Assessing the impact of climate change on Indian agriculture: Evidence from major crop yields
Abstract
This study empirically examines the effect of climate change on the yields of primary food as well as non-food crops in India. The present study uses annual time-series data of seven major crops such as rice, wheat, pulses, rapeseeds and mustard, cotton, sugarcane, and groundnut for 58 years (1961–20 17) to assess the influence climatic variables namely rainfall, maximum, and minimum temperatures on crop yields. The empirical findings of the study indicate that a significant effect on major crop yields from rainfall, maximum and minimum temperatures, but the level of impact varies across the crops studied. A rise in rainfall has an adverse effect on food crops except for pulses, however it has a positive relationship with non-food crops throughout the study period. Further, the average maximum temperature has a positive influence on food and non-food crops excluding rice. The average minimum temperature has an adverse impact on non-food crops, but it has a positive association with food crops. The adverse effects of climatic factors on crop yields may be likely to pose severe implications for food and nutritional security. Conclusively, this study recommends taking adaptation activities to cope with the adverse impacts of climate change.
1 INTRODUCTION
There is an increasing scientific agreement that natural and anthropogenic activities have made a significant contribution to rising atmospheric concentrations of greenhouse gases (Pachauri & Meyer, 2014). The surge of greenhouse gases has enhanced the natural greenhouse effect, which in turn has caused more extensive warming of the earth's atmosphere and surface (Tol & De Vos, 1998). An increased warming has resulted in the intensity and frequency extreme climatic situations causing climate change (Agba, Adewara, Adama, Adzer, & Atoyebi, 2017). Though it is debatable about the sources climate change, the evidence of alterations in rainfall, temperatures, and extreme climatic events is unquestionable (Nath & Mandal, 2018). Furthermore, there are numerous of mechanisms by which “global climate change” may affect human prosperity. The most direct and indeed most considerable impact may come through agriculture particularly food production (Kaufmann & Snell, 1997). Likewise, among several affected by climate change agriculture is considered to be the most vulnerable sector because of its fragile relationship with environment (Raju & Phanindra, 2018). Globally, changes in temperatures and rainfall patterns are likely to have major effects on agricultural output (Lobell, Schlenker, & Costa-Roberts, 2011). For instance, extreme temperature frequency and intensity are anticipated to increase, which can destruct food production systems (Deryng, Conway, Ramankutty, Price, & Warren, 2014). The occurrences of both floods and droughts are in the same way expected to increase in intensity and frequency in the coming years as a result of climate change, which could reduce crop productivity (Lesk, Rowhani, & Ramankutty, 2016).
Although climate variability and change is a universal phenomenon, however, the adverse effects and climate-induced vulnerability are felt asymmetrically based on the adaptive capability of individual countries (Guiteras, 2009). Historically, significant portions studies on agricultural impacts of climate change have concentrated on developed countries (for instance, United States); however, developing countries are more vulnerable to climate change (Adger, Huq, Brown, Conway, & Hulme, 2003; ). Indian crop sector is one of the most sensitive and exposed areas to climate change due to less adaptive capacity to cope with it (Birthal, Khan, Negi, & Agarwal, 2014). The assessment of climate change effects on crop productivity in India is highly important where a substantial share of the population depends on farming for their livelihood and sustenance (Pattanayak & Kumar, 2014). As the changes in climatic factors directly affect agriculture, it is essential to investigate the effects of changes in climatic conditions on farm productivity. Hence, this study aims to investigate the effects of climate change major crop yields in India. The primary motivation for this study arises from the inordinate contribution of crop production to Indian economy and also the crop growing sector susceptibility to climate change.
1.1 Climate change and agriculture in India
Agriculture sector plays a greater role in economic development in India. Although a significant decrease in its portion of gross domestic product, the Indian agriculture sector continues an important part of the economy for the reason that of its strategic significance for food safety, job creation, and poverty reduction. Moreover, this sector still employs almost 52% of the labor force in the country. India is among most affected countries in terms of climate change and natural hazards due to its inadequate arable land, higher population, and dependence on agriculture, monsoon dependent farming, limited technological, and financial development for adaptation to climate change (Birthal et al., 2014). In particular, the crop productivity is likely to experience significant yield loss in coming days due to climate variability and extreme weather events for instance droughts and floods (Gupta, Sen, & Srinivasan, 2014).
