This study evaluates the impacts of climate change on rainfed maize (Zea mays) yield and evaluates different agro-adaptation measures to counteract its negative impacts at Sikkim, a Himalayan state of India. Future climate scenarios for the 10 years centered on 2025, 2055 and 2085 were obtained by downscaling the outputs of the HadCM3 General Circulation Model (GCM) under for A2 and B2 emission scenarios. HadCM3 was chosen after assessing the performance analysis of six GCMs for the study region. The daily maximum and minimum temperatures are projected to rise in the future and precipitation is projected to decrease (by 1.7 to 22.6% relative to the 1991–2000 baseline) depending on the time period and scenarios considered. The crop simulation model CERES-Maize was then used to simulate maize yield under future climate change for the future time windows. Simulation results show that climate change could reduce maize productivity by 10.7–18.2%, compared to baseline yield, under A2 and 6.4–12.4% under B2 scenarios. However, the results also indicate that the projected decline in maize yield could be offset by early planting of seeds, lowering the farm yard manure application rate, introducing supplementary irrigation and shifting to heat tolerant varieties of maize.

INTRODUCTION

Climate change is a significant environmental, social and economic threat for human beings and the planetary ecosystem in general (e.g. Intergovernmental Panel on Climate Change (IPCC) 2007). Contemporary studies on climate change in the Himalayan region have shown significant change in temperature and precipitation patterns (Shrestha et al. 1999; Du et al. 2004; Seetharam 2008; Mishra et al. 2013). These studies suggest probable increases in temperature and ambiguous changes to precipitation patterns, with some locations associated with increasing precipitation trends and others with decreasing precipitation trends (Babel et al. 2013; Shrestha et al. 2014a; Yadav et al. 2015). Alcamo et al. (2007) also demonstrated that as extreme events become more frequent and intense, the impact on agriculture and food security is magnified.

The agricultural sector is directly influenced by climatic variables as the physiological processes of plants are directly linked to climate inputs. The alteration of the magnitude and pattern of temperature and precipitation affects the crop water cycle by changing the growth period, photosynthesis ability and respiration rates. Increases to temperature and decreases in precipitation increases water stress on crops and has follow on negative impacts on food security (Tao et al. 2003a, 2003b), with previous studies (e.g. O'Neill 2007; Lobells et al. 2008) showing that change in temperature of up to 2 °C in the dry season can substantially reduce crop yields.

Agriculture is the main economic activity of the people in Sikkim, a state of India located in the Himalayan foothills, and is practiced by 65% of the population. Owing to the undulating topography in East Sikkim, maize (Zea mays) is an important crop grown in 55% of the total cultivated area (GIAHS 2009). Rainfed agriculture is dominant in Sikkim since water storage for irrigation is a persistent problem due to physiographic complexity (Sikkimstat Statistical database for Sikkim state 2013). Overdependence of farmers on rainfed maize production has led to a significant decline in product ion and productivity (ICIMOD 2010). Site-specific impact assessment studies at hilly regions are highly complicated. Several studies appraising future impacts of climate change on maize yield in East Indian and Nepalese basins using various yield simulation models (Malla 2008; Nayava & Gurung 2010; Panda et al. 2012; Joshi & Chaturvedi 2013) have shown that, without adaptation strategies, maize yield is projected to decrease significantly.

In agriculture, the role of adaptation is decisive as it can reduce the magnitude of impacts associated with climate change (Reidsma et al. 2010). Deb et al. (2014) evaluated impacts of climate change on maize production in the Himalayan region of India and emphasized the need for effective adaptation consisting of changing sowing dates, change in fertilizer application rate and change in cultivars as measures to counter the negative impacts of climate change. A study on super-ensemble based probabilistic projection approach was applied to project maize productivity and water use in North China plain by Tao & Zhang (2010) with the results revealing that early and late planting of temperature sensitive and high-temperature tolerant varieties, respectively, are suitable and effective adaptation strategies. Moradi et al. (2013b) evaluated irrigation and planting date management as adaptation measures for maize cultivation under climate change in Iran. In addition, crop substitution is also observed to be a suitable agro-adaptation option under climate change (Rezaei et al. 2013). Several studies also have validated that suitable adaptation options are usually region specific and need to be appraised with respect to the location (Bryan et al. 2009; Tao & Zhang 2010).

Despite the progress in assessment of climate change implications in agricultural production, there is a dearth of exhaustive studies that examine the long-term effects of climate change on regional scale maize production for the Himalayan region. The present study focuses on potential impacts of climate change on maize production and evaluation of several agro-adaptation measures to minimize the adverse impacts using two crop simulation models in hilly Himalayan terrains.

MATERIALS AND METHODS

Study area and maize cultivation

The study area (Sikkim, a state of India) lies at the foothill of the Himalayas in the northeast part of India (approximately 27.33 °N and 88.62 °E). The altitude of the study sites is 1400 m above mean sea level (amsl) (Figure 1). Precipitation in the monsoon season (May–September) contributes 85% of the 2200–3900 mm total annual rainfall. Maximum and minimum daily temperature varies between 17–24 °C and 7–13 °C, respectively. The study was conducted based on meteorological data collected from meteorological stations setup by the Indian Meteorological Department (IMD) network. The major climatic characteristics of the region (temperature, reference evapotranspiration (ETo) and precipitation) for Gangtok station are shown in Figure 2(a) and (b). Sowing of seed is mostly performed from mid-February to mid-March depending on light showers during planting. Location specific planting of the seed depends on precipitation and the optimum temperature which is favorable for the maize in the region (Basnet et al. 2003). NLD-White composite variety of maize is mostly grown in the region which has a growing cycle of 128 days. It is essentially used for poultry feed and self-consumption. It is also exported to other parts of India and gives an excellent monetary return. Other major cultivars grown in the region include Sethi Makai and NAC 6004 which have a growing cycle of 115 and 93 days, respectively. Single cropping practice is followed by local farmers due to unavailability of water for irrigation during dry months.

Figure 1

Map of the study area with location of Gangtok station contemplated in this study.

Figure 1

Map of the study area with location of Gangtok station contemplated in this study.

Figure 2

Observed (a) maximum and minimum temperature, and (b) monthly precipitation and ETo for Gangtok station.

Figure 2

Observed (a) maximum and minimum temperature, and (b) monthly precipitation and ETo for Gangtok station.

Field experimental data

Field experimental data on maize were collected from the Indian Council of Agricultural Research (ICAR) research complex located at Gangtok which was used for model calibration and validation. The experiments were carried out in 2004 and 2005 at Gangtok for NLD-White cultivar of maize. The field experiments were laid out in complete block design. The seeds for the experimental trails were derived from the state government agricultural department. Sowing of seeds was done on 21 February and 3 March followed by tasseling of maize on 11 April and 25 April and silking on 19 April and 2 May for 2004 and 2005, respectively (Table 1). Finally the harvesting of the maize was done on 29 June and 15 July for the corresponding years. Weeding was performed a week after sowing (WAS), 4 WAS, 9 WAS and 1 week prior to harvest in all the trials. Farm yard manure (FYM) was applied at the rate of 12 t ha−1 and 10 t ha−1 for 2004 and 2005, respectively. All the trials were rainfed; therefore, no irrigation water was applied. The effective rainfall for the growing season was observed to be 795 and 859 mm for 2004 and 2005, respectively. In addition, the crop evapotranspiration was calculated to be 220 and 212 mm for the corresponding years which was significantly lower than the effective rainfall for the station. These experiments were conducted to evaluate the effect of planting dates and various management practices on yield and yield components.

Table 1

Crop characteristics collected from ICAR, Gangtok used to calibrate and validate crop models.

