How Far Will Climate Change Affect Future Food Security? An Inquiry into the Irrigated Rice System of Peninsular India

Abstract

Climate change poses a great challenge to food security, particularly in developing nations where important food crops such as rice and wheat have been grown in large quantities. The study investigates food security using an integrated approach, which comprises forecasting future rice production using the AquaCrop model and demand for rice using an economic model. The proposed approach was evaluated in the Cauvery delta zone in the eastern part of Tamil Nadu, which is a major rice-growing hotspot in peninsular India. Our results showed that the future rice productivity of the Cauvery delta region would be reduced by 35% between 2021 and 2040 and by 16% between 2041 and 2050. However, the supply–demand gap addressing food security in the Cauvery delta zone is positive for the future, as evidenced by the availability of surplus rice of 0.39 million tonnes for the period 2021–2030 and 0.23 million tonnes and 0.35 million tonnes for the periods 2031–2040 and 2041–2050, respectively. Nevertheless, as the neighboring regions are relying on rice production from the Cauvery delta, this surplus rice production is potentially not sufficient to meet the demand of the state as a whole, which suggests climate change may pose a severe threat to the food security of the Tamil Nadu State. These findings emphasize the necessity of performing regional-level food security assessments with a focus on developing location-specific policy options to mitigate the adverse effects of climate-induced anomalies on food security.

1. Introduction

Climate change is undeniable, and poses a severe threat to food production and, consequently, food security [1,2,3,4]. The temperature rises, changes in precipitation patterns, and an increase in the frequency and intensity of extreme events are all attributed to the changing climatic conditions [4]. Agriculture is one of the primary sectors that are most vulnerable to climate change-induced disparities [5,6]. For instance, the increase in the duration and magnitude of heat stress and water stress alters the agricultural growing season [7,8,9]. Article 2 of the United Nations Framework Convention on Climate Change (UNFCCC) recommends strengthening the global response to mitigate the adverse effects of climate change.

Numerous studies have examined the linkages and consequences of climate change and food security at various spatial scales [10,11,12,13,14,15,16,17,18,19]. Climate change affects food security in complex ways, which affects food production directly through changes in agroecological conditions and indirectly by affecting the growth and distribution of incomes, thus creating demand for agricultural produce [9,20,21,22]. Changes in rainfall and temperature are the prime drivers that determine agricultural productivity [9,23,24] and induce significant anomalies in food production [25]. Further, the occurrence of extreme climate events that induce frequent flash floods may also result in substantial yield losses [9,26]. With rising concerns over food security, predicting future food production for the changing climate is essential.

The impacts of climate change are likely to be severe for India [27,28] owing to its heavy dependence on agriculture. Despite the modest contribution to the national gross domestic product (GDP), India’s agriculture sector remains vital [29], sustaining the livelihood of nearly 70% of the Indian population [30]. The Intergovernmental Panel on Climate Change (IPCC) projected that temperature is likely to rise by 2 to 4.7 °C, with the most probable level being around 3.3 °C by the year 2100. Similarly, an increase in the frequency and intensity of extreme weather events has a substantial influence on food production, eventually, food security [31]. Furthermore, the impact of climate change further extends to the water table, rising salinity, irrigation water quality, soil fertility, soil moisture, and the increasing resistance of pests to pesticides in various parts of India, causing great concern [32,33]. The problem is further aggravated by diminishing per capita availability of arable land and sluggish climate change adaptation, which will lower the food supply, exacerbating the challenge of fulfilling food demand [34]. Even though India’s economy has been flourishing in recent times, agricultural production is still at an alarming level. As per the Global Hunger Index 2020, India ranks 94th out of 107 countries with a score of 27.2, which falls under the severe category [35].

Rice feeds approximately 557 million people in Asia [15], which accounts for nearly 87 percent of global rice consumption; consumed primarily as a staple food and feeds more people than any other crop [36]. Projections also suggest that rice consumption will increase to 555 million tonnes in 2035 [37]. Rice has been cultivated both under irrigated as well as rainfed conditions; nonetheless, the former accounts for 75 percent of global rice production, which signifies the importance of irrigated rice on global food security. Nevertheless, climate change could utterly threaten rice production; studies that have quantified the climate change impacts on rice yield in different regions have shown both positive [17,24,38,39] and negative impacts [15,25].

The impact of climate change on food security is generally quantified by employing either crop simulation models or economic models. The crop simulation models are widely used by agronomists and meteorologists to quantify the impact, which has been assessed by employing simulation processes under different climate conditions [40,41,42,43,44]. On the other hand, economists deploy the Ricardian model [45,46] and/or the panel data approach [47,48,49] to examine the significance of climate variables on food security. However, both approaches quantify the impact on the supply side, assuming that either reduction or increment in food production causes food security and does not account for the demand side. Indeed, the increment in agricultural production as a result of climate change would not mean better food security unless demand has been accounted for in the estimation. A comprehensive study assessing the impact of climate change on food security by accounting for both supply of food and the demand for food is still pending.

In this study, irrigated rice is used as an indicator crop for food security assessment in the Cauvery delta region, which is a major rice-growing hotspot in peninsular India. The overarching aim of this study is to evaluate the future food security of the Cauvery delta zone. The specific objectives are (i) to assess the impact of future climate on rice production using the crop-modeling approach; (ii) to estimate the future rice demand using an economic model; (iii) to evaluate the food security of the Cauvery delta zone for the next few decades.

