensitive sector, agriculture needs adaptation to sustain global food security, whereas, being a source of greenhouse gas emission, mitigation strategies are also required in this sector (Loboguerrero et al. 2019; Aryal et al. 2020). Both the adaptation and mitigation actions require a willingness to change behavior, and for that, it is necessary to perceive that climate change is happening and action is necessary (Niles & Mueller 2016). A large body of literature tracked farmers’ adaptation strategies along with their perception of climate change, declaring that perceiving climate change is the first step to adaptation (Sahu & Mishra 2013; Tambo & Abdoulaye 2013; Nath 2014).

The assessment of farmers’ perceptions naturally raises the question: to what extent do farmers’ perceptions match reality? In this line, researchers argued that farmers’ perceptions should be compared with the trends in instrumental records (Roco et al. 2015; Hasan & Kumar 2020). The extent to which perceptions match the meteorological trends would indicate how accurately farmers perceive the changes in the climate. However, it is worth noting that, while researchers rely on instrumental records and statistical techniques to detect climate change, farmers, especially in developing countries, rely on their past experiences, indigenous knowledge system, and memory to construct their perceptions, often without any prior knowledge of climate change. Therefore, what the farmers perceive as a change in climate might not always reflect the reality that statistical tests claim to be and the researchers and policymakers are concerned about.

The divergence between practitioners’ and farmers’ perspectives on climate change might arise for various reasons. Firstly, it could be due to the choice of statistical tests for assessing the climatic trends. Generally, the trend direction, magnitude, and significance are highly dependent on the robustness of statistical tests, fulfillment of test assumptions, the quality of meteorological data (e.g., missing value, outlier, homogeneity, etc.), and its length. Therefore, before aligning farmers’ perceptions with actual climatic trends, it is essential to carefully analyze the trends in meteorological variables. Secondly, if the meteorological trends are assessed accurately, then divergence might arise due to different dimensions of the societal setup and the individual's ability to detect climate change. Age, education, family size, income, farm size, access to extension services, and media are often reported as significant determinants of farmers’ accurate perceptions (Roco et al. 2015; Uddin et al. 2017; Hasan & Kumar 2020; Imran et al. 2020). Besides, the severity of some extreme climatic events might leave a long-term imprint on an individual's mind in such a way that it can easily lead to misperception of climatic trends (Platt et al. 2021). Also, some studies reported that infrastructural facilities impact farmers’ perception of climate change (Nath 2014; Niles & Mueller 2016).

In line with the nexus between climate change and agriculture, India is among those developing countries which might face major agricultural losses if appropriate adaptation and mitigation strategies are not implemented (Cline 2008). Currently, different Indian states have developed ‘State Action Plans on Climate Change’ to aid adaptation in agriculture and other sectors; India has also endorsed the Paris Agreement by pledging to reduce emission intensity and increase additional carbon sinks (Jaiswal 2017). However, several scholarly studies (Leiserowitz & Thaker 2012; Keller et al. 2020) have revealed that the majority of Indians still have very limited knowledge of climate change. Indian farmers being in the frontline to face the worst hits of climate change, if not perceive the changes accurately, it might hinder their ability to adopt appropriate adaptations by themselves as well as their support for governmental initiatives aimed at mitigation and adaptation.

Climate change and Indian farmers’ perceptions: a systematic review of literature

So far, several empirical research works have been carried out in different parts of India to investigate farmers’ perceptions of climate change. To determine how accurately Indian farmers perceived climate change, it is necessary to compile the scattered information. For this purpose, a systematic literature review (SLR) was undertaken to acquire information on some relevant research questions as presented in Table 1 and to identify the knowledge gaps for accelerating the contemporary field of research.

Table 1

Formulation of questions and the process followed in the SLR

Review of literature
Question formulation a. What changes do the farmers perceive in the meteorological variables, and for what time period?
b. Do farmers’ perceptions match the trends in meteorological variables?
c. How have the meteorological data been analyzed?
d. What factors determine farmers’ perception of climate change? 
Development of search string for database search (‘Climat* Chang*’ OR ‘Climat* extrem*’ OR ‘Climat* uncertainit*’ OR ‘Vulnerabilt’ OR ‘Maximum temperature ris*’ OR ‘Minimum temperature ris*’ OR ‘Temperature ris*’ OR ‘rainfall variabilit*’ OR ‘Sea level ris*’) AND (‘Farm*’ OR ‘Farm level’ OR ‘Agricultur*’ OR ‘Cultivat*’ OR ‘Smallhold*’ OR ‘Farm* communit*’) AND (‘Perception’ OR ‘Belie*’ OR ‘attitud*’) AND (‘India’) 
Screening of the identified records (155) • Literatures excluded based on title screening (n=113)
• Literatures screened by abstract (n=42) and literatures excluded (n=12)
• Full text evaluated for eligibility (n=30) and excluded (n=7) 
Studies included in the review process • Studies included by full-text evaluation (n=20)
• Studies included by searching reference list (n=3)
• Studies included by searching citation list (n=1)
• Studies included from authors’ personal desk (n=1) 
Review of literature
Question formulation a. What changes do the farmers perceive in the meteorological variables, and for what time period?
b. Do farmers’ perceptions match the trends in meteorological variables?
c. How have the meteorological data been analyzed?
d. What factors determine farmers’ perception of climate change? 
Development of search string for database search (‘Climat* Chang*’ OR ‘Climat* extrem*’ OR ‘Climat* uncertainit*’ OR ‘Vulnerabilt’ OR ‘Maximum temperature ris*’ OR ‘Minimum temperature ris*’ OR ‘Temperature ris*’ OR ‘rainfall variabilit*’ OR ‘Sea level ris*’) AND (‘Farm*’ OR ‘Farm level’ OR ‘Agricultur*’ OR ‘Cultivat*’ OR ‘Smallhold*’ OR ‘Farm* communit*’) AND (‘Perception’ OR ‘Belie*’ OR ‘attitud*’) AND (‘India’) 
Screening of the identified records (155) • Literatures excluded based on title screening (n=113)
• Literatures screened by abstract (n=42) and literatures excluded (n=12)
• Full text evaluated for eligibility (n=30) and excluded (n=7) 
Studies included in the review process • Studies included by full-text evaluation (n=20)
• Studies included by searching reference list (n=3)
• Studies included by searching citation list (n=1)
• Studies included from authors’ personal desk (n=1) 

The selection of literature was from the Scopus database following the guidelines of PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses) statement (Moher et al. 2009). This database has several advantages, such as advanced search function, extensiveness (indices over 5,000 publishers), quality control, multidisciplinarity, and combats predatory publishing (Gusenbauer & Haddaway 2019). The database was searched through some pre-defined keywords (Table 1) from 25th January to 3rd November 2021. The initial search produced 155 results in total. A screening of the titles, followed by the abstracts, was preliminarily done to identify relevant studies. After careful examination of the full text of 30 studies, 20 studies were included for the review. Additional records were identified by checking the reference, citation lists, and authors’ personal desks. The list of exclusion–inclusion criteria set for screening the studies is presented in Table S1, and the list of selected studies is presented in Appendix S1. The finding from the SLR is presented in Table S2 and discussed below.

What changes do the farmers perceive in the meteorological variables, and for what time period?

The assessment revealed that an increase in temperature and a decrease in rainfall are frequently perceived by the farmers as a change in climatic conditions (Table S2). Among the reviewed studies, 18 reported an overall increase in temperature, whereas eight and five studies reported summer and winter temperature increases, respectively. On the other hand, in only one study, most respondents reported no change in the winter temperature. Concerning the perception of changes in rainfall, most of the studies (20) reported a decrease and an increase reported in three studies (Table S2). However, there is no consensus among the studies regarding the time period over which farmers’ perception has been elicited. In total, 18 studies reported the time period in which the farmers were asked to state the change, of which the maximum stated a change of 30 and 20 years.

