Adaptation to Climate Change among Farmers in Drought-Hit Region of Khyber Pakhtunkhwa, Pakistan


 The agriculture sector is the most important, however, most vulnerable to climate change in Asia and the Pacific. Droughts are one of the leading sources of disaster risks that can alter yield levels and cause sizeable productivity losses in agricultural products. In response, the farmers are adopting several climate risk management strategies. Therefore, this study aims to find out farmers’ preferences for climate change adaptation strategies and their potential determinants. Data were collected from randomly selected 200 farmers in District Nowshehra, Khyber Pakhtunkhwa. A multivariate probit model was used to assess the impacts of various factors on farmers’ decisions to adopt a particular adaptation strategy. The findings of the study revealed that rainwater harvesting, change in the planting dates, soil conservation, ponds building, and terraces with spillways were the dominant adaptation strategies found to combat the ill effects of droughts in the study area. Furthermore, the study highlighted the potential role of socio-economic attributes in the adoption of these adaptation strategies. The findings suggested that the government may extend its support to these farmers in making ponds and providing them the storage of water facilities.


Introduction
Climate change is one of the most challenging issues of the 21st century having widespread impacts on the societies and environment (IPCC, 2018). It is posing a serious risk to the global community particularly the developing nations who are continuously facing a vulnerable situation with the increase in intensity and frequency of extreme weather events (Chowdhury et al., 2020). A recent report by the IPCC, (2018) highlighted that by 2050, around 150 million people will be displaced due to climate change-induced phenomena around the world. In the last 50 years, around 79% of the global disasters involved water, weather, and climate-related hazards are accounted for 56% of mortalities with a 75% share in the economic loss (IFRC, 2020). The last decade (2010-2019) saw an increase in the frequency of weather and climate change-related disaster events by around 9% and 14% compared to the previous decades i.e. 2000-2010and 1990-2000, respectively (IFRC, 2020. Drought is among the many climatic extremes that possess severe impacts observed all over the world with an increasing trend of warming. The demand for precipitation is expected to increase globally owing to the increase in warming conditions (Cook et al., 2014). Unlike other hazards such as oods, earthquakes, and tsunamis, drought can occur anywhere in the world. Furthermore, drought evolves slowly ranging from few months to years of prolonged conditions of water scarcity. The direct consequences of short-term droughts may be loss of crop production, poor pasture growth, and decline in fodder supplies along with long-lasting impacts such as forest res (Mishra and Singh, 2010). Droughts are considered the most damaging of all hazards owing to their long-lasting socio-economic impacts (Sivakumar et al., 2014). For instance, in the case of United States, drought causes an average of about US$ 6 to US$ 8 billion per year (NCDC, 2015). The recent statistics of the average annual economic impact from drought in Europe suggest an increasing trend with an estimated average annual economic loss of €6.2 billion per year (EEA, 2010). Climate-induced hazards such as oods, droughts, res are the most damaging for the agriculture sector in Pakistan (Ullah Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js and Shah, 2019). The agriculture sector in Pakistan in general and in Khyber Pakhtunkhwa (KP) in particular is confronting with severe oods and weather-related risks (Ahmed et al., 2019).
Droughts impact all sectors of the economy. However, the agriculture sector is the most vulnerable sector to climate change in Asia and the Paci c (ADB, 2009). Climate-induced disasters affect the availability of water to farmers in many regions of the world (Klopper et al., 2006). The risks faced by farmers due to climate change ranges from personal, nancial, social, environmental, and others (Iqbal et al., 2016). Farmers' are required to adapt their agricultural activities to maintain their production and reduce their vulnerability to climate change (Hassan and Nhemachena, 2008). The adaptation of farmers to climate change is in response to their perceived vulnerability with the intention to reduce risks associated with farming (Arbuckle et al., 2013). Climate-induced disasters and risks might bring serious negative consequences for agriculture and people associated with agriculture for their livelihood (Gerkensmeier and Ratter, 2018). Therefore, adaptation strategies in the local context are very important to be studied in developing countries like Pakistan.
Keeping in mind the importance of agriculture to the economy and rural livelihoods, the signi cance of climate change adaptation strategies is crucial (Ali and Erenstein, 2017). Adaptation in the agriculture sector involves practices against changes in the climatic situation. These practices may be in the form of anticipatory adaptation, autonomous adaptation, planned adaptation, even public and private adaptation, and reactionary or reactive adaptation practices (Roka, 2019). The farm-level adaptation strategies are essential for effective policymaking as a tool to manage agriculture-related risks arising from climate change (Hassan and Nhemachena, 2008). Adaptation in the agriculture sector is also important for reducing the impacts of increasing temperature and precipitation across the world (Pradeep and Mendelsohn, 2007). Thus it is equally essential that efforts should be made to increase the resilience and adaptive capacity of a system to overcome the negative impacts of climate change with limiting consequences and damages (Food, 2003). Among the many policy options, adaptation is one of the key options for reducing the harmful consequences of climate change on agriculture (Ozor et al., 2012). An important factor for informing policies aimed at promoting successful adaptation strategies in agriculture is to understand farmers' perceptions regarding climate change, their adaptation strategies, and related factors (Jianjun et al., 2015).
There is extensive literature available on farmers' risk preferences and their role in decisions related to agriculture production (Feder, 1980;Just and Zilberman, 1983;Adger et al., 2009). Studies on risk preferences and adaptation to climate change indicate that the socio-cognitive processes and risk preferences of those involved in decision-making play a crucial role in the promotion of adaptation decisions (Frank et al., 2011;Tam and McDaniels, 2013). Therefore, policymakers need to understand the linkage between farmers' risk preferences and how they respond to the climate change adaptation in order to design effective adaptation strategies (Chavas et al., 2010 and in-depth insights into the practices of farmer's decision-making under risk. The ndings of the study will guide the government agencies to help these farmers in drought situations.