  Calibration Validation 
Sowing date 21 February 2004 3 March 2005 
Plant spacing 45 × 50 cm 50 × 50 cm 
Tasselling date 11 Apr 2004 25 Apr 2005 
Silking date 19 Apr 2004 2 May 2005 
Maturity date 29 June 2004 15 July 2005 
Treatment Rainfed Rainfed 
FYM application 12 t/ha 10 t/ha 
Effective rainfall 293 mm 298 mm 
Crop evapotranspiration 245 mm 296 mm 
  Calibration Validation 
Sowing date 21 February 2004 3 March 2005 
Plant spacing 45 × 50 cm 50 × 50 cm 
Tasselling date 11 Apr 2004 25 Apr 2005 
Silking date 19 Apr 2004 2 May 2005 
Maturity date 29 June 2004 15 July 2005 
Treatment Rainfed Rainfed 
FYM application 12 t/ha 10 t/ha 
Effective rainfall 293 mm 298 mm 
Crop evapotranspiration 245 mm 296 mm 

Soil properties

Data on physical and chemical properties of soil were collected from Land Development Department (LDD), Government of Sikkim which included depth wise information of the soil. The dataset collected consists of physical characteristics of the soil such as, soil texture, hydraulic conductivity, field capacity and permanent wilting point of the soil. The chemical properties included the cation exchange capacity (CEC), organic carbon content, pH, potassium, nitrogen and phosphorous (Table 2). The physical and chemical properties of the soil play a significant role in the crop production due to its ability to alter the nutrients and water holding capacity especially in the mountain ecosystems (Debnath et al. 2012; Deb et al. 2013).

Table 2

Physio-chemical properties of the soil with respect to the depth of the experimental plot

  Soil depth (cm)
 
Attributes 0–20 21–60 61–76 77–120 
Sand (%) 55 57 44 42 
Silt (%) 31 26 29 22 
Clay (%) 14 17 27 36 
FC (cm3 water/cm3 soil) 0.22 0.23 0.28 0.32 
PWP (cm3 water/cm3 soil) 0.10 0.12 0.16 0.20 
Saturated hydraulic conductivity (cm h−11.57 1.06 0.41 0.22 
Bd (g cm−31.49 1.46 1.37 1.32 
CEC (cmol kg−110.4 23.7 22.3 20.2 
pH 3.7 4.7 
OC (g kg−120 15 
N (ppm) 172 155 142 118 
P (ppm) 3.3 3.6 4.5 4.2 
K (ppm) 166 174 156 180 
  Soil depth (cm)
 
Attributes 0–20 21–60 61–76 77–120 
Sand (%) 55 57 44 42 
Silt (%) 31 26 29 22 
Clay (%) 14 17 27 36 
FC (cm3 water/cm3 soil) 0.22 0.23 0.28 0.32 
PWP (cm3 water/cm3 soil) 0.10 0.12 0.16 0.20 
Saturated hydraulic conductivity (cm h−11.57 1.06 0.41 0.22 
Bd (g cm−31.49 1.46 1.37 1.32 
CEC (cmol kg−110.4 23.7 22.3 20.2 
pH 3.7 4.7 
OC (g kg−120 15 
N (ppm) 172 155 142 118 
P (ppm) 3.3 3.6 4.5 4.2 
K (ppm) 166 174 156 180 

Meteorological data

Daily meteorological data including maximum and minimum (air temperature, precipitation, reference evapotranspiration, sunshine hours, and average relative humidity) were collected for the period of 1975–2010 for the Gangtok meteorological station. The daily solar radiation for the study area was calculated based on the Angstrom formula as in Equation (1). 
formula
1
where Rs is solar radiation in MJ m−2 day−1, n and N are actual and maximum possible duration of sunshine, respectively, in hour and Ra is extraterrestrial radiation (MJ m−2 day−1). N and Ra for the considered meteorological stations are derived from standard charts provided in Allen et al. (1998).

Crop model

CERES-Maize

The Decision Support System for Agrotechnology Transfer (DSSAT) v 4.0 Cropping System Model (CSM) particularly CERES-Maize model was chosen for the impact assessment due to its wide area of application and its robustness to simulate crop output parameters (Jeffrey et al. 2010). The model can simulate 27 crops with similar input and output files. It has been applied extensively to analyze the impacts of climate change (Tubiello et al. 2002; Attri & Rathore 2003; Babel et al. 2011). DSSAT-CERES for Maize is a dynamic simulation model which works on the understanding of the physiological processes in the crop growth. Cultivar-specific genetic coefficients are required in determining the growth and phenological development of a particular crop depending on photoperiod, thermal time and dry matter partitioning. It assumes dry matter accumulation as a function of photosynthetic active radiation, hence solar radiation is the most essential input to the model. Actual dry biomass at a particular stage is a function of the stress coefficient of water or nitrogen stress and initial accumulated biomass and the actual leaf area index. More details on DSSAT-CERES can be found in Jones et al. (2003) and Hoogenboom et al. (2012).

Crop model development

The input of the CERES-Maize model includes climatic data including maximum, minimum temperature, solar radiation and precipitation; soil data including depth wise physical and chemical properties of the soil namely sand, silt, clay content, pH, CEC, water holding capacity, organic carbon and PNK (phosphorous, nitrogen and potassium) content; crop management data including the irrigation amount and schedule, and the planting and maturity details of the crop. The model was calibrated by iterating the simulation while changing the variety specific six phenological coefficients to match the simulated outputs with the observed yield for the year 2004 as in Table 1. The coefficients were adjusted based on the reference values provided in the user manual. While modelling, the sowing date of maize was assumed to be 15 February based on the criteria that the suitable minimum temperature (10 °C) for germination is reached on that particular day. In addition, the input variables used in modelling consisted of climatic, soil and crop management data. The validation of the model was done by simulating the maize yield using the calibrated parameters to match the observed yield for the year 2005 (Table 1). The output variables considered for the calibration and validation process included grain yield, total biomass, number of leaves at maturity and days to maturity. Furthermore, these predicted values were compared to the observed values for those particular years and percent change was calculated to assess the model performance.

Climate change scenario selection

The Intergovernmental Panel on Climate Change (IPCC) has defined emission scenarios based on various socio-economic, technological advancement and energy use for their adoption in evaluation of projected climate change (Intergovernmental Panel on Climate Change (IPCC) 2007). The Special Report on Emission Scenarios (SRES) was divided into four main storylines (A1, A2, B1 and B2). The A1 scenario family describes a future world of rapid economic growth, global population that peaks at the mid-century and declines thereafter and rapid introduction of new and more efficient technologies. The A1 storyline was further divided into three main sub groups A1F1 (fossil fuel intensive ∼960 ppm CO2 concentration), A1T (predominantly non-fossil fuel ∼585 ppm) and A1B (balanced ∼710 ppm). The A2 storyline defines a very heterogeneous world with economic development regionally oriented and per capita economic growth and technological change more fragmented and slower (∼850 ppm). The B1 storyline describes a convergent world with the same global population that peaks in mid-century and declines thereafter as in A1 storyline, with however emphasis on the global solutions to economic, social and environmental sustainability including improved equity but without additional climate initiatives (∼550 ppm). On the other hand, B2 storyline describes a world in which the emphasis is on local solutions to economic, social and environmental sustainability. It also describes the world with a continuously increasing population at a lower rate than A2 scenario, intermediate levels of economic development and less rapid and more diverse technological change than in B1 and A1 storylines (∼625 ppm) (Intergovernmental Panel on Climate Change (IPCC) 2007). For India, a fast developing nation, industrial growth is expected to be very rapid in the near future which can lead to accumulation of high concentration of greenhouse gases by the end of the century and therefore A2 scenario was considered for the study. In addition, another expectation is that in the future people may prioritize environmental sustainability and place more emphasis on the local solutions to economic and social perspectives leading to less industrial growth and hence B2 scenario is more suitable. Moreover, several studies on climate change impact assessment in the Himalayan region also have considered these two scenarios and were given the primary importance because of their applicability in the region (Babel et al. 2013; Shrestha et al. 2014b; Deb et al. 2014).