Do farmers’ perceptions match the trends in meteorological variables?

Maximum studies (24 out of 25) validated farmers’ perceptions by comparing those with the meteorological data. However, the period over which farmers’ perception and meteorological data have been compared for validating the perception is of a similar length only in two studies. In maximum studies, authors reported the alignment of farmers’ perception (i.e., of 30 years or less) with instrumental records by analyzing meteorological data of relatively longer duration (i.e., 60 years or more). Out of these 24, 20 studies reported that the maximum of the sample farmers’ perception of temperature and rainfall changes fully matched with the meteorological data (Table S2). Whereas two studies reported discrepancies for winter temperature and rainfall, and one study found that farmers’ perceptions are closely matched only with the nearest station records.

How have the meteorological data been analyzed?

Seventeen studies reported alignment/non-alignment of farmers’ perception with meteorological data by means of analyzing the available meteorological records. Whereas seven studies reported the match on the basis of available published literature only. For analyzing the meteorological data, 15 studies opted for the statistical trend analysis method. Among these, eight studies used the non-parametric Mann–Kendall (MK) test, one study used MK and its modified version, five studies opted for linear trend analysis, and the newly developed Sen's innovative trend analysis was used in one study (Table S2). Whereas four studies analyzed meteorological data based on other methods such as mean difference to detect the change. Meanwhile, the analysis methods remained undetectable in six studies.

What factors determine farmers’ accurate perception of climate change?

Among the 25 reviewed studies, 14 tried to explain factors responsible for farmers’ accurate perception of climate change (Table S2). However, most of the studies (12 out of 14) are qualitative in nature and focused on various factors related to farmers’ experience and memory that helped them recognize the change in climate by comparing the present climatic conditions with different past events. Only two studies opted for statistical analysis to identify different socio-economic and demographic factors (e.g., income, education, family type, age, gender, access to irrigation, and availability of weather information) associated with farmers’ perception of climate change.

  • The review identified the following limitations in the existing literature:

First, different studies communicated the concerned period with climate change to farmers in different ways, and they frequently compared the perception with the long-term meteorological data. Since long-term climatic trends often differ from short-term trends, comparing farmers’ perceptions with short-term climatic trends is more rational. Second, when it came to analyzing meteorological data, most research relied solely on parametric linear regression and the non-parametric MK test without taking into account their assumptions. Although the last one is less sensitive and largely used for trend detection, it assumes unautocorrelated observation. Therefore, modified MK tests are more appropriate in the presence of autocorrelation, which is overlooked in most studies. Besides, several researchers recommended multiple statistical analyses to confirm the trends in climatic parameters (Khaliq et al. 2009; Sonali & Kumar 2013). Third, much of the existing research is qualitative, focusing on various physiological aspects that influence farmers’ perceptions. However, several researchers (Roco et al. 2015; Uddin et al. 2017; Asare-Nuamah & Botchway 2019; Hasan & Kumar 2020; Imran et al. 2020; Mbwambo et al. 2021) have emphasized the importance of statistical models in order to bridge the knowledge gap between policymakers and farmers. Therefore, to better understand why certain farmers perceive climate change more precisely than others, more quantitative studies are required, which are scarce in the literature.

This study, therefore, focussed on first identifying the trends in local climate using a combination of statistical and graphical tests; second, compared these trends with farmers’ perception; and, lastly, identified the factors driving farmers’ accurate perception. Integrating farmers’ perceptions in conjunction with meteorological data would bridge knowledge gaps between qualitative and quantitative approaches and contribute to further debates in the contemporary literature. Furthermore, the findings would assist planners and policymakers in comprehending and incorporating local farmers’ perspectives, as well as channeling funds to the appropriate group of individuals to enable a behavior change in support of adaptation and mitigation.

The study was conducted in the sub-Himalayan West Bengal of India. Bhutan borders a large portion of this area to the north, which might justify the alternate name of ‘Dooars’ (i.e., gateway to Bhutan) for this region. Distinct geophysical and socio-economic characteristics of this area, such as its location as a tilted plain (piedmont zone) at the base of the Himalayan steep slopes, acidic and porous soil with low water holding capacity, lack of irrigation facilities, recurring floods, flash floods, shifting river courses, a higher concentration of scheduled caste and scheduled tribe population, poverty, illiteracy, and a lack of industrial and infrastructural development, increased its vulnerability to climate change (Jana 2012; Datta & Behera 2022). Despite a series of setbacks for the agricultural sector, farming continued to be an important source of livelihood for the majority of the population in this region. However, with relatively sluggish growth in the adoption of modern tools and techniques, agricultural practices have remained traditional and subsistence-oriented (Sarkar 2013; NABARD 2014). Paddy, jute, and potato are the principal crops, with Areca nuts lately gaining much popularity as a commercial crop (Sit et al. 2011). Recent studies showed a significant change in the long-term climatic pattern in this region (Datta & Das 2019a, 2019b). Given these circumstances, this region is identified as a stressed region under the Indian state of West Bengal's State Action Plan on Climate Change (SAPCC), requiring immediate attention from researchers and policymakers.

Sampling and primary data collection

The primary data were collected through household surveys from six villages located in the sub-Himalaya, namely, Dhumpara, Pradhanpara, Ballalguri, Uttar Rangali Baznar, Dalsingpara, and Turturi Khanda (Figure 1). These six villages are identified as climate-induced vulnerable villages in the district disaster management plan, since these are susceptible to a number of climate-induced adversities like floods, flash floods, inundations, and so on. Climate change is thought to amplify these existing disastrous events (Mishra et al. 2019). The following formula developed by Cochran (1977) was used to estimate the sample size:
(1)
where N is the estimated sample size, z is the critical value of desired confidence level, a is the estimated proportion of an attribute present in the population, b=1−a and e is the allowable error. For the present study, assuming the maximum variability, equal to 50% (a=0.5) and taking a 95% confidence level with 5.65% allowable error (expressed as decimal 0.0565), the estimated sample size is 300. Therefore, the present study used data from 50 farming households that were chosen using simple random sampling from each of the six villages.
Figure 1

Location of the study villages.

Figure 1

Location of the study villages.

Close modal

The households were interviewed from March to July 2021 using a carefully designed pre-tested semi-structured schedule. Face-to-face interviews were undertaken since the literacy rate in this area is low, and farmers might require explanations to fully comprehend the concept of climate change. Pretesting was undertaken in February 2021 with a total of 20 non-sample respondents through a pilot study. Based on the experiences gained from the pilot study, the questions were further modified to enhance clarity. The questions mainly focused on gathering information on farmers’ personal and socio-economic characteristics, perception of climate change, and farming practices. In case farmers were not familiar with the concept of climate change, clear instructions were given to them for expressing their views in terms of increase, decrease, or no change in the climatic condition over the last 25–30 years.

Most of the survey respondents were considered to be the head farmer of the households, and in their absence, the second head farmer was interviewed. Lastly, it should be mentioned that all of the household interviews were carried out following the Government guidelines for the COVID-19 situation. Besides, only those households were interviewed who were willing to share information.

