Earth Syst Environ (2017) 1:19 https://doi.org/10.1007/s41748-017-0021-y

REVIEW ARTICLE

A Review of Food Security and the Potentials to Develop Spatially Informed Food Policies in Bangladesh

1,2 1 Iffat Ara • Bertram Ostendorf

Received: 30 June 2017 / Accepted: 24 October 2017 / Published online: 8 November 2017 Ó Springer International Publishing AG 2017

Abstract Keywords Bangladesh Á Food security Á Food availability Á Background Food security globally depends primarily on Food access Á Food utilization Á Climate change Á Policy three components: food availability, food access, and food utilization. Regional variations of these components may affect food security via spatial differences in natural, social 1 Introduction or economic conditions and the interaction of these in a complex environmental system. Food security is a complex and important global issue that is Purpose It is important to understand the regional varia- affected by environmental system and human actions tion of food security, particularly where and under what (Simelton et al. 2012;TiwariandJoshi2012). Many natural natural and socio-economic circumstances people become variables, including climate change, may influence food vulnerable to low food security in a country. production, while socio-economic variables are mostly Methods This article provides an overview of food security responsible for food access in a country (Burke and Lobell in Bangladesh in terms of the three main components, 2010; Harvey and Pilgrim 2011). One most recent estimate identifies knowledge gaps in present food security research, identifies about 795 million undernourished people in the reviews possible impacts of climate change on food secu- world, with one-third of these people living in South Asia, rity, and sourced a wide range of spatio-temporal data including Bangladesh (FAO 2015). Bangladesh is facing relevant for food security. challenges in attaining and maintaining food security because Results The study highlights potentials and indicates dif- of its large population, numbering 152 million at the last ferent processes to develop spatially informed food policies population census from 2011 (BBS 2011a). In addition, in a country, particularly focuses on Bangladesh. This will numerous drivers including poverty (Kam et al. 2005), mal- contribute to improved food security by considering nutrition (Bose and Dey 2007;Hossainetal.2005), frequent regional food security conditions, region-specific deficits, disasters (Mirza 2002; Ninno and Lundberg 2005), agricul- climate change, other future risks, and devises actions tural management practices (Ara et al. 2016) and the effect of related to the respective components. climate change (Karim et al. 1996;Sarkeretal.2012; Ruane Conclusion The study concludes that different processes et al. 2013;Araetal.2017) influence food security across the can provide a foundation for policy development and these country (Habiba et al. 2015a). will advance research-policy linkage to improved food According to the Food and Agricultural Organization security. (FAO), a country achieves food security when it ensures adequate food along with a nutritious diet for the entire & Iffat Ara population (FAO 2006). Thus, the concept of food security [email protected] indicates a situation in a particular place and time where 1 School of Biological Sciences, The University of Adelaide, people live without hunger or fear of undernourishment. Adelaide 5005, Australia Food security is constituted by the three components of 2 Department of Geography and Environment, Jahangirnagar food availability, food access and food utilization (FAO University, Savar, Dhaka 1342, Bangladesh 2006). Until 1980, food availability was seen as the key 123 19 Page 2 of 18 I. Ara, B. Ostendorf component to ensure food security in Bangladesh and the (MoFDM 2008), in particular, do not address regional details government prioritized food production to overcome food effectively to reduce food security challenges in a complex shortages (Elahi and Ara 2008). However, Sen’s (1981) environmental system in the country. A spatially informed analysis clearly indicates that food availability is not the food security assessment is needed to advance the research- only component of food security in the country. Low food policy linkage in the country by considering such a system. security can result from an individual’s lack of ability to The objectives of this paper are: (1) to provide an overview of access enough food because of poor socio-economic con- food security in Bangladesh; (2) to examine the national data ditions or inadequate food distribution in a place. All food sources relevant for food security; and (3) to address a way security components are now considered as important in forward for a spatially informed food policy development. existing food policies in Bangladesh (MoFDM 2006; Ahmed 2015). Globally, several studies have analyzed the food security 2 Food Security in Bangladesh components individually, leading to improved food policies in their respective countries (Ray 2007;Rocha2009; Bashir and Bangladesh is located in the Ganges–Brahmaputra delta Schilizzi 2015). Similarly, in Bangladesh food security has with an area of 147,570 km2 (BBS 2014). At present, the been analyzed in terms of food availability (Mainuddin and administrative geography of Bangladesh is divided into Kirby 2015; Parvin and Wakilur 2009), food access (Hossain eight major administrative divisions, 23 greater region, 64 et al. 2005;Thorne-Lymanetal.2010,BoseandDey2007; districts (zilas), 545 sub-districts (489 upazilas and 56 Faridi and Wadood 2010) and food utilization (Pitt 1983; thanas), and 4550 unions (GoB 2016). Figure 1 indicates Hels et al. 2003;RahmanandIslam2014). Whilst these the administrative boundaries of Bangladesh (greater findings successfully inform food policy, the full complexity region, districts, and sub-districts). Food security varies of food security cannot be understood and addressed effec- across Bangladesh and administrative levels, particularly tively using the components individually for a better policy sub-districts, are typically used to show the regional (Pinstrup-Andersen 2009). Some assessments (Mainuddin and Kirby 2015; Parvin and Wakilur 2009)haveconsidered the temporal dimension of food security in Bangladesh, but the spatial aspects of food security have mostly been over- looked. Spatial patterns of food security including these components have received attention from food security scholars globally (Belesky 2014;Ray2007;Iizumietal. 2013). However, existing food security studies for Bangla- desh are not adequately addressing the spatio-temporal dynamics of food security components in the country. To strengthen food security research in Bangladesh and in many countries globally spatial analysis is required to assess all food security components. Diverse regional and environmental conditions affect food security, but these have received insufficient attention in terms of both research and food policy in Bangladesh. Food production varies due to natural conditions (including climate change) and socio-economic conditions (Faisal and Parveen 2004). Climate change may affect food production in a spatial manner in Bangladesh (MoEF 2005, 2009). Many studies have principally focused on crop production to understand the impact of climate change on food availability (Karim et al. 1996;Sarkeretal.2012;Ruaneetal.2013; Parvin and Ahsan 2013;Sarkeretal.2014;Aminetal.2015). Hence, these studies have substantial limitations in considering climate change impact on regional food security. Beside climate change, other regional factors such as bio-physical, manage- ment, and socio-economic conditions may affect food secu- rity. Both the National Food Policy (NFP) (MoFDM 2006) Fig. 1 Administrative boundaries of Bangladesh. Source: BBS and the National Food Policy Plan of Action (NFPPA) (1981, 2014) 123 A Review of Food Security and the Potentials to Develop Spatially Informed Food Policies in… Page 3 of 18 19 variations of food security (GoB and WFP 2004). This section of the paper documents food security in Bangladesh and highlights food security conditions including: (a) spa- tial variations of food security, (b) food availability, (c) food access, and (d) food utilization. a b 2.1 Spatial Variation of Food Security

The Food Security Atlas of Bangladesh compiled by GoB and WFP (2004) represents the first major initiative to assess regional variations of food security at the sub-dis- trict level in Bangladesh. The atlas uses indicators such as low agricultural production, limited infrastructure, poverty, seasonal unemployment, limited public and health facili- ties, unawareness regarding hygiene and sanitation prac- c tices, and the frequent occurrence of natural hazards. The data used in the atlas for map production were collected from the Bangladesh Population Census, 2001 (BBS 2001), the Local Government Engineering Department (LGED), d and the Bangladesh Agricultural Research Council (BARC). To visualize the overall spatial pattern of food security, these indicators were ranked at sub-district level Relative food security from 1 to 511 (470 upazila and 41 urban thanas included) Very low and summed. The resulting map (Fig. 2) shows distinct Low clusters of very low to low levels of food security in the Medium 05010025 Km following regions: (a) the Northwest, (b) the Sylhet Basin, High (c) the Chittagong Hill Tracts (Chittagong, Bandarban, Rangamati and Khagrachhari), and (d) the Coastal Fig. 2 Spatial distribution of levels of food security in Bangladesh at sub-district level, 2001. The key areas of very low to low food Regions. Whilst the atlas is very useful for understanding security, a the Northwest, b the Sylhet Basin, c the Chittagong Hill the spatial variations of food security determinants across Tracts (Chittagong, Bandarban, Rangamati and Khagrachhari), and the country, the maps are based on 2001 census data d the Coastal region of Bangladesh. Source: GoB & WFP (2004) without subsequent updates and without the possibility of understanding the spatio-temporal dynamics of different (particularly rice, wheat, and maize) by assuming national indicators. The regional context was represented in the estimates of seed, feed, and wastage, 11.58% (MoF 1991). atlas by using the household data aggregated to respective The Food Grain Requirement can be estimated from pop- sub-districts, but this did not show the variations of food ulation numbers by assuming a constant grain consumption security indicators within individual sub-districts. per day per person (here we used, 16 oz or 453.66 g; FPMU 1991). Both the NFGP and FGR increased between 2.2 Food Availability 1971 and 2010 (Fig. 3). Bangladesh closed the food gap in 1999 when the food grain requirement dropped below the Food availability refers to the sufficiency of food that is net total food grain production (Talukder and Chile 2011). supplied through domestic production or imports (includ- This demonstrates the success of government initiatives to ing food aid) (FAO 2006). Both food production and food overcome the food gap by raising local food production, requirements in Bangladesh have changed over time. particularly rice production. Bangladesh typically depends on grain, particularly rice, The national production of rice substantially increased wheat, and maize, which meet the majority of food from 9774 tons in 1971–72 to 33,889 tons in 2011–12 requirements. When national food production does not (BBS 2012a). This was primarily due to the introduction of meet the national food requirement, a nation experiences a High Yield Varieties (HYV) from the early 1970s (Deb food gap. The persistence of a food gap in Bangladesh until et al. 2009) and an increase in irrigated land for rice cul- 1999 can be visualized by comparing the Net total Food tivation since 1991 (Parvin and Wakilur 2009). The area Grain Production (NFGP) with the Food Grain Require- under irrigation increased from 1.64 to 6.15 million hec- ment (FGR), shown in Fig. 3. Net total Food Grain Pro- tares between 1981 and 2012 (BBS 2012b). Approxi- duction is estimated from local food production mately, 65% of the cultivated land was irrigated in 2012, 123 19 Page 4 of 18 I. Ara, B. Ostendorf

Fig. 3 Trends in national Net

Total Food Grain Production ton) ( 30000 Net total food grain production (NFGP)

(NFGP) and Food Grain t n

Requirement (FGR) of e Food grain requirement (FGR) m

Bangladesh over the period of e r i

1971–2010 u 25000 q e r

n i a r g

d 20000 o o f

d n a

n

o 15000 i t c u d o r p

n 10000 i a r g d o o f

l 5000 a t o t

t e 1971 1980 1990 2000 2010 N contributing substantially to increased production; pre- Regardless of these positive achievements, undernourish- dominantly rice (Rahman and Mondal 2015). Cropping ment still exists in the country, as 60 million people are not intensity also increased from 151 to 183% for the same able to consume the minimum daily food intake required period (BBS 2011b). Local food production varies for a healthy life (FAO 2013). Furthermore, levels of regionally due to environmental conditions and varied poverty vary substantially across the country (BBS et al. management practices, including cropping intensity and 2010; Kam et al. 2005). The poorest areas are the North- irrigation (Brammer 2000; Rahman 2010; Alauddin and west, the Chittagong Hill Tracts and the Coastal Region of Sharma 2013). However, these were not considered in Bangladesh, where food security was found to be critically previous food security assessments (Mainuddin and Kirby low in 2001 (Fig. 2). Income inequality exists in Bangla- 2015; Parvin and Wakilur 2009) to analyze national food desh (Hossain et al. 2005) and sources of income (agri- availability. Although Bangladesh now maintains relative cultural and non-agricultural) and average agricultural self-sufficiency in rice production, this has not been wage rates vary geographically in the country (GoB and accompanied by substantial rises in the availability of other WFP 2004; BBS et al. 2010). foods. Food availability in Bangladesh may face challenges Food access often varies between rural and urban areas in the future because of low production of other non-grain in Bangladesh (Pitt 1983; Hels et al. 2003). Figure 4 shows foods, import management of food (Bishwajit et al. 2013) trends in per person per day total food intake (grams) by and, regional variations of local food production due to residence type (rural and urban area) (HIES, various years, climatic and other environmental conditions. BBS 2010) and required food intake (MoHFW and BNNC 1997) in Bangladesh over the period of 1983–85 to 2010. 2.3 Food Access The national level per person per day food intake was sourced from various Household Income and Expenditure Food access refers to the economic, social, political and Survey (HIES) reports and BBS (2010) during 1983–2010. legal rights of individuals to adequate resources for getting The food items considered are rice, other cereals, vegeta- sufficient food for a nutritious diet (FAO 2006). Poverty bles and potatoes, pulses, oils and fats, sugar, fruits, fish, has decreased in Bangladesh since 1992 (FAO 2013). Per meat, eggs, and milk. Figure 4 shows the total daily food capita Gross Domestic Production (GDP) (adjusted to US$ intake in grams by summing these food items and illus- in 2014) has increased over the last 52 years from $100 in trates the changes of total food intake in rural, compared 1962 to $1089 in 2014 (BBS 2014). Employment genera- with urban areas, over the period 1983–2010. Overall, per tion has increased through public and private sector pro- person per day total food intake has increased in both areas grams and the number of the extreme poor reduced from (Fig. 4). However, food intake in rural and urban areas has 44 million in 2000 to 26 million in 2010 (FAO 2013). always been lower than the required food intake of 934

123 A Review of Food Security and the Potentials to Develop Spatially Informed Food Policies in… Page 5 of 18 19

Fig. 4 Trends in per person per Rural area day total food intake (grams) by 1000 Urban area

residence type (rural and urban )

s Required food intake area) in Bangladesh during m a

1983–85 to 2010. Data collected r

g 950 ( from (HIES, various year, BBS e

2010) and compared to required k a t

food intake given by MoHFW n i

900

and BNNC (1997) d o o f

l a

t 850 o t

y a d

r

e 800 p

n o s r

e 750 p

r e P 700

1983-85 1991-92 1995-96 2000 2005 2010 grams per person per day (MoHFW and BNNC 1997). In 2251 kcal (kilocalorie ) during 1962–64 to 2318 kcal in rural areas, it has increased from 741 g in 1983–85 to 2010 (HIES 2010). It remains below the FAO recom- 906.4 g in 2010. Food intake in urban areas showed a mended required average calorie consumption of 2410 kcal fluctuating trend during the period, but increased overall per person per day for adults from all sources of food (FAO from 763 g in 1983–85 to 871 g in 2010 (Fig. 4). These 2008). changes may have occurred due to changes in consumption At present, Bangladesh is experiencing a substantial patterns of different types of foods and rural–urban variation in food utilization from different food categories demographic dissimilarities in the country during the (FAO 2013). Figure 5 indicates the national level per period. person per day calorie consumption and desirable calorie Access to infrastructure resources, particularly total consumption in kcal from 1991–92 to 2010 for rice and delivery distances of food within a region, the location of other major food items. The national level per person per growth centers, and access to electricity (Scaramozzino day food intake was sourced from the reports of the 2006; Shin 2010) may affect food access in Bangladesh. Household Income and Expenditure Survey (HIES) over Except for the Chittagong Hill Tracts and Coastal Region the period of 1991–92 to 2010 for different food items in of Bangladesh, the national average travel time to the grams. The per person per day desirable food intake was nearest growth center is approximately 1 h (GoB and WFP given by NFPCSP (2013) for these food items in grams. 2004). It is likely that spatial variations of infrastructure Data from both sources were converted into kcal for rice development and socio-economic conditions may influence and other major food items including wheat, potato, fish, food security. However, little attention has been given to pulses, vegetables, meat, eggs, milk, and fruits by using these conditions previously in assessing food security in their nutrition values (Rahman and Islam 2014). However, Bangladesh. this estimation did not include food intake from other cereals, oils, and fat. Calorie consumption from major food 2.4 Food Utilization items other than rice is well below the desirable calorie consumption over the entire period (per capita 1026 kcal Food utilization encompasses the nutritional well-being of for these items) (NFPCSP 2013). Rice consumption mar- a person that comes from adequate food consumption from ginally decreased from 1991 (1518 kcal) to 2010 a diversity of food sources, food storage, and processing, (1335 kcal); however, calorie consumption from rice is still including clean water and healthy environments (FAO higher than the desirable calorie consumption standard (per 2006). Utilization of food has always been unsatisfactory capita 1124 kcal from rice) (NFPCSP 2013). This repre- nationally in Bangladesh. The national average per person sents a dietary imbalance in food utilization in Bangladesh. per day nutrition intake from a variety of food sources has Several factors influence the utilization of food. For only slightly increased over the last 50 years from instance, education levels affect the understanding of the

123 19 Page 6 of 18 I. Ara, B. Ostendorf

Fig. 5 Average per person per Rice day calorie consumption (kcal) from rice and other food major Other major food items ) items in Bangladesh during l a 1500 1991–92 to 2010, Calorie c k ( consumption (kcal) and n desirable food consumption o i t

(kcal) data collected from (BBS, p

HIES, various year) and m u (NFPCSP 2013) respectively s n 1000 o c

e i r o l a c

y a d

r 500 e p

n o s r e

Per p 0

1991-92 1995-96 2000 2005 2010 Desirable calorie Consumption (kcal) nutritional value of foods (Sajjad et al. 2014). Individual (BMD 2012) were interpolated and aggregated at greater health conditions affect the capacity of the body to absorb region level using a Geographic Information System (GIS) food and access to clean and safe drinking water and to show spatial variations of rainfall (Fig. 6a) and average sanitary facilities are critical in food utilization. In Ban- temperature (Fig. 6b) in Bangladesh. Most of the eastern gladesh, 70% of the population had access to safe drinking part of the country has high rainfall; averaging water in 1992. This reached 85% by 2012, and the 2688–3192 mm over the period (Fig. 6a). The southern advancement in sanitation has improved (FAO 2013); part of the country has slightly higher temperatures com- however, these averages may not be representative for all pared with the northern part of the country (Fig. 6b). regions in Bangladesh. A strong geographical association Although the spatial variations in long-term temperature was found in three anthropometric indicators of undernu- are small, the differences may influence food security trition (stunting, wasting and underweight) at the division across the country in the long term and are similar in level (FAO 2013); with Sylhet division in the Northwest magnitude to the climate changes predicted to occur during being the worst affected (BDHS 2013). Regional variation the next decades (Ara et al. 2016, 2017). of dietary imbalance and other factors related to food uti- Several studies have analyzed the impact of climate lization may influence food security in Bangladesh and it is change on food production, particularly rice production in important to evaluate these. Bangladesh as a means of food security evaluation. Table 1 indicates various assessments on the impact of climate 2.5 Food Security and Climate Change change on rice yield/production in Bangladesh for different in Bangladesh rice eco-types (Aus, Aman, and Boro). Assessments based on crop models such as Simulation Models, Crop Envi- The impact of climate change on food security is an ronment Resource Synthesis (CERES), Decision Support important concern globally (Rosenzweig and Parry 1994; System for Agro-technology Transfer (DSSAT) and Mul- Deryng et al. 2014; Burke and Lobell 2010; Iizumi et al. tifactor Impact Analysis predicted a decrease in rice yield 2013) including Bangladesh (Habiba et al. 2015b). This due to temperature increase (Karim et al. 1996; Basak et al. section of the paper primarily focuses on food security and 2010; Ruane et al. 2013). On the other hand, empirical climate change in order to consider the potential geo- assessments showed mixed responses from climate change graphic impact of climate change on food security com- for different rice eco-types, based on historic evidence ponents in a country, Bangladesh (Rimi et al. 2009; Sarker et al. 2012, 2014; Amin et al. Climatic variables, particularly rainfall and temperature, 2015). Some studies assessed the combined effect of both vary across Bangladesh over the period 1981–2010 climatic (rainfall and temperature) and management prac- (Fig. 6). The station-based rainfall and temperature data tices (irrigation and cropping intensity) variables on rice

123 A Review of Food Security and the Potentials to Develop Spatially Informed Food Policies in… Page 7 of 18 19

a b

Average rainfall (mm) Average temperature (°C) Very low (1619 - 1911) Very low (25.20 - 25.48) Low (1912 - 2232) Low (25.48 - 25.70) Medium (2233 - 2687) Medium (25.71 - 25.86)

High (2688 - 3192) High (25.86 - 25.96) 0 50 100km

Fig. 6 Spatial distributions of average rainfall (a) and average temperature (b) at greater regions level in Bangladesh during 1981–2010. Classes represent quartiles yield. (Ara et al. 2016) whereas others primarily focused on reduce the risk of local low production due to climate the benefit of transplanting date changes for Boro rice change for different crops (MoEF 2005, 2009; Islam et al. alone (Mahmood 1998; Mahmood 2003). However, there is 2013a). Unfortunately, the spatial detail in existing significant uncertainty concerning changes in rice produc- understanding of climate change impacts still limits region- tion/yield associated with climate change (Karim et al. specific actions (Islam et al. 2013b). In addition, food 1996; Ruane et al. 2013) as the effects vary in the different security is strongly influenced by natural disasters includ- studies (Rimi et al. 2009; Sarker et al. 2012, 2014; Amin ing flood, drought, cyclones, and riverbank erosion, which et al. 2015). Climate change may also indirectly affect food will likely increase with climate change (Karim et al. 1990; production in coastal areas because of salinity intrusion and Brammer 1999). Natural disasters can make areas vulner- increased yield variability causing reduced food availabil- able to reduced production, can cause damage to infras- ity (Parvin and Ahsan 2013; Sarwar and Islam 2013). tructure and communication networks and can lead to a Table 1 shows that some assessments include the loca- displacement of people (Islam and Sumon 2013; Coirolo tion to ascertain the regional variations of climate change et al. 2013). In Bangladesh, floods influence the food dis- impact on rice yield, particularly at the broad climatic zone tribution system through their impact on roads and trans- (Sarker et al. 2014) or agro-ecological zones (Ruane et al. port networks. However, the regional influence of natural 2013). Using a multifactorial mixed model analysis, tem- disasters on food security in Bangladesh is little studied. perature significantly decreased yield by 0.14 t/ha rice In addition to food availability, climate change may also yield per 1 °C; thus, even small spatial temperature dif- affect food access and food utilization in Bangladesh. As ferences may have noticeable effects on crop yield aver- agricultural production becomes susceptible to environ- ages (Ara et al. 2016). Such a spatial difference of mental and socio-economic conditions, farmers may temperature may affect other crop (wheat, potato, pulses, change preferences for crop production (Rosenzweig and and vegetables) production. Binswanger 1993; Simelton et al. 2012; Iizumi and Both the National Adaptation Program of Action Ramankutty 2015). This eventually influence the avail- (NAPA) and the Bangladesh Climate Change Strategy and ability of food for local consumption; all having a potential Action Plan (BCCSAP) recognize the importance of spa- influence on household income (Burke and Lobell 2010). tially variable climate influences on food security and Climate change is likely to become an important part of indicate the need of region-specific adaptation measures to food security in Bangladesh. In certain districts in

123 19 123 Table 1 Assessments on impact of climate change on rice production/rice yield in Bangladesh ae8o 18 of 8 Page Authors of the Year Rice Models Variables Impact on Rice Location Climate change General comment assessmentsa,b,c eco-types* (scenarios* /historic evidence) Aus Aman Boro

Karim et al.a 1996 Yield CERES-Rice Temperature (average) -- - 6 districts GCM Rice yield changes are varied in (1996) Temperature (average) and -- CCCM six districts due to different simulation Co2 GFDL Only Co2 ?? ? Mahmoodc 1998 Productivity YIELD Climate 2 districts Comparison of two models for (1998) CERES Pedological (Mymensingh modifying the transplanting date and Barisal) of Boro only Hydrological Agronomic Climate station ((lat, long) Mahmood 2003 Yield CERES-Rice Climate Climate Station Which region would get benefit c et al. (2003) Pedological locations (16) from a transplanting date change for Boro cultivation Hydrological Agronomic Climate station ((lat, long)) Rimi et al.b 2009 Production Statistical Temperature (seasonal 0 -?1district GCM-GFDL-TR Changes in rice yield varied due to (2009) analysis max) (Shatkhira) UKTR different climate scenarios Temperature (seasonal min) 0 -? HadCM2 Rainfall (seasonal total) ?- - Soil -?/-?/- Basak et al.c 2010 Yield DSSAT Temperature (seasonal - 12 districts PRECIS Reduction of Boro rice variety (2010) max) (BR3 and BR14) yield in all Temperature (seasonal min) locations Rainfall Sarker et al.b 2012 Yield Statistical Temperature (seasonal ? 0 - National Historic evidence Maximum and minimum (2012) analysis max) temperature are more Temperature (seasonal min) 0 -? pronounced compared with rainfall Total Rainfall (seasonal ?? 0 total) Ruane et al. a 2013 Production Multifactor Temperature ( seasonal -- - 16 agro-ecological GCM Substantial changes in rice yield

(2013) impact max) zones for climate scenario in different Ostendorf B. Ara, I. analysis Temperature (seasonal min) -- - zones were measured alone Rainfall (seasonal mean) -- Sea level rise was not simulated for all regions Co2 ?? ? Flood -- Sea level rise -- - eiwo odScrt n h oetast eeo ptal nomdFo oiisin Policies Food Informed Spatially Develop to Potentials the and Security Food of Review A Table 1 continued

Authors of the Year Rice Models Variables Impact on Rice Location Climate change General comment assessmentsa,b,c eco-types* (scenarios* /historic evidence) Aus Aman Boro

Sarker et al. b 2014 Yield Statistical Temperature (seasonal 0 ?-7 climatic zones Historic evidence Rice yield changes will be higher (2014) analysis max) for Aman compared with Boro Temperature (seasonal min) - 00 and Aus due to climate change Rainfall - 00 Region ?? ? Time (year) ?? ? Amin et al. b 2015 Yield Statistical Temperature (seasonal 0 --National Historic evidence Findings indicated impact of (2015) analysis max) climatic variables on rice yield Temperature (seasonal min) 0 0 ? but regional variation of rice and these climatic variables were Rainfall (seasonal mean) 0 - 0 overlooked Humidity (seasonal mean) ?? 0 Sunshine (seasonal mean) 0 0 ? Time (year) 0 ? 0 Ara et al.c 2016 Yield Statistical Temperature (annual 23 greater regions Historic evidence Analyzed the combined impact of (2016) analysis average) climate and management Rainfall (annual total) practices on rice yield. Increase in annual mean temperature and Cropping intensity total rainfall will decrease rice Area under irrigation yield aAssesments modeled rice production/yield changes due to changes of climatic variables, (?) increase, (-) decrease bAssesments indicated impacts of individual climate variables on rice production/rice yield, (?) positive effect, (-) negative effect, (0) no significant effect cAssesments do not consider all rice ecotypes Climate Scenarios considered: CCCM climate and carbon cycle modeling group, GCM general circulation model, GFDL geophysical fluid dynamic laboratory, GFDL-TR geophysical fluid … dynamics laboratory transient, HadCM2 Hadley center during 1995 and 1996 using the second version of the United Kingdom Meteorology Office’s Unified Model, PRECIS Providing regional climates for impacts studies ae9o 18 of 9 Page 123 19 19 123 Table 2 Components, issues, determinants of food security and the available geographic and temporal data sources on these in Bangladesh nationally ae1 f18 of 10 Page Components Issues Literature that quotes determinants Determinants (primary and Possible Data Geographic data Temporal data availability Proxy) Sources availability

Food Production Mainuddin and Kirby (2015), Moslehuddin Food production BBS, BARC Greater region 1947–2010 availability et al. (2015) Soil SRDI, BCA National 2006 Seed quality BBS Greater region 1991–1997 Management Abraham et al. (2014), Iizumi and Cropped area BBS Greater region 1981–2010 practices Ramankutty (2015), Hadgu et al. (2009), Means of irrigation BBS Greater region 1981–2010 Tiwari and Joshi (2012), Deryng et al. (2011), Alauddin and Sharma (2013), Use of fertilizer BBS Greater region 1990, 1991 Rahman (2010), Nasrin et al. (2015), Use of pesticides BBS Greater region 1976–1987 Ahmad et al. (2014), Yengoh (2012), Ara Groundwater availability WARPO, IWM Groundwater well 1985–2010 et al. (2016), Kirby et al. (2016) station Climatic Rosenzweig and Parry, 1994; Parry et al., Temperature BMD Weather station 1948–2012 variables 1999; Deryng et al., 2014; Aggarwal and Rainfall BMD Weather station 1948–2012 Singh, 2010; Basak and Alam, 2013 Humidity BMD Weather station 1972–2012 Sunshine BMD Weather station 1972–2012 Disasters Atkins (2009), Simelton et al. (2012), Ninno Flood DIMC, CEGIS, National, district, Historic record of flood and Lundberg (2005), Dewan et al. (2017), SPARRSO sub-district Holle and Islam (2017) Drought DIMC, CEGIS, National, district, Historic record of drought SPARRSO sub-district Cyclone DIMC, CEGIS, National, district, Historic record of cyclone SPARRSO sub-district Food access Demography Bose and Dey (2007), Scaramozzino (2006), Activity rate BBS Greater region 1981a Felker-Kantor and Wood (2012) District, sub-district 1991a, 2001a, 2011b Sex ratio BBS Greater region 1981a District, sub-district 1991a, 2001a, 2011b Dependency ratio BBS Greater region 1981a District, sub-district 1991a, 2001a, 2011b Age structure BBS Greater region 1981a District, sub-district 1991a, 2001a, 2011b Household structure BBS Greater region 1981a District, sub-district 1991a, 2001a, 2011b Economics Bose and Dey (2007), Scaramozzino (2006) Land ownership HIES District, sub-district 1991–1992, 1995–1996, 1999–2000, 2004–2005, 2010 Income HIES District, sub-district 1991–1992, 1995–1996, 1999–2000, 2004–2005, 2010

Assets HIES District, sub-district 1991–1992, 1995–11996, 1999–2000, Ostendorf B. Ara, I. 2004–2005, 2010 Expenditure on food HIES District, sub-district 1991–1992, 1995–1996, 1999–2000, 2004–2005, 2010 Employment Scaramozzino (2006) Labour force structure BBS Greater region 1981a District, sub-district 1991a, 2001a, 2011b BBS Greater region 1981a eiwo odScrt n h oetast eeo ptal nomdFo oiisin Policies Food Informed Spatially Develop to Potentials the and Security Food of Review A Table 2 continued

Components Issues Literature that quotes determinants Determinants (primary and Possible Data Geographic data Temporal data availability Proxy) Sources availability Agriculture vs. non-agriculture District, sub-district 1991a, 2001a, 2011b activities Rural–urban Hossain et al. (2005), Thorne-Lyman et al. Ratio of people living in urban BBS Greater region 1981a differences (2010), Pitt (1983) and rural areas District, sub-district 1991a, 2001a, 2011b Infrastructure Scaramozzino (2006), Shin (2010) Access to roads BBS Greater region 1981a District, sub-district 1991a, 2001a, 2011b LGED District, sub-district 2015 Number of schools BBS Greater region 1981 Districts 1991a, 2001a, 2011b Access to electricity BBS Greater region 1981a District 1991a, 2001a, 2011b Access to information technology BTRC District 2001–2014 Food Calorie intake Bose and Dey (2007), Frongillo et al. (2003) Food consumption HIES, FPMU District, sub-district 1991–1992, 1995–1996, 1999–2000, utilization Hels et al. (2003), Hillbruner and Egan 2004–2005, 2010 (2008), Roetter and Keulen (2007) Health Scaramozzino (2006) Fertility BBS Greater region 1981a District 1991a, 2001a, 2011b Mortality BBS Greater region 1981a District 1991a, 2001a, 2011b Anthropogenic indicators HIES District, sub-district 1991–1992, 1995–1996, 1999–2000, 2004–2005, 2010 Education Sajjad et al. (2014) Literacy rate BBS Greater region 1981a District, sub-district 1991a, 2001a, 2011b Water and Sajjad et al. (2014), Roetter and Keulen Access to drinking water BBS Greater region 1981a sanitation (2007) District 1991a, 2001a, 2011b Access to sanitary facilities BBS Greater region 1981a … District 1991a, 2001a, 2011b Proxy determinants are indicated in italics a Population and housing census year. Administrative boundary changes occurred between 1981 and 1991 census years b Note that not all community series for districts from the 2011 Population and Housing census were available at the time of article publication ae1 f18 of 11 Page 123 19 19 Page 12 of 18 I. Ara, B. Ostendorf

Bangladesh, farmers are currently practicing various Bangladesh Telecommunication Regulatory Commis- indigenous techniques to adjust with the local climatic sion (BTRC) anomalies to produce different crops (Alauddin and Sarkar Bangladesh Space Research and Remote Sensing 2014). However, national level strategies need to improve Organization (SARRSO) for different areas to combat climate change. Although the Centre for Environmental & Geographic Information government has launched several initiatives to facilitate Services (CEGIS) climate change adaptation, more interventions are still Disaster Incidence Database of Bangladesh (DMIC) required to develop region-specific resilience (MoEF 2009; Food Security and Nutrition Data Portal (FPMU) MoFDM 2008). It is important to understand how to design a Household Income and Expenditure Survey (HIES) long-term food security plan for Bangladesh that includes a Local Government Engineering Department (LGED) consideration of regional climate change and other envi- Soil Research Development Institute (SRDI) ronmental system impacts on all food security components. Water Resource Planning Organization (WARPO) Such a plan may reduce the degree and severity of low food Administrative boundaries (Fig. 1) are typically fol- security nationally in future. The following section of the lowed by the above-mentioned sources for published spa- paper explores the possible national data sources relevant to tio-temporal data including greater regions, districts, and food security, which may help to develop such a plan by sub-districts analyzing this data to improve food security in the country. Table 2 presents the issues and determinants (primary and proxy) of food security under the three food security components and indicates the available national spatio- 3 National Data Sources on Food Security temporal data on food security determinants for Bangla- in Bangladesh desh. The spatial resolution of food security datasets vary from administrative unit-level aggregates to point-based Data provide the fundamental evidence for assessment of measurements. To be useful in spatial analysis, station- all food security components, but often, data is difficult to based data (e.g., climate, groundwater) needs to be pre- obtain and manage. Many factors influence its quality and pared using appropriate spatial interpolation techniques. adequacy, including its use for objectives that were not Temporal data availability also varies for different years. intended during collection, correlations between determi- Time series data are available for most of the primary food nants, and its derivations for different purposes (Pinstrup- security determinants, but it is only possible to get proxy Andersen 2009; Scaramozzino 2006). Food security com- determinants from Population and Housing Census reports ponents often rely on multiple proxies in a country (Al- for 1981, 1991, 2001 and 2011. Some primary determi- wang et al. 2001). For instance, fertility and mortality nants (e.g., Crop production, irrigation, cropping intensity) (frequently used in demographic research) are relevant in have been used in national food security assessments pre- many food security analyses as proxies for population viously (Mainuddin and Kirby 2015; Parvin and Wakilur health. The present study explores the availability of 2009); however, these assessments overlooked many directly measured (primary) and proxy determinants for the important primary determinants (Table 2). In addition, different food security components. Most determinants whilst proxy determinants are considered as important in were sourced from the recent national and international food security assessment globally (refer to relevant cita- literature under different food security issues, including tions in Table 2), little attention so far has been given to production, management practices, climatic variables, dis- proxy determinants in Bangladesh. It is important to con- asters, demography, economics, employment, rural–urban sider all primary and proxy determinants to strengthen food differences, infrastructure, calorie consumption, health, security research in the country. education, water, and sanitation. Although Table 2 shows that spatio-temporal data on The present study identifies a wide range of possible food security determinants can be achieved in Bangladesh national data sources on important determinants (primary from past records, such data relevant to food security and proxy) at the spatio-temporal level of Bangladesh. The components are often ignored in food security assessments. following sources have been identified for data relevant to Most previous food security assessments were based on food security analysis: primary data (Frongillo et al. 2003; Lewis 1993; Coates Bangladesh Agricultural Research Council (BARC) et al. 2010; Thorne-Lyman et al. 2010) and HIES data Bangladesh Bureau of Statistics (BBS) (Faridi and Wadood 2010; Hossain et al. 2005; Bose and Bangladesh Country Almanac (BCA) Dey 2007), which are collected through questionnaire Bangladesh Meteorological Department (BMD) surveys at household/individual level. However, the studies Bangladesh Household Income and Expenditure Survey based on these data do not show the spatio-temporal (HIES) dynamics of food security in Bangladesh. The data 123 A Review of Food Security and the Potentials to Develop Spatially Informed Food Policies in… Page 13 of 18 19 identified in the present study (Table 2) allow for the Figure 7 identifies three separate parts in the process variations of different determinants to be considered over towards improved evidence-based, regional decision-mak- space and time. The following section addresses a way ing food policy. This includes spatio-temporal data on food forward for food security analysis that can use spatially security determinant, region-specific deficits, climate explicit data (Table 2) to understand the spatial and tem- change, and other future risks to understand present and poral pattern of each food security components. future food security conditions regionally. Further, this will address possible regional actions, which will be required for a respective risk in order to improve food security. The 4 Spatially Informed Food Policy Development: following section describes different processes according A Way Forward to Bangladesh context as an example, using spatio-tem- poral data mentioned in the previous section. The present study addresses a way forward for food secu- rity analysis that highlights a potential path towards a 4.1 Part 1: food security data analysis and mapping spatially informed food policy development. However, this will depend on data availability at aggregated spatial and The approach starts with the spatial data analysis and temporal resolution in different countries globally. mapping (Fig. 7). This has been done before in many

Spatio-temporal data supporting analysis of food security

Food security mapping

Component 1: Component 2: Component 3: Food availability Food access Food utilization Issues: Issues: Issues: Food production Demography Food consumption Management practices Employment Health Climatic variables Economics Education Disasters Rural-urban differences Water and sanitation Determinants: (Table 2) Infrastructure Determinants: (Table 2) Part 1 Spatial data analysis and mapping Determinants: (Table 2)

Regional deficits and risks Climate spatio- Global food change security uncertainty Part 2 Considering temporal evidence Political International uncertainty trade

Actions: Spatially informed food security policy based Part 3 Evidence- policy

Fig. 7 Spatio-temporal data on food security and processes towards evidence-based spatially informed food policy development

123 19 Page 14 of 18 I. Ara, B. Ostendorf countries (Tiwari and Joshi 2012; Yengoh 2012; Simelton influences on food security in Bangladesh. Many of these et al. 2012; Qin et al. 2013), but updated and new data factors are currently being researched (Mainuddin and provides the foundation for substantial improvement over Kirby 2015; Kirby et al. 2016) in the country and need to the time. For instance, a prominent past example is the be included in future policies. External factors that are National Food Security Atlas using data from the 2001 outside of direct government influence include global Population Census (GoB and WFP 2004). Several new economic factors such as global annual food production datasets have since become available that will allow an and international trade of agricultural or other commodities update, which in itself is an important contribution to food that can influence individual food security components in a security policy development as data is available for dif- country. There is also political uncertainty that may cause ferent epochs, providing the means for a spatial visual- changes in support and lack of policy continuity. As the ization of changes in critical food security components in complexity in this part is high, it is essential to provide Bangladesh. A detailed temporal data have been collected easy access to evidence in the form of spatio-temporal data for all three key components of food security (Table 2; (Part 2, Fig. 7). Fig. 7, Part 1). Although some harmonization of data will Climate change uncertainty influences all aspects of be necessary because of differences in spatial resolution or food security. Both flooding and cyclones are likely to temporal availability of the data, the results from the data increase with future climate change and have spatially review above indicate that this is possible. differing risk profiles associated with all components of The mapping of food availability includes food pro- food security, production, distribution, housing, and health. duction, management practices, climatic variables, and Some of these interactions are well understood (Shin 2010; disasters. Food access analysis is based on the human Sajjad et al. 2014; Hadgu et al. 2009; Shi and Tao 2014); dimension of food security, including socio-economic others lack evidence, which may successively reduce conditions, access to resources and communication in dif- uncertainties in modeling when it becomes available ferent regions. Food utilization analysis mainly considers through research. This research is crucial to build and the food consumption, health, education, water and sani- apply quantitative models that use risk appraisals to test tation of a region. Part 1 results in a visualization of the policy options for diverse future scenarios in a country. geographic variation of components of all of the three food security determinants across the country over time. 4.3 Part 3: development of a spatially informed food policy 4.2 Part 2: identifying spatio-temporal evidence for deficits and risks associated with all food The third part (Part 3, Fig. 7) suggests that regional food security components security also requires regionally adapted actions to address specific food security components. Actions may include The resulting outcomes from different assessments of food but are not limited to region-specific adaptation measures security components from part 1 provide the spatio-tem- for food availability, varied livelihood options for food poral evidence to identify deficits and risks. For example, access, and diversified diet preferences and awareness an assessment of food availability will help to identify the building related to health and hygiene environments for biophysical limitations of food production of a country proper food utilization. The present study advocates a regionally. Food access evaluations will help to ascertain spatially informed food security policy by considering the socio-economic and infrastructure inadequacies of food regional level deficits, future risks, and devises actions access. Food utilization assessments may indicate a lack of related to the respective components to improve food nutrition and dietary imbalances. All these outcomes will security for an individual country. The overview of Ban- help to understand existing regional risks for food security gladesh food security indicates that the country has in a country. achieved significant increases in rice production. Calorie This part includes explicitly spatio-temporal interactions consumption has increased marginally at the national level. in a given environmental system of a country, which have Both production and consumption have increased, along not received sufficient consideration in past policy devel- with imports of food from foreign countries in emergen- opments (Ray 2007; Rocha 2009; Bashir and Schilizzi cies, and this has played a vital role in the past for 2015; Habiba et al. 2015c). Examples of possible interac- improving food security. Regardless of these positive tions are off-site effects, i.e., water usage upstream that conditions at national level, the present food policy needs influences the water availability downstream or potentially to be improved by incorporating spatial representations of unsustainable use of groundwater (Ahmad et al. 2014; Ara environmental system and socio-economic conditions to et al. 2016) and economic development that indirectly tackle regional food security challenges (MoFDM influences movement of people and goods with indirect 2006, 2008). A comprehensive policy response to food 123 A Review of Food Security and the Potentials to Develop Spatially Informed Food Policies in… Page 15 of 18 19 security that fully addresses spatio-temporal variations of opportunities for crop intensification for food security in food security components is likely to contribute to Bangladesh. Water Resour Manage 28(5):1415–1429 Ahmed FRS (2015) Policy support and institutional dimensions of improving food security there. food security. In: Habiba U, Hassan RAW, Abedin AM, Shaw R (eds) Food security and risk reduction in Bangladesh. Springer, Tokyo 5 Conclusions Alauddin M, Sarker MAR (2014) Climate change and farm-level adaptation decisions and strategies in drought-prone and ground- water-depleted areas of Bangladesh: an empirical investigation. The present study has shown that food security is not only Ecol Econ 106:204–213 an issue of producing food. Food access and food utiliza- Alauddin M, Sharma BR (2013) Inter-district rice water productivity tion are also very important components of food security in differences in Bangladesh: an empirical exploration and impli- cations. Ecol Econ 93:210–218 Bangladesh. Most food security determinants vary spatially Alwang J, Siegel PB, Jorgensen SL (2001) Vulnerability: a view from in Bangladesh and this variation should be considered for different disciplines. worldbank.org/curated/en/2001/06/1637776/ regional food security assessment. Similarly, the possible vulnerability-view-different-disciplines. Accessed 15 Mar 2016 geographic impact of climate change needs to be consid- Amin MR, Zhang J, Yang M (2015) Effects of climate change on the yield and cropping area of major food crops: a case of ered. The study identified available spatio-temporal data on Bangladesh. Sustain Sci 7:898–915 food security in Bangladesh, which have not been used Ara I, Lewis M, Ostendorf B (2016) Spatio-temporal analysis of the previously for regional food security assessment. impact of climate, cropping intensity and means of irrigation: an Bangladesh has not yet developed spatially informed assessment on rice yield determinants in Bangladesh. Agric Food Secur 5(1):1–11 policies to improve its food security, even though there are Ara I, Lewis M, Ostendorf B (2017) Understanding the spatially documented spatial differences in food security (GoB and variable effects of climate change on rice yield for three ecotypes WFP 2004). Region-specific analysis is required for future in Bangladesh, 1981–2010. Adv Agric 2017:11. https://doi.org/ evidence-based policy development globally including 10.1155/2017/6287156 Atkins PJ (2009) Famine. In: Kitchin R, Thrift N (eds) International Bangladesh. The present study has developed a way for- encyclopedia of human geography. Elsevier, Oxford ward for regional food security assessment that includes Basak JK, Alam K (2013) Impacts of carbon dioxide emission and analysis of existing spatio-temporal data but also explicitly subsequent rise of temperature on rice production in bangladesh: stresses interaction of this data with other forms of data. implications for food security. Int Res J Environ Sci 2:60–67 Basak JK, Ali MA, Islam MN, Rashid MA (2010) Assessment of the Whereas the first part can be used directly to understand effect of climate change on boro rice production in Bangladesh regional variations of factors determining food security, the using DSSAT model. J Civ Eng 38(2):95–108 second part highlights potential connections with factors Bashir M, Schilizzi S (2015) Food security policy assessment in the with high global or national uncertainty. These components Punjab, Pakistan: effectiveness, distortions and their perceptions. Food Secur 7(5):1071–1089 need some consideration to obtain spatially informed food BBS (1981) Bangladesh population census, administrative report. policy in a country. This will contribute to the design of GoB, Dhaka robust policy and action plans, capable of addressing food BBS (2001) Bangladesh population census. GoB, Dhaka security regionally in more effective ways than are possible BBS (2010) Report of the household income & expenditure survey 2010. Ministry of Planning, GoB, Dhaka at present in Bangladesh and elsewhere. The government BBS (2011a) Bangladesh population census report. GoB, Dhaka can then give priority to areas where existing food security BBS (2011b) Agricultural statistical year book of Bangladesh. GoB, deficits and future risks are high and devise the necessary Dhaka region-specific actions to improve food security. BBS (2012a) Agricultural statistical year book of Bangladesh. GoB, Dhaka BBS (2012b) Statistical year book of Bangladesh. GoB, Dhaka BBS (2014) Statistical Year Book of Bangladesh. GoB, Dhaka BBS, Wfp & WB (2010) Poverty maps of Bangaldesh: key findings. References GoB, World Food Program and The World Bank, Dhaka BDHS (2013) Demographic and Health Survey. Dhaka, Bangladesh: Abraham B, Araya H, Berhe T, Edwards S, Gujja B, Khadka RB et al National Institute of Population Research and Training (2014) The system of crop intensification: reports from the field (NIPORT), Mitra and Associates, and ICF International on improving agricultural production, food security, and Belesky P (2014) Regional governance, food security and rice resilience to climate change for multiple crops. Agric Food reserves in East Asia. Glob Food Secur 3(3–4):167–173 Secur 3(1):1–12 Bishwajit G, Sarker S, Kpoghomou M-A, Gao H, Jun L, Yin D et al Aggarwal PK, Singh AK (2010) Implications of global climatic (2013) Self-sufficiency in rice and food security: a South Asian change on water and food Security. In: Ringler C, Biswas AK, perspective. Agric Food Secur 2(1):1–6 Cline S (eds) Global change: impacts on water and food security. BMD (2012) Climate Database. Bangladesh meteorological depart- Springer, Berlin Heidelberg ment, Dhaka Ahmad MU, Kirby M, Islam M, Hossain MJ, Islam MM (2014) Bose ML, Dey MM (2007) Food and nutritional security in Groundwater use for irrigation and its productivity: status and Bangladesh: going beyond carbohydrate counts. Agric Econ Res Rev 20(2):203–225

123 19 Page 16 of 18 I. Ara, B. Ostendorf

Brammer H (1999) Agricultural disaster management in Bangladesh. GoB (2016) Bangladesh National Portal. Dhaka, Bangladesh. whttp:// University Press Limited, Dhaka www.bangladesh.gov.bd/. Accessed 26 Aug 2016 Brammer H (2000) Agro-ecological aspect of agricultural research in GoB and WFP (2004) The Food Security Atlas of Bangladesh. Bangladesh. University Press Limited, Dhaka Bangladesh Planning Commission and the United Nations World Burke M, Lobell D (2010) Climate effects on food security: an Food Programme, Dhaka overview. In: Lobell D, Burke M (eds) Climate change and food Habiba U, Abedin MA, Hassan AWR, Shaw R (eds) (2015a) Food security. Advances in Global Change Research, vol 37. Springer, security and risk reduction in Bangladesh. Springer, Japan Dordrecht Habiba U, Abedin MA, Shaw R (2015b) Introduction and overview of Coates J, Webb P, Houser R, Rogers B, Wilde P (2010) ‘‘He said, she food security and risk reduction issues. In: Habiba U, Hassan said’’: who should speak for households about experiences of RAW, Abedin AM, Shaw R (eds) Food security and risk food insecurity in Bangladesh? Food Secur 2(1):81–95 reduction in Bangladesh. Springer, Tokyo Coirolo C, Commins S, Iftekharul Haque I, Pierce G (2013) Climate Habiba U, Abedin MA, Shaw R (2015c) Future approach of food change and social protection in Bangladesh: are existing security, suatainable development, environment and resource programmes able to address the impacts of climate change? managemnet and risk redustion. In: Habiba U, Hassan RAW, Dev Policy Rev 31(s2):074–090 Abedin AM, Shaw R (eds) Food security and risk reduction in Deb UK, Hossain M, Jones S (2009) Rethinking food security Bangladesh. Springer, Tokyo strategy: self-sufficiency of self-reliance. http://econpapers. Hadgu K, Rossing WH, Kooistra L, van Bruggen AC (2009) Spatial repec.org/paper/esswpaper/id_3a3458.htm. Accessed 10 Mar variation in biodiversity, soil degradation and productivity in 2016 agricultural landscapes in the highlands of Tigray, northern Deryng D, Sacks W, Barford C, Ramankutty N (2011) Simulating the Ethiopia. Food Secur 1(1):83–97 effects of climate and agricultural management practices on Harvey M, Pilgrim S (2011) ‘The new competition for land: food, global crop yield. Glob Biogeochem cycles 25(2). https://doi. energy, and climate change. Food Policy 36(Supplement 1):S40– org/10.1029/2009GB003765 S51 Deryng D, Conway D, Ramankutty N, Price J, Warren R (2014) Hels O, Hassan N, Tetens I, Thilsted SH (2003) Food consumption, Global crop yield response to extreme heat stress under multiple energy and nutrient intake and nutritional status in rural climate change futures. Environ Res Lett 9(3):034011 Bangladesh: changes from 1981–1982 to 1995–96. Eur J Clin Dewan A, Hossain MF, Rahman MM, Yamane Y, Holle RL (2017) Nutr 57(4):586–594 Recent Lightning-related fatalities and injuries in Bangladesh. HIES (2010) Household income and expenditure survey. GoB, Dhaka Weather Clim Soc Hillbruner C, Egan R (2008) Seasonality, household food security, Elahi KM, Ara I (2008) Understanding the Monga in Northern and nutritional status in Dinajpur, Bangladesh. Food Nutr Bull Bangladesh. Academic Press and Publishers Library, Dhaka 29(3):221–231 Faisal IM, Parveen S (2004) Food Security in the face of climate Holle RL, Islam A (2017) Lightning fatalities in Bangladesh in May change, population growth, and resource constraints: implica- 2016. In: Proceedings of the 8th Conference on the Meteoro- tions for Bangladesh. Environ Manage 34(4):487–498 logical Applications of Lightning Data. 2017 American Mete- FAO (2006) The state of food insecurity in the world 2006: orological Society Annual Meeting Seattle, Washington, eradicating world hunger—taking stock ten years after the world pp 22–26 food summit. Rome, Italy: Food and Agriculture Organization. Hossain M, Naher F, Shahabuddin Q (2005) Food security and ftp://ftp.fao.org/docrep/fao/009/a0750e/a0750e00.pdf. Accessed nutrition in Bangladesh: progress and determinants. Electron J 23 Dec 2015 Agric Dev Econ 2(2):103–132 FAO (2008) FAO Methodology for the measurement of food Iizumi T, Ramankutty N (2015) How do weather and climate deprivation: updating the minimum dietary energy requirement. influence cropping area and intensity? Glob Food Secur 4:46–50 Rome, Italy: Food and Agriculture Organization. http://www. Iizumi T, Sakuma H, Yokozawa M, Luo J-J, Challinor AJ, Brown ME fao.org/fileadmin/templates/ess/documents/food_security_statis et al (2013) Prediction of seasonal climate-induced variations in tics/metadata/undernourishment_methodology.pdf. Accessed 28 global food production. Nature Clim Change 3(10):904–908 Aug 2016 Islam A, Sumon A (2013) Integration of climate change adaptation, FAO (2013) A compendium of country reports on metadata for disaster risk reduction and social protection in bangladesh: National Agricultural Statistics: Bangladesh. http://faorap-apcas. contemporary views and experiences. In: Shaw R, Mallick F, org/rdes/PPT/Compendium2.pdf. Accessed 29 Aug 2016 Islam A (eds) Climate change adaptation actions in Bangladesh. FAO (2015) The state of food insecurity in the world, 2015. Rome, Springer, Tokyo Italy: Food and Agriculture Organization http://www.fao.org/3/ Islam A, Shaw R, Mallick F (2013a) National adaptation program of a-i4646e.pdf. Accessed 02 Sep 2016 action. In: Shaw R, Mallick F, Islam A (eds) Climate change Faridi R, Wadood SN (2010) An econometric assessment of adaptation actions in Bangladesh. Springer, Tokyo household food security in Bangladesh. Bangladesh Dev Stud Islam A, Shaw R, Mallick F (2013b) Bangladesh climate change 33(3):97–111 strategy and action plan. In: Shaw R, Mallick F, Islam A (eds) Felker-Kantor E, Wood C (2012) Female-headed households and food Climate change adaptation actions in Bangladesh. Springer, insecurity in Brazil. Food Secur 4(4):607–617 Tokyo FPMU (1991) Study on the estimation of seed, feed and post-harvest Kam SP, Hossain M, Bose ML, Villano LS (2005) Spatial patterns of wastage of food grain crops in Bangladesh Dhaka, Bangladesh: rural poverty and their relationship with welfare-influencing Food Planning and Monitoring Unit, Ministry of Food, GoB. factors in Bangladesh. Food Policy 30(5–6):551–567 http://pdf.usaid.gov/pdf_docs/PNABJ023.pdf. Accessed 04 Sep Karim Z, Hussain S, Ahmed M (1996) Assessing impacts of climatic 2016 variations on foodgrain production in Bangladesh. Water Air Frongillo EA, Chowdhury N, Ekstro¨m EC, Naved RT (2003) Soil Pollut 92(1):53–62 Understanding the experience of household food insecurity in Kirby J, Mainuddin M, Mpelasoka F, Ahmad M, Palash W, Quadir M rural Bangladesh leads to a measure different from that used in et al (2016) The impact of climate change on regional water other countries. J Nutr 133(12):4158 balances in Bangladesh. Clim Change 135(3–4):481–491