Study Area and Sampling
The study in hand was conducted in the Khyber Pakhtunkhwa (KP) province of Pakistan. KP province was purposively selected for this study as farmers in KP face serious challenges of low water use e ciency and lack modern technology, skills, and knowledge to engage in high-value agriculture value chains (World Bank, 2019). Therefore, there is a need for an improved and more climate-resilient community and water management system in KP (World Bank, 2019). District Nowshera of the KP province was purposively selected due to its lowest irrigated land among the ve central districts of Khyber Pakhtunkhwa (Government of Khyber Pakhtunkwa, 2014). According to population census 2017, the Nowshera district has a total population of 1,518,540. Among the three Tehsils (sub-districts), of Nowshera, Tehsil Pabbi was purposively selected as the study area for the present research, and data were collected from three purposively selected Union Councils (UCs) namely Spinkhak, Jaroba, and Jalozai Mera as these union councils are mostly exposed to droughts (Adaptation Fund, 2019). Based on the data from population census 2017 and using Yamane, (1967) formula (Eq. 1), a total of 202 sample respondents were selected.
Among the three villages, the sample was proportionally distributed. A list of farm households was prepared from the voter lists, villages heads, and village council Nazims. Farmers were randomly selected from the list for the survey. Two respondents were dropped from the analysis due to incomplete data. Where: N = population size, n = sample size and e = precision value set as 5%.

Data Collection
Data were collected through a semi-structured questionnaire. Data were collected on farmers' socioeconomic and demographic factors and their perceptions about climatic risk sources.

Data Analysis
Behavioral response models concerning more than two possible outcomes are either multinomial or multivariate probit models. A total of eight strategies were explored the farmers were practicing in the study area. These were rainwater harvesting, soil conservation, change in planting date, terraces with spillways, ponds, livestock, multiple cropping, and modern irrigation techniques. However, based on frequencies of adoption of the mentioned adaptation strategies among the selected sampled farmers, only ve adaptation strategies (rainwater harvesting, soil conservation, change in planting dates, terraces with spillways and ponds) were considered for the present research study. If the adaptation strategies have any possible contemporaneous relationship, then multivariate probit models are used (Ullah and Shivakoti, 2014). A Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js multivariate probit model (Eq. 2) was employed to assess the impacts of various independent variables on the decision to adopt the adaptation strategies among the sampled farmers.
Where Yij ( j = 1,...,m) represent the risk management alternatives (in our case m = 5 ) faced by the i th ij is a 1×k vector of observed variables that affect the decisions to adopt a particular adaptation strategy, β j is a k ×1 vector of unknown parameters (to be estimated), and ϵ ij is the unobserved error term. In this speci cation, each Y j is a binary variable and thus Eq. 2 is actually a system of m equations (m = 5 in this case) to be estimated: are ve latent variables underlying each of the adaptation strategy's adoption decisions such that y j = 1 if y * j > 0; and 0 otherwise.

Dependent Variables
The focus of this research is on ve dominant strategies found in the study area i.e. on-farm rainwater harvesting, soil conservation tillage, change planting dates, terraces with spillways (TSP), and ponds.
These variables were binary in nature: 1 = adoption, 0 = no adoption.