Secondary climate data and detection of trends

The secondary data for precipitation and temperature were obtained from the Climatic Research Unit gridded Time Series, CRU TS v. 4.05 (https://crudata.uea.ac.uk). The daily or sub-daily level observed meteorological data from a global observatory network of different National Meteorological Services and other external agents are used to construct the gridded data at a finer resolution of 0.5°×0.5°. As mentioned in Harris et al. (2020), the dataset has been constructed using angular-distance weighting (ADW) interpolation. Also, the CRU data have been compared with different widely used global-level datasets, e.g., University of Delaware (UDEL), Global Precipitation Climatology Center (GPCC), climate research temperature TEM data (CRUTEM), and consistency were found. In the Indian context, Robertson et al. (2013) and Rao et al. (2014) have given preference to CRU data due to its reasonable degree of closeness with the data of the National Data Centre (NDC) India and the India Meteorological Department (IMD). This dataset is currently used to detect climatic trends over several Indian regions (Rao et al. 2014; Datta & Das 2019b; Sharma et al. 2021). Three grid points near the study villages were selected for analysis and are presented in Figure 1. To detect the trends and trend magnitudes in meteorological variables, non-parametric MK and Modified Mann–Kendall (MMK), Theil–Sen Slope Estimator (TSSE), and the newly developed Şen's Innovative Trend Assessment (both the graphical and statistical approach) were employed.

MK and MMK tests

The non-parametric MK (Mann 1945; Kendall 1975) test is particularly suitable for analyzing the trends in climatic parameters since this test does not assume any particular distribution and is also applicable where there is a missing value and outliers in the data (Sonali & Kumar 2013). However, the test is affected greatly by the presence of autocorrelation in the time series, which is very common in hydro-climatological data. In this study, autocorrelation was checked in the RStudio version 4.0.5, and in the case of an autocorrelated series, the MMK test as proposed by Hamed & Rao (1998) was applied. This test modifies the variance of the original MK test and calculates the corrected z value. More details regarding the test statistics and formula could be found in the work of Sonali & Kumar (2013) and Datta & Das (2019b).

Theil–Sen slope estimator (TSSE)

The TSSE is also a non-parametric method to estimate the trend magnitude. The method was initially developed by Theil (1950) and later modified by Sen (1968). This method is compatible with the outliers and missing values present in the dataset (Gilbert 1987). The formula for calculating the TSSE can be found in the work of Datta & Das (2019a).

Graphical Innovative Trend Assessment (G-ITA)

In order to reduce the methodological glitches either from the statistical perspective or from any other uncertainties, Şen (2012) developed a G-ITA method, which is compatible even in the presence of significant autocorrelation coefficients. The basis of G-ITA is that if two random time series are identical, their plot against each other shows scattered points along a 1:1 (45°) straight-line on the Cartesian coordinate system (Şen 2012). In order to detect trends, firstly, the entire time series has to be divided into two equal segments, followed by sorting them in ascending order. When data points cross the 1:1 line and fall above, a monotonic increasing trend is indicated, and vice versa. If the data points scatter both sides of the 1:1 line, it indicates a non-monotonic trend (Şen 2012; Datta & Behera 2021).

Statistical Innovative Trend Assessment (S-ITA)

Similar to the G-ITA, in S-ITA also, the time series is divided into two equal halves and sorted ascendingly. The slope, Sm, can be estimated considering the arithmetic averages of the ascendingly ordered two-halves (Şen 2017) as expressed in the following equation:
(2)
In this equation, n is the number of observations, μ2 and μ1 are the mean of the second half and first half, respectively. Here, n/2 is the time difference between the two-halves. Based on the slope's sign (+ and −), the trend direction, i.e., positive or negative, is inferred. Since the probability density function (PDF) of the slope stands with the normal distribution, the standard deviation of the slope can be expressed as:
(3)
where is the cross-correlation between the ascendingly ordered two-halves’ arithmetic averages. Scri is the confidence limits of a standard normal PDF at α percent significance level, and Şen (2017) recommended obtaining the (1−α) percent confidence limits for the trend slope as follows:
(4)
where σs is the standard deviation of the slope, Sm, and estimated using Equation (2). The null hypothesis, H0 (there is no significant trend), is accepted if the calculated slope ranges within the upper and lower CL boundary (Equation (4)), or else an alternative hypothesis, Ha (there is a significant trend), will be inferred.

The climate change perception model

In recent times, there has been a spurt in identifying factors influencing a ‘clear perception’ of climate change (Roco et al. 2015; Hasan & Kumar 2020). These studies emphasized an unambiguous separation between the farmers who accurately perceived climate change and those who did not and measured the perception by a dummy variable, in which a value of 1 was assigned if a farmer perceived the change in certain meteorological variables in accordance with the results of trend analysis and a value of 0 otherwise. The purpose of such strict distinction is to notify policymakers about the factors that reduce the knowledge gap between farmers’ perceptions and scientific discourse. In this study, a farmer is considered to have a clear perception of climate change (y=1) if all of the following conditions relating to the changes over the last 30 years are met in line with the meteorological trends:

  • i.

    The farmer declares a change in the local climate.

  • ii.

    The farmer perceives a change in the summer temperature.

  • iii.

    The farmer perceives a change in the winter temperature.

  • iv.

    The farmer perceives a change in monsoon rainfall.

In this study, we considered the declaration of a change in the local climate as the first step to recognize the changing climatic conditions. Next, we emphasized on the two key meteorological variables that are responsible for altering several conditions like evapotranspiration, water availability, the occurrence of floods, and droughts, which eventually affect agricultural production and, consequently, the livelihoods and food security. Further, the perception of temperature change is considered for two distinct conditions, i.e., hot summer (April–September) months and cold winter (December–February); whereas for the rainfall, we only considered the monsoon (June–September) season; since near about 78 %of the rainfall occurs in this season. Given the poor educational development in this region, inquiring about changes in the prominent seasonal conditions would enable effective communication with the farmers about climate change. On the other hand, if the clear perception was only defined by one of the four criteria listed above, the dependent variable would be highly concentrated, with little variation, and would fail to produce the necessary clear balance between 0 and 1 s (Roco et al. 2015).

A binary logistic regression (BLR) model was employed to assess the significant variables that determined farmers’ accurate perception (Table 2). The choice of all the explanatory variables is based on the experience gained from the uniqueness of the study area, consultations with agricultural experts who are well-informed of local realities, and theoretical background from past studies (Table 2). Before estimating the BLR, it is essential to check the multicollinearity among the independent variables since BLR is sensitive to the high correlation among the predictor variables (Pallant 2013). This was attained through computing the variance inflation factor (VIF) and tolerance limit. A relatively high tolerance varying from 0.65 to 0.89 and the VIF much below the usually accepted limit of 10 suggests that multicollinearity is not a major issue in our study (Table S3). The specification of the BLR is as follows:
(5)
where Yi is a dichotomous dependent variable, is the Y-intercept, is the coefficients to be estimated, and X is the explanatory variables hypothesized to influence farmers’ perception of climate change, and is an error term.
Table 2

Choice of explanatory variables for analyzing the factors influencing farmers’ clear perception of climate change