123 A Review of Food Security and the Potentials to Develop Spatially Informed Food Policies in… Page 17 of 18 19

Lewis DJ (1993) Going it alone: female-headed households, rights, Rahman KMM, Islam MA (2014) Nutrition-sensitive agriculture in and resources in rural Bangladesh. Eur J Dev Res 5(2):23–42 Bangladesh: a review. Food Secur 6(5):671–683 Mahmood R (1998) Air temperature variations and rice productivity in Rahman R, Mondal SM (2015) Role of water resource management in Bangladesh: a comparative study of the performance of the YIELD ensuring food security. In: Habiba U, Hassan RAW, Abedin AM, and the CERES-Rice models. Ecol Model 106(2–3):201–212 Shaw R (eds) Food security and risk reduction in Bangladesh. Mahmood R, Meo M, Legates DR, Morrissey ML (2003) The Springer, Tokyo CERES-Rice model-based estimates of potential monsoon Ray R (2007) Changes in food consumption and the implications for season rainfed rice productivity in Bangladesh. Prof Geogr food security and undernourishment: india in the 1990s. Dev 55(2):259–273 Change 38(2):321–343 Mainuddin M, Kirby M (2015) National food security in Bangladesh Rimi RH, Rahman SH, Karmakar S, Hussain SG (2009) Trend to 2050. Food Secur 7(3):633–646 analysis of climate change and investigation on its probable Mirza MQ (2002) Global warming and changes in the probability of impacts on rice production at Sathkhira, Bangladesh. Pak J occurrence of floods in Bangladesh and implications. Glob Meteorol 6:37–50 Environ Change 12(2):127–138 Rocha C (2009) Development in national policies for food and MoEF (2005) National Adaptation Programme of Action (NAPA). nutrition security in Brazil. Dev Policy Rev 27(1):51–66 Ministry of Environment and Forest, GoB, Dhaka Roetter RP, Keulen H (2007) Food Security. In: Roetter RP, Keulen MoEF (2009) Bangladesh Climate Change Strategy and Action Plan H, Kuiper M, Verhagen J, Laar HH (eds) Science for agriculture (BCCSAP). Ministry of Environment and Forest, GoB, Dhaka and rural development in low-income countries. Springer, MoF (1991) The estimation of seed, feed and post-harvest wastage of Netherlands foodgrain crops in Bangladesh. Ministry of Food, GoB, Dhaka Rosenzweig MR, Binswanger HP (1993) Wealth, weather risk and the MoFDM (2006) National Food Policy (NFP). Ministry of Food and composition and profitability of agricultural investments. Econ J Disaster Management, GoB, Dhaka 103(416):56–78 MoFDM (2008) National Food Policy Plan of Action (NFPPA). Rosenzweig C, Parry M (1994) Potential impact of climate change on Dhaka, Bangladesh: Food Planning and Monitoring Unit world food supply. Nature 367(6459):133–138 (FPMU), Ministry of Food and Disaster Management, GoB Ruane AC, Major DC, Yu WH, Alam M, Hussain SG, Khan AS et al MoHFW and BNNC (1997) Bangladesh national plan of action for (2013) Multi-factor impact analysis of agricultural production in nutrition Dhaka, Bangladesh: Ministry of Health and Family Bangladesh with climate change. Glob Environ Change Welfare and Bangladesh National Nutrition Council, GoB 23(1):338–350 Moslehuddin AZM, Abedin MA, Hossain MAR, Habiba U (2015) Sajjad H, Nasreen I, Ansari SA (2014) Assessment of spatio-temporal Soil health and food security: perspective from Southwestern dynamics of food security based on food security index analysis: Coastal region of Bangladesh. In: Habiba U, Hassan RAW, a case from Vaishali District, India. J Agric Sustain Abedin AM, Shaw R (eds) Food security and risk reduction in 5(2):125–152 Bangladesh. Springer, Tokyo Sarker MAR, Alam K, Gow J (2012) Exploring the relationship Nasrin S, Bergman Lodin J, Jirstro¨m M, Holmquist B, Andersson between climate change and rice yield in Bangladesh: an Djurfeldt A, Djurfeldt G (2015) Drivers of rice production: analysis of time series data. Agric Syst 112:11–16 evidence from five Sub-Saharan African countries. Agric Food Sarker MAR, Alam K, Gow J (2014) Assessing the effects of climate Secur 4(1):1–19 change on rice yields: an econometric investigation using NFPCSP (2013) Desirable Dietary Pattern for Bangladesh. Dhaka, Bangladeshi panel data. Econ Anal Policy 44(4):405–416 Bangladesh: National Food Policy Capacity Strengthening Sarwar MGM, Islam A (2013) Multi hazard vulnerabilities of the Programme, Bangladesh Institute of Research and Rehabilitation coastal land of Bangladesh. In: Shaw R, Mallick F, Islam A (eds) in Diabetes, Endocrine and Metabolic Disorders (BIRDEM) and Climate change adaptation actions in Bangladesh. Springer, GoB Tokyo Ninno DC, Lundberg M (2005) Treading water: the long-term impact Scaramozzino P (2006) Measuring Vulnerability to Food Insecurity. of the 1998 flood on nutrition in Bangladesh. Econ Hum Biol ESA Working Paper No. 06-12. Rome. http://www.fao.org/3/a- 3(1):67–96 ah630e.pdf. Accessed 3 Mar 2015 Parry M, Rosenzweig C, Iglesias A, Fischer G, Livermore M (1999) Sen A (1981) Poverty and famine an essay on entitlement and Climate change and world food security: a new assessment. Glob deprivation. Oxford University Press Limited, Delhi Environ Change 9(Supplement 1):S51–S67 Shi W, Tao F (2014) Vulnerability of African maize yield to climate Parvin GA, Ahsan SMR (2013) Impacts of climate change on food change and variability during 1961–2010. Food Secur security of rural poor women in Bangladesh. Manag Environ 6(4):471–481 Qual Int J 24(6):802–814 Shin M (2010) A geospatial analysis of market integration: the case of Parvin L, Wakilur RM (2009) Impact of irrigation on food security in the 2004/5 food crisis in Niger. Food Secur 2(3):261–269 Bangladesh for the past three decades. J Water Res Prot Simelton E, Fraser EG, Termansen M, Benton T, Gosling S, South A et al 3:216–225 (2012) The socioeconomics of food crop production and climate Pinstrup-Andersen P (2009) Food security: definition and measure- change vulnerability: a global scale quantitative analysis of how ment. Food Secur 1(1):5–7 grain crops are sensitive to drought. Food Secur 4(2):163–179 Pitt MM (1983) Food preferences and nutrition in rural Bangladesh. Talukder D, Chile L (2011) Estimation of population and food grain Rev Econ Stat 65(1):105–114 production in Bangladesh by 2020: a simple moving average Qin Y, Liu J, Shi W, Tao F, Yan H (2013) Spatial-temporal changes approach to a time series analysis. Bangladesh e-J Sociol of cropland and climate potential productivity in northern China 8(2):4–16 during 1990–2010. Food Secur 5(4):499–512 Thorne-Lyman AL, Valpiani N, Sun K, Semba RD, Klotz CL, Rahman S (2010) Six decades of agricultural land use change in Kraemer K et al (2010) Household dietary diversity and food Bangladesh: effects on crop diversity, productivity, food avail- expenditures are closely linked in rural Bangladesh, increasing ability and the environment, 1948–2006. Singap J Trop Geogr the risk of malnutrition due to the financial crisis. J Nutr 31(2):254–269 140(1):182S–188S

123 19 Page 18 of 18 I. Ara, B. Ostendorf

Tiwari P, Joshi B (2012) Natural and socio-economic factors affecting Bertram Ostendorf is a Pro- food security in the Himalayas. Food Secur 4(2):195–207 fessor at School of Biological Yengoh GT (2012) Determinants of yield differences in small-scale Sciences and director of the food crop farming systems in Cameroon. Agric Food Secur Spatial Information Group 1(1):1–17 (SIG) of Adelaide University, Australia. His main research interest is in evidence-based Iffat Ara is a Research Fellow decision-making for natural (Spatial Science) at the School resources management at the of Land and Food, University of border of environmental and Tasmania, the Australia and agricultural sciences. He spe- Tasmanian Institute of Agricul- cialises in spatial environmental ture (TIA). She was awarded a sciences, GIS, and spatial deci- Ph.D. in Science at the School sion support. His research of Biological Sciences, Univer- relates understanding causes of sity of Adelaide, Australia. Her spatial heterogeneity of the environment and how this can be used to research focused on regional increase both economic production and the integrity of the environ- variability of food security ment (i.e., landscape capability, biodiversity, and economic return at determinants and how to deal different spatial and temporal scales). Before joining Adelaide with this variability when University he worked as Scientific Assistant in University of Bayr- developing policies to improve euth, Germany; as a Lecturer in University of Innsbruck, Austria; as a food security. She obtained her Senior Researcher/Project Leader at the European Academy, Bol- Bachelor (Honours) and Master’s degree in Geography and Envi- zano/Bozen, Italy, and as Research Officer at CSIRO Wildlife and ronment at the Jahangirnagar University, the Bangladesh. She has Ecology/Sustainable Ecosystems, Atherton, Australia. He received his more than eight years of research and teaching experiences at dif- Doctor of Philosophy from University of California, Davis and San ferent universities in Bangladesh including the Stamford University of Diego State University, U.S.A. His teaching goal is to increase Bangladesh, the Jagannath University, and the Jahangirnagar awareness about the needs for spatial evidence in environmental University. She received research grants, awards, and fellowships decision making. Currently, he coordinates three subjects related to from Government of Bangladesh, Commonwealth Government of spatial information generation and management: ’Spatial Information Australia and CSIRO. Ara has a number of publications in reputed and Land Evaluation’, ’GIS for Environmental Management’, and international peer reviewed journals. Her academic interests encom- ’GIS for Agricultural Sciences’ at University of Adelaide, Australia. pass geographic information system (GIS) and remote sensing applications in managing environment, agriculture and natural resources.

123 Earth Systems and Environment (2018) 2:119–132 https://doi.org/10.1007/s41748-018-0037-y

ORIGINAL ARTICLE

A Study on Land Suitability for Rice Cultivation in Khordha District of Odisha (India) Using Remote Sensing and GIS

Sudhansu S. Rath1 · Jagabandhu Panda1 · R. Annadurai2 · Sachikanta Nanda2

Received: 23 November 2017 / Accepted: 7 February 2018 / Published online: 13 February 2018 © Springer International Publishing AG, part of Springer Nature 2018

Abstract With the global population on the rise, it is important to address the increasing demand for food. According to FAO (Food and Agriculture Organization, United Nations), by 2050, the developing countries must double their food production to meet the growing demand. Proper land utilization can be one of the solutions for this problem. In view of this, the current study focussed on land suitability analysis for Khordha district of Odisha (India) for rice crop. This study estimated that the amount of land suitable for rice cropping was 195,731 ha against the currently cultivated land of 122,183.38 ha. Therefore, there was a possibility of more amount of land that could be available for rice cultivation in Khordha district than the currently cultivated area. In order to perform this exercise, the land use and land cover data from IRS (Indian remote sensing satellite), soil nutrient parameters like pH values and nitrogen, potassium, phosphorous and organic carbon contents were considered. In addition, the climatic parameters such as near surface temperature, rainfall and number of rainy days were taken into account. The unused land identifed in Khordha district in this study might be utilized for cultivating rice crop in this region.

Keywords Land suitability · Rice · Remote sensing · GIS · Soil · Rainfall · LULC

1 Introduction are signifcantly afected by LULC changes and result in land degradation, which consequently led to a decline in soil Agriculture remains the largest employment sector in devel- productivity (Biro et al. 2013). In order to perform the land oping country like India. Since food-security is essential suitability analysis for a crop cultivation, changes in LULC for human establishment and quite important for a popu- need to be analysed besides soil nutrient parameters and few lous country like India, international agriculture agreements relevant meteorological variables using remote sensing and are needed for this purpose. The concept of food security GIS (Geographic Information System) techniques (Bhagat includes providing both physical and economic access to et al. 2009; Gumma et al. 2009; Kihoro et al. 2013; Kuria food that meets dietary needs and their food preferences. et al. 2011; Samanta et al. 2011). Thus, it is important to have an integrated approach for Agriculture contributes about 40% towards Gross improving the ability of countries to plan and monitor the National Product (GNP) and provides livelihood to about better use and management of their land resources in order 70% of the population in India. Rain-fed agro-ecosystem to increase agricultural productivity without compromising occupies a prominent position in the Indian agriculture (Lal the land and environmental safety. Therefore, studies related 2008). Nearly two-thirds (or ~ 67%) of the nation’s cropped to land-use and land cover (LULC) changes for analysing area is under rain fed agriculture. The potential of rain-fed the land suitability bear importance. The ‘soil properties’ crop production is dependent on many factors including suit- able climatic conditions, appropriate soil nutrients, slope of the land, etc. In general, crop cultivation and agricultural * Jagabandhu Panda [email protected]; [email protected] yield are functions of several factors including weather, soil characteristics and package of practices (POP). In India, 1 Department of Earth and Atmospheric Sciences, National weather plays a major role in crop yield since extreme con- Institute of Technology Rourkela, Rourkela, Odisha 769008, ditions viz. droughts, foods, etc. can have signifcant impact. India Several researchers all over the world adopted remote 2 Department of Civil Engineering, SRM University, Chennai, sensing and GIS-based techniques not only for rice India

Vol.:(0123456789)1 3 120 S. S. Rath et al. cultivation (Dengiz 2013; Gumma et al. 2009; Kihoro et al. GIS techniques. The study included: (1) developing long- 2013; Kuria et al. 2011; Samanta et al. 2011) but also for term database of climate to assess the variation in length of other crops (Bhagat et al. 2009). Satellites in recent years rainy season by considering rainfall and temperature using have emerged as a vital tool for generating the biophysi- GIS, (2) generating spatial soil data base in terms of the cal information (e.g. topography and slope), which helps to micronutrient content in the soil and (3) assessing the pre- evolve the optimal land use plan for sustainable development vailing LULC of the district (mainly agriculture land) and of an area. The digitally processed remotely sensed LULC evaluating the possible amount of land that can be used par- from moderate resolution imaging spectrometer (MODIS) ticularly for rice crop. was used in global, regional and local level studies by sev- eral researchers (Friedl et al. 2002; Hansen et al. 2000; Wes- sel et al. 2004). Satellite data from Indian remote sensing 2 Study Area and Geographical Location satellites (IRS series) including RESOURCESAT-1 (IRS P6) and 2 in recent times are quite useful in agriculture- Agriculture is the dominant sector to Odisha’s economy with related studies that involve LULC change analysis in Indian a contribution of ~ 30% to the Net State Domestic Product context (Bhagat et al. 2009; Choudhury et al. 2006; Devi (NSDP). About 73% of total workers are engaged in agri- and Kumar 2008; Kumar et al. 2013; Sathish and Niranjana culture including 44.3% cultivators and 28.7% agricultural 2010). These satellites have sensors like LISS III, LISS IV labourers. It may be noted that people living in rural areas and AWiFS, which have better spectral and radiometric reso- (~ 87% of total population) mostly depend upon the agri- lutions and are capable of covering larger area so that the culture sector and serve as the driver of the major portion of acquired information can be helpful for all agricultural stud- Odisha’s economy. Though the contribution of agriculture to ies. Again, GIS plays very important role in handling vast NSDP has signifcantly declined from 67% in 1951 to ~ 30% data including those of soil, climatic database and LULC in 2008, the percentage of workforce engaged in agriculture obtained from satellite for such types of studies. Analysing has remained somewhat unchanged. However, the scenario all the data in statistical, spatial and temporal forms and is changing in recent times (post 2010) and gradually less providing outputs in the form of maps and attributes are the people are being engaged in agriculture sector. On the other operations those can be done by GIS. Thus, use of remote hand, still there is an overcrowding in agriculture without sensing and GIS in land suitability analysis for rice crop cul- any perceptible increase in production. tivation appears to be quite convincing. Several researchers The reason of insignifcant increase in agricultural yield used this approach over diferent parts of the globe including could be due to the erratic south-west monsoon that can Tana delta (Kuria et al. 2011) and great Mwea (Kihoro et al. never be relied upon for the desired amount of precipita- 2013) of Kenya, central Anatolian region of Turkey (Dengiz tion and its regional distribution. For any sound agricul- 2013), inland valley wetlands of Ghana (Gumma et al. 2009) tural planning at regional scale, the seasonal defciencies and Morobe province in the Papua New Guinea (Samanta must be known to the planners. It may be noted that the et al. 2011). Over India, very few studies have been carried considered area of interest is located in the eastern part of out in this direction. One of such studies done by Bhagat India and depends upon monsoonal rainfall for agriculture. It et al. (2009) focused on Himachal Pradesh region and car- may be noted that the crop productions in Odisha are mostly ried out the land suitability analysis for cereal production afected by seasonal droughts and foods every year. Taking using GIS techniques. However, there was hardly any study into account this aspect, the present study focused on Odisha available, which used the remote sensing and GIS techniques in general and Khordha district in particular since the state in order to perform land suitability analysis for rice cultiva- has high scope for development in agricultural planning. tion. Thus, the present study is one of few studies to use such Odisha is surrounded by the Bay of Bengal on one side approach in order to perform land suitability analysis for rice and on the other sides by neighbouring states like West Ben- cultivation focusing on Khordha district of Odisha, India. gal, Jharkhand, Chhattisgarh and Andhra Pradesh (Fig. 1). Since fewer number of studies was performed over Odi- Among 30 districts of Odisha, Khordha is chosen for the sha region using satellite-derived measurements including study of land suitability analysis for rice crop looking into LULC for agricultural aspects, the present study intended account the data availability and signifcance of the loca- to address and analyse few relevant issues in this direction. tion. It may be noted that rice has been the principal food The current analysis will probably be helpful in this sense, crop of Odisha (earlier name Orissa) since fourteenth cen- especially for encouraging the farmers to grow appropriate tury A. D. or earlier (Panda 1997). The geographical area of crops in their land apart from the traditional ones. Khordha district lies between 19°63′N to 20°44′N latitude The present study aimed at performing the ‘land suit- and 84°91′E to 86°04′E longitude. Keeping in view of the ability analysis’ by adopting some relevant FAO principles tradition and food habits of the region, the present study (FAO 1978, 1994, 1996) besides using remote sensing and bears its signifcance as far as rice crop is concerned.

1 3 A Study on Land Suitability for Rice Cultivation in Khordha District of Odisha (India) Using… 121

Fig. 1 Location map showing the study area 1 3 122 S. S. Rath et al.

Table 1 Satellite data specifcations used for LULC classifcation Satellite Image Name IRS-P6 IRS P6 LISS III Sensor LISS III Date of acquisition 26th December 2010 Path 106 Geo-referencing Row 058 IGFOV (instantaneous geometric feld of 23.5 m view) Spectral bands (μm) B2 0.52–0.59 B3 0.62–0.68 Area of Interest B4 0.77–0.86 Masked B5 1.55–1.70 Swath 141 km Spectral resolution 23.5 m Unsupervised classification 3 Data and Methodology

The current study of land suitability analysis for rice crop Assignment of classes carried out by using the available climatic data for rain- fall and ‘near surface temperature’, soil analysis data of Khordha district and LULC data. The satellite derived LULC data from IRS-P6 LISS III for the year 2010 was Land use/Land cover considered (Table 1). The resolution for the desired LISS Map III images was 23.5 m. These images were used for prepar- ing the LULC map for the study area. The methodology Fig. 2 Flowchart for LULC Map preparation for the preparation of LULC map is presented in the form of a fow chart and illustrated in Fig. 2. The topographical map of the study area from ‘Survey 3.1 Creation of Soil and Climatic Database of India’ was used for geo referencing purpose and the delineation of the boundary. The satellite data from SRTM Soil samples (100) were collected from 10 blocks of (Shuttle Radar Topography Mission) was used for slope Khordha district (10 samples from each block) during the analysis in order to understand and depict the prevailing feldwork. The samples were collected from sub-surface topography of the region. ‘ERDAS IMAGINE’ software layer below 15 cm in the paddy felds during the fallow was adopted for processing and information extraction period. These paddy felds are located in the 10 blocks from satellite images. ‘ArcGIS’ was used for spatial data (Balianta, Bhubaneswar, Balipatna, Jatni, Khordha, Begu- analysis and various thematic map preparation. Hand held nia, Bolagarh, Tangi, Banapur and Chilika) of Khordha ‘GPS’ device was used to obtain the geographical coor- district of Odisha, India (Fig. 1). The sampling locations dinates of the observed feld locations during the ground were chosen by considering a homogeneous distribution of truthing study for collecting soil samples and LULC them in the respective blocks. For each sampling location, information. the felds were divided into several homogenous units and Attribute data sets for meteorological variables such as surface litters were removed. With the help of spade, ‘V’ temperature and precipitation were collected from district shaped cut was made up to a depth of 15 cm. Slices of soil agricultural ofce and India Meteorological Department was taken out from the sides of ‘V’ shaped cut. All the soils (IMD) for the period 2002–2012. Spatial interpolation of collected from a single location were thoroughly mixed and rainfall data collected for 10 climatic stations was car- roots, stones and foreign materials removed. Quartering was ried out using ‘ArcGIS’. In addition to these, IMD data done for each location and required amount of soil placed sets, TRMM (Tropical Rainfall Measuring Mission) rain- in sampling bags. Tagging was done for each bag with sam- fall considered for Khordha region. The satellite derived ple number, location and some previous crop information. monthly averaged TRMM rainfall data were considered The samples were then sent to the soil-testing laboratory at for assessing the availability of rainfall for cultivation of Orissa University of Agriculture and Technology (OUAT) rice crop. for physical and chemical analysis. In addition to the 100 samples collected, few supplementary soil maps prepared

1 3 A Study on Land Suitability for Rice Cultivation in Khordha District of Odisha (India) Using… 123 by OUAT were used. The whole data set was incorporated connected to the famous wetland ecosystem, “the Chilika in the GIS platform for ‘soil database’ preparation. Lake”. Thus, the prevailing water bodies are noticeable in In order to prepare the climatic database for rainfall the southern part of the study area. The eastern part of the analysis, daily rainfall was collected from IMD and district district is mostly covered by urban infrastructures, which agricultural ofce. Monthly average rainfall values were include the capital city ‘Bhubaneswar’ and the neighbour- obtained from the daily data for the years 2002 to 2012. ing regions. The western part of the district is covered These observations were used as input in the attribute tables by natural vegetation as notifed reserved ‘forest lands’. of the point location of the ten blocks as climatic station The ‘forest land’ also extends to the northern part of the rainfall data within GIS platform. It may be noted that rain- region. The mixed vegetation area can be noticed near the fall is a highly signifcant piece of hydrologic data and is ‘forest lands’. The agricultural lands are distributed all recorded through comprehensively designed rainfall station over the district randomly (Fig. 3). networks. However, the rainfall records are often incomplete The satellite image of the study area was categorized because of missing data in the measured period, or insuf- with 45 classes through unsupervised classifcation for fcient number of stations in the study region. To overcome determining the LULC distribution (refer Fig. 2 for meth- this problem, the inverse distance weighting (IDW) interpo- odology). The LULC types were reduced to fve broad lation technique was adopted. Interpolation estimates were classes like water body, forest area, mixed vegetation, agri- made based on available values at nearby locations weighted cultural land and urban areas (Fig. 3). Accordingly, the only by distance from the considered location. total area covered under ‘water body’ category was found to be 6399.27 ha and ‘forest cover’ 27,718.79 ha. How- ever, ‘mixed vegetation’ and ‘agricultural land’ cover sig- 4 Results and Discussion nifcantly larger areas, i.e., 107,722.67 and 122,183.38 ha, respectively. The satellite image used for this analysis is Adequate agricultural exploitation of the climatic potentials for the harvesting month for rice crop, i.e. December. In and maintenance of land productivity largely depend on soil Khordha district, the primary crop being the rice crop, fertility and the management of soils that is ecologically the quantity of agricultural land obtained from the image sustainable. For assessing the suitability of soils for crop analysis is considered as amount of land currently used for production, soil requirements of crops must be known. These the rice cultivation (i.e. 122,183.38 ha). The built up area requirements (Table 2) must be understood within the con- that comprised mostly the capital city Bhubaneswar and text of limitations imposed. In view of this, the prevailing surroundings, covers 11,859.41 ha. The main purpose of soil characteristics of Khordha district were analysed along this LULC map preparation was to fnd out the amount of with LULC and the climatological variation of temperature land currently under rice cultivation. This calculated area and rainfall in order to support the suitability analysis. is based on the LULC distribution during 2010. Moreover, the same might have changed in the due course of time 4.1 Prevailing Land Use and Land Cover because of urban expansion in Bhubaneswar and Khordha town region. However, it would not infuence the suitabil- The study area (Fig. 1) comes under “East and South- ity analysis for rice crop cultivation as far as the quality of Eastern Coastal Zone” agro ecological sub-region. Spatial the land is concerned. It might still infuence quantitatively variability in the LULC distribution over Khordha is quite since the rice production is infuenced by the reduction in evident from the Fig. 3. The southern part of the district is agricultural land availability.

Table 2 Land evaluation Requirement Highly suitable Moderately suitable Not suitable parameters and rice crop requirements adopted from FAO Nitrogen 150–200 kg/ha 100–150 kg/ha < 50 kg/ha framework (FAO 1976, 1983) Phosphorous 2–3 kg/ha 1–2 kg/ha 0–1 kg/ha Potassium 30–40 kg/ha 20–30 kg/ha 10–20 kg/ha Soil pH 5–6 4 < 4 Soil organic carbon 0.5–0.75% 0.25–0.5 0–0.25 Rainfall > 950 mm 550–950 mm < 550 mm Slope 0–3% 3–10% > 10% Soil texture C, CL, SiCL, SiC, SiL C, L, SCL LS Soil depth Very deep Deep Shallow Number of dry months 2–4 5–6 > 7

1 3 124 S. S. Rath et al.

Fig. 3 Land use and land cover map (2010) of Khordha district, Odisha

4.2 Soil and Slope 6, 7, 8). All the three types of soil are having potential for rice cultivation. An exclusive soil map is not included in Soil quality parameters such as pH, nitrogen content, this study since the relevant soil nutrient maps are used organic carbon content, potassium content and phospho- in discussing the findings. rous content were determined from chemical analysis ‘Soil pH’ is an indication of the acidity or alkalinity of carried out at OUAT soil-testing laboratory. For this pur- soil. The most accurate method for determining the ‘soil pose, standard procedure mentioned in the soil-testing pH’ is by a ‘pH meter’. The efect of soil pH is signifcant on methods manual (MMSTI 2011) was followed. Further, the solubility of minerals or nutrients. Slightly acidic soils the soil-testing results were analysed quantitatively. These having a pH value of 6–7 are better for paddy cultivation. data were validated with the available soil nutrient maps However, it was found to be grown in a wide range of pH (1:50,000 scale) of Khordha district. After validation, the values varying from 4 to 8 (FAO 1976). The soil pH values soil thematic maps were prepared for the said nutrients. of the whole district Khordha varied within the accepted According to the soil category map (from OUAT), there range 4.5–7.5 and found to be suitable for rice cultivation are generally three types of soils available in the Khordha (Fig. 4). The soil pH is varied by taking into account the district, i.e. (1) Alfisols, (2) Ultisols and (3) Entisols. amount of water present or the submergence level of soil. The Alfisols consisted of the deltaic alluvial soil in the Accordingly, it was realized that most of the agricultural eastern part and red loamy soils are present in the north- land (Fig. 3) had soil pH in the range 4.5–6.5 (Fig. 4), indi- western parts of Khordha. These soils are light textured, cating the region to be quite suitable for rice cultivation. usually free from lime, deficient in nitrogen, phosphate Nitrogen is also a critical element in proftable rice cul- and soil organic carbon. The pH of the soil varied from tivation. Too little nitrogen results in low yield and proft, 6.0 to 7.0. Ultisols included laterite and/or lateritic soil, while too much can increase vulnerability to insects and red and yellow soils and usually seen in the northern and diseases, high levels of sterility and reduced yield. Although north central part of the district. The pH of these soils nitrogen is the most abundant element in the atmosphere, (Ultisols) range from 4.5 to 6.0. The third category of plants cannot use it until it is naturally processed in the soil, soil Entisols, included the coastal alluvial soils along the or added as fertilizer. Fields with no recent history of leg- Chilika lake and those in the central part of the district. umes require a minimum of 90 kg of Nitrogen per hectare The texture of this type of soil is generally sandy to loamy (or 200 kg/ha urea) prior to achieve the desired nitrogen and usually lack in nitrogen, phosphoric acid and humus, level. Thus, it was important to understand the soil Nitro- which could be seen in the soil nutrients map (Figs. 4, 5, gen content of Khordha district before performing suitability

1 3 A Study on Land Suitability for Rice Cultivation in Khordha District of Odisha (India) Using… 125

Fig. 4 Soil pH map of Khordha district, Odisha

Fig. 5 Representation of soil nitrogen content distribution in Khordha district, Odisha

analysis for the rice crop cultivation. The spatial distribution found to have soil Nitrogen content in the range 100–200 kg/ of soil Nitrogen content is illustrated in Fig. 5. The availabil- ha (Fig. 5). This is again quite favourable for rice cultiva- ity of the soil Nitrogen content in the units of kg/ha deter- tion and does not need much application of Nitrogen-based mined the amount of fertilizer to be applied in the rice feld. fertilizers. Most of the agricultural land in Khordha district (Fig. 3) was

1 3 126 S. S. Rath et al.

Fig. 6 Representation of soil organic carbon distribution in Khordha district, Odisha

Fig. 7 Spatial distribution of soil potassium content in Khordha district of Odisha

Soil organic carbon, the major component of organic mat- soil organic carbon is quantifed by percentage value and in ter in soil, is extremely important in all relevant processes. this study, the highest amount was found to be 0.75% located The annual rate of loss of organic matter can vary greatly, at the northern part of the district (Fig. 6) near the Mahanadi depending on cultivation practices, the type of plant/crop river basin. The majority of the land has a moderate to high cover, drainage status of the soil and weather conditions. The amount of soil organic carbon content ranging from 0.25

1 3 A Study on Land Suitability for Rice Cultivation in Khordha District of Odisha (India) Using… 127

Fig. 8 Spatial distribution of soil phosphorous content in Khordha district of Odisha

to 0.75% (Fig. 6) mostly overlapping with the designated shown in Fig. 9. The map revealed that most of the area of region of agricultural land when compared with the LULC Khordha district had a slope within a range of 0–6% and map of Khordha district (Fig. 3). very few areas had a slope of > 30% usually for mountain- Soils having adequate Potassium allow plants to develop ous regions located in the south-western part. Thus, most of rapidly and outgrow plant diseases, insect damage and pro- the land satisfed the rice crop requirement in terms of slope tect against winter freeze damage. Soil minerals are rich in i.e. < 10% according to FAO guidelines (Table 2). potassium naturally. Clay soils usually contain more miner- A comprehensive analysis of the soil nutrient availabil- als and tend to have higher levels of potassium, making it ity (Figs. 4, 5, 6, 7, 8) and the requisite slope information most suitable for rice cultivation. On the other hand, sandy (Fig. 9), along with the consideration of their required ranges textured soils tend to have lower amounts of potassium and (Table 2), indicates that the region is quite capable of host- are unsuitable for rice cultivation. The requirement of potas- ing rice cultivation. However, it does not guarantee that the sium nutrient in soil is similar to the nitrogen in amount as region is suitable for this purpose since the climatic factors far as rice cultivation is concerned. Analysis of the spatial such as the prevailing temperature and abundancy in rainfall distribution of Potassium content in the soils of Khordha also play a signifcant role. district revealed that the values were mostly < 280 kg/ha (Fig. 7). Moreover, most of the agricultural lands had the 4.3 Rainfall and Temperature potassium content within 118–280 kg/ha. Adequate phosphorous availability for plants stimulates Several climatic factors including temperature, solar radia- early growth and accelerates maturity. The active phospho- tion or sunlight, rainfall and cyclonic storms impact rice rous pool that is relevant for the rice crops is the one that yield by directly infuencing the physiological processes and can be released into the soil solution but is generally small in indirectly through diseases and insects (Yoshida 1981). The comparison to the fxed phosphorous types. A brief analysis important aspects of rice cultivation such as crop growing of the soil phosphorous thematic map of Khordha district period, productivity and stability, get afected by the said (Fig. 8) indicated that most part of the region had adequate climatic factors in diferent ways over tropics. Though tem- amount (< 5 kg/ha) required for rice crop. perature is usually favourable for rice growth throughout Apart from the soil nutrient analysis, slope analysis was the year in the tropical regions, irrigation is not available in also carried out using SRTM data and given as input to the most of the places and, therefore, the rice cultivation usually GIS platform during the land suitability analysis. The slope starts with rainy season or monsoon. Since abundant sun- map was prepared from 90 m resolution SRTM data and light is available over tropics, analysis of climate data sets

1 3 128 S. S. Rath et al.

Fig. 9 Slope map of Khordha district of Odisha. The specifed percentages (as shown in the legend in diferent colours) rep- resent the slopes in fve ranges

for meteorological variables such as near-surface tempera- the maximum temperature is ~ 34 °C over Cuttack-Bhu- ture and rainfall for a specifc location can indicate whether baneswar region as illustrated by their study. The increase the region is suitable for cultivation of a particular crop or in temperature results in decrease in yield over this region. not. In view of this, near-surface temperature and rainfall The present study being observational based, found that over Khordha district were considered during 2002–2012 during kharif season, the decrease in average temperature taking into account the data availability. occurred until July agreeing with the fndings of Krishnan Though temperature-specifc analysis is relatively less, et al. (2007). However, the subsequent rise was found to few studies like that of Krishnan et al. (2007) emphasized be until September instead of October as demonstrated by upon the impact of temperature on rice yield by taking into Krishnan et al. (2007) though ~ 0.9 °C diference in temper- account various stages of growth. According to this study, ature observed between the mean values of these 2 months temperature is usually on higher side during the sowing stage (Fig. 10). The possible reasoning could be consideration followed by a decrease in average temperature in the tillering of diferent durations (in terms of year) for monthly vari- stage and increase in temperature is observed during panicle ation of average temperature. According to Krishnan et al. initiation and fowering stage. During the fowering stage, (2007), the higher value of monthly average temperature

Fig. 10 Variation of monthly mean temperature (blue line with circular marker) and rainfall (red line with diamond marker) over Bhubaneswar, and TRMM rainfall (green dotted line with triangular marker) over Khordha District of Odisha

1 3 A Study on Land Suitability for Rice Cultivation in Khordha District of Odisha (India) Using… 129 during October corresponds to fowering stage after which During June–October, it appeared from IMD observation, the variation should show a declining trend. Though there is TRMM data and the Odisha rainfall monitoring system that small diference in the monthly average temperature between Khordha district received signifcant amount of rain in these September and October months, the values were in the range months as obvious from the monthly variations (Figs. 10, 27.5–29 °C, which should be sufcient for fowering. In 11). The same could further be realized by analysing spa- addition, the range of monthly average temperature during tial distribution of annual rainfall (Fig. 12) by considering kharif period was within 25–31 °C (Fig. 10), which might be appropriate interpolated values through GIS (IDW interpo- considered as reasonable and favourable for rice cultivation. lation is used). The interpolated result provided the values In order to understand signifcance of rainfall during of rainfall in every point location inside the boundary of the kharif period, observational data from IMD and TRMM study area and indicated that moderate amount of accumu- were considered for illustrating the monthly variations dur- lated annual rainfall received in the agricultural area during ing 2002–2012 (Fig. 10). The variation of monthly average 2002–2012 supporting the requirement for rice cultivation. IMD rainfall over Bhubaneswar showed a maximum value of 391.5 mm and the range during June–October period varied 4.4 Rice Crop Requirements and Suitability Analysis between 204.8 and 166.1 mm. However, the spatio-tempo- rally averaged TRMM rainfall over whole Khordha district Various requirements of rice crop (Table 2) in the agro cli- showed a maximum value of ~ 338.6 mm and the range matic region of Khordha district were collected from FAO varied 210.8–149.7 mm during the said period. Though guidelines (FAO 1983) and included in GIS analysis. Soil the maximum rainfall occurrence did not coincide as far as thematic maps or layers (Figs. 4, 5, 6, 7 8) and interpolated month is concerned, the variations were quite similar. rainfall accumulated values (Fig. 12) created with ArcGIS In order to support the argument of rain fed rice crop- were combined in order to perform the suitability analysis ping in Khordha, it was also important to know the number for the rice crop in the study area. The methods followed of rainfall days in a month. Therefore, the number of days for the suitability analysis illustrated in the fowchart shown in which rainfall occurred over Khordha district was col- in Fig. 13. All the layers separately converted to raster and lected from Odisha rainfall monitoring system for the period weighted overlay analysis was carried out using the raster 2008–2012 as per the availability of data. Monthly aver- layers. Specifed ranking was provided for highly, moder- age number of rainy days was computed by considering the ately and not suitable lands as ‘rank 1’, ‘rank 2’ and ‘rank non-zero values of daily rainfall over the district (Fig. 11). 3’, respectively. The overlaying procedure was repeated in There were ~ 127 days in which rainfall occurred in a year. order to discard the urban area and forest cover since they However, more than 100 rainfall days were found during the were not meant for agriculture. Accordingly, the fnal suit- 5 months: June to October period, and each month had more ability analysis map was generated and shown in Fig. 14. than 10 rainy days. The maximum number of rainy days was The suitability map shows that about 103,934 ha of land in July and August (Fig. 11) agreeing with the maximum is highly suitable for rice cultivation in Khordha district. average rainfall during these 2 months (Fig. 10). Thus, the However, a sizable area available is moderately suitable for 5 months with maximum number of rainy days over Khordha the rice cultivation and found to be 91,797 ha. A reasonably matched with the main rice cultivation period supporting the less amount of land (3216 ha) was found to be not suitable agriculture practice over this region. for rice cultivation due to degradable soil conditions and

Fig. 11 Illustrating average number of rainfall days in a month for Khordha district of Odisha by taking into account 5 years of data from 2008 to 2012

1 3 130 S. S. Rath et al.

Fig. 12 Spatial distribution of rainfall (spatially interpo- lated with IDW method) over Khordha district of Odisha during 2002–2012

Survey of India Soil Satellite Image Meteorological Data 5 Concluding remarks Topo-sheet Map IRS P6 LISS III Rainfall & A regional land suitability mapping is carried out for Khordha district of Odisha, India. The study takes into Base Map & Digital Soil Land use/Land Interpolation of Slope Map Database cover Map Rainfall account LULC distribution for the year 2010 in order to com- pute the amount of agricultural land and classify into three categories those are highly, moderately or not suitable for rice crop cultivation. From this analysis it is found that the

Agro edaphic units Agro climatic Units total amount of agricultural land is currently 122,183.38 ha and is relatively less as compared to the amount of land (195,731 ha) that is potentially suitable for rice cultivation, which is either highly or moderately suitable type. There- fore, it is concluded that this region has more potential for Overlay Crop Requirement Analysis rice cropping as compared to the presently perceived and cultivated area. The present analysis fnds that 103,934 ha is

Crop Suitability highly suitable and 91,797 ha of land is moderately suitable Map for rice cropping. The suitability analysis is performed by considering soil Fig. 13 Illustration of land suitability analysis for rice crop through nutrient contents as well as the available near-surface tem- fow-chart perature and rainfall data. The soil sample analysis suggests that there is high amount of nitrogen content (100–200 kg/ ha) and the potassium content in the soil is mostly < 280 kg/ insufcient rainfall. However, those areas might be made ha, while most of the agricultural lands have reasonable reasonably suitable by proper soil treatment and fertilizer Potassium content within the range 118–280 kg/ha. The and irrigation application. This calculation did not include phosphorous content in the soil is mostly found to be < 5 kg/ urban and forest areas. ha and is within the desired range or adequate for the rice cultivation. Similarly, the soil organic content is within the desired range 0.25–0.75% and the soil pH is in the appropri- ate range (i.e. 4.5–7.5) as well. Besides these aspects, the

1 3 A Study on Land Suitability for Rice Cultivation in Khordha District of Odisha (India) Using… 131

Fig. 14 Suitability map for rice crop for Khordha district of Odisha

water retention capacity should be high to make the soil the present study is quite helpful from agriculture point of suitable for the rice cultivation. In view of these, it is real- view, especially the farmers who are interested in rice farm- ized that the prevailing soil in Khordha district of Odisha is ing. Therefore, the land evaluation, LULC mapping as well quite appropriate to meet the desired nutrient requirements as climatic analysis would help the farmers to identify their for the cultivation of rice. lands for optimum use in this region. Though this study is In addition to the soil characteristics and appropriate limited to Khordha district, it can still be performed for agricultural land availability, climatic conditions must be other regions of the country as well as for other crops in favourable for the cultivation of rice. The observed monthly order to have better land utilization and improved and smart average temperature during June–October, the main period farming. Consequently, the agricultural productivity can be of rice cultivation, is found to be within the favourable lim- increased in order to meet the demand of food due to popula- its, i.e. 25–31 °C. Further, the long-term data analysis for tion growth and ensure food security. rainfall suggests that Khordha district receives a substantial Further, the changing climate can have a significant amount of rainfall annually ranging from 1070 to 1634 mm impact on agricultural productivity including that of rice in and during June to October, the received amount lies within India (Mall and Aggarwal 2002; Aggarwal and Mall 2002). the range 210.8–149.7 mm as evident from the satellite The southern and western parts of India, which have rela- measurements from TRMM. The receipt of substantial rain- tively lower temperatures as compared to the northern and fall during more than 100 rainy days within the 5-month eastern regions, are more likely to show greater sensitivity to period of June–October and the prevailing suitable tempera- rice yields in the changing climate scenario (Aggarwal and ture is quite encouraging for rice cultivation since they meet Mall 2002). In view of this, future studies must include the the desired climatic requirements. Though the present study impact of climate change on rice cultivation. However, it is does not take into account the availability of sunlight or day important to be quite careful when results from the impact length or sunshine, it can still be reasonable to state that the assessment studies are used while considering the mean region receives signifcant amount of solar radiation during changes in climatic parameters. the kharif season as well as harvesting period since it lies in the tropical region. The land suitability analysis suggests that there is a siz- References able area available in Khordha district for rice cultivation and the unrealized area (currently not under cultivation) Aggarwal PK, Mall RK (2002) Climate change and rice yields in may be considered by the farmers for this purpose. Thus, diverse agro-environments of india. II. Efect of uncertainties in

1 3 132 S. S. Rath et al.

scenarios and crop models on impact assessment. Clim Change Hansen MC, DeFries RS, Townshend JR, Sohlberg R (2000) Global 52:331–343 land cover classifcation at 1 km spatial resolution using a clas- Bhagat RM, Singh S, Sood C, Rana RS, Kalia V, Pradhan S, Immerzeel sifcation tree approach. Int J Remote Sens 21(6–7):1331–1364 W, Shrestha B (2009) Land suitability analysis for cereal produc- Kihoro J, Bosco NJ, Murage H (2013) Suitability analysis for rice tion in Himachal Pradesh (India) using Geographical Information growing sites using a multicriteria evaluation and GIS approach System. J Indian Soc Remote 37(2):233–240 in great Mwea region, Kenya. Springer Plus 2(1):1–9. https://doi.​ Biro K, Pradhan B, Buchroithner M, Makeschin F (2013) Land use/ org/10.1186/2193-1801-2-265 land cover change analysis and its impact on soil properties in Krishnan P, Swain DK, Bhaskar BC, Nayak SK, Dash RN (2007) the northern part of Gadarif region, Sudan. Land Degrad Dev Impact of elevated CO 2 and temperature on rice yield and meth- 24(1):90–102 ods of adaptation as evaluated by crop simulation studies. Agric Choudhury I, Chakraborty M, Santra SC, Parihar JS (2006) Characteri- Ecosyst Environ 122(2):233–242 zation of agroecosystem based on land utilization indices using Kumar S, Patel NR, Sarkar A, Dadhwal VK (2013) Geospatial remote sensing and GIS. J Indian Soc Remote 34(1):23–37 approach in assessing agro-climatic suitability of soybean in rain- Dengiz O (2013) Land suitability assessment for rice cultivation based fed agro-ecosystem. J Indian Soc Remote 41(3):609–618 on GIS modeling. Turk J Agric For 37(3):326–334 Kuria D, Ngari D, Waithaka E (2011) Using geographic information Devi GMS, Kumar KSA (2008) Remote sensing and GIS application systems (GIS) to determine land suitability for rice crop growing for land quality assessment for cofee growing areas of karnataka. in the Tana delta. J Geogr Reg Plan 4(9):525 J Indian Soc Remote 36:89–97 Lal R (2008) Managing soil water to improve rainfed agriculture in FAO (1976) A framework for land evaluation. Soils Bulletin 32. Food India. J Sustain Agric 32(1):51–75 and Agriculture Organization of the United Nations, Rome, Italy. Mall RK, Aggarwal PK (2002) Climate change and rice yields in ISBN 92-5-100111-1. Available online at http://www.fao.org/ diverse agro-environments of india. I. Evaluation of impact docre​p/x5310​e/x5310​e00.HTM assessment models. Clim Change 52:315–330 FAO (1978) Report on the agro-ecological zones project, vol 1. Meth- MMSTI (2011) Methods manual soil testing in India, Department of odology and results for Africa. World Soil Resources Report 48/1, Agriculture and Cooperation, Ministry of Agriculture, Govern- FAO, Rome, p 158 ment of India, New Delhi FAO (1983) Land evaluation of rainfed agriculture. Soils Bulletin, 52. Panda SK (1997) Land and agriculture in medieval Orissa (1000–1600 Food and Agriculture Organisation, Rome, p 237 A.D.). In: Patnaik NR (ed) Economic history of Orissa, Chap- FAO (1994) AEZ in Asia. In: Proceedings of regional workshop on ter 11. ISBN: 8173870756. Indus Publishing Company, New agro-ecological zones methodology and applications, Held at FAO Delhi Regional Ofce for Asia and the Pacifc (RAPA), Bangkok, Thai- Samanta S, Pal B, Pal DK (2011) Land suitability analysis for rice land, 17–23 November 1991, p 259 cultivation based on multi-criteria decision approach through GIS. FAO (1996) Guidelines: agro-ecological zoning. Soils Bulletin 73. Int J Sci Eme Tech 2(1):12–20 FAO, Rome, p 78 Sathish A, Niranjana KV (2010) Land suitability studies for major Friedl MA, McIver DK, Hodges JCF, Zhang XY, Muchoney D, Strahler crops in Pavagada taluk, Karnataka using remote sensing and GIS AH, Woodcock CE, Gopal S, Schneider A, Cooper A, Baccini A, techniques. J Indian Soc Remote 38(1):143–151 Gao A, Schaaf C (2002) Global land cover mapping from MODIS: Wessel KJ, De Fries RS, Dempewolf J, Anderson LO, Hansen AJ, algorithms and early results. Remote Sens Environ 83(1):287–302 Powell SL, Moran EF (2004) Mapping regional land cover with Gumma M, Thenkabail PS, Fujii H, Namara R (2009) Spatial mod- MODIS data for biological conservation: examples from the els for selecting the most suitable areas of rice cultivation in the Greater Yellowstone Ecosystem, USA and Para State, Brazil. Inland Valley Wetlands of Ghana using remote sensing and geo- Remote Sens Environ 92:67–83 graphic information systems. J Appl Remote Sens 3(1):033537- Yoshida S (1981) Fundamentals of rice crop science. International Rice 1–033537-21. https​://doi.org/10.1117/1.31828​47 Research Institute, Los Baños, Laguna, Philippines, p 269

1 3 Earth Systems and Environment (2018) 2:133–143 https://doi.org/10.1007/s41748-018-0038-x

ORIGINAL ARTICLE

Potential for Phytoextraction of Cu by Sesamum indicum L. and Cyamopsis tetragonoloba L.: A Green Solution to Decontaminate Soil

Hira Amin1 · Basir Ahmed Arain1 · Muhammad Sadiq Abbasi2 · Taj Muhammad Jahangir3 · Farah Amin4

Received: 18 August 2017 / Accepted: 7 February 2018 / Published online: 16 February 2018 © Springer International Publishing AG, part of Springer Nature 2018

Abstract Phytoextraction is a plant based-technique for removing toxic heavy metals from polluted soil. The experiment reported in this paper was undertaken to study the basic Cu phytoextraction potential of Sesamum indicum in comparison with Cya- mopsis tetragonoloba for remediation of Cu contaminated soil in the framework of a pot-experiment. Plants were subjected to seven Cu concentrations (0, 25, 50, 100, 150, 200, and 300 mg kg−1 soil) for 12 weeks. The morphological (i.e. growth) and biochemical (i.e. chlorophyll) parameters of both the plant species were observed throughout the experimental period; the phytoextraction efciency of S. indicum and C. tetragonoloba were also determined. Most growth parameters were reduced under high Cu stress. Our results shows that at low concentration (25 mg Cu kg­ −1) all the growth and biochemical parameters were increased but at elevated Cu concentrations, root length, shoot length, and biomass (fresh and dry) were all signifcantly decreased (p < 0.05). Chlorophyll contents also declined with increasing concentrations of Cu, when com- pared with control. A consistent increase of Cu accumulation in root and shoot of both S. indicum and C. tetragonoloba with rising concentrations of Cu in soil was noted for all tested treatments. In this study, both plant species showed quite high Cu tolerance and accumulation efciency, even though C. tetragonoloba have higher Cu accumulation and tolerance indices than that of S. indicum. At 300 mg Cu ­kg−1, the highest Cu concentration was found in the root (282.08 mg Cu ­kg−1) followed by leaf (105.78 mg Cu kg­ −1), stem (65.30 mg Cu kg­ −1), and pod (8.13 mg Cu kg­ −1) of S. indicum. In contrast, C. tetragonoloba had highest Cu concentration primarily in the root (158.45 mg Cu ­kg−1) followed by the stem (154.73 mg Cu ­kg−1), leaf (152.32 mg Cu ­kg−1), and pod (8.13 mg Cu ­kg−1). Considering rapid growth, high biomass, tolerance, accumula- tion efciency, bioconcentration factor (BCF) > 1, bioaccumulation coefcient (BAC) > 1 and translocation factor (TF) > 1 established C. tetragonoloba as a potential candidate plant for the decontamination of slightly Cu-polluted soil where the growth of plants would not be impaired and the extraction of Cu could be maintained at satisfying levels. Therefore, the present study suggested that C. tetragonoloba could possibly be used as a viable tool for phytoextraction.