2. Materials and Methods

2.1. Study Region

The Cauvery delta zone is one of the agroclimatic zones of India, located in the southern part of the country, and has been considered as the rice bowl of the Tamil Nadu State. The delta region contributes 30 percent of the state’s rice production, a significant food security source for the state. Rice is the major crop of this zone, accounts for 65 percent of the cropping area, and has been cultivated primarily under irrigated conditions. The zone lies in the eastern part of Tamil Nadu and is presented in Figure 1.

The annual rainfall of the Cauvery delta zone is 1051.27 mm, and it receives 50 percent of its annual rainfall during the northeast monsoon, majorly covering the months of October, November, and December. The zone is mostly warmer in April, May, and June, during which the temperature will reach 36 °C. Figure 2 shows the comprehensive climate profile of the Cauvery delta zone.

A rice-fallow-pulse cropping system is principally practiced in the delta region, in which pulse crops are cultivated after the harvest of the rice crop. Hence, rice (65 percent), black gram (14 percent), and green gram (10 percent) are the major crops in the delta region (Figure 3).

Rice is cultivated in four seasons in the Cauvery delta zone, namely Kuruvai, Samba, Thaladi, and Navarai. The cropping calendar (Figure 4) for an agricultural year starts from the Kuruvai season, which starts in June. During the Kuruvai season, short-duration varieties of rice have been cultivated. Farmers with well irrigation cultivate the Kuruvai crop since canal water reaches the delta by the last week of June or the first week of July. Samba season starts during the first week of August and ends in mid-January, which accounts for the highest share of cropped area (

Table 1. Season-wise cropped area of the Cauvery delta zone.

Period	Rice Cultivated Area (Million ha)
	Kuruvai	Samba	Thaladi
1989–1998	0.074	0.370	0.353
1999–2008	0.076	0.327	0.064
2009–2018	0.086	0.333	0.079

Note: data represent the decadal average of the respective periods. Source: Department of Agriculture, Government of Tamil Nadu.

) among the cropping seasons. During the Samba season, long-duration varieties are cultivated owing to climatic factors of the zone, and it is observed that the canal water is available for the entire samba season. Hence, canal irrigation is predominantly followed during Samba rice cultivation. Thaladi cropping season starts in September and ends during mid-January. Farmers who raise the Kuruvai crop opt for Thaladi and cultivate either short or medium-duration rice cultivars. Since canal water is available until January, it is the major source of irrigation for cultivation during the Thaladi season. Farmers who raise short-duration cultivars during Thaladi opt for Navarai, which starts in December and ends in mid-March. The farmers who have access to supplementary irrigation opt for the Navarai season and cultivate short-duration cultivars.

The farmer who raises either Samba or medium-duration varieties during Thaladi cultivates black gram and green gram as rice fallow. Both are cultivated during mid-January (after the Samba/Thaladi crop), harvested last week of March, and raised mostly under rainfed conditions. The major cropping system, followed in the Cauvery delta zone is (i) rice–pulses; (ii) rice–rice–pulses; (iii) rice–rice–rice.

2.2. Data and Sampling

This study used four types of data, namely, climate data, farmers’ survey data, consumer survey data, and secondary data (i.e., area, production, and yield) for food security assessment. The first one is the climate data, which is projected based on the regional climate model (RCM)–GFDL at the one-degree resolution and data were originally processed and extracted from a previously published study [50].

The second type are the farmer survey data, which are the primary data collected through a random sampling procedure. The farm information, such as crop variety, seed rate, date of sowing, date of harvesting, yield, soil type, number of irrigations, source of irrigation, etc., were collected through personal interviews using well-structured and pre-tested interview schedules. A total of 450 samples were collected from nine villages (i.e., 50 samples from each village) in the Cauvery delta zone. However, concerns have been accounted for the distribution of samples across both the delta regions (i.e., six villages from the old delta region and three villages from the new delta region) and the data collected pertain to the agriculture year of 2017–2018.

The consumption data at the household level were accessed from the 68th round (2011–2012) of the National Sample Survey (NSS). It provides per capita monthly consumption (in terms of quantity and value) of food items in a 30-day recall period. This study accounted for 750 sample households from the national sample survey pertaining to the Cauvery delta zone. Finally, the study used secondary data, the data related to the cropping patterns of the Cauvery delta zone, which were collected from the Director of Economics and Statistics, Ministry of Agriculture and Farmers Welfare, Government of India [51], and the population data, which are projected, based on census 2011 [52].

2.3. Food Security Assessment Framework

This study accounts for (i) the projections on the rice supply and (ii) the projections on demand for rice during 2021–2050 (Figure 5). The projections on the rice supply were simulated for future climatic conditions using the AquaCrop model. The demand for rice includes the estimation of the expenditure elasticity of rice and the estimation of demand for rice. The difference in supply and demand will reflect in food security.