Independent variables
Research studies have reported that several socio-economic factors in uence the adaptation strategies in agriculture in a different context. The socio-economic attributes considered for the present study include

Ethical Consideration
The study has been approved by the ethical research committee of Institute of Management Sciences Peshawar. Moreover, written consents were obtained from the study particpants for taking part in interviews.

Descriptive Statistics of the study variables
Summary statistics of the variables considered for the study are presented in Table 1. The ndings revealed that the most prevalent adaptation strategy to cope with the adverse impacts of droughts was rainwater harvesting, with a mean value of 0.86 is ranked as the most dominant strategy to combat the ill impacts of drought in the study area ( Table 1). The second most dominant strategy was reported as changing planting dates with a mean value of 0.67 (Rank II). Soil conservation, ponds, and TSP were also practiced by the farmers. It implies that the farmers make preferences when they are adopting different strategies to combat the climate change induced droughts. Most of the farmers in the study area were of a higher age as shown by the mean value of 49.80 years. Less than half were literate (who can read and write) and experienced.
The average distance from home to the farm was 705.5 meters. Dependent family members were 3.8 per household. Income from the farm and non-farm activities were falling in the range of PKR.5000-10,000: Rented land was less (18%) and mostly farmers (mean = 0.82) owned the land. The respondents knew about climate change.
Loading [MathJax]/jax/output/CommonHTML/fonts/TeX/fontdata.js     Shah, (2019). They revealed that educated farmers were more likely to practice adaptation strategies to climate-induced oods in Pakistan. Off-farm income was positively associated with soil conservation practices in the study area (Table 3). Off-farm income is an additional source that enables the farmers to adopt climate change (FAO, 2014). Similarly, land ownership was positively (0.146) associated with soil conservation practices in the study area. The ndings for land ownership are in agreement with Freudenberger and Miller, (2010). They revealed that land ownership and property rights are important determinants of climate change adaptation. Likewise, the variable 'erratic rain' was also found to be positive and signi cant with coe cient value of 0.463 and p-value < 0.10. Perceived rains as source of risk is important determinant of drought adaptation strategy. The farmers who considered rains a major source of risk in the study area, they more likely to adopt the soil conservation. Likewise, the farmers were using drought resistant seeds and were practicing soil conservation for the better crop production. Moreover, the farmers have gained from their counterparts that climate adaptation increases their crop production. Therefore, they were going for soil conservation as adaptation strategy to droughts. Table 3

Factors In uencing Ponds
The results of the model show that age has a negative association with the pond as an adaptation strategy to droughts. It is argued that as the farmers' age increases, their knowledge about agricultural practices also increases (Ullah, 2014

Factors In uencing Terraces with Spillways
Soil water management practices can increase resilience in the context of climate change (Kosmowski, 2018). in uenced adaptation to climate-induced oods in Pakistan. The landholding size of farmers positively in uenced their adoption of terraces with spillways. It implies that farmers with large landholdings are more likely to adopt this strategy. Knowledge about climate change was found positive and signi cant with a coe cient value of 0.554 and p-value < 0.05. Those farmers who had knowledge about climate change were practicing terraces with spillways. According to Olazabal et al., (2018), successful adaptation to climate change needs knowledge, and understanding the complex structure and functioning.
Climate change is a real threat to farmers around the globe. Agriculture is highly dependent on including high and low temperatures, rainfall, good weather, wind intensity, and many other variables. Estimates show that climate change might reduce global agriculture productivity up to 17% by 2050 (Calmon and Feltran-Barbieri, 2019). Climate change adaptation is the action to global warming, which helps to reduce the vulnerabilities in the social

Conclusion
Climate-induced disasters are very common around the world and affecting human societies in several ways particularly through variation in the productivity of the agriculture sector. In response to these climate risks and the drought-affected areas, the farmers are practicing several risk management activities.
Farmers make choices of using strategy, which is easily and readily available at the time of climatic risk situations. Sometimes, they are making combinations of these strategies. The study areas are prone to severe droughts and are mostly rain-fed, having no irrigation system. It has been explored that the age of the farmers, their literacy, experience, dependent family members, landholding size, and land ownership, perceptions of erratic nature of rains, and climate change knowledge are important determinants of climate change adaptation strategies. Five dominant strategies were identi ed among the eight available strategies in the study area. These strategies were ranked and found in order as rainwater harvesting, soil conservation, change in the planting dates, building ponds, and terrace with spillways. The ndings are helpful for the government in policymaking and designing programs to support farmers when they are in a drought situation. The government may extend its support to these farmers in making pond and water storage facilities.

Declarations Con ict of interest
Authors declare that they have no con ict of interest.