Keywords Soil pollution · Copper · Accumulation · Translocation · Phytoextraction

1 Introduction

* Hira Amin Agricultural soil contamination due to hazardous heavy met- [email protected] als is a signifcant environmental problem. The accumulation of substantial toxic metals in soil profle adversely afects 1 Institute of Plant Sciences, University of Sindh, Jamshoro 76080, Pakistan agricultural production. As a consequence of heavy metal contaminated soil, the harmful metals turn into a part of 2 Department of Mathematics and Statistics, Quaid-e-Awam University of Engineering, Science and Technology, a natural way of life and furthermore cause serious risk to Nawabshah 67480, Pakistan plants, animals, humans, and entire soil condition including 3 Institute of Advanced Research Studies in Chemical soil organisms (microbes) (Nagajyoti et al. 2010; Singh and Sciences, University of Sindh, Jamshoro 76080, Pakistan Prasad 2014). 4 National Centre of Excellence in Analytical Chemistry, Rapid industrialization and urbanization, improper University of Sindh, Jamshoro 76080, Pakistan use and transfer of unsafe metal containing squanders,

Vol.:(0123456789)1 3 134 H. Amin et al. inappropriate utilization of manure, commercial fertilizers In this way, the sustainable clean-up of Cu-contaminated and pesticides are the primary sources for heavy metal entry soil to confne its impact on the environment is required. into the agricultural soil (Dmitry et al. 2015; Amel et al. Remediation of substantial metal contaminated soil by 2016). Toxic heavy metals, including lead, copper, zinc, cad- conventional physical and chemical techniques are well- mium, chromium, nickel, cobalt, and iron are major envi- reported in literature but these are insufcient and unaccep- ronmental contaminant components, especially in regions table for agricultural lands as they require huge venture and with high anthropogenic activities (Nagajyoti et al. 2010; innovative resources (Oh et al. 2013). Along these lines, a Subhashini et al. 2013). plant-based technology has been developed as a cure for Among heavy metals, copper (Cu) is the most abundant the manageable control of elevated levels of diferent toxic transition heavy metal in earth’s crust, known as a coinage heavy metals in soil. metal, that exists in the monovalent and divalent oxidation Phytoextraction (or else known as phytoaccumulation/ states. Cu has assumed an essential part of human social, phytoabsorption/phytosequestration) a promising, cheaper, cultural, modern industrial, and technical advancement since eco-efcient technique for the remediation of heavy metal early times. Like other transition metals (e.g. Ag and Au) polluted soil has attracted immense consideration in recent Cu has broadly utilized as a part of an extensive variety of times. Phytoextraction is the take-up of heavy metals from activities, for example, fungicides and pesticides, wood addi- soil to root by metal-accumulating plants and their transfer tives, pigments used in glasses, paints and furthermore as a to easily harvestable parts (i.e. aboveground shoot/biomass) superconductor (Rebecca 2006). Cu is considered as basic (Seema et al. 2015; Amanullah et al. 2016). essential micronutrient (trace element) in low concentration The plant species appropriate for phytoextraction purpose for animal (Wintz et al. 2002) and plant biological functions should have great tolerance to the harmful impacts of the (Mahmood and Islam 2006; Muhammad et al. 2015). In an substantial metals in soil, high accumulation and transfer enzyme chemical reaction, Cu acts as cofactor and activa- of metals from below-ground to above-ground parts, fast tor enlightening enzymes/substrate metal complex (Mildvan growth and development rate, high production of above- 1970) and in addition, serves as catalysts for homogeneous ground biomass (shoot) and must be simple to cultivate and and heterogeneous chemical reactions. Cu phytotoxicity in harvest (Shabani and Sayadi 2012; Hazrat et al. 2013). Sev- plants changed the take-up of other essential nutrients. In eral workers have reported the phytoremediation potentials plants, excess Cu is more toxic relatively with other essen- of plant species belonging to botanicals families, specifcally tial trace elements, such as Zn and Mn which are possibly the Brassicaceae, Asteraceae, Fabaceae, Poaceae and Cheno- toxic in excess (Dresler et al. 2014). High take-up of Cu podiaceae. Even phytoremediation potentials by Chlorophy- well beyond plant prerequisites brings about toxic impacts ceae are well documented in the literature (Gawronski and on plant normal functions (Monni et al. 2000). Excess Cu Gawronska 2007; Anjum et al. 2014; Balaji et al. 2014a, b, has become deleterious to plants and is responsible for 2016). The heavy metal take-up limit, accumulation, exclu- reduced seed germination, stunted root and shoot develop- sion, compartmentation and mechanisms of metal tolerance ment, reduced yield, oxidative stress generation and forma- difer among various plant species and furthermore between tion of reactive oxygen species (ROS) (Stadtman and Oliver diferent parts of plants (Sharma et al. 2014). 1991; Azooz et al. 2012) that causes damaging infuence to Sesame (Sesamum indicum L.) is an important oilseed metabolic pathways and harm to macromolecules (Hegedus crop of family Pedaliaceae (Elleuch et al. 2007). Sesame et al. 2001). seeds are considered to be the oldest oilseed crop well- Soil contamination by Cu is reliant on both natural and known to man, highly resistant to drought conditions and also anthropogenic sources. Agricultural soil receives a sub- have the ability to grow where most of the crops are unsuc- stantial amount of Cu from environmental contamination cessful to cultivate (Dawodu et al. 2014). Recent studies that came about because of anthropogenic activities which illustrate that sesame seeds have been widely used as alter- frequently utilize Cu for agricultural farming, industrial and native feedstock for biodiesel production. The methyl ester mechanical purposes (Jiang et al. 2004). The inappropriate, extracted from sesame seeds can efectively be utilized as inequitable and untreated utilization of Cu-containing pes- petrodiesel (Ahmad et al. 2011). The viscosity and density ticides to manage plant pests, diseases and infections have of methyl esters of sesame seed oil were observed to be near brought about the surplus accumulation of Cu in agricultural that of diesel, while the heating (7.5%) and calorifc (5.4%) soil profle (Mackie et al. 2012; Muhammad et al. 2015). values were lower than that of diesel to some extent (Saydut Some of the distinguished anthropogenic activities including et al. 2008). mining and smelting operations, industrial and urban activi- Moreover, guar (Cyamopsis tetragonoloba L.) is a ties, inorganic–organic fertilizers, liming, sewage sludge and remarkable legume crop plant of family Fabaceae. The wastewater for irrigation are the most common sources of Cu endosperm of guar seeds contains gum, which has to pick up to the agricultural soil (Herawati et al. 2000). signifcance as a non-food product (Ashraf et al. 2002). The

1 3 Potential for Phytoextraction of Cu by Sesamum indicum L. and Cyamopsis tetragonoloba L.:… 135 utilization of fuid (gum) extracted from guar in fracking for Soil samples were collected from equidistant (2 m) inter- oil has extended the signifcance of guar around the world. val and mixed well to make single uniform bulk soil. The The oil business industry has begun utilizing extracted fuid soil samples were air dried for about 15 days and ground from guar in pressure-driven cracking of rocks or coal beds with pestle and mortar to pass through a size of 2 mm mesh (process of hydraulic fracturing) attributable to its high and use for further study. In order to increase soil porosity, consistency and profciency for petroleum and natural gas sandy-loam soil samples were mixed together with sand in extraction (Deepak et al. 2014; Abidi et al. 2015). A key 3:1 proportion. factor of using legumes for phytoremediation is their func- tion in giving additional N-compounds to the soil, there- 2.3 Soil Sample Measurements fore enhancing soil richness and fertility and sustaining the development of plants and soil organisms (Xiuli et al. 2013). For characterization, Soil pH was measured with a pH- This study was set up to investigate a plant species that meter (InoLab-WTB GmbH; Weilheim, Germany) using could tolerate high toxicity of Cu treatment levels in soil. glass electrode at 1:2 (w/v) ratio of soil to water suspension Both the plant species, i.e. S. indicum and C. tetragonoloba (Rachit et al. 2016). The electrical conductivity (EC) was have fast growth and development rate also high above- measured with an electrical conductivity meter (WTW— ground biomass. It was estimated that plant species with 330i) at the 1:2 (w/v) ratio of soil to water suspension high phytotolerance could then be exploit for their phytoex- (Rachit et al. 2016). Organic matter (OM) and organic car- traction potential. In this way, considering above mentioned bon (OC) (%) were measured according to Walkley and points, the objectives of the present study are to (1) observe Black (chromic acid titration) method (Fanrong et al. 2011). the phytotoxicity of Cu on morphological and biochemical parameters of S. indicum and C. tetragonoloba, (2) examine 2.4 Preliminary Screening for Cu Treatment Levels the Cu accumulation efciency in below and above ground biomass growing on a Cu polluted soil and (3) to investigate For the selection of Cu treatment levels, diferent concentra- the possibility of using S. indicum and C. tetragonoloba for tions of ­CuCl2 (0, 50, 100, 200, 500, 700, 1000, 1500 and Cu phytoextraction. 2000 mg kg−1) were undertaken in the preliminary screening of the S. indicum and C. tetragonoloba for 20 days. In light of the Cu phytotoxicity, morphological growth and develop- 2 Materials and Methods ment of the plant seedlings, the subsequent concentration levels (0, 25, 50, 100, 150, 200 300 mg kg−1) were fnally 2.1 Seed Procurement and Surface Disinfection selected (Table 1). Process 2.5 Pot Experiment Seeds of two plant species, i.e. Sesamum indicum (sesame) variety TH-6, from the Institute of Oilseeds Research Pro- Plastic pots were flled with 5 kg sieved soil, after which soil gram and Cyamopsis tetragonoloba (guar) variety BR-99 was artifcially spiked with Cu (aqueous solution) using CuCl­ 2 were procured from the Institute of Fodder Research Pro- salt with increasing concentration levels (25, 50, 100, 150, gram, National Agricultural Research Centre (NARC), 200 300 mg Cu kg­ −1) to each pot with three replicates and Islamabad, Pakistan. Seeds were surface cleaned (disin- set aside for 15 days to get stability. The uncontaminated soil fected) with 0.1% ­HgCl2 for 10 min and rinsed seven times without Cu spiking was utilized as control (T0). Experimental with deionized (DI) water, to keep away from any microbial pots were arranged in a complete randomized design (CRD). infection (Pourakbar et al. 2007). Following 15 days of stabilization, soil in pots was mixed well, and 20 surface disinfected (i.e. sterilized) seeds were sown in 2.2 Test Soil Sample Collection each replicate pot. One week (7 days) after seed germination, seedlings were thinned down to fve for each pot. A plastic Soil samples (with sandy-loam texture) were collected plate was placed below the pot for the collection of liquid (lea- from uncontaminated farming felds situated in Jamshoro, chate) that drains out, which was returned back to the pot at Sindh, Pakistan, at depth of 0–15 cm using hand shovel. subsequent watering. Overall investigation was performed in

Table 1 Cu treatment levels Heavy metal Salt used Treatments (mg kg−1 soil) selected for pot experiment T1 T2 T3 T4 T5 T6 T7

Copper (Cu) CuCl2 0 25 50 100 150 200 300

1 3 136 H. Amin et al. a greenhouse for a period of 12 weeks. The phytotoxic efects 2.10 Quality Control and Quality Assurance of metal displayed by plants were apparently noted throughout the test time frame. Plant species were harvested after 90 days All the glassware utilized throughout the present investigation of germination. Likewise, plant morphological (growth) and was comprised of high-quality Pyrex glass material which has biochemical (chlorophyll) parameters were measured. Soil test great resistance to acid. The analytical grade reagents with samples (in triplicate) were also collected for investigation of a certifed purity of 99% and metal standard stock solution Cu by an Atomic Absorption Spectrophotometer (Perkin- (1000 ppm) for AAS analysis were secured from E. Merck Elmer, AAnalyst 800). (Germany). Working standards were prepared by proper dilu- tions of standard stock solutions with double-distilled water. 2.6 Germination Percentage (%) 2.11 Plant Sample Preparation, Digestion and Cu Germination percentage is an estimation of the viability of Determination seeds, calculated as total number of germinated seeds to the total number of seeds sown expressed in percentage (Talebi To determine Cu accumulation in diferent plant tissues (i.e. et al. 2014): root, stem, leaf and pod), harvested plant parts were rinsed Total no. of germinated seeds thoroughly by means of tap water, then with deionized (DI) Germination percentage (%) = × 100. Total no. of seeds sown water to clean adhered components of soil and then oven-dried at 80 °C till steady weight. The oven-dried plant tissues were 2.7 Morphological Parameters ground carefully using an electric grinder and passed through a 1.0-mm mesh strainer. The ground plant tissue samples Plant samples were carefully uprooted from each treatment pot were digested by HNO­ 3 along with HCIO­ 4 mixed at a ratio to quantify morphological (growth) parameters. Plant root and of 3:1 (v/v) according to the protocols devised by Altaf et al. shoot lengths were measured using metric scale. Fresh weights (2017). 0.5 g of plant sample was digested with 12 ml of 3:1 of root and shoot were measured as well with the assistance HNO­ 3/HClO4 di-acid mixture on the hot plate. After cooling, of analytical weight balance. Plant samples were air dried for the digested solution was fltered through Whatman’s flter 1 week. After that plants were oven-dried at 80 °C to attain paper and fnally the volume made up to mark 50 ml by adding a constant weight and at that point their dry weights were deionized (DI) water. recorded. The quantifcation of copper (Cu) in respective tissues was carried out by atomic absorption spectrophotometer (Perkin- 2.8 Chlorophyll Contents Elmer, AAnalyst 800) provided with a copper cathode lamp, under optimum analytical conditions for the estimation of cop- Extraction of chlorophyll content, in completely extended foli- per. The optimum conditions for AAS used throughout these age (leaf) from each replicate pot, was carried out by taking studies are given in Table 2. The standard calibration method 0.5 g of fresh leaf material, ground with 10 ml of 80% acetone. was adopted for the quantifcation of results and triplicate sam- Following fltration, 1 ml of the suspension was diluted with an ples were run to insure the precision of quantitative results. extra 2 ml of acetone, and optical density was determined with The concentration of Cu as well as accumulation in plant root a UV–Visible spectrophotometer (Biochrom Libra S22), using and shoot was calculated according to Monni et al. (2000): two wavelengths (663 and 645 nm) against blank. Chlorophyll AAS interpretation (reading)×dilution factor Cu Conc.(mg∕kg) = , content was evaluated by Arnon (1949). dry wt. of plant tissues (root, stem, leaf, pod)

2.9 Determination of Tolerance Index (TI)

Table 2 Measurement conditions of F-AAS for copper (Cu) determi- Tolerance index (TI) was expressed as the ratio between nation growth parameters (root/shoot length, root/shoot fresh and dry weight) of the plant species in contaminated soil in relation to Parameters Values the growth parameters of plants from non-polluted soil calcu- Wave length (nm) 327.4 lated using the following equation (Wilkins 1978): Hollow cathode lamp current (mA) 5.0

Flame type Air–C2H2 Growth parameter TI (%) = Cu contaminated soil × 100. Background correction On Growth parameterControl soil Slit-width (nm) 1.0 Flame condition Oxidizing Expansion factor 1

1 3 Potential for Phytoextraction of Cu by Sesamum indicum L. and Cyamopsis tetragonoloba L.:… 137

Cu Acc. (μg∕plant) = Cu conc. × dry wt. of plant tissues. 2.14 Statistical Data Analysis

2.12 Soil Sample Preparation, Digestion and Cu All conducted tests were carried out with three replicates Determination and the data were statistically analyzed with PASW­ ® Sta- tistics 18 (SPSS Inc., Chicago, IL, USA). For the compari- Soil samples were air dried at room temperature, ground, son of treatment means, analysis of variance (ANOVA) was mixed well, and kept in plastic (polyethylene) sealed lock performed and to observe the signifcance diference among bags used for subsequent metal analysis. Digestions of soil treatment means Duncan’s multiple range Post Hoc tests samples were done using aqua regia method. To quantify the were applied at a signifcance level of p < 0.05. Cu content in soil, sample of 1 g soil was digested by means of wet acid digestion method through ­HNO3 along with HCl in proportion of 3:1 (v/v) and heated on a hot plate for 2 h 3 Results and Discussion at a temperature of 110 °C until the solution became clear. After cooling, the volume was completed to 50 mL by adding 3.1 Characterization of Tested Soil distilled water. The solution was fltered through Whatman’s flter paper and consequently, examined for Cu contents with The tested soil was sandy loam in texture with an average Atomic Absorption Spectrophotometer. pH value of 6.89 ± 0.04 and electrical conductivity (EC) of 1662 ± 11 µS cm−1. Organic carbon (OC) content of the 2.13 Evaluation of Phytoextraction Efciency soil was 2.20%, while organic matter (OM) of the tested soil was found to be 3.79%. Among soil properties, mobil- To evaluate the phytoextraction potential of S. indicum and C. ity and bioavailability of Cu strongly depend on pH of soil tetragonoloba, the following factors were calculated according and organic matter (OM) contents present in soil (Bravin to Rohan et al. 2013. et al. 2012). Soil pH has direct afect on the solubility of heavy metals together in soil as well as soil solution. Adri- 2.13.1 Bioconcentration Factor (BCF) ano (2001) reported that, both the mobility and Cu bioa- vailability increased with decreased soil pH, while organic The bioconcentration factor was expressed as the ratio of Cu matter (OM) makes available a variety of organic chemical concentration in plant roots in relation to Cu concentration in substances to the soil solution that carried out function as soil medium, calculated as follows: chelating agents (chelates) and enhances mobility and avail- [Cu] root ability of metal to plants (McCauley et al. 2009; Fanrong Bioconcentration factor [BCF]= . [Cu] soil et al. 2011). Therefore, in accordance with soil properties, Cu is more mobile and more bioavailable. 2.13.2 Bioaccumulation Coefcient (BAC) 3.2 Cu‑Induced Phytotoxic Efects The bioaccumulation coefcient was determined as the ratio of Cu concentration in plant shoots to that of Cu concentration Copper is an essential micronutrient at low concentration; in soil medium, calculated as follows: the maximum values for all tested growth and biochemi- cal parameter (chlorophyll content) in two plant species [Cu] shoot −1 Bioaccumulation coefficient [BAC]= . were observed at 25 mg kg , but gradual increase in Cu [Cu] soil concentration signifcantly (p < 0.05) reduced these param- eters. In the current investigation, the germination percent- 2.13.3 Translocation Factor (TF) age of S. indicum and C. tetragonoloba seed was afected signifcantly (p < 0.05) at 300 mg Cu ­kg−1 as compared to The translocation factor was measured as the ratio of Cu con- control (Table 3). The reduced germination percentages (20 centration in plant shoots in relation to Cu concentration in and 75%) were recorded at 300 mg Cu ­kg−1 in S. indicum plant roots, calculated as follows: and C. tetragonoloba, respectively. It has been well docu- [Cu] shoot mented in the literature that germination is a fundamental Translocation factor [TF]= . [Cu] root process in the life a plant species to decide the impacts of Cu toxicity. According to Li et al. (2005) seed is the only stage in the whole plant life well protected against the heavy metal toxicity. The outermost layer of seed acts like a bar- rier that prevents the entrance of substantial heavy metal such as Cu inside the embryo from toxic soil environment

1 3 138 H. Amin et al.

Tables 3 Phytotoxic efects of Cu on growth parameters of Sesamum indicum L. and Cyamopsis tetragonoloba L.

Plant species Cu applied Germination (%) Root length (cm) Shoot length (cm) Root fresh weight Shoot fresh Root dry weight Shoot dry weight (mg ­kg−1) (g ­plant−1) weight (g ­plant−1) (g ­plant−1) (g ­plant−1)

S. indicum 0 90.00a ± 13.23 12.26b ± 0.43 113.80b ± 5.67 11.50b ± 0.61 35.93ab ± 4.32 7.32b ± 0.67 19.67b ± 2.11 25 90.00a ± 0.00 14.17a ± 0.25 125.77a ± 3.86 15.83a ± 0.72 40.78a ± 3.19 9.63a ± 0.48 23.10a ± 1.15 50 90.00a ± 5.00 11.26c ± 0.93 93.96c ± 1.30 10.10c ± 0.18 31.13bc ± 1.21 6.86b ± 1.61 15.44c ± 1.10 100 90.00a ± 5.00 9.30d ± 0.38 87.67d ± 1.39 8.25d ± 0.45 27.27cd ± 2.03 5.11c ± 0.84 11.55d ± 0.58 150 70.00b ± 18.03 8.94d ± 0.26 78.40e ± 2.33 7.33de ± 0.66 25.59cd ± 5.06 4.37c ± 0.47 10.59d ± 1.21 200 25.00c ± 10.00 8.68d ± 0.13 77.69e ± 1.36 6.83ef ± 0.50 21.79de ± 3.72 3.63cd ± 0.57 7.99e ± 1.53 300 20.00c ± 5.00 7.43e ± 0.21 73.11e ± 1.17 5.81f ± 0.88 17.89e ± 1.73 2.81d ± 0.33 6.13e ± 1.12 C. tetragonoloba 0 95.00a ± 5.00 14.07b ± 0.78 117.04b ± 1.20 14.26b ± 0.73 39.11b ± 4.50 8.28b ± 1.05 21.32b ± 1.19 25 95.00a ± 0.00 20.17a ± 1.16 135.13a ± 4.49 16.17a ± 0.67 44.26a ± 2.56 11.03a ± 0.91 25.17a ± 2.10 50 90.00ab ± 5.00 14.05b ± 1.08 111.86b ± 0.36 10.45c ± 0.67 34.18c ± 1.03 10.55a ± 0.77 16.66c ± 1.14 100 90.00ab ± 0.00 13.25b ± 1.00 102.27c ± 6.90 9.07d ± 0.85 29.26d ± 3.73 7.16bc ± 0.71 15.51cd ± 1.14 150 85.00bc ± 5.00 11.03c ± 1.42 96.29cd ± 1.18 8.10de ± 0.91 28.25d ± 1.01 6.04c ± 1.03 13.14de ± 2.76 200 80.00cd ± 5.00 10.02cd ± 0.96 91.78d ± 2.89 7.05ef ± 0.50 27.18d ± 1.00 4.63d ± 0.55 11.40ef ± 1.16 300 75.00d ± 5.00 9.04d ± 0.80 81.08e ± 1.77 6.64f ± 0.15 26.11d ± 0.95 4.10d ± 0.10 9.03f ± 0.83

Similar letters in same column are statistically non-signifcant according to Duncan’s Multiple Range Test (p < 0.05); data are means (n = 3 ± SD) a Signifcantly highest followed by later alphabets for lower means and ensures the protection of embryo from Cu phytotoxicity. growing shoot that results yellowing of leaf, i.e., chlorotic Sesame (S. indicum) and guar (C. tetragonoloba) seeds are symptoms due to mineral nutrients defciency and eventually able to germinate in the presence of low to moderate level leading to stunted plant growth (Muhammad et al. 2015). of Cu concentrations in soil. The consideration of past stud- Cu contamination showed signifcant (p < 0.05) afects ies reported by Ansari et al. (2013) on phytotoxicity of Cu on both fresh and dry weights (biomass) of S. indicum and in seed germination of various plants species signifes the C. tetragonoloba at higher concentration (Table 3). Cu tox- variability of heavy metal tolerance and resistance within icity at 300 mg kg−1 decreased root fresh weight (5.81 and the same and among diferent plant species. 6.64 g ­plant−1) and shoot fresh weight (17.89 and 26.11 g Root and shoot lengths (Seedling’s height) are among plant­ −1) in S. indicum and C. tetragonoloba, respectively. the most important determinants of plant morphologi- The dry biomass follows the same trend as fresh biomass. At cal (growth) parameters. In this study, the root and shoot higher concentration (300 mg kg−1) Cu stress reduced root lengths in terms of growth parameter were signifcantly dry weight (2.81 and 4.10 g ­plant−1) and shoot dry weight (p < 0.05) afected under Cu stress (Table 3). The elevated (6.13 and 9.03 g plant­ −1) in S. indicum and C. tetragonoloba, Cu concentration has direct infuence on plant morphology. respectively. Plant biomass is an excellent indicator for In S. indicum and C. tetragonoloba the longest roots (14.17 describing the growth and developmental changes of plants and 20.17 cm) and shoots (125.77 and 135.13 cm) were within the prospect of heavy metal toxicity. The reduction found in lower treatment at 25 mg Cu ­kg−1, respectively. in plant biomass might be related with disturbed metabolic At 300 mg Cu kg­ −1 treatment, root length decreased by activities because of decreased take-up of fundamental min- 7.43 and 9.04 cm while shoot length reduced by 73.11 and eral nutrients when developed under Cu toxicity (Muham- 81.08 cm in both S. indicum and C. tetragonoloba, respec- mad et al. 2015). Moreover, plants species which can gener- tively. Heavy metal stress is related to a common process of ate high shoot (above-ground) biomass and have the capacity plant growth inhibition or with retarded growth. Elongation to accumulate heavy metals could be utilized for phytoex- of plant root and shoot has demonstrated a notable sensitiv- traction purposes including exclusion of heavy metals from ity to over excess Cu present in soil. Barbosa et al. (2013) contaminated soil. Various examinations demonstrate the have reported that height of maize plant directly (linearly) phytotoxic efects of increased levels of Cu on plant bio- reduced with higher Cu treatments. Increased level of Cu in mass (fresh and dry) cultivated in Cu contaminated soil. the soil decreases root length which directly infuences root Our outcomes for Cu phytotoxicity were obvious from hin- growth and specifc superfcial area, decreasing the absorp- dered growth and development and also reduced fresh and tion capacity of water and nutrients. The shoot length was dry weights that are in consonance with a similar studies on decreased consequently because of Cu hindrance with meta- maize seedlings under Cu stress (Dresler et al. 2014). bolic system of the plants which reduced mineral elements Chlorophyll contents decreased signifcantly (p < 0.05) uptake and increased substantial amount of Cu inside the with steady raise of Cu concentrations from 25 to 300 mg

1 3 Potential for Phytoextraction of Cu by Sesamum indicum L. and Cyamopsis tetragonoloba L.:… 139

Cu ­kg−1 (Fig. 1). In S. indicum and C. tetragonoloba, the Tolerance indices (TIs) were also afected by Cu toxicity. maximum amount of chlorophyll contents were measured Both plant species had diferent tolerance indices (TIs) under at 25 mg Cu kg­ −1, while the lowest concentration of chlo- Cu stress (Table 4). In this study, C. tetragonoloba was more rophyll a (0.07 and 0.09 mg g−1 f.w.), chlorophyll b (0.08 tolerant to Cu stress than S. indicum. At 300 mg kg−1 Cu and 0.06 mg g−1 f.w.) and total chlorophyll (0.15 mg g−1 treatment, S. indicum and C. tetragonoloba had the TIs for f.w.) was at 300 mg Cu ­kg−1, respectively. Cu concentration root lengths (60.66 and 64.49%) and shoot lengths (64.36 in excess amount showed distinctive phytotoxic symptoms and 69.28%), root fresh weights (50.53 and 46.67%) and in foliage (leaves) of various plants species. The reduction shoot fresh weights (50.17 and 67.22%), root dry weights in photosynthetic pigments is likely because of chloroplast (38.63 and 49.97%) and shoot dry weights (31.43 and damage during development phase upon exposure of Cu in 42.29%), respectively. Metal tolerance of plant is a funda- soil system (Ali et al. 2015). Kabata-Pendias and Pendias mental requirement to fnd out the plant-metal interactions (2001) have presented a strong evidence regarding the chlo- prior to exploit for phytoextraction purpose. As reported by rophyll biosynthesis reduction which might be linked with Salt et al. (1998), a plant species utilized for phytoreme- the destruction of photosynthetic organization at thylakoid diation must have high tolerance and metal accumulation level and also the hindrance of Cu with organized system of capacity in their harvested biomass. Therefore, plant toler- chlorophyll (Wodala et al. 2012). ance to substantial metal toxicity is evaluated in accordance with their root or potential shoot development limitations by the metal present in a medium (Ali et al. 2002). Growth and development hindrance might be a typical response of plant to heavy metal stress and is likewise a prominent factor amongst the most imperative agricultural indices for sub- stantial heavy metal stress tolerance (Monni et al. 2000). According to Audet and Charest (2007), if TI values < 1, this indicates that the plant experienced a stress owing to metal contamination with a net reduced in plant biomass. By contrast, TI > 1 suggested that plant species have devel- oped tolerance with a net increase in biomass (hyper-accu- mulator). If TI values equal to 1, the plant is unafected by metal pollution, indicating no diference relative to control treatments.

3.3 Cu Concentration in Plant Tissues

The Cu concentrations among the diferent plant tissues (root, stem leaf, and pod) of both plant species are presented in Table 5. In S. indicum, the maximum concentration of Cu accumulated in the root: 282.08 mg Cu ­kg−1 followed by leaf: 105.78 mg Cu ­kg−1, stem: 65.30 mg Cu ­kg−1, and pod: 8.13 mg Cu ­kg−1 at 300 mg Cu ­kg−1 treatment. How- ever, in C. tetragonoloba Cu accumulated primarily in the root: 158.45 mg Cu kg­ −1 followed by stem: 154.73 mg Cu ­kg−1, leaf: 152.32 mg Cu ­kg−1, and pod: 8.13 mg Cu ­kg−1 at 300 mg Cu kg­ −1 treatments. The elevated Cu contents in the plant tissues (e.g. root, stem, leaf and pod) are noticeably associated with the increasing metal concentration in the soil environment. Studies have demonstrated the take-up of metals; their distribution and translocation to various plant Fig. 1 Efect of Cu stress on photosynthetic pigments chlorophyll-a parts and also the extent of tolerance is reliant on the metal, (a), chlorophyll-b (b) and total chlorophyll (a + b) (c), on S. indicum its bioavailability, the plants species and their metabolic and C. tertragonoloba after 12-week growth in soil medium with systems (Rohan et al. 2013). The Cu accumulation capacity varying concentrations of Cu. Similar letters are statistically non-sig- exceptionally difers among various plants species, based on nifcant according to Duncan’s Multiple Range Test (p < 0.05); data are means (n = 3 ± SD). Superscript a represents signifcantly highest their availabilities in the soil and also infuenced by diferent followed by later alphabets for lower means soil conditions, as reviewed by Muhammad et al. (2015).

1 3 140 H. Amin et al.

Table 4 Efect of Cu stress on the tolerance indices (TIs) of Sesamum indicum L. and Cyamopsis tetragonoloba L. Plant species Cu applied Tolerance indices (mg ­kg−1) Root length (%) Shoot length (%) Root fresh weight Shoot fresh Root dry weight Shoot dry weight (%) weight (%) (%) (%)

S. indicum 25 115.60a ± 2.00 110.73a ± 7.15 137.76a ± 5.92 114.73a ± 18.19 131.99a ± 7.95 118.65a ± 16.99 50 91.95b ± 9.22 82.69b ± 4.03 87.94b ± 4.27 87.26b ± 7.76 95.15b ± 29.85 79.00b ± 8.81 100 75.99c ± 5.74 77.17bc ± 4.15 71.93c ± 6.78 77.07bc ± 14.85 70.67bc ± 17.04 59.25c ± 7.93 150 73.05c ± 4.67 69.01cd ± 4.15 63.64de ± 2.40 72.43bc ± 18.89 59.91cd ± 7.16 54.39c ± 9.84 200 70.91c ± 3.53 68.41cd ± 4.35 59.48ef ± 4.89 61.53bc ± 15.00 49.47cd ± 3.78 41.28cd ± 11.08 300 60.66d ± 3.09 64.36d ± 3.50 50.53f ± 7.86 50.17c ± 6.39 38.63d ± 6.03 31.43d ± 7.05 C. tetragonoloba 25 143.83a ± 15.27 115.45a ± 3.03 113.44a ± 1.46 113.98a ± 12.70 133.97a ± 9.91 118.64a ± 16.51 50 99.83b ± 4.95 95.59b ± 1.29 73.42b ± 5.37 87.95b ± 7.28 128.19a ± 9.10 78.51b ± 9.93 100 94.05bc ± 2.44 87.38c ± 5.86 63.61c ± 5.15 75.01bc ± 8.27 88.11b ± 19.89 72.98bc ± 8.06 150 78.15cd ± 6.06 82.28cd ± 1.86 56.84c ± 5.71 72.98bc ± 9.87 74.64bc ± 22.03 61.79bcd ± 13.59 200 71.55d ± 10.87 78.44d ± 3.24 49.42d ± 2.32 70.16c ± 9.36 56.05c ± 3.68 53.69cd ± 7.57 300 64.49d ± 8.69 69.28e ± 1.56 46.67d ± 2.96 67.22c ± 6.67 49.97c ± 5.19 42.29d ± 1.71

Similar letters in same column are statistically non-signifcant according to Duncan’s Multiple Range Test (p < 0.05); data are means (n = 3 ± SD) a Signifcantly highest followed by later alphabets for lower means

Tables 5 Cu concentration, bioconcentration factor (BCF), bioaccumulation factor (BAC) and translocation factor (TF) of Sesamum indicum L. and Cyamopsis tetragonoloba L. Plant species Cu applied Cu concentration BCF BAC TF (mg ­kg−1) Root (mg ­kg−1) Stem (mg ­kg−1) Leaf (mg ­kg−1) Pod (mg ­kg−1)

S. indicum 25 33.17f ± 2.80 11.09f ± 1.24 9.40f ± 0.94 2.20f ± 0.10 1.33a ± 0.11 0.91a ± 0.09 0.69a ± 0.12 50 60.33e ± 4.04 18.32e ± 0.64 17.37e ± 0.97 3.32e ± 0.30 1.21b ± 0.08 0.78b ± 0.02 0.65a ± 0.03 100 114.40d ± 5.71 31.30d ± 1.47 36.22d ± 6.99 4.35d ± 0.87 1.14bc ± 0.06 0.72bc ± 0.06 0.63a ± 0.07 150 162.33c ± 6.43 41.32c ± 1.51 54.94c ± 2.44 5.53c ± 0.38 1.08bc ± 0.04 0.68c ± 0.01 0.63a ± 0.02 200 210.45b ± 10.32 50.30b ± 1.13 70.56b ± 4.46 7.10b ± 0.36 1.05cd ± 0.05 0.64cd ± 0.02 0.61a ± 0.03 300 282.08a ± 3.88 65.30a ± 0.01 105.78a ± 4.87 8.13a ± 0.06 0.94d ± 0.01 0.60d ± 0.01 0.64a ± 0.00 C. 25 21.43f ± 1.89 18.23e ± 0.95 12.69e ± 1.28 0.91f ± 0.20 0.86a ± 0.08 1.27a ± 0.08 1.49c ± 0.09 tetragonoloba 50 40.45e ± 1.31 33.33d ± 3.20 18.70d ± 0.46 2.00e ± 0.10 0.81ab ± 0.03 1.08b ± 0.07 1.34c ± 0.12 100 74.40d ± 2.71 74.88c ± 4.51 22.29d ± 1.67 3.20d ± 0.95 0.74b ± 0.03 1.00b ± 0.05 1.35c ± 0.11 150 91.32c ± 6.14 74.98c ± 4.27 74.21c ± 3.51 4.17c ± 0.06 0.61c ± 0.04 1.02b ± 0.02 1.69b ± 0.14 200 128.73b ± 3.29 107.33b ± 12.19 104.60b ± 6.66 7.24b ± 0.16 0.64c ± 0.02 1.10b ± 0.05 1.70b ± 0.03 300 158.45a ± 2.68 154.73a ± 10.77 152.32a ± 2.45 8.13a ± 0.06 0.53d ± 0.01 1.05b ± 0.03 1.99a ± 0.02

Similar letters in same column are statistically non-signifcant according to Duncan’s Multiple Range Test (p < 0.05); data are means (n = 3 ± SD) a Signifcantly highest followed by later alphabets for lower means

3.4 Cu Accumulation in Root and Shoot i.e. metal concentration and biomass, for accurate quantity measurements (Vymazal 2016). A considerable increase of Along with concentrations, the overall amount of metals Cu accumulation in root and shoot per plant varied with accumulated in the above-ground biomass (i.e. shoot) is con- respect to Cu concentrations in soil for both plant species sidered as the fundamental parameter to assess the plant’s (Fig. 2). In this study, both S. indicum and C. tetragonoloba capability for phytoextraction (Hanen et al. 2010). Hence, accumulated more Cu contents in shoot than root. Root for the evaluation of accumulation potential, it is important accumulation of Cu in S. indicum and C. tetragonoloba was to consider plant biomass. Consequently, metal accumula- increased from 25 to 300 mg Cu ­kg−1. The highest value for tion in plant biomass most probably relies upon both factors, Cu accumulation in S. indicum and C. tetragonoloba root

1 3 Potential for Phytoextraction of Cu by Sesamum indicum L. and Cyamopsis tetragonoloba L.:… 141

higher than 1, under heavy metal stress, are considered as good phytoextractor whereas, those with bioconcentration factor and translocation factor values lower than 1 are not suitable candidate for phytoextraction. Following the crite- ria, plants species with bioconcentration factor (BCF) val- ues > 1 and translocation factor (TF) values < 1 would likely be suitable for phytostabilization (Mendez and Maier 2008). Between tested plant species, S. indicum had BCF val- ues > 1 from 25 to 200 mg Cu ­kg−1 and BAC and TF val- ues < 1 at all treatments, indicating that S. indicum can be identifed as phytostabilizer and utilized for phytostabiliza- tion of Cu-polluted soil. In contrast, C. tetragonoloba had the BCF values < 1 from 25 to 300 mg Cu ­kg−1, while both bioaccumulation coefcient (BAC) and translocation factor (TF) values > 1 were found at all treatments, indicating that C. tetragonoloba could be a high-efciency plant for Cu translocation from root to the shoot and used as a valuable tool for phytoextraction of Cu from soil.

4 Conclusions

The investigation concluded that no plant species were iden- tifed as metal hyperaccumulator. However, C. tetragonoloba Fig. 2 Accumulation of Cu in root (a) and shoot (b) of S. indicum had considerably higher Cu accumulation than S. indicum and C. tertragonoloba after 12-week growth in soil medium with in light of its better growth and development, efcient toler- varying concentrations of Cu. Similar letters are statistically non-sig- ance and accumulation efciency. Moreover, the notewor- nifcant according to Duncan’s Multiple Range Test (p < 0.05), Data are means (n = 3 ± SD). Superscript a represents signifcantly highest thy estimation of BCFs, BACs and TFs recommend that followed by later alphabets for lower means C. tetragonoloba is a prospective candidate for remediat- ing Cu-polluted soil in quick and successive fushes than S. indicum. In addition, the two plant species have economic (793.48 and 649.75 μg ­plant−1) was found at 300 mg Cu and ecological values. With the use of these plant species in kg­ −1, respectively. Likewise, shoot accumulation of Cu in the remediation of metal polluted soil, the immense positive S. indicum and C. tetragonoloba was also increased from attributes are that the cost is low in contrast with other physi- 25 to 300 mg Cu kg­ −1. The highest value for Cu accumu- ochemical techniques, and can expel contaminations from lation in S. indicum and C. tetragonoloba shoot (1096.98 soil and diminish their development towards groundwater, and 2849.73 μg plant­ −1) was also found at 300 mg Cu ­kg−1, manage the soil properties and may enhance soil quality and respectively. proftability. Furthermore, after harvesting the metal accu- mulated biomass could be burned (incinerated) and reduced 3.5 Phytoextraction Potential for metal recovery and would also be utilized as biofuel. Further study is required to understand the mechanisms of The plant species appropriate for phytoextraction or phyto- Cu absorption in plants. stabilization can be identifed by elucidating the accumula- tion potential and translocation behaviors of heavy metals within plant and soil system. The Phytoextraction efciency for S. indicum and C. tetragonoloba was quantifed by evalu- References ating the bioconcentration factor (BCF), bioaccumulation coefficient (BAC) and translocation factor (TF) values Abidi N, Liyanage S, Auld D, Imel RK, Norman L, Grover K, Angadi (Table 5). According to Fitz and Wenzel (2002), the suitable S, Singla S, Trostle C (2015) Challenges and opportunities for increasing guar production in the United States to support uncon- criteria for plants species used in phytoextraction of metal ventional oil and gas production. In: Uddameri V et al (eds) contaminated soil should have the bioconcentration factor, Hydraulic fracturing impacts and technologies. CRC Press, Boca bioaccumulation coefcient and translocation factor values Raton, pp 207–226

1 3 142 H. Amin et al.

Adriano DC (2001) Trace elements in terrestrial environments: biogeo- Deepak M, Sheweta B, Bhupendar SK (2014) Guar gum: processing, chemistry, bioavailability, and risks of metals, 2nd edn. Springer, properties and food applications—a review. J Food Sci Technol New York 51:409–418 Ahmad M, Ullah K, Khan MA, Ali S, Zafar M, Sultana S (2011) Quan- Dmitry IB, Alexander SL, Naser AA, Ahmad I, Eduarda P (2015) titative and qualitative analysis of sesame oil biodiesel. Energy Evaluation of zinc accumulation, allocation, and tolerance in Zea Sources Part A Recovery Util Environ Ef 33:1239–1249 mays L. seedlings: implication for zinc phytoextraction. Environ Ali NA, Bernal MP, Ater M (2002) Tolerance and bioaccumula- Sci Pollut Res. https​://doi.org/10.1007/s1135​6-015-4698-x tion of copper in Phragmites australis and Zeamays. Plant Soil Dresler S, Hanaka A, Bednarek W, Maksymiec W (2014) Accumula- 239:103–111 tion of low-molecular-weight organic acids in roots and leaf seg- Ali S, Shahbaz M, Shahzad AN, Khan HA, Anees M, Haider MS, ments of Zea mays plants treated with cadmium and copper. Acta Fatima A (2015) Impact of copper toxicity on stone-head cabbage Physiol Plant 36:1565–1575 (Brassica oleracea var. capitata) in hydroponics. PeerJ 3:e1029 Elleuch M, Besbes S, Roiseux O, Blecker C, Attia H (2007) Qual- Amanullah M, Ping W, Amjad A, Mukesh KA, Altaf HL, Quan W, ity characteristics of sesame seeds and by products. Food Chem Ronghua L, Zengqiang Z (2016) Challenges and opportunities 103:641–650 in the phytoremediation of heavy metals contaminated soils: a Fanrong Z, Shafaqat A, Haitao Z, Younan O, Boyin Q, Feibo W, Guop- review. Ecotoxicol Environ Saf 126:111–121 ing Z (2011) The infuence of pH and organic matter content in Altaf HL, Zengqiang Z, Zhanyu G, Amanullah M, Ronghua L, Mukesh paddy soil on heavy metal availability and their uptake by rice KA, Tanveer AS, Farhana K, Ping W, Feng S, Junchao Z, Hui plants. Environ Pollut 159:84–91 H (2017) Potential use of lime combined with additives on (im) Fitz WJ, Wenzel WW (2002) Arsenic transformation in the soil–rhizo- mobilization and phytoavailability of heavy metals from Pb/Zn sphere–plant system, fundamentals and potential application of smelter contaminated soils. Ecotoxicol Environ Saf 145:313–323 phytoremediation. J Biotechnol 99:259–278 Anjum NA, Umar S, Iqbal M (2014) Assessment of cadmium accu- Gawronski SW, Gawronska H (2007) Plant taxonomy for phytoreme- mulation, toxicity, and tolerance in Brassicaceae and Fabaceae diation. In: Advanced Science and Technology for Biological plants—implications for phytoremediation. Environ Sci Pollut Decontamination of Sites Afected by Chemical and Radiological Res 21:10286–10293 Nuclear Agents. Springer, Dordrecht, pp 79–88 Ansari MKA, Oztetik E, Ahmad A, Umar S, Iqbal M, Owens G (2013) Hanen Z, Tahar G, Abelbasset L, Rawdha B, Rim G, Majda M, Souhir Identifcation of the phytoremediation potential of Indian mustard S, Stanley L, Chedly A (2010) Comparative study of Pb-phy- genotypes for copper, evaluated from a hydroponic experiment. toextraction potential in Sesuvium portulacastrum and Brassica Clean Soil Air Water 41:789–796 juncea: tolerance and accumulation. J Hazard Mater 183:609–615 Arnon DI (1949) Copper enzymes in isolated chloroplasts: polyphenol Hazrat A, Ezzat K, Muhammad AS (2013) Phytoremediation of heavy oxidase in Beta vulgaris. Plant Physiol 24:1–15 metals—concepts and applications. Chemosphere 91:869–881 Ashraf MY, Akhtar K, Sarwar G, Ashraf M (2002) Evaluation of arid Hegedus A, Erdei S, Horvath G (2001) Comparative studies of H­ 2O2 and semi arid ecotypes of guar (Cyamopsis tetragonoloba) for detoxifying enzymes in green and greening barley seedings under salinity tolerance. J Arid Environ 52:473–482 cadmium stress. Plant Sci 160:1085–1093 Audet P, Charest C (2007) Heavy metal phytoremediation from a meta- Herawati N, Suzuki S, Hayashi K, Rivai IF, Koyoma H (2000) Cad- analytical perspective. Environ Pollut 147:231–237 mium, copper and zinc levels in rice and soil of Japan, Indonesia Azooz MM, Abou-Elhamd MF, Al-Fredan MA (2012) Biphasic efect and China by soil type. Bull Environ Contam Toxicol 64:33–39 of copper on growth, proline, lipid peroxidation and antioxidant Jiang LY, Yang XE, He ZL (2004) Growth response and phytoextrac- enzyme activities of wheat (‘Triticum aestivum’ cv. Hasaawi) at tion of copper at diferent levels in soils by Elsholtzia splendens. early growing stage. Aust J Crop Sci 6:688–694 Chemosphere 55:1179–1187 Balaji S, Kalaivani T, Rajasekaran C (2014a) Bio sorption of zinc and Kabata-Pendias A, Pendias H (2001) Trace elements in soils and plants, nickel and its efect on growth of diferent Spirulina strains. Clean 3rd edn. CRC Press, Boca Raton Soil Air Water 42(4):507–512. https://doi.org/10.1002/clen.20120​ ​ Li W, Khan MA, Yamaguchi S, Kamya Y (2005) Efects of heavy met- 0340 als on seed germination and early seedling growth of Arabidopsis Balaji S, Kalaivani T, Rajasekaran C, Shalini M, Siva R, Singh RK, thaliana. Plant Growth Regul 46:45–50 Akthar MA (2014b) Arthrospira (Spirulina) species as bio adsor- Mackie KA, Müller T, Kandeler E (2012) Remediation of copper in bents for lead, chromium and cadmium removal—a compara- vineyards—a mini review. Environ Pollut 167:16–26 tive study. Clean Soil Air Water 42(12):1790–1797. https​://doi. Mahmood T, Islam KR (2006) Response of rice seedlings to copper org/10.1002/clen.20130​0478 toxicity and acidity. J Plant Nutri 29:943–957 Balaji S, Kalaivani T, Shalini M, Sankari M, Priya RR, Siva R, McCauley A, Jones C, Jacobsen J (2009) Soil pH and organic matter. Rajasekaran C (2016) Biomass characterisation and phylogenetic Nutrient management modules 8, #4449-8. MontanaState Univer- analysis of microalgae isolated from estuaries: role in phycore- sity Extension Service, Bozeman, Montana, pp 1–12 mediation of tannery efuent. Algal Res 14:92–99. https​://doi. Mendez MO, Maier RM (2008) Phytostabilization of mine tailings in org/10.1016/j.algal​.2015.12.016 arid and semiarid environments—an emerging remediation tech- Barbosa RH, Tabaldi LA, Miyazaki FR, Pilecco M, Kassab SO, Biga- nology. Environ Health Perspect 116:278–283 ton D (2013) Foliar copper uptake by maize plants: efects on Mildvan AS (1970) Metal in enzymes catalysis. In: Boyer DD (ed) The growth and yield. Cienc Rural 43:1561–1568 enzymes, vol 11. Academic Press, London, pp 445–536 Bravin MN, Garnier C, Lenoble V, Gérard F, DudalY Hinsinger P Monni S, Salemma M, Millar N (2000) The tolerance of Empetrum (2012) Root-induced changes in pH and dissolved organic matter nigrum to copper and nickel. Environ Pollut 109:221–229 binding capacity afect copper dynamic speciation in the rhizos- Muhammad A, Shafaqat A, Muhammad R, Muhammad I, Farhat A, phere. Geochim Cosmochim Acta 84:256–268 Mujahid F, Muhammad ZR, Muhammad KI, Saima AB (2015) Dawodu FA, Ayodele OO, Bolanle-Ojo T (2014) Biodiesel production The efect of excess copper on growth and physiology of impor- from Sesamum indicum L. seed oil: an optimization study. Egypt tant food crops: a review. Environ Sci Pollut Res. https​://doi. J Pet 23:191–199 org/10.1007/s1135​6-015-4496-5

1 3 Potential for Phytoextraction of Cu by Sesamum indicum L. and Cyamopsis tetragonoloba L.:… 143

Nagajyoti PC, Lee KD, Sreekanth TVM (2010) Heavy metals, occur- Sharma S, Singh B, Manchanda VK (2014) Phytoremediation: role rence and toxicity for plants: a review. Environ Chem Lett of terrestrial plants and aquatic macrophytes in the remediation 8:199–216 of radionuclides and heavy metal contaminated soil and water. Oh K, Li T, Cheng HY, Xie Y, Yonemochi S (2013) Development of Environ Sci Pollut Res 22:946–962 proftable phytoremediation of contaminated soils with biofuel Singh A, Prasad SM (2014) Remediation of heavy metal contaminated crops. J Environ Prot 4:58–64 ecosystem: an overview on technology advancement. Int J Environ Pourakbar L, Khayami M, Khara J, Farbidina T (2007) Physiological Sci Technol. https​://doi.org/10.1007/s1376​2-014-0542-y efects of copper on some biochemical parameters in Zea mays L. Stadtman ER, Oliver CN (1991) Metal-catalyzed oxidation of proteins. seedlings. Pak J Biol Sci 10:4092–4096 Physiological consequences. J Biol Chem 266:2005–2008 Rachit K, Verma KS, Meena T, Yashveer V, Shreya H (2016) Phy- Subhashini V, Swamy AVVS, Hema Krishna R (2013) Pot experiment: toextraction and bioconcentration of heavy metals by Spinacia to study the uptake of Zinc by diferent plant species in artifcially oleracea grown in paper mill efuent irrigated soil. Nat Environ contaminated soil. World J Environ Eng 1:27–33 Pollut Technol 15:817–824 Talebi S, Nabavi KSM, Sohani DAL (2014) The study efects of heavy Rebecca RC (2006) Copper: inorganic and coordination chemistry, metals on germination characteristics and proline content of Triti- Encyclopedia of inorganic chemistry. Wiley, New York. https​:// cale (Triticoseale Wittmack). Int J Farm Allied Sci 3:1080–1087 doi.org/10.1002/04708​62106​.ia052​ Vymazal J (2016) Concentration is not enough to evaluate accumula- Rohan D, Mayank V, João P, Paul MS (2013) Spatial distribution of tion of heavy metals and nutrients in plants. Sci Total Environ heavy metals in soil and fora associated with the glass industry 544:495–498 in North Central India: implications for phytoremediation. Soil Wilkins DA (1978) The measurement of tolerance to edaphic factors Sediment Contam Int J 22:1–20 by means of root growth. New Phytol 80:623–633 Salt DE, Smith RD, Raskin I (1998) Phytoremediation. Annu Rev Plant Wintz H, Fox T, Vulpe C (2002) Responses of plants to iron, zinc and Phys 49:643–668 copper defciencies. Biochem Soc Trans 30:766–768 Saydut A, Duz MZ, Kaya C, Kafadar AB, Hamamci C (2008) Transes- Wodala B, Eitel G, Gyula TN, Ördög A, Horváth F (2012) Monitor- terifed (Sesamum indicum L.) seed oil as a biodiesel fuel. Biore- ing moderate Cu and Cd toxicity by chlorophyll fuorescence and sour Technol 99:6656–6660 P700 absorbance in pea leaves. Photosynthetica 50:380–386 Seema JP, Promith B, Suman B, Lakshmi B, Namratha (2015) Phytore- Xiuli H, Safyih T, Pin X, Marc JO, Hend AA, Christopher R, Gehong mediation of copper and lead by using sunfower, indian mustard W (2013) Phytoremediation of heavy and transition metals aided and water hyacinth plants. Int J Sci Res (IJSR) 4:2319–7064 by legume-rhizobia symbiosis. Int J Phytoremediat. https​://doi. Shabani N, Sayadi MH (2012) Evaluation of heavy metals accumu- org/10.1080/15226​514.2013.77327​3 lation by two emergent macrophytes from the polluted soil: an experimental study. Environmentalist 32:91–98

1 3 Earth Systems and Environment (2018) 2:455–475 https://doi.org/10.1007/s41748-018-0059-5

ORIGINAL ARTICLE

What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan

Ayesha Qaisrani1 · Muhammad Awais Umar1 · Ghamz E Ali Siyal1 · Kashif Majeed Salik1

Received: 14 March 2018 / Accepted: 29 June 2018 / Published online: 10 July 2018 © The Author(s) 2018

Abstract Rural livelihoods in semi-arid Pakistan are increasingly exposed to climate impacts such as rising temperatures, erratic rainfalls and more intense and frequent climate-related extreme events. This is introducing new risks to the already vulner- able and marginalised societies that lack development and have high poverty rates. This study uses the IPCC Livelihood Vulnerability Index (LVI) approach to analyse the determinants of household livelihood vulnerability defning vulnerability in terms of exposure, sensitivity and adaptive capacity. It also determines various adaptation responses that farmers apply and elucidates the reasons why some farmers choose not to adapt to climate change. It focuses on three semi-arid districts in Pakistan (Faisalabad, D.G. Khan and Mardan) and uses a sample of 150 rural agricultural households. As per the LVI scores, D.G. Khan is the most vulnerable district to climate change impacts, followed by Mardan and Faisalabad, respectively. Results show that (a lack of) adaptive capacity plays quite an important role in shaping households’ livelihood vulnerability for any given degree of exposure and sensitivity. Besides lower exposure and sensitivity to climate change, extremely low levels of adaptive capacity make Mardan more vulnerable to climate change compared to Faisalabad. The paper argues on people-centric development for rural areas through strengthening of agriculture sector as well as providing rural household opportunities for of-farm livelihoods.