The study made the following assumptions to estimate robust and reliable results:

(i)

The technology is assumed to be constant during the study period;

(ii)

The cropping area is assumed to be constant during the study period;

(iii)

Except for climate, all the other factors (i.e., technology, input usage, etc.) that are associated with crop yield are assumed to be constant during the study period;

(iv)

Rice is being assumed as the primary food crop for the Cauvery delta region during the study period;

(v)

The consumption behavior and the real income of the consumer of the Cauvery delta zone are assumed to be constant during the study period;

(vi)

The contribution from the upper catchment to the delta’s irrigated rice would remain unaltered or constant in the future.

2.3.1. Projections on Supply of Rice through AquaCrop Modeling

The study used the AquaCrop model to estimate crop-yield projections for future climatic conditions. AquaCrop demands minimum input data, which are easily obtainable; nevertheless, it achieves a great balance between simplicity, output accuracy, and robustness [53]. AquaCrop simulates crop growth and yield based on climate parameters (i.e., rainfall, temperature, carbon dioxide concentration, and reference evapotranspiration), crop characteristics (phenology, green canopy cover, root depth, harvest index, water productivity (WP), and stress responses), soil (field capacity, saturated hydraulic conductivity, permanent wilting percentage), and management (irrigation, plant density, fertility) factors. AquaCrop projects the crop yield by multiplying biomass with the harvest index (Figure 6). The crop biomass is simulated from the daily transpiration rate, reference evapotranspiration, and water productivity [53,54].

Model validation

Model validation was conducted by using the default conservative parameters [54,55] calibrated by FAO for rice. However, the phenological characteristics of local cultivars and local management parameters that characterize the Cauvery delta zone were also taken into account for reliable estimates. Rice cultivar parameters were collected from the previous studies conducted in the Cauvery delta zone (refer to

Table 2. Crop parameters of the four major rice cultivars used for this study.

Particulars	Unit	CR1009D	CR1009T	ADT38	ADT43	ASD16	Source
Initial canopy cover	%	2.5	4.95	7.5	9.9	9.9	[56]
Plant density	plants per ha	500,000	330,000	500,000	660,000	660,000	[56,57,58]
Maximum canopy cover (CCx)	%	80	85	85	85	80	[56]
Reference harvest index (HIo)	%	41	41	42	40	35	[56,59]
From transplanting to recovered transplant	Days	7	7	7	7	7	[60]
From transplanting to maximum canopy cover	Days	65	62	39	30	32	[60]
From transplanting to starting senescence	Days	120	100	90	70	75	[60]
From transplanting to maturity	Days	150	125	110	88	93	[60]
Maximum effective rooting depth	Meter	0.5	0.5	0.5	0.5	0.5	[58]

. Source). The details of the phenological characteristics of the local cultivars considered in this study are listed in Table 2.

Scenario generation for future yield simulation

In order to achieve reliable and robust results, the following scenarios were constructed.

Season: In the Cauvery delta zone, rice has been cultivated in four major seasons, namely Kuruvai, Samba, Thaladi, and Navarai. The same has been considered for constructing the scenarios for crop model estimation.

Soil: The Cauvery delta region is predominantly covered by two soil types. The old delta region is primarily covered by clay soil, whereas sandy loam soil is largely present in the new delta region. The soil characteristics of Rice Research Station, Aduthurai, have been taken into account for crop model estimation for the old delta region, whereas the soil characteristics of Soil and Water Conservation Institute, Thanjavur, have been used for the new delta region.

Cultivars: The cultivars which are cultivated predominantly in the Cauvery delta zone were also considered to construct scenarios. Based on the survey results, four major cultivars were identified, and the details of the cultivars are presented in

Table 3. Major rice cultivars used for yield estimation using the AquaCrop model.

Cultivar	Season	Basin
ADT43	Kuruvai	Old delta
ASD16	Kuruvai/Navarai	Old/New delta
CR1009	Samba	Old/New delta
ADT38	Thaladi	Old/New delta

.

Sowing window: Crop yield significantly varies with the date of sowing under climate variability conditions. The farmer survey found that the date of sowing varies across the first month of the cropping season. Hence, the first month of a crop season was divided into three sowing windows following 10-day intervals [61], and the middle of the sowing (5th day) window was counted for the date of sowing. The details of the sowing windows of multiple seasons used to construct the modeling scenarios for yield estimation are presented in

Table 4. Sowing window considered for AquaCrop model simulation.

Season	Date of Sowing
	1st Sowing Window	2nd Sowing Window	3rd Sowing Window
Kuruvai	27 June 2021	7 July 2021	17 July 2021
Samba (Transplanted)	9 September 2021	19 September 2021	29 September 2021
Samba (Direct sown)	5 August 2021	15 August 2021	25 August 2021
Thaladi	30 September 2021	10 October 2021	20 October 2021
Navarai	26 December 2021	6 January 2022	16 January 2022

.

Type of sowing: The type of sowing determines the duration of the crop as well as the production, which is significant under climate variability conditions. In the Cauvery delta zone, rice is cultivated both as a transplanted crop and as a direct-seeded crop. The farmers’ survey highlighted that the direct-seeded rice shortens the duration from 10 to 15 days. Hence, that type of sowing was also taken as a parameter to construct scenarios for crop model estimation.

To estimate the most reliable and robust results, the crop model needs to capture all the possible variables that determine crop production. Hence, 30 scenarios were constructed for crop model simulation, and the details of the scenarios used in this study are presented in

Table 5. Scenarios employed for AquaCrop model simulation.