GCM and National Centers for Environmental Prediction (NCEP) reanalysis predictors

General circulation models (GCMs) project the change in climate under different emission scenarios. However, the spatial resolution of the GCM projections is coarse (typically 250 × 250 km) and their representativeness for a particular area or variable is associated with significant uncertainties and limitations (Parry et al. 2007; Stainforth et al. 2007; Koutsoyiannis et al. 2008, 2009; Blöschl & Montanari 2010; Montanari et al. 2010; Verdon-Kidd & Kiem 2010; Kiem & Verdon-Kidd 2011; Stephens et al. 2012). This makes it necessary to have the selection of an appropriate GCM prior to further analysis. In the study six GCMs (Table 3) were analyzed to select the most suitable GCM for the study site. The performance evaluation of the GCMs was conducted based on precipitation since it is the most uncertain variable. Three indicators were used for the evaluation: coefficient of determination (R2), root mean square error (RMSE) and mean bias error (MBE).

Table 3

List of the six GCMs used for performance evaluation for the study area

Name of GCM Country Center 
ECHAM5/MPI-OM Germany Max-Plank Institute for Meteorology 
CCSM3 USA University Corporation for Atmospheric Research 
UKMO-HadCM3 UK UK Meteorological Office 
CSIRO-MK3.0 Australia Commonwealth Scientific and Industrial Research Organization (CSIRO) 
CGCM3.1 Canada Canadian Centre for Climate Modeling and Analysis 
MIROC3.2 Japan Meteorological Research Institute 
Name of GCM Country Center 
ECHAM5/MPI-OM Germany Max-Plank Institute for Meteorology 
CCSM3 USA University Corporation for Atmospheric Research 
UKMO-HadCM3 UK UK Meteorological Office 
CSIRO-MK3.0 Australia Commonwealth Scientific and Industrial Research Organization (CSIRO) 
CGCM3.1 Canada Canadian Centre for Climate Modeling and Analysis 
MIROC3.2 Japan Meteorological Research Institute 

It was observed that three GCMs (UKMO-HadCM3, ECHAM5/MPI-OM and CGCM3.1) reproduced statistically reliable results considering daily and average monthly precipitation. Satisfactory R2 (in range of 0.21–0.66) was observed for the three GCMs. Nonetheless, highest R2 was obtained for HadCM3 ranging from 0.31 to 0.64. In the case of A2 scenario, HadCM3 showed lowest RMSE for average monthly precipitation. Similarly, for the corresponding scenario, lowest MBE was further observed for HadCM3 with an absolute value of 2.36 mm for the corresponding variable. With the established results, HadCM3 was considered for the study area.

Predictor variables of National Centers for Environmental Prediction (NCEP), also called reanalysis predictors, and HadCM3 predictors analogous to two SRES A2 and B2 were downloaded from http://www.cics.uvic.ca/scenarios/sdsm/select.cgi. The predictors are directly used for the downscaling tool as the NCEP reanalysis data was previously normalized according to the HadCM3 grid size. The predictor files used are as follows: NCEP_1961–2001: File contains 41 years of daily observed predictor data derived from the NCEP reanalysis data; H3A2a_1961-2099 and H3B2a_1961-2099: Files contain 139 years of daily GCM predictor data derived from HadCM3 A2 and B2 experiments, respectively.

Downscaling coarse resolution GCM data

The outputs of GCMs had very coarse spatial resolution therefore cannot represent sub-grid features like topography, land use as required in assessment studies consequently makes it unsuitable for basin or regional scale studies. The existing gap of spatial resolution among GCMs and regional scale studies was resolved by downscaling the climate variables (Giorgi & Mearns 1991). Statistical downscaling presumes relationship among large-scale predictor variables and station level climate variables. In the present study, statistical downscaling was used for its computational simplicity as compared to dynamic downscaling (Prudhomme et al. 2002; Najafi et al. 2011).

Although several approaches are available for statistical downscaling, Statistical Down Scaling Model (SDSM v4.2) (Wilby & Dawson 2007) was used for this study due to its wide application on global scale (Xu et al. 2008; Toews & Allen 2009; Mahmood & Babel 2013; Babel et al. 2013) and its satisfactory performance over other techniques (Khan et al. 2006; Kiem et al. 2008; Samadi et al. 2013; Hassan et al. 2013). The principle of downscaling by SDSM includes developing multiple linear regression transfer function among the large-scale predictors and local climate variables (Wilby et al. 2002). The major steps to be performed are: (1) screening of large scale predictors; (2) calibrating transfer functions generated based on observed variables; (3) validation of the model results, and (4) generating scenario. A major limitation exists in downscaling by SDSM since it assumes the selected predictors based on the predictor–predictand relationship in the calibration process will be unaltered under future climate which is not justifiable in many cases (Segui et al. 2010).

Calibration and validation of SDSM

Before calibration, 26 predictor variables were screened to select the best set of predictors which exhibit good correlation with the predictand. The criteria for screening predictors set was that at least one predictor should significantly correlate with the predictand variable at a significance level of 0.05%. A correlation matrix developed among the predictors and predictand provides the base for the selection of the predictors which were more significantly related to the predictand. Table 4 presents the relationship among the global scale predictand (maximum temperature) and the local scale predictors (surface zonal velocity, 500 hpa geopotential height and near surface relative humidity) at a monthly scale for the Gangtok station. It is clear that the Pearson correlation coefficient varies among the predictand and the predictors varies with respect to months. Highest relationship is observed for 500 hpa geopotential height for the month of April. The selected predictors used in this study for further analysis are presented in Table 5. As mentioned earlier, surface zonal velocity, 500 hpa geopotential height and near surface relative humidity is found to be the most influencing predictors for maximum temperature during the calibration process of the model. Interestingly, those predictors are also found to have significant relationship with minimum temperature and precipitation where surface zonal velocity is found to have very poor correlation for the latter. Monthly and conditional options were selected for maximum and minimum temperature for calibration of the model. However, unconditional option was selected for precipitation as the amount is dependent on intermediate processes like wet/dry days (Wilby et al. 2002). The historic meteorological data 1975–1990 and 1991–2000 were used for calibration and validation, respectively, for validation of SDSM. The validation was performed for maximum temperature (Tmax), minimum temperature (Tmin) and precipitation (Precp). Simulated data were compared with the observed for mean daily value and standard deviation (SD) for the corresponding value of the station data.

Table 4

Generated correlation matrix during screening of the predictor variables process for daily maximum temperature at Gangtok station

Analysis period: 01 January 1975–31 December 1990 
Significance level: 0.05 
Predictand: Tmax Gtk 1975–2000 data 
Predictors January February March April May June July August September October November December 
Ncepp_uas.dat 0.01 0.32 0.27 0.02 0.32 0.14 0.09 0.01 0.02 0.19 0.29 0.04 
Ncepp500as.dat 0.43 0.01 0.12 0.54 0.02 0.02 0.01 0.14 0.21 0.11 0.38 0.31 
Nceprhumas.dat 0.00 0.12 0.00 0.21 0.12 0.03 0.04 0.08 0.05 0.01 0.12 0.02 
Analysis period: 01 January 1975–31 December 1990 
Significance level: 0.05 
Predictand: Tmax Gtk 1975–2000 data 
Predictors January February March April May June July August September October November December 
Ncepp_uas.dat 0.01 0.32 0.27 0.02 0.32 0.14 0.09 0.01 0.02 0.19 0.29 0.04 
Ncepp500as.dat 0.43 0.01 0.12 0.54 0.02 0.02 0.01 0.14 0.21 0.11 0.38 0.31 
Nceprhumas.dat 0.00 0.12 0.00 0.21 0.12 0.03 0.04 0.08 0.05 0.01 0.12 0.02 
Table 5

Summary of selected predictors for the Gangtok station

Station Predictand Predictors 
Gangtok Maximum temperature Surface zonal velocity (P_u) 
500 hpa geopotential height (P500) 
Near surface relative humidity (rhumas) 
Minimum temperature Mean sea level pressure (mslp) 
Surface zonal velocity (P_u) 
Near surface relative humidity (rhumas) 
Precipitation 500 hpa geopotential height (P500) 
Near surface relative humidity (rhumas) 
Station Predictand Predictors 
Gangtok Maximum temperature Surface zonal velocity (P_u) 
500 hpa geopotential height (P500) 
Near surface relative humidity (rhumas) 
Minimum temperature Mean sea level pressure (mslp) 
Surface zonal velocity (P_u) 
Near surface relative humidity (rhumas) 
Precipitation 500 hpa geopotential height (P500) 
Near surface relative humidity (rhumas) 

Future climate projections

A calibrated model was used for generating the future climate variables for future time periods at the respective stations. Maximum, minimum temperature and precipitation were considered for two scenarios A2 and B2 from HadCM3 for the downscaling process. A2 represents the world with continuously increasing population with a regional focus for development in terms of economy and culture, whereas B2 scenario represents a world integrating and focusing on local solution to economic, social and environmental sustainability (Carter 2007). A2 and B2 represent two different socio-economic and cultural trends which are very suitable in the context of the mountainous regions of South Asia and therefore both scenarios were analyzed in this study. 20 ensembles were generated for both temperature and precipitation for the period of 1961–2099. Averages of all the ensembles were then selected for further analysis. With the downscaled results for each station, the changes in maximum/minimum temperature and precipitation for the time periods 2021–2030, 2051–2060 and 2081–2090 (represented as 2025s, 2055s and 2085s, respectively) relative to baseline period of 1991–2000 (1995s) were calculated. The shift in climatic variables for the state (Sikkim) was calculated based on averaging the changes of variables from the three representative stations.