Keywords Climate change · Vulnerability · Adaptation · Rural livelihoods · Semi-arid regions

1 Introduction in addition to the characteristics of natural resources such as soil quality, fertility, and water availability (Belliveau et al. Rural areas are characterised by high dependence on agricul- 2006). Climate change and its impacts add another dimen- ture, low human development levels, low adaptive capacities sion by introducing risks that further complicate livelihood and receive little attention from policymakers (Skoufas et al. activities (Salik et al. 2015). These impacts may be pre- 2011; Dasgupta et al. 2014; Panthi et al. 2015). According to sented through either a slow onset of climate/environmen- the IPCC Fifth Assessment Report (Dasgupta et al. 2014), tal changes afecting rural livelihoods,1 or through extreme climate change introduces grave implications for rural areas events such as (amongst others) droughts, foods, and heat through a direct toll on rural livelihoods. Rural economies waves (Dasgupta et al. 2014). in semi-arid regions are primarily natural resource based. The adverse impacts of climate change on farm-based In this regard, rural livelihoods are shaped by complex livelihoods are manifested through shifts in cropping seasons political, economic, institutional and biophysical conditions and a loss in agricultural productivity (IPCC 2014). This not (Abid et al. 2016). This implies that farmers’ livelihoods are only translates into loss of income, but also exacerbates food afected by factors such as government policy on agriculture, insecurity among the rural population. Especially vulnerable the taxing structure, availability of extension programmes, are those households who are engaged in subsistence farm- subsidies, market structures, connectivity and infrastructure ing. Furthermore, climate extremes resulting in widespread destruction have livelihood consequences that may reso- * Ayesha Qaisrani nate long after the event has passed. Through loss of lives, [email protected]; ayesha‑[email protected] 1 Based on agriculture and ecosystems such as (amongst others) the 1 Sustainable Development Policy Institute (SDPI), Islamabad, natural trend of environmental degradation, population pressures and Pakistan land use changes.

Vol.:(0123456789)1 3 456 A. Qaisrani et al. property and livelihoods, people lose their human, physical, Pakistan, owing to its geographical location in arid and natural and fnancial capital that has long-lasting repercus- semi-arid zones,2 is expected to witness high temperatures sions on income and income-generating means (Gray and as impacts of climate change make themselves more evident Mueller 2012; Ngwa et al. 2015). It is estimated that between (Arif 2007; Hussain 2010; Abid et al. 2016). Afzaal and 2000 and 2005 alone, natural disasters led to a loss of about Haroon (2009) postulated in their study that the rate of area- 250 million livelihoods per year globally (Feron 2012, cited weighted annual temperature rise in Pakistan is speeding up. in Ngwa et al. 2015). The threats associated with climate The rate of average annual temperature change was 0.2 °C change impacts may lock the already marginalised rural between 1907 and 1945 per decade, which by 2007 had gone population in poverty traps as their livelihoods are exposed up to 0.53 °C per decade.3 In a comprehensive assessment, to additional risks, thereby perpetuating poverty, deprivation Salik et al. (2015) explain future climate trends in Pakistan and livelihood insecurity (IPCC 2014; Ngwa et al. 2015). based on country-specifc climate scenarios. Using RCP8.54 In view of the enhanced vulnerability of semi-arid regions scenario, it was projected that during 2030–59, average to climate change (IPCC 2014), this paper attempts to iden- annual temperature is expected to rise by 2 °C or slightly tify the key determinants of vulnerability in three semi-arid less in semi-arid regions (with base period 1970–1999), districts of Pakistan using an indicator-based approach. especially in the irrigated plains. Precipitation has the prob- By providing a comparative vulnerability ranking for the ability of declining in the monsoon belt, comprising the selected districts, it also aims to encourage investment in arid and semi-arid regions, however, uncertainty related adaptation strategies that are better suited to the needs of the to precipitation trends is high. Salik et al. (2015) project a population. This paper uses the IPCC-Livelihood Vulner- decrease in average precipitation in the frst half of the year ability Index for developing site-specifc vulnerability scores (up to − 17%) and an increase in the second half of the year that portray the unique aspects that determine districts’ vul- (up to + 12%). As agriculture is largely concentrated in the nerability to climate change. arid and semi-arid regions of Pakistan, these areas are the breadbasket of the country. In that sense, climate risks to 1.1 Rural Economy in Semi‑Arid Pakistan these areas may afect the already discouraging state of food security in the country (Shah et al. 2005). Being primarily an agrarian country, 61% of the population Semi-arid lands in Pakistan predominantly feature irri- of Pakistan lives in the rural areas and over 45% derives its gated agriculture. However, in southern semi-arid districts of income from agriculture (The World Bank 2015). However, Pakistan such as D.G. Khan, agricultural farms are also irri- Pakistan’s economy is transforming from agrarian and rural gated through hill torrents (Dawn 2007). It is speculated that to a more industrial- and service-based economy, leading climate change would render river fows highly unpredict- to a declining share of agriculture in the Gross Domestic able due to increasingly early and rapid melting of glaciers. Product (GDP) (GoP 2016). Despite the staggering agricul- In some years, the increase in early melting of Hindu Kush tural contribution to GDP, its importance cannot be denied Himalaya (HKH) glaciers will be coupled with monsoon in terms of the employment it generates. The rural livelihood period in the country, often leading to increased fooding profle of Pakistan shows that households’ livelihood activi- in the rivers (Hasson et al. 2017). In other years, projec- ties are somewhat diverse, but the prime source of income tions also show that delayed and decreased glacier melting is mostly agriculture (Irfan 2007; Hamid and Afzal 2013). may reduce river fows, thereby unfavourably afecting crop Women’s role in the rural economy is crucial, yet it production (ibid.). goes largely unnoticed because of strong patriarchal norms One fourth of the cultivable area of Pakistan sufers from (Samee et al. 2015). Rural women of all ages actively par- wind and water erosion, salinity, inundation and water log- ticipate in agricultural production, livestock rearing and cot- ging (Irfan 2007). Agricultural productivity, particularly for tage industries. However, much of the work they perform is unpaid or underpaid and mostly ignored in public policy and development planning. Women are involved in farming 2 during various stages such as seed bed preparation, weeding, About 70% of Pakistan’s geographic area can be characterised as semi-arid, receiving annual rainfall of about 250 mm (Alam 2000; and crop harvesting (Ishaq and Memon 2016), yet despite Qaisrani 2015). all their eforts, they are merely considered ‘helpers’ to their 3 The area-weighted anomaly has been calculated using 1961–1990 male family members and get little to no remuneration. as base period. 4 The RCP8.5 combines assumptions about high population and rela- tively slow income growth with modest rates of technological change and energy intensity improvements, leading in the long term to high energy demand and GHG emissions in absence of climate change policies. RCP8.5 thus corresponds to the pathway with the highest greenhouse gas emissions.

1 3 457 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan winter crops such as wheat, often sufers due to heat waves (IPCC 2001; Cutter 2003; Adger et al. 2003; Hahn et al. (Qin et al. 2014; Saeed et al. 2015). A study by the Asian 2009; Shah et al. 2013 etc.). In the context of sustainable Development Bank (2012) observes that subsequent impacts livelihoods, vulnerability assessment is often undertaken of climate change may adversely afect the production of through an indicator-based approach. Adger et al. (2003) cereal crops in Pakistan. With the River Indus being the used an indicator-based approach to understand how various lifeblood of semi-arid regions, alternate periods of droughts factors interact to explain vulnerability signifcantly. As per (1999–2012) and foods (2010 onwards) have immensely the conceptual framework developed by the authors, vulner- increased the vulnerability of livelihoods based on river ability cannot be defned in terms of singular indicators nor fows (Asian Development Bank 2012). The changing pat- it is static; rather diferent factors combine diferently in a terns of monsoon rains increase the risk of fash-fooding in specifc context to determine a system’s vulnerability in a the hill-torrent zones of the country (Hussain 2010). As a dynamic way. result, extreme events have become more frequent and less Hahn et al. (2009) calculated livelihood vulnerability predictable in the recent years (GoP 2013). for two districts of Mozambique through two alternative methods: the Livelihood Vulnerability Index (LVI) and the IPCC–LVI. The authors used primary data collected at the 2 Literature Review household level. Both methods produced similar results related to the vulnerability status of the two districts. This For any adaptation efort to be initiated in response to cli- methodology categorises incidence of natural disasters, cli- mate change impacts, vulnerability assessment is often the mate variability, early warnings and monetary loss due to preliminary step. Fussel (2007) described three approaches climate events under exposure; food, water and health condi- to understand vulnerability of a system: (1) a risk-hazard tions under sensitivity; and socio-demographic, livelihood approach that looks into the risk that a system experiences strategies and social networks under adaptive capacity. The as a result of exposure to a particular hazard; (2) a social study provides a practical approach to identify factors that constructive approach that considers the socioeconomic play a signifcant role in explaining the vulnerability of the dynamics which shape the ability of a system to respond to areas. The ad hoc selection of indicators can be debated, but any shock; and (3) an integrated approach that combines the at the same time, such a context-based choice of indicators earlier two forms to integrate the hazard exposure as well as actually proves to be helpful in capturing the local aspects the socioeconomic capacity to respond to that hazard. These that determine the districts’ vulnerability. classifcations are similar to the work of Turner et al. (2003) Hahn et al. (2009) have ofered a replicable methodol- who classifed vulnerability into three approaches: risk- ogy which has been used by many researchers to determine hazard model, pressure and release model and an expanded vulnerability in diferent contexts (e.g. Pandey and Jha, vulnerability model. Exploring climate change-related vul- 2012; Shah et al. 2013; Panthi et al. 2015, Adu et al. 2017, nerability fnds its relevance in the integrated approach (Fus- etc.). Pandey and Jha (2012) used the approach for two com- sel 2007; Adger 2006) or the expanded vulnerability model munities in rural India in the Lower Himalayan region and (Turner et al. 2003), both of which consider the synergies provided a comparative analysis of the strengths of rural between the human and biophysical systems. mountainous livelihoods. The study found that owing to the A number of vulnerability assessment frameworks have similar ecological settings, the diference in exposure and been utilised over the past decade that aspire to provide a sensitivity between the two regions was ofset, portraying a metrics for quantifying vulnerability. Preston et al. (2011) somewhat similar picture of overall vulnerability. Similarly, gave a critical review of 45 vulnerability mapping studies Shah et al. (2013) employed the Livelihood Vulnerability through the context of (1) goals of assessment, (2) vulner- Index for the case of two rural communities in wetlands of ability framework used, (3) technical methodologies applied Trinidad and Tobago. Apart from providing a community for assessing vulnerability and (4) users, benefciaries and comparison, the study also compared vulnerability scores on participants of the assessment activity. The review showed gender basis and found that female-headed households are that a lack of consensus on the appropriate methodology and more vulnerable as per the index, compared to male-headed framework often leads to choice of methods based on ease of households. Panthi et al. (2015) applied the methodology use rather than the efectiveness of the approach. According for the case of agro-livestock owners in three ecologically to the authors, the objective of conducting the vulnerability diferent districts in Nepal. Besides throwing light on the assessment directs the efciency of the method used. site-specifc determinants of vulnerability, the study found Although methods of vulnerability assessment may often that a lack of income and livelihood diversifcation opportu- be debated, studies focusing on vulnerability agree that nities is the prime causes of high vulnerability in all districts. exposure to risk, sensitivity to damage and the capacity to Applying the same methodology for two communities in recover are essential elements of determining vulnerability Ghana, Adu et al. (2017) identifed that both regions had

1 3 458 A. Qaisrani et al. diferent determinants of vulnerability. While the Wenchi This research applies the IPCC Livelihood Vulnerability community’s high vulnerability was due to high climate Index to study the risks imposed by climate-related extreme variability, high incidence of natural disasters and shortage events on agricultural households in semi-arid regions. The of food and water (i.e. high exposure and sensitivity), low IPCC Livelihood Vulnerability Index ofers a tested tech- adaptive capacity in the Techiman region, primarily due to nique that has the ability to identify the determinants of limited livelihood strategies, was the defning factor of vul- vulnerability and at the same time possessing the capacity nerability for this community. to compare scores among diferent units (Adu et al. 2017). Such index-based analyses have certain precedence over It also allows contextualising the choice of indicators to the other methodologies that rely on secondary data as they are local scenario. The power of providing household level anal- better designed to provide a localised viewpoint, giving con- ysis is also a positive feature of this methodology (Adu et al. text-based insight for local needs and adaptation responses 2017). We take guidance from work done by Hahn et al. required. As vulnerability is a socially constructed subject (2009) and Panthi et al. (2015) and apply it to the specifc (Hinkel 2011), studies based on primary data collection context of semi-arid regions of Pakistan. or those employing a mix of primary and secondary data are better posited to provide insight on socially determined drivers of vulnerability. This paper aspires to identify the 4 Computational Method determinants of livelihood vulnerability in three semi-arid districts of Pakistan and also aims to compare the scores For computing the IPCC-Livelihood Vulnerability Index, we to encourage investments on issue-oriented adaptation have categorised components of socio-demographic profle, strategies. livelihoods, health, social networks, food, water and natural disasters and climate variability into aspects exposure, sen- sitivity and adaptive capacity (Hahn et al. 2009; Adu et al. 3 Methods 2017). The data used in the computation of subcomponents have been measured at diferent scales, so it is essential to The Third Assessment Report of the IPCC (2001) defnes rescale/normalise data before measuring vulnerability index climate change vulnerability as: (Hahn et al. 2009). For this purpose, the min–max normali- sation technique (Patro and Sahu 2015) has been applied The degree to which a system is susceptible to, or using the following formula: unable to cope with adverse efects of climate change, S − S including climate variability and extremes. Vulnerabil- = d min . Indexsd S S ity is a function of the character, magnitude, and rate max − min of climate variation to which a system is exposed, its S S sensitivity, and its adaptive capacity. Here d is defned as the original value of a variable, and min and Smax refect the minimum and maximum values of that In its most simple form, vulnerability is the tendency to variable, respectively. be adversely afected (Kelly and Adger 2000; IPCC 2001; In addition, to determine the internal validity of the iden- Fussel 2007; Opiyo et al. 2014). Since the primary aim of tifed variables, Chronbach’s Alpha Test was applied. With this study is to understand vulnerability as an outcome, the test score of 0.675, it is drawn that the included variables rather than as a factor that shapes an outcome (i.e. risk) are internally consistent.5 The balance weighted average (IPCC 2014), we consider vulnerability of a system within approach was, then, applied (Beccari 2016) which provided three components: equal weighted index for each sub-component. Following the normalisation of variables, method of 1. Exposure to a risk or a hazard. aggregation was employed to fnd out the values of expo- 2. Sensitivity to that risk or hazard. sure, sensitivity and adaptive capacities using the following 3. Capacity to respond to that hazard either by coping, equation: recovering or adapting from the situation (IPCC 2001; n w P Smit and Wandel 2006; Reed et al. 2013). ∑i Pi di CF = =1 . d n w ∑i=1 pi These drivers of vulnerability difer according to geo- graphic location, economic situation, socio-political scenar- ios, psychological conditions, infrastructural development, institutional capacities as well as individual characteristics such as gender, age, health, and education (Zarafshani et al. 5 Co-efcient values close to 0.70 or higher represent internal con- 2016). sistency (Babin and Zikmund 2015).

1 3 459 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan

Here, CF­ d is defned as the contributor factor such as 3. UC 34 Bala Garhi, Mardan. exposure, sensitivity and adaptive capacity for each of the districts. The major components of each of the contributing These UCs were identifed in consultation with subject factor are expressed by Pdi, weights of the components are expert stakeholders as well as by drawing on the knowledge expressed by wpi and n refects the number of major com- of locals. ponent in each of the contributing factors. All individual indicators were equally weighted. Using equal weights for 4.1.1 Faisalabad constructing composite indices is the most popular method (Beccari 2016). Diferent weights for each proxy variable Located in Central Punjab, District Faisalabad has a com- leave room for subjective value judgment as experts and paratively thriving economy compared to the other two researchers may give diferent value to variables (Hudr- study sites. More than half (52.2%) of the district’s popu- likova 2013). A weighted average was obtained for each of lation resides in the rural areas (PBS 2018). Drawing on the major components as per the above equation. Follow- the descriptive facts from the survey, 64% of respondents ing that, IPCC–LVI formula, given below, was applied to in Faisalabad had diversifed incomes. About 66% of the combine the weighted averages to obtain scores for each of households were headed by individuals with at least second- the districts (Deressa et al. 2008). IPCC–LVI ranges from 0 ary level education. Wheat is commonly grown as a win- (least vulnerable) to 1 (most vulnerable). The computations ter crop, cultivated by more than 90% of the respondents. were conducted on the software titled Statistical Package for During the summer season, sugarcane is the most popularly Social Sciences (SPSS). grown crop, often in combination with cotton or fodder. IPCC − LVI = (Exposure + Sensitivity) − Adaptive Capacity. Farmers primarily cultivate winter and summer crops for the sale of products, while saving some portions of wheat for To test the signifcance of diference in the values of indi- household consumption. In 10% households, women were t cators, we applied sample -test to the indicators comprising actively involved in labour force activities. Working women the index for all possible pairs of the three study sites (Adu in Faisalabad district were mostly salaried, engaged in gov- et al. 2017; Salik et al. 2017). Majority of the indicators ernment jobs (such as village school or local health clinic) show signifcance diference among the indicators between as well as private jobs (including work as domestic help). the three sites. Table 5 in “Appendix” shows the p values of the diferences in indicators between the sites. 4.1.2 Khan This study presents a holistic picture of rural agricultural households’ vulnerability to climate change and climate Located in South Punjab, the semi-arid district of D.G. Khan extremes in three semi-arid sites of Pakistan. The research hosts 80.9% of its population in rural areas (PBS 2018). A adopts a case study approach for each site to defne its vul- descriptive analysis from the survey depicts that even though nerability in terms of exposure, sensitivity and adaptive 80% of the respondents rely on more than one source of capacity, while also throwing light on the adaptation strate- income, agriculture is still the lifeline of their livelihoods. gies available to the farmers. Table 1 presents the selection Wheat is the most popular winter crop grown in the district of indicators for these three components and provides their (harvested by 94% of the respondents) which is often cou- criteria for justifcation: pled with sugarcane as a secondary crop. In summers, cot- ton is the most commonly harvested crop (grown by 68% of 4.1 Study Sites and Data Collection respondents) which is often grown in combination with rice and fodder. About 30% of the households surveyed had at With respect to the focus on semi-arid regions in Pakistan, least one female member active in the labour force. Majority the study areas include two districts of Punjab (Faisalabad of these women were involved in agriculture, i.e. managing and D.G. Khan) and one district of Khyber Pakhtunkhwa 6 livestock or helping their male household members in farm- (KP) (Mardan). Figure 1 shows the map of Pakistan with ing activities. the study sites highlighted. Union Councils within the dis- trict were purposely selected that had witnessed climate 4.1.3 Mardan extreme events in the past: Mardan is located in the Khyber Pakhtunkhwa (KP) prov- 1. UC 108 Rattar Chattar, Faisalabad. ince of Pakistan. Majority population (81.5%) resides in 2. UC 28 Kala, D.G Khan. rural areas (PBS 2018). As much as 82% of the house- holds surveyed showed income fows from more than one stream, while the prime occupation remained agriculture. 6 For more details about site selection, refer to Salik et al. (2017). Most households had one or more members working as

1 3 460 A. Qaisrani et al. IPCC. ( 2001 ) Salik et al. ( 2015 ) Salik et al. Porter and Semenov ( 2005 ) Porter and Semenov ( 2015 ) Salik et al. Powell and Reinhard, 2016 and Reinhard, Powell Salik et al. ( 2015 ) Salik et al. References Porter and Semenov ( 2005 ) Porter and Semenov ( 2015 ) Salik et al. Panthi et al. ( 2015 ) et al. Panthi ( 2009 ) Hahn et al. ( 2015 ) Salik et al. Hahn et al. ( 2009 ) Hahn et al. Hutton and Menne ( 2014 ) Hutton Hahn et al. ( 2009 ) Hahn et al. Panthi et al. ( 2015 ) et al. Panthi Salik et al. ( 2015 ) Salik et al. Zachariadis ( 2016 ) to climate change impacts climate change to will be risk of drought/water shortage will be risk which of drought/water enhanced exposure leads to increases theincreases risk yield crop to sure to climate change impacts climate change to sure exposure the risk yield crop to the households sensitivity to climate change impacts climate change to to higher sensitivity to implies higher sensitivity implies diverse fresh water resources for irrigation for resources water fresh diverse Higher the loss, greater will be the exposure Higher the frequency of dryHigher the frequency months higher Increase in theIncrease of precipitation variability - Higher the loss, greater crop will be the expo Higher the frequency, higher will be the Higher the frequency, Functional relationship Functional Increase in temperature variability increases increases variability in temperature Increase More frequent events lead to higher exposure lead to events frequent More This is to assess the occupational diversity of assess theThis is to occupational diversity the skills, higher will be the Less diverse Higher cost implies higher sensitivity implies Higher cost More injuriesMore or death higher exposure refect More disabled persons in theMore household lead Limited access to good-quality drinking water drinking good-quality Limited access to water Higher sensitivity because of lower access to access to Higher sensitivity because of lower Department Department Department Department Household survey Pakistan Meteorological Meteorological Pakistan Pakistan Meteorological Meteorological Pakistan Household survey Pakistan Meteorological Meteorological Pakistan Source of dataSource Pakistan Meteorological Meteorological Pakistan Household survey Household survey Household survey Household survey Household survey Household survey Household survey last extreme event extreme last months during precipita - spring season having tion < 5 mm and in summer precipitation = 0 mm plete crop failure due to last climate extreme last due to failure crop plete monthly temperature above 30 °C above temperature monthly last 10 years (derived from the from total number of (derived 10 years last reported theclimate disasters households in by the 10 years) past skill other than farming for the household for as a result of extreme event (used to assess the (used to event of extreme as a result on human body) disaster of any impact physical bled member to good quality drinking good quality to water to canal water to Average monetary loss in PKR incurred due to Average Number of extreme dry as the months of extreme explained Number Monthly variability in average precipitation in average variability Monthly Percentage of households that com - experienced Percentage Number of extreme hot months with average months hot with average of extreme Number Explanation Monthly variability in average temperature in average variability Monthly Average number of climate extreme events in the events number of climate extreme Average Percentage of households that do not have any any of households thathave do not Percentage Average health-related cost per (in PKR) month health-related cost Average Percentage of households with an injuryPercentage or death Percentage of households with a physically disa - of households with a physically Percentage Percentage of households that do not have access of households thathave do not Percentage Percentage of households that do not have access of households thathave do not Percentage Indicators for IPCC Livelihood Vulnerability Index Vulnerability IPCC Livelihood for Indicators body Monetary loss Extreme dry Extreme months Physical damage to human to damage Physical Precipitation variability Hot months Hot Temperature variability Temperature Climate extremes/disaster Professional skills Professional Cost on health Cost issues Physical disability Physical Access to water to Access Source of irrigation Source Crop failure Crop 1 Table and subcomponents Indicators Exposure Sensitivity

1 3 461 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan ( 2015 ). Cutter et al. ( 2008 ) Cutter et al. Hahn et al. ( 2009 ) Hahn et al. Panthi et al. ( 2015 ) et al. Panthi Brooks and Adger ( 2006 ) and Adger Brooks Panthi et al. ( 2015 ) et al. Panthi References Hahn et al. ( 2009 ) Hahn et al. ( 2016 ) et al. Zarafshani Connolly-Boutin and Smit Connolly-Boutin Barr ( 2010 ) et al. ( 2015 ) et al. Panthi household to adjust more quickly to any any to quickly more adjust household to and stressor shock by providing support in the form of physical support of physical in the providing form by sharing, and experience help, information etc. awareness of hazards and preparedness and preparedness of hazards awareness contributes adaptive to ultimately which capacity extreme events that leads to lower sensitivity that lower leads to events extreme capacity capacity will be the adaptive capacity will be the adaptive be the adaptive capacity be the adaptive More savings enhance the capacity of the savings More Social networking enhances adaptive capacity enhances adaptive Social networking Communication media helps to enhance Communication media helps to Houses made of concrete are more resistant to to resistant more are Houses made of concrete Income diversifcation increases adaptive adaptive increases Income diversifcation Functional relationship Functional More crop diversity, higher will be adaptive higher will be adaptive diversity, crop More More education of household heads, higher More Higher the age dependency ratio, lower will lower dependency ratio, Higher the age Household survey Household survey Household survey Household survey Household survey Source of dataSource Household survey Household survey Household survey any kind any involved in local politics, social relief work, work, in local politics, social relief involved social a strong associations and have local level circle internet and TV of temporary material source of income source households + 1), e.g.: a household that grows households + 1), e.g.: a household that grows crop cane, rice,sugar wheat and maize will have = 1/(4 + 1) 0.20 index diversity ary and above level of education level ary and above of 15–64 Percentage of households who have savings of savings of households who have Percentage Percentage of households who are actively actively of households who are Percentage Percentage of households having access to phone, access to of households having Percentage Percentage of households that have a house made of households that have Percentage Percentage of households who have more than more one of households who have Percentage Explanation The inverse of (the number of crops grown by the by of (the grown number of crops The inverse Percentage of household heads who have second - of household heads who have Percentage Ratio of the population above or below the of age the of age or below of theRatio population above (range 0–1) (range Savings Social networking Average crop diversity index index diversity crop Average Access to media/information to Access Type of house Type Income diversifcation Education Dependency ratio 1 Table (continued) and subcomponents Indicators Adaptive capacity Adaptive

1 3 462 A. Qaisrani et al.

daily wage labourer in the non-farm sector. Only 18% of the household heads had education up to secondary school level or higher. Like the other two districts, the most commonly grown winter crop is wheat, while in summer cultivating maize is popular, sometimes in combination with rice. The main purpose of farming both summer and winter crops is for household consumption. Although female members of households in Mardan were better educated than women References Hahn et al. ( 2009 ) Hahn et al. Schefran et al. ( 2011 ) et al. Schefran Adger et al. ( 2003 ) et al. Adger in D.G. Khan, none of them were actively participating in labour force activities. The household survey intended to refect on farmers’ perceptions and experiences related to climate change, agricultural risks they are exposed to, coping mechanisms, livelihood strategies and adaptation options being used. Data were collected during December 2016 and January 2017. The research sample constituted agricultural households, i.e. those households whose primary source of income is from the agriculture sector. In each district, 50 rural house- holds were interviewed, regardless of their land ownership status. The respondents were household heads chosen ran- domly among the agricultural households in the villages and help in improving adaptive capacity adaptive help in improving tive capacity of farming households capacity of farming tive as a source of income which contributes of income which as a source to capacity adaptive Functional relationship Functional Awareness of and access to these of and access to services Awareness Investments in agriculture the will raise adap - Investments Commercial agricultureCommercial agriculture refects consent was taken from them by giving a brief introduction regarding the purpose of the survey. As the target respond- ents were household heads, the survey team mostly came across male respondents. However, females were involved through qualitative in-depth interviews to understand their lifestyle in the villages. Secondary data were also used for estimating climate indicators. The study used average monthly data for tem- Source of dataSource Household survey Household survey Household survey perature and precipitation for the time period 1961–2014 obtained from Pakistan Meteorological Department (PMD). Faisalabad and D.G. Khan districts have meteorological sta- tions. However, due to the absence of any meteorological station in Mardan, data from the nearest meteorological sta- tion of Risalpur were used, which is at a distance of 17 km from Mardan. Key Informant Interviews (KIIs) were conducted with experts and government ofcials in district development and agriculture departments. The purpose of conducting these KIIs was to develop an understanding of the depart- ments’ initiatives in addressing climate change, agriculture and migration concerns, as well as the challenges they face. or have access to government-initiated climate government-initiated access to or have plans management disaster remittances on agriculture production for sale of product for production In total, nine KIIs were conducted. Explanation Percentage of households have awareness about awareness of households have Percentage Percentage of households willing to invest their invest of households willing to Percentage Percentage of households that use agriculturePercentage

5 Results

5.1 Livelihood Vulnerability Index

Results of IPCC–LVI show varying levels of vulnerability experienced by farmers in the three study sites (Table 2). management plans management ment tion Figure 2 shows the vulnerability triangle for the three Government-initiated disaster disaster Government-initiated - invest Use of remittances for Purpose of agriculture- produc 1 Table (continued) and subcomponents Indicators study sites. Detailed results of the subcomponents of the

1 3 463 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan

Fig. 1 Map of Pakistan with study sites. Source: Authors’ own

Table 2 IPCC–LVI scores for Faisalabad, D.G. Khan and Mardan. IPCC–LVI for each of the study site are given in Table 4 in Source: Authors’ own Appendix. The Index entails components refecting the farm- IPCC–LVI Faisalabad D.G. Khan Mardan ers’ exposure to climate risks, their sensitivity to climate impacts and their levels of adaptive capacity that refects Exposure 0.336 0.406 0.251 their abilities to cope with climate shocks. Sensitivity 0.382 0.412 0.274 A composite score shows that D.G. Khan has the high- Adaptive Capacity 0.699 0.504 0.467 est livelihood vulnerability (0.314), followed by Mardan IPCC–LVI 0.019 0.314 0.058 (0.058). Faisalabad is the least vulnerable of the three dis- tricts with a score of 0.019. D.G. Khan shows the highest levels of exposure and sensitivity, and is better than Mardan Exposure in terms of adaptive capacity. Despite being more exposed 0.7 0.6 and sensitive to climate change than Mardan, Faisalabad 0.5 has comparatively higher scores of adaptive capacity that 0.4 decreases its overall livelihood vulnerability. 0.3 0.2 The prime factor leading to D.G. Khan’s higher sensitiv- Faisalabad 0.1 ity is the highest incidence of complete crop failure. All D.G Khan 0 farmers who participated in the survey lost their standing Mardan crops during the foods of 2010 and 2011. In comparison, Adapve Sensivity in both Faisalabad and Mardan, 70% of farmers experienced capacity complete crop failure. D.G. Khan’s higher exposure is also a result of higher number of extreme hot months that it experi- ences compared to the other two districts. The district has Fig. 2 Vulnerability triangle based on IPCC–LVI. Source: Authors’ endured longer and more frequent dry spells which add to own its higher exposure. The Index shows that loss in monetary

1 3 464 A. Qaisrani et al. terms was also highest in D.G. Khan as a result of climate Observed changes in the climate extreme events. Higher exposure of both, Faisalabad and 30% D.G. Khan, is also explained by the higher level of variabil- 25% ity in precipitation and temperature that may refect more 20% intense impacts of climate change in these areas. 15% Similarly, D.G. Khan has the highest score for sensitivity 10% (0.412) to climate change impacts, followed by Faisalabad 5% Faisalabad 0% D.G Khan (0.382) and then Mardan (0.274). More people in D.G. Khan Mardan (96%) do not have access to canal water for irrigating their agricultural lands compared to Faisalabad and Mardan. This percentage is generally low in Mardan, with only 16% farm- ers without access to canal water, while the remaining 84% irrigate their crops through a canal. The type of housing structure also shapes the sensitivity of households to climate impacts. In D.G. Khan, about 50% households reside in mud Fig. 3 Farmers’ perceptions on climate change. Source: Author’s own or thatch houses, leading to their greater susceptibility to destruction during extreme events. Faisalabad ranks second in terms of sensitivity owing to a high percentage of people D.G Khan who do not have access to safe drinking water and fewer 30 people with diversifed professional skills, i.e. skills other than farming. Proper access to canal water lowers Mardan’s 28 sensitivity to climate change and related events. 26 In terms of adaptive capacity, Faisalabad (0.699) scores 24 far better than the other two districts (0.504 for D.G. Khan and 0.467 for Mardan). Faisalabad’s higher adaptive capac- 22 ity is primarily due to more educated household heads. 20 Additionally, Faisalabad also boasts better connectivity as more than 70% households have access to information and 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 communication technology. The district also demonstrates Temperature higher saving patterns with more than 80% households having savings of some kind compared to D.G. Khan and Fig. 4 Average annual temperature over time in D.G. Khan (1961– Mardan where approximately 60% households have savings. 2012) Mardan has the lowest adaptive capacity among the three districts. The two main features that shape this low adap- Faisalabad tive capacity are the strikingly low level of awareness about 30 government-initiated plans for disaster risk management and 28 low reliance on agriculture for income-generating purposes. 26 5.2 Farmers’ Vulnerability Experiences 24 22 This study inquired about farmers’ perceptions regarding 20

climate change and climate variability. This refects their 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 understanding of the changes that are taking place and the Temperature expected impacts these changes might have on their liveli- hoods. Figure 3 shows the common changes that farmers Fig. 5 Average annual temperature over time in Faisalabad (1961– have observed in the climate of their respective locations. A 2012) common observation in all three locations was that tempera- tures are rising and rainfalls are declining. In Punjab districts (Faisalabad and D.G. Khan), a considerable percentage of years 1961–2012 (Figs. 4, 5, 6). While the graphs do not farmers (approximately 21%) perceived changes in the num- refect clear upward trends in the temperature, wide fuctua- ber of hot and cold days. tions in the average annual temperature are observed for all Average annual temperature trends drawn from station three study sites. Longer term trends may be better able to data from the respective study sites are available for the demonstrate changes in annual temperatures. IPCC (2014)

1 3 465 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan

Mardan to food water. About 77% farmers also lost their stock of 29 food and fodder as a result of foods. The last food in D.G. 27 25 Khan inficted losses of over PKR 400,000 (USD 4000) to 23 36% of the respondents. 21 Torrential rains in Mardan have led to fash and river- 19 ine foods afecting the district severely. Hailstorms in the 17 15 district have also led to heavy losses for the farmers. About 83% farmers faced complete crop failure due to foods, and 1961 1964 1967 1970 1973 1976 1979 1982 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 63% farmers were severely afected due to disease or death Temperature of livestock. Rising temperatures were regarded as one of the important factors leading to livestock diseases. As a result Fig. 6 Average annual temperature over time in Mardan (1961– of supply shortage and difculty in access, 63% complained 2012). Source: Pakistan Meteorological Department that food stocks in the village markets declined. Heavy rains also damaged the stock of food and fodder for about 67% projections for semi-arid areas also predict temperature rises of households. These statistics indicate the inter-linkages over the next century. between climate extreme events and potential rise in food The most common source of climate information among insecurity. Furthermore, 60% of the respondents incurred the farmers varied in all three regions. In Faisalabad, the monetary losses under PKR 50,000 (USD 500), while 30% most common source was media through which 43% peo- had to bear losses from PKR 50,0000 to PKR 100,000 (USD ple obtained information related to climate change. This 500–1000). included information such as rising temperatures, rainfall predictions and warnings related to climate extremes. In 5.3 Adaptation Strategies for Farmers D.G. Khan, 40% people relied on their social networks for climate information, while in Mardan, 66% people perceived With the rise of climate change risks, farmers have devel- changes in the environment through their own personal oped diverse measures to reduce their livelihood risks and experiences. This refects that while households in all three vulnerabilities. In case of sudden extreme events, farmers districts are aware of changes in the climate, the quality and use certain coping strategies to manage the losses and dam- accuracy of such information may vary depending upon the age faced. As per the survey, borrowing money from friends source of information. and family is the most common strategy used in all three In Faisalabad, shortage of water due to infrequent rains sites in the wake of any kind of disaster. In addition, other emerged as the most severe climate-related issue, which is factors such as socioeconomic status, education, and insti- also closely tied to intense heat waves over the past 10 years. tutional support also shape the choice of a coping measure Resultantly, as many as 80% respondents reported a decline (Coulibaly et al. 2015). in crop productivity and in some cases complete crop fail- In D.G. Khan, 88% of respondents claimed being dis- ure. Spread of contaminated water was also cited as a major placed from their settlements for some time as a result of issue, and 44% farmers reported that they experienced dis- a climate extreme event (mostly foods), refecting their ease or death of livestock, 53% experienced loss of working high susceptibility to such events. As an emergency coping hours, and thereby, decline in income following a climate mechanism, many farmers (26%) tend to sell their livestock extreme event. During the last climate event, about 40% for getting ready-cash to fulfl their immediate needs. About farmers incurred fnancial losses of less than PKR 50,000 11.2% of farmers reported that in such hard times, they rely (USD 500), and about 22% sufered from losses between on government-provided support in the shape of relief initia- PKR 50,000 and 100,000 (USD 500–1000). Approximately tives to sustain themselves. Furthermore, to compensate for 16% farmers even bore monetary losses amounting PKR the loss in agricultural production, farmers reported increas- 400,000 (USD 4000) and above. ing their use of inputs such as fertilisers and pesticides (51%) In D.G. Khan, foods, including fash-foods and riverine and diversifying their crop production (28%). About 11% foods, were ranked as extreme events with the most severe also reported increasing the use of labour and technologi- impacts on farmers’ livelihoods. About 96% people reported cal input in eforts to increase production. In fact, it was complete crop failure as the severest climate impact that observed that in about 4% households, women were more makes them vulnerable. Similarly, 60% experienced decline involved in labour activities after the massive foods of 2010. in crop productivity as a result of extreme heat, water short- In Faisalabad, the most common response to cope with age and climate-related disasters. Health hazards and inju- climate extremes (after borrowing money from family and ries, and infrastructural damages were also common issues friends) is selling livestock (16%) and other assets such as faced by 60% and more than 80% farmers, respectively, due jewellery and motorbike (10%). Many farmers complained

1 3 466 A. Qaisrani et al. that in such difcult times, they have to sell their livestock About 48% expressed their interest in shifting to another at cheaper rates to have enough to feed themselves. On the livelihood source with high preferences for attaining a sala- other hand, to manage the losses being incurred due to the ried job or starting their own small business (that includes slow onset climate changes—rising temperatures, erratic establishing a small shop or providing transport services). rainfall, declining soil fertility—farmers tend to intensify In comparison with the other locations, the percentage their use of inputs such as fertilisers and pesticides (42.7%). of farmers who do not make any eforts to adapt to climate Diversifying crop production is the second most common change was highest in Mardan (14%). Approximately 69% measure, while 18% farmers reported that members of their of the farmers who do not opt for adaptive measures also households have migrated to other locations as a means responded that they do not fnd it useful to take up adapta- of diversifying their income sources. About 14% farmers tion measures. Most of them were of the view that adap- reported introducing an alternative mode of irrigation to tation strategies are not efective when a disaster hits and compensate for water shortage. if they invest in adaptation, then they have to bear greater Similar to the other two districts, farmers in Mardan who losses. In response to the scenario if adverse climate impacts have experienced climate extreme events cope with disas- continue to increase, 64% farmers reported that they would trous impacts by borrowing money from their friends and prefer to shift away from agriculture to other means of earn- relatives (39.2%), selling livestock (17.6%) and other assets ing incomes. Most farmers voiced their preference of invest- such as motorbike, jewellery and property (9.5%). About ing in their own small start-up if they have to move away 12.2% also reported reducing their consumption levels in from agriculture. response to the difcult situation. With regards to long-term Government-provided facilities were reported more in adaptation strategies, increasing the use of inputs was the districts of Punjab, whereas in Mardan, people reported that most common (40.6%), followed by diversifcation of crop there were no government-initiated facilities available for production (29.7%), similar to the trend in other sites. About climate adaptation. In Faisalabad, 35% people reported that 19% farmers also reported that as agriculture becomes less they had access to government-provided early warning sys- proftable due to climate change impacts, they migrate from tems and 23% reported that they received trainings on how the villages in search of economic opportunities elsewhere to adapt to changing climatic scenario. In D.G. Khan, 37% for diversifying their livelihoods, while 12.5% reported shift- farmers stated that they had received some subsidy, credit ing to other sources of incomes while staying in the villages. or insurance from the government owing to the district’s In D.G. Khan about 6.1% farmers in the survey reported repeated exposure to foods. About 35% also reported that that they do not use any adaptation measure against adverse the government has provided them with new crop varieties climate impacts and 50% shared that they do not see the need that are more resistant to climate change impacts. Table 3 to adapt. They explained that even if they use any adaptive provides an overview of the government facilities provided measures, they all go to waste when foods hit and wash in the study sites based on KIIs with the Agriculture Exten- away all their crops, damage their settlements and result sion Departments of the respective districts. in death of family members and livestock. Given the situa- To summarise, it has been found that in semi-arid regions tion that climate impacts continue to rise in frequency and of Pakistan, the most common methods of adapting agricul- intensity, only 28% farmers responded that they would con- tural livelihoods to climate impacts are intensifying the use sider shifting away from agriculture to another source of of agricultural inputs such as pesticides and fertilisers, and income. On investigating further, it was found that farmers diversifying crop varieties. These results are consistent with in D.G. Khan believe that they have no means to shift to the earlier fndings of Khan et al. (2011). In addition, migra- another livelihood source and agriculture is the only activity tion is a common adaptation response to climate impacts, yet that they are familiar with. Nevertheless, those who would its incidence varies for diferent locations. Climate extremes choose to switch to another livelihood reported they would over time generally tend to weaken income generation from work as daily wage labour. agriculture, in response to which many households decide In Faisalabad, 6.7% farmers revealed that they are not to send a family member to the city to seek economic oppor- doing anything to adapt to climate change impacts for their tunities (Salik et al. 2015). In contrast, high costs of adapta- agricultural livelihoods. Amongst them, 57% reported that tion and lack of information on how best to adapt to certain they do not have the required skills or knowledge about the climate impacts came out as the main impediments for those kind of strategies to use, while 29% responded that they can- farmers who do not take any adaptive measures for their not aford to use adaptive measures as they are expensive. livelihoods (Ali and Erenstein 2017).

1 3 467 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan

Table 3 Salient features of agricultural support provided by District Ofce of Agricultural Extension. Source: Authors’ own based on KIIs Faisalabad D.G. Khan Mardan

Training of farmers on cropping practices and Provision of new seed varieties to major farm- Awareness raising through publication of usage of pesticides (after lab testing) ers (large landholders are willing to take agricultural newsletter (Zarat Nama), risk. If new varieties give good results, small electronic messages, trainings, seminars, and farmers are also encouraged to try new seeds) feld work Provision of new seed varieties to 191 villages Soil sampling of land to advise what to grow Provision of door-to-door consultancies to in the district due to changing soil quality farmers regarding crop productivity, fertilis- ers, price/marketing schemes Provision of farm implements/agricultural Implementation of Punjab Kisaan Package Provision of new wheat seeds at subsidised tools (4-5 per UC) (interest-free loan scheme) rates Provision of 4000 vegetable toolkits for Registration of farmers for E-credit that Provision of early warning systems (However, kitchen gardening with a subsidy of PKR 50 provides small farmers (with up to 5 acres of the agriculture ofcer reported a lack of per packet land) with loan of up to PKR 25,000 per acre. capacity to provide trainings on food adap- tation mechanisms)

6 Discussion seeds, and soil fertility are depleted, thus amplifying their vulnerabilities. In addition, as small farmers mainly fulfl 6.1 Climate Change, Extreme Events their household food requirements from farms, loss of crops and Agricultural Vulnerability may worsen food insecurity (Wilkinson and Peters 2015). On a macro-scale, increasing incidents of crop fail- As defned in the methodology section, livelihood vulner- ures resulting from climate extremes can have detrimental ability to climate change is not merely a function of exposure impacts on national food and fbre production and afect the to climate impacts. It is defned by a number of pre-existing whole agricultural value chain. Considering that Pakistan’s factors—poverty levels, education, awareness, and liveli- industrial sector is largely agro-based, this may even result hood diversifcation—all of which shape the sensitivity and in setbacks for the industrial sector (Wilkinson and Peters adaptive capacity of a household (Abid et al. 2016). As con- 2015; FAO 2015). Recurring foods and prolonged droughts cluded by Abid et al. (2016), this research refects that each not only depreciate peoples’ livelihood-earning capabilities, study site has varying associated factors that determine the they also hamper the government’s development eforts and level of livelihood vulnerability experienced by farmers. A poverty reduction initiatives (Looney 2012). This may push striking fnding is that rural livelihood vulnerability can be vulnerable communities into perpetual states of deprivation. caused by a lack of adaptive capacity, despite its exposure Results indicate that sensitivity of agricultural livelihoods and sensitivity to climate change impacts being lower. This is highly dependent upon the source of water for irrigation, is refected by Mardan’s vulnerability score that is higher making those who do not have reliable irrigation sources than Faisalabad district despite its exposure and sensitiv- more vulnerable (Salik et al. 2015; Zachariadis 2016). This ity being lower than both D.G. Khan and Faisalabad. This is particularly important when the agricultural dependency implies that high exposure and sensitivity to climate risks on water is high but infrastructure is either poor or inacces- can be ofset by a stronger adaptive capacity, rendering the sible to marginalised farmers (Panthi et al. 2015). In terms of population better positioned to counterweigh the negative sensitivity of livelihoods to climate change impacts, human efects (Hahn et al. 2009; Adu et al. 2017). capital plays a fundamental role, especially if taken in terms According to the fndings, the most pronounced impact of of health. Health issues decrease the working capacities of climate disasters that defnes the exposure of livelihoods is labour and result in reduced number of working days (Hahn complete crop failure. As observed in the study sites, climate et al. 2009). Respondents (males and females) in all three extreme events such as hydrological disasters (rain storm, sites were dissatisfed with the health facilities available at foods and drought-like conditions) and heat waves have the village level. They reported that the dispensaries and/ caused instability in agriculture-based livelihoods through or the basic health units available in the villages were not complete crop failure. Floods and torrential rains wash equipped with sufcient medicines. In addition, the avail- away the standing crop, while intense heat impacts the rate ability of doctors in the facilities was also irregular, which of evapotranspiration and soil moisture, often resulting in often forced them to travel long distances to the next towns decline in crop production (Rasul et al. 2012). This is par- or district capital to consult a doctor or purchase medicines. ticularly alarming as the increasing magnitude and intensity While this issue was raised by both men and women, the lat- of such events erode the livelihoods of small landholders ter seemed to be more concerned about the defcit in health (Rasul et al. 2012). Their agricultural assets such as crops, provisions in their areas. This refects concern for their

1 3 468 A. Qaisrani et al. children and family members sufering from diseases that community networks) faster than those whose social ties are can be prevented if health facilities are adequately available. weak (Adger et al. 2003). D.G. Khan, with the lowest score Sensitivity to climate vulnerability is also defned by the in the social network indicator, has lower adaptive capac- skill diversity of the population (Hahn et al. 2009; Shah et al. ity compared to Faisalabad. This indicates that community 2013). This study assumed that households that possess members are less integrated in terms of reliance on each skills and capacities in addition to farming are less vulner- other for support during times of need (Adger 2001). able to climate change impacts as they can easily use these It was assumed that higher rates of female labour force additional skills to be absorbed in another profession in case participation would refect less vulnerability (Muthoni and climate change renders farming more and more susceptible Wangui 2013). However, the impacts on vulnerability are (Hahn et al. 2009). This is also refected in the study wherein structured by the type of labour force activity that women many households in all three sites were concerned about not are involved in. For instance, by comparing the cases of having any skill other than farming in case they decide to Faisalabad and D.G. Khan, it can be concluded that higher migrate to the cities. labour force participation of women in agricultural activi- As mentioned, higher adaptive capacity of a population ties in D.G. Khan actually enhances the overall vulnerabil- may balance out the high exposure and sensitivity to climate ity of the community due to higher dependence on natural change. Factors that contribute to higher adaptive capacity resources (Samee et al. 2015; Batool and Saeed 2017). On included education of household head, access to information the other hand, involvement in activities that are not too nat- and communication technologies and strong social networks ural resource intensive reduces their vulnerability. Women (Hahn et al. 2009; Connolly-Boutin and Smit 2015; Panthi in D.G. Khan were generally more involved in livestock et al. 2015). Higher education level refects higher ability to management, crop sowing, weeding and harvesting, water receive, interpret and comprehend risks and hazards, thus, management, collecting wood and biofuel. In Faisalabad, it positively impacts adaptive capacity (Byrne 2014). As the women were more involved in service sector and in educa- household head is responsible for making the fnal house- tion and health sectors or serving as domestic helpers in hold level decisions, his/her education level determines the private homes. capacity of the household to anticipate risks and respond to While labour force participation is higher, women gen- them through changes in livelihood strategies (Panthi et al. erally are less involved in decision-making processes and 2015). have limited access to information and facilities (Samee The population of Faisalabad had more access to informa- et al. 2015). Education levels of women in D.G. Khan are tion and communication technologies than D.G. Khan and also lower than those in the other two districts. In contrast, Mardan, refecting its higher ability to obtain information though fewer women were economically active in Faisal- quicker in case of any climatic or non-climatic risks. Greater abad, they were more educated and involved in government dissemination of weather and climate news in Faisalabad or private sector jobs. This refects weaker impact of cli- could be because of the presence of the Regional Agrome- mate change on their livelihood activities, thus, their lesser teorological Centre in the city, along with the Ayub Agricul- vulnerability. In Mardan, despite higher education levels ture Research Centre7 in the city. The proximity of source of women compared to D.G. Khan, women did not partake of information and the greater integration of communication in any income-generating activities given the culture and technologies refect higher adaptive capacity in Faisalabad. norms of the village. Similarly, social capital is an important pillar in defning adaptive capacity of a household (Salik et al. 2015; Qais- 6.2 Adapting Agriculture to Climate Change rani 2015). Kinship, relationships and strong community and Climate Extremes networks can act as a support system at times of distress and can also refect stronger fow of information and news In light of the above vulnerability scenario, identifcation (Schefran et al. 2011). Strength of social network is also of the most suitable adaptation measures is pivotal to deter- useful in climate information dissemination, as is refected mine the pathways to resilient economic development in by statistics from D.G. Khan where majority of people rely semi-arid regions. This research fnds that most farmers are on disaster or climate updates from their networks through already adjusting their livelihoods to climate impacts. These a phone call or message. People with strong social networks adjustments include short-term, distress responses known have been observed to survive calamities and rebuild their as coping strategies as well as long-term modifcations in lives with the help social capital (family ties, relatives, and their livelihood activities that can reduce their risk to climate impacts in the future (Khan et al. 2011; Ullah et al. 2018). In the wake of a climate shock, farmers tend to adopt 7 Pakistan Meteorological Department http://namc.pmd.gov.pk/ramc- distress measures such as borrowing from friends and fam- faisa​labad​.php. ily, selling livestock and other assets or reducing household