Scenario	Season	Type of Sowing	Sowing Date	Cultivar	Soil	Basin
S1	Kuruvai	Transplanted	27 June 2021	ADT43	Paddy soil	Old delta
S2	Kuruvai	Transplanted	7 July 2021	ADT43	Paddy soil	Old delta
S3	Kuruvai	Transplanted	17 July 2021	ADT43	Paddy soil	Old delta
S4	Kuruvai	Transplanted	27 June 2021	ASD16	Sandy loam	New delta
S5	Kuruvai	Transplanted	7 July 2021	ASD16	Sandy loam	New delta
S6	Kuruvai	Transplanted	17 July 2021	ASD16	Sandy loam	New delta
S7	Samba	Transplanted	9 September 2021	CR1009	Paddy soil	Old delta
S8	Samba	Transplanted	19 September 2021	CR1009	Paddy soil	Old delta
S9	Samba	Transplanted	29 September 2021	CR1009	Paddy soil	Old delta
S10	Samba	Transplanted	9 September 2021	CR1009	Sandy loam	New delta
S11	Samba	Transplanted	19 September 2021	CR1009	Sandy loam	New delta
S12	Samba	Transplanted	29 September 2021	CR1009	Sandy loam	New delta
S13	Samba	Direct sown	5 August 2021	CR1009	Paddy soil	Old delta
S14	Samba	Direct sown	15 August 2021	CR1009	Paddy soil	Old delta
S15	Samba	Direct sown	25 August 2021	CR1009	Paddy soil	Old delta
S16	Samba	Direct sown	5 August 2021	CR1009	Sandy loam	New delta
S17	Samba	Direct sown	15 August 2021	CR1009	Sandy loam	New delta
S18	Samba	Direct sown	25 August 2021	CR1009	Sandy loam	New delta
S19	Thaladi	Transplanted	30 September 2021	ADT38	Paddy soil	Old delta
S20	Thaladi	Transplanted	10 October 2021	ADT38	Paddy soil	Old delta
S21	Thaladi	Transplanted	20 October 2021	ADT38	Paddy soil	Old delta
S22	Thaladi	Transplanted	30 September 2021	ADT38	Sandy loam	New delta
S23	Thaladi	Transplanted	10 October 2021	ADT38	Sandy loam	New delta
S24	Thaladi	Transplanted	20 October 2021	ADT38	Sandy loam	New delta
S25	Navarai	Transplanted	26 December 2021	ASD16	Paddy soil	Old delta
S26	Navarai	Transplanted	6 January 2022	ASD16	Paddy soil	Old delta
S27	Navarai	Transplanted	16 January 2022	ASD16	Paddy soil	Old delta
S28	Navarai	Transplanted	26 December 2021	ASD16	Sandy loam	New delta
S29	Navarai	Transplanted	6 January 2022	ASD16	Sandy loam	New delta
S30	Navarai	Transplanted	16 January 2022	ASD16	Sandy loam	New delta

.

Accounting for postharvest losses and milling

For reliable estimates, postharvest losses have also been taken into account; thus, this study accounted for 9 percent [62] postharvest loss of rice. Furthermore, rice is only consumed in processed forms; raw rice must be milled before being consumed. Hence, the study used a conversion factor of 66.60 percent [63] to convert brown rice to white rice.

2.3.2. Projections on Demand for Rice through the Demand Model

The model that has been adopted in this study to estimate the demand for rice has been used in many studies and the mathematical description of the model is described below:

$$D_{i, t}=d_{i,0}\ast N_t{\left(1+r^{* }{\eta }_i^y\right)}^t$$

(1)

where

$D_{i, t}$ is the total household demand for $ith$ commodities of the selected region for the year ${}^{‘}{t}^{’}$

$d_{i,0}$ is per capita demand of ith commodities during the base year 2011–2012

‘$r$’ is the growth in per capita GDP between the ${}^{‘}{0}^{’}$ and ${}^{‘}{t}^{’}$ periods

${\eta }_i^y$ is the expenditure elasticity of demand of the $ith$ commodity, and

$N_t$ is the projected population during the year ${}^{‘}{t}^{’}$

Estimation of elasticity—Almost Ideal Demand System (AIDS) Model

The study employed the almost ideal demand system (AIDS) model of [64] to estimate expenditure elasticity. The AIDS model derives the share equations in an n-good system from a specific cost function.

$$w_i={\alpha }_i+{{\displaystyle \sum }}_{j=1}^n{\gamma }_{ij}\ln p_j+{\beta }_{i}ln\left(\raisebox{1ex}{$X$}\!\left/ \!\raisebox{-1ex}{$P$}\right.\right)$$

(2)

where $w_i$ is the share of the $ith$ good, ${\alpha }_i$ is the constant coefficient in the $ith$ share equation, ${\gamma }_{ij}$ is the slope coefficient of the $jth$ good in the $ith$ share equation, and $p_j$ is the price on the $jth$ good. $X$ is the total expenditure on the system of goods provided by

$$X={{\displaystyle \sum }}_{i=1}^n{p}_i{q}_i$$

(3)

where $q_i$ is the quantity demanded of the $ith$ good; $P$ is the price index.