FactorsVariablesDescriptionSourceMeanSD.
Personal characteristics Age Continuous, age of the farmers in years Maddison (2006); Roco et al. (2015)  48.88 12.81 
Gender Dummy, 1=male, 0 otherwise Ofuoku & Campus (2011); Liu et al. (2013)  0.80 0.40 
Community Dummy, 1=tribal, 0 otherwise Field experience 0.43 0.50 
Education Continuous, years of formal schooling Piya et al. (2012); Nath (2014)  3.03 3.76 
Access to information Access to newspaper Dummy, 1=yes, 0 otherwise Piya et al. (2012); Roco et al. (2015)  0.13 0.34 
Access to television Dummy, 1=yes, 0 otherwise Karki et al. (2020)  0.41 0.49 
Access to extension service Dummy, 1=yes, 0 otherwise Bryan et al. (2013); Debela et al. (2015)  0.07 0.26 
Farm characteristics Landholding size Continuous, total landholding size in hectare Akanda & Howlader (2015); Uddin et al. (2017)  0.87 0.99 
Land tenure Dummy, 1=secure, 0 otherwise Roco et al. (2015)  0.81 0.39 
Access to irrigation Dummy, 1=yes, 0 otherwise Nath (2014); Roco et al. (2015)  0.45 0.50 
Flood affectedness Dummy, 1=yes, 0 otherwise Spence et al. (2011)  0.22 0.41 
Elephant crop-raiding Dummy, 1=yes, 0 otherwise Field experience 0.62 0.49 
FactorsVariablesDescriptionSourceMeanSD.
Personal characteristics Age Continuous, age of the farmers in years Maddison (2006); Roco et al. (2015)  48.88 12.81 
Gender Dummy, 1=male, 0 otherwise Ofuoku & Campus (2011); Liu et al. (2013)  0.80 0.40 
Community Dummy, 1=tribal, 0 otherwise Field experience 0.43 0.50 
Education Continuous, years of formal schooling Piya et al. (2012); Nath (2014)  3.03 3.76 
Access to information Access to newspaper Dummy, 1=yes, 0 otherwise Piya et al. (2012); Roco et al. (2015)  0.13 0.34 
Access to television Dummy, 1=yes, 0 otherwise Karki et al. (2020)  0.41 0.49 
Access to extension service Dummy, 1=yes, 0 otherwise Bryan et al. (2013); Debela et al. (2015)  0.07 0.26 
Farm characteristics Landholding size Continuous, total landholding size in hectare Akanda & Howlader (2015); Uddin et al. (2017)  0.87 0.99 
Land tenure Dummy, 1=secure, 0 otherwise Roco et al. (2015)  0.81 0.39 
Access to irrigation Dummy, 1=yes, 0 otherwise Nath (2014); Roco et al. (2015)  0.45 0.50 
Flood affectedness Dummy, 1=yes, 0 otherwise Spence et al. (2011)  0.22 0.41 
Elephant crop-raiding Dummy, 1=yes, 0 otherwise Field experience 0.62 0.49 
In general, a logistic regression model can be characterized as follows (Hosmer & Lameshow 2010):
(6)

Holding other independent variables constant, for a unit increase in a particular independent variable, each estimated coefficient would change in the log-odds, log (p/(1−p)), of the dependent variable (Jin et al. 2016). For testing the goodness of fit, Omnibus tests of model coefficients (OTMC) and Hosmer–Lemeshow (H–S) goodness-of-fit test were employed. For the OTMC, a value less than 0.05 is generally considered highly significant, and a value more than 0.05 of the H–S test statistic indicates a good fit for the model (Pallant 2013). In our case, the OTMC value was 0.00, and the χ2 value for the H–S test was 3.64 with a significance level of 0.89, therefore indicating good support for the BLR model employed in this study.

Trends in meteorological variables

The changes in the temperature and precipitation over the study area are presented in Table 3. The MK/MMK and the S-ITA tests show a significant increasing trend in both the summer and winter months. Although the slope values are slightly intense in S-ITA than in the TSSE, there is no difference in the trend direction. The TSSE values show an increase of 0.02 °C/year in all the study villages during summer, whereas, for the S-ITA, it varied from 0.02 to 0.03 °C/year. In case of the winter, the slope value in TSSE varied from 0.01 to 0.03 °C/year and from 0.03 to 0.04 °C/year in S-ITA. The G-ITA also supports the results of the statistical tests and shows a monotonic increase in both seasons (Figure 2). The results align with the previous studies conducted over India (Sonali & Kumar 2013; Purnadurga et al. 2018), other South Asian regions (Rahman & Lateh 2016; Nawaz et al. 2020), and the West Bengal State Action Plan on Climate Change (WBSAPCC). Especially after 1970, India has witnessed a clear upward trend in temperature, although there were spatial variations in the trend magnitude. As projected by numerous general circulation models (GCMs), the temperature will continue to rise further, and it is likely to increase more in the winter months than in the summer (Almazroui et al. 2020).

Table 3

Trends in meteorological variables

Grid pointsTrend tests and farmers’ perceptionSummer temperatureWinter temperatureMonsoon precipitation
G1 MK/MMK 1.963** 1.154 −1.722* 
TSSE 0.02 0.01 −8.91 
G-ITA MI MI MD 
S-ITA 0.02*** 0.03*** −8.46*** 
Farmers’ perception Increasing Decreasing Decreasing 
G2 MK/MMK 2.124** 1.545 −1.759* 
TSSE 0.02 0.02 −11.35 
G-ITA MI MI MD 
S-ITA 0.03*** 0.03*** −9.53*** 
Farmers’ perception Increasing Decreasing Decreasing 
G3 MK/MMK 2.569** 1.83* −1.641 
TSSE 0.02 0.03 −9.09 
G-ITA MI MI MD 
S-ITA 0.03*** 0.04*** −7.81*** 
Farmers’ perception Increasing Decreasing Decreasing 
Grid pointsTrend tests and farmers’ perceptionSummer temperatureWinter temperatureMonsoon precipitation
G1 MK/MMK 1.963** 1.154 −1.722* 
TSSE 0.02 0.01 −8.91 
G-ITA MI MI MD 
S-ITA 0.02*** 0.03*** −8.46*** 
Farmers’ perception Increasing Decreasing Decreasing 
G2 MK/MMK 2.124** 1.545 −1.759* 
TSSE 0.02 0.02 −11.35 
G-ITA MI MI MD 
S-ITA 0.03*** 0.03*** −9.53*** 
Farmers’ perception Increasing Decreasing Decreasing 
G3 MK/MMK 2.569** 1.83* −1.641 
TSSE 0.02 0.03 −9.09 
G-ITA MI MI MD 
S-ITA 0.03*** 0.04*** −7.81*** 
Farmers’ perception Increasing Decreasing Decreasing 

MK/MMK, z value of the Mann–Kendall/Modified Mann—Kendall test; TSSE, Theil–Sen slope estimator; G-ITA, Graphical Innovative Trend Assessment; S-ITA, Statistical Innovative Trend Assessment; MI, monotonic increasing; MD, monotonic decreasing.

*Statistically significant trends at 90% significance level.

**Statistically significant trends at 95% significance level.

***Statistically significant trends at 99% significance level.

Figure 2

Graphs showing the results of G-ITA for the summer temperature, winter temperature, and monsoon precipitation at three grid points (G1, G2, and G3).

Figure 2

Graphs showing the results of G-ITA for the summer temperature, winter temperature, and monsoon precipitation at three grid points (G1, G2, and G3).

Close modal

In the case of precipitation, both the statistical tests (MK/MMK, S-ITA) as well as the G-ITA (Figure 2) show a monotonic decreasing trend over the study area, which are also highly significant (Table 3). In the TSSE, it varied from −8.91 to −11.35 mm/year, whereas, for the S-ITA, it varied from −7.81 to −9.53 mm/year. Similar to our findings, a decrease in the Indian summer monsoon precipitation has been reported in a number of previous studies (Datta & Das 2019b; Singh et al. 2019). Also, the WBSAPCC identified decreasing monsoon precipitation over the study region. Monsoon being an important source of water for agriculture and to a range of other sectors as well as for the streamflow, it has attracted the attention of numerous researchers to track its changes. Studies discussed the weakening of monsoon in light of different anthropogenic activities (Guo et al. 2015; Singh et al. 2019). Whereas some studies stressed on its variability due to different natural forcings (Taraphdar et al. 2018; Das et al. 2020). Although, in the GCMs, the monsoon is expected to be intense over the Indian region (Menon et al. 2013; Almazroui et al. 2020), it has remained a debatable issue (Saha et al. 2014; Sabeerali et al. 2015; Aadhar & Mishra 2020). The limited spatial resolution, simplified physics, and thermodynamics of the GCMs, along with the complexities of monsoon, are possibly responsible for such (Sabeerali et al. 2015; Saha & Ghosh 2019).