Yield simulation for future climate

The calibrated and validated model was used to simulate the future yield under fictional scenarios for Gangtok station in order to determine the consequence of escalated temperatures and CO2 levels on maize yield. Additional weather parameters (rainfall, solar radiation and relative humidity) were assumed to be the same as that of the baseline period of 1991–2000. The yield was simulated for different combinations of CO2 concentrations (390, 400, 500, 600, 700 and 800 ppm) with Tmax and Tmin (+1, +2, +3, +4, +5 and +6 °C).

In addition, the model was also used to simulate the future yield of maize under the two climate change scenarios downscaled at station scale. The relative changes in the yield for the scenarios, referring to yield predicted for the current climate, was used for the study rather than absolute yields. In order to represent the overall impact of climate change on maize productivity, relative changes of yield and other output components for future climate referring to present climate had been simulated based on the historical dataset.

Agro-adaptation measures

Several plausible adaptation strategies were also evaluated to minimize the negative impacts of climate change on maize productivity in the region. Alternate sowing dates, proper nutrient management, introducing supplementary irrigation practices and broaching heat-tolerant varieties were evaluated by the calibrated CERES-Maize model. Simulations for different sowing dates ranging from −34 to +37 days at an interval of 7 days compared to the current sowing date of 15 February were considered. Various application rates of FYM from 50 to 160% compared to the recommended application rate (12 t/ha) was also considered for this study. Investigation of different irrigation water application rates (10, 20, 30 and 40 mm) at 25-day interval was further done to find the optimum amount required under climate change. The suitability of other heat tolerant composite maize cultivars namely Sethi Makai and NAC 6004 (Basnet et al. 2003) was also assessed for the study site in order to evaluate higher yield under climate change relative to the traditionally grown cultivar NLD-White. The sowing date of the seeds in the case of the adaptation measures, except for changes in sowing dates, was considered to be the same as that of the baseline period of 15 February for the future climate. The effect on average yield prolonging for a decade was compared with the yield in the period for normal rainfed conditions. Percent change in the yield compared to that of baseline conditions determined the appropriate adaptation measures for the selected region to mitigate severe climate change impact.

RESULTS AND DISCUSSION

Climate scenarios

Validation of SDSM

In the case of simulated Tmax, maximum difference among observed and simulated temperatures is observed to be 1.1 °C in March. Similarly, the highest deviation in SD is noted in October with magnitude of 0.4 °C (Figure 3). In addition, the difference in the simulated and observed temperature for Tmin, is observed to be relatively lower with a magnitude of 0.6 and 0.3 °C for SDs for April and December, respectively. This reflects that the model shows higher magnitude of variation during the transition period from winter to summer and higher SD during winter periods. Furthermore, in the case of precipitation, the highest variation in the simulated and observed mean daily rainfall is noticed for July and June with a magnitude of 7.97 and 2.7 mm for difference and SD, respectively. Nevertheless, calculated R2 and RMSE for maximum, minimum temperatures and precipitation suggest the model can simulate results in good agreement with observed data.

Figure 3

Comparison of observed and simulated (a) mean daily maximum temperature, (b) mean daily minimum temperature, and (c) mean daily precipitation for each month during the validation period (1991–2000) at Gangtok station.

Figure 3

Comparison of observed and simulated (a) mean daily maximum temperature, (b) mean daily minimum temperature, and (c) mean daily precipitation for each month during the validation period (1991–2000) at Gangtok station.

Future temperature and precipitation

An increasing trend in annual temperature is observed for the future time windows. The highest changes observed in Tmax are for 2085s with a magnitude of +1.91 and +1.59 °C for A2 and B2 scenarios, respectively, compared to baseline temperature (Table 6). It is also evident that for Tmin, a higher magnitude in the change is observed relative to Tmax for the corresponding time period and scenarios. The maximum change observed is for A2 scenario with a magnitude of 1.98 °C in 2085s. It is worth noting that for 2025s, Tmax shows higher change relative to Tmin.

Table 6

Projected changes in future maximum, minimum temperature and average annual precipitation at Gangtok station under A2 and B2 scenarios

Baseline (1995s) Time period Scenario A2 Scenario B2 
Tmax (°C) Projected (°C) Change (°C) Projected (°C) Change (°C) 
19.84 2025s 20.54 0.70 20.38 0.54 
2055s 20.98 1.14 20.81 0.97 
2085s 21.75 1.91 21.43 1.59 
Tmin (°C) Projected (°C) Change (°C) Projected (°C) Change (°C) 
10.85 2025s 11.49 0.64 11.38 0.53 
2055s 12.12 1.27 11.84 0.99 
2085s 12.83 1.98 12.59 1.74 
Precp (mm) Projected (mm) Change (%) Projected (mm) Change (%) 
3707 2025s 3644 − 1.7 3907 5.4 
2055s 3347 − 9.7 3381 − 8.8 
2085s 2869 − 22.6 2984 − 19.5 
Baseline (1995s) Time period Scenario A2 Scenario B2 
Tmax (°C) Projected (°C) Change (°C) Projected (°C) Change (°C) 
19.84 2025s 20.54 0.70 20.38 0.54 
2055s 20.98 1.14 20.81 0.97 
2085s 21.75 1.91 21.43 1.59 
Tmin (°C) Projected (°C) Change (°C) Projected (°C) Change (°C) 
10.85 2025s 11.49 0.64 11.38 0.53 
2055s 12.12 1.27 11.84 0.99 
2085s 12.83 1.98 12.59 1.74 
Precp (mm) Projected (mm) Change (%) Projected (mm) Change (%) 
3707 2025s 3644 − 1.7 3907 5.4 
2055s 3347 − 9.7 3381 − 8.8 
2085s 2869 − 22.6 2984 − 19.5 

Analysis of precipitation suggests the study area is expected to experience a declining trend in average annual rainfall under climate change. The future precipitation is projected to lower by 22.6 and 19.5% for A2 and B2 scenarios, respectively, for 2085s. The magnitude of precipitation is observed to lower from 3707 mm to 2869 and 2984 mm for the corresponding period and scenarios. Surprisingly, an increase in precipitation is observed for 2025s in the case of B2 scenario. The underlying environment friendly assumption for B2 scenario can be the causative factor for this increase of precipitation. It is also noteworthy that the increase in precipitation can lead to frequent flash flooding in the downstream during monsoon period (Shrestha et al. 2014c). These trends imply an early seasonal forecasting of precipitation would be beneficial at station scale for better water management practices.

Crop model set up

CERES-Maize model was calibrated and validated for the rainfed cropping season of maize by fine tuning the calibration parameters in the model. One year of extensive field experimental data were used for each of calibration and validation process. Table 7 represents the calibrated values of the parameters for the NLD-White variety of maize for the study site. Calculated percent change in the yield components for calibration and validation process illustrate CERES-Maize simulates output variables in good agreement with the observed ones and the model is set for future yield projection. The variation in yield is observed to be −3.88 and 1.73% for calibration and validation, respectively. Similarly, a minimum variation is also observed for total biomass, number of leaves and days to maturity (Table 8). However, the maximum variation is noted for number of leaves, 5.79 and 4.44% for calibration and validation, respectively, is due to inability of the model because of the complex phenological calculations driving the simulated values. However, the errors are still within an acceptable range.

