Climate change farm based autonomous adaptation measures and its impact on wheat crop productivity in Punjab, Pakistan

Climate induced hazards has increased production threats to agriculture where such environmental risks can signicantly reduced by adapting adaptation measures. In farm based adaptation measures specically the autonomous adaptation aspect has not properly focused particularly in developing countries like Pakistan. This study attempted to investigate on-farm autonomous adaptation (OFAA) measures of wheat farmers to climate change and estimate its impact on wheat yield and total return. This research work used the data of 480 wheat farmer from production based six categorized higher vulnerable districts of Punjab, Pakistan. Probit model was employed to examine adaptation decisions determinants of farmers and Propensity Score Matching used for to investigate the impact of OFAA practices on outcomes of wheat crop. Estimates indicated as wheat farmers to reduce the unfavorable effects of climate change have applied different OFAA strategies such as seed varieties changing, fertilizer management, variations in cultivation dates and management of supplementary irrigated water. Probit model estimates illustrated as climate information access, credit access, ownership of land, off-farm income and ownership of tubewell considered signicant determinants regarding adaptation decisions of farmers. Estimates of propensity score matching indicated as farmers those applied almost one OFAA strategy obtained 310kg per hectares higher wheat yield as US$82.75 per hectares return rather than those not applied any adaptation strategies. Adapter wheat farmers applied multiple adaptation strategies obtained higher yield and return rather than adapter farmers used limited adaptation strategies. Results imply as in developing sustainable farmers livelihood, adequate food production, reducing crops variability and climate change adverse impacts it is mandatory to applications of OFAA measures. The study also elaborated as for developing farm-level adaptation there is signicant role of climate information, farmers schooling, access of credit and adequate irrigation. required in (PSM)


Introduction
Climate change frequent dynamics in the couple of decades has increased severity and incidence of natural disasters such as ood, landslides, drought, cyclones and earthquakes (Teo et  Pakistan an agrarian economy where almost 70% population depends on agriculture for their livelihood it provides employment to 38.5% labor force and shares 19.2% GDP of the country (PBS, 2021). During the couple of decades because of frequent climatic variations country faced higher yield losses particularly the cereal crops which subsequently raised emerging issue of food security in the country PBS, 2020). In cereal crops, wheat is main cereal crop of the country which provides staple food diet to population (Ahmad et  Ahmad et al., 2020) whereas no research work in scenario of Pakistan have not addressed the aspect of autonomous adaptation application for local poverty reduction and food security. In developing country like Pakistan where State based institutions are not capable to support farming communities, farming community least resilient and agriculture higher susceptible to climate variation adaptation on farm level and its usefulness have higher signi cance such aspect need to address. In addressing this research gap this research work paying attention (a) to investigate wheat farmer's on-farm climate change autonomous adaptation strategies (b) to estimate impact of adaptation on wheat crop yield and total return (c) to investigate its applications for local poverty reduction and food security. This study is subdivided in to four segments as introduction illustrated in rst segment, second segment discussed material and method. Third segment of the study highlighted results and discussion whereas last segment elaborated the conclusion and suggestions.

Selection of study area
Punjab province is publicly recognized the land of ve rivers, augmented with fertile lands and located in central region of the country (GOP, 2021). Punjab among four provinces of the country more preferably focused for this research because of several considerable bases. Firstly, Punjab is foremost agricultural GDP contributor 52% and represents 53% population of the country (PBS, 2020). Secondly, Punjab by producing ¾ of country's cereals production is known food basket of the country (GOP, 2021). Thirdly, Punjab produces 77% wheat of the country (BOS Punjab, 2020) whereas currently facing consecutive and signi cant yield reduction owing to induced factors of climate change. Lastly, wheat production is more susceptible in the region due to consecutive increase in erratic rainfall, rising temperature and frequent hailstorms in recent decades (PBS, 2021). This scenario has severe threat for livelihood of millions of wheat farmers' households and food security relying for their employment and nutrition in the region. Kharif and Rabi are two major cropping seasons comprising with major and minor crops (BOS Punjab, 2019). Based on share of province in wheat production six districts were chosen with various wheat production based categorizations. In scenario of wheat production, Dera Ghazi Khan and Layyah from low wheat producing districts, Muzaffargarh and Vehari from medium while Rahim Yar Khan and Bahawalnagar from high wheat producing districts were purposively chosen for the study as indicated in gure 1(BOS Punjab, 2021). In the study, all six selected districts, risk variations experience, socioeconomic, geographical, climatic and structural characteristics were documented. In provincial environment climatic variation were estimated as cold in winter and hot in summer having the average temperature of 33.9°C in summer and average temperature of 8.7°C in winter (PMD, 2019). In Punjab, rainfall disperse pattern was estimated where 68% erratic and routine rainfall mostly expected in monsoon season (PMD, 2019; BOS Punjab, 2020). Major cash (sugarcane and cotton) and cereals crops (wheat, rice and maize) major share is produced in Punjab province which severely affected from climate induced factors (PBS, 2021).