Imprints of climate change on local farmers’ perceptions

Most of the surveyed farmers (90.7%) declared a change in the local climate. There was a general agreement between the summer temperature rise and farmers’ perception (Table 3). The majority of the farmers (91.7%) felt that the summer months were getting more hot in the study region. However, the winter temperature was not accurately perceived by a large number of sample farmers across the study villages. 72.3 and 2% of farmers declared a decrease and no change in winter temperature, respectively, which is contrary to the trends in meteorological data. Our findings are similar to the findings of Platt et al. (2021) in Indian Western Himalayas and Piya et al. (2012) in Nepal Himalayas. Both of these studies reported that a majority of respondents perceived summer temperature rise accurately; however, they disagreed while asked for winter temperature rise.

Since farming often includes activities to be performed on the field, therefore, the farmers could be more sensitive to warming in the summer because the ambient temperature is already high, and any further rise would undoubtedly create discomfort while performing farming activities. The study of Howe et al. (2012) and Platt et al. (2021) also indicated that the chances of perceiving higher temperatures are more in the summer. In contrast, it is obvious that the temperature drops in the winter, making it difficult for certain people to carry out their daily tasks, which may drive them to express the winter temperature trend contrary to meteorological records. Besides, Figure 3 depicts the temperature anomalies in the study area showing negative anomalies for the winter temperature in the recent 3 years, which could have impacted the farmers’ perception and led them to overlook the long-term trend. Besides, in December 2018, some areas of Darjeeling and Kalimpong received sudden snowfall after more than a decade (The Times of India 2019), and such sudden snowfall and associated decrease in temperature might have left a significant impact on the farmers’ minds. The effect of memorable climatic events often misleads people regarding the long-term climate trend (Weber & Stern 2011). However, a slightly negative anomaly is also present in the summer temperature (Figure 3) but might have less impact on farmers’ perception than the experience of heat stress while being exposed to the fields. In open-ended questions, several farmers expressed their discomfort due to the scorching heat of summer, which they had not experienced before.

Figure 3

Graphs showing the anomalies for the summer temperature, winter temperature, and monsoon precipitation at three grid points (G1, G2, and G3).

Figure 3

Graphs showing the anomalies for the summer temperature, winter temperature, and monsoon precipitation at three grid points (G1, G2, and G3).

Close modal

An overwhelming percentage of the farmers reported decreased precipitation during the monsoon season, which concurs with the meteorological trends identified in this study (Table 3, Figure 4). The previous studies were conducted over India (Nath 2014; Datta & Behera 2021), as well as other Himalayan regions (Chaudhary & Bawa 2011; Piya et al. 2012), and also reported similar perceptions. The ‘visual salience’ of rainfall perhaps facilitated the perception of the decreased quantity of rainfall. Several farmers expressed their past experience of water accumulation in the yards and stated that now there is a remarkable decrease in the amount of rainwater received from the monsoon season. Indeed, these experiences offer a framework for pictorial comparison of the monsoon over 25–30 years. The role of visual salience in the perception of climate change is also described in the study of Vedwan (2006), Piya et al. (2012), and Platt et al. (2021). Also, respondents stated that the soil of their agricultural fields has become much drier than before, and they observed cracks forming very frequently as a result of low rainfall. Since the soil in the study area is porous and has a limited water holding capacity, the lower rainfall, along with increased temperatures, might just have exacerbated soil dryness. Consequently, the experience of increased soil dryness perhaps made the respondents who perceive the decreasing trend of monsoon rainfall clearly.

Figure 4

Farmers’ perception of climate change.

Figure 4

Farmers’ perception of climate change.

Close modal

Factors influencing the probability of a clear perception of climate change

This study stressed on identifying factors responsible for no knowledge gap and a clear perception of climate change and employed a BLR model for this purpose. The Nagelkerke R2 is 0.504, indicating the model explained 50% of the variance in the dependent variable and correctly classified 84% of the cases.

Personal characteristics

As climate change is a long-term change in the climatic condition, different studies usually show that older farmers are more likely to have long experience, and they could demonstrate the changes more accurately than the younger farmers (Maddison 2006). However, the study of Byg & Salick (2009) reported that where the changes in climatic conditions are more of a recent phenomenon, the association between perception and age might be different than usual. In this study, we found that age is negatively and significantly associated with the accuracy of perception (Table 4). This finding aligns with the finding of Roco et al. (2015), in which the authors described this phenomenon as a result of the ‘recency effect’, where the recent climatic events impact more on the minds of younger people than older. Besides, gender is also an important variable that often creates differences in climate change perception. Studies found both a positive and negative association between gender and perception. Some studies indicated that male farmers have more access to social networks and training; therefore, they perceive climate change more accurately than their female counterparts (Ofuoku & Campus 2011; Ndambiri et al. 2012). However, the study of Liu et al. (2013) in the USA and Hasan & Kumar (2020) in Bangladesh reported that female farmers perceived climate change more accurately than male farmers and were more concerned about the negative impacts of climate change on their farms. In our study, male farmers are expected to perceive the climatic changes more accurately than female farmers (Table 4).

Table 4

Estimates of the BLR model

VariablesCoefficientSEOdds ratio
Age −0.07*** 0.02 0.94 
Gender 0.91* 0.50 2.49 
Community 0.67 0.42 1.95 
Education −0.02 0.06 0.98 
Access to newspaper 3.84*** 0.62 46.32 
Access to television −0.94** 0.44 0.39 
Access to extension service 1.42 0.86 4.16 
Landholding size −0.80** 0.39 0.45 
Land tenure 0.60 0.65 1.83 
Access to irrigation −1.06** 0.46 0.35 
Flood affectedness −0.75 0.50 0.47 
Elephant crop-raiding 1.47*** 0.50 4.36 
Nagelkerke R2 0.504   
VariablesCoefficientSEOdds ratio
Age −0.07*** 0.02 0.94 
Gender 0.91* 0.50 2.49 
Community 0.67 0.42 1.95 
Education −0.02 0.06 0.98 
Access to newspaper 3.84*** 0.62 46.32 
Access to television −0.94** 0.44 0.39 
Access to extension service 1.42 0.86 4.16 
Landholding size −0.80** 0.39 0.45 
Land tenure 0.60 0.65 1.83 
Access to irrigation −1.06** 0.46 0.35 
Flood affectedness −0.75 0.50 0.47 
Elephant crop-raiding 1.47*** 0.50 4.36 
Nagelkerke R2 0.504   

*p<0.1, **p<0.05, and ***p<0.01.

In the case of the other variables in farmers’ personal characteristics, the study did not find any significant association with the accuracy of climate change perception (Table 4). Although education is expected to increase the chances of perceiving climatic changes more accurately, it is not always universal (Karki et al. 2020). In the study area, the literacy rate is low, and the mean schooling years of the farmers were not even above the primary level (Table 2), which might not be enough to help them perceive climate change accurately. This argument is in line with the study of Piya et al. (2012) and Nath (2014). Whereas farmers from tribal communities are generally more connected with the natural environment and expected to have more clear observation of climatic trends. However, the present study did not find any such association (Table 4).