Table 7

Calibrated parameters of CERES-Maize for NLD-White cultivar of maize

Parameters P1a P2b P5c G2d G3e PHINTf 
Calibrated values 268 0.28 856.6 265 6.5 56 
Parameters P1a P2b P5c G2d G3e PHINTf 
Calibrated values 268 0.28 856.6 265 6.5 56 

aThermal time from seeding to end of juvenile phase (expressed in degree days above a base temperature of 8 °C) during which the plant is not affected by photoperiod.

bLevel to which development (expressed in days) is generally deleted for each hour increase in photoperiod above the longest photoperiod at which development proceeds at a maximum rate (which is considered to be 12.5 hours).

cThermal time from silking to physiological maturity (expressed in degree days above a base temperature of 8 °C).

dMaximum possible kernel number per plant.

eKernel filling rate during the linear grain filling stage and under optimum conditions (mg/day).

fPhylochron interval; the interval in thermal time (degree days) between successive leaf tip appearances.

Table 8

Performance of CERES-Maize during calibration and validation processes at Gangtok station

  Calibration
 
Validation
 
Variables Observed Simulated Change (%) Observed Simulated Change (%) 
Yield (t ha−13.88 4.03 − 3.86 4.03 3.96 1.73 
Total biomass (t ha−18.43 8.64 − 2.50 8.78 8.71 − 0.56 
Number of leaves at maturity 19 17.9 5.79 18 17.2 4.44 
Days to maturity 134 138 − 2.98 135 140 − 3.70 
  Calibration
 
Validation
 
Variables Observed Simulated Change (%) Observed Simulated Change (%) 
Yield (t ha−13.88 4.03 − 3.86 4.03 3.96 1.73 
Total biomass (t ha−18.43 8.64 − 2.50 8.78 8.71 − 0.56 
Number of leaves at maturity 19 17.9 5.79 18 17.2 4.44 
Days to maturity 134 138 − 2.98 135 140 − 3.70 

Literature shows CERES-Maize has been extensively used for climate change impact assessment studies due to its wide applicability and robustness. Researchers have applied the model for simulation of maize yield at various agro-ecological zones at global scale. Zhi-Qing & Da-Wei (2008) applied CERES-Maize in Northeast China plains to assess the future impacts of changes in climate and its variability on food production. In a separate study, Meza et al. (2008) applied the same model for assessing the climate change implication on irrigated maize in the Mediterranean climates along with evaluation of double cropping as an adaptation option under climate change. CERES-Maize is also applied for prediction of maize yield within a particular season for rainfed corn yields in Delaware, USA by Quiring & Legates (2008). The application of CERES-Maize is also extended to further evaluation of various agro-adaptation measures for maize production under climate change, for instance by Moradi et al. (2013a, 2013b) and Rezaei et al. (2013).

Effect of temperature and carbon dioxide (CO2) levels at fixed intervals

The simulation results show a significant declining trend is observed in maize yield for increased temperature. Antithetically, an increase in CO2 levels is beneficial for yield at any particular temperature. Considering the current level of CO2 concentration (390 ppm), the model simulated an average yield reduction of 3.72% for a temperature increase of 1 °C. Additionally, for the corresponding CO2 concentration an increase of 6 °C above ambient temperature results in a yield reduction of 27.54%. The loss in yield can be attributed to the heat stress due to increase in temperature during the tassel initiation and silking phase in maize. On the contrary, an increase of 25.31% in maize yield is observed for the ambient temperature at 800 ppm CO2 concentration (Table 9). Moreover, with higher CO2 concentration and corresponding temperature (+6 °C), a relatively low increase in yield is observed (5.46%) compared to that of ambient temperature, thus implying that although higher CO2 uplifts the crop growth rate by accelerating photosynthesis process the heat injury caused by high temperature is inevitable. Similar results in Thailand and India for rice were also obtained by Krishnan et al. (2007) and Babel et al. (2011).

Table 9

Predicted meana change (%) in potential yield under different temperature and CO2 scenarios for NLD-White variety of maize for Gangtok station simulated by CERES-Maize

  CO2 concentration (ppm)
 
Temperature increment (°C) 390 400 500 600 700 800 Average 
4.71 9.18 15.38 19.6 25.31 12.36 
− 3.72 3.23 8.19 12.16 18.11 18.61 9.43 
− 9.43 − 2.23 6.70 11.17 14.64 13.15 5.67 
− 12.66 − 5.46 2.98 7.94 11.91 12.16 2.81 
− 14.89 − 5.96 1.74 4.47 10.17 11.41 1.16 
− 20.35 − 12.90 − 2.98 − 2.73 4.47 8.19 − 4.38 
− 27.54 − 14.14 − 12.66 − 4.22 − 0.74 5.46 − 8.97 
Average − 12.66 − 4.68 1.88 6.31 11.24 13.47 2.59 
  CO2 concentration (ppm)
 
Temperature increment (°C) 390 400 500 600 700 800 Average 
4.71 9.18 15.38 19.6 25.31 12.36 
− 3.72 3.23 8.19 12.16 18.11 18.61 9.43 
− 9.43 − 2.23 6.70 11.17 14.64 13.15 5.67 
− 12.66 − 5.46 2.98 7.94 11.91 12.16 2.81 
− 14.89 − 5.96 1.74 4.47 10.17 11.41 1.16 
− 20.35 − 12.90 − 2.98 − 2.73 4.47 8.19 − 4.38 
− 27.54 − 14.14 − 12.66 − 4.22 − 0.74 5.46 − 8.97 
Average − 12.66 − 4.68 1.88 6.31 11.24 13.47 2.59 

aChanges are averaged for all the years (2001–2010).

Impact of predicted GCM scenarios on maize yield

The projected changes in maize productivity under different climate change scenarios downscaled from HadCM3 suggest a significant decline in maize yield relative to baseline maize productivity (Figure 4). A reduction in maize yield ranging from 10.7 to 18.2% and 6.4 to 12.4% is observed for the three time windows in the cases of A2 and B2 scenarios, respectively. The highest maize production is observed for 2025s under B2 scenario (3.71 t ha−1) and lowest is noted for 2085s under A2 scenario (3.24 t ha−1) relative to the baseline yield of 3.96 t ha−1. As anticipated, higher reduction in yield is observed for 2085s under both scenarios. It is also observed that the variance in maize yield increases with the progress of time windows considered. The higher increase in the maximum temperature during the tasselling and reproductive phases leading to the increase in the evaporative demand can be attributed to the increased yield loss (Figure 5). Also, lower magnitude of the projected precipitation for future time windows under climate change during the respective critical stages contributing to unavailability of sufficient water to meet the evaporative demand of maize can also be attributed to yield loss. Another yield loss contributing factor can be attributed to the higher proportion of sandy soil in the root zone which has lower water holding capacity leading to deep percolation of the precipitation water ultimately leading to lower water availability and therefore hampering the critical stages of the maize growth cycle (Harrison et al. 2011). Nevertheless, climate change will have a significant implication on the maize productivity if agricultural practices remain the same in the future.

Figure 4

Projected change in yield of NLD-White variety of maize under climate change (Error bars show the variance of change in maize yield across different years).

Figure 4

Projected change in yield of NLD-White variety of maize under climate change (Error bars show the variance of change in maize yield across different years).

Figure 5

Changes in average monthly minimum and maximum temperatures for 2085s relative to baseline obtained from downscaling future climate data for the cropping season.

Figure 5

Changes in average monthly minimum and maximum temperatures for 2085s relative to baseline obtained from downscaling future climate data for the cropping season.