Deaton and Muellbauer [64] also suggested a linear approximation of the nonlinear AIDS model by specifying a linear price index given by

$$\ln P={{\displaystyle \sum }}_{i=1}^n{w}_{i}ln{p}_i$$

(4)

which gives rise to the linear approximate AIDS (LA-AIDS) model. In practice, the LA-AIDS model is more often estimated than the nonlinear AIDS model. Conservation entails the following restrictions on the parameters in the nonlinear AIDS model:

$${{\displaystyle \sum }}_{i=1}^n{\alpha }_i=1, {{\displaystyle \sum }}_{i=1}^n{\beta }_i=0, {{\displaystyle \sum }}_{i=1}^n{\gamma }_{ij}=0$$

(5)

Homogeneity is satisfied if and only if, for all $i$

$${{\displaystyle \sum }}_{j=1}^n{\gamma }_{ij}=0$$

(6)

Symmetry is satisfied if

$${\gamma }_{ij}={\gamma }_{ji}$$

(7)

Parameters obtained from the AIDS model can be used to measure the elasticity of consumption goods.

(i)

Own-price elasticity

$$E_{ii}=-1+\frac{b_{ii}}{w_i}-c_i$$

(8)

(ii)

Cross-price elasticity

$$E_{ij}=\frac{b_{ij}}{w_i}-\frac{c_i}{w_i}{w}_j$$

(9)

(iii)

Income elasticity

$${\eta }_i=1+\frac{c_i}{w_i}$$

(10)

The nature of the demand for food could directly be inferred from the signs of the AIDS parameters. Commodities with positive parameters ($c_i$ > 0) are income elastic, and those with negative parameters ($c_i$ < 0) are income inelastic. Equally, commodities with negative own-price parameters ($b_{ij}$ < 0) are price elastic, and those with positive parameters ($b_{ij}$ > 0) are price inelastic. The price coefficients ($b_{ij}$) imply a change in the $ith$ budget share owing to a proportionate change in the price keeping the real income constant.

This study assessed food security using a crop model and economic modeling approach. The AquaCrop model simulated 30 field scenarios for estimating future rice yields. Rice demand was estimated using the economic model, which requires per capita consumption, per capita income, income elasticity, and population of the respective year. The income elasticity was estimated by employing the Almost Ideal Demand System (AIDS) model. The per capita consumption was derived from the consumption data, and the information on population was retrieved from the census of India. The results of the future projections are presented in terms of both old and new delta regions representing the next three decades (2021–2050).

3. Results

3.1. Characteristics of the Projected Climate Data

This study used GFDL climate data to predict the impact of climate change on rice production in the irrigated ecosystem, and the characteristics of the climate data are presented in Figure 7. It can be seen that during decade I (2021–2030), the new delta region (Figure 7A) received its maximum rainfall in September (112 mm), followed by October (106 mm), April (97 mm), November (96 mm) and August (95 mm). This implies that the region receives most of its rainfall during the northeast monsoon, starting from October to December. The region is warmer during April (39 °C) and May (40 °C), and for the rest of the months, the temperature ranges between 31 °C and 37 °C. The annual rainfall (

Table 6. Rainfall and temperature change comparison for both old and new delta regions.

Particulars	New Delta	Old Delta
Normal
Annual rainfall (mm)	1022	1067
Maximum temperature (°C)	33	33
Projected
Decade I (2021–2030)
Annual rainfall (mm)	817	881
Change in rainfall (%)	20	17
Maximum temperature (°C)	35	35
Change in temperature (°C)	2	2
Decade II (2031–2040)
Annual rainfall (mm)	944	1005
Change in rainfall (%)	8	6
Maximum temperature (°C)	35	35
Change in temperature (°C)	2	2
Decade III (2041–2050)
Annual rainfall (mm)	751	773
Change in rainfall (%)	26	27
Maximum temperature (°C)	36	37
Change in temperature (°C)	3	4

) of the region is 817 mm in decade I (2021–2030), which is significantly less than its normal rainfall (1022 mm), and the change is −20 percent. It is further observed that the temperature of the new delta region increases by 2 °C from 33 °C to 35 °C during decade I (2021–2030). The climate characteristics of the old delta region (Figure 7D) during decade I reveal that the climate of this region is identical to the new delta region, implying that in both regions, similar climate conditions prevail during decade I (2021–2030).

Figure 7B shows the climograph of the new delta region during decade II (2031–2040). Similar to decade I (2021–2030), the region receives most of its rainfall during the northeast monsoon and is warmer during April (39 °C) and May (40 °C). The annual rainfall (Table 6) of the region is 944 mm, which is 8 percent less than its normal rainfall (1022 mm). The mean temperature of the region is 35 °C, which is 2 °C higher than its normal temperature (33 °C). It can be observed (Figure 7E) that there is no significant change in the climate pattern of the old delta region during decade II (2031–2040) in comparison with decade I (2021–2030); the region receives most of its rainfall during the northeast monsoon, and is warmer during April and May. However, the magnitude of the rainfall (Table 6) is higher (1005 mm/annum) during decade II (2031–2040) than in the previous decade (2021–2030)

Figure 7C,F display the climate of both the new delta region and the old delta region of the Cauvery delta zone during decade III (2041–2050). It can be seen that the climate pattern of both the regions of the Cauvery delta zone is not identical to the previous decades (i.e., decade I and decade II). Instead of receiving most of its rainfall during the northeast monsoon, the rainfall is spread across the months. Furthermore, both regions receive significantly less rainfall—the new delta region receives 751 mm/annum, and the old delta region receives 773 mm/annum—comparatively. It is further observed that both regions are comparatively warmer than the previous decades (i.e., decade I and decade II), and the temperature rise is 3 °C in the new delta region and 4 °C in the old delta region against their respective normal temperatures.