Many previous studies have presented that India is anticipated to experience one of the world's highest losses of agricultural productivity in line with perceived climate change patterns and projected scenarios. The projections of climate change for India up to 2,100 shown that an overall upsurge in temperature by 2–4°C, and there might be no significant change in rainfall magnitude (Kavikumar, 2010). Similarly, it anticipated that a rise in average temperature will be 3–6°C, and precipitation will rise by 15–40% over India by the completion of the 21st century (National Communication Project, 2004). With regard to temperature, Intergovernmental Panel on Climate Change has anticipated the mean temperature upturn to be between 1.1–6.4°C by the completion of this century. The yield loss of rice crop owing to increased temperature of 1 to 2°C could cause in 3–17% in different parts of the country (Aggarwal & Mall, 2002). Another research finds that, after taking into consideration the influence of carbon fertilization, the predicted loss of agricultural output in India by 2,100 ranges from 10–40% (Aggarwal, 2008). Auffhammer, Ramanathan, and Vincent (2012) revealed that rice production is packed down by the famines and extreme precipitation in rain-fed zones of India during 1966–2002. Likewise, Saravanakumar (2015) projected that by end-of-century impacts for rice output could decline 10% relative to the reference line yield (1971–2009).
Therefore, it is essential for finding the association between climatic factors and primary crop yields in India over empirical analysis. Hence, the main objective of the present study is to empirically evaluate the effects of climate related variables on major crop yields in India. The residual part of the paper is structured as follows. Section 2 presents a brief literature review associated with climate and crop yields. The data sources and empirical methodology are described in Section 3. Section 4 provides the empirical findings and discussion. The last section presents conclusions and policy implications of the study.
2 LITERATURE REVIEW
The adverse effects of climate change and its variability on agriculture have attracted the primary concern of scholars, economists, and policymakers around the world. As a result, a large number of empirical studies have evaluated the association between climate change and crop yields using different methodologies. There are three major approaches to evaluate the effects of climate change on crop growing sector (Guiteras, 2009; Sarker, Alam, & Gow, 2014). These are (a) crop modeling approach (biophysical) that is also called as a production function approach, (b) Ricardian approach, and (c) econometric approach. The studies on effects of climate change on agriculture using these three major methodologies are discussed below.
Traditionally, crop simulation approaches are the most extensively used techniques to evaluate climate change impacts on crop production in the world. Crop modeling experiments investigating the effect of climatic factors are carried out in a controlled environment, or well-organized scientific labs tend to be location-specific (Sarker et al., 2014). However, these approaches do not take into account of the adaptive capacity as well as coping behavior of the crop growers, and hence, the adverse effects are overestimated, and positive impacts are underestimated (Guiteras, 2009; Mendelsohn & Dinar, 1999; Mendelsohn, Nordhaus, & Shaw, 1994). There is a considerable consensus among the crop modeling studies on the overall effects of climate change on crops across the world (Kim et al., 2015; Phillips, Lee, & Dodson, 1996; Tan & Shibasaki, 2003; Valizadeh, Ziaei, & Mazloumzadeh, 2014), and there might be distinction in the effects estimated for particular regions. In case of India, the most important regions for the crop modeling studies are the Indo-Gangetic Planes, Northwest India, and several other regions. Pathak et al. (2003) and Lal, Singh, Rathore, Srinivasan, and Saseendran (1998) have evaluated the effects of climatic factors on wheat and rice yields in Indo-Gangetic Planes and Northwest India, respectively. They found that wheat and rice productivity is more vulnerable to less radiation and greater nighttime temperature. Aggarwal and Mall (2002) projected the effects of climate change on rice by employing ORYZA1N as well as CERES-rice crop simulations across India for different zones. They found that a rise in temperature of 1 to 2°C and existing CO2 concentrations could lead to a 3–17% reduction in productivity across separate zones. Similarly, Kumar et al. (2011) evaluated the likely effects of climate change on both rain-fed and irrigated rice across the Western Ghats, northeast and coastal parts of India. But the rain-fed area of rice in the region could be reduced by up to 10%. Rain-fed rice is most susceptible in the northeast area, with a range of effects dropping from −35% to +5%. A latest study by Mukherjee and Huda (2018) found that potential output trend in Terai, Coastal Saline, and New Alluvial Zones of West Bengal have been declining. And also they suggested that improvement of productivity can be accomplished through the adoption, temperature tolerant seeds, and more sophisticated crop management methods.