Data collection and sampling framework
In sample selection of wheat growers from the study area multistage sampling method was used by involving random selection and strati ed approach. In the rst stage, Punjab more preferably considered because of higher wheat production province in the country. In the second stage, in using the strati ed sampling approach six districts were chosen according to categorized low, medium and high wheat producing areas of Punjab. In particular, districts were categorized in province such as low wheat producing zone (below 600,000 tonnes), medium wheat producing zone (600,000 to 900,000 tonnes) and high wheat producing zone (above 900,000tonnes). In each categorized wheat production zone, two districts were randomly selected in the third stage. In fourth stage, two tehsils from each district were randomly chosen and from each tehsil two union councils were randomly chosen.
In the last stage, from each union council two villages were randomly chosen and ten wheat farmers were randomly preferred for data collection comprising 480 total wheat farmer from study area. From each province eighty wheat farmers were interviewed as sampling framework illustrated in Table 1.
[ Table 1] In the study area, a well-structured questionnaire was applied for face to face interview and data collection from sample household's wheat farmers. To ensuring the accuracy and relevancy of the data a pre-tested survey was conducted prior to collecting the data from study area. Farmers farming practices, socioeconomic characteristics, farm outcomes related to production outcomes and adaptation practices were main feathers included in the questionnaire. Author himself and four enumerators the students of COMSATS University Vehari prior trained for data collection in the study area. Data collection regarding farmer's response about wheat crop outcomes and adaptation practices and data collected from July 2020 to October 2020. In the scenario of study objectives and usage of collected data farmers were well informed and motivated to participate in data collection with accuracy of information. Farmers involve themselves warmly about data collection questionnaire and 27 participants refused to take part which were replaced to others farmers from study area.

OFAA evaluation impact on wheat productivity
Crop return and production is feasible to increase through using the effective adaptation measures. In estimating the adaptation impact on wheat production, wheat farmers were rstly categorized in to two types' non-adapter and adapter farmers. Non-adapter were considered those farmers who have not applied any OFAA adaptation measure and adapter those who at least applied single adaptation measure for wheat crop to reduce climate change severity. In such estimation, in the Table 3 (1) In the above equation propensity score elaborated as p whereas X ik indicated the observable attributes, whereas pr as adaptation probability to climate change U i . There is similar conditional to distribution regarding to non-adapter and adapter in the model of propensity score matching (Frolich, 2007;Thavaneswaran and Lix, 2008).
In measuring the treatment impact on variables outcomes propensity score matching estimation engaged in to ve stages. In the rst stage, related to theoretical assumptions initially selected the list of observable covariates (socioeconomic threats of farmers) for pre-testing. Secondly, propensity score or probabilities are estimated through observable covariates of regressing related to outcome variables (adaptation determinants) (Ali and Abdulai, 2010). In the third stage, matching method is applied to anticipated propensity score matching of non-adapter and adapter (Kassie et al., 2011). Treatment casual effect (adaptation) is estimated the variables outcome (wheat yield) in the fourth stage (Rubin, 2001). In the last stage, propensity score evaluation estimation is in uence related to climate change strategies adoption. Attributes of institutional services, farm status, farmers socioeconomic characteristics were chosen as observable covariates. In Table 2, list with details of observable covariates of farmers, with measurement unit and description has elaborated. Explanatory variables impact on adaptation decisions with expected direction indicated with the associated sign. [ In above equation, in population average treatment effect (ATE) indicated as t, control group as Y 2 and treatment as Y 1 which is this study related to non-adapter and adapter wheat growers. In such scenario as indicated prior ATE, as estimates may be biased. Furthermore, ATT can be observed that illustrates the precise estimates after contrasting the treatment and controlled groups. Average treatment effect on treated (ATT) as illustrated in the equation below In above equation, ATT reported as notion T whereas p(X) as after calculated propensity score. In sequence of calculating the p-