3.2. Projections on Rice Production

The study found that there is a significant variance in yield, soil, and cropping pattern over space between the new delta and the old delta regions of the Cauvery delta zone. Consequently, the results of the rice supply projections are presented in terms of space and time.

Rice yield projections from the AquaCrop model

The future rice yield of the Cauvery delta zone during 2021–2050 was estimated using the AquaCrop model, and the results are shown in Figure 8. It can be seen that during decade I (Figure 8A), the rice yield of the new delta region is 5.85 tonnes per ha, and the highest yield is observed during Samba season (6.94 tonnes/ha) followed by Thaladi (6.52 tonnes/ha), Navarai (4.50 tonnes/ha) and Kuruvai (4.49 tonnes/ha). It is further noticed that the rice yield is comparatively lower in the old delta region, and the yield is 2.35 tonnes per ha, where the highest yield is observed during the Navarai season (3.49 tonnes/ha), while the lowest is in the Kuruvai season (1.65 tonnes/ha).

The results for decade II (Figure 8B) reveal that similar to decade I, the rice yield during decade II is comparatively higher in the new delta region (6.14 tonnes/ha). In the new delta region, the highest yield is observed in the Samba season (7.23 tonnes/ha), while the lowest is noticed in the Kuruvai season (4.65 tonnes/ha). Conversely, the old delta region accounted for the lowest yield (2.31 tonnes/ha), and it could be seen that across the different crop seasons of the old delta region, the maximum yield is observed in the Navarai season (3.51 tonnes/ha) followed by Thaladi (2.80 tonnes/ha), Samba (1.80 tonnes/ha) and Kuruvai (1.61 tonnes/ha). The yield estimates for decade III (Figure 8C) reveal that the rice yield follows a similar trend to the previous decades; the yield is comparatively higher in the new delta region (6.14 tonnes/ha) against the old delta region (2.31 tonnes/ha). As far as crop season is concerned, in the new delta region, the maximum yield is observed in the Samba season (7.48 tonnes/ha) and the lowest in the Kuruvai season (4.57 tonnes/ha), while in case of the old delta region, the highest yield is noticed in the Thaladi season (4.00 tonnes/ha) and the lowest in the Kuruvai season (2.47 tonnes/ha).

Keeping the other variables (i.e., factors of rice production) constant, the crop model estimates the rice yield with respect to climate factors. Thus, any deviation in the estimated rice yield against the normal yield (4.55 tonnes/ha) of the zone is primarily the result of climate factors, and the results are depicted in Figure 9. The results reveal that during decade I (Figure 9A), the rice yield increases by 28.64 percent in the new delta region, and the maximum yield increase is observed during Samba season (52.60 percent) followed by Thaladi (43.39 percent), while the rice yield declined by one percent during Kuruvai and Navarai season against its normal yield. Conversely, the impact of climate on rice yield is negative in the old delta region, and the rice yield declines by 48.40 percent against its normal yield, and the highest impact is observed during Kuruvai (63.82 percent), followed by Samba (51.23), Thaladi (50.76 percent), and Navarai (23.26 percent). During decade II (Figure 9B), the rice yield in the new delta region increases by 34.86 percent, and the highest increment is noticed during Samba season (59.00 percent) followed by Thaladi season (49.15 percent), while the impact is marginal during Navarai (2.72 percent) and Kuruvai (2.29 percent). The rice yield declines by 49.30 percent in the old delta region against its normal yield during decade II, and the reduction is maximum in the Kuruvai season (64.65 percent) followed by Samba (60.35 percent), Thaladi (38.39 percent), and Navarai (22.75 percent). The results for decade III (Figure 9C) show that in the new delta region, the rice yield increases by 38.49 percent against its normal yield during decade III. It is further observed that the increment in the rice yield is highest in the Samba season (64.38 percent), followed by Thaladi (54.29 percent), while it is marginal in the case of Navarai (6.96 percent) and Kuruvai (0.33 percent).

Conversely, the rice yield declined by 27.67 percent in the old delta region during decade III, and the maximum reduction is noticed during the Kuruvai season (45.71 percent), followed by Samba (32.24 percent), Navarai (16.00 percent), and Thaladi (12.15 percent).

Rice production estimate

The projections of rice production have been derived by multiplying the projected yield with the cultivated area (i.e., the long-term average area under cultivation) of the region, and the results are presented in

Table 7. Projections of future rice production (supply) of the Cauvery delta region for the period 2021–2050.