Access to information

Newspaper, television, and extension service agents are considered important sources of getting information on climate change and could play a significant role in generating awareness (Piya et al. 2012; Bryan et al. 2013; Debela et al. 2015; Roco et al. 2015). The results of the BLR model show that access to the newspaper is significantly associated with perceiving climate change clearly (Table 4). Controlling for other factors in the model, farmers who have access to the newspaper are over 46.32 times more likely to perceive climate change. Currently, in popular Indian newspapers, there is a gradual increase in reporting the climate change-related topics, and among these, the impact of climate change on agriculture is highlighted most frequently (Keller et al. 2020). The regionally circulated newspapers might also follow a similar pattern and substantially contributed to farmers’ perception. However, in the study villages, only 13% of farmers read the newspaper on a regular basis (Table 2).

On the other hand, a relatively higher percentage of farmers (41%) have access to television news (Table 2), but it is negatively associated with the probability of perceiving climate change. Currently, day-to-day weather forecasts are broadcasted on the news channels, although there is no program to give seasonal forecasts combined with agricultural advice. Studies (Aram & Nivas 2015; Khatun & Chaudhuri 2021) found that climate change-related news reports are event-specific in Indian news channels. News on climate change increased particularly in the years of IPCC reports publication and meetings of the Conference of the Parties (COP) and fell sharply later on. Besides, impacts of climate change are region-specific, and visuals of catastrophic events in other regions might create confusion among individuals about the local climatic trends. Unlike the study of Bryan et al. (2013) and Debela et al. (2015), we did not find any significant influence of extension services on the perception of climate change (Table 4). In general, the information-seeking behavior is observed to be low among the farmers of the study area, and agricultural practices still being traditional; farmers prefer consulting with their neighbor farmers.

Farm characteristics

The relationship between landholding size and clear perception of climate change is negative and significant in the present study (Table 4). The finding is similar to Uddin et al. (2017), whereas dissimilar to Akanda & Howlader (2015), who reported a significant positive association. In general, large landholders have more access to several agricultural inputs such as quality seeds, fertilizer, pesticides, and irrigation facilities, due to which they could be less sensitive to climatic exposures, and that might have influenced the farmers’ clear perception of climate change. Whereas smallholders having a lack of access to the equipment required to safeguard their production might have experienced more production loss due to the impacts of climate change and thus have a clearer perception than the large landholders. Whereas land tenure did not show any significant association with farmers’ perception. The results are contrary to Roco et al. (2015). Thus, farmers with secure tenure might not always have a clearer perception of climate change.

Further, the results show a negative relation between farmers’ perception and access to irrigation facilities. Farmers with access to irrigation in their farms are 0.35 times less likely to perceive climate change than the farmers without irrigation (Table 4). The findings are similar to Roco et al. (2015). Climate change is more likely to affect those farmers who do not have irrigation facilities on their farms, and these farmers are more vulnerable to the decreasing monsoon rainfall in the study villages. Venkateswarlu & Singh (2015) indicated that the farmers would need more access to irrigation since the monsoon is expected to become more irregular, and crop water demands are expected to increase as temperatures rise. The previous study of Nath (2014) also indicated that with the provision of irrigation, farmers often do not feel the shortage of rainwater and thus might not perceive the climatic trends accurately.

The impact of disaster experience can impact an individual's perception of climate change (Bergquist et al. 2019). Flooding being such a disaster that is often linked to climate change might impact an individual's perspective on climate change. Spence et al. (2011) found that previous flood experiences increased individuals’ concern for climate change. However, experiencing floods can also lead to stress and anxiety (Bergquist et al. 2019), which might interfere with farmers’ ability to perceive climate change more accurately. Karki et al. (2020) indicated that farmers with experience with natural disasters might focus more on climate variability and emphasis on that particular disastrous event and, therefore, might not perceive climatic trends in accordance with the meteorological data. However, we did not find any significant influence of flood experience on farmers’ probability of clearly perceiving climate change (Table 4).

While climate-related disaster experiences and their effect on climate change perception are documented in the literature, it is still not well understood whether non-climatic environmental concerns distract attention from perceiving farmers’ perception of climate change. When a large literature shows that elephant crop-raiding creates distress among farmers (Osborn & Parker 2003; Montgomery et al. 2021), we tried to explore whether such experience has any impact on farmers’ perception of climate change or not. Results show that it has a positive and significant association (Table 4). While climate-induced disasters might influence farmers’ perception negatively, non-climatic stressors might not influence in such a way. Farmers stated that elephant crop-raiding has been increasing over the last 8–10 years in the study villages, and they felt that elephants do not have enough food in the forests due to environmental changes. Such concerns might have positively influenced the farmers to perceive the climatic changes more clearly. A recent study (Naha et al. 2020) showed that incidents of crop-raiding increase in the dry period when rainfall gets scanty in the forest. As there is a decreasing rainfall trend in the study area, forest streams and vegetation might have been negatively impacted and increased the incidents of crop raidings by elephants.

The success of agricultural adaptation to climate change and mitigation activities requires collaboration among diverse stakeholders, including policymakers, governmental and non-governmental organizations, academicians, and, of course, the farmers. In this line, an assessment of climatic trends coupled with farmers’ perception of change in the local climatic conditions (that are often required in behavioral change to support adaptation and mitigation) can identify the loopholes between the understandings of different stakeholders and enhance the collaborative efforts. From this point of view, the present study analyzed the changes in two key meteorological variables (temperature and precipitation) by employing a combination of robust statistical and graphical trend analysis tests, explored the relevance of these trends from a societal point of view by aligning farmers’ perception with the identified trends, and also identified the factors responsible for the probability of farmers’ clear perception of climate change.

The analysis of meteorological data shows that there is significant warming both in the summer and winter seasons; also, the monsoon precipitation is showing a decreasing trend. The majority of the farmers’ perceptions aligned with the meteorological trends in summer temperature and monsoon precipitation. However, divergence occurred mostly in the case of winter temperature. Negative winter temperature anomalies and incidents of snowfall in the nearby hills during the last 3 years might have left a strong imprint on the farmers’ minds. Further analysis revealed that gender, access to newspapers, and experience of elephant crop-raiding are significantly associated with the perception of climate change, whereas age, landholding size, access to irrigation, and television are negatively associated.

The results of the study are particularly important for strengthening the WBSAPCC, where the identified meteorological trends concur with our findings; however, since the majority of the farmers’ do not have a clear perception of climate change, especially the change in winter temperature, policymakers need to first generate awareness among the farmers for successful implementation of adaptation initiatives. In this case, the provision and dissemination of information on climate change, seasonal and weekly climate forecasts with probable impact on the major crops through mass media and extension agents are essential for bridging the knowledge gap. Further, more specific action is required for informing the older and female farmers in this area. Our findings are strong enough to support policy actions in the sub-Himalayan West Bengal, and they might be extended to other locations with similar physical and socio-economic conditions. Future studies could focus on mixed-method approaches to better understand both the quantitative and qualitative dimensions of farmers’ perceptions.

The authors are grateful for the editor's and anonymous reviewers’ comments, which helped enhance the paper's quality and presentation. The authors would also like to thank Soumik Das, Amiya Basak, Satyajit Das, Ananda Chanda, Subhankar Saha, Prasanta Das, Suhrid Kundu, and others for their support during the survey planning and execution of fieldwork. The authors also express their gratitude to all the respondents, agricultural experts, and key informants who generously shared their expertise and knowledge.