The outputs of this study can be compared to those of Twine et al. (2013) for Midwest USA under climate change where a decline in yield is observed except for dry years in the future where an insignificant increase is noted. In addition, an average change of −9.5% in yield for 2046–2065 was observed by Tachie-Obeng et al. (2013) for Ghana based on simulation done from nine GCMs. Furthermore, study conducted in China by Tao & Zhang (2010) also indicates a potential reduction of 13.2–19.1% is expected in North China Plain during 2050s if proper adaptation is not being taken into consideration. Kucharik & Serbin (2008) suggests under future climate, for every degree rise in temperature in the US Corn Belt region, corn and soybean yields are potentially expected to reduce by 13 and 16%, respectively. A separate study by Meza et al. (2008) illustrates, maize yield is expected to be affected by climate change with yield reductions between 10 and 30% based on the probable scenarios. Also, it will affect the rate of development, irrigation needs and growing cycle of maize in Mediterranean region. A global scale study by Ray et al. (2012) indicates maize productivity in the Himalayan region is either stagnated or never improved based on location for the current climate relative to the past. In general, the region is expected to have a reduction in maize productivity due to the projected alteration in magnitude of precipitation and increase in temperature.

Adaptation measures to enhance the maize yield

Various studies have validated that climate change is expected to have significant negative implications on agriculture and economics without adaptation and mitigation strategies (Smit & Skinner 2002). Identification of appropriate adaptation measures involves evaluation of the climatic variables and their influence on crop productivity. Proxy crop management practices comprising different planting dates (Moradi et al. 2013a; Tachie-Obeng et al. 2013), rate of FYM application (Shrestha et al. 2014a), supplementary irrigation (Moradi et al. 2013b) and change in cultivar (Babel et al. 2011) were investigated at Gangtok site as adaptation options to climate change. Furthermore, the literature also suggests that very little knowledge of all these adaptation measures and their application have been done for higher altitudes. Also, although changing the farming practice from traditional to heat and stress tolerant cultivars is also well documented, however its application as a climate change adaptation strategy is not being applied successfully in many parts of the world (Rezaei et al. 2013). Hence, our results provide comprehensive information on the plausible agro-adaptation measures and outputs that can be used in other regions to counteract the negative impacts of climate change in agriculture.

Change in planting date

Simulation of yield suggests early shifting of planting date is beneficial for maize cultivation under climate change for both scenarios and time windows (Figure 6). Among the various sowing dates evaluated, for A2 scenario 1 February is observed to increase maize yield by 5% for 2025s. In addition, for the corresponding scenario, shifting the planting date to 25th and 18th January can enhance yield by 13.4 and 22.5% compared to yield for the initial sowing date (15 February) for 2055s and 2085s, respectively. In the case of B2 scenario, 1 February is noted suitable for 2025s with an increase of 12.5% in potential yield. Similarly for 2055s and 2080s, planting on 25th January and 18th can increase the yield by 11.4 and 11%, respectively, for the corresponding scenario. Early shift of planting date helps the critical stages (silking and tasselling) avoid high temperature stress. In addition, it also lowers the evaporative demand of the plants such that although less precipitation water is available, it is sufficient.

Figure 6

Change in maize yield (%) for NLD-White cultivar with different transplanting dates for (a) A2 and (b) B2 scenarios at Gangtok site. (Error bars show variance of change in maize yield across different years for different transplanting dates).

Figure 6

Change in maize yield (%) for NLD-White cultivar with different transplanting dates for (a) A2 and (b) B2 scenarios at Gangtok site. (Error bars show variance of change in maize yield across different years for different transplanting dates).

Change in FYM application

The soil of Sikkim is enriched with high organic matter, in addition the current trend shows farmers prefer to apply more manure for better production (Debnath et al. 2012), and therefore proper management of FYM application is paramount in the context of climate change (Babel et al. 2011). Contrasting maize yield is observed for varying FYM application rate at Gangtok site (Figure 7). For the baseline period (1995s), an application of 110% FYM compared to current rate (12 t ha−1) improves the potential yield up to 1.72%. In addition, for 2025s under A2 scenario by curtailing the application rate to 80% enhances the maize yield by 7.5% relative to the yield simulated by present application rate for that period. However, for 2055s and 2085s, 60% FYM application rate is found optimum with an increase of 6.2 and 7.1%, respectively, for the corresponding scenario. An application rate of 80, 70 and 60% for 2025s, 2055s and 2085s can boost the yield by 2.6, 7.7 and 5.9% respectively, for B2 scenario. Due to the presence of higher organic matter in the soil, the intake of positive charged nitrogen (N), potassium (K) and phosphorous (P) is lowered by the carbon ions present in the soil and therefore in order to counteract this, lower application of FYM enhances the readily available ions after disaggregation and gets easily absorbed in soil colloids influencing the fertilizer use efficiency (Shrestha et al. 2014a). Also, presence of higher organic matter in clay soil leads to water logged condition for the plants and thus addition of higher FYM contributes to wilting and hence yield is reduced (Deb et al. 2014). The results of the present study show that a site-specific nutrient management system introduced by International Maize and Wheat Improvement Center will be beneficial for accelerating the nitrogen-use efficiency for forthcoming time periods under climate change (Olesen et al. 2011).

Figure 7

Change in maize yield (%) for NLD-White cultivar with different FYM application rates for (a) A2, and (b) B2 scenario at Gangtok site. (Error bars show variance of change in maize yield across different years for different FYM application rates).

Figure 7

Change in maize yield (%) for NLD-White cultivar with different FYM application rates for (a) A2, and (b) B2 scenario at Gangtok site. (Error bars show variance of change in maize yield across different years for different FYM application rates).

Effect of supplementary irrigation on grain yield

It is evident that Gangtok site is expected to have a potential reduction of 22.6 and 19.5% in precipitation for A2 and B2 scenarios, respectively, relative to the baseline period by the end of the 21st century and consequently a significant reduction in yield of rainfed maize is also observed. Although the magnitude of precipitation is observed to be lower than the baseline period, however the magnitude is higher than the crop evapotranspiration. In addition, the soil properties in the root zone suggest lower water holding capacity at the site and therefore introduction of a supplementary irrigation system was tested. Results suggest for the current study site under the A2 scenario, quadric application of 20, 30 and 40 mm for 2025s, 2055s and 2085s during the growing period is observed to increase the maize yield by 17.1, 20.7 and 38%, respectively, relative to rainfed cultivation. For the B2 scenario, 30 mm (four applications) is observed to maximize the yield by 12.6, 15.1 and 17.6%, respectively, for the corresponding time windows (Figure 8). The property of water to alter the canopy temperature irrespective of outside air temperature may be the contributing factor for the increase in maize yield under future climate. In addition, due to the poor water holding capacity of the soil strata, supplemental application of irrigation enhances the water intake to the plants and therefore the crop evapotranspiration is met and the yield is met up to the potential.

Figure 8

Change in maize yield (%) for NLD-White cultivar with different irrigation water application rates for (a) A2, and (b) B2 scenarios at Gangtok site. (Error bars show variance of change in maize yield across different years for various irrigation water depth).

Figure 8

Change in maize yield (%) for NLD-White cultivar with different irrigation water application rates for (a) A2, and (b) B2 scenarios at Gangtok site. (Error bars show variance of change in maize yield across different years for various irrigation water depth).

Change in maize variety

Fostering new maize cultivars which are consistent in developing kernels at high temperature and water stress has proved to be an important agro-adaptation measure to mitigate climate change (Tachie-Obeng et al. 2013). In this context, we simulated the maize yield for two composite varieties Sethi Makai and NAC 6004 under rainfed conditions for historical and future climate conditions. The calibrated parameters for the two cultivars in order to calibrate the model are provided in Table 10. The simulated maize yield under future climate with current management practice for the two cultivars illustrate an increase of 17.18 to 66.07% and 17.54 to 44.14% for A2 and B2 scenarios, respectively, relative to the yield simulated for NLD-White. Furthermore, shifting from NLD-White to NAC 6004 suggest a significant increase of 32.68% in yield for the historical period (Figure 9). The heat tolerant cultivars are observed to be less affected by heat injury in spite of an increase in temperature. The special properties of these cultivars is to change the leaf orientation, short and long-term stress avoidance, ability to alter the membrane lipid composition and cooling by transpiration. Increase in yield by varying cultivar also has been observed by Tingem & Rivington (2009) at Cameroon and Travasso et al. (2006) in South America.