Particulars	Decade I (2021–2030)	Decade II (2031–2040)	Decade III (2041–2050)
New Delta
Yield (tonnes/ha)	5.85	6.14	6.30
Area (million ha)	0.089	0.089	0.089
Production (million tonnes)	0.520	0.545	0.560
Old Delta
Yield (tonnes/ha)	2.35	2.31	3.29
Area (million ha)	0.419	0.419	0.419
Production (million tonnes)	0.983	0.966	1.38
Cauvery Delta Zone
Area (million ha)	0.508	0.508	0.508
Production (million tonnes)	1.50	1.51	1.94

. The area under rice cultivation in the new delta region is assumed as 0.089 million ha during the projected period, and its production is 0.520 million tonnes during decade I (2021–2030), while it is 0.545 million tonnes and 0.560 million tonnes in decade II (2031–2040) and decade III (2041–2050), respectively.

The rice cultivation area of the old delta region is 0.419 million ha for the projected period, and its production is 0.983 million tonnes during decade I (2021–2030), and it is 0.966 million tonnes, and 1.38 million tonnes during decade II (2031–2040), and decade III (2041–2050), respectively. Thus, it is assumed that rice will be cultivated on 0.508 million ha in the Cauvery delta zone during the projected period. Its production is 1.50 million tonnes during decade I (2021–2030), and it is 1.51 million tonnes and 1.94 million tonnes during decade II (2031–2040) and decade III (2041–2050), respectively.

Impact of climate change on rice production

The impact of climate change on rice production was derived by taking the difference between projected rice production and normal rice production (i.e., decadal average), and the results are presented in Figure 10. It can be seen that the impact of climate change on rice production is positive in the new delta region, and the impact is highest during decade III (38 percent), followed by decade II (35 percent) and decade I (29 percent). Conversely, the impact of climate change on rice production is negative in the old delta region, and it is maximum during decade II (49 percent), followed by decade I (48 percent) and decade III (28 percent). Thus, the impact of climate change on rice production in the Cauvery delta region is negative during the predicted periods, and it is 35 percent during decade I and decade II, whereas it is 16 percent during decade III.

3.3. Projections of Demand for Rice

The parameters such as monthly per capita consumption (MPC), monthly per capita consumption expenditure (MPCE), expenditure elasticity, expenditure share, and marginal expenditure share explain the magnitude of the rice consumption behavior of the people of the Cauvery delta zone and the results presented in

Table 8. The magnitude of rice consumption behavior of the people in the Cauvery delta region.

Particulars	Value
MPC (in Kg)	9.72
MPCE (in INR)	137.05
Expenditure elasticity	0.82
Expenditure share (%)	22.08
Marginal expenditure share (%)	18.08

. It can be seen that the monthly per capita consumption of rice in the Cauvery delta zone was 9.72 kg, which cost 137.05 INR and shared 22.08 percent of the food consumption expenditure basket. It is further observed that the expenditure elasticity was 0.82, which implies the less income elastic behavior of the people in the Cauvery delta zone.

The demand projections of the Cauvery delta zone were derived using parameters such as projected population, estimated per capita consumption, and expenditure elasticity. The results are presented in

Table 9. Projections of future rice demand of the Cauvery delta region for 2021–2050.

Period	Population (Million)	Elasticity	Demand (Million Tonnes)
Decade I (2021–2030)	9.69	0.82	0.609
Decade II (2031–2040)	10.33	0.82	0.774
Decade III (2041–2050)	10.78	0.82	0.933

. It can be seen that the projected population of the Cauvery delta zone is 9.69 million during decade I, and it will be 10.33 million and 10.78 million during decade II and decade III, respectively. As far as demand is concerned, the projected demand for rice in the Cauvery delta zone is 0.609 million tonnes during decade I. During decades II and III, the projected rice demand is 0.774 million tonnes and 0.933 million tonnes, respectively.

3.4. Projections on Food Security through the Rice Supply–Demand Assessment

The produced rice (i.e., brown rice) needs to be processed (milled) for final consumption (i.e., white rice), and studies [63] revealed that the milling percentage of rice is an average of 66.6 percent, and it has been used for deriving the supply of white rice.

Table 10. Projections of future rice demand–supply gap of Cauvery delta region.

Period	Future Rice Production (Million Tonnes)	Future White Rice Availability for Consumption * (Million Tonnes)	Future Rice Demand (Million Tonnes)	Rice Supply and Demand Difference (Million Tonnes)
Decade I (2021–2030)	1.50	1.01	0.609	0.392
Decade II (2031–2040)	1.51	1.06	0.774	0.232
Decade III (2041–2050)	1.94	1.30	0.933	0.358

* Note: the milling percentage is 66.6 percent.

presents the gap between the demand for and supply of rice in the Cauvery delta zone during the projected period. The results reveal that the supply of rice (i.e., white rice) is 1.01 million tonnes during decade I, and it is 1.06 million tonnes and 1.30 million tonnes during decade II and decade III, respectively. Concerning the demand–supply gap of rice in the Cauvery delta zone, it observed that it would be 0.392 million tonnes during decade I and it is 0.232 million tonnes and 0.358 million tonnes during decade III, respectively.

4. Discussion

Climate anomalies in temperature, rainfall patterns, and the frequency and severity of extreme weather events can reduce crop productivity and lead to food shortages and price increases. Thus, climate change may destabilize food security, which may have a detrimental effect on developing countries such as India. This study examined the impact of climate change on food security by examining the imbalance between future rice demand and supply for the period 2021–2050 of the Cauvery Delta Zone. The study has found that there is a significant variance in yield, soil, and cropping pattern over space between the new delta and old delta regions of the Cauvery delta zone.