Data cannot be made publicly available; readers should contact the corresponding author for details.

Aadhar
S.
&
Mishra
V.
2020
On the projected decline in droughts over South Asia in CMIP6 multimodel ensemble
.
Journal of Geophysical Research: Atmospheres
125
(
20
),
e2020JD033587
.
Akanda
M. G. R.
&
Howlader
M. S.
2015
Coastal farmers’ perception of climate change effects on agriculture at Galachipa Upazila under Patuakhali District of Bangladesh
.
Global Journals of Science Frontier Research: Agriculture and Veterinary
15
(
4
),
30
39
.
Almazroui
M.
,
Saeed
S.
,
Saeed
F.
,
Islam
M. N.
&
Ismail
M.
2020
Projections of precipitation and temperature over the South Asian countries in CMIP6
.
Earth Systems and Environment
4
(
2
),
297
320
.
Aram
I. A.
&
Nivas
G. P.
2015
How does the Indian television cover climate change?
Current Science
108
(
8
),
1420
1421
.
Aryal
J. P.
,
Sapkota
T. B.
,
Khurana
R.
,
Khatri-Chhetri
A.
,
Rahut
D. B.
&
Jat
M. L.
2020
Climate change and agriculture in South Asia: adaptation options in smallholder production systems
.
Environment, Development and Sustainability
22
(
6
),
5045
5075
.
Bergquist
M.
,
Nilsson
A.
&
Schultz
P.
2019
Experiencing a severe weather event increases concern about climate change
.
Frontiers in Psychology
10
,
220
.
Bryan
E.
,
Ringler
C.
,
Okoba
B.
,
Roncoli
C.
,
Silvestri
S.
&
Herrero
M.
2013
Adapting agriculture to climate change in Kenya: household strategies and determinants
.
Journal of Environmental Management
114
,
26
35
.
Cline
W. R.
2008
Global warming and agriculture
.
Finance & Development
45
(
001
),
23
27
.
Cochran
W. G.
1977
Sampling Techniques
, 3rd edn.
John Wiley & Sons
,
New York, NY, USA
.
Datta
P.
&
Behera
B.
2021
What caused smallholders to change farming practices in the era of climate change? Empirical evidence from Sub-Himalayan West Bengal, India
.
GeoJournal
,
1
17
,
https://doi.org/10.1007/s10708-021-10450-1
.
Datta
P.
&
Behera
B.
2022
Assessment of adaptive capacity and adaptation to climate change in the farming households of Eastern Himalayan foothills of West Bengal, India
.
Environmental Challenges
,
100462
,
https://doi.org/10.1016/j.envc.2022.100462
.
Gilbert
R. O.
1987
Statistical Methods for Environmental Pollution Monitoring
.
John Wiley & Sons
,
New York
.
Guo
L.
,
Turner
A.
&
Highwood
E.
2015
Impacts of 20th century aerosol emissions on the South Asian monsoon in the CMIP5 models
.
Atmospheric Chemistry and Physics
15
(
11
),
6367
6378
.
Hamed
K. H.
&
Rao
A. R.
1998
A modified Mann-Kendall trend test for autocorrelated data
.
Journal of Hydrology
204
(
1–4
),
182
196
.
Hosmer
D. W.
&
Lameshow
S.
2010
Applied logistic regression. Wiley series in probability and statistics. Wiley, New York
.
Harris
I.
,
Osborn
T. J.
,
Jones
P.
&
Lister
D.
2020
Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset
.
Scientific Data
7
(
1
),
1
18
.
Hasan
M. K.
&
Kumar
L.
2020
Meteorological data and farmers’ perception of coastal climate in Bangladesh
.
Science of the Total Environment
704
,
135384
.
Hosmer
D. W.
&
Lameshow
S.
2010
Applied logistic regression. Wiley series in probability and statistics. Wiley, New York
.
Howe
P. D.
,
Markowitz
E. M.
,
Lee
T. M.
,
Ko
C. Y.
&
Leiserowitz
A.
2012
Global perceptions of local temperature change
.
Nature Climate Change
3
(
4
),
352
356
.
Imran
M.
,
Shrestha
R. P.
&
Datta
A.
2020
Comparing farmers’ perceptions of climate change with meteorological data in three irrigated cropping zones of Punjab, Pakistan
.
Environment, Development and Sustainability
22
(
3
),
2121
2140
.
Jaiswal
A.
2017
India Leads on Climate Action as Trump Exits Paris Agreement
. .
Jana
A. K.
2012
Backwardness and political articulation of backwardness in the North Bengal Region of West Bengal
. In:
Rethinking State Politics in India
(A. Kumar, ed.)
,
Routledge
,
New Delhi, India
, pp.
161
204
.
Keller
T. R.
,
Hase
V.
,
Thaker
J.
,
Mahl
D.
&
Schäfer
M. S.
2020
News media coverage of climate change in India 1997–2016: using automated content analysis to assess themes and topics
.
Environmental Communication
14
(
2
),
219
235
.
Kendall
M. G.
1975
Rank Correlation Methods
.
Charles Griffen
,
London
.
Khatun
R.
&
Chaudhuri
M.
2021
Environmental reporting in Indian television
.
Consortium: An International Journal of Literary and Cultural Studies
1
(
1
),
42
54
.
Leiserowitz
A.
&
Thaker
J.
2012
Climate Change in the Indian Mind
. .
Liu
Z.
,
Smith
W. J.
&
Safi
A. S.
2013
Rancher and farmer perceptions of climate change in Nevada, USA
.
Climatic Change
122
(
1
),
313
327
.
Loboguerrero
A. M.
,
Campbell
B. M.
,
Cooper
P. J.
,
Hansen
J. W.
,
Rosenstock
T.
&
Wollenberg
E.
2019
Food and earth systems: priorities for climate change adaptation and mitigation for agriculture and food systems
.
Sustainability
11
(
5
),
1372
.
Maddison
D.
2006
The perception of and adaptation to climate change in Africa. CEEPA Discussion Paper No. 10, Centre for Environmental Economics and Policy in Africa, University of Pretoria, Hatfield, South Africa
.
Mann
H. B.
1945
Non-parametric tests against trend
.
Econometrica: Journal of Economy and Society
13
(
3
),
245
259
.
Mendelsohn
R.
2009
The impact of climate change on agriculture in developing countries
.
Journal of Natural Resources Policy Research
1
(
1
),
5
19
.
Menon
A.
,
Levermann
A.
,
Schewe
J.
,
Lehmann
J.
&
Frieler
K.
2013
Consistent increase in Indian monsoon rainfall and its variability across CMIP-5 models
.
Earth System Dynamics
4
(
2
),
287
300
.
Mishra
A.
,
Appadurai
A. N.
,
Choudhury
D.
,
Regmi
B. R.
,
Kelkar
U.
,
Alam
M.
,
Chaudhary
P.
,
Mu
S. S.
,
Ahmed
A. U.
,
Lotia
H.
&
Fu
C.
2019
Adaptation to climate change in the Hindu Kush Himalaya: stronger action urgently needed
. In:
The Hindu Kush Himalaya Assessment
. (
P. Wester. A. Mishra, A. Mukherji, A. B. Shrestha, eds.
),
Springer
,
Cham
, pp.
457
490
.
Moher
D.
,
Liberati
A.
,
Tetzlaff
J.
,
Altman
D. G.
&
Group
P.
2009
Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement
.
PLoS Medicine
6
(
7
),
e1000097
.
Montgomery
R. A.
,
Raupp
J.
,
Mukhwana
M.
,
Greenleaf
A.
,
Mudumba
T.
&
Muruthi
P.
2021