Table 10

Calibrated parameter values for the two cultivars at Gangtok site

Cultivars P1 P2 P5 G2 G3 PHINNT 
Sethi Makai 266.1 0.13 851.2 298 6.2 67 
NAC 6004 255.6 0.21 889.6 334 7.6 58 
Cultivars P1 P2 P5 G2 G3 PHINNT 
Sethi Makai 266.1 0.13 851.2 298 6.2 67 
NAC 6004 255.6 0.21 889.6 334 7.6 58 
Figure 9

Yield of different maize cultivars for future periods under (a) A2, and (b) B2 scenarios at Gangtok site. (Error bars show variance of maize yield across different years for different cultivars).

Figure 9

Yield of different maize cultivars for future periods under (a) A2, and (b) B2 scenarios at Gangtok site. (Error bars show variance of maize yield across different years for different cultivars).

CONCLUSIONS

This study investigates the changes in the future climate and its impacts on rainfed maize productivity at Gangtok site in Sikkim state of India. Six GCMs (ECHAM5, CCSM, HadCM3, CSIRO-MK3.0, CGCM3.1 and MIROC3.2) were first evaluated and the best GCM (HadCM3) based on statistical evaluation was selected and applied for the impact assessment. The coarse resolution of the GCM was downscaled to station level by statistical downscaling tool (SDSM) by employing large scale GCM predictors. Statistical evaluation of downscaling procedure indicates the model represents the mean and extremes of the temperature and precipitation in good agreement with the observed data. Downscaling climate variables suggest increase in average annual maximum temperature by 1.91 and 1.59 °C for A2 and B2 scenarios, respectively, for 2085s relative to 1995s. A relatively higher increase in minimum temperature is observed for the considered scenarios and time period with a magnitude of 1.98 and 1.74 °C. In addition, precipitation is expected to reduce by 1.7 to 22.6% for all time periods and scenarios considered relative to 1995s.

The outputs of the crop simulation model CERES-Maize suggest a significant decline in the maize productivity ranging from 10.7 to 18.2% and 6.4 to 12.4% for A2 and B2 scenarios, respectively, relative to the baseline yield of 3.96 t ha−1 for 1995s. Simulation of yield for various CO2 concentration level and incremental temperature suggests higher CO2 is beneficial for maize productivity at a constant temperature. On the contrary, high temperature at a constant level of CO2 concentration illustrates lower yield. The study also reveals early shifting in the planting dates to 1 February, 25 and 18 January for 2025s, 2055s and 2085s, respectively, enhances the potential maize yield for both scenarios. Also, lowering the FYM application rate to 80% for 2025s and 60% for 2055s and 2085s for A2 scenario whereas 80, 70 and 60% for B2 scenario enhances the yield. In addition, four applications of 20, 30 and 40 mm, respectively, irrigation enhances the yield under future climate relative to rainfed condition for A2 scenario, whereas four applications of 30 mm irrigation is observed to be suitable for the B2 scenario. Finally, shifting from NLD-White to composite varieties such as NAC 6004 and Sethi Makai shows an increase in yield from 17.2 to 66.1 and 17.54 to 44.14% for A2 and B2 scenarios, respectively, under future climate. The results of this study will provide policy makers and key stakeholders with an insight that can be used to develop a decision support tool to optimize the maize cultivation under the changing climate regimes for Sikkim state of India. Moreover, the outputs can also be used as a guideline for adaptation measures under future climate change scenarios in other regions of the world.

ACKNOWLEDGEMENTS

The authors are wholeheartedly thankful to Indian Council of Agricultural Research (ICAR), Sikkim center and Indian Meteorological Department (IMD) for providing the necessary data to carry out this research study.