The initial climate analysis reveals that during decade I (2021–2030), both the new delta and old delta regions receive most of their rainfall during the northeast monsoon, and their annual rainfall is 817 mm and 881 mm, respectively. Both the regions are warmer during April (39 °C) and May (40 °C), and for the rest of the months, the temperature ranges between 31 and 37 °C. Thus, during decade I (2021–2030), the new delta and old delta regions receive less rainfall than their respective normal rainfalls, and the reduction is 20 percent and 17 percent, respectively. Further, the temperature of both regions is found to be increased by 2 °C. During decade II (2031–2040), both regions receive most of their rainfall in the northeast monsoon. However, the reduction in rainfall against normal rainfall is less, comparatively. The rainfall reduction is 8 percent and 6 percent in the new delta and old delta regions, respectively, while temperature increases by 2 °C in both the regions of the Cauvery delta zone. During decade III (2041–2050), unlike decade I (2021–2030) and decade II (2031–2040), the rainfall distribution is not normal—northeast monsoon will not receive most of the rainfall—rather, the rainfall will eventually distribute across months. Further, both regions receive less rainfall than their normal rainfall, and the reduction is 26 percent in the new delta region and 27 percent in the old delta region.

The temperatures increases by 3 °C in the new delta region against its normal temperature, while the increase is 4 °C in the case of the old delta region. Thus, during 2021–2050, the rainfall of both regions of the Cauvery delta zone decreases against the normal rainfall, ranging between 6 and 27 percent. Conversely, the temperature of both the Cauvery delta regions increases, ranging from 2 to 4 °C. Many studies documented that climate change features increasing rainfall, increasing temperature, and increasing frequency of extreme climate events. However, the climate data used in this study features decreases in rainfall in the near future. These findings are contradictory to the findings of Geethalakshmi et al. [65]. This deviation may be due to the adoption of a different climate model for forecasting climate data, where Geethalakshmi et al. [65] employed RegCM3, whereas, following the findings of Samiappan et al. [50], this study uses GFDL data. Samiappan et al. [50] suggest that the application of the GFDL model could be comparatively appropriate in the case of the Cauvery delta zone. In addition, the choice of crop model used in this study could also be another possibility for contradicting results. AquaCrop relies on a few easily obtainable variables.

The future yield estimates from AquaCrop revealed that there is a significant difference in yield response in both delta regions. Climate change boosts rice yield in the new delta zone while lowering yield in the old delta region. The variable response between the regions against climate change is due to the soil character of the respective regions. For instance, the paddy soil of the old delta region has a higher water-holding capacity than the sandy loam soil of the new delta region, which results in more water in the field than the required amount of water during heavy rainfall, and causes the reduction in rice yield. The yield reduction in the old delta region is 48.40 percent during decade I (2021–2030), 49.30 percent during decade II (2031–2040), and 27.67 percent in decade III (2041–2050). Indeed, these findings support the cultivation of a flood-tolerant variety (CR1009 sub) specifically released for the Cauvery delta zone [66].

These results imply that the rice demand will be met in the near future (next 30 years), as the rice production is surplus in the Cauvery delta zone. Nevertheless, the Cauvery delta zone is considered the rice bowl of the State of Tamil Nadu, accounting for 30 percent of the rice production. This surplus rice production is not sufficient for the State of Tamil Nadu, resulting in a significant threat to the food security of the State of Tamil Nadu. Thus, the changing climate may pose a significant threat to the food security of the State of Tamil Nadu. To ensure rice sustainability and food security in the face of climate change, it is important to adopt practices that increase resilience in rice-farming systems, such as the use of flood-tolerant varieties, precision agriculture techniques, and sustainable water management [67].

This study is the first of its kind, which measures food security by accounting for both demand for and supply of rice; it captures all possible variables to estimate robust and reliable results. However, this study is not free from limitations; this study made a few assumptions, and in the case of deviation from these assumptions, the estimates may not be reliable. The methodological framework of this study is not suitable for measuring food security at the regional level; regional markets are not closed and have potential leakage for both the demand and supply of food. Thus, the methodology is most applicable to either the state or national levels. Further, this study used consumption data belonging to the year 2011–2012, since the consumption survey of the NSSO has not yet been published for recent years. Hence, the demand estimates of the study may not capture the present consumption behavior. This study does not implement the bias correction on projected climate data owing to the unavailability of long-term reliable historical weather data. Henceforth, the validity of the climate model used in this study for projecting the climate of the Cauvery delta zone warrants further verification.

5. Conclusions

This study employed an integrated modeling approach to measure the impact of climate change on food security, calculating demand using an economic model and supply with a crop model. Our findings indicate that the peninsular river system’s (Cauvery) food security will be jeopardized in the near future. Several studies have reported that climate change has a minimal impact on irrigated rice ecosystems; nevertheless, the delta region may encounter a significant threat to rice production in the future as a result of changing climate. The old delta region may experience ~50% yield reduction during 2021–2040. Furthermore, two contrasting yield projections from the old and new delta regions revealed that the implications of climate change are location specific. Overall, this study necessitates the development of more flood-tolerant cultivars, standardized operating methods for the discharge of river water for irrigation during the monsoon, and the construction of improved drainage systems at the farm level. These could be the potential mitigation measures required to minimize the consequences of climate change.