1 3 469 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan consumption. These measures are generally meant to bufer attaining a salaried job, doing non-farm wage labour or start- for the immediate adverse impacts (Smith et al. 2000). ing a small business. However, all respondents mentioned Such responses, in fact, reduce farmers’ future capacity to that the prospects of diversifying livelihoods were low in invest and render their livelihoods more insecure (Boansi their villages. About 20% households reported using migra- et al. 2017). In this study, the most popular modes of cli- tion as a strategy to ofset the loss of income as a result of mate adaptation are increasing use of farm inputs such as a climate crisis. This difers from the usual phenomenon of pesticides and fertilisers and using more improved (heat- displacement when families are uprooted from their homes resistant or drought tolerant) crop varieties, in addition to to settle in other locations as a result of a disaster (Ferris, various site-specifc measures such as shifting to other liveli- 2014). Migration of one or more family members refects a hoods, migrating for work, and introducing other modes of planned response to a shock in income. However, it was also irrigation (Khan et al. 2011). As numerous climate change observed that migration is often considered as a last resort impacts manifest themselves on populations, people often when other strategies fail. For agriculture-based households, adopt multiple strategies to ofset future risk. For example, when natural resources are pressurised due to factors such in D.G. Khan, farmers are exposed to recurring foods as as climate change, migration often becomes the last logical well as intense heat waves. In such a scenario, opting one option when alternative economic opportunities are limited strategy may not ensure security of livelihoods in a sustain- (Penning-Rowsell et al. 2013; Arsenault et al. 2015). able manner. Using a mix of heat- and food-tolerant crop Besides the farmers who are taking adaptation meas- varieties and getting crop insurance may beneft the farmers ures, there were also some who were not taking any action more in times of oscillating weather and climate conditions to adapt to climate change. It was observed that lack of (Boansi et al. 2017). employing an adaptive strategy stems from two main rea- Institutional support is also pivotal in shaping farmers’ sons: lack of fnancial resources to introduce a change in adaptive capacity (Hahn et al. 2009). This can be drawn from practices and a lack of awareness of the efectiveness of an the case of Mardan, where despite low exposure and low adaptation strategy. Generally, farmers who did not adapt, sensitivity to climate change and extreme events, underlying especially those in D.G. Khan and Faisalabad, perceived low levels of adaptive capacity has afected its livelihood that it is unfruitful to invest in agricultural adaptation as vulnerability adversely. A major factor behind low adaptive climate extremes erode all their eforts. This implies that capacity is the lack of government support to the community. these farmers may not have enough information regarding Although the District Ofce of Agriculture in Mardan claims the usefulness of adaptation strategies that can withstand to provide agricultural extension services in the district, yet extreme events. Abid et al. (2015) also came to the conclu- the community surveyed was not receiving any government sion that usually farmers are unaware of the type of adap- support to help them bufer against climate manifestations. tation needed and, therefore, believe that adaptation will In contrast, communities in Faisalabad and D.G. Khan were be inefective in their case. While some farmers admitted receiving government support in the shape of training and receiving government-provided training as mentioned ear- awareness sessions, early warnings, provision of new crop lier, this fnding indicates that some proportion of farmers varieties, and access to insurance, subsidies and credit. The are not receiving relevant trainings in the context of climate district agricultural department of D.G. Khan was appreci- disasters. This highlights the need of the local and district ated for the soil sampling they conducted for agricultural government emphasising more on trainings and awareness land and held awareness-raising sessions on what to grow on sessions about adaptation practices and ensuring that these the soil with respect to the changing soil quality. The district trainings are imparted to all farmers, especially targeting administration was also said to be active in disseminating areas that are hard-hit by climate extremes. Afordability of early warnings in case of riverine foods through the police applying adaptation strategies is another facet that constrains and announcements in mosques. Communities were cogni- farmers from using the most efective methods. High-input zant of the positive impacts of these services provided to prices, high prices of new weather-resistant crop varieties, them, but demanded enhanced government support in terms low penetration of credit facilities and widespread poverty of subsidy on purchase of inputs. They admitted receiving in rural areas also restrict farmers’ capacities to invest in information about new and improved seeds to withstand adaptation practices (Abid et al. 2015). changing climatic conditions, but complained that those new seeds were not afordable for the small farmers. In Mardan, farmers expressed their need of loan and credit to support 7 Conclusion and Policy Recommendations agricultural activities. This study found that farmers prefer to opt for of-farm This study uses the IPCC livelihood vulnerability approach activities as a means of diversifcation. Their preferences to understand the important causal factors behind exposure, vary according to the site, including strategies such as sensitivity and adaptive capacity of various rural semi-arid

1 3 470 A. Qaisrani et al. regions in Pakistan that explain farmers’ livelihood vul- activities, local governments must create opportunities for nerability to climate change. According to this study’s them to access well-paid work in villages, such as supporting IPCC–LVI ranking-based results, D.G. Khan appears as the women-owned cottage industries by providing them train- most vulnerable district, Mardan as the second most vulner- ings and access to credit and markets. Women should also able, while Faisalabad as the least vulnerable among the be formally integrated in the value chains and eforts must three districts in Pakistan. An important fnding is that (a be undertaken to reduce the wage gap between men and lack of) adaptive capacity plays quite an important role in women. shaping households’ livelihood vulnerability on any given This study found that climate-resilient seed varieties cur- degree of exposure and sensitivity. rently available to small farmers are quite expensive, espe- As agricultural livelihoods become increasingly vulner- cially for small farmers. In this regard, agricultural exten- able to climate change impacts and risks, farmers resort to sion services need to provide subsidised crop varieties for various strategies to ofset the ill impacts on their earnings. farmers at afordable prices, control the spread of pests and Besides, distress coping strategies adopted to respond to cli- diseases, and ensure proper storage of farm outputs. Further- mate extremes, farmers also employ certain long-term adap- more, agricultural extension services need to broaden their tive measures to compensate for their depreciating liveli- reach, especially in KP, where people are still not receiv- hoods. In addition to increasing farm inputs and diversifying ing government services, and are thus, unaware of the many crop varieties, rural households also consider diversifying positive, livelihood enhancing schemes ofered by the pro- their livelihoods through out-migration. Rural economies vincial government. need people-centric development to build climate resilience, Currently, the potential of migration as a resilience- i.e. growth that diminishes poverty and reduces deprivation. enhancing adaptation strategy is not recognised in devel- Livelihood vulnerability of farmers can be ofset by reduc- opment plans. Policies facilitating planned migration could ing climate sensitivity and enhancing their adaptive capacity support improved climate adaptation for migrant families, to climate change impacts. In light of this study, it is clear and mitigate their risk of displacement. Since migration is that livelihood resilience of rural agricultural households not an important response/adaptation strategy, the national and only needs policy attention for strengthening of the agricul- sub-national governments should mainstream migration and ture sector, but also facilitation in diversifying livelihood climate change adaptation into National Adaptation Plans activities. (NAPs) and sub-national integrated development plans An important cause of sensitivity to climate change in (Local Adaptation Plans of Action—LAPAs). rural semi-arid areas of Pakistan is the variability and uncer- tainty of irrigation water supply for agricultural purposes. Acknowledgements This study has been produced by the Sustain- able Development Policy Institute as part of the long-term research This variability might also be caused due to weak and dete- agenda of the Pathways to Resilience in Semi-arid Economies (PRISE) riorating irrigation infrastructure. The disruption in water research project. PRISE holds the intellectual property rights to the supply can be reduced by adopting innovative irrigation and research conducted for this paper. water-saving/harvesting techniques, and a more efective role of the irrigation department with enhanced interactions Compliance with Ethical Standards with farmers. As climate-related (extreme) events become more evident Conflict of interest On behalf of all authors, the corresponding author and severe, it is increasingly essential to rely on rural youth states that there is no confict of interest. (including men and women) for coping with these risks, and Open Access This article is distributed under the terms of the Crea- adapting their livelihoods to climate change. This can be tive Commons Attribution 4.0 International License (http://creat​iveco​ done by improving their capacity to understand vulnerable mmons.org/licen​ ses/by/4.0/​ ), which permits unrestricted use, distribu- aspects of rural livelihoods, ability to efectively use new tion, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the scientifc information as well as local knowledge to antici- Creative Commons license, and indicate if changes were made. pate and combat climate change risks and stressors such as foods, droughts, and heat waves. Similarly, rural labour also needs to have better and updated skills through trainings so Appendix that they are able to choose alternative (non-farm) oppor- tunities which are less or not vulnerable to climate change. See Tables 4 and 5. Due to women’s active involvement in agricultural activi- ties, especially in D.G. Khan, their vulnerability to climate change is high. To encourage their participation in non-farm

1 3 471 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan

Table 4 IPCC–LVI scores for each study site Indicators and subcomponents Explanation Faisalabad D.G Khan Mardan

Exposure Climate extremes/disaster Average number of frequent climate extreme events in the last 0.129 0.142 0.172 10 years (based on survey) Temperature variability Monthly variability in temperature 0.232 0.248 0.158 Hot months Number of extreme hot months with temperature above 30 °C 0.519 0.607 0.463 Precipitation variability Monthly variability in total precipitation 0.209 0.220 0.171 Extreme dry months Number of extreme dry months explained as the month during 0.667 0.722 0.315 spring season having precipitation < 5 mm and in summer precipi- tation = 0 mm Monetary loss Average monetary loss incurred due to last extreme event 0.055 0.111 0.012 Crop failure Percentage of households facing complete crop failure due to cli- 0.700 1.000 0.700 mate extreme Physical damage to human body Percentage of household with an injury or death as a result of natu- 0.180 0.200 0.020 ral disaster (used to assess the physical impact of any disaster on human body) Average 0.336 0.406 0.251 Sensitivity Source of irrigation Percentage of households that do not have access to canal water 0.600 0.960 0.160 Access to water Percentage of households not having access to good-quality drink- 0.360 0.020 0.000 ing water Physical disability Prevalence of any permanent disability in the household 0.100 0.100 0.120 Cost on health issues Average health-related cost per month for the household 0.119 0.085 0.137 Professional skills Percentage of households not having any skill other than farming 0.700 0.420 0.440 Type of house Percentage of households having house made of temporary material 0.020 0.500 0.320 Average crop diversity index (range The inverse of (the number of crops grown by the households + 1), 0.778 0.798 0.738 0–1) e.g.: a household that grows sugar cane, rice, wheat and maize will have crop diversity index = 1/(4 + 1) = 0.20 Average 0.382 0.412 0.274 Adaptive capacity Dependency ratio Ratio of the population 0–14 and 65 and above years of age over the 0.530 -0.154 0.333 population between 15 and 64 years of age Education Percentage of household heads who have secondary and above level 0.660 0.160 0.180 of education Income diversifcation Percentage of households who have more than one source of income 0.640 0.800 0.820 Access to media/information Percentage of households having access to phone, Internet and TV 0.713 0.353 0.433 Social networking Percentage of households who are actively involved in: local poli- 0.730 0.595 0.680 tics, social relief work, local level associations and have a strong friendship circle Savings Percentage of households who have savings of any kind 0.840 0.600 0.680 Government-initiated disaster man- Percentage of households aware about or have access to govern- 0.52 0.52 0 agement plans ment-initiated climate disaster management plans Use of remittances for investment Percentage of households willing to invest their remittances on 0.740 0.740 0.857 agriculture Purpose of agriculture production Percentage of households that use agriculture production for sale of 0.92 0.92 0.22 products Average 0.699 0.504 0.467 IPCC–LVI (Exposure + sensitivity) − adaptive 0.019 0.314 0.058

1 3 472 A. Qaisrani et al.

Table 5 Testing signifcance of indicators among the three study sites Indicators and subcomponents Explanation Faisalabad D.G Khan Faisalabad Mardan D.G khan Mardan

Exposure (p value 0.000 < 0.05) Climate extremes/disaster Average number of frequent climate 0.489 0.034** 0.132 extreme events in the last 10 years (based on survey) Temperature variability Monthly variability in temperature 0.534 0.291 0.768 Hot months Number of extreme hot months with tem- 0.044** 0.271 0.005** perature above 30 °C Precipitation variability Monthly variability in total precipitation 0.002** 0.000* 0.000* Extreme dry months Number of extreme dry months explained 0.342 0.000* 0.000* as the month during spring season having precipitation < 5 mm and in summer precipitation = 0 mm Monetary loss Average monetary loss incurred due to last 0.019** 0.000* 0.000* extreme event Crop failure Percentage of households facing complete 0.000* 1.000 0.000* crop failure due to climate extreme Physical damage to human Percentage of household with an injury or 0.801 0.007** 0.004** body death as a result of natural disaster (used to assess the physical impact of any disas- ter on human body) Sensitivity Source of irrigation Percentage of households that do not have 0.000* 1.000 0.000* access to canal water Access to water Percentage of households not having access 0.001** 0.001** 0.320 to good-quality drinking water Physical disability Prevalence of any permanent disability in 0.801 0.007** 0.004** the household Cost on health issues Average health-related cost per month for 0.156 0.563 0.044** the household Professional skills Percentage of households not having any 0.004** 0.008** 0.842 skill other than farming Type of house Percentage of households having house 0.001* 0.001* 0.068*** made of temporary material Average crop diversity index The inverse of (the number of crops grown 0.058*** 0.001* 0.001* (range 0–1) by the households + 1), e.g.: a household that grows sugar cane, rice, wheat and maize will have crop diversity index = 1/ (4 + 1) = 0.20 Adaptive capacity Dependency ratio Ratio of the population 0–14 and 65 and 0.000* 0.006** 0.000* above years of age over the population between 15 and 64 years of age Education Percentage of household heads who have 0.000* 0.000* 0.000* secondary and above level of education Income diversifcation Percentage of households who have more 0.076*** 0.043** 0.801 than one source of income Access to media/information Percentage of households having access to 0.011** 0.022** 0.752 phone, Internet and TV Social networking Percentage of households who are actively 0.080*** 0.042** 0.699 involved in local politics, social relief work, local level associations and have a strong friendship circle Savings Percentage of households who have savings 0.007** 0.062*** 0.410 of any kind

1 3 473 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan

Table 5 (continued) Indicators and subcomponents Explanation Faisalabad D.G Khan Faisalabad Mardan D.G khan Mardan

Government-initiated disaster Percentage of households aware about 1.000 0.001* 0.001* management plans or have access to government-initiated climate disaster management plans Use of remittances for invest- Percentage of households willing to invest 1.000 0.149 0.150 ment their remittances on agriculture Purpose of agriculture produc- Percentage of households that use agricul- 1.000 0.001* 0.001* tion ture production for sale of products

References change and other concerns. Occasional Paper no. 27. University of Guelph, Canada Boansi D, Tambo JA, Müller M (2017) Analysis of farmers’ adaptation Abid M, Schefran J, Schneider UA, Ashfaq M (2015) Farmers’ per- to weather extremes in West African Sudan Savanna. [Online] ceptions of and adaptation strategies to climate change and their https://doi.org/10.1016/j.wace.2017.03.001​ . Accessed 8 June 2017 determinants: the case of Punjab province, Pakistan. Earth Syst Byrne TR (2014) Household adaptive capacity and current vulner- Dyn 6:225–243 ability to future climate change in rural Nicaragua. Department Abid M, Schilling J, Schefran J, Zulfqar F (2016) Climate change vul- of Geography, University of Lethbridge, Alberta nerability, adaptation and risk perceptions at farm level in Punjab, Connolly-Boutin L, Smit B (2015) Climate change, food security, Pakistan. Sci Total Environ 547:447–460 and livelihoods in Sub-Saharan Africa. Reg Environ Change ADB (2012) Addressing climate change and migration in Asia and the 16(2):385–399 Pacifc. Asian Development Bank, Manila Coulibaly JY, Gbetibou GA, Kudhlande G, Sileshi GW, Beedy TL Adger WN (2001) Social capital and climate change. Tyndall Centre (2015) Responding to crop failure: understanding farmers’ cop- for Climate Change Research. Working Paper No. 8 ing strategies in Southern Malawi. Sustainability 7:1620–1636 Adger WN (2006) Vulnerability. Glob Environ Change 16(3):268–281 Cutter SL (2003) The vulnerability of science and the science of vul- Adger WN, Huq S, Brown K, Conwoy D, Hulme M (2003) Adapta- nerability. Ann Assoc Am Geogr 93(1):1–12 tion to climate change in the developing world. Prog Dev Stud Cutter SL, Barnes L, Berry M, Burton C, Evans E, Tate E, Webb J 3(3):179–195 (2008) A place-based model for understanding community resil- Adu DT, Kuworno JKM, Anim-Somuah H, Sasaki N (2017) Appli- ience to natural disasters. Global Environ Change 18(4):598–606 cation of livelihood vulnerability index in assessing smallholder Dasgupta P, Morton JF, Dodman D, Karapinar B, Meza F, Riverra- maize farming households’ vulnerability to climate change in Ferre MG (2014) Rural areas. In: Field CB, Barros VR, Dokken Brong-Ahafo Region of Ghana. Kasetart J Soc Sci 39:22–32 DJ, Mach KJ (eds) Climate change 2014: impacts, adaptation and Afzaal M, Haroon MA (2009) Interdecadel oscillations and the warm- vulnerability, part A: global and sectoral aspects, contribution of ing trend in the area—weighted annual mean temperature of Paki- working group II to the ffth assessment report of the intergov- stan. Pak J Meteorol 6(13):13–19 ernmental panel on climate change. Cambridge University Press, Ali A, Erenstein O (2017) Assessing farmer use of climate change Cambridge, pp 613–657 adaptation practices and impacts on food security and poverty in Dawn (2007) Sustainable agriculture in rain-fed areas.Dawn News Pakistan. Clim Risk Manag 16:183–194 [Online]. https​://www.dawn.com/news/25761​2. Accessed 31 Arif M (2007) Livelihood rights of peasants and rural workers. South May 2017 Asia Partnership-Pakistan, Lahore Deressa T, Hassan RM, Ringler C (2008) Measuring Ethiopian farm- Arsenault MP, Azam MN, Ahmad S (2015) Riverbank erosion and ers’ vulnerability to climate change across regional states. IFPRI migration in Bangladesh’s Char Lands. In: Mallick B, Etzold B Discussion Paper 806. International Food Policy Research Insti- (eds) Environment, migration and adaptation: evidence and poli- tute, Washington, D.C. tics of climate change in Bangladesh. AH Development Publish- FAO (2015) Pakistan and FAO: Partnering to Improve Food Security ing House, Dhaka and Livelihoods. Food and Agriculture Organisation of the United Babin BJ, Zikmund WG (2015) Exploring marketing research. Cengage Nations, Islamabad Learn 24:322 Feron D (2012) Monsoon foods in Assam India. Europa [Online]. Batool S, Saeed F (2017) Mapping the cotton value chain in Pakistan: a http://ec.europ​a.eu/echo/index​_en.htm. Accessed 6 June 2017 preliminary assessment for identifcation of climate vulnerabilities Ferris E (2014) Changing times: the international response to internal and pathways to adaptation. Pathways to Resilience in Semi-arid displacement in colombia. Brookings-LSE project on internal dis- Economies (PRISE) and Sustainable Development Policy Institute placement. The Brookings Institution, Washington, D.C. (SDPI), Islamabad Fussel HM (2007) Vulnerability: a generally applicable conceptual Barr R, Fankhauser S, Hamilton K (2010) Adaptation investments: a framework for CC research. Glob Environ Change 17:155–167 resource allocation framework. Mitig Adapt Strat Glob Change GoP (2013) National disaster risk reduction policy. Ministry of Climate 15(8):843–858 Change and National Disaster Management Authority, Govern- Beccari B (2016) A comparative analysis of disaster risk, vulnerabil- ment of Pakistan, Pakistan ity and resilience composite indicators [Online]. PLOS Current GoP (2016) Pakistan economic survey 2015–16. Ministry of Finance, Disasters 1. http://currents.plos.org/disas​ ters/artic​ le/a-compa​ rativ​ ​ Government of Pakistan, Pakistan e-analysis-of-disas​ ter-risk-vulne​ rabil​ ity-and-resil​ ience​ -compo​ site-​ Gray CL, Mueller V (2012) Natural disasters and population mobility indic​ators​/. Accessed 20 Feb 2018 in Bangladesh. In: Proceedings of the National Academy of Sci- Belliveau S, Bradshaw B, Smit B, Reid S, Ramsey D, Tarleton M, ences of the United States of America (PNAS), 109(16). Univer- Sawyer B (2006) Farm-level adaptation to multiple risks: climate sity of California, Los Angeles

1 3 474 A. Qaisrani et al.

Hahn MB, Riederer AM, Foster SO (2009) The livelihood vulnerability Penning-Rowsell E, Sultana P, Thompson PM (2013) The ‘Last index: a pragmatic approach to assessing risks from climate vari- Resort’? Population movement in response to climate-related ability and change—a case study in Mozambique. Glob Environ hazards in Bangladesh. Environ Sci Policy 27s:s44–s59 Change 19:74–88 Porter JR, Semenov MA (2005) Crop responses to climatic variation. Hamid YA, Afzal J (2013) Gender, water and climate change: the case Philos Trans R Soc Lond B Biol Sci 360(1463):2021–2035 of Pakistan. Global Water Partnership, Pakistan Water Partner- Powell JP, Reinhard S (2016) Measuring the efects of extreme weather ship-South Asia, Islamabad events on yields. Weather Clim Extremes 12:69–79 Hasson S, Bohner J, Lucarini V (2017) Prevailing climatic trends and Preston BL, Emma Y, Westaway RM (2011) Putting vulnerability to run-of response from Hindukush-Karakoram-Himalaya, Upper climate change on the map: a review of approaches, benefts, and Indus Basin. Earth Syst Dyn 8:337–355 risks. Sustain Sci 6(2):177–202 Hinkel J (2011) “Indicators of vulnerability and adaptive capacity”: Qaisrani A (2015) Connecting the dots: linking climate change resil- towards a clarifcation of the science–policy interface. Global ience to human capital. Pathways to resilience in semi-arid econo- Environ Change 21(1):198–208 mies. Sustainable Development Policy Institute, Islamabad Hudrlikova L (2013) Composite indicators as a useful tool for interna- Qin D, Plattner GK, Tignor M, Allen SK, Boschung J, Nauels A et al tional comparison: the Europe 2020 example. Prague Econ Pap (2014) Climate change 2013: the physical science basis. In: 22(4):459–473 Stocker T (ed) Contribution of working group I to the ffth assess- Hussain S (2010) Food security and climate change assessment in ment report of the intergovernmental panel on climate change. Pakistan. Special Program on Food Security and Climate Change, Cambridge University Press, Cambridge Oxfam Novib Rasul G, Afzal M, Zahid M, Bukhari AA (2012) Climate change Hutton G, Menne B (2014) Economic evidence on the health impacts in Pakistan: focus on Sindh Province. Pakistan Meteorological of climate change in Europe. Environ Health Insights 8:43–52 Department, Islamabad IPCC (2001) Climate change 2001: impacts, adaptation, and vulner- Reed MS, Podesta G, Fazey I, Geeson N, Hessel R, Hubacek K (2013) ability—contribution of working group II to the third assess- Combining analytical frameworks to assess livelihood vulner- ment report of the intergovernmental panel on climate change. ability to climate change and analyze adaptation options. Ecol In: McCarthy JJ, Canziani OF, Leary NA, Doken DJ, White KS Econ 94:66–77 (eds) Cambridge University Press, Cambridge Saeed F, Salik KM, Ishfaq S (2015) Climate change and heat waves: IPCC (2014) Climate change 2014: impacts, adaptation, and vulner- rural to urban migration in Pakistan. A silent looming crisis. ability. In: Field CB, Barros VR, Dokken DJ, Mach KI, Mastran- Environment and Climate Change Unit, Sustainable Development drea MD, Bilir TE (eds) Summaries, frequently asked questions, Policy Institute, Islamabad and cross-chapter boxes. A working group II contribution to the Salik KM, Jahangir S, Zahdi WZ, Hasson S (2015) Climate change ffth assessment report of the intergovernmental panel on climate vulnerability and adaptation options for the coastal communities change. World Meteorological Organisation, Geneva of Pakistan. Ocean Coast Manag 115:61–73 Irfan M (2007) Poverty and natural resource management in Pakistan. Salik KM, Qaisrani A, Umar MA, Ali SM (2017) Migration futures Pak Dev Rev 4(46):691–708 in Asia and Africa: economic opportunities and distributional Ishaq W, Memon SQ (2016) Roles of women in agriculture: a case efects—the case of Pakistan. Pathways to Resilience in Semi- study of rural Lahore, Pakistan. J Rural Dev Agric 1(1):1–11 arid Economies (PRISE) Report. Sustainable Development Policy Kelly PM, Adger WN (2000) Theory and practice in assessing vulner- Institute, Islamabad ability to climate change and facilitating adaptation. Clim Change Samee D, Nosheen F, Khan HN, Khowaja IA, Jamali K, Paracha PI 47(4):325–352 et al (2015) Women in agriculture in Pakistan. Food and Agricul- Khan MA, Hussain A, Mehmood I, Hassan S (2011) Vulnerability ture Organisation of the United Nations, Islamabad to climate change: adaptation strategies and layers of resilience Schefran J, Marmer E, Sow P (2011) Migration as contribution to in semi-arid zones of Pakistan. Pak J Agric Res 24(1–4):65–74 resilience and innovation in climate adaptation: social network Looney R (2012) Economic impacts of the foods in Pakistan. Contemp and co-development in Northwest Africa. Appl Geogr 33:119–127 South Asia 20(2):225–241 Shah H, Khan MA, Akmal N, Sharif M (2005) Livelihood assets and Muthoni J, Wangui E (2013) Women and climate change: strategies livelihood strategies of small farmers in salt range: a case study for adaptive capacity in Mwanga District, Tanzania. Afr Geogr of Pind Dadan Khan, District Jhelum, Pakistan. Pak J Agric Sci Rev 32(1):59–71 42:1–2 Ngwa KA, Ndoh MI, Balgah RA, Fomunyam RT (2015) Natural disas- Shah KU, Dulal HB, Johnson CA, Baptiste AK (2013) Understand- ters, vulnerability and livelihood security in rural Cameroon. Int ing livelihood vulnerability to climate change: applying the Rev Basic Appl Sci 3(9):108–121 liveliood vulnerability index in Trinidad and Tobago. Geoforum Opiyo FE, Wasonga OV, Nyangito MM (2014) Measuring household 47:125–137 vulnerability to climate-induced stresses in Pastoral Rangelands of Skoufas E, Essama-Nssah B, Katayama R (2011) Too little, too late: Kenya: implications for resilience programming, research, policy welfare impacts of rainfall shocks in Rural Indonesia. Policy and practice. Springer, Pastoralism Research Working Paper 5615. The World Bank, Washington, Pandey R, Jha SK (2012) Climate vulnerability index—measure of cli- D.C. mate change vulnerability to communities: a case of rural Lower Smit B, Wandel J (2006) Adaptation, adaptive capacity and vulner- Himalaya, India. Mitig Adapt Strategy Glob Change 17:487–506. ability. Glob Environ Change Hum Policy Dimens 16:282–292 https​://doi.org/10.1007/s1102​7-011-9338-2 Smith B, Burton I, Klein RJ, Wandel J (2000) An anatomy of adapta- Panthi J, Aryal S, Dahal P, Bhandari P, Krakauer NY, Pandey VP tion to climate change and variability. Clim Change 45:223–251 (2015) Livelihood vulnerability approach to assessing cli- The World Bank (2015) World development indicators: Pakistan. The mate change impacts on mixed agro-livestock smallholders World Bank, D.C. around the Gandaki River Basin in Nepal. Reg Environ Change Turner BL 2nd, Kasperson RE, Matson PA, McCarthy JJ, Corell 16(4):1121–1132 RW, Christensen L et al (2003) A framework for vulnerabil- Patro S, Sahu KK (2015) Normalization: a pre-processing stage ity analysis in sustainability science. Proc Natl Acad Sci USA [Online]. Cornell University Library. Available from: https://arxiv​ ​ 100(14):8074–8079 .org/abs/1503.06462.​ Accessed 12 Feb 2018

1 3 475 What Defnes Livelihood Vulnerability in Rural Semi‑Arid Areas? Evidence from Pakistan

Ullah W, Nihei T, Nafees M, Zaman R, Ali M (2018) Understand- Zachariadis T (2016) Climate change in Cyprus: review of impacts and ing climate change vulnerability, adaptation and risk perceptions outline of an adaptation strategy. Cyprus University of Technol- at household level in Khyber Pakhtunkhwa, Pakistan. Int J Clim ogy, Springer. https​://doi.org/10.1007/978-3-319-29688​-3 Change Strat Manag 10(3):359–378 Zarafshani K, Sharaf L, Azadi H, Passel SV (2016) Vulnerability Wilkinson E, Peters K (2015) Climate extremes and resilient poverty assessment models to drought: toward a conceptual framework. reduction: development designed with uncertainty in mind. Over- Sustainability 8(588):1–21 seas Development Institute, London

1 3 Earth Systems and Environment (2018) 2:499–514 https://doi.org/10.1007/s41748-018-0068-4

ORIGINAL ARTICLE

Unpacking Climate Impacts and Vulnerabilities of Cotton Farmers in Pakistan: A Case Study of Two Semi‑arid Districts

Samavia Batool1 · Fahad Saeed1,2,3

Received: 5 April 2018 / Revised: 1 August 2018 / Accepted: 2 August 2018 / Published online: 14 August 2018 © The Author(s) 2018

Abstract This paper aims to contribute to the understanding of climate risks and vulnerability facing cotton farmers in semi-arid regions of Pakistan. Given the ever-increasing climate change impacts on cotton production in Pakistan, especially in semi- arid regions where water scarcity puts additional pressure on water sensitive agricultural livelihoods, we have conducted this study to identify climate risks facing cotton farmers in two semi-arid districts of Punjab province (average annual con- tribution to total cotton production is 80%), based on various climate indicators (such as temperature, rainfall and climate extremes.). A mix of qualitative and quantitative methods has been used to explore factors of vulnerability and comparative vulnerabilities. In the cotton production stage, we found that vulnerability to climate change decreases with increase in the size of the landholding, mainly because large farmers have more fnancial resources at their disposal to deal with adverse climate impacts, such as crop damages and losses. Adaptive capacity, on the other hand, is found to be one of the signifcant factors determining the overall vulnerability at the household level, as level of exposure and sensitivity do not difer across diferent sized households. In other words, indicators of adaptive capacity, such as access to fnancial resources, diversifed livelihoods and access to weather information plays a major role in reducing vulnerability against climate change.

Keywords Vulnerability · Adaptation · Value chain · Climate change

1 Introduction has a share of 1.0% in the GDP of Pakistan ((GoP) Govern- ment of Pakistan 2017). While cotton production declined Climate change is exacerbating the existing challenges of the signifcantly during 2015–2016, primarily due to food (a agriculture sector (IPCC 2014), especially for agro-based major impact of climate change), the policy focus has shifted economies, such as Pakistan, which has severe implications to agriculture and especially to the cotton sector. This is also for productivity, livelihoods and economic growth (Siddique in recognition that climate change is projected to exacerbate et al. 2012). Agriculture is the backbone of Pakistan’s econ- risks in the future, which would have serious implications omy. It accounts for 19.5% of the gross domestic product for the export economy of Pakistan (Batool and Saeed 2017). (GDP), employs 42.3% of the total labour force and provides Moreover, textile industry, a major industrial sector of the raw material for several value-added sectors ((GoP) Govern- country, is largely dependent on the cotton crop for produc- ment of Pakistan 2017). Within agriculture, cotton is one of tion and hence any adverse impact on local cotton crop has the major sectors driving economic growth of the country. implications for textile sector (in the form of reduced input It contributes 5.2% to the agricultural value addition and supply) and in turn, for the national economy (Batool and Saeed 2017). Studies based on the station observations found an * Samavia Batool increasing trend of temperature in Pakistan (Río et al. [email protected] 2013; Sheikh et al. (2009); ADB (2017); Sadiq and Qureshi (2010); Khattak and Ali 2015; Ali et al. 2016). Moreover, 1 Sustainable Development Policy Institute, Islamabad, Pakistan future projections based on climate modelling also sug- gest a continuation of this trend (Batool and Saeed 2017; 2 Center for Excellence in Climate Change Research, King Abdul-Aziz University, Jeddah, Saudi Arabia Saeed and Athar 2017). Besides mean temperature, climate extremes such as heat waves also show an increasing trend 3 Climate Analytics, Berlin, Germany

Vol.:(0123456789)1 3 500 S. Batool, F. Saeed both in the past and in future projection for Pakistan (Maida features, a one-size-fts-all solution cannot work (Hinkel and Rasul 2010; Saeed et al. 2017). Furthermore, studies 2011). Hence, there arises a need for vulnerability assess- also point towards an increase in temporal and spatial vari- ments that are specifc to location and sector that can guide ability of precipitation both in the past as well as future (Ali practitioners and policy makers to devise and implement et al. 2016; Ikram et al. 2016; Salma et al. 2012; Hussain and targeted policy actions (Panthi et al. 2016). Lee (2014); Siddique et al. 2012). This has consequences Limited literature exists on farm level livelihood vul- not only in terms of droughts and dry spells, but combine nerabilities in Pakistan. Rehman et al. (2018) hint towards efect of snow/ice melting and heavy monsoon precipitation deteriorating incomes as a result of changing climate. Food makes the country very vulnerable to fooding events. As a security vulnerability arising out of climate change is briefy result of these climate impacts, there are a huge number of highlighted in Ullah (2017). CIAT and World Bank (2017) reported evidences of increased poverty, food insecurity and also highlight infrastructure damages and employment losses health deterioration (Asian Development Bank 2017). Not in the aftermath of climate extreme events. Asian Develop- only this, livestock, forestry, water, energy, transport and ment Bank (2017) also found signifcantly negative impact agriculture sector in particular have found to be adversely of changing climate on the livestock, which is one of the afected by climate change to a large extent (Asian Develop- major sources of livelihood for farmers in Pakistan. Apart ment Bank 2017). This in turn, have huge implications for from these, a major concern of the farmers now is climate the livelihoods of millions of people associated with these impacts on crop production. Crop production in Pakistan is sectors and their value chains1 (Batool and Saeed 2017). projected to reduce by 8–10% by 2040 as a result of increase Cotton is considered to be a highly sensitive crop and in temperature (Dehlavi et al. 2015). In terms of adapta- climate change (both gradual change and extreme events) tion, Abid et al. (2015) found that adaptation practices at is expected to have profound impacts on its productivity local level pertains to simple measures such as changing (Ton 2011). Cotton crop is particularly vulnerable to high of crop sowing date, as opposed to advance management temperatures, CO­ 2 concentration in the atmosphere, low technologies. water availability, high atmospheric evaporation rate and In this backdrop, this study focuses on understanding heat stress (MacKerron 2001; Richardson et al. 2002; Singh the climate risks and vulnerabilities facing cotton farmers et al. 2007; Bange 2007). A comparative analysis of climate in Pakistan. The fndings of this paper can help generate impact on cotton crop across several countries, including evidence on how cotton farmers in semi-arid regions are China, Pakistan, Australia, the US and Brazil among others, afected by climate impacts, and more importantly, how indicates that a likely decrease in rainfall along with rising social diferentiation determines the level of vulnerabil- temperatures will likely cause the cotton’s demand for water ity between diferent landholders and locations. Enhanced to increase in all of the study areas (Ton 2011). While the understanding on these issues would be benefcial in design- pace of glacial depletion is getting faster, there is a likeli- ing targeted policy interventions for building resilience of hood that water in the Indus River (shared by Pakistan, India cotton farmers in the face of ever-increasing climate risks. and China), which is a major source of irrigation for nearby The targeted focus on semi-arid regions is because of the cotton producing regions, would start to run dry by the end fact that Pakistan’s water resources are under great threat of of this century (Kakakhel 2015). climate change, especially in semi-arid regions (Salik et al. These studies exclusively highlight direct impacts of cli- 2015). mate change on crop production (including productivity and The paper is organized as follow; Sect. 2 details the meth- yield) and do not take into account impacts on livelihoods of odology adopted for this study, Sect. 3 presents extensive the farmers. However, given huge macro-economic implica- discussion on results and Sect. 4 concludes the paper with tions of climate change and agricultural production being policy recommendations for enhancing resilience of cotton an entry point for climate impacts, it is important to dive farmers, to climate risks. deep into the micro-level impacts to identify and address key climate risks. In addition to understanding future climate risks, vulnerability assessment is the key to build resilience 2 Data and methodology through targeted interventions which can help to reduce identified vulnerabilities. While every region has faced This research paper follows the mixed methods research with varying degrees and types of climate vulnerabilities, approach that integrates both qualitative and quantitative depending on geographical, social, economic and cultural tools to explore research questions in depth. The broader aim of this paper is to understand and explore climate risks facing cotton farmers for promoting climate-resilient live- 1 A value chain is a cumulative process through which a product lihoods and economic development. In line with this, the gains value at each step before reaching end users. paper will address the following two research questions:

1 3 Unpacking Climate Impacts and Vulnerabilities of Cotton Farmers in Pakistan: A Case Study of… 501 a. What are the current and future climate change risks at of close-ended questions related to household (household the cotton production stage? members, education, gender ratio) and farm (irrigation pat- b. What are some of the factors of diferential vulnerability tern, type of crops, level of production, climate impacts) across various socio-economic groups and locale? level information. Key Informant Interviews (KIIs) and Focused Group Discussions (FGDs) were also done to We started our analysis with identifcation of future cli- gain further insights into the vulnerability at community mate risks for cotton production in Pakistan and particularly level. Dera Ghazi Khan (DGK) and Faisalabad (FSD) were for study sites, using a global gridded crop model called selected as study sites, based on their high contribution to EPIC (Environmental Policy Integrated Climate Model) the total cotton production of the country (42% of the cotton (Williams et al. 1989) which is an agro-ecosystem model production in semi-arid regions in Punjab comes from these running with a daily time step. It simulates crop develop- two districts). Three union councils (UC) were selected in ment and yield, hydrological, nutrient and carbon cycles and DGK (namely, Kala, Mana Ahmadani and Mor Jhangi) and a wide range of crop management activities. It takes inputs one in FSD (namely 91), based on their cotton production of minimum and maximum temperature (°C), precipitation averages in the last 10 years. The selection of UCs was car- −2 (mm), global radiation (MJ m ), CO­ 2 concentration, dif- ried out to capture the diversity in climate risks and vulner- ferent soil properties, crop heat requirement. In addition, abilities within the study sites, with DGK UCs (Kala, Mana it also takes diferent parameters related to soil properties Ahmadani, Mor Jhangi) being sensitive to hill torrent and and crop management as input. The data used for this paper riverine fooding, while FSD UC (91) to heat extremes. The was produced as a part of ISIMIP (Inter-Sectoral Impact detailed description of the UCs is presented in “Appendix Model Inter-comparison Project) initiative (Warszawski 2”. et al. 2014). A sampling framework was developed in collaboration For these simulations, underlying assumptions include with the local (district level) agriculture department. A full the use of Representative Concentration Pathway (RCP) 8.5 list of cotton farmers (140–160 per UC) was developed for scenario, the use of Shared Socio-Economic Pathway (SSP) each UC and then a minimum of 100 randomly selected 2 scenario, the representation of full irrigation, and the rep- farmers were interviewed in each UC. To ensure that the resentation of ­CO2 fertilization. In ISIMIP, all crop models survey captured diferent land tenure systems as well, the (including EPIC) are forced with only fve Global Climate following categorization was made for each UC: landless Models (GCMs) which are HadGEM2-ES, MIROC-ESM- farmers (tenants,2 sharecroppers3 and contractors4) small CHEM, IPSL-CM5A-LR, GFDL-ESM-2 M and NorESM1- farmers (holding less than 12 acres of land), medium farm- M. Hence, the present analysis is based on the ensemble ers (holding more than 12 acres but less than 25 acres) and of these fve climate models which is presented by taking large farmers (holding more than 25 acres).5 A minimum of the relative diference of four future period (2016–2035, 75 landholders6 and 25 in each UC were interviewed. The 2036–2055, 2056–2075, 2076–2095) against historical category-based data allowed us to make comparisons across period of 1981–2000. Spatially averaged future yields are diferent landholdings and location. also presented for the provinces of Punjab and Sindh which Table 1 further elaborates the sample size as per each together accounts for more than 95% of country’s cotton pro- category of farmers. While DGK has a greater share of cot- duction. The data from a global gridded crop model EPIC, ton production than FSD, 75% of the sample was selected forced with fve diferent global climate models (GCMs) from DGK whereas FSD accounted for 25% of the sample. runs is obtained from ISIMIP (Inter-Sectoral Impact Model Inter-comparison Project) database (Warszawski et al. 2014). For this simulation, underlying assumptions include the use of Representative Concentration Pathway (RCP) 8.5 sce- nario, the use of Shared Socio-Economic Pathway (SSP) 2 scenario, the representation of full irrigation, and the repre- 2 sentation of ­CO2 fertilization. In earlier literature, RCP 8.5 A person who occupies a land rented from a landlord. and SSP2 are both referred as ‘business as usual’ and ‘mid- 3 A tenant farmer who gives a part of each crop as rent. dle of the road’, respectively (Fricko et al. 2017). Moreover, 4 A daily wage labourer working for landholder under a seasonal among the four IPCC AR5 scenarios, RCP 8.5 is the highest contract. emission scenario. 5 The categorization is done based on the ofcial categories of difer- For farm level climate impacts information, we carried ent farmers done by the Pakistan Bureau of Statistics. 6 out an extensive survey of 436 farming households (cotton It is also important to note that classifcation of farmers was done on the basis of total landownership and not on cotton cultivation area farmers) in two semi-arid districts of Punjab province (see as it is difcult to fnd farmers cultivating cotton on more than 25 full map in “Appendix 1”). The questionnaire was composed acres of land.

1 3 502 S. Batool, F. Saeed

Table 1 Sample size based on land tenure system economic resources, livelihood diversifcation, social sup- Union councils Total port institutions, etc. (Weis et al. 2016). After selecting the proxy variables for Exposure, Sensi- 91 Kala Mana Mor Jhangi tivity and Adaptive capacity (“Appendix 4”), we normal- Ahmadani ised the variables based on the functional relationship of Landowners 77 78 79 75 309 variables with vulnerability, using min–max normalisation Sharecroppers 10 7 8 9 34 (Iyengar and Sudarshan 1982). In case of a positive rela- Tenant and 11 20 15 10 56 tion (for example vulnerability vs. exposure and sensitive employed as wage capacity), Eq. 1 was used, whereas Eq. 2 was used in case labourers of negative relationship (for example, vulnerability vs. adap- Contractors 9 10 8 10 37 tive capacity). Total 107 115 110 104 436 X − min(X ) ij ij , Yij = (1) max(Xij)−min(Xij) 2.1 Calculation of vulnerability index

max(X )−X Both top-down (e.g. climate modelling-based approaches) ij ij , Yij = (2) and bottom up approaches (focus on what causes communi- max(Xij)−min(Xij) ties to be vulnerable) exist to study climate vulnerabilities (Hinkel et al. 2014; Dessai and Hulme 2004; Van Aalst et al. where Xij denotes the value of indicator j (j = 1,2,3…n) in the 2008). Although a combination of both would be an ideal i village (i = 1,2,3…n) and Yij is the normalised score. The situation to identify vulnerabilities, lack of site-specifc cli- normalised values lie between 0 and 1. mate data allowed us to opt a bottom-up approach to iden- In the next step, equal weights were assigned to each indi- tify climate vulnerabilities facing cotton farmers. One of cator using simple average of normalised score and vulner- the major advantage of using a bottom-up approach is that ability index was obtained using Eq. 3: it helps in the identifcation of vulnerable groups and difer- ∑j Xij + ∑j Yij + ∑j Zij ences in vulnerabilities (even at small spatial scale) (Hinkel VI = , (3) K et al. 2014). Moreover, to carry out vulnerability assessment, we fol- where Xij , Yij and Zij are indicators used as proxy variables lowed the vulnerability framework defned in IPCC AR4 i.e. for sensitivity, exposure and adaptive capacity. The vulner- Vulnerability = ƒ (exposure, sensitivity, adaptive capacity) ability score obtained tells us the comparative vulnerabilities This implies that vulnerability of an individual or house- across various households. Using this methodology, we have hold is directly proportional to exposure and sensitivity, derived comparative vulnerabilities: whereas it is inversely proportional to adaptive capacity. This relationship has been endorsed by various experts includ- a. Based on UCs ing Adger (2006); Weis et al. 2016; Metzger and Schröter b. Based on landholding (2006), etc. Exposure in particular, is related to the changes in climatic parameters (their intensity and frequency) and To obtain a score based on each UC or landholding, we its potential impacts on resources. It answers the question aggregated the sum of individuals in a particular UC or of ‘what is exposed?’ and refers to the presence of people, landholding. The fnal score for each component explains livelihoods, species or ecosystems, environmental services the level of vulnerability of people residing in a particular and resources, infrastructure, or economic, social, or cul- UC. UCs and Landholders were then ranked based on the tural assets in places that could be adversely afected (IPCC vulnerability score. First rank represents extreme vulner- 2014). On the other hand, sensitivity is more dependent on ability, whereas vulnerability decreases with the increase socio-economic factors (gender, decision making power, in rank (5th rank = lowest vulnerability). The value of each mobility options, community structure, etc.), that may or component (Exposure, sensitivity and adaptive capacity) may not reduce the adverse impacts of climate change (Car- ranges between 0 and 1, where 1 means most vulnerable dona et al. 2012). Adaptive capacity plays a positive role and 0 means least vulnerable. in decreasing vulnerability against climate threat through Finally, in terms of limitations, this paper focuses on two adjustments in current behaviours. Some major attributes districts only and the fndings cannot be generalized for Pun- of adaptation include education level, networks (that pro- jab or Pakistan. The fndings, however, are refective of the motes social learning and knowledge exchange), access to overall situation in similar regions.

1 3 Unpacking Climate Impacts and Vulnerabilities of Cotton Farmers in Pakistan: A Case Study of… 503

Fig. 1 Mean projected relative changes (in %) in cotton yield relative to 1981–2000 for Pakistan using global gridded crop model EPIC forced by HadGEM2-ES, MIROC-ESM-CHEM, IPSL-CM5A-LR, GFDL-ESM-2M, NorESM1-M

3 Results and discussion because it is also called as business-as-usual scenarios. The negative impacts of climate change on cotton yield start to 3.1 Future cotton production trends appear right from the frst-time slab (2016–2035) in northern region of Punjab. However, it is important to note that there Cotton production provides a direct entry point for climate is a slight increase in the central to southern Punjab region impacts, which then trickle down to the associated value (Blue area) which is mainly attributed to the efects of ­CO2 chains (Batool and Saeed 2017). We have developed the fertilization (McGrath and Lobell 2013), which tends to have impact modelling for the cotton crop to see how cotton yield a positive efect on crop yield (depending on crop type and might change over time under a changing climate. other factors). However, a consistent decrease in the cotton Figure 1 depicts future changes in the cotton yield over yield is witnessed in other time slabs towards the end of the Pakistan relative to the base period 1981–2000, by taking the century. In the last time slab from 2076 to 2095, an acute average of the whole ensemble (EPIC forced with 5 GCMs), reduction of around 60–80.0% can be seen in most of the for four diferent time slabs. As mentioned earlier, RCP 8.5 cotton producing areas of Punjab and Sindh, especially in is the most extreme scenario in the suite of scenarios devel- Punjab, which accounts for 80.0% cotton production of the oped for IPCC AR5. We based our results on this scenario country.

1 3 504 S. Batool, F. Saeed

no signifcant change in mitigation or adaptation practices is made. The cotton yield is projected to decline by at least 60.0% by 2096, which will not only afect the livelihood of the cotton farmers but the impacts will likely have a trickle- down efect on the associated industries as well as the overall economy of Pakistan (Batool and Saeed 2017).

3.2 Vulnerability assessment at farm level

Fig. 2 Future change in cotton yield in Punjab calculated as relative diference (in %) from the annual averaged value of historical period As mentioned earlier, we adopted IPCC AR4 vulnerability (1981–2000) framework which is presented in Sect. 2. This section briefy outlines the indicators used to calculate exposure, sensitivity and adaptive capacity of the landholders and landless cotton As mentioned in Sect. 2, the purpose of analysing the farmers. data from global crop model is to have an idea about the future impact of global warming on the yield of cotton crop 3.2.1 Exposure at the national level in future. Since the provinces of Pun- jab and Sindh together accounts for more than 95% of the Pakistan is severely exposed to climate risks as evident from country’s cotton production, hence we show future yield of climate events in the past and future threats identifed in cotton averaged over these two province in Fig. 2 as relative the literature (Asian Development Bank 2017). This section diference (in %) from the base period. These are plotted discusses the climate indicators used to derive the level of by taking the spatial average of annual cotton yield over exposure of diferent landholders to climate change in our the provinces of Punjab and Sindh. A moving average of study sites. 10 years has been applied to smoothen the data by removing year-to-year variability. The analysis is carried out by tak- 3.2.1.1 Frequency of climate events As a measure of expo- ing the relative diference of four future period (2016–2035, sure to climate risks, we have analysed perception-based 2036–2055, 2056–2075, 2076–2095) against historical data (due to lack of downscaled climate data for each district period of 1981–2000. The dark lines represent median of under consideration) on the occurrence of climate extreme the ensemble, whereas the shaded area represents the full events in our study sites. range of the ensemble. Table 2 provides information about the frequency and The fgure projects a sharp decline in cotton yield by the intensity of various primary and secondary climate events end of this century, in a business as usual scenario, in which

Table 2 Number of times Climate indicators Never experienced Experienced 1–3 Experienced 3–6 Experi- climate events experienced times times enced 6–10 during the last 10 years (% in times parenthesis) Primary/frst order climate events Drought 232 (53.2) 117 (26.8) 69 (15.8) 18 (4.1) Floods 182 (41.7) 119 (27.3) 80 (18.3) 55 (12.6) Heat wave 66 (15.1) 224 (51.4) 131 (30.0) 15 (3.4) Monsoon ­variabilitya 98 (22.5) 246 (56.4) 88 (20.2) 4 (0.9) Hailstorm 418 (95.9) 17 (3.9) 1 (0.2) 0 Secondary/second order climate events/impacts Pest attack 33 (7.6) 171 (39.2) 184 (42.2) 48 (11.0) Erosion 195 (44.7) 171 (39.2) 62 (14.2) 8 (1.8) Waterlogging 390 (89.4) 30 (6.9) 15 (3.4) 1 (0.2) Salinisation 409 (93.8) 19 (4.4) 5 (1.1) 3 (0.7) Weeds 189 (43.3) 172 (39.4) 65 (14.9) 10 (2.3) Price shocks 94 (21.6) 121 (27.8) 166 (38.1) 55 (12.6)

a In context of our study, Monsoon variability refers to the changes in the monsoon rainfall, i.e. either too much rainfall in one go or excessive rainfall over a course of a few days (more than average rainfall)

1 3 Unpacking Climate Impacts and Vulnerabilities of Cotton Farmers in Pakistan: A Case Study of… 505 experienced by cotton farmers in DGK and FSD. At the Table 3 Mean and standard deviation of reported climate events in overall sample scale, heat wave,7 monsoon variability, diferent UCs during the last 10 years resultant pest attack and price shocks after these events are 91 Kala Mana Mor Jhangi witnessed at a higher intensity and scale relative to other cli- Ahmadani mate-related shocks. 51.4 and 56.4% of the respondents said Flood that they experienced heat wave and monsoon variability, Mean 0 4 1 5 respectively, between 1 and 3 times during the last 10 years. Std. deviation 0.0 2.4 1.1 2.8 Similarly, 42.0% of the respondents witnessed pest attacks Drought more than 3 times during the last 10 years. Mean 3 1 2 0 The data also highlights that a large proportion of Std. deviation 3.2 1.8 1.9 0.9 respondents said that they have never experienced foods, Heat wave droughts, erosion, waterlogging, etc. which provides a Mean 3 3 3 2 rationale to develop policy interventions to safeguard these Std. deviation 1.8 1.9 2.0 2.1 farmers, who are currently not under the threat of foods and Monsoon variability droughts, as they could be faced with such challenges with Mean 2 2 2 2 climate change in coming decades. Moreover, there is a large Std. deviation 1.6 1.7 2.0 1.5 percentage of cotton farmers who said that they have faced pest attack, erosion, heat wave and foods more than 1 time and up to 10 times during the last 10 years. Geographic location is a major factor determining the ber of female labourers employed (at the respondents’ land) exposure to climate change. DGK is highly vulnerable to for farm activities is almost double than male labourers in food risk due to close proximity to River Indus but FSD is case of UC 91 and Mana Ahmadani. not exposed to any such threat. Even within DGK, proximity of communities to Indus River defnes the level of expo- 3.2.2.2 Percentage of farmers (respondents) afected by cli‑ sure to foods. Moreover, some communities reported being mate events Figure 3 below highlights percentage of sam- particularly vulnerable to hill-torrent food risk, depend- pled cotton farmers afected by various climate events in ing on their location. Within our sample size, 22.0% of the diferent UCs in DGK and FSD during the last 10 years. In respondents were solely afected by riverine foods while the case of Kala, 95.7% of the respondents said that they 26.2% were afected by hill torrents. 10.1% of the respond- were afected by pest attack and foods in the last 10 years, ents reported to have been afected by both hill torrents and followed by heat wave, prices shock and monsoon vari- riverine foods. ability. Similar climate risks have been faced by UC Mana Based on the data, we fnd that overall foods, drought, Ahmadani. UC Mor Jhangi is more afected by foods monsoon variability and heat wave (as primary indicator, (almost 99.1% of the respondents afected by food), price leading to other issues) are the major issues faced by cot- shocks and monsoon variability. While no episode of food ton farmers. Analysis of location based climate risks high- has been recorded in FSD, UC 91 is more afected from pest lights that UC Kala and Mor Jhangi are more exposed to attack (99.0%) and heat wave (80.8%). Heat wave is one of foods as compared to UC 91 and Mana Ahmadani (Table 3). the many reasons for pest attacks in Pakistan (Zulfqar et al. These UCs have issues pertaining to heat wave and rainfall 2010; Ton 2011; Baig and Amjad 2014). variability. If we segregate our data on the basis of UCs and climate indicators, we fnd that there is a large proportion of cotton 3.2.2 Sensitivity farmers that perceive high sensitivity to climate hazards. For example, in Table 4, we have clustered all those farmers Some of the factors determining the sensitivity of the cotton who have witnessed the climate event at least once in the farmers in DGK and FSD are discussed below. last 10 years under ‘sensitive to climate change. We have also developed another category of farmers who have expe- 3.2.2.1 Number of male and female labourers involved rienced various climate events more than 3 times during the in agricultural activities A large number of females are last 10 years and labelled them as ‘severely sensitive’. Then involved in agriculture in DGK and FSD. Women bring girls based on these, we have categorised the percentage of popu- with them for cotton picking and the picking activity serves lation of farmers exposed. Results are presented in the form as a networking platform for women at village level. Num- of a sensitivity matrix where green represents the population exposure of less than 30.0%, yellow represents population 7 Heat wave refers to prolonged period of excessive heat along with afected is greater than 30.0% but less than 50.0%, and red high humidity. represents more than 50.0% of afected population.