Relatively recently, a number of empirical studies have been conducted to quantify the effects of climatic factors on crop yields using the “Ricardian approach.” Ricardian methodology estimates the association between land revenue or land value and agro-climate variables by using cross-sectional information (Mendelsohn et al., 1994). The major advantage of the Ricardian method is that it captures adaptation actions of farmers that influence land value as measured by farm income or net revenue. As a result, this approach has been effectively applied in a huge number of nations as follows: Taiwan (Chang, 2002), United States (Mendelsohn et al., 1994), and China (Wang et al., 2009). However, a limited number of studies applied the Ricardian framework in the context of Indian agriculture. Some of the studies using cross-sectional models include Dinar et al. (1998), Sanghi and Mendelsohn (2008), Kumar (2011), Kar and Das (2015), and Mishra, Sahu, and Sahoo (2016). Furthermore, a systematic study has done by Kumar and Parikh (2001) to evaluate the impact of climate sensitivity on crop productivity in Indian agriculture. However, the major of the Ricardian framework lies in its failure to include omitted variables such as soil quality and crop grower's unobservable skills (Barnwal & Kotani, 2013; Guiteras, 2009). Moreover, due to inefficient land markets, use of this approach might provide biased results in developing countries (Sarker et al., 2014).
In more recent times, the use of panel as well as time series data approach studies have been increasing much rapidly to evaluate the effects of climate change on crop yield. This approach pioneered by Auffhammer, Ramanathan, and Vincent (2006) and Deschênes and Greenstone (2007). Using these methods, several studies have assessed the effect of weather change on crop output across the world (for instance see Sarker et al., 2014; Agba et al., 2017; Zhang, Zhang, & Chen, 2017; Sbaouelgi, 2018; Attiaoui & Boufateh, 2019). Some of the studies employing the econometric models to examine effects of weather variation on crop productivity in India include Guiteras (2009), Moorthy, Buermann, and Rajagopal (2012), Barnwal and Kotani (2013), Gupta et al. (2014), Birthal et al. (2014), Farook and Kannan (2016), Nath and Mandal (2018), and Pal and Mitra (2018). Lastly, the overview of the literature reported that changes in climate have an adverse effect on food and non-food production. Maximum studies explore empirically the effect of climatic change on farm output based on single crop or more so. However, majority of the studies used crop modeling approach, but fewer investigations have been done using time series data. Moreover, there is no single comprehensive study using time series data of the recent period (till 2017), which is evident that significant rise in extreme climatic events. On the other hand, quantifying the overall influence of environmental change on primary food grain as well as non-food crops are a vital empirical issue. With this backdrop, the present study evaluates the effects of climate factors on seven major crops in India.
3 DATA AND METHODOLOGY
3.1 Data sources
The main aim of this study is to empirically explore the impact of climate change on the productivity of major crops in India for the time period 1961 to 2017. To achieve this objective, this study uses the annual time-series secondary data collected from various sources. This study selected seven food and non-food crops such as rice, wheat, pulses, rapeseeds and mustard, cotton, sugarcane, and groundnut. The data on agricultural productivity and crop-wise cultivated area were collected from Handbook of Statistics on Indian Economy, Reserve Bank of India. Information on Climatic variables namely actual rainfall, maximum temperature, and minimum temperature were gathered from the Ministry of Statistics and Programme Implementation, Government of India. The statistical descriptions (summary statistics) of the independent and explanatory variables for India were presented in Table 1.