Farmers socioeconomic characteristics descriptive statistics
Climate change strategies non-adapter, adapter differences and pooled sample farm base and socioeconomic characteristics mean values illustrated in Table 3. In study area majority of wheat farmers 78%(374) adopted OFAA climate change adaptation measures whereas adapter were relatively younger (45 years), higher schooling (8years) rather than average schooling (7years) and age (48 years) in sample area. Adapter farmers have higher household size (8members) and primary occupation of agriculture (88%) rather than average household size (7members) and primary occupation (75%) in study area. Almost 49% non-adapter farmer's considered agriculture primary source of family income so less conscious to manage adaptation rather than adapter farmer. In study area, average land holding (7acres) and 90% farmers having land ownership whereas adapter farmers have higher land holding (9acres) and 97% were owner of land illustrating as adapter have higher ownership and holding land than non-adapter farmers. In study area, limited irrigation 13.7% feasible with canal water whereas 86.3% managed by farmer through tubewell because of that majority 90% of adapter farmers having ownership of tubewell rather than 32% non-adapter farmers. These illustrations indicated as water supply enhances the farmer's capacity of managing climate risks of wheat crop such as in winter shortage of water is managed through tubewell water supply.
[ Table 3] In study area, adapter farmers considered more resourceful regarding assets as holding average 8animals and farm income almost 49 thousands monthly higher than non-adapter with livestock (4.32 animals) and 24 thousands monthly income. Climate information access, advisory services and credit access considered signi cant variables regarding institutional services in the study area. Information accessed through formal and informal sources whereas farm based formal sources organizations/institutions by private and public considered more feasible in this study. Estimates illustrated as farmers have higher access of climate change information 60% and advisory services 57% rather than credit access 36% because of higher advancements in technologies such as mobile phones and internet access. Mostly organizations and institutions provide advisory services and climate information through SMS/WhatsApp applications which becomes more feasible for farmers (Khan et al., 2020; Ahmad and Afzal, 2021). These results illustrated as adapter farmers as contrast to non-adapter farmers more resourceful regarding assets because of higher more access of information and higher income due to more consideration regarding climate change adaptations and access higher yield of crops (Zhai et al., 2018).

Farmer adapted OFAA strategies
In the study area to overcoming the climate change risks farmers applied multiple OFAA adaptation strategies as particularly indicated in the gure 2. Overall wheat farmers more speci cally focused OFAA seed variety changing (73%), managing fertilizer Farmers from higher and medium wheat production areas more preferably focused the changing crop varieties, managing fertilizer and supplementary irrigation climate smart adaptation strategies whereas low wheat production farmers preferred to plot resizing, changing irrigation time and harvesting and cultivation changing. Adapted based strategies measures of wheat farmers were classi ed in to four categorizes as higher adapter (applied>4 adaptation measures), medium adapter (applied 3 or 4 adaptation measures) small adapter (applied 2 adaptation measures) and non-adapter (not applied any adaptation measure) for assessing diversi cation application measures. Majority of farmers (34%) were higher adapter, (28%) medium adapter and (16%) small adapter as illustrated in gure 3 illustrating as majority of wheat farmers in the study area have applied multiple adaptation measures to climate change. Almost one-fth (22%) wheat farmers in the study area not focused any climate-induced adaptation measure in the study area. Higher frequency of climate-induced non-adapter and small adapter was estimated in the low production region whereas majority of medium and high adapter farmers belong to high and medium wheat production region. These estimates indicated as in low yield districts farmers having limited diversi cation adaptations measures whereas higher diversi cation adaptation among wheat farmers was estimated in medium and high wheat yield production districts which alternately causes the yield differences in the various districts and regions.