1 3 506 S. Batool, F. Saeed

UC KALA (DGK) UC MANA AHMADANI (DGK) Pest attack 95.7% Pest attack 96.4% Flood 95.7% Heat wave 94.5% Heat wave 88.7% Monsoon variability 80.0% Price shock 80.0% Price shock 75.5% Monsoon… 76.5% Drought 64.5% Weeds 65.2% Weeds 59.1% Soil erosion 47.8% Soil erosion 57.3% Drought 40.0% Flood 37.3% Water logging 9.6% Water logging 10.9% Salinisation 6.1% Hailstrom 10.0% Hailstrom 2.6% Salinisation 8.2% 0% 50% 100% 0% 50% 100%

UC MOR JHANGI (DGK) UC 91 (FSD) Flood 99.0% Pest attack 99.1% Price shock 87.5% Heat wave 81.3% Monsoon… 83.7% Price shock 71.0% Pest attack 77.9% Monsoon… 70.1% Heat wave 74.0% Drought 66.4% Soil erosion 62.5% Soil erosion 55.1% Weeds 48.1% Weeds 54.2% Water logging 22.1% Salinisation 5.6% Drought 15.4% Hailstrom 0.0% Salinisation 4.8% Water logging 0.0% Hailstrom 3.8% Flood 0.0% 0% 50% 100% 0% 50%100%

Fig. 3 Percentage of farmers reported to be afected by diferent climate change indicators during the last 10 years

Table 4 Level of sensitivity to climate change as per UCs Climate event Level of exposure UCs 91 Kala Mana Ahmadani Mor Jhangi

Drought Sensitivity to drought risk 66.4% 40.0% 65.5% 15.4% Severely sensitive to drought risk 44.9% 14.9% 21.9% 1.0% Floods Sensitivity to food risk N/A 95.7% 37.3% 99.0% Severely sensitive to food risk N/A 53.0% 1.8% 69.2% Heat wave Sensitivity to heat wave risk 81.3% 88.7% 94.5% 74.0% Severely sensitive to heat wave risk 35.5% 33.9% 36.4% 27.9% Monsoon variability Sensitivity to monsoon variability risk 70.1% 76.5% 80.0% 83.7% Severely sensitive to monsoon variability risk 17.8% 20.0% 23.6% 23.1%

This table is based on the perception data. Farmers were asked about the number of years they were afected by a particular climate indicator during the last 10 years

According to these measures, in UC 91 44.9 and 35.5% of heat wave (36.4% of the farmers) and drought (21.9% of the the farmers are characterized as severely sensitive to risk of farmers) can be found. A large proportion of farmers in Mor drought and heat wave, respectively. Conversely, farmers in Jhangi (99.0%) reported that they are sensitive to food risk UC Kala appear to be severely sensitive to foods (53.0%), while out of these, 69.2% are severely sensitive to the same. heat wave (33.9%) and monsoon variability (20.0%). Heat wave and monsoon variability is also one of the severe Largely, 95.0% of the cotton farmers said that they are sen- risks facing farmers in Mor Jhangi. sitive to food hazard and 88.7% of the farmers are sensi- tive to heat wave. In Mana Ahmadani, severe sensitivity to

1 3 Unpacking Climate Impacts and Vulnerabilities of Cotton Farmers in Pakistan: A Case Study of… 507

25.00% 23.50% 22.70% 23.10% 22.40% 21.80% 21.50% 20.90% 20.90% 20.90% 19.60% 20.20% 18.70% 19.10% 19.10% 19.20% 19.20% 20.00% 17.80% 18.30% 15.70% 15.50% 15.00%

10.00% of respondets

% 5.00%

0.00% 91 KALA MANA AHMADANI MOR JHANGI

very poor poor middle rich very rich

Fig. 4 Percentage of farmers as per wealth categories/UCs

3.2.2.3 Percentage of population dependent on canal water UC level, UC Kala has the least number of farmers depend- for irrigation In our sample size, we found that 45% of the ent on only a single source of income, i.e. farming. On the farmers use only tube well water for irrigation whereas 50% other hand, UC Mana Ahmadani and Mor Jhangi has the of the farmers use both canal and tube well water. Due to most percentage of farmers with diversifed income sources. variability in canal water availability, only around 5% of the Similarly, small and large landholders are more inclined farmers rely solely on canal water for irrigation. Segregation towards income diversifcation, according to our sample. of data as per UC shows that Mana Ahmadani and 91 have There is, however, no clear trend in case of landless farmers. the largest number of farmers who also use canal water for irrigation. 3.2.3.3 Wealth status To derive sensitivities based on wealth status, we have calculated the wealth index. The 3.2.3 Adaptive capacity methodology has been explained in “Appendix 3”. Wealth index divides our respondents into fve categories, i.e. very A multitude of studies concludes that adaptive capacity plays rich, rich, middle, poor and very poor. Landless farmers are a major role in reducing vulnerabilities and building resil- categorised into very poor and poor categories. Small and ience to climate impacts. As indicated earlier, access to edu- medium farmers are represented in poor, middle and rich cation and resources (both physical and fnancial), livelihood category whereas large landholders are mostly categorised diversifcation and social networks that promotes knowledge into very rich category (see “Appendix 3”). exchange are some of the key attributes of adaptation (Weis As we are interested to compare wealth diferences across et al. 2016; Mendoza et al. 2014). Using these as indicators, UCs, we fnd that UC Kala and Mana Ahmadani have the we have built an indicator of adaptive capacity to compare largest percentage of very poor and poor cotton farmers diferential adaptive capacities among sampled farmers and (Fig. 4). However, Kala also has the largest number of rich UCs. cotton farmers, followed by UC More Jhangi. In terms of sensitivity, interventions should be targeted at more vulner- 3.2.3.1 Education In terms of education of the respondents able farmer’s groups, i.e. very poor and poor. (cotton farmers), UC Mor Jhangi and 91 has the largest num- ber of college graduate cotton farmers (around 30–40%). On 3.2.3.4 Access to fnancial services and post‑disaster com‑ the other hand, UC 91 has a highest percentage (38%) of pensation Out of a sample of 436 farmers, only 34.4% have uneducated cotton farmers, followed by UC Kala (48%). a bank account. 39.7% of farmers said that they have access to crop loans and insurance. However, almost 18.0% of the 3.2.3.2 Livelihood diversifcation Within our sample, farmers reported that they do not prefer to take loans despite agriculture is the primary source of income for 96% of the having access to fnancial services. Major reasons cited for households. However, 57.3% of the households have a major not taking loans is high interest rates on borrowing and low second source of income. Out of those, 21.8% of the house- capacity to return loans with interest. A small proportion holds rely on livestock for livelihood after agriculture. Other (1.0%) of the respondents also reported religious reasons for major secondary sources of income include government job not taking loans. (7.6%), construction (6.2%) and shop keeping (4.1%). At

1 3 508 S. Batool, F. Saeed

Fig. 5 Results of the vulner- ability index (comparative bar chart)

The level of compensation after food events for each 3.3 Vulnerability Index UC is directly proportional to the severity of the food. For example, a high number of farmers received compensation Since the main objectives of this paper is to see how cli- after 2010 foods in More Jhangi (90.0%) and Kala (52.2%) mate vulnerabilities difer across various groups of farm- as compared to farmers in Mana Ahmadani (13.2%). ers and UCs. In this section, we will analyse if there are any diferences in vulnerability to climate change among 3.2.3.5 Access to weather information With regards to diferent groups of landholders (small, medium and large) weather information, 81.7% of the total respondents reported as well as farmers across diferent UCs. The relationship that they receive weekly or monthly updates on weather. used for the calculation of vulnerability Index is presented in There is a statistically signifcant relationship (correlation ‘Data and Methodology’ section. The proxy variables used is signifcant at the 0.05 level; p value 0.021) between land to calculate vulnerability index are discussed in detail in size and receipt of weather information such that the number previous sections. Functional relationship of these variables of farmers who receive monthly/weekly updates increase with elements of vulnerability is defended in “Appendix 3”. with the increase in the size of land. For example, 79.0% of Appendix 4 explains the steps used to derive the vulner- small farmers said they receive weather updates as opposed ability index. to 85.5 and 92.2% of medium and large farmers. 3.3.1 Comparative vulnerabilities across diferent 3.2.3.6 Early warning In our sample, only 12.4% of the farm- landholdings ers from food-afected areas reported that they had received warning prior to the 2010 food. Among those who reported Figure 5 summarises the overall results of the vulnerability to have been warned, 83.6% said that the warning included index. The value of each component ranges between 0 to 1, information about the severity of food. While UC Kala and where 1 means most vulnerable and 0 means least vulner- Mor Jhangi had severe episodes of fooding in 2010, 32.2% able. The fgure shows that landless farmers are most vulner- of the respondents from Kala and 16.3% from Mor Jhangi able to climate change, followed by the categories of small reported that they did not receive a warning before the food. landholders, medium landholders and large landholders. The

1 3 Unpacking Climate Impacts and Vulnerabilities of Cotton Farmers in Pakistan: A Case Study of… 509 vulnerability of landholders is related to the level of adaptive 4 Conclusion and way forward capacity as provided by access to fnancial services, strong for promoting climate resilient cotton networks which allows them to gain knowledge on new production in Pakistan and adaptive agricultural practices, etc. Larger landholders are more likely to have access to these sources of adaptive The findings of this paper provide crucial policy entry capacity whereas small landholders and landless farmers are points, which can help build resilience of the cotton farm- less likely to have large social networks, access to credit, or ers in Pakistan that are under continued threat of climate other assets such as livestock. change. Cotton sector has sufered huge losses as a result Based on the components of vulnerability, we also fnd of adverse impacts of climate in the past few years and will that there is less variation in exposure and sensitivity to cli- continue to be afected by large extent, as depicted by our mate change between landholders and landless. On the other climate projections, if adaptation measures are not taken. hand, large diferences in adaptation were found among both While it is crucial to understand underlying features of these groups which suggests that adaptation capacity shapes vulnerability to promote adaptation, we fnd that vulnerabil- vulnerability to climate change in the case of landholders ity to climate change decreases with increased size of land- and landless cotton farmers in semi-arid regions of Pakistan. holding. More importantly, wealth plays an important part While exposure and sensitivity to climate change are partly in promoting adaptation decisions at household level but it determined by external factors such as household depend- does not always ensure adaptation decisions. In other words, ency ratio and number of climate events experienced at farm not all wealth farmers adapt to climate impacts. In terms of level, adaptation decision-making can be promoted through policy, it implies that there is a need for farmer level aware- targeted policy interventions that build institutional capaci- ness raising about climate change and its implications for ties and promote knowledge creation and sharing. agricultural productivity. Currently, there is no formal mech- anism to disperse climate information top down, where it is 3.3.2 Comparative vulnerabilities across UCs needed the most. Information is disseminated through social networks, which beneft large landholders with large social We fnd that farmers from UC Kala are the most vulnerable networks. This information gap regarding current climate to climate change, followed by Mana Ahmadani, 91 and Mor risks and adaptation requirements results in major losses in Jhangi. A higher level of vulnerability implies high exposure productivity and limited adaptation to future climate risks. and sensitivity to climate change coupled with low levels of Second, despite varying level of vulnerability across land- adaptive capacity to cope with climate change. holdings, we fnd that all landless and landholders are vul- UC Kala is highly exposed to food risks and has a rela- nerable to climate change. Flood being a recurrent climate tively low percentage of farmers with livelihood diversi- risk for Pakistan ends in loss of livelihoods of millions of fcation (45.0%), has larger households (having up to 20 farmers across the country. This not only results in massive members) and the highest number of non-educated cotton food insecurity but also increase in poverty as medium and farmers. Similarly, Mana Ahmadani is also highly vulnerable small landholders are further pushed down the poverty line to several climate change, including monsoon variability, as they fail to recover from food damages. Crop insurance, drought and heat wave that lead to pest attacks. A major per- a potential tool to deal with fnancial losses, is although centage of the farmers (82.7%) are dependent on canal water available (through public banks) but have extremely limited for irrigation which makes Mana Ahmadani highly sensitive outreach. In this context, crop insurance tools need to be to climate change. UC 91 has almost the same characteristics developed that caters for both short and long-term climate as Mana Ahmadani and is particularly vulnerable to heat related losses faced by farmers. Targeted insurance tools stress. But since the level of exposure and sensitivity is not should be developed for landless daily wage laborers. as high, this UC has a relatively lower vulnerability score. Finally, data collected on the access to weather infor- On the other hand, farmers in Mor Jhangi, who are the mation also suggest extremely limited access of farmers to most afected in terms of foods (as shown by exposure) is weather and climate information services. Weather informa- found to be least vulnerable as compared to other UCs. This tion infrastructure at the local level should be upgraded and is due primarily to the relatively high adaptive capacity of eforts should be made to develop easy access of farmers the farmers in this UC as Mor Jhangi has the highest number to information data. Again, training of farmers on how to of college graduate farmers. It also has the highest rate of interpret and utilize climate data for efective adaptation is livelihood diversifcation primarily because of high risk of required at the local level. food every year due to which people do not rely solely on agricultural income. Acknowledgements This study is based on the Pathways to Resilience in Semi-Arid Economies (PRISE) project, funded by Canada’s Interna- tional Development Research Centre (IDRC) and the UK’s Department

1 3 510 S. Batool, F. Saeed for International Development (DFID) through the Collaborative Adap- riverine fooding from the Indus river. We interviewed cot- tation Research Initiative in Africa and Asia (CARIAA). RISE holds ton farmers in two villages, namely Patti Makwal and Basti the intellectual property rights to the research conducted for this paper. Raimen, with 60 and 400 households, respectively. Major Compliance with ethical standards occupation of the villagers include farming, daily wage labourer and foreign employment. Major crops planted in Conflict of interest On behalf of all authors, the corresponding author the villages include cotton, sugarcane, rice and wheat. Patti states that there is no confict of interest. Makwal mostly has small farmers whereas Basti Raimen has a large number of medium-sized farmers (having more Open Access This article is distributed under the terms of the Crea- than 12.5 acres of land per household). Agricultural pro- tive Commons Attribution 4.0 International License (http://creat​iveco​ duction in this UC is particularly afected by foods, heavy mmons.org/licen​ ses/by/4.0/​ ), which permits unrestricted use, distribu- tion, and reproduction in any medium, provided you give appropriate rainfall and pest attack. Although majority of the farmers credit to the original author(s) and the source, provide a link to the produce cotton, there is no cotton farmer’s organization at Creative Commons license, and indicate if changes were made. the UC level. There are around 6–7 pesticide and fertilizer Appendix 1: Map of the study sites

Appendix 2: District profle suppliers, located at a distance of 3 km from the village. Farmers believe that the number of providers is sufcient to Dera Ghazi Khan cater to the demands of local farmers. However, they think that the reach of government extension department needs to Brief profle of the selected UCs of DGK is given here: be enhanced. Kala is located between DGK canal and river Indus. It Mana Ahmadani is further away from the Indus river and is afected by foods from rod khoi (hill torrents) as well as is moderately afected by foods from hill torrents. Within

1 3 Unpacking Climate Impacts and Vulnerabilities of Cotton Farmers in Pakistan: A Case Study of… 511 this, we have covered a number of villages, namely Basti Proxy variables Functional relationship with garbi, Bhabay wala, Kotla Ahmed khan, Basti noor wahi, vulnerability Hala and Basti Foja, based on cotton production fgures. The Sensitivity to climate change population in these villages ranges from 500 (50 households) Number of male and female Vulnerability ↑ with the ↑ in num- to 6000 people (500 households). Major occupations include labourers in agricultural lands ber of labourers agricultural production, daily wage labour (farm workers and Number of farmers afected by Vulnerability ↑ with the ↑ in small factory workers) and foreign labour. Wheat, sugarcane, Floods, droughts, heat stress, number of farmers afected by cotton and tobacco are the major crops of these villages. monsoon variability and climate events hailstorm Major crop issues include rainfall variability, hailing and Number of population Vulnerability ↑ with the ↑ in num- pest attack. dependent of canal water for ber of farmers using canal water Mor Jhangi is located at the western side of the river irrigation Indus and is severely afected by foods from the Indus river. Adaptive capacity of the farming household Basti Malana, being a large village and severely afected by Level of education of the Vulnerability ↓ with the level of 2010 foods, was the only village covered under this survey. respondents education It has an average of 1800 households and has 2500 acres of Highest level of education at Vulnerability ↓ with the ↑ highest agricultural land. It has a good mix of small, medium and the household level level of education ↓ large sized farmers, but large farmers dominate the cotton Livelihood diversifcation Vulnerability with access to of farm employment production. The majority of agricultural land was destroyed Wealth status Vulnerability ↓ with the level of during the 2010 food. Increase in temperature and resultant wealth outbreak of pest is another major issues facing crop produc- Access to fnancial resources Vulnerability ↓ with the access to tion in this region. fnancial resources Access to weather updates Vulnerability ↓ with the access to Faisalabad weather updates Early warning Vulnerability ↓ with early warning Cotton production has declined by 30.0% in Faisalabad since a Sum of total does not add to the total number of respondents as 1991. Cotton farmers have shifted to sugarcane production. multiple options exists for sources of information We covered one UC in Faisalabad having a large number of cotton farmers. Farmers in UC 91 still produce a compara- tively larger yield of cotton. Village Danabad was selected Appendix 4: Wealth index categories as a study site. There is not a single episode of fooding as per landholdings recorded since 1981. This site was chosen to assess other climate indicators such as temperature change and rainfall variability, which have signifcant impact on crop produc- Following the methodology used by Vyas and kumaranay- tion and quality. ake (2016), we have constructed the wealth index, using the Principal Components Analysis (PCA). PCA helps in the identifcation of small uncorrelated variables in a large data- Appendix 3: Indicators used set, defnes their relationship, summarizes the data and pre- for the construction of vulnerability index vent loss of even minute information (Devkota et al. 2014, Filmer and Prittchet 2001). PCA can be mathematically represented as follow: Proxy variables Functional relationship with x − x̄ x − x̄ x − x̄ 1 1 2 2 ⋯ m m vulnerability yj = �1 + �2 + + �m , s1  s2  sm  Exposure to climate change­ a* x̄ s Frequency of foods Vulnerability ↑ with the ↑ in where m and m are the mean and standard deviation of asset frequency of the event xm , and is the weight for each variable. Frequency of droughts Vulnerability ↑ with the ↑ in This division of data creates weighted components or frequency of the event factors that allows for interpretation of smaller components Frequency of heat wave Vulnerability ↑ with the ↑ in from large datasets. Table below highlights the variables frequency of the event used for the construction of the PCA, which includes land ↑ ↑ Frequency of monsoon vari- Vulnerability with the in ownership, regular household items and access to irrigation ability frequency of the event facilities. Frequency of hailstorm Vulnerability ↑ with the ↑ in frequency of the event

1 3 512 S. Batool, F. Saeed

Descriptive statistics Total variance explained Mean Std. deviation Com- Initial eigen values­ a Extraction sums of ponent squared loadings Small farmer 0.43 0.495 Total % of Vari- Cumula- Total % of Cumula- Medium farmer 0.11 0.313 ance tive % Vari- tive % Large farmer 0.23 0.424 ance Tractor 0.33 0.47 TV 0.48 0.5 1 764.883 99.638 99.638 2.47 13.724 13.724 Radio 0.16 0.368 2 0.425 0.055 99.694 Phone 0.57 0.496 3 0.345 0.045 99.739 Smartphone 0.39 0.487 4 0.321 0.042 99.78 Computer 0.06 0.245 5 0.25 0.033 99.813 Canal water 0.55 0.498 6 0.227 0.03 99.843 Use tube well 0.95 0.209 7 0.212 0.028 99.87 Own tube well 0.85 0.361 8 0.175 0.023 99.893 Bank account 0.34 0.476 9 0.15 0.02 99.912 Weather information 0.64 0.479 10 0.129 0.017 99.929 Educated 0.89 0.31 11 0.119 0.016 99.945 Income diversifcation 0.58 0.493 12 0.106 0.014 99.959 Land ownership 0.77 0.421 13 0.086 0.011 99.97 Acres 0.69 0.111 14 0.08 0.01 99.98 15 0.067 0.009 99.989 16 0.048 0.006 99.995 17 0.038 0.005 100 18 4.60E−16 5.99E−17 100

1 3 Unpacking Climate Impacts and Vulnerabilities of Cotton Farmers in Pakistan: A Case Study of… 513

References Hussain MS, Lee S (2014) Long-term variability and changes of the precipitation regime in Pakistan. Asia Pac J Atmos Sci 50(3):271–282 Abid M, Schefran J, Schneider UA, Ashfaq M (2015) Farmers’ per- Ikram F, Afzaal M, Bukhari S, Ahmed B (2016) Past and future trends ceptions of and adaptation strategies to climate change and their in frequency of heavy rainfall events over Pakistan. Pakistan J determinants: the case of Punjab province, Pakistan. Earth Syst Meteorol 12(24):57–78 Dyn 6:225–243 IPCC (2014) Climate Change 2014: impacts, adaptation, and vulner- Adger N (2006) Vulnerability. Global Environ Change 16(2006):268– ability Part B: regional aspects. In: Barros VR, Field CB, Dok- 281 (Elsevier) ken DJ, Mastrandrea MD, Mach KJ, Bilir TE, Chatterjee M, Ebi Ali S, Khattak M, Khan D, Sharif M, Khan H, Ullah A, Malik A (2016) KL, Estrada YO, Genova RC, Girma B, Kissel ES, Levy AN, Predicting future temperature and precipitation over Pakistan in MacCracken S, Mastrandrea PR, White LL (eds) Contribution of the 21st century. J Eng Appl Sci (JEAS) 35:61–76 (University of working group II to the ffth assessment report of the intergov- Engineering and Technology, Peshawar) ernmental panel on climate change. Cambridge University Press, Asian Development Bank (2017) Climate change profle of Pakistan. Cambridge, p 688 Report. Manila, Philippines Iyengar NS, Sudarshan P (1982) A method of classifying regions from Baig MA, Amjad S (2014) Climate change impact on major crops of multivariate data. Econ Polit Wkly 17:2048–2052 Pakistan: a forecast for 2020, Institute of Business Management Kakakhel S (2015) The Indus River Basin and climate change. Crite- Bange M (2007) Efects of climate change on cotton growth and devel- rion Quarterly 10(3) opment. The Australian Cotton Grower. June issue. Green Mount Khattak S, Ali S (2015) Assessment of temperature and rainfall trends Press, Australia in Punjab province of Pakistan for the period 1961-2014. J Himal Batool S, Saeed F (2017) Mapping the cotton value chain in Pakistan: Earth Sci 48:42–61 a preliminary assessment for identifcation of climate vulnerabili- MacKerron D (2001) Climate change and global crop productivity. ties & pathways to adaptation. PRISE working paper. Sustainable Edited by Reddy KR, Hodges HF. Wallingford, UK: CABI Pub- Development Policy Institute, Islamabad lishing (2000), pp. 471. Experimental Agriculture, 37(2), 269–274 Cardona OD, van Aalst MK, Birkmann J, Fordham M, McGregor G, Maida Z, Rasul G (2010) Rise in summer hest index over Pakistan. Perez R, Pulwarty RS, Schipper ELF, Singh BT (2012) Determi- Pakistan J Meteorol 6:12 nants of risk: exposure and vulnerability. In: Field CB, Barros McGrath JM, Lobell DB (2013) Regional disparities in the CO2 ferti- V, Stocker TF, Qin D, Dokken DJ, Ebi KL, Mastrandrea MD, lization efect and implications for crop yields. Environ Res Lett Mach KJ, Plattner G-K, Allen SK, Tignor M, Midgley PM (eds) 8(1):014054 Managing the risks of extreme events and disasters to advance Mendoza MET, The BD, Naret H, Ballaran VG, Arias JKB (2014) climate change adaptation A special report of working Groups I Assessing vulnerability to climate change impacts in cambodia, and II of the intergovernmental panel on climate change (IPCC). the philippines and vietnam: an analysis at the commune and Cambridge University Press, Cambridge, pp 65–108 household level. Environ Sci Manag 17(2):78–91 CIAT and World Bank (2017) Climate-Smart Agriculture in Pakistan. Metzger MJ, Schröter D (2006) Towards a spatially explicit and quan- CSA Country Profles for Asia Series. International Center for titative vulnerability assessment of environmental change in Tropical Agriculture (CIAT); The World Bank. Washington, D.C., Europe. Reg Environ Change 6:201–216 p 28 Panthi J, Aryal S, Dahal P, Bhandari P, Krakauer NY, Pandey VP Dehlavi A, Groom B, Gorst A (2015) Climate change adaptation in the (2016) Livelihood vulnerability approach to assessing cli- indus ecoregion: a microeconometric study of the determinants, mate change impacts on mixed agro-livestock smallholders impacts, and cost efectiveness of adaptation strategies. World around the Gandaki River Basin in Nepal. Reg Environ Change Wide Fund for Nature (WWF) Pakistan, Islamabad 16(4):1121–1132 Dessai S, Hulme D (2004) Does climate adaptation policy need prob- Rehman MA, Siddique BN, Hashmi N, Masud K, Adeel M, Rameez abilities? Clim Policy 4:107–128 M, Khan A, Dawodd KM, Shah SAA, Karim M (2018) Climate Devkota JU, Singh S, Prajapati C, Hada B (2014) Principal components change impact on rural livelihoods of small landholder: a case of analysis of the socioeconomic conditions of biogas users - with Rajanpur, Pakistan. Int J Appl Agric Sci 4(2):28–34. https​://doi. examples from Nepal. Int J Renew Energy Res 4(3):655–664 org/10.11648​/j.ijaas​.20180​402.11 Filmer D, Pritchett LH (2001) Estimating wealth effects without Richardson AG, Reddy KR, Boone ML (2002) Sensitivity analysis of expenditure data or tear: an application to educational enrolments climate change impacts on cotton production using the GOSSYM in states of India. Demography 38(1):115–132 crop model. Int J Biotronics 31:25–49 Fricko O, Havlik P, Rogelj J, Klimont Z, Gusti M, Johnson N, Kolp Río S, Iqbal A, Cano-Ortiz M, Herrero A, Hassan L, Penas A (2013) P, Strubegger M, Valin H, Amann M, Ermolieva T, Forsell N, Recent mean temperature trends in Pakistan and links with tel- Herrero M, Heyes C, Kindermann G, Krey V, McCollum DL, econnection patterns. Int J Climatol 33:277–290 Obersteiner M, Pachauri S, Rao S, Schmid E, Schoepp W, Riahi Sadiq N, Qureshi M (2010) Climatic variability and linear trend models K (2017) The marker quantifcation of the Shared Socioeconomic for the fve major cities of Pakistan. J Geogr Geol 2:83–92 Pathway 2: a middle-of-the-road scenario for the 21st century. Saeed F, Athar H (2017) Assessment of simulated and projected cli- Global Environ Change 42:251–267 mate change in Pakistan using IPCC AR4-based AOGCMs. Theor (GoP) Government of Pakistan (2017) Economic survey of Pakistan Appl Climatol 130:1–14 2016–2017. Ministry of Finance Saeed F, Almazroui M, Islam N, Khan MS (2017) Intensifcation of Hinkel J (2011) Indicators of vulnerability and adaptive capacity: future heat waves in Pakistan: a study using CORDEX regional towards a clarifcation of the science policy interface. Global climate models ensemble. Nat Haz 87:1635–1647 Environ Change 21:198–208 Salik KM, Ishfaq S, Saeed F, Noel E, Syed Q (2015) Pakistan Country Hinkel J, Schipper L, Wolf S (2014) A Framework for Climate Change Situation Assessment. PRISE working paper Vulnerability Assessments. Deutsche Gesellschaft für Internation- Salma S, Rehman S, Shah MA (2012) Rainfall trends in diferent cli- ale Zusammenarbeit (GIZ) GmbH, India mate zones of Pakistan. Pakistan J Meteorol 9(17):34–47

1 3 514 S. Batool, F. Saeed

Sheikh MM, Manzoor N, Adnan M, Ashraf J, Khan AM (2009) Climate Warszawski L, Frieler K, Huber V, Piontek F, Serdeczny O, Schewe profle and past climate changes in Pakistan. Research Report No. J (2014) The inter-sectoral impact model intercomparison pro- GCISC-RR-01, Global Change Impact Studies Centre, Islamabad ject (ISI–MIP): project framework. Proc Natl Acad Sci USA Siddique R, Samad G, Nasir M, Jalil H (2012) The impact of climate 111:3228–3232 change on major agricultural crops: evidence from Punjab, Paki- Weis SWM, Agostini VN, Roth LB, Gilmer B, Schill SR, Knowles JE, stan. PIDE Working Paper Blyther R (2016) Assessing vulnerability: an integrated approach Singh RP, Vara PV, Sunita K, Giri SN, Reddy KR (2007) Infuence for mapping adaptive capacity, sensitivity, and exposure. Clim of high temperature and breeding for heat tolerance in cotton: a Change 136(3):615–629 review. Adv Agronomy 93:313–385 Williams JR, Jones CA, Kiniry JR, Spanel DA (1989) The EPIC crop Ton P (2011) Cotton and climate change: impacts and options to miti- growth model. Trans ASAE 32:497–511 gate and adapt. International Trade Centre (ITC) Geneva: ITC, Zulfqar MA, Sabri MA, Raza MA, Hamza A, Hayat A, Khan A (2010) 2011. xii, 32 p. (Technical paper) Efect of Temperature and Relative Humidity on the Population Ullah S (2017) Climate change impact on agriculture of Pakistan— Dynamics of Some Insect Pests of Maize. Pakistan J Life Soc a leading agent to food security. Int J Environ Sci Nat Resour Sci 8:16–18 6(3):555690. https​://doi.org/10.19080​/IJESN​R.2017.06.55569​0 van Aalst MK, Cannon T, Burton I (2008) Community level adaptation to climate change: the potential role of participatory community risk assessment. Global Environ Change 18(1):165–179

1 3 Earth Systems and Environment (2018) 2:525–535 https://doi.org/10.1007/s41748-018-0062-x

ORIGINAL ARTICLE

The National Climate Change Policy of Pakistan: An Evaluation of Its Impact on Institutional Change

Muhammad Mumtaz1

Received: 20 May 2018 / Revised: 8 July 2018 / Accepted: 15 July 2018 / Published online: 21 July 2018 © Springer Nature Switzerland AG 2018

Abstract Climate change is a reality. It is happening and posing adverse impacts globally as well as on Pakistan. To efectively respond to this ubiquitous threat, Pakistan formulated the National Climate Change Policy (NCCP) in 2012 and operationalized it in 2013. Yet it encourages further analyses and evaluations so as to identify any un-addressed or unidentifed measures in the policy, to examine the established policy measures more in depth, and to ensure its efective implementation. This study presents a qualitative analysis of the NCCP. To undertake the analysis in a recognized and systematic fashion, the policy document is evaluated against a criterion set by Cheung et al. (Aust Health Rev 34:405–413, 2010). The main characteristics of this criterion include accessibility, policy background, policy goals, resources, monitoring and evaluation, public oppor- tunities, and obligations. This study contributes to the literature to understand the critical aspects of a climate policy from a developing and one of the most afected countries due to climate change. The study is important to explore the strengths and shortcomings of the policy. Additionally, our study contributes by setting a framework of novel insights by utilizing a new criterion for analyzing a climate policy. The analysis provides valuable inputs to the subnational governments in Pakistan which are actually responsible for implementation of climate and other related policies. Our evaluation found that NCCP ofers some strengths but the document has certain weaknesses too. The policy is a promising document which provides directions and guidelines to the subnational governments for establishing their policies and efective actions plans. It provides a proper mechanism of monitoring the implementation activities in the country. Moreover, it covers important sectors and emphasizes on integration of sectoral policies with climate change policy. The policy presents a reasonable mechanism to enhance the human and institutional capacity. However, it lacks realistic and comprehensive backing for established goals and objectives. For instance, it proposes some measures which are not practically actionable. One of such measures suggests to protecting the glaciers which is not possible keeping in view the existing military confict between Pakistan and India in the region. This shows that the policy lacks to base on empirical research. The fndings of the study will be very helpful for policy makers and climate experts while revising or revisiting the policy document. This analysis provides valuable lessons to provincial governments in Pakistan while framing their provincial climate change policies and action plans. The study opens new research areas and avenues for further evaluations and analysis for the NCCP and upcoming provincial climate change policies in Pakistan.

Keywords Policy · Climate change · Analysis · Pakistan

1 Introduction There are evidences that climate change is happening around the world. These evidences include rise in tempera- Climate change is a reality. According to Intergovernmental ture, changing rain and snowfall patterns, rising sea-level, Panel on Climate Change (IPCC) ffth assessment report shrinking sea-ice, and melting glaciers (Singh and Singh humans are responsible for this unsustainable situation. 2012; Weitzman 2009). Climate change poses serious threats to the world * Muhammad Mumtaz (Leiserowitz 2005). From future climate change perspec- [email protected] tive, both human and natural systems are at risk (Walker et al. 2014). There are evidences that climate change will 1 Fundacao Getulio Vargas Escola de Administracao de continuously pose threats throughout this century despite Empresas de Sao Paulo, Sao Paulo, Brazil

Vol.:(0123456789)1 3 526 M. Mumtaz taking certain successful steps by international community inter-ministerial coordination, regional and international coop- to curb greenhouse gas emissions (Wilson 2006). Many eration are also envisioned in the policy. countries around the world, especially, the developing The NCCP also highlights some other prominent policy countries are more at risk due to climate change. measures. For instance, it emphasis on regional and inter- Climate change is posing negative impacts on South national cooperation on climate change, to get benefts from Asia (Kelkar and Bhadwal 2007; Sivakumar and Stefanski international fnancial mechanism and establishment of vari- 2011). This region is the most disaster-prone region in ous fnancial institutions. Pakistan is one of the developing the world (Sivakumar and Stefanski 2011). South Asia is countries, which have prepared such a comprehensive policy afected annually by climate extremes: increase in aver- on climate change (Yusuf 2011). However, it is important to age temperature, rise in sea level, and the recession of evaluate the policy for its efective implementation. glaciers, decrease trend in annual mean rainfall and change Establishment of the NCCP is a positive step towards tack- in precipitation are the evidences of this change (Sterrett ling climate change in Pakistan it requires proper scrutiny and 2011). Such climatic impacts severely threaten the liveli- analysis. It is evident that the NCCP is not evaluated much hood of poor people living in these areas (Morton 2007). being a nascent policy. In our literature review, we came across Asian region, Pakistan is highly sensitive to the impacts of some notable analysis conducted by Khan (2012) and Mumtaz climate change (Wassmann et al. 2009). Pakistan is ranked (2013). in the list of top 10 most vulnerable countries (Khan and According to Khan (2012), weaknesses of the NCCP are Samiullah 2015). identifed: “In its [the NCCP] present form it is difcult to Climate change is happening and causing adverse impacts implement as the NCCP has not meaningfully involved key on Pakistan. Geographically, Pakistan is located in a region stakeholders”. Relevant literature, which was used for this where the impact of climate change is being felt quite seri- analysis, indicates that the NCCP requires analysis and evalu- ously (Malik et al. 2012). Pakistan is facing serious threats ation, especially in academic circles, for its validation and due to climate change in the form of many disasters like improvement. While acknowledging the importance of devel- floods, droughts, and other natural calamities (Banoori oping the NCCP (Khan 2012) stresses that the NCCP must 2012). These disasters leave social, environmental, and be analyzed. economic impacts. For example, on the heal of the 2010 Some shortcomings in the NCCP in the review of Khan foods, it is estimated that more than 20 million people were (2012) were identifed, such as the absence of analyses on afected, about 1.88 million houses damaged, 1767 persons the NCCP, and self-demand of the NCCP for its analysis and were killed or missing, and 2865 persons injured (Kurosaki evaluation dictate that the NCCP must be evaluated. Similarly, et al. 2011). Mumtaz (2013) stressed that for proper evaluation, the NCCP Pakistan is among those countries which were badly requires more study not only to explore the weaknesses but to afected in 2012 due to climate change (Germanwatch 2014). validate the studies that have taken place for the NCCP. Livelihoods of millions of people, water, food and energy There are possibilities that the nascent policy may have security, are in state of danger in Pakistan due to climate overlooked some important aspects. By undertaking detailed change (Aftab and Hickey 2010). Due to high vulnerability, and rigorous analyses of the NCCP, overlooked aspects it is imperative for Pakistan to confront the consequences of would be uncovered in addition to any inappropriate pro- climate change within due course of time. posals in the NCCP. The in-depth analysis and evaluation of In Pakistan, climate change has attained serious attention a policy is helpful for successful implementation. It is likely by the government for realizing its sensitivities and vulner- that fndings of the study would play a constructive role not abilities (Rasul et al. 2011). Like rest of the world, Pakistan only for the revision of the NCCP but it will also provide a responded climate change by taking various initiatives in the solid background for provincial climate policies and upcom- form of climate policy and action plans. Pakistan launched ing climate change action plans in Pakistan. its frst climate change policy in 2012. The formulation of The objective of the study is to analyze the NCCP docu- the National Climate Change Policy (NCCP) was a positive ment so as to highlight its strengths and shortcomings. For development to deal with climate change in the country. The the analysis, we employ Cheung et al.’s (2010) framework policy proposes more than 120 policy measures covering which is a well-suited model to analyse any public policy. diferent areas. The NCCP is a multi-sector policy which provides mitiga- tion and adaptation measures. The policy stresses on devel- 2 Evaluation of Climate Policies opment sectors such as agriculture, transport, human health, energy, forestry and disaster preparedness. The policy also Climate policies are a concrete response to the impacts of emphasizes on raising awareness, technology transfer and climate change. Climate change is considered a compara- capacity building and institutional strengthening. Additionally, tively young and new area of public policy making (Urwin

1 3 The National Climate Change Policy of Pakistan: An Evaluation of Its Impact on Institutional… 527 and Jordan 2008). Many countries around the world have Informal evaluation performed by non-state actors is established or are establishing climate policies to address considered more critical to analyze policies (Weiss 1993). the negative consequences of climate change. However, it is Informal evaluation may employ efective and robust criteria imperative to investigate the efectiveness and practicability (Mickwitz 2013) to focus on policy side efects (see Vedung of these policies by evaluating and analyzing such policies. 2013). The informal evaluation is conducted to expose the Vedung (1997) defnes evaluation as the ‘careful retro- weak aspects of policies to put pressure on policy makers spective assessment of the merit, worth, and value of admin- to respond. Informal evaluation also has weaknesses. For istration, output and outcome of government interventions, instance, the evaluators may not have deep knowledge of the which is intended to play a role in future practical action whole policy process and they may overlook some important situations’. The evaluation and analysis of climate change aspects or political discourses through which policy emerged policies are taking place around the world. However, the (Weiss 1993). Moreover, the informal evaluation may not evaluation of environmental/climate policy has developed be recognized much unless it becomes a mean to produce at a slower place as compared to other policy areas such as public pressure. education and welfare (Knaap and Kim 1998; Mickwitz and In general, policy evaluation has gone through vari- Birnbaum 2009). ous stages before its emergence in the nineteenth century The related literature informed us that academics and (Crabbe´ and Leroy 2008). Initially, it was supposed to assist climate change practitioners are more interested to evaluate national parliaments in monitoring the lawfulness of gov- and analyze the performances of existing climate policies ernment actions. After the Second World War, its emphasis (Haug et al. 2010). These evaluations are in the form of non- was transferred to assess more administrative, managerial, scientifc evaluations and scientifc evaluations. The non- and economic inquires. Lastly, from 1990s onwards, politi- scientifc evaluations include evaluations commissioned by cal questions related to public support for policies were NGOs and governments, etc., whereas scientifc evaluations becoming the focus of investigations. During this period are in the form of books and peer-reviewed journal articles, many popular evaluation criteria emerged. In the evaluation etc. There are some other notions explaining the diference literature, these evaluation parameters are indicated in the in evaluations. form of efectiveness and goal attainment, cost efective- Various scholars emphasized to distinguish between ness, efciency, legal acceptability, legitimacy, fairness, and formal evaluation which is driven by government and the coordination with other policies (Crabbe´ and Leroy 2008; informal evaluation which is society driven. Hildén et al. Kraft and Furlong 2010). (2014) defne formal evaluation as ‘state-led’ and informal Various evaluation techniques have been used to evaluate evaluation as ‘evaluation activities by non-state actors’. and analyze climate policies. The most common indicators Weiss (1993) diferentiates as ‘inside evaluation’ which used to evaluate the climate policies are identifed as efec- is conducted by people ‘inside’ government, and ‘outside tiveness and/or goal achievement, efciency, and cost efec- evaluation’ by actors not linked with government. These tiveness (Huitema et al. 2011). They further argue that some formal and informal evaluations have their own strengths other criteria are also used like fairness, coordination with and weaknesses. other policies, and legitimacy. However, these were used As far as formal evaluation is concerned, it is argued that far less frequently as compared to previous ones. Moreo- the evaluators are in a better position to assess the policy ver, some other frameworks are being utilized to analyze keeping in view they are quite familiar with policy process the climate policies. For instance, critical discourse analysis, and circumstances in which the policy emerged (Toulem- argumentative discourse analysis set by Maarten Hajer et al. onde 2000; Weiss 1993). These evaluators pay much atten- Climate policy evaluations and analysis are conducted tion to the key motivator factors for policy evaluation (Weiss around the world. These evaluations are being produced 1993). However, there are some weaknesses as well for for- either by non-scientifc evaluations or scientifc evaluations. mal evaluation. Everyone is free to conduct such evaluations. The practition- It is pointed out that evaluation by government actors ers of climate change and academia are the leading evalua- may not be so critical as compared to the evaluation con- tors of climate policies. One of the studies in literature indi- ducted by non-stake actors (Weiss 1993). It is further argued cates that universities and independent research institutes, that formal actors apparently seek the evidences that suit to followed by consultancy frms are the most active evalua- their established hypotheses or views on a policy by way of tors for climate policies (Huitema et al. 2011). According to a ‘confrmation bias’ (Nickerson 1998). Chelimsky (2006) Lehtonen (2005) and Martinuzzi (2004), evaluation practices notes that if the evaluation brings some unfavorable results, have the potential to act as a new form of environmental the governmental actors can suppress them so that to keep governance. It is noted that there is still a very long way to them away from public debate. Likewise, informal evalua- go before the climate policy evaluation is fully realized. The tion also has strengths and weaknesses. process is quite fexible and the practices of climate policy

1 3 528 M. Mumtaz evaluation will continue to develop and it will bring new not a much analyzed policy, especially in academic circles. models and dynamics for the evaluation. That is why this policy document is selected for evaluation. The criteria set by Cheung et al. (2010) is important for policy evaluation. According to Cheung et al. (2010), the 3 Methodology established criterion is important for evaluating a policy proposal to improve the policy document and for its efec- The type of methodology adopted by any research depends tive implementation. Therefore, it completely fulflls the upon the central research objective and questions. Our demands of our study. research problem is to evaluate the NCCP. We adopted In their study, Cheung et al. (2010) used the established qualitative research design which includes three steps. In framework for health policy. However, they suggested that the frst step, a content analysis of the NCCP was done. For this framework can be used to evaluate any policy document this purpose, the NCCP document was accessed, read and irrespective of its area. Ellahi and Zaka (2015) used this understood in detail. In the second step, we choose Cheung framework to analyze Higher Education Policy Framework et al. (2010) framework for comprehensive evaluation of the for Open and Distance Education in Pakistan. Ellahi and NCCP. The NCCP document is evaluated against a certain Zaka (2015) indicated in their study that they consulted two criteria established by Cheung et al. (2010) given in Table 1. experts to validate the criteria for evaluation and both the In the third step, the study fndings are validated through experts agreed that the given criterion is well established semi-structured interviews with policy and climate experts to evaluate any policy. Therefore, it has provided a solid in Pakistan. In these interviews with policy experts, we dig backing to use the intended framework for our study. Addi- out the motivations for establishing the NCCP. tionally, we contacted the author(s) through email about the The proposed study is signifcant as policy analysis is use of their framework for evaluation of the NCCP. They important to provide a solid backing for efective implemen- strongly encouraged and suggested that the given frame- tation. The evaluation of the NCCP is in and of itself vital. work is well poised for evaluation of the NCCP. Therefore, This is because policy analysis is an important task (Majone Cheung et al. (2010) is an appropriate and well-suited frame- 1977). This vitality of policy analysis becomes essentiality work for our study. important when a policy under question is un-analyzed or The proposed framework established by Cheung et al. scantly analyzed. This notion equally applies to the NCCP. (2010) includes accessibility, policy background, pol- The NCCP, as far as our literature review is concerned, is icy goals, resources, monitoring and evaluation, public

Table 1 Criteria for analyzing the policy document

Accessibility The policy document is accessible (hard copy and online) Policy background The source of policy is explicit 1. Authority (persons, books, articles, or other sources of information) 2. Quantitative or qualitative analysis 3. Deduction (premises that have been established from authority). The policy encompasses some set of feasible alternatives Goals The goals/objectives are explicitly stated: The goals are concrete enough to be evaluated later The goals are clear in intent and in the mechanism with which to achieve the desired goals, yet does not attempt to prescribe what the change must be The outcomes of the goals are clearly stated Resources Financial resources are addressed (e.g., estimated fnancial resources and their cost) Human resources are addressed Organizational capacity is addressed Monitoring and evaluation The policy indicated monitoring and evaluation mechanism The policy nominated a committee or independent body to perform the evaluation The outcome measures are identifed for each objective/goals The data collected for evaluation collected before, during, and after the introduc- tion of the new Policy Follow-up takes place after a sufcient period to allow the efects of policy change to become an evident criteria for evaluation are adequate or clear Public and political opportunities The population supports the actions Multiple stakeholders are involved Primary concern of stakeholders and acknowledged to obtain long-term support Obligations The obligations of various implementations are specifed—who has to do what

1 3 The National Climate Change Policy of Pakistan: An Evaluation of Its Impact on Institutional… 529 opportunities and obligations. The framework is explained measures, it is suggested to opt for high-level technology in Table 1. but keeping in view the economic condition of Pakistan to acquire such technology is not possible. The policy back- ground does not adequately provide the sources of the given 4 Analysis of the Study background information. Moreover, based on given informa- tion, it can be inferred that systematic literatures is not well In this section, we analyse the NCCP through the proposed reviewed. framework given in Table 1. 4.3 Goals 4.1 Accessibility In the NCCP documents, the goal and objectives are explic- The frst stage of the evaluation is accessibility of the policy itly mentioned. The objectives mentioned in the policy are document. In case of the NCCP, this condition is fulflled as follows: as the policy document is available on the website of the ministry of climate change. The hard copy of the policy 1. To pursue the sustained economic growth by appropri- document can be obtained from the relevant department of ately addressing the challenges of climate change. the ministry. Moreover, the hard copy of the policy can also 2. To integrate climate change policy with other related be obtained from the other organizations working in area national policies. of climate change. These organizations include Sustainable 3. To focus on pro-poor gender-sensitive adaptation while Development Policy Institute, Global Change Impact Studies also promoting mitigation to the extent possible in a Centre, Leadership for Environment and Development Paki- cost-efective manner. stan, International Union for Conservation of Nature, etc. 4. To ensure water security, food security and energy Moreover, the policy document is also available on inter- security of the country in the face of challenges posed net. Therefore, researchers, stakeholders or anybody who is by climate change. interested to obtain the NCCP document can acquire easily. 5. To minimize the risks arising from expected increase in frequency and intensity of extreme events: foods, 4.2 Policy Background droughts, tropical storms, etc. 6. To strengthen inter-ministerial and inter-provincial The background information of the NCCP is available. The decision-making and coordination mechanism on cli- document contains background information stating and mate change. explaining why this document is being prepared. It also pro- 7. To facilitate efective use of the opportunities, particu- vides some data in the various sections but without any spe- larly fnancial, available both nationally and interna- cifc source. An attempt is made to set a background but it is tionally. not considered to be a comprehensive background. There is 8. To foster the development of appropriate economic room available to improve its valuable background. Being a incentives to encourage public and private sector nascent policy, it may lack to provide detailed background investment in both adaptation and mitigation measures. of the policy. However, some important sources are well 9. To enhance the awareness, skill and institutional capac- explored and certain valuable documents were consulted ity of relevant stakeholders. before establishing the NCCP. For instance, the Task Force 10. To promote conservation of natural resources and long- Report on Climate Change (TFCC) 2010 is one of the build- term sustainability. ing blocks for the creation of the NCCP. The TFCC is a com- prehensive document prepared by the planning commission The policy has established the main goal “To ensure that of Pakistan. The TFCC is a document mainly prepared by climate change is mainstreamed in the economically and the federal government. However, relevant stakeholders were socially vulnerable sectors of the economy and to steer Paki- consulted from all provinces while preparing the document. stan towards climate resilient development”. The above ten Therefore, the claim of the NCCP for fair representation mentioned objectives are set to achieve the main stated goal. from provinces and other stakeholders seems fne and which However, some objectives are not explicitly addressed in is an appreciable step. the policy. The justifcation appears weak for some objec- The NCCP proposes certain measures without any statis- tives while proposing policy measures. For instance, in tical evidences for such recommendations. It lacks to provide objective 6, it proposes to strengthen the inter-ministerial considerable statistical evidences that need to be explored. decision-making and coordination mechanisms but it does It provides very general recommendations at some points not appear relevant keeping in view that climate change is which are not practically actionable. For instance, at some a provincial subject and provinces are actually responsible