| Variable | Observation | Mean | Standard deviation | Min | Max |
|---|---|---|---|---|---|
| Rice yield | 57 | 1,642.68 | 501.78 | 862 | 2,578 |
| Rice area | 57 | 40.92 | 2.99 | 34.69 | 45.54 |
| Wheat yield | 57 | 2,079.09 | 758.85 | 730 | 3,371 |
| Wheat area | 57 | 23.28 | 5.18 | 12.57 | 31.19 |
| Pulses yield | 57 | 566.02 | 99.06 | 377 | 841 |
| Pulses area | 57 | 23.22 | 1.69 | 20.35 | 29.99 |
| Cotton yield | 57 | 242.09 | 129.14 | 103 | 532 |
| Cotton area | 57 | 8.59 | 1.55 | 6.46 | 13.08 |
| Groundnut yield | 57 | 971.98 | 281.37 | 554 | 1,868 |
| Groundnut area | 57 | 791.35 | 288.76 | 300 | 1,397 |
| R&M* yield | 57 | 4.83 | 1.42 | 2.87 | 7.32 |
| R&M area | 57 | 6.87 | 0.98 | 4.59 | 8.71 |
| Sugarcane yield | 57 | 59,875.7 | 9,999.12 | 40,336 | 79,650 |
| Sugarcane area | 57 | 3.55 | 0.90 | 2.05 | 5.15 |
| Rainfall | 57 | 1,153.06 | 108.43 | 930.1 | 1,401.4 |
| Maximum temperature | 57 | 29.56 | 0.52 | 28.72 | 31.63 |
| Minimum temperature | 57 | 19.45 | 0.49 | 18.62 | 21.28 |
- Abbreviation: R&M, rapeseed and mustard.
3.2 Empirical methodology
The Intergovernmental Panel on Climate Change stated that change in climatic factors is emerging as one of the 21st century's serious problems. However, this study evaluates the effects of climatic factors on primary crop yields in India. To examine the impact of climate change on crop productivity, seven important food and non-food crops have been taken. Those are rice, wheat, pulses, rapeseeds and mustard, cotton, sugarcane, and groundnut. These seven crops largely cultivated based on monsoons and any change in climate, particularly precipitation and temperature would affect the yields of these crops significantly.
(1)εt is the error term. α is the intercept, and β1 to β4are the regression coefficients. This regression estimation has been done in STATA to fit the Equation 1.
4 RESULTS AND DISCUSSION
The following analysis deals with the empirical results and discussion of the study. Table 2 presents the empirical results of rice, wheat, and pulses crops. The R-squared value specifies that 87 % of the variation in the rice yield is explained by climate variables in India. Both minimum temperature and cropped area of rice are statistically significant. Moreover, area and minimum temperature are positively associated with rice yield. Higher minimum temperature leads to a higher yield of rice. Although rainfall and maximum temperature do not appear to be statistically significant, these factors are negatively related with the rice yield. This implies that higher rainfall and maximum temperature would mean lower the rice yield rates. The rice results are not consistent with Gupta et al. (2014) but more or less similar with Birthal et al. (2014). The reason for not getting similar results might be the different data series and time periods.
| Food crops | |||
|---|---|---|---|
| Variable | Rice | Wheat | Pulse |
| Area | 124.440*** | 127.044*** | 10.260 |
| Rainfall | −0.329 | −0.063 | 0.132 |
| Maximum temperature | −57.788 | 113.757 | 109.967** |
| Minimum temperature | 344.841** | 101.907 | 14.903 |
| Constant | −88.509*** | −62.830** | −34.826*** |
| No of observations | 57 | 57 | 57 |
| R-squared | 0.876 | 0.933 | 0.522 |
- *** p < .01
- ** p < .05.
The empirical results of the wheat crop are presented in Table 2. The R-squared value specifies that climate variables in India explain 93% of the variation in the wheat yield. All the climatic variables namely minimum and maximum temperatures and actual rainfall are not statistically significant in the wheat model. However, the cropped area of wheat is statistically significant and positively related with wheat yield. Although minimum and maximum temperatures do not appear to be statistically significant, these factors are negatively related with the wheat yield. Conversely, rainfall has an adverse impact on wheat yield. The wheat results are not consistent with Birthal et al. (2014). Furthermore, the empirical results of the pulses are presented in Table 2. The R-squared value indicates that 52% of the variation in the pulses yield is explained by climate variables. All the climatic variables namely minimum temperature and rainfall are not statistically significant in the pulses model. But only maximum temperature is statistically significant at the 5% level. However, all the explanatory variables are positively related with pulses yield. This implies that higher rainfall and temperatures lead to a higher yield of pulses in India.