Adaptation impact on wheat crop yield and return with cost-bene t analysis
Adapter and non-adapter wheat crop average outcomes comparison illustrated with major differences of wheat yield and return of total crop as indicated in Table 3. Signi cant variation regarding non-adapter and adapter estimated related to input cost that inspires to prerequisite the analysis of cost bene t. Moreover, for estimating the diversi cation adaptation impact on outcome of crop, comparison was estimated means of wheat yield, cost of total input, return of total crop and gains of total pro t in four categorized farmers groups and generated the results. Wheat farmers were categorized in to four groups for estimating pro tability across intensity of adaptation as such measure were taken for estimating adaptation pro tability in the research work of Arfanuzzaman et al., (2021).
[ Figure 4] In study area among various types of climate induced strategies adapter farmers the mean (maund per acre) of wheat yield and total return illustrated in gure 4 and 5. Estimates indicated as wheat crop yield and total return per acre increased from small to medium and higher adapter because higher intensity of adaptation strategies increases yield of wheat production and total return.
Remarkably, more prevalent increasing trend related to districts of higher production zone that is the reason of higher adaptive capacity to application multiple adaptation strategies. Another reason is that farmers in high yield region having higher nancial resources so allocate additional nancial resources for crop inputs. Additional resources allocation for farm inputs particularly based on adaptation of diversi cation causes ultimately return their input cost.
[ [ Table 4] In farming practices irrigation access through canal and tubewell considered more essential factor regarding wheat cultivation as illustrated the more positive and signi cant association with farmer decisions of adaptations. Estimates of marginal effect indicated as wheat farmers with own tubewell irrigation access implement 19% more climate induced adaption measures rather than those farmers having no own irrigation access. In winter season shortage of irrigation water adversely effects wheat crop yield so adequate irrigation water access to wheat crop through tubewell considered main climate change adaption strategy for higher yield and coping severe effects of climate change as results are consistent with the research work of Alauddin

Adaptation causal effect on wheat crop yield and return
Propensity score was rstly estimated and then NNM matching technique was applied to estimate the casual in uence of adaptation (treatment) on non-adapter (control) and adapter (treated) group. Table 5  [ Table 5] In the same scenario ATT has also reported as treatment group also obtained raised wheat yield crop 3.14 maunds per acre and 310.23kg per acre whereas wheat crop total return PKRs 6927 per acre and US$82.75 per hectares as contrast to controlled groups.
ATT lower value is because of potential bias reduction which affects impact of adaptation on wheat yield. ATT illustrated that adapter farmers regardless of their socioeconomic status mostly raised wheat yield and obtained higher total wheat crop return rather than non-adapter wheat farmers in the study area. Adaptations measures reduces the severe impacts of climate change also positive and signi cantly in uences the wheat crop yield and total wheat crop return as empirically justi ed from these estimates in

Propensity score evaluation
In the scenario of estimating casual in uence of treatment on variables outcomes it is prerequisite to evaluate results adequacy. In such method, rstly values of p>chi 2 , LR chi 2 and Pseudo-R 2 were estimated after and before matching the p-score p(X). In Table 6, statistics regarding test illustrated that values of p>chi 2 , LR chi 2 and Pseudo-R 2 signi cantly decreased afterward matching propensity score of respondents. Subsequently, median and mean biases also estimated to evaluate the results adequacy in term of reducing biases of selection.
[ Table 6] Estimates illustrated median biases reduced 59 to 7% and mean biases reduced from 64 to 10% elaborating signi cant reduced bias of external and internal covariates.  Table 7 as given below. [

Conclusion And Suggestions
In Pakistan millions of households livelihood and nutrition based on wheat cultivation whereas in the current decade notable decline in wheat crop production estimated because of severe climate change-induced factors such as erratic rains, dynamic temperature and extreme natural events. In the region, wheat productivity can mitigated from climate-induce severe impacts by implementing on-farm autonomous adaptations (OFAA). This study attempted to investigate farmers OFAA strategies and evaluated their impact on wheat crop production and its returns by application of collect data from production based lower, medium cannot be generalized to other districts because of some regional, environmental and socioeconomic disparities.

Funding
This study has no funding from any institution or any donor agency.
Competing Interests        Climate change adaptation intensity its impact on wheat yield