1 3 530 M. Mumtaz for implementation of climate-related policies. In objective mechanism will be framed at federal and provincial levels. 8, it points out that economic incentives would be ofered It also mentions that the NCCP will be revised after every to public and private sector investment to promote adapta- 5 years based on empirical data provisions and recommenda- tion strategies. This is a very important objective to enhance tions given by established committees at federal and provin- adaptation strategies considering the climatic condition cial level. However, after the 18th constitutional amendment, in the country. However, this is not addressed properly in the future of the NCCP is unclear because the respective policy measures. Although policy goal and objectives are provinces are establishing their own climate change policies mentioned clearly but to evaluate them specifcally using and action plans. quantitative evaluation seems difcult. There is no evidence The policy explores the signifcance of monitoring. How- for any alternative objective to the stated objectives. Moreo- ever, it does not provide outcome measures for given objec- ver, some of objectives are just mentioned as example above tives and goal. The policy highlights that the NCCP will but without solid backing in the form of proposed policy be revised at fve-year intervals but it does not explicitly measures. Therefore, it is needed to revise the weak objec- mention about data collection before or during the policy tives while revisiting the policy. implementation. However, implicitly it can be said that the established committees may have a task to collect data so 4.4 Resources that they may revise the policy accordingly. Moreover, the policy does not mention any specifc criteria about evalua- The policy accepts that Pakistan is lacking financial tion and monitoring. resources, human resources and institutional capacity at the moment. However, the policy describes some measures to 4.6 Political and Public Opportunity address these challenges. As far as fnancial resources are concerned, the policy dictates how these resources would be The policy mentions the importance and involvement of generated. It is stated that being the signatory to the United all relevant stakeholders. It claims that related stakeholders Nations Framework Convention on Climate Change, various were consulted. It is stated that all the provincial govern- other institutions fnancial assistance can be taken. However, ments, federal government, various ministry and depart- it is a fact that Pakistan could not gain too much from Clean ments, non-governmental organizations and civil socie- Development Mechanism as compared to other Asian coun- ties played an important role for establishing the policy. It tries. Therefore, it is too early to say that Pakistan will gain has acknowledged various federal departments, provincial much from these institutions. There is no given fnancial departments, and other independence bodies for their con- cost to community nor did any estimate fnancial resources tributions and inputs. However, it does not mention to what for the implementation of the policy is mentioned. All the extent these agencies were involved in the formation of pol- measures regarding fnancial resources are very general and icy and what were their stances. Second, it does not explain statistically no data is provided in the policy. who are the real stakeholders and how they were engaged? In terms of human resources, the policy provides certain It is also not clear what the primary concerns of these stake- options. For instance, it is described that the human capacity holders were and whether they were addressed or not? would be enhanced by sending young scientists and students to renowned international institutions so as to develop cli- 4.7 Obligations mate change expertise. Likewise, revision and development of new curriculum on climate change will be introduced in It is indicated that various federal and provincial bodies educational institutions in Pakistan. including other relevant departments are responsible for the Regarding organizational capacity, the policy docu- task to implementation. A complete hierarchy for imple- ment indicated the need to establish more institutions. For mentation committees at Federal and Provincial level is instance, establishment of National Climate Change Com- indicted in the policy. In the policy document, it is stated mission, introduction of climate change cells at federal and that the federal government shall develop an “Action Plan” provincial levels for improving the coordination among the for implementation. The federal government already estab- relevant departments. The policy proposes certain measures lished implementation framework for the NCCP in 2013. to manage these resources. However, time will tell how it The framework was developed for 2014–2030 to effec- will be managed and implemented. tively implement the NCCP. It proposed priority (within 2 years) measures, short-term (within 5 years) measures, 4.5 Monitoring and Evaluation medium-term (within 10 years) measures, and long-term (within 20 years) measures. The actual implementation The policy describes the importance of monitoring and mechanism of the NCCP is very slow and we could see any evaluation. It explains that how the policy implementation substantial achievements for implementation of proposed

1 3 The National Climate Change Policy of Pakistan: An Evaluation of Its Impact on Institutional… 531 measures of the framework document. For instance, the pri- UNEP Governing Council in 1995. Kakakhel was elected ority actions should have been taken within 2 years of time. member of the Executive Board of the UNFCCC ‘Clean However, these actions are not taken at all or just partially Development Mechanisms’ for 2009–10. Again for the years taken. The driver behind the weak implementation can be 2011–12, he was representing the Non-Annex parties of Asia the devolution of the subject to the province. Therefore, the Pacifc. He is a member of the Board of Governors for the provinces are establishing their own policies and implemen- Sustainable Development Policy Institute in Islamabad since tation frameworks at a provincial level. The time will defne 2009 and has served as its Chairperson since 2013. Moreo- how subnational governments have set their frameworks for ver, he was one of the senior members of the TFCC estab- actions to handle climate change in their respective domains. lished by the Government of Pakistan in 2008. All the provincial governments, Azad Jammu and Kash- Ali Tauqeer Sheikh is the chief executive ofcer of Lead- mir, Gilgit-Balistan, Federally Administered Tribal Areas ership for Environment and Development in Pakistan. He and local governments will have their own plans and strate- is deeply involved in sustainable development, particularly gies for efective implementation of climate change poli- in poverty-environment nexus, climate vulnerabilities and cies and action plans. The subnational entities are getting equitable development. Mr. Ali remained the Asia Direc- guidelines from established implementation framework for tor for Climate and Development Knowledge Network. He the NCCP at the federal level. At the federal level, it was has attached to a number of international organizations such indicated that the implementation committee will oversee as the Asian Development Bank, European Commission, the implementation of climate change actions in Pakistan. Packard Foundation, Rockefeller Foundation, and The Asia However, it does not mention that who has to do what and Foundation. Ali has vast experience in training and facili- what would be their authority with responsibility. Table 2 tating multi-sectoral and multi-disciplinary expert groups highlights the complete summary of this study. on policy planning, leadership development, and consensus building. The creation of the NCCP is fully encouraged by both the 5 Two Comparative Expert Views experts. They viewed that the NCCP should have been estab- on Motivation and Efcacy of the NCCP lished earlier keeping in view the Pakistan’s vulnerability to climate change. For instance, Kakakhel indicated that ‘… To validate the fndings that are produced in this study, Pakistan was way behind India and Bangladesh in creating semi-structured interviews were conducted with two policy a policy framework concerning climate change which had and climate experts in Pakistan to ascertain the motivations developed their national climate change policies in 2008.’ and efcacy of the NCCP, as a supplement to the policy The reason for this delay may well be considered in the con- analysis conducted by the author. These two experts were text of the security environment in the country following the chosen based on their involvement in global climate change advent of war in neighboring Afghanistan. policy as well as their intimate familiarity with the Paki- Regarding the process of the NCCP, the experts had stani context. Neither respondent was directly engaged in divergent opinions which are evaluated in terms of their the formulation of the policy itself to avoid any confict of embedded assumptions. Ali described that the stakehold- interest. The following two experts were consulted based ers’ identifcation was opportunistic rather than being sys- on these criteria: tematic in terms of the scope of various interest groups Shafqat Kakakhel is a senior retired Pakistani diplo- which needed to be consulted. Provincial engagement was mat. He served as the UN Assistant Secretary General and not undertaken with the level of care that was demanded Deputy Executive Director of the United Nations Environ- of a country with still fragile national political integration. ment Programme (UNEP). He was elected president of the However, both expert respondents were optimistic that the

Table 2 Complete summary of Criteria Fulflled Better but needs Not fulflled the study improvement

1. Accessability √ 2. Policy background √ 3. Goals √ 4. Resources √ 5. Monitoring and evaluation √ 6. Political and public opportunities √ 7. Obligations √

1 3 532 M. Mumtaz establishment of the NCCP was a positive initiative in so 6 Discussion far that it raised awareness of climate change policy; many of its weaknesses can be overcome by revisiting and revis- The summary of the analysis in Table 2 shows that major- ing the policy from time to time. ity of established criteria is met by the NCCP. Accessibil- Both expert respondents strongly negated the view ity to the policy is its major strength. However, the rest that the policy was created to get funds from donors. For of the criteria included the policy background, goals and example, Kakakhel explained that ‘it is unfair to attribute objectives, resources, monitoring and evaluation, political the formulation of the NCCP solely to a quest for donor and public opportunities, and obligations are needed to funds’. To him, the NCCP represents an important mile- revise for its improvement. stone in the development of Pakistan’s climate regime on The policy document is easily accessible for everyone the basis of the report of the TFCC in October 2008. He in soft and hard copy form but the inclusion of policy further argued that ‘the credit for the preparation of the background is needed for its improvement. Comprehensive NCCP should go to the civil society activists who persis- policy background is one of the key strengths for a policy tently called for an overarching policy framework on cli- as it provides complete backing for a policy document. In mate change…’. Likewise, Ali maintained that there is no case of the NCCP, it may lack scientifc backing. There evidence that the policy was framed with the motivations is no provision of any evidence in the NCCP to consult to secure funds from donor agencies. Both experts have any empirical research. Therefore, it is dire need for any same opinion that it is not justifed to say that the policy policy document to have statistical background and com- was established to get funds. prehensive goals and objectives for establishing achievable The implementation of public policies, especially, envi- policy measures. ronmental policies remained a challenge for Pakistan. It is Some of the policy’s objectives are not realistically important to know the progress for implementation of the established. Although, the objectives are highlighted in NCCP. The federal government established implementation the policy but the policy lacks to provide the backing for framework for the NCCP in 2013 and set priority, short- these objectives in the form of policy measures. The real- term-, medium-term-, and long-term actions. If you see the istic and doable goals and objectives along with resources priority actions which should have been taken, we cannot see are major strengths for an efective and optimal policy. any substantial achievements. Most of these actions are not Resources including fnancial and human resources are taken at any level. Ali described that these are ambitious tar- very important for an efective implementation of a policy. gets without any appropriate budget. He further maintained In case of the NCCP, it is highlighted that these resources that the federal government may work for framework but it are important for the policy and it proposes certain meas- does not need to set implementation targets for provinces ures to address them. Pakistan’s dangling economy and keeping in view addressing climate change is the responsi- allocation of weak budget to manage climate change is bility of the provinces. a point of concern for countering climate change in the Kakakhel responded that ‘…the NCCP and the frame- country. work for its implementation will remain unfulfllable wish Subsequently, monitoring and evaluation is emphasized lists unless they are elaborated in carefully prepared pro- in the policy and the hierarchy is established in the policy. jects or programs and are implemented through federal and This is the strength of the NCCP. The monitoring and eval- provincial governments and non-state stakeholders’. How- uation is of extreme importance for addressing the weak- ever, he is optimistic about implementation of the NCCP nesses of a policy especially in its implementation phase. describing that ‘We hope that the institutional mechanisms The proposed evaluation for NCCP is mainly based on enshrined in the landmark National Climate Change Bill internal evaluation. However, emphasis on external evalu- approved by the Parliament in March 2017, namely the Cli- ation is not described in the policy which could be the real mate Change Council, the Climate Change Authority and the test about success and failure of the policy. Climate Change Fund, would ensure the full and efective The policy document acknowledges the importance implementation of the NCCP as well as Pakistan’s Intended and involvement of stakeholders. There are various fed- Nationally Determined Contribution (INDC), thereby ena- eral and provincial departments along with some non- bling Pakistan to contribute to the implementation of the governmental organizations which are indicated in the Paris Agreement’. policy. However, there is no clarity as to who is the real The expert opinions tell us that the creation of the NCCP stakeholder. Second, the document lacks to provide the was the need of the hour for Pakistan taking an account its concerns of any stakeholder and appropriate measures for high vulnerability. However, certain weak aspects of the such concerns. Therefore, these weaknesses of the policy policy were discussed which need to be addressed to attain should be addressed on an urgent basis for its efective and desirable results of the policy.

1 3 The National Climate Change Policy of Pakistan: An Evaluation of Its Impact on Institutional… 533 long-lasting implementation. To address the weaknesses of relevant stakeholder while dealing climate change in the of policy, it is important because the subnational govern- country. Apart from multiple strengths of the policy, it also ments are getting valuable inputs from the NCCP, while has some shortcomings. Some policy measures are very devising provincial climate change policies and action generic in nature and they are not prepared with detailed plans. research and scientifc backing. For instance, the policy Further to this, with regard to obligations, the NCCP proposes to protect glaciers which seem quite complicated identifes the hierarchy for the implementation at federal, keeping in view the complexity of military confict between provincial and local levels. However, there is no indication Pakistan and India in the region. who will do what tasks and when it would be done. Keep- Moreover, it proposes some measures which are not prac- ing in view the signifcance of implementation plans of any tically implementable. It is not possible to utilize such meas- policy, there is a dire need to present the rational and spe- ures keeping in mind the dangling economy of Pakistan. The cifc implementation mechanism. In case of the NCCP, such policy claims that all relevant stakeholders were consulted. coherent and specifed mechanisms at the moment may be However, there is no indication that Pakistan’s military was lacking and which should be dealt with. also part of such consultation although, there is a strong The fndings are validated by the expert opinions. The linkage between national security and climate change. The experts are in the views that the NCCP is a positive move to NCCP is a promising initiative but it requires improvements. tackle climate change in Pakistan. However, they suggested The policy must be revised to address its shortcomings and that there is need to establish proper institutional arrange- to incorporate new strategies as per requirements to address ments, ensure involvement of related stakeholder, and to the current situation of climate change in the country. establish action plans at the right level. The federal Ministry of Environment was renamed as ministry of climate change in 2012. The important tasks of the ministry are to establish climate change policies, action 7 Conclusion and Final Remarks plans to counter climate change, and to deal climate change matters at international forum. It also coordinates between Pakistan is amongst the highly vulnerable countries to cli- inter- and intra-provincial governments in the context of cli- mate change; it is among the few developing countries who mate change and global warming. The ministry promotes have formulated their climate change policies. The estab- research and technical cooperation, and liaises with inter- lishment of the NCCP was a laudable step to efectively national donors/agencies. countering climate change in Pakistan. After 18th consti- Pakistan ofcially launched the NCCP in 2013. The min- tutional amendment in Pakistan, the implementation of cli- istry then created an action plan to implement the measures mate change policy and other related policies are the respon- and actions proposed in the NCCP. In 2013, the ministry sibility of subnational governments. However, the NCCP was degraded to a division. However, the government real- which is prepared at federal level plays an instrumental role ized that it requires a full ministry so as to move forward to provide a strong backing for establishing climate change for efective implementation of the NCCP and the ministry policies and action plans in the provinces. was reinstated in 2015. Therefore, the implementation of The NCCP was created at the national level while engag- the NCCP was a major motivation factor to reinstate the ing the respective provinces and considering the require- ministry. ments of all the provinces. Therefore, the respective prov- In 2017, the climate change act was passed which “will inces can acquire valuable policy measures from the NCCP fast-track measures needed to implement actions on the while devising their provincial climate change policies. The ground” in the country which is behind the efective imple- NCCP ofers a systemic monitoring and coordination mecha- mentation of climate actions. New institutional arrangements nism for provinces as the national government is responsible are set: Pakistan Climate Change Council, Pakistan Climate for dealing all the international treaties and legal obliga- Change Authority, and Pakistan Climate Change Fund. tions at international forums. Therefore, coordination is very The council is a decision-making body chaired by either important to present the case of Pakistan at the international the prime minister or a person nominated by him. The gov- level. ernment appoints federal and provincial ministers, chief It is important to note that Pakistan has a weak capac- ministers and chief secretaries as members of the council. ity to deal with climate change. The policy proposes some The Climate Change Authority is an autonomous govern- concrete and handful measures for promotion of institutional ment department composed of scientists, academics, indus- and human capacity. The policy is poised to enhance climate trialists, agriculturalists and serving and retired government research and scientifc evidences so that the innovative strat- servants, etc. The task of the authority is to formulate adap- egies can be incorporated for efective results. Moreover, it tation and mitigation policies and projects designed to meet strongly emphasizes citizen engagement and involvement Pakistan’s obligations under international climate accords

1 3 534 M. Mumtaz like the Paris Agreement. The implementation of the NCCP Lehtonen M (2005) OECD environmental performance review pro- is the driving force for all these new institutional arrange- gramme. Accountability (f)or learning? Evaluation 11:169–188 Leiserowitz AA (2005) American risk perceptions: is climate change ments in the country. dangerous? Risk Anal 25:1433–1442 Majone G (1977) Technology assessment and policy analysis. Policy Compliance with ethical standards Sci 8(2):173–175 Malik SM, Awan H, Khan N (2012) Mapping vulnerability to climate Conflict of interest It is declared that there are no conficts of interest change and its repercussion on human health in Pakistan. Glob with respect to this research, its authorships, and/or publication of the Health 2012(8):31 article. Martinuzzi A (2004) Sustainable development evaluations in Europe—market analysis, meta evaluation and future chal- lenges. J Environ Assess Policy Manag 6(4):411–442 Mickwitz P (2013) Policy evaluation. In: Jordan A, Adelle C (eds) Environmental Policy in the EU: Actors, Institutions and Pro- References cesses. Routledge, London, pp 267–286 Mickwitz P, Birnbaum M (eds) (2009) Environmental Program and Policy Evaluation: Addressing Methodological Challenges. Aftab E, Hickey GM (2010) Forest administration challenges in Paki- New Directions for Evaluation Number 122. San Francisco, stan: the case of the Patriata reserved forest and the new Murree CA: Jossey-Bass. development. Int For Rev 12:97–105 Morton JF (2007) The impact of climate change on small- Banoori W (2012). Pakistan: climate report. http://www.thefr​ontie​ holder and subsistence agriculture. Proc Natl Acad Sci rpost​.com/artic​le/16361​3. Accessed July 2017 104(50):19680–19685 Chelimsky E (2006) The purposes of evaluation in a democratic Mumtaz M (2013) National climate change policy of Pakistan: an society. In: Shaw IF, Greene JC, Mark MM (eds) Handbook analysis. MPhil thesis, National Defence University, Pakistan of evaluation: policies, programs and practices. Sage, London Nickerson RS (1998) Confrmation bias: a ubiquitous phenomenon Cheung KK, Mirzaei M, Leeder S (2010) Health policy analysis: in many guises. Rev Gen Psychol 2(2):175–220 a tool to evaluate in policy documents the alignment between Rasul G, Chaudhry Q, Mahmood A, Hyder K, Dahe Q (2011) Gla- policy statements and intended outcomes. Aust Health Rev ciers and glacial lakes under changing climate in Pakistan. Pak 34:405–413 J Meteorol 8(15):1–11 Crabbé A, Leroy P (2008) The handbook of environmental policy Singh BR, Singh O (2012) Study of impacts of global warming on evaluation. Earthscan’, London climate change: rise in sea level and disaster frequency. In: Ellahi A, Zaka B (2015) Analysis of higher education policy frame- Singh BR (ed) Global warming—impacts and future perspec- works for open and distance education in Pakistan. Eval Rev tive. ISBN 978-953-51-0755-2. Publication date: September 39(2):255–277 19, 2012 under CC BY 3.0 license. http://www.intec​hopen​ Germanwatch (2014) Global climate change risk index. German- .com/books​/globa​l-warmi​ng-impac​ts-and-futur​e-persp​ectiv​e. watch, Berlin Accessed Aug 2017 Haug C, Rayner T, Jordan A, Hildingsson R, Stripple J, Monni Sivakumar MVK, Stefanski R (2011) Climate change in South Asia. S et al (2010) Navigating the dilemmas of climate policy in In: Lal R, Sivakumar MVK, Faiz SMA, Mustafzur-Rahman Europe: evidence from policy evaluation studies. Clim Change AHM, Islam KR (eds) Climate change and food security in 101(3–4):427–445 South Asia. Springer, Netherlands, pp 13–30 Hilden M, Jordan A, Rayner T (2014) Climate policy innova- Sterrett C (2011) Review of climate change adaptation practices in tion: developing an evaluation perspective. Environ Politics South Asia. Oxfam research reports, Climate Concern, Australia 23(5):884–905 Toulemonde J (2000) Evaluation culture(s) in Europe: diferences Huitema D et al (2011) The evaluation of climate policy. Policy Sci and convergence between national practices. Vierteljahrsheft- 44(2):179–198 ezur Wirtschaftsforschung DIW 69(3):350–357 Kelkar U, Bhadwal S (2007) South Asian regional study on climate Urwin K, Jordan A (2008) Does public policy support or under- change impacts and adaptation: Implications for human devel- mine climate change adaptation? Exploring policy interplay opment. Human development report, United Nations Develop- across diferent scales of governance. Glob Environ Change ment Programme 18(1):180–191 Khan HR (2012) A review of national climate change policy of Paki- Vedung E (1997) Public policy and program evaluation. Transaction stan. SPO discussion paper series, pp 1–13. Available at SPO Publishers, New Brunswick National Centre 30-A Nazim-ud-din Road F-10/4 Islamabad Vedung E (2013) Six models of evaluation. In: Araral E, Fritzen S, 44000 Pakistan. http://www.spopk​.org. Accessed 15 July 2013 Howlett M et al (eds) Routledge Handbook of Public Policy. Khan MA, Samiullah M (2015) Financing for disaster risk reduction Routledge, London, pp 387–400 in Pakistan. In: Rehman AU et al (eds) Disaster risk reduction Walker BJ, Adger WN, Russel D (2014) Institutional barriers to cli- approaches in Pakistan, disaster risk reduction. Springer, Japan mate change adaptation in decentralised governance structures: Knaap GJ, Kim TJ (1998) Introduction: environmental program transport planning in England. Urban Stud 52(12):2250–2266 evaluation. Framing the subject, identifying issues. In: Knaap Wassmann R, Jagadish S, Sumfeth K, Pathak H, Howell G, Ismail A, GJ, Kim TJ (eds) Environmental program evaluation. A primer. Serraj R, Redona E, Singh R, Heuer S (2009) Chapter 3 regional University of Illinois Press, Urbana, pp 1–20 vulnerability of climate change impacts on asian rice production Kraft ME, Furlong SR (2010) Public policy Politics, analysis, and and scope for adaptation. In: Advances in agronomy, vol 102. alternatives. CQ Press, Washington Academic Press, pp 91–133 Kurosaki T, Khan H, Shah MK, Tahir M (2011) Natural disasters, Weiss CH (1993) Where politics and evaluation research meet. Eval relief aid, and household vulnerability in Pakistan: evidence Pract 14(1):93–106 from a pilot survey in Khyber Pakhtunkhwa. PRIMCED discus- Weitzman ML (2009) On modeling and interpreting the economics sion paper series 12, Hitotsubashi University of catastrophic climate change. Rev Econ Stat 91(1):1–19

1 3 The National Climate Change Policy of Pakistan: An Evaluation of Its Impact on Institutional… 535

Wilson E (2006) Adapting to climate change at the local level: the Muhammad Mumtaz is a PhD research scholar in Public Administra- spatial planning response. Local Environ 11(6):609–662 tion & Government at Fundação Getulio Vargas, Sao Paulo, Brazil. His Yusuf S (2011) Pakistan’s frst climate change policy ready. http:// PhD study investigates climate adaptation governance at subnational beta.dawn.com/news/63177​9/Pakis​tans-first​-clima​te-chang​ level in Pakistan. His area of research is climate governance at national, e-polic​y-ready​. Accessed Nov 2017 subnational, and local level.

1 3 Earth Systems and Environment (2018) 2:553–562 https://doi.org/10.1007/s41748-018-0065-7

ORIGINAL ARTICLE

Vulnerability Assessment of Formal and Informal Credit Borrowers: In Flood Prone Zone of Punjab, Pakistan

Ghamz E Ali Siyal1 · Muhammad Zahid Siddique1 · Syed Mohsin Ali Kazmi2

Received: 6 April 2018 / Accepted: 30 July 2018 / Published online: 29 August 2018 © Springer Nature Switzerland AG 2018

Abstract In Pakistan, farmers are facing climatic and non-climatic challenges in retaining their output and proft. Their dependence on credit is increasing with passage of time. Previously, the factors behind availing credit policy are discussed in diferent studies, but there is no study found that empirically compares vulnerability of formal and informal farmers on the basis of credit facility in food prone zone. For estimation and analysis, quantitative and qualitative measures are used. The quantitative measures are based on estimating vulnerability indices. The data were collected through survey of 146 households through multistage-stratifed random sampling technique. The qualitative data focus group discussion which helps in sampling and investigate factors behind vulnerability of farmers. Results suggest that formal credit borrowers are more vulnerable than informal credit borrowers in terms of higher exposure and sensitivity and lower adaptive capacity. Formal credit borrow- ing farmers are afected by inefciencies namely, late declaration of calamity, defective loss assessment, slow procedure of claim disbursement, meager amount of claims, political infuence in listing of disaster afected village, and delayed credit payment operations. The higher vulnerability of formal credit borrowers raises need for overall improvement of credit policy of Crop Loan Insurance Scheme.

Keywords Vulnerability · Floods · Formal and informal credit policy

1 Introduction 2013; Chen 2011). Because a large segment of population in developing countries is dependent upon income generated by The agriculture sector of almost all developing countries agriculture farming (Choudhury et al. 2015), poverty rates has remained signifcantly afected by increasing uncertainty are triggering by these uncertainties. of crop from sowing to harvest. The major causes of this The agriculture sector is the backbone of Pakistan’s econ- uncertainty are natural extreme events, epidemics, shortage omy; it remained the major source of income for 80% of of irrigated water, unavailability of pesticides, lack of access small landholders (Ali 2013). Contribution of agriculture to credit, lack of information about farming practices due to comes from crops (31% major and 11% from minor crops) the absence of research institutes in their communities and and livestock (Syed Ali Raza et al. 2012). lack of access to basic facilities like health, and educational In case of Pakistan, climate change has become a threat, facilities (Rahman et al. 2012; Raju and Chand 2008; Ali especially for agriculture sector (Abid et al. 2015), because performance of this sector is dependent on suitable weather conditions. These climate conditions are responsible for Electronic supplementary material The online version of this afecting production and prices of agriculture which infu- article (https​://doi.org/10.1007/s4174​8-018-0065-7) contains ences economic growth (GoP 2017). The farming commu- supplementary material, which is available to authorized users. nity faces climate-related risks, namely, extreme minimum * Ghamz E Ali Siyal and maximum temperature, diseases in animal diseases, pest [email protected] attack, and changing crop time periods, and water shortages. These outcomes of climate change have resulted to lower 1 National University of Sciences and Technology (NUST), crop and livestock output (Abid et al. 2015). Islamabad, Pakistan Since the 1950s, Pakistan has faced 24 food disasters 2 Sustainable Development Policy Institute (SDPI), Islamabad, resulting in severe losses of US $ 18 billion (GoP-INDC Pakistan

Vol.:(0123456789)1 3 554 G. E. A. Siyal et al.

2016). The trends of climate change and increased fre- governments to help out farmers. However, issues of limited quency of climate-induced extremes events (like, foods and access to credit and program design, sometimes, do not lead droughts) have increased agriculture losses, i.e., 10.63 mil- to positive impacts. The major reasons for limited access to lion acres of crops destroyed including 38.12 million people credit are collateral issues, and high markups which become afected, 3.45 million houses damaged, food security, rising barrier for availing formal credit in Pakistan (Akram et al. sea level, and erosion of coastal zones (GoP 2016). These 2008). It is one of the barriers for having formal lending. disasters were treated by allocating 5.8 and 7.6% of total In the light of above literature review, signifcance of expenditure in the federal budget in 2015 (GoP-INDC 2016). credit policy cannot be denied. It has already been estab- The Government of Pakistan (GoP) has actively involved lished that formal credit policy is more restrained by institu- institutional reforms by launching Federal Flood Commis- tional standard of procedures (sops) as compared to informal sion (FFC) in January 1977 (GoP 2018) and National Disas- credit sources. However, this study flls gap left out by the ter Management Authority (NDMA), and National Disaster previous studies. This study not only compares vulnerability Management Commission (NDMC) in 2010 (Zeeshan and assessment of formal and informal borrower in context of Khan 2015) (Table 1). food prone zone but also investigates factors behind inef- Along with institutional reforms, Mandatory Crop Loan fectiveness of national credit policy, i.e., Crop Loan Insur- Insurance Scheme (CLIS) was launched in 2008 by Ministry ance Scheme (CLIS). These factors help to fnalize aims of of Finance, GoP. The main objective of this policy was to this study, smooth out the income of farmer by provision of fnancial The aim of study is to assess the vulnerability of formal support against natural disasters (SBP 2014a, b). Govern- and informal borrowers in food prone zones of Pakistan. A ment of Pakistan is working on the formulation of National comparison is made between farmers who availed Crop Loan Agriculture Insurance Scheme (NAIS) to support farming Insurance Scheme (CLIS)—both from formal and informal community to increase their resilience against extreme sectors. This comparison helps us to estimate vulnerability events (like foods) through climate change adaptation strat- of farmers living in the same hotspot area. egies mainly crop insurance without loan criterion (Gop 2017). Because, climate change shock afects ecological 1.1 Objectives conditions in rural areas, especially agriculture productiv- ity shifts. It leads to harming income of poor and marginal 1. To conduct vulnerability assessment of formal and infor- farmers (Saeed et al. 2015). mal farmers residing in food prone zones of Punjab, The major adaptations performed by farming commu- Pakistan nity are altering crop types, varieties, and planting dates. 2. To determine factors behind vulnerability of formal and However, these adaptations are facing constraints like lack informal farmers residing in food prone zones of Pun- of fnancial resources, constrained access to provision of jab, Pakistan institutional services like, credits, farm inputs, machinery, marketing services, weather forecasting, and information (Abid et al. 2015). Particularly, the subsistence and small 2 Data and Methodology farmers’ are facing severe limitations to aford basic inputs of cropping, namely, high quality seeds, sufcient fertilizer, 2.1 Vulnerability Index and farm implements. These farmers have lower income, lower savings, and lower capital formations (Saboor et al. The expected utility theory describes about theoretical 2009). The farmers need fnancial support for dealing with framework of study. This theory was given by John Von their social needs, purchasing farm inputs, and enhancing Neumann and Oscar Morgenstern in year 1944 that discus crop productivity. Credit plays a pivotal role in agriculture, about the development of expected utility model. This theory as it not only supports small farmers but also medium and based on decision making to deal with risky situation, like, large farmers for improving farm output. The access to credit prizes and lotteries, etc. (Levin 2006). The farming com- is becoming challenging for farmers because of its limited munity, like other individuals, makes decisions to minimize availability of formal credit to farmers (Saqib et al. 2018). risk and deal with challenges related to production, market- In fact, these farmers need credit for two reasons: produc- ing, etc. These decisions help to understand their inclination tion and development. The credit for production purpose is towards adaptation policies (Nantui et al. 2012, Barry 1984). used for purchasing inputs namely, seeds, fertilizers, pesti- The challenge of climate change infuences farmers’ decision cides, water (‘Abiyano’), and fuel charges. The microfnance for adaptations. However, these adaptations are subject to programs increase welfare of participating households, espe- socioeconomic factors, farm characteristics, and changes in cially credit policies are generally designed and supported by climate factors (Deressa et al. 2008). To measure adaptive

1 3 Vulnerability Assessment of Formal and Informal Credit Borrowers: In Flood Prone Zone of Punjab,… 555 capacities of communities facing climate risk, vulnerability 2.2 Construction of Livelihood Vulnerability Index assessment tool is considered to be an efective tool (Hahn et al. 2009). This tool comprises of three components, The vulnerability indices were based on the review of namely, exposure, sensitivity, and adaptive capacity (Ibid). numerous research studies and expert opinions during con- The concept of vulnerability assessment is a new and structing it. Some of important studies used are namely, assessment practice evolved to measure impact of climate Hahn et al. (2009), Panthi et al. (2016) and Salik et al. change. It has been used for issues such as livelihoods, food (2015). The defnition for vulnerability is based on IPCC’s security, natural hazards, global environment change, cli- defnition which defnes it as comprised of exposure, sensi- mate change, and others (Füssel and Klein 2006). The most tivity, and adaptive capacity. widely quoted defnition of vulnerability is given by IPCC These subcomponents of exposure, sensitivity and adap- assessment report, as, “it is an integrated measure of the tive capacity are defned by variable that are supported from expected magnitude of adverse efects to a system caused the literature review. The variable selection under these sub- by a given level of certain external stressors” (McCarthy components with their literature support is given below in et al. 2001). Table 2. The alternate studies used diferent frameworks for esti- To formulate index, we need to convert variables from mating Vulnerability Index of communities. The study of diferent scales to single scale. This process is known as Antwi et al. (2015) estimates community vulnerability of normalization of variables for using in the index. The basic four districts of Ghana. The Vulnerability Index was com- step for normalization is the min–max normalization tech- prised of four subcomponents like engineering, socio-eco- nique (Poverty and Human Development Initiative 2011): nomic, ecological, and political, and these subcomponents S S Index = di min . were further comprised of other indicators. Another study sd S − S (1) of Adeyemi (2014) also estimates vulnerability of farmers max min to food disaster in Oyo State, Nigeria. The Vulnerability Each subcomponent of index is normalized through above Index was based on MOVE framework which estimates the formula given in Eq. (1). Here, Sdi is defned as the original subcomponents of socio-economic characteristics, capaci- value of a variable for each respondent, and Smin and Smax ties, food exposure, and susceptibility. refect the minimum and maximum values of overall vari- There is another type of framework developed for estimat- able, respectively. Second step for the calculation of index ing impact of vulnerability on livelihoods of people or com- is method of aggregation to fnd out the values of expo- munities. It is known as sustainable livelihood framework sure, sensitivity, and adaptive capacities using the following which comprises of assets of household, namely, natural, equation: social, fnancial, physical, and human capital (Chambers and n Conway 1992). The combination of exposure, sensitivity and ∑ wpiPdi CF = i=1 . adaptation strategies by household becomes an efective tool d n (2) ∑i=1 wpi for measuring risk due to climate change (Hahn et al. 2009). The study of Hinkel (2011) elaborates that vulnerability Each subcomponent of Vulnerability Index is aggregated assessment has become common tool for academic research- through above formula given in Eq. (2). Where, CF­ d is the ers, policy makers, and organizations. The common pur- contributor factor like; exposure, sensitivity, and adaptive poses of vulnerability assessment are six problems, namely, capacity. Where, wpi is the weightage of one of the major identifcation of mitigating targets, second, identifcation contributing factor and Pdi is the major component for dis- of vulnerable people, community, or regions, third, raising trict d indexed by ‘i’ (Panthi et al. 2016). In the above for- awareness, fourth, allocation of adaptation funds, ffth, mon- mula, ‘n’ refects the number of major components in each itoring of adaptation policy, and sixth, conducting scientifc of the contributing factors. Similar aggregation occurs for research. Among all aforementioned problems, only identi- Livelihood Vulnerability Index which comprises of socio- fcation of vulnerable people, communities, and regions is demographics, credit, climate variability, food, water, liveli- most considerable. Therefore, policy makers should rightly hood strategies, health, social networking, and natural dis- introduce variables in vulnerability assessment. However, asters components. After obtaining the values of exposure, as described in the study of Hahn et al. (2009), it has been sensitivity, and adaptive capacities using the above equa- used in the variety of circumstance, for example, analyzing tion, we applied the following IPCC formula to obtain the the early warning system, mapping for targeting food aid, results of IPCC-Livelihood Vulnerability Index (IPCC-LVI). and other issues of poverty, health status, biodiversity, and IPCC-LVI ranges from − 1 (least vulnerable) to 1 (most globalization. vulnerable): IPCC-LVI =(Exposure-adaptive capacity)×Sensitivity.

1 3 556 G. E. A. Siyal et al. ), Salik et al. ( 2015 ) ( 2009 ), Salik et al. Salik et al. ( 2015 ) Salik et al. Salik et al. ( 2015 ) Salik et al. Salik et al. ( 2015 ) Salik et al. Panthi et al. ( 2016 ) et al. Panthi ), Salik et al. ( 2015 ) ( 2014 ), Salik et al. Porter et al. Panthi et al. ( 2016 ) et al. Panthi Salik et al. ( 2015 ) Salik et al. Salik et al. ( 2015 ) Salik et al. Hahn et al. ( 2009 ) Hahn et al. Pachauri et al. ( 2014 ) et al. Pachauri Hahn et al. ( 2009 ) Hahn et al. Adegible Panthi et al. ( 2016 ) et al. Panthi Adegbile (2014) Adegbile ), Hahn et al. ( 2016 ), Hahn et al. et al. Panthi References Panthi et al. ( 2016 ) et al. Panthi exposure exposure risk of droughts/water shortage leads to risk which of droughts/water enhanced exposure the risk yield crop to adaptive capacity adaptive capacity to crop yield crop to land Higher the range of temperature higher will be of temperature Higher the range Higher the range of temperature higher will be of temperature Higher the range Higher the frequency of dryHigher the frequency months higher will be Increase in theIncrease of precipitation variability increases Higher the age dependency ratio, lower will be the lower dependency ratio, Higher the age Higher the frequency higher will be theHigher the exposure frequency Higher the frequency, higher will be the exposure Higher the frequency, Higher the experience, higher will be adaptive Higher the higher will be adaptive experience, Higher the will be the loss, more sensitivity Increase in temperature variability increases the increases variability risk in temperature Increase Higher the value, higher will be sensitivity of farm higher will be sensitivity of farm Higher the value, Higher food insecurity, higher will be sensitivity insecurity, Higher food Increase in distanceriskIncrease reduces of being exposed Higher cost implied, higher sensitivity implied, Higher cost More frequent events leads to higher exposure leads to events frequent More relationship Functional Higher the health issues, higher sensitivity 10 °C month during precipita - spring season having tion < 5 mm and in summer precipitation = 0 mm school year of age over the population between 15 and the population between over of age year of age 64 years with temperature above 30 °C above event food staf during climate extremes food household in the last 10 years (based on survey) 10 years in the last Monthly average diurnal range min. temperature average Monthly Monthly average diurnal range min. temperature average Monthly Number of extreme dry as the months of extreme explained Number Percentage of household who did not attended of household who did not Percentage Monthly variability in total precipitation in total variability Monthly Ratio of the population 0–14 and 65 and above of theRatio population 0–14 and 65 above Number of extreme cold months (below − cold months (below of extreme Number Number of extreme hot months hot with temperature of extreme Number Number of years involved in agriculture involved of years Number Monthly max temperature variability max temperature Monthly Average monetary loss incurred extreme last due to Average Monthly min temperature variability min temperature Monthly Average time impact on farmers farm by water by farm on farmers time impact Average Percentage of household having limited supply of limited supply of household having Percentage Distance between river and land river Distance between Average health related cost per month for the per month health cost for related Average Average number of frequencies of climate extremes of climate extremes number of frequencies Average Explanation Households having disability in family Households having Components of Vulnerability Index of Vulnerability Components staf during climate extremes since the past 3 months?since the past Education Dependency ratio Experience of agriculture Household having limited supply of food of food limited supply Household having Monetary loss Number of days food water stayed on land stayed food water of days Number Any health issue/disability in your family family health issue/disability in your Any Cost on health Cost issues Diurnal max. temperature Diurnal min. temperature Extreme dryExtreme months Precipitation variability Cold months Hot months Hot Adaptive capacity Adaptive Min temperature variability Min temperature Min temperature variability Min temperature Near to river to Near Sensitivity 1 Table Exposure Climate extremes/disaster and subcomponents Indicators

1 3 Vulnerability Assessment of Formal and Informal Credit Borrowers: In Flood Prone Zone of Punjab,… 557 Panthi et al. ( 2016 ) et al. Panthi Adegible Panthi et al. ( 2016 ) et al. Panthi Chikaire et al. ( 2016 ) et al. Chikaire Arshad et al. ( 2016 ) Arshad et al. Hahn et al. ( 2009 ) Hahn et al. Hahn et al. ( 2009 ) Hahn et al. Hahn et al. ( 2009 ) Hahn et al. Panthi et al. ( 2016 ) et al. Panthi Panthi et al. ( 2016 ) et al. Panthi References capacity adaptation toward crop insurance crop adaptation toward toward crop insurance crop toward providing support in the form of physical help, support of physical in theproviding form sharing and experience information, etc. capacity of hazards and preparedness which ultimately ultimately which and preparedness of hazards capacity contributes adaptive to Higher borrowing higher will adaptive capacity higher will adaptive Higher borrowing Higher visits higher will be adaptive capacity Higher visits higher will be adaptive Support from government will increase adaptive adaptive will increase Support government from More awareness of crop insurance, more chances of chances more insurance, of crop awareness More More willingness more chances of adaptation chances willingness more More Higher the foods higher chances of selling assets Higher the foods higher chances Social networking enhances adaptive capacity by capacity by enhances adaptive Social networking More diversity of livestock, higher will be adaptive higher will be adaptive of livestock, diversity More Higher crop diversity will increase adaptive capacity adaptive will increase diversity Higher crop Communication media helps to enhance awareness enhance awareness Communication media helps to Income diversifcation increases adaptive capacity adaptive increases Income diversifcation Functional relationship Functional past 5 years past culture extension ofcer extension culture calamity like food calamity like crop insurance crop in: local politics, social relief work, and local-level and local-level work, in: local politics, social relief friendship a strong circle associations, and have season Internet, and TV source of income source Borrowing credit to cope with any type of shock in type of shock cope with to credit Borrowing any Percentage of people who agreed for visits of agri of people who agreed- for Percentage Received any government support after government natural any Received Awareness can increase chances of availing CLIS of availing chances can increase Awareness More willingness more chances of adaptation toward of adaptation toward chances willingness more More Selling assets for adjusting with adjusting food shocks for Selling assets Percentage of households who are actively involved involved actively of households who are Percentage Households having more than one livestock type than more one livestock Households having Households planting more than in single Households planting more one crop Percentage of households having access to phone, access to of households having Percentage Percentage of households who have more than more one of households who have Percentage Explanation Took loan Took Visits of agriculture ofcer Visits extension Other government support Other government Awareness of crop insurance in CLIS insurance of crop Awareness Willingness of CLIS Willingness Selling assets during Selling assets emergencies Social networking Livestock diversity Livestock Crop diversity Crop Access to media/information to Access Income diversifcation 1 Table (continued) and subcomponents Indicators

1 3 558 G. E. A. Siyal et al.

2.3 Data Collection Scheme are 66 in number. The farmers of both categories have diferent land size. Both farming categories include 2.3.1 Survey Area rich and poor farmers. The rationale behind it was to gen- eralize reasons behind vulnerability between formal and Over a past decade, district Sargodha has faced riverine informal borrowers. The time period of survey is 1 month, fooding through Jhelum and Chenab Rivers; particularly for i.e., during March 2017. During survey, farmers were busy years 2010, 2011, and 2014 as per annual food commission in cultivating kharif crop. report (Siyal 2018). This study has considered most recent A multistage stratifed random sampling has been used foods year of 2014 for Sargodha. to draw the sample size. Stratifcation is most suitable This study comprises of primary data analysis which is technique, especially when heterogeneity and sample based on the feld survey of food afected villages of Sar- biases prevail in the data. This is the best suited sam- godha district. The feld survey was based on multistratifed pling technique. Similarly, studies of Saqib et al. (2018) random sampling technique. In the frst stage, ‘Sargodha’ used multistage-stratifed random sampling technique for district was selected based on availability of data. In the sec- increasing efectiveness of targeted population. At the frst ond stage, out of seven tehsils (Behra, Bhalwal, KotMomin stage, the survey team selected food afected three tehsils Sargodha, Shahpur, Silanwali, and Sahiwal), three (Behra, and villages of Sargodha district. In the next stage, formal Sargodha, and Sahiwal) were randomly selected. Out of each borrowers and informal borrowers were selected through three tehsils, one village was randomly selected, namely, focus group discussion to identify targeted farming com- Aqil Shah Village, and Lakhiwaal Village, and Sheikh- munity. Then, those farmers were randomly selected to purkona. In the third stage, listing of formal and informal conduct survey. farmers was done through focus group discussions. Out of Along with primary survey of farmers, three focus group listing, we randomly selected 146 farmers, in total. This discussions with farmers were conducted in aforementioned survey took almost 3 months, from January to March 2017 villages. These focus group discussion had more than eight (Fig. 1). farmers which were inquired about overall credit policy and vulnerabilities faced by foods. 2.3.2 Cropping Profle in Sargodha District

The season of Rabi crop ranges from October to Decem- 3 Results ber months and harvested in months of March–April. The majority of farming community grows ‘Wheat’ as their The scale for Livelihood Vulnerability Index is comprised of major Rabi crop. Other common crops are oranges, veg- 0–1 values. Here, ‘0’ shows least vulnerability and ‘1’ shows etables, and fodder. The farmers have started a new trend highest vulnerability (Fig. 2). of cropping of ‘Rose’ plants during Rabi season. The chart of VI tells us about Vulnerability of farmers During foods, kharif crops faced severe damages dur- who have taken loan “Loanee” and who have not taken loan ing their harvesting. The season of Kharif season starts from formal source, i.e., “Non-Loanee or Informal loan bor- from April–May and harvested in August–September. The rower”. The Vulnerability Index is comprised of three com- majority of farming community cropped ‘sugarcane’ as ponents, namely, exposure, sensitivity, and adaptive capac- their major crop in kharif season, and other crops sown ity. The results show that exposure and sensitivity of loanee are cotton, fodder, and rice. The watermelon is a new crop is higher than non-loanee, but adaptive capacities of loanee trend introduced recently which is sown by a few farmers. are lower than non-loanee (Table 3). The Kharif crops get severely damaged by foods when The empirical results of Vulnerability Index are based it is about to harvest. Sometimes, fooding along with on the formula of (E + S (1 − A)) in which ‘E’ shows expo- monsoon rainfall, like food of year 2010, afects Rabi sure, ‘S’ shows sensitivity, and ‘A’ shows adaptive capacity. crops also during their sowing season. The overall result of Composite Vulnerability Index (CVI) Out of total sample of 146 farmers, the category which shows that non-loanee farmers are less vulnerable than loa- availed Crop Loan Insurance Scheme (CLIS) is ‘80’ which nee farmers. is termed as ‘formal loanee‘. The loanee farmers borrowed On the basis of Comer et al. (2012) and Hammill et al. loan against major crops (Wheat, Cotton, Sugarcane, and (2013), we can see that component of exposure of loanee Rice). We considered loan borrowers who took loan with and non-loanee farmers is low as its values laid between in past 5 years. The crop loan is taken for major crops 0.0 ≤ CVI ≥ 0.3. Similarly, component of sensitivity of (wheat, rice, sugarcane, and cotton) by farmers in district loanee and non-loanee farmer is showing medium level of Sargodha. Similarly, the term ‘Informal loan borrowers’ sensitivity. Finally, adaptive capacity component of loanee are farmers who are not availing Crop Loan Insurance shows medium level of adaptive capacity but for non-loanee

1 3 Vulnerability Assessment of Formal and Informal Credit Borrowers: In Flood Prone Zone of Punjab,… 559

Table 2 Construction of Components Formal borrowers Informal borrowers Livelihood Vulnerability Index Source: Authors’ calculation Socio-demographic profle 0.45225 0.4855 Livelihood Strategies 0.468 0.532 Social network 0.7208 0.2782 Health issues 0.36 0.3566 Food 0.539 0.461 Water 0.641 0.661 Natural disaster 0.296 0.251 Credit 0.462 0.537 Climate change 0.2978 0.2978 Sum 4.23685 3.5623 Livelihood Vulnerability Index (LVI) 0.47076 0.39581

has high adaptive capacity. The overall values of Composite vulnerable) to 1 (most vulnerable). The results in the above Vulnerability Index (CVI) show the medium level of vulner- table show that loanee farmers are more vulnerable than ability for farming categories, loanee and non-loanee. non-loanee against foods even after availing CLIS. This result was supported by the previous Composite Vulner- 3.1 Inter‑Governmental Panel on Climate Change ability Index (CVI). (IPCC) Vulnerability Index 3.2 Livelihood Vulnerability Index (LVI) The IPCC Vulnerability Index is calculated through formula of (exposure-adaptive capacity) × sensitivity. The result of The Livelihood Vulnerability Index (LVI) is based on socio- IPCC Vulnerability Index is given below (Table 3). demographic profle, livelihood strategies, social network, The scale for IPCC Vulnerability Index is given in the health issues, food, water, natural disasters, and credit and paper of Hahn et al. (2009). The scale ranges from − 1 (least

Fig. 1 Data site map Source: Map constructed by authors

1 3 560 G. E. A. Siyal et al. climate change (Fig. 3). The values of each component are Pakistan. The data site of this study is three food prone given in Table 4. tehsils of Sargodha district, namely, Behra, Sahiwal, and Formal borrowers are having higher livelihood vulner- Shahpur. This study conducted followed multistratifed sam- ability than informal borrowers. The index of Composite pling technique focus group discussion in aforementioned Vulnerability Index (CVI) shows that informal loan bor- study area to identify targeted farming community. After rowing farmers are less vulnerable on basis of Livelihood that, this paper, the results of vulnerability indices show Vulnerability Index. On the basis of assets categories, formal that Formal Credit borrowers (farmers) are more vulner- loan borrowing farmers are weaker on the basis of Socio- able as compared to informal credit borrowing farmers. demographic profle, livelihood strategies, water resources, The reasons for their higher vulnerability are delayed credit and credit facilities. Particularly, adaptive capacity of com- approval duration, late declaration of calamity, inefective ponent of Composite Vulnerability Index (CVI) highlights loss assessment, slow procedure of claim disbursement, reasons of lower adaptive capacities of formal borrowers. meager amount of claims, and political infuence in listing The subcomponents are namely, years of experience in agri- of disaster afected village. The study shows that the farm- culture, access to phone/TV/Internet, crop and livestock ing community is, in general, well informed of the relevant diversity, social activeness, willingness for crop insurance, climatic changes that afect their business prospects. They and government support. confrmed that they had observed climatic changes and dis- The reasons for higher vulnerability of loanee farmers asters in last 10 years. These climatic changes include irreg- can be understood from focus group discussions (FGDs), as ular rainfall, decrease in rainfall, increase in temperature, given in the following table. changes in number of hot and cold days, frequent climate Claim receiving duration: formal credit borrowers (farm- extremes, and more humidity. The farming community also ers) described delayed payment of insurance claims; almost noted increase in climate-induced disasters namely, foods, 6 months to 1 year. Due to delayed payments, farmers faced hailstorms, and pest attack. The major outcomes of climatic issues in fnancing of next crop. Farmers used personal sav- changes are decrease in crop productivity or quality, increase ings, took another loan, or sold household assets which exac- in chilling temperature, increase in pest attack, increase in erbate fnancial stress on them. day light for crops, and decrease in water for cropping. In Ineffective loss assessment: formal credit borrowers addition, outcomes of foods are crop destruction, prevalence (farmers) described that amount of claims was not sufcient of water borne diseases, increase in salinity, damage of road/ as compared to losses which they faced. Mostly, insurance infrastructure, shortage of food items, death of livestock, companies do not conduct individual loss assessment of spread of contaminated water, spread of disease in livestock, farm. Farmers suggested that government should ask insur- physical damage of settlement, and decline in crop yield. ance companies to provide us either 50% of loss recovery or 50% of input cost recovery through reimbursement. 4.1 Recommendations Crop loss: formal credit borrowers (farmers) described that crop loss was higher than claim amount. Formal credit On the basis of research fndings and discussions, it is clear borrowers (farmers) described that chances of loss for wheat, that there is need for improvement in this policy which will cotton, and vegetable crop were 100%. However, for rice, not only reduce vulnerability and improve resilience of farm- it was 70% chances of loss, and for sugarcane, it was 50%. ers against climate-induced disasters (especially, foods) but Apart from major crops, neither vegetables nor oranges were also efectively allocate subsidized incentive to right farmer. insured against disasters under CLIS. Therefore, this study recommends that: Political infuence: farmers from both categories; formal and informal loanee considered political infuence was also a major factor in declaration of disaster afected villages. Source: focus group discussions (FGDs) and key informant interviews.