Non-food crops estimates are made using the multiple regression approach. The contribution of the climate variables to the cotton, groundnut, rapeseed and mustard and sugarcane produce are presented in Table 3. The R-squared value specifies that climate variables in India explain 76% of the variation in the cotton yield. The statistical significance and sign of the estimated coefficients for the regressors are found to be different across the four non-food crop models. Rainfall and maximum temperature have a positive effect on cotton yields, and these variables are statistically significant. Further, minimum temperature is adversely related to cotton yield in India. This implies that that cotton yields decrease due to increase in maximum temperature. However, minimum temperature is statistically insignificant. The cotton results are somewhat consistent with Hebbar, Venugopalan, Prakash, and Aggarwal (2013) and Padakandla (2016).
| Non-food Crops | ||||
|---|---|---|---|---|
| Variable | Cotton | Groundnut | R&M | Sugarcane |
| Area | 58.508*** | −107.024*** | 134.114*** | 9,141.419*** |
| Rainfall | 0.214** | 0.940*** | 0.020 | 11.708* |
| Maximum temperature | 118.393** | 147.251 | 317.049*** | 6,209.374* |
| Minimum temperature | −42.336 | 154.475 | −110.947 | −3,921.56 |
| Constant | −35.159*** | −65.888*** | −76.073*** | −935.46 |
| No of observations | 57 | 57 | 57 | 57 |
| R-squared | 0.769 | 0.571 | 0.834 | 0.804 |
- Abbreviation: R&M, rapeseed and mustard.
- *** p < .01
- ** p < .05
- * p < .10.
Estimated R-squared at 0.57 indicate the overall significance of the groundnut model. The maximum and minimum temperatures have a positive effect on the groundnut yields in model, but these variables are not statistically significant. Similarly, mean rainfall have a positive impact on groundnut yields, and it is statistically significant. Based on the empirical results of groundnut, it can be said that the climatic variables have positive effect on productivity in India. The R-squared value indicates that 83% of the variation in the rapeseed and mustard yield is explained by climate variables. Maximum temperature and mean rainfall have a positive effect on rapeseed and mustard yields, but only maximum temperature is statistically significant. Furthermore, minimum temperature is negatively related to rapeseed and mustard yield in India, but it is not statistically significant. Moreover, the empirical results of sugarcane are presented in Table 3. The R-squared value specifies that climate variables in India explain 80% of the variation in the sugarcane yield. Mean rainfall and maximum temperature have a positive effect on sugarcane yields, and these variables are statistically significant. Furthermore, minimum temperature is adversely related to sugarcane yield in India. However, minimum temperature is statistically insignificant. The sugarcane results are to some extent consistent with Kumar, Sharma, and Ambrammal (2015).
5 CONCLUSIONS AND POLICY IMPLICATIONS
The study analyzes the impact of climatic variables on the productivity of major food and non-food crops in India during 1961–2017. The present study taken seven major crops namely rice, wheat, pulses, rapeseeds and mustard, cotton, sugarcane, and groundnut. These seven crops constituted approximately 70% of the overall gross cropped area of India. Study found that the statistical significance and sign of the estimated coefficients for the explanatory variables are found to be differ across the four non-food crops. An increase in rainfall has an adverse effect on food crops except for pulses; however, it has a positive relationship with non-food crops during the study period. Further, the average maximum temperature has a positive influence on food and non-food crops excluding rice. The average minimum temperature has an adverse impact on non-food crops, but it has a positive association with food crops. Conclusively, this found that crop yields are impacted differently with different climatic variables in India. This study recommends taking adaptation activities to cope with the adverse impacts of climate change. More specifically, this study recommends crop insurance, stress-tolerant variety seeds, irrigation facilities, proper credit, and modern inputs to deal with the possible losses to farmers due to climate change.
Biography
Raju Guntukula is currently a Senior Research Fellow (UGC) in Economics at the School of Economics, University of Hyderabad, Telangana, India. His area of expertise is Economics of climate change, Agricultural economics and Environmental economics with an application of econometric methods. He has published several research articles in reputed national and international journals.