The efficacy of interventions to protect crops from raiding elephants
.
Ambio
51
,
716
727
.
NABARD
2014
District Profile Appendix-1 – NABARD
.
Available from: https://www.nabard.org/ (accessed 4 January 2022)
.
Nath
K. P.
2014
Analyzing Farmers’ Perception of and Adaptation to Climate Change in Assam
.
A thesis submitted to the Indian Institute of Technology Kharagpur for award of the degree of Doctor of Philosophy
.
Ndambiri
H. K.
,
Ritho
C.
,
Mbogoh
S. G.
,
Nyangweso
P. M.
,
Ng'ang'a
S. I.
,
Muiruri
E. J.
,
Kipsat
M. J.
,
Kubowon
P. C.
,
Cherotwo
F. H.
&
Omboto
P. I.
2012
Analysis of Farmers’ Perceptions of the Effects of Climate Change in Kenya: The Case of Kyuso District
.
No. 304-2016-4806
, pp.
309
328
.
Ofuoku
A. U.
&
Campus
A.
2011
Rural Farmers’ Perception of Climate Change in Central Agricultural Zone of Delta State, Nigeria
.
Pallant
J.
2013
SPSS Survival Manual
.
McGraw-Hill Education
,
UK
.
Piya
L.
,
Maharjan
K. L.
&
Joshi
N. P.
2012
Perceptions and realities of climate change among the Chepang communities in rural mid-hills of Nepal
.
Journal of Contemporary India Studies: Space and Society, Hiroshima University
2
(
5
),
35
50
.
Platt
R. V.
,
Ogra
M.
,
Kisak
N.
,
Manral
U.
&
Badola
R.
2021
Climate change perceptions, data, and adaptation in the Garhwal Himalayas of India
.
Climate and Development
13
(
2
),
95
106
.
Purnadurga
G.
,
Lakshmi Kumar
T. V.
,
Koteswara Rao
K.
,
Rajasekhar
M.
&
Narayanan
M. S.
2018
Investigation of temperature changes over India in association with meteorological parameters in a warming climate
.
International Journal of Climatology
38
(
2
),
867
877
.
Rahman
M. R.
&
Lateh
H.
2016
Spatio-temporal analysis of warming in Bangladesh using recent observed temperature data and GIS
.
Climate Dynamics
46
(
9–10
),
2943
2960
.
Rao
B. B.
,
Chowdary
P. S.
,
Sandeep
V. M.
,
Rao
V. U. M.
&
Venkateswarlu
B.
2014
Rising minimum temperature trends over India in recent decades: implications for agricultural production
.
Global and Planetary Change
117
,
1
8
.
Robertson
A. W.
,
Bell
M.
,
Cousin
R.
,
Curtis
A.
&
Li
S.
2013
Online tools for assessing the climatology and predictability of rainfall and temperature in the Indo-Gangetic Plains based on observed datasets and seasonal forecast models
. In:
CGIAR Research Program on Climate Change, Agriculture and Food Security (CCAFS). Working Paper No. 27
,
Cali, Colombia
.
Roco
L.
,
Engler
A.
,
Bravo-Ureta
B. E.
&
Jara-Rojas
R.
2015
Farmers’ perception of climate change in Mediterranean Chile
.
Regional Environmental Change
15
(
5
),
867
879
.
Sabeerali
C. T.
,
Rao
S. A.
,
Dhakate
A. R.
,
Salunke
K.
&
Goswami
B. N.
2015
Why ensemble mean projection of south Asian monsoon rainfall by CMIP5 models is not reliable?
Climate Dynamics
45
(
1–2
),
161
174
.
Saha
A.
&
Ghosh
S.
2019
Can the weakening of Indian monsoon be attributed to anthropogenic aerosols?
Environmental Research Communications
1
(
6
),
061006
.
Saha
A.
,
Ghosh
S.
,
Sahana
A. S.
&
Rao
E. P.
2014
Failure of CMIP5 climate models in simulating post-1950 decreasing trend of Indian monsoon
.
Geophysical Research Letters
41
(
20
),
7323
7330
.
Sarkar
B. C.
2013
Delineation of agricultural regions in Jalpaiguri District, West Bengal
.
Golden Research Thoughts
2
(
11
),
2013
.
Sen
P. K.
1968
Estimates of the regression coefficient based on Kendall's tau
.
Journal of the American Statistical Association
63
(
324
),
1379
1389
.
Şen
Z.
2012
Innovative trend analysis methodology
.
Journal of Hydrologic Engineering
17
(
9
),
1042
1046
.
Şen
Z.
2017
Innovative trend significance test and applications
.
Theoretical and Applied Climatology
127
(
3–4
),
939
947
.
Sharma
C.
,
Shukla
A. K.
&
Zhang
Y.
2021
Climate change detection and attribution in the Ganga-Brahmaputra-Meghna river basins
.
Geoscience Frontiers
12
(
5
),
101186
.
Singh
D.
,
Ghosh
S.
,
Roxy
M. K.
&
McDermid
S.
2019
Indian summer monsoon: extreme events, historical changes, and role of anthropogenic forcings
.
Wiley Interdisciplinary Reviews: Climate Change
10
(
2
),
e571
.
Sit
A. K.
,
Acharya
G. C.
&
Thomas
G. V.
2011
Arecanut Based Cropping Systems in Sub Himalayan Terai Region of West Bengal. Arecanut Based Cropping/Farming Systems
.
Central Plantation Crops Research Institute
,
Kasaragod
, p.
138
.
Published by Director Central Plantation Crops Research Institute (Indian Council of Agricultural Research)
.
Spence
A.
,
Poortinga
W.
,
Butler
C.
&
Pidgeon
N. F.
2011
Perceptions of climate change and willingness to save energy related to flood experience
.
Nature Climate Change
1
(
1
),
46
49
.
Tambo
J. A.
&
Abdoulaye
T.
2013
Smallholder farmers’ perceptions of and adaptations to climate change in the Nigerian savanna
.
Regional Environmental Change
13
(
2
),
375
388
.
Taraphdar
S.
,
Zhang
F.
,
Leung
L. R.
,
Chen
X.
&
Pauluis
O. M.
2018
MJO affects the monsoon onset timing over the Indian region
.
Geophysical Research Letters
45
(
18
),
10011
10018
.
The Times of India
2019
Snow After a Decade Drives Tourists to Darjeeling for a White Year-End
.
Available from: https://timesofindia.indiatimes.com/ (accessed 14 January 2022)
.
Theil
H.
1950
A rank-invariant method of linear and polynomial regression analysis, part 3
.
Proceedings of Koninalijke Nederlandse Akademie Van Weinenschatpen A
53
,
1397
1412
.
Uddin
M. N.
,
Bokelmann
W.
&
Dunn
E. S.
2017
Determinants of farmers’ perception of climate change: a case study from the coastal region of Bangladesh
.
American Journal of Climate Change
6
(
1
),
151
165
.
Venkateswarlu
B.
&
Singh
A. K.
2015
Climate change adaptation and mitigation strategies in rainfed agriculture
. In:
Climate Change Modelling, Planning and Policy for Agriculture
. (
Singh. A. K., Dagar. J. C., Arunachalam. A., Gopichandran. R., Shelat. K. N., eds.
).
Springer
,
New Delhi
, pp.
1
11
.
This is an Open Access article distributed under the terms of the Creative Commons Attribution Licence (CC BY-NC-ND 4.0), which permits copying and redistribution for non-commercial purposes with no derivatives, provided the original work is properly cited (http://creativecommons.org/licenses/by-nc-nd/4.0/).

Supplementary data