4 Conclusion and Policy Recommendations

The signifcance of credit cannot be denied under rising challenges to farming community in Pakistan (Abid et al. 2015, Saboor et al. 2009; Saqib et al. 2018). This study conducted the vulnerability assessment of formal and infor- mal credit borrowing farmers in the food prone zones of Fig. 2 Composite Vulnerability Index Source: Author’s fndings

1 3 Vulnerability Assessment of Formal and Informal Credit Borrowers: In Flood Prone Zone of Punjab,… 561

Table 3 Vulnerability Index Exposure Sensitivity Adaptive capacity (composite) Source: Authors’ estimations Lonaee 0.2881 0.3524 0.4729 Non-loanee 0.2781 0.3376 0.5516 Composite Vulnerability Index Formal CVI = 0.3892 Informal CVI = 0.3547 (E + S (1 − A))

Fig. 3 Livelihood Vulnerability Index Source: Author’s fndings

Table 4 IPCC Vulnerability Index Source: Authors’ estimations References

IPCC Vulnerability Index Abid M, Schefran J, Schneider UA, Ashfaq M (2015) Farmers’ per- Formal loan borrowers − 0.06512352 ceptions of and adaptation strategies to climate change and their Informal loanee borrowers − 0.0923336 determinants: the case of Punjab province, Pakistan. Earth Syst Dyn 6(1):225 Adeyemi AO (2014) Vulnerability of farmers to food disaster in Akinyele Local Government Area, Oyo State of Nigerai. West • Government should expedite procedure of Crop Loan African Science Service Centre on Climate Change and Adapted Land Use. https://www.wasca​ l.org/speci​ al-pages​ /fle-show/conta​ ​ Insurance Scheme (CLIS) which will help in conducting iner/downl​oad/?tx_ggfl​elibr​ary_pi1%5Bfl​e%5D=132&cHash​ loss assessment timely and efectively. =44308​543e9​f2567​eefa8​7afb7​42d4d​79. Accessed 27 Dec 2017 • Government should also consider declaration of other Akram W, Hussain Z, Sial MH, Hussain I (2008) Agricultural credit natural calamities like excessive rain, hailstorm, and constraints and borrowing behavior of farmers in rural Punjab. Eur J Sci Res 23(2):294–304 pests attack under CLIS along with foods. Ali A (2013) Farmers’ willingness to pay for index based crop insur- • Introducing crop insurance policy without criteria of loan ance in Pakistan: a case study on food and cash crops of rain-fed at least in hotspot areas to willing farmers. areas. Agric Econ Res Rev 26(2):241–248 • Increasing the amount of claims by covering either 50% Antwi EK, Boakye-Danquah J, Owusu AB, Loh SK, Mensah R, Boafo YA, Apronti PT (2015) Community vulnerability assessment of input cost of cropping or 50% of loss faced by farmers. index for food prone savannah agro-ecological zone: a case study • Initiation of weather or yield-based index rather than of Wa West District, Ghana. Weather Clim Extremes 10:56–69 indemnity-based crop loan insurance. It will speed up Arshad M, Kächele H, Krupnik TJ, Amjath-Babu TS, Aravindak- the process of efective and individual losses assessment shan S, Abbas A, Müller K (2016) Climate variability, farmland value, and farmers’ perceptions of climate change: implications to provide claims before cropping of next season. for adaptation in rural Pakistan. Int J Sustain Dev World Ecol 24(6):532–544. https://doi.org/10.1080/13504​ 509.2016.12546​ 89​ Barry PJ (ed) (1984) Risk management in agriculture. Iowa State Uni- Compliance with ethical standards versity Press, Ames Chambers R, Conway G (1992) Sustainable rural livelihoods: practical concepts for the 21st century. Institute of Development Studies Conflict of interest On behalf of all authors, the corresponding author (UK), London states that there is no confict of interest.

1 3 562 G. E. A. Siyal et al.

Chen YT (2011) Weather index-based rice insurance: a pilot study of Panthi J, Aryal S, Dahal P, Bhandari P, Krakauer NY, Pandey VP nine villages in Zhejiang province, China. ETH Zurich, Zurich (2016) Livelihood vulnerability approach to assessing cli- Chikaire JU, Tijjani AR, Abdullahi KA (2016) The perception of rural mate change impacts on mixed agro-livestock smallholders farmers of agricultural insurance as a way of mitigation against around the Gandaki River Basin in Nepal. Reg Environ Change climate change variability in Imo State, Nigeria. Int J Agric Policy 16(4):1121–1132 Res 4(2):17–21 Porter J, Dessai S, Tompkins E (2014) What do we know about UK Choudhury A, Jones J, Okine A, Choudhury RL (2015) Drought trig- household adaptation to climate change? a systematic review. gered index insurance using cluster analysis of rainfall afected by Clim Change 127:371–379 climate change. J Insur Issues 39(2):169–186 Poverty O, Human Development Initiative (2011) Training material Comer PJ, Young B, Schulz K, Kittel G, Unnasch B, Braun D, Ham- for producing national human development reports: the Human merson G, Smart L, Hamilton H, Auer S, Smyth R, Hak J (2012) Development Index (HDI) Climate change vulnerability and adaptation strategies for natu- Rahman Z, bin Dost MK, Wassan AA (2012) Socio-economic prob- ral communities: piloting methods in the Mojave and Sonoran lems of farmers in Pakistan: focused area Tauluka Pano Akil Dis- deserts. Report to the U.S. Fish and Wildlife Service. Nature trict Sukkur. Kuwait Chapter Arab J Bus Manag Rev 1(5):94–103 Serve, Arlington, VA Raju SS, Chand R (2008) A study on the performance of National Deressa TT, Hassan RM, Ringler C, Alemu T, Yesuf M (2008) Analy- Agricultural Insurance Scheme and suggestions to make it more sis of the determinants of farmers’ choice of adaptation methods efective. Agric Econ Res Rev 21(1):11–19 and perceptions of climate change in the nile basin of ethiopia. Saboor A, Hussain M, Munir M (2009) Impact of micro-credit in alle- International Food Policy Research Institute, Washington, DC viating poverty: an insight from rural Rawalpindi, Pakistan. Pak Füssel HM, Klein RJ (2006) Climate change vulnerability assessments: J Life Soc Sci 7(1):90–97 an evolution of conceptual thinking. Clim Change 75(3):301–329 Saeed DF, Salik KM, Ishfaq S (2015) Climate change and heat-waves: GoP (2016) Annual food report. Ofce of the Chief Engineering Advi- rural to urban migration in Pakistan. A silent looming crisis. sor and Chairman Federal Flood Commission, Islamabad SDPI, Policy Brief no. 47. https​://sdpi.org/publi​catio​ns/fles​/ GoP (2017) Agriculture, Pakistan Economic Survey 2016–17, Ministry Clima​te-Chang​e-and-Heat-Waves​-Rural​-to-Urban​-Migra​tion-in- of Finance. http://www.fnan​ce.gov.pk/surve​y/chapt​ers_16/02_ Pakis​tan-A-silen​t-Loomi​ng-Crisi​s.pdf. Accessed Jan–July 2017 Agric​ultur​e.pdf Salik KM, Jahangir S, ul Hasson S (2015) Climate change vulnerability GoP (2018) Federal Flood Commission. Ofce of Chief Engineering and adaptation options for the coastal communities of Pakistan. Adviser/Chairman Federal Flood Commission. http://www.fc. Ocean Coast Manag 112:61–73 gov.pk/. Accessed May 2017 Saqib SE, Kuwornu JK, Panezia S, Ali U (2018) Factors determining GoP-INDC (2016) Pakistan’s intended nationally determined contri- subsistence farmers’ access to agricultural credit in food-prone bution. Government of Pakistan. http://www4.unfcc​c.int/Submi​ areas of Pakistan. Kasetsart J Soc Sci 39(2):262–268 ssions/INDC/Publi​ shed%20Doc​ ument​ s/Pakis​ tan/1/Pak-INDC.pdf​ . SBP (2014a) Prudential regulations for agriculture fnancing. Agri- Accessed Aug 2017 culture Credit and Microfnance Department. http://www.sbp.org. Hahn MB, Riederer AM, Foster SO (2009) The Livelihood Vulner- pk/publi​catio​ns/prude​ntial​/PRsAg​riApp​roved​.pdf. Accessed 21 ability Index: a pragmatic approach to assessing risks from cli- May 2017 mate variability and change—a case study in Mozambique. Glob SBP (2014b) Indicative credit limits and eligible items for agri- Environ Change 19(1):74–88 fnancing. Agriculture Credit and Microfnance Department, SBP, Hammill A, Bizikova L, Dekens J, McCandless M (2013) Comparative Karachi analysis of climate change vulnerability assessments: lessons from Siyal GA (2018) Farmers see room for improvement in Crop Loan Tunisia and Indonesia. Competence Centre for Climate Change, Insurance Scheme. The Express Tribune. https​://tribu​ne.com.pk/ Bonn, Germany. https://gc21.giz.de/ibt/var/app/wp342​ deP/1443/​ story/16574​ 00/2-farme​ rs-see-room-impro​ vemen​ t-crop-loan-insur​ ​ wpcon​tent/uploa​ds/fileb​ase/va/vulne​rabil​ity-guide​s-manua​ls- ance-schem​e/. Accessed Apr 2018 repor​ts/Compe​rativ​e-analy​sis-of-clima​techa​nge-vulne​rabil​ity- Syed Ali Raza M, Ali Y, Mehboob F (2012) Role of agriculture in asses​sment​s.pdf economic growth of Pakistan. Int Res J Fin Econ 83:1–8 Hinkel J (2011) “Indicators of vulnerability and adaptive capacity”: Zeeshan A, Khan MB (2015) Implementation of disaster risk reduc- towards a clarifcation of the science-policy interface. Glob Envi- tion policy in Pakistan—an evidence from Sialkot. ISSRA paper ron Change 21(1):198–208 2015. http://www.ndu.edu.pk/issra​/issra​_pub/artic​les/issra-paper​ ​ Levin J (2006) Choice under uncertainty. Standford University. https​ /ISSRA​_Paper​s_Vol7_Issue​I_2015/07-Imple​menta​tion-of-Disas​ ://web.stanf​ord.edu/~jdlev​in/Econ%20202​/Uncer​taint​y.pdf. ter-Risk.pdf. Accessed Feb 2017 Accessed Jan–July 2017 McCarthy JJ, Canziani OF, Leary NA, Dokken DJ, White KS (eds) (2001) Climate change 2001: impacts, adaptation and vulnerabil- Ghamz E Ali Siyal MS in Economics from National University of Sci- ity. Cambridge University Press, Cambridge ences and Technology (NUST), Islamabad, Pakistan. Nantui MF, Bruce SD, Yaw OA (2012) Adaptive capacities of farmers to climate change adaptation strategies and their efects on rice Muhammad Zahid Siddique Assistant Professor in School of Social production in the northern region of Ghana. Russ J Agric Soc Sciences and Humanities (S3H) at National University of Sciences and Econ Sci 11(11):1–9 Technology (NUST), Islamabad, Pakistan. Pachauri RK, Allen MR, Barros VR, Broome J, Cramer W, Christ R et al (2014) Climate change 2014: synthesis report. In: Contribu- Syed Mohsin Ali Kazmi Data Analyst in Sustainable Development tion of Working Groups I, II and III to the ffth assessment report Policy Institute (SDPI), Islamabad, Pakistan. of the Intergovernmental Panel on Climate Change. IPCC, p 151

1 3 Earth Systems and Environment (2018) 2:621–632 https://doi.org/10.1007/s41748-018-0075-5

ORIGINAL ARTICLE

Understanding the Interdependence Between Worker Livelihoods and Decent Work at Certifed and Non‑Certifed Mango Orchards in Pakistan

Muhammad Bilal Ahsan1 · Mubashir Mehdi1 · Burhan Ahmad1 · Umar Sohaib Ahmad1

Received: 2 June 2018 / Accepted: 24 September 2018 / Published online: 15 October 2018 © Springer Nature Switzerland AG 2018

Abstract There is little evidence of published literature regarding the incorporation of on-farm workers into global fresh produce value chains. Agriculture, a major sector in Pakistan, employs 50% of the total labor force, and contributes 21 and 30% to GDP and GNP, respectively. Two-thirds of the population resides in rural areas. Mango is the second major fruit product in the country. The horticulture industry of Pakistan in general, and the mango sector specifcally, has been the focus of research and development eforts by governmental and various international developmental organizations. A participatory value chain approach has been adopted to enhance technological advancements and increase the competitiveness of farm enterprises. The adoption of best practices is slow due to a lack of attention to occupational health measures, lack of training and low wages among on-farm workers. Market incentives, compliance standardization and increasing demand for certifed products has driven mango farms in Pakistan to seek certifcation by organizations such as GLOBALG.A.P., which has established provisions for decent work to focus on worker rights and safety. A sustainable livelihood for workers can be achieved by upgrading the value chain system through institutional reforms and an emerging philosophy of decent work put forth by the International Labour Organization. In this study, the efects of decency in the value chain on the income of on-farm workers were assessed at both GlobalG.A.P.-certifed and non-certifed orchards in Sindh and southern Punjab. Because of the food quality and safety standards required for certifed orchards within the global value chain system, these compliant orchards ofer greater potential for improving the livelihoods of on-farm workers. To overcome constraints, extensive awareness campaigns should be launched within the mango industry by institutional supportive bodies.

Keywords Decent work · Global value chains · Pakistan mango industry · Sustainable livelihood · GlobalG.A.P. compliance · On-farm workers

1 Introduction

Pakistan is an agriculture-based country of 190 million people. Fifty percent of the total labor force is directly engaged in agri- culture; among these, half are women (Government of Pakistan 2016). Rural areas contribute more than two-thirds of the total. * Muhammad Bilal Ahsan However, most of these rural populations are economically [email protected] stressed and socially deprived in terms of low wages. Growers Mubashir Mehdi are the most common victims of seasonal and weather efects, [email protected] price variations, and pest attacks, as well as demand and sup- Burhan Ahmad ply fuctuations in traditional value chain systems. The horti- [email protected] culture industry of Pakistan in general, and the mango sector Umar Sohaib Ahmad specifcally, has been the focus of research and development [email protected] eforts by governmental and developmental organizations 1 Institute of Business Management Sciences, University including the Australian Centre for International Agricultural of Agriculture, Faisalabad, Pakistan Research (ACIAR), United Nations Industrial Development

Vol.:(0123456789)1 3 622 M. B. Ahsan et al.

Organization (UNIDO), Food and Agriculture Organization of 1.2 Goals and Objectives the United Nations (FAO), International Centre for Develop- ment and Decent Work (ICDD) and the United States Agency Efective post-harvest value chain management is essential for International Development (USAID). A participatory value for dealing with problems such as low quality, the perish- chain development approach has been espoused to improve able nature of fruits, and pest issues. Mangoes of the highest technological advancement and competitiveness among these quality can only be obtained with the most productive and farm enterprises. An advanced ACIAR “whole-of-chain” efcient labor, which requires functional upgrading in terms approach has been instrumental in transforming the value of social and economic aspects of decent work to ensure chain strategy in the context of on-farm industry development. the highest satisfaction under the sustainable development agenda of the ILO, enabling them to be part of a competitive 1.1 Problem Statement global value chain. Hence, on-farm worker issues are garner- ing prominent attention among international development “Waged on-farm agricultural workers” can be defned as agencies and global certifcation systems (ILO 2016; ICDD men or women who are employed in the orchards, felds, 2014; Global GAP 2016). Therefore, the implementation livestock units, greenhouses and packhouses (i.e. process- of labor standards at certifed mango orchards is gaining ing facilities) at the farm level. They do not own or rent the ground in terms of social, economic and physical aspects. land or the instruments they use (Adato and Meinzen-dick The current study was designed by generating a hypothesis 2002; Hurst et al. 2005). They can be grouped into three that social upgrading under decent work agenda would lead main categories: permanent workers, seasonal workers and to economic upgrading and benefts in premium-quality migrant workers (Ghani 2012). We can further categorize value chains distributed to on-farm workers at certifed them as highly skilled or low-skilled workers based on orchards (Chambers and Conway 1992). their involvement in technical operations or ordinary work, The following objectives were set out for the study: respectively. They receive low wages and are considered a deprived class of rural Pakistan. On-farm workers account 1. Comparative analysis of socioeconomic factors and for 55% of total occupational injuries (Government of Paki- decent work indicators in certifed and non-certifed stan 2016). They receive in-kind wages such as rice, wheat orchards. or other agricultural crops as an alternative to real wages for 2. Assessment of the interdependence between the liveli- their services (Scherrer 2016; Ghani 2012). They are desig- hoods (income) of on-farm workers and decent work nated as a depressed class because of their complete depend- indicators. ence on landlords/owners and lack of ability to organize. 3. Recommendations for appropriate policy reforms to Consumers in the developed world are becoming increas- improve the on-farm industry. ingly conscious of both quality and working conditions at the point of production (Hurst et al. 2005; FAO 2015; Wog- num et al. 2011). GlobalG.A.P., an international certifcation system aimed at ensuring good agricultural practice (GAP), 2 Mango Industry and Value Chains and led by some of the top supermarket chains, has set out in Pakistan comprehensive compliance criteria for growers to meet the needs of this competitive global value chain. Most certifed Mango (Mangifera indica), also known as the “king of farms in developing economies prefer to export to developed fruits”, is native to Burma and eastern India. Mango is the countries (Collins and Dunne 2007; Raynolds et al. 2004). second highest fruit product in Pakistan, after citrus, and According to Sustainable Development Goal (SDG) 8 of Pakistan ranks ffth in mango production worldwide (FAO the International Labour Organization (ILO), decent work 2015). Numerous varieties are being produced in our coun- is associated with the efectiveness of many physical, eco- try, which are Dodheri, Anwar Ratol, Langra, Siroli, Malda, nomic and social indicators for functional upgrading of value Alphonso, Fajri, Gulab Khas, Golden, Swarnarekha and Ban- chains, including economic development of certifed farms ganapalli categorized by size, taste, acidity, shape, aroma and and social upgrading for workers (Kadigi et al. 2007; ILO color. Two varieties, Chaunsa and Sindhri, have the greatest 2015). These aspects of on-farm labor rights are periodically commercial importance, in both local and global markets. inspected by consultants to assess compliance. Requirements include occupational safety and health training, basic hygiene to prevent disease, protective clothing to prevent skin cuts 2.1 Industry and exposure to hazardous substances, frst aid kits, warn- ing signs in hazardous areas, and other social and economic The main export markets for Pakistani mangoes are Saudi measures for labor welfare (Global GAP 2016). Arabia, the United Arab Emirates, the UK, Qatar, Iran,

1 3 Understanding the Interdependence Between Worker Livelihoods and Decent Work at Certifed… 623

Afghanistan and Oman (Government of Pakistan 2009; Gha- the growing pressure for graded, certifed and standard- foor et al. 2009). Numerous initiatives have been undertaken ized products and price incentives, some farm enterprises by the Pakistani government and many international agen- in Pakistan are obtaining international certifcation such as cies to develop the export value chain in order to empower GLOBALG.A.P., an on-farm certifcation system to enable growers and exporters. However, traditional chains are entrance into the global value chains. GLOBALG.A.P. cer- facing many issues including global certifcation systems tifcation has been acquired by 24 mango farms in Pakistan (Hazard Analysis and Critical Control Points [HACCP] and (PHDEC 2016). GLOBALG.A.P.), pest issues, poor working conditions and lack of institutional reforms. Orchards in Pakistan can be grouped into two general 3 Conceptual and Empirical Framework categories, non-certifed and certifed, based on the quality of the mangoes produced. Traditional value chain mangoes A review of the literature on the incorporation of on-farm from non-certifed orchards are identifed by classical stand- workers found preliminary evidence of the contribution by ards and measurements of quality including aroma, appear- agricultural on-farm workers to the development of sustaina- ance and size, and are typically sold in local retail roadside ble agricultural practices (Hurst et al. 2005). At non-certifed markets (CSF 2007; Mehdi et al. 2016). These traditionally orchards, teenagers and children were found to be working harvested mangoes are full of sap burns, blemishes, harmful under poor health and safety conditions on fruit farms, and chemicals and pesticides. The use of wooden sticks is com- workers were not allowed to forms unions (HRW 2002), mon practice in these orchards, which may lead to internal whereas wage rates, and occupational safety and health injury of the fruit. Thirty to forty percent of post-harvest (OSH) conditions were improving at certifed orchards. losses occur in traditional value chains (Collins and Iqbal 2010). Wooden boxes are used, exposing the mangoes to 3.1 Conceptual Framework pathogens. In addition, overflling of boxes increases bruis- ing of the fruit and physical damage. Poor working conditions, with lack of access to water and other basic needs, as well as exposure to health hazards, have 2.2 Traditional Versus Best‑Practices Value Chain been highlighted in many previous studies (Collins 2009; and Role of Decent Work Certifcation Hale and Wills 2007; Smith et al. 2005; Oxfam International 2004). Both institutional supportive bodies and international According to SDG 8 of the ILO, decent work is defned as compliance agencies can contribute signifcantly to sociode- “work that is productive and delivers a fair income, security mographic improvement among the many value chain actors. in the workplace and social protection for families, better Global agro-based value chain standards are driven by prospects for personal development and social integration, producers or consumers concerned about decent work con- freedom for people to express their concerns, organize and ditions for workers; this helps to promote the functional participate in the decisions that afect their lives and equal- upgrading of on-farm workers because of consumer con- ity of opportunity and treatment for all women and men” cerns about point of production and brand visibility, par- (ILO 2015). ticularly in value chains that are consumer-driven. This In contrast, some GAP-certifed orchards are now restruc- study was developed by assuming the general hypothesis turing product fows, supplying mangoes to supermarkets that upgrading of traditional value chain systems through in large cities such as Islamabad, Karachi, Faisalabad and compliance with food quality and safety standards under Lahore, premium-quality outlets and international markets, an international compliance structure would lead to social because of additional quality attributes including blemish- and economic upgrading, as shown in Fig. 1. To test this free fruit, ethylene ripening, prestige, and cardboard packag- hypothesis, interdependence among diferent decent work ing (ACIAR 2007; Mehdi et al. 2014). These orchards take indicators and livelihoods were assessed at the bottom of suitable care of mangoes under compliance measures estab- the global value chains, while diferent social, economic and lished by GLOBALG.A.P. (a certifcation system governing physical decent work indicators were identifed at diferent on-farm practices). Measures such as these not only increase levels of the value chain to determine the improvement in market value, such as organic labeling, but may reduce pre- upgraded value chains. and post-harvest losses. A comparison of key activities per- In the case of food and fresh fruit commodities, safety formed in certifed (decent work) and non-certifed orchards and quality measures are of great concern for supermarkets is described in Table 1. and consumers in developed countries (Serrat 2008; Hum- Higher beneft-cost ratios are achieved by these certi- phrey and Memedovic 2006). Preliminary evidence shows fied orchards relative to non-certified orchards because that social upgrading (unionization) has contributed signif- of premium-quality products (Mehdi et al. 2014). Due to cantly to economic improvements as well as greater adoption

1 3 624 M. B. Ahsan et al. ) for ) for 2 safety equipment, chemical storage in separate places, and proper training provided provided training places, and proper in separate storage chemical equipment, safety workers farm for of an iron pole of variable length (cut-and-convey type of mango picker) in order in order type of mango picker) length pole of variable of an iron (cut-and-convey sap burns, reduce to using special Transport internal injury infection. and fungal ameni - with decent work of fruit, sort easy to workers by one layer withcrates only aid kits, doctors frst the to clothing according of the nature work, ties (protective on site, etc.) sapping, turning of mangoes on specifcally designed racks, is frequently used. is frequently sapping, turningdesigned racks, of mangoes on specifcally (Ca(OH) and dipping in 0.5% lime water de-stemming Other involve methods 2–3 min ards to some degree, as the ripening process has seen a paradigm shift towards some degree, to ards as the ripening shift has seen a paradigm process towards chambers ethylene are applied at certifed orchards to improve the shelf life and to protect from pests pests from protect and to the shelf life applied at certifed improve are to orchards and transported- in card and then packed as fruit chains, such value fies in global buyers attributes by and quality the to set according size, variety boxes board Certifed orchards observe proper for worker occupational safety measures such as such measures occupational safety worker for proper Certifed observe orchards GLOBALG.A.P.-certifed orchards (under international orchards standards) GLOBALG.A.P.-certifed After ascertaining means the maturity is conducted by harvesting of the fruit, safe Certifed orchards de-sap the fruit through a variety of ways, where physical de- physical where Certifed de-sap the fruit orchards of ways, through a variety In certifed stand the- compliance ripening orchards, has shifted process organic to Before packing, treatments such as the fungicide carbendazim or hot water treatment treatment as thewater fungicide carbendazim or hot such treatments packing, Before mango skin, which attracts insects. Transport of fruit from mango tree to packing of fruit insects. Transport attracts packing to mango tree mango skin, which from folds of fruits quantity in many with extra is done in a pannier, area/packhouse losses post-harvest leads to other which each over sanitary and phytosanitary measures such as de-sapping, a process of removing sanitary of removing as de-sapping, a process and phytosanitary such measures the the sap burns plant sap from fruit. attack, of fungal the Sap increases chance and blemishes on fruit the use of calcium carbide (carcinogen), known locally as masala, to accelerate accelerate as masala, to locally known the use of calcium carbide (carcinogen), the ripening process. certifed orchards. Mangoes are packed in wooden boxes over their capacity, their capacity, over boxes in wooden certifed packed Mangoes are orchards. infection of fungal chance and increased in deterioration results in quality which Of-season activities include pruning/trimming, irrigating plowing, and fertilizing Hand-picking or through wooden sticks. Ugly blemishes, sap burns appear on the Ugly sticks. or through wooden Hand-picking Non-certifed orchards Non-certifed orchards do not adhere to protocols such as export quarantines, or as export quarantines, such protocols to adhere do not Non-certifed orchards Mangoes are ripened by traditional methods, which carry which methods, traditional ripenedMangoes are as by health such hazards, Packing experts have a specifc role in on-farm industry in on-farm a specifc role at both non-certifed and experts have Packing On-farm value chain activities Source: feld observations and group feld observations activities Source: discussions chain value On-farm harvest handling harvest Pre-harvest practices Pre-harvest Harvesting Preliminary post- 1 Table Ripening Packaging

1 3 Understanding the Interdependence Between Worker Livelihoods and Decent Work at Certifed… 625

Fig. 1 Conceptual framework International compliance (Global GAP)

ON-Farm Safety & Quality Standards Sustainable DECENT ON- Livelihoods WORK Farm AGENDA Workers Economic Upgrading

Institutional Support (Ministry of Labor and Human resource)

of OSH standards for workers (Freeman 1981, 1982; Hirsch economic factor of decent work (if available, E1 = 1; other- 1997). Government is often the principal actor in sustain- wise = 0), and E2 is the compensation payments as dummy able development initiatives due to its regulatory powers variable (if paid, then E2 = 1; otherwise = 0. and infuence on rural industry reforms (World bank 2017). The comprehensive formula for the log–log regression model Eq. (2) is stated as

3.2 Empirical Framework ln ln ln ln IL = A + 1 C1 + 2 C2 + 3 C3 + 4S1 + S + P + P + E + E +  (3) A model in empirical form can be characterized as 5 2 6 1 7 2 8 1 9 2 = , , , IL f C SD PD ED (1) 4 Data where C is the continuous variables of socioeconomic attrib- utes, SD is the social context of decent work as dummy vari- A case study methodology was employed to gain an in- able, PD is dummy variables such as physical and human depth understanding of how decency in an upgraded value decent work indicators, and ED is the economic aspect of chain afects on-farm worker income in both GAP-certifed decent work. and other orchards. From the two Pakistani provinces of Equation (1) above can be amended as Sindh and Punjab, three major mango-producing regions (Multan, Hyderabad and Rahimyarkhan) were selected I = AlnC 1 lnC 2 lnC 3S 4S 5P 6P 7E 8E 9 L 1 2 3 1 2 1 2 1 2 (2) for collection of data. These provinces account for 95% where IL is on-farm worker income (USD), C1 is the edu- of total mango production in Pakistan. A well-structured cational level of the on-farm workers (grades passed), C2 is questionnaire was developed by focusing on the social, worker experience in the on-farm industry (years), C3 is the economic and physical impact of decent work. age of on-farm workers (years), S1 is the social dummy vari- Fifteen certifed orchards were selected using simple able for class discrimination (if discrimination, then S1=1; random sampling. A total of 200 on-farm workers from otherwise = 0), S2 is the dummy variable for unionization (if both these orchards and other non-certifed orchards were 2 unions present, then S = 1; otherwise = 0), P1 is the physical interviewed using purposive sampling. In-depth focus dummy variable for OSH (if provided, P1= 1; otherwise = 0), group discussions were conducted with the growers and P2 is the human aspect dummy specialization (if provided, farm contractors. 2 then P = 1; otherwise = 0, E1 is the financial services

1 3 626 M. B. Ahsan et al.

The analysis consisted of the following: A signifcant diference was observed between GAP- compliant and non-compliant orchards in rural Pakistan. In • A comparative analysis of GAP-compliant and non- certifed orchards, 100% of on-farm workers were union- compliant orchards by employing descriptive statistics ized, while only 4.3% were represented by labor unions in techniques including frequencies, percentages and means non-compliant orchards (Table 2). Class-based wage deter- • Log–log regression analysis performed to assess the mination was seen in 16% of non-compliant orchards and interdependence between the income of workers and 10% of certifed orchards. Unionization has consequences socioeconomic and decent work variables. for collective bargaining power, social dialogue, social secu- rity and the ability to voice concerns, as well as many other According to Econometrics, log–log regression mod- provisions in a decent work agenda. els are preferred when dealing with development-related Compensation payments are awarded to workers who are approaches and variables (Gujarati 1995). disabled, become ill or are injured during working hours. These comprise numerous types of disbursements, including worker injury benefts, and maternity and paid leave. Most 4.1 Descriptive Statistics and Decent Work growers allocate workers a piece of land upon which to build Indicators their mud houses. On-farm accommodation rates were very high at GAP-compliant orchards, with about 90% having As described earlier, descriptive statistics were used for mud houses, while a rate of 60% was found in the case of comparative analysis of decent work indicators and soci- non-compliant orchards, as indicated in Table 2. Ninety- odemographic factors. Seventy-four percent of seasonal four percent of all non-compliant orchard workers were workers were employed at certifed orchards, where 16.7% able to ofcially own their own houses, versus only 57% were permanent workers at non-compliant orchards, as at compliant orchards. Forty-three percent and 7% lived in shown in Table 2. rented houses in GAP-compliant and non-certifed orchards, Eighty-three percent workers were seasonal at non-com- respectively. pliant orchards and 25.7% were permanent workers at certi- fed orchards. Thus, a greater number of seasonal workers worked at certifed orchards versus non-certifed orchards. 5 Results and Conclusions Economic upgrading in terms of producing premium-quality products correlates with social upgrading of on-farm work- As GAP-compliant orchards are driven by an international ers, along with better earnings in local and global markets. certifcation system and supermarkets in developed coun- Most on-farm workers in the traditional on-farm industry in tries, they produce premium-quality products, obviously Pakistan are economically depressed and socially deprived, with a decent price for farm products, ultimately devel- as shown in Table 2, but circumstances are better in GAP- oping a huge potential for decent working conditions for compliant orchards. workers. However, there is a lack of literature supporting

1 3 Understanding the Interdependence Between Worker Livelihoods and Decent Work at Certifed… 627

Table 2 Sociodemographic Description Non-certifed Certifed orchards Total distribution and decent work orchards (%) (%) indicators Source: author’s Number % calculations from survey 2016 Worker type Permanent 16.7 25.7 115 57.5 Seasonal 83.3 74.3 84 42 Decent work indicators Labor unions 28.26 100 66 33 Professional training 31.43 66.67 85 42.5 Occupational safety and health 4.26 100 110 55 Class-based discrimination 15.71 10 22 11.5 Financial services 36.4 64.3 148 74 Compensation payments 27.7 66 94 47 Lifestyle Migrant 18.6 36.7 66 33 Three full meals at home 100 67.1 155 77.5 Clean water access 78 94 165 82.5 Cement house 40 10 38 19 Mud house 60 90 162 81 Rental house 6.667 42.86 65 32.5 Owned house 93.33 57.14 136 68 Joint family 68.6 86.7 149 74.5 Aford children’s education 78 81 157 86.5 Married 67.1 73.3 138 69 Income (USD)/month ≤$50.00 31.4 0 44 22 $50.01–100.00 50 13.3 79 39.5 $100.01–150.00 14.3 23.3 65 32.5 $150.01–200.00 2.86 43.3 10 5 $200.01–350.00 1.43 20 5 2.5 Experience (years) <10 46.67 77.1 12 6 11–20 20 17.14 145 72.5 20–25 22.33 1.42 42 21 No experience 10 4.28 3 1.5 Age (years) ≤18 4.5 0 18 9 19–25 50 47.1 71 35.5 25–35 26.67 35.7 79 39.5 35–45 23.33 17.1 24 12

Frequencies include total sample (200) the notion that social upgrading leads to economic upgrad- 5.1 Decent Work and Livelihoods (Income) Models ing of on-farm workers (Brown 2007; Locke et al. 2007). (DLM) The research question should be investigated by determining the relationships between economic and social upgrading at Two separate log–log regression models were used to farm level (Barrientos et al. 2010), particularly by inspect- determine the efects of diferent decent work variables on ing the circumstances (decent value chain safety and quality worker income. All assumptions of ordinary least squares standards) under which social development may progress (OLS) like normality of data, linearity, perfect collinearity, towards economic upgrading. homoscedasity were tested using SPSS Statistics version 22 software (IBM Corp., Armonk, NY, USA). Variables used in the livelihood models are explained in Table 3.

1 3 628 M. B. Ahsan et al.

Table 3 Description of Model variables Minimum Maximum Mean SD variables used in livelihood models Source: author’s own Income (PKR)/month (DV) 4800 ($48) 32,000 ($320) 11,046.0 ($110.4) 6072.01 ($60.72) calculations Age (years) 12 45 26.39 7.812 Education (grades) 0 14 4.17 3.660 Experience (years) 0 21 6.00 4.723 Compensation payments 0 1 0.39 0.490 Occupational safety and health 0 1 0.30 0.461 Class discrimination 0 1 0.22 0.416 Labor unions 0 1 0.50 0.503 Specialized professional 0 1 0.30 0.461 Financial services 0 1 0.34 0.476

DV dependent variable

Table 4 Comparative results Model variables IV (income) Certifed orchards Non-certifed orchards of log–log regression models Source: author’s estimations Coefcients Signifcance Coefcients Signifcance from model (*p < 0.10, NS **p < 0.05, ***p < 0.01) Age 0.787 0.006*** 0.137 0.276 Education 0.071 0.166NS 0.087 0.023** Experience 0.064 0.051* 0.065 0.185NS Occupational safety and health 0.291 0.070* 0.01 0.174NS Professional training 0.323 0.049** 0.224 0.155NS Class discrimination − 0.341 0.008*** − 0.222 0.006*** Labor union 0.397 0.081* 0.185 0.013** Compensation payments 0.225 0.031** 0.074 0.324NS Financial services 0.045 0.656NS 0.197 0.040** Constant 6.619 0.000*** 8.277 0.000*** R2 0.909 0.739 Adjusted R2 0.826 0.546 F value 10.534 1.012 Durbin–Watson statistics 1.572 9.162

IV independent variables

To assess the overall signifcance of the model, analy- indicated in Table 4. For non-certifed orchards, the relation- sis of variance (ANOVA) was performed by assuming that ship was non-signifcant. all coefcients were equal to zero. The F value was very Education is a key factor in the level of technical skill high (35.101; p < 0.000), indicating that both models were and opportunity for promotion. It is a major contributor highly accurate overall. Thus our independent variables have to a country’s socioeconomic development (Griliches and adequate explanatory power for regression of our depend- Mason 1972). For non-certifed orchards, the education coef- ent variable income of on-farm workers, a major indicator fcient also indicated a positive relationship with income, of livelihood. The model hit rates were 73.9 and 90.8% for and was found to be highly signifcant (p < 0.05), with a certifed and non-certifed orchards, respectively, as the coef- 1% increase in schooling years resulting in an increase in fcient of determination (R2) showed values of 0.785 and income of 0.08%, as denoted in Table 4, while certifed 0.908 (Table 4). According to the econometrics literature, orchards showed a non-signifcant efect. it is better to use adjusted R2 if our model has incorporated Experience is also an important factor, with more experi- more than one independent variables (Gujarati 1995). enced workers likely to earn a higher wage (Mincer 1974). The age coefcient showed a positive relationship with In this study, at certifed orchards, a 1% increase in experi- wage in certifed orchards and was signifcant at the 1% con- ence was found to increase income signifcantly, by 0.06% fdence level (p < 0.01), with a 1% increase in age resulting (p > 0.10), while it was non-signifcant in the case of non- in an increase in income of 0.78%, all else being equal, as certifed orchards (p > 0.10), keeping the efect of other

1 3 Understanding the Interdependence Between Worker Livelihoods and Decent Work at Certifed… 629 factors constant (Table 4). Experience increases labor tech- While professional training for specialization was signif- nical skills to efectively deal with various technical issues cant at the 5% confdence level (p < 0.05), with a p value of and practices at the farm level and enables workers to elevate 0.049, the model revealed that the average income for work- their work quality. ers receiving professional training was 0.32% greater (Bar- tel 1995). These are known as skilled agricultural workers, 5.2 Economic Upgrading Through Decent Work who are able to assume specifc roles as a result of this on- farm industry training, such as machine operators, packag- Global value chains are transforming their processes to ing experts, graders, supervisors and harvesters, and these enable the production of premium-quality and cost-efective career-oriented jobs afect income. products. A positive correlation may be assumed between social and economic upgrading (Sono 2013; Gundersen 5.4 Social Upgrading Through Decent Work 2003), particularly with regard to increased production capacity of workers and greater fexibility. The most ef- Some studies, however, have revealed negative relationships cient and productive labor can thus be achieved by worker for social and economic upgrading under specifc conditions satisfaction during and after leaving the job. (Astill and Grifth 2004; Raworth 2004). Scientists should Under the economic umbrella of decent work, compensa- focus more on the social aspect of work than on the work- tion payments are an important factor. This variable showed place with regard to long-term benefts (Nuwayhid 2004). a positive relationship with labor wages: workers receiving Unionization is a major factor in wage determination and compensation payments, such as illness and maternity leave, addresses decent work framework provisions including leaves for special occasions and worker injury benefts, had worker representation, social dialogue, and freedom to 0.22% higher average income at certifed orchards, keep- organize and associate, which are major points for social ing the efect of other factors constant (p < 0.05), as shown integration and personnel development according to ILO in Table 4. In contrast, at non-certifed orchards this vari- SDG 8 (Freeman 1984; Diewert 1974; ILO 2015). able was non-signifcant. Financial services at non-certifed Labor unions were found to have a signifcant efect orchards at the farm level had a signifcant efect at a conf- in both types of orchards (p < 0.05). Average income for dence level of 5% (p < 0.05), showing that workers having unionized workers was 0.39 and 0.18% higher at certifed fnancial services available received 0.14% higher average and non-certifed orchards, respectively (Table 4) (Hirsch income, all else equal (Table 4). For certifed orchards, this et al. 1997; Hirsch and Macpherson 2003; Mishel and Wal- variable showed a non-signifcant efect. ters 2003). Class discrimination may contribute to social downgrading. This variable was highly signifcant at both 5.3 Physical and Human Upgrading Through Decent certifed and non-certifed orchards at a 1% confdence Work level (p < 0.01). Workers who are subjected to discrimina- tion based on class may receive on average 0.34 and 0.22% A sustainable mango value chain system can be achieved by less income at GAP-compliant and GAP-non-compliant integrating ILO SDGs into traditional production networks, orchards, respectively. Hence, worker exploitation may particularly for workers and farmers with small holdings. cause social downgrading. This is a major factor afecting GAP-compliant orchards are implementing OSH and difer- payment of wages in the developing world, especially in ent food quality standards. There is a common misperception the South Asian countries (Banerjee and Knight 1985; Das that these OSH standards increase production costs, which and Dutta 2007; Best and Mamic 2008). These workers may discourage the implementation of such standards, but deserve attention by institutional supportive bodies and in fact it increases the social and economic efciency of the government, but no direct labor law exists for them in enterprise by improving its competitiveness. Pakistan, although India has special legislation for on- This variable showed a signifcant efect at a 10% con- farm workers. fdence level (p < 0.10), with a p value of 0.07. This model predicts that, keeping the efect of other factors constant, the availability of amenities such as separate chemical rooms, 5.5 Conclusions and Policy Recommendations frst aid kits, safety equipment, food hygiene, protective clothing, hazard signs, doctor contracts and proper sanita- • Social upgrading has implications for economic tion facilities may result in 0.29% greater average income improvement in the mango value chain with regard at certifed orchards, as shown in Table 4. This variable to a decent work environment and improved worker showed a non-signifcant efect at non-compliant orchards. income for premium-quality mango production at certi- fed orchards.

1 3 630 M. B. Ahsan et al.

• Decent work infuences the efects of various socioeco- Bartel AP (1995) Training, wage growth, and job performance: evi- nomic indicators of on-farm worker livelihoods, which dence from a company database. J Labor Econ 13(3):401–425 Best S, Mamic I (2008) Global agri-food chains: employment and ultimately informs the value chain development process. social issues in fresh fruit and vegetables. International Labour • With the global compliance drive to certify orchards Organization, Geneva under the food quality and worker safety standards, these Brown D (2007) Globalization and employment conditions study. orchards have greater potential to improve the livelihoods World Bank, Washington (social protection discussion paper 0708) of on-farm workers. Chambers R, Conway G (1992) Sustainable rural livelihoods: practi- • Growers are reluctant to adopt decent work practices due cal concepts for the 21st century. IDS, Brighton (IDS discus- to the common misperception that maintaining compli- sion paper 296) ance standards will increase the cost of production. In Collins JL (2009) Threads: gender, labor, and power in the global apparel industry. University of Chicago Press, Chicago reality, meeting compliance measures will ultimately Collins RJ, Dunne AJ (2007) A rapid supply chain appraisal approach improve farm competitiveness. for agribusiness development projects. In: II International sym- • Socially upgraded labor contributes positively to eco- posium on improving the performance of supply chains in the nomic improvement for workers and farms in terms of transitional economies paper no. 794 Sep 23, pp 73–80 Collins R, Iqbal M (2010) Integrating post-harvest, marketing and higher wages and premium-quality products, respec- supply chain systems for sustainable industry development: the tively. Pakistan mango industry as work in progress. In: Batt PJ (ed) • This study has utilized an integrated approach to ana- Third international symposium on improving the performance lyze social, economic and OSH aspects of decent work of supply chains in the transitional economies, 4–8 July 2010, Kuala Lumpur, Malaysia in global production networks, with regard to the diverse CSF (2007) Horticulture action plan: background paper. Competitive income gap among both growers and workers. Hence, Support Fund. Ministry of Finance, Government of Pakistan, there is an urgent need to ensure that benefts are equally Islamabad distributed among workers and producers. Das MB, Dutta P (2007) Does caste matter for wages in the Indian labor market?. World Bank, Washington (draft paper) • Government should develop a more effective policy Diewert WE (1974) The effects of unionization on wages and for socioeconomic development to focus on both labor employment: a general equilibrium analysis. Econ Inq unions to represent workers, and farm-level initiatives to 12(3):319–339 increase awareness (to minimize class-based discrimina- FAO (2015) FAOSTAT, Food and Agriculture Organization of the United Nations (FAO) October 03, 2015 tion at farms). Freeman RB (1981) The efect of unionism on fringe benefts. ILR • There is huge potential for functional upgrading (labor- Rev 34(4):489–509 intensive), because workers with low skill levels lean Freeman RB (1982) Union wage practices and wage dispersion towards migration to fnd healthier work environments within establishments. ILR Rev 36(1):3–21 Freeman RB (1984) Longitudinal analyses of the efects of trade where they can obtain better working conditions. unions. J Labor Econ 2(1):1–26 • Technical occupations are being promoted in a decent Ghafoor A, Mustafa K, Mushtaq K, Abedullah (2009) Cointegra- way at the certifed orchards for trained workers. Voca- tion and causality: an application to major mango markets in tional training institutes may also initiate short diploma Pakistan. Lahore J Econ 14:85–113 Ghani MA (2012) Article: labour laws and agriculture workers. courses for on-farm professions. Pakistan. Business recorder. Available at ​ https​://fp.breco​rder. com/2012/01/20120​11511​42813​/. Accessed 15 Jan 2012 Global GAP (2016) Database, certifed orchards in Pakistan. Avail- able at https​://www.Globa​lGAP.org References Government of Pakistan (2009) Pakistan economic survey 2013–14. Government of Pakistan, Finance Division, Economic Afairs Wing, Islamabad ACIAR (2007) A constraint analysis of Pakistan mango supply chain. Government of Pakistan (2016) Agricultural statistics of Pakistan Australian Centre for International Agricultural Research, 2013–14. Pakistan Bureau of Statistics, Statistics Division, Canberra Islamabad Adato M, Meinzen-Dick R (2002) Assessing the impact of agri- Griliches Z, Mason WM (1972) Education, income, and ability. J cultural research on poverty using the sustainable livelihoods Political Econ 80(3, part 2):74–103 framework. International Food Policy Research Institute, Gujarati DN (1995) Basic econometrics, 3rd edn. McGraw-Hill Pub- Washington lishing Company, Maidenheach Astill A, Grifth M (2004) Clean up your computer: working condi- Gundersen B (2003) Unions and the well-being of low-skill work- tions in the electronics sector. CAFOD, London ers. Dissertation, George Warren Brown School of Social Work, Banerjee B, Knight JB (1985) Caste discrimination in the Indian Washington University urban labour market. J Dev Econ 17(3):277–307 Hale A, Wills J (2007) Women working worldwide: transnational net- Barrientos S, Geref G, Rossi A (2010) Economic and social upgrad- works, corporate social responsibility and action research. Glob ing in global production networks: developing a framework for Netw 7(4):453–476 analysis. Int Labor Rev 150(3–4):319–340

1 3 Understanding the Interdependence Between Worker Livelihoods and Decent Work at Certifed… 631

Hirsch BT (1997) Unionization and economic performance: evidence Serrat O (2008). The sustainable livelihoods approach, DC: Asian on productivity, profts, investment, and growth. Fraser Institute, Development Bank. Knowledge Solutions. November, 15. https​ Vancouver ://www.adb.org/Knowl​edges​oluti​ons Hirsch BT, Macpherson DA (2003) Union membership and coverage Smith S, Tallontire A, Dolan C, Barrientos S (2005) Reaching the database from the current population survey: note. ILR Review marginalised? Gender value chains and ethical trade in African 56(2):349–354 horticulture. Dev Pract 15(3–4):559–571 Hirsch BT, Macpherson DA, DuMond JM (1997) Workers’ compen- Sono HS (2013) The efects of Adra’s mango project on livelihoods sation recipiency in union and nonunion workplaces. ILR Rev in the Yilo Krobo district. Doctoral dissertation, University of 50(2):213–236 Ghana) HRW (2002) Tainted harvest. child labor and obstacles to organizing Wognum PM, Bremmers H, Trienekens JHVD, Vorst JGAJ, Bloemhof on Ecuador’s banana plantations. Human Rights Watch, New York JM (2011) Systems for sustainability and transparency of food Humphrey J, Memedovic O (2006) Global value chains in the agri- supply chains—current status and challenges. Adv Eng Inform food sector. A report submitted to (FAO) Food and Agriculture 25(1):65–76 Organizations of United Nations. https​://agris​.fao.org World Bank (2017) Annual report. Available at http://pubdo​cs.world​ Hurst P, Termine P, Karl M (2005) Agricultural workers and their bank.org/en/90848​15074​03754​670/Annua​l-Repor​t-2017-WBG. contribution to Sustainable Agriculture and Rural Development, pdf International Labour Organization (ILO), International Union of Food, Agricultural, Hotel, Restaurant, Catering, Tobacco and Allied Workers’ Associations (IUF). Report submitted to (FAO) Muhammad Bilal Ahsan has Food and Agriculture Organizations of United Nations. 03 April passed Master of Science in ICDD (2014) Supply Chain Governance: A decent work approach to Agribusiness with high distinc- optimize mango value chain in Pakistan. Project details. Avail- tion and currently, serving as able at http://www.uni-kassel.de/einri​ chtun​ gen/inter​ natio​ nal-cente​ ​ Lecturer at the University of r-for-developmen​ t-and-decen​ t-work-icdd/resea​ rch/resea​ rch-2015-​ Agriculture, Faisalabad (Bure- 2019/1-globa​l-agric​ultur​al-produ​ction​-syste​ms-gaps.html wala Campus). His research ILO (2015) International Labour Organization, Goal #8, agenda for interests are sustainable rural sustainable development; decent work and economic growth. development, value chain analy- International Labour Organization, Geneva. Available at ​https​:// sis and consumer behavior. www.ilo.org/wcmsp5/group​ s/publi​ c/---ed_norm/---relco​ nf/docum​ ​ ents/meeti​ngdoc​ument​/wcms_54583​7.pdf ILO (2016), Decent Work Programme in Pakistan, SDG N0. 8, Inter- national Labour Organization. Available at​ https​://www.ilo.org/ wcmsp​5/group​s/publi​c/---ed_mas/---progr​am/docum​ents/gener​ icdoc​ument​/wcms_56102​5.pdf Kadigi RM, Mdoe NS, Senkondo E, Mpenda Z (2007) Efects of food safety standards on the livelihoods of actors in the Nile perch Mubashir Mehdi is Assistant Pro- value chain in Tanzania (No. 2007: 24). DIIS working paper fessor at the Institute of the Busi- Locke R, Kochan T, Romis M, Qin F (2007) Beyond corporate codes ness Management Sciences Uni- of conduct: work organization and labour standards at Nike’s sup- versity of Agriculture, pliers. Int Labour Rev 146(1–2):21–40 Faisalabad, Pakistan. He holds a Mehdi M, Adeel A, Ahmad Z, Abdullah M, Hussain F (2014) Efec- Ph.D. in agribusiness from the tiveness of a “whole of chain” approach in linking farmers to University of Queensland, Aus- market: a case of Pakistan mango market. UMK Procedia 1:57–62 tralia. His research focuses on Mehdi M, Ahmad B, Yaseen A, Adeel A, Sayyed N (2016) A compara- agricultural marketing, manage- tive study of traditional versus best practices mango value chain. ment, and overall supply chain Pak J Agri Sci 53(3):733–742 analysis. He has been actively Mincer J (1974) Schooling, experience, and earnings. Human Behavior engaged in industry-focused and Social Institutions No. 2 R&D and capacity building Mishel L, Walters M (2003) How unions help all workers. EPI Brief- activities. ing Paper# 143 Nuwayhid IA (2004) Occupational health research in develop- ing countries: a partner for social justice. Am J Public Health 94(11):1916–1921 Oxfam International (2004) Trading away our rights: women working in global supply chains. Oxfam International, Oxford PHDEC (2005) Mango marketing strategy, Pakistan Horticultural Development and Export Company, Lahore Raworth K (2004) Trading away our rights: women working in global supply chains. Oxfam Policy Pr Priv Sect 1(1):1–52 Raynolds LT, Douglas M, Peter LT (2004) Fair trade cofee: building producer capacity via global networks. J Int Dev 16(8):1109–1121 Scherrer C (2017) Addressing the decent work defcit in agricultural supply chains. Paper presented in 5th conference on regulating for decent work, 3–5 July, ILO Geneva

1 3 632 M. B. Ahsan et al.

Burhan Ahmad earned his PhD Umar Sohaib Ahmad passed his Degree from Norway and is Master of Science degree from working as Assistant Professor in the University of Agriculture, the Institute of Business Man- Faisalabad, and serving as agement Sciences, University of research assistant in USAID- Agriculture, Faisalabad, Paki- funded project in Pakistan. His stan. His research interests are research interest is an analysis of international trade and the WTO, the marketing system of diferent global marketing, agricultural fruit crops. value chains, market functioning, price volatility and foreign direct investment. His recent publica- tion is ‘Spatial diferences in rice price volatility: A Case of Paki- stan’ in The Pakistan Development Review (PDR) Vol.56 (3), (2017).

1 3