REFERENCES

REFERENCES
Alcamo
J.
Dronin
N.
Endejan
M.
Golubev
G.
Kirilenko
A.
2007
A new assessment of climate change impacts on food production shortfalls and water availability in Russia
.
Glob. Environ. Change
17
,
429
444
.
Allen
R. G.
Pereira
L. S.
Raes
D.
Smith
M.
1998
Guidelines for Computing Crop Water Requirements, Irrigation and Drainage Paper 56
.
Food and Agricultural Organization of the United Nations
,
Rome
.
Attri
S. D.
Rathore
L. S.
2003
Simulation of impact projected climate change on wheat in India
.
Int. J. Climatol.
23
,
693
705
.
Babel
M. S.
Bhusal
S. P.
Wahid
S. M.
Agarwal
A.
2013
Climate change and water resources in the Bagmati River basin, Nepal
.
Theor. Appl. Climatol.
115
(
3–4
),
639
654
.
Basnet
B. S.
Avaste
R. K.
Bhutia
K. G.
2003
Present status of maize cultivation in Sikkim and future strategies
.
ENVIS Bulletin: Himalayan Ecology
11
,
17
25
.
Blöschl
G.
Montanari
A.
2010
Climate change impacts – throwing the dice?
Hydrol. Process.
24
,
374
381
.
Bryan
E.
Deressa
T. T.
Gbetibouo
G. A.
Ringler
C.
2009
Adaptation to climate change in Ethiopia and South Africa: options and constraints
.
Environ. Sci. Policy
12
,
413
426
.
Carter
T. R.
2007
IPCC-TGICA: general guidelines on the use of scenario data for climate impact and adaptation assessment (TGICA), version 2
.
Debnath
P.
Deb
P.
Sen
D.
Pattannaik
S. K.
Sah
D.
Ghosh
S. K.
2012
Physico-chemical properties and its relationship with water holding capacity of cultivated soils along altitudinal gradient in Sikkim
.
Intl. J. Agric. Env. Biotech.
5
,
99
102
.
Du
M. Y.
Kawashima
S.
Yonemura
S.
Zhang
X. Z.
Chen
S. B.
2004
Mutual influence between human activities and climate change in the Tibetan Plateau during recent years
.
Global Planet. Change
41
,
241
249
.
GIAHS
2009
Sikkim-Himalaya-Agriculture: Improving and scaling up of the Traditionally Managed Agricultural Systems of Global Significance (Sikkim State – India). http://www.fao.org/nr/giahs/other-systems/other/asia-pacific/en (Accessed 26 July 2013)
.
Harrison
L.
Michaelsen
J.
Funk
C.
Husak
G.
2011
Effects of temperature changes on maize production in Mozambique
.
Clim. Res.
46
,
211
222
.
Hassan
Z.
Shamsudin
S.
Harun
S.
2013
Application of SDSM and LARS-WG for simulating and downscaling of rainfall and temperature
.
Theor. Appl. Climatol
.
116
(
1–2
),
243
257
.
Hoogenboom
G.
Jones
J. W.
Wilkens
P. W.
Porter
C. H.
Boote
K. J.
Hunt
L. A.
Singh
U.
Lizaso
J. L.
White
J. W.
Uryasev
O.
Royce
F. S.
Ogoshi
R.
Gijsman
A. J.
Tsuji
G. Y.
Koo
J.
2012
Decision Support System for Agrotechnology Transfer (DSSAT) Version 4.5 [CD-ROM]
.
University of Hawaii
,
Honolulu
,
Hawaii
.
ICIMOD
2010
Rural Livelihoods and Adaptation to Climate Change in the Himalayas; INFORMATION SHEET #5/10. Prepared by International Centre for Integrated Mountain Development Publications Unit, November 2010
,
Kathmandu
,
Nepal
.
Intergovernmental Panel on Climate Change (IPCC)
2007
Summary for policy makers. In: Parry ML, Canzani OF, Palutikof JP, et al. Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change
,
Cambridge University Press
,
Cambridge
,
UK
.
Jeffrey
W.
White
J. W.
Jones
J. W.
Porter
C.
McMaster
G. S.
Sommer
R.
2010
Issues of spatial and temporal scale in modeling the effects of field operations on soil properties
.
Oper. Res. Int. J.
10
,
279
299
.
Jones
J. W.
Hoogenboom
G.
Porter
C. H.
Boote
K. J.
Batchelor
W. D.
Hunt
L. A.
Wilkens
P. W.
Singh
U.
Gijsman
A. J.
Ritchie
J. T.
2003
DSSAT Cropping System Model
.
Eur. J. Agron.
18
,
235
265
.
Joshi
K.
Chaturvedi
P.
2013
Impact of climate change on agriculture
.
Oct. J. Environ. Res.
1
,
39
42
.
Khan
M. S.
Coulibaly
P.
Dibike
Y.
2006
Uncertainty analysis of statistical downscaling methods
.
J. Hydrol.
319
,
357
382
.
Kiem
A. S.
Ishidaira
H.
Hapuarachchi
H. P.
Zhou
M. C.
Hirabayashi
Y.
Takeuchi
K.
2008
Future hydroclimatology of the Mekong River basin under climate change simulated using the high resolution Japan Meteorological Agency (JMA) AGCM
.
Hydrol. Process.
22
,
1382
1394
.
Koutsoyiannis
D.
Efstratiadis
A.
Mamassis
N.
Christofides
A.
2008
On the credibility of climate predictions
.
Hydrol. Sci. J.
53
,
671
684
.
Krishnan
P.
Swain
D. K.
Bhaskar
B. C.
Nayak
S. K.
Dash
R. N.
2007
Impact of elevated CO2 and temperature on rice yield and methods of adaptation as evaluated by crop simulation studies
.
Agric. Ecosyst. Environ.
122
,
233
242
.
Lobells
D. B.
Burke
M. B.
Tabaidi
C.
Mastrandrea
M. D.
Falcon
W. P.
Naylor
R. L.
2008
Prioritizing climate change adaptation need for food security in 2030
.
Science
319
,
607
610
.
Malla
G.
2008
Climate change and its impact on Nepalese agriculture
.
J. Agric. Environ.
9
,
62
71
.
Mishra
B.
Babel
M. S.
Tripathi
N. K.
2013
Analysis of climatic variability and snow cover in the Kaligandaki River Basin, Himalaya, Nepal
.
Theor. Appl. Climatol.
116
(
3–4
),
681
694
.
Montanari
A.
Blöschl
G.
Sivapalan
M.
Savenije
H.
2010
Getting on target
.
Pub. Ser. Rev.: Sci. Tech.
7
,
167
169
.
Moradi
R.
Koocheki
A.
Mahallati
M. N.
2013a
Adaptation of maize to climate change impacts in Iran
.
Mitig. Adapt. Strateg. Glob. Change
19
(
8
),
1223
1238
.
Moradi
R.
Koocheki
A.
Mahallati
M. N.
Mansoori
H.
2013b
Adaptation strategies for maize cultivation under climate change in Iran: irrigation and planting date management
.
Mitig. Adapt. Strateg. Glob. Change
18
,
265
284
.
Olesen
J. E.
Trnka
M.
Kersebaum
K. C.
Skjelvag
A. O.
Seguin
B.
Peltonen-Sainio
P.
Rossi
F.
Kozyra
J.
Micale
F.
2011
Impacts and adaptation of European crop production systems to climate change
.
Eur. J. Agron.
33
,
96
112
.
O'Neill
T.
2007
National Geographic Magazine
.
National Geography Society October 2007
,
Washington, DC
.
Panda
R. K.
Alam
J.
Nandgude
S.
2012
Effect of climate variability on maize yield and evaluation of coping strategies using the crop growth model
.
Int. J. Clim. Change: Impacts and Responses
3
,
71
94
.
Parry
M. L.
Canziani
O. F.
Palutikof
J. P.
van der Linden
P. J.
Hanson
C. E.
2007
Climate Change 2007: Impacts, Adaptation and Vulnerability. Contribution of Working Group II to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change
,
Cambridge University Press
,
Cambridge
,
UK
, pp.
976
.
Prudhomme
C.
Reynard
N.
Crooks
S.
2002
Downscaling of global climate models for flood frequency analysis: where are we now? Hydrol
.
Process.
16
,
1137
1150
.
Ray
D. K.
Ramankutty
N.
Mueller
N. D.
West
P. C.
Foley
J. A.
2012
Recent patterns of crop yield growth and stagnation
.
Nat. Commun
.
3
,
1293
.
Rezaei
E. E.
Gaiser
T.
Siebert
S.
Ewert
F.
2013
Adaptation of crop production to climate change by crop substitution
.
Mitig. Adapt. Strateg. Glob. Change
.
doi:10.1007/s11027-013-9528-1
.
Samadi
S.
Wilson
C. A. M. E.
Moradkhani
H.
2013
Uncertainty analysis of statistical downscaling models using Hadley Centre Coupled Model
.
Theor. Appl. Climatol.
114
(
3–4
),
673
690
.
Seetharam
K.
2008
Climate change scenario over Gangtok
.
Mausam
59
,
361
366
.
Shrestha
S.
Deb
P.
Bui
T. T. T.
2014a
Adaptation strategies for rice cultivation under climate change in Central Vietnam
.
Mitig. Adapt. Strateg. Glob. Change
doi:10.1007/s11027-014-9567-2
.
Shrestha
S.
Khatiwada
M.
Babel
M. S.
Parajuli
K.
2014b
Impact of climate change on river flow and hydropower production in Kulekhani hydropower project of Nepal
.
Environ. Process
1
(
3
),
231
250
.
Sikkimstat Statistical database for Sikkim state
.
2013
Government of India http://www.sikkimstat.com/agriculture/2/stats.aspx (Accessed on 15 August 2013)
.
Smit
B.
Skinner
M. W.
2002
Adaptation options in agriculture to climate change: a typology
.
Mitig. Adapt. Strateg. Glob. Change
7
,
85
114
.
Stainforth
D. A.
Allen
M. R.
Tredger
E. R.
Smith
L. A.
2007
Confidence, uncertainty and decision-support relevance in climate predictions
.
Philos. Trans. R. Soci. Lond. A.
365
,
2145
2161
.
Tao
F.
Yokozawa
M.
Hayashi
Y.
Lin
E.
2003b
Future climate change, the agricultural water cycle, and agricultural productions in China
.
Agric. Ecosyst. Environ.
95
,
203
215
.
Tingem
M.
Rivington
M.
2009
Adaptation for crop agriculture to climate change in Cameroon: turning on the heat
.
Mitig. Adapt. Strateg. Glob. Change
14
,
153
168
.
Travasso
M. I.
Magrin
G. O.
Baethgen
W. E.
Castano
J. P.
Rodriguez
G. R.
Pires
J. L.
Gimenez
A.
Cunha
G.
Fernandes
M.
2006
Adaptation Measures for Maize and Soybean in Southeastern South America
.
AIACC Working Paper No. 28
pp.
16
31
.
Twine
T. E.
Bryant
J. J.
Richter
K. T.
Bernacchi
C. J.
McConnaughay
K. D.
Morris
S. J.
Leakey
A. D. B.
2013
Impacts of elevated CO2 concentration on the productivity and surface energy budget of the soybean and maize ecosystem in the Midwest USA
.
Glob. Change Biol.
19
,
2838
2852
.
Verdon-Kidd
D. C.
Kiem
A. S.
2010
Quantifying drought risk in a non-stationary climate
.
J. Hydrometeorol.
11
,
1019
1031
.
Wilby
R. L.
Dawson
C. W.
2007
SDSM 4.2 – a Decision Support Tool for the Assessment of Regional Climate Change Impacts:
User Manual. Available from https://co-public.lboro.ac.uk/cocwd/SDSM/main.html (accessed on 23 April 2014)
.
Wilby
R. L.
Dawson
C. W.
Barrow
E. M.
2002
SDSM – a decision support tool for the assessment of regional climate change impacts
.
Environ. Model Softw.
17
,
147
159
.
Yadav
S.
Deb
P.
Kumar
S.
Pandey
V.
Pandey
P. K.
2015
Trends in major and minor meteorological variables and their influence on reference evapotranspiration for mid Himalayan region at East Sikkim, India
.
J. Mt. Sci
.
doi:10.1007/s11629-014-3238-3
.