Journal of Agriculture and Food Research 11 (2023) 100480

Contents lists available at ScienceDirect

Journal of Agriculture and Food Research

journal homepage: www.sciencedirect.com/journal/journal-of-agriculture-and-food-research

Future climate implication on maize (Zea mays) productivity with adaptive

options at Harbu district, Ethiopia
Fikru Chekole Chekole, Adem Mohammed Ahmed *
College of Agriculture, Wollo University, P.O. Box 1145, Dessie, Ethiopia

A R T I C L E I N F O A B S T R A C T

Keywords: Maize (Zea mays L.) is one of the important food crops in Ethiopia with significant area coverage but the yield is
Adaptation very below as compared to its genetic potential. The major factors for the low productivity of maize are water
Amhara region deficit, low soil fertility and heat stress. Climate change may further reduce future maize productivity. This study
Crop model
was conducted with the following objectives (1) to evaluate the performance of the CERES-maize model for
Drought
Cultivars
simulating phenology, growth and yield of maize (2) to assess impact of climate changes on maize productivity in
Nitrogen two time periods (2030s and 2060s) and (3) to explore adaptive measures for maize crop under the changing
climate. The CERES-maize model was first calibrated using the phenology, growth, and yield data obtained from
field experiments. Future rainfall and temperature data were projected using the 17 General Circulation Models
(GCMs) using the two climate change scenarios (RCP4.5 and RCP8.5). Well-adapted maize cultivar (melkasa-2)
was used as a test crop. Three maize cultivars (short, medium and long maturity duration) and four levels of
nitrogen (0, 46, 92 and 138kgNha− 1) were evaluated as adaptive measures to impact of climate change on maize.
The model calibration result showed that the root mean square error (RMSE) between the observed and the
predicted values for anthesis, physiological maturity, grain yield and aboveground biomass yield were 1 day, 3
days, 112kgha− 1, and 814kgha− 1, respectively indicated well agreement between the simulated and observed
values. The values during the model evaluation, were 1 day, 2.52 days, 540.9kgha− 1 and 755.3kgha− 1 for the
respective parameters. The overall results showed that the crop genetic coefficients were properly determined.
The result from impact analysis showed that anthesis, physiological maturity, grain yield and aboveground
biomass yields will likely decrease in the 2030s and 2060s under both RCP scenarios. Mean grain yield may
decrease by 14.15% in the 2030s and by 17.2%, in the 2060s under both RCPs. The management scenario
indicated that changes in cultivars and use of nitrogen fertilizer will likely increase maize productivity under
future climate conditions. We concluded that future climate may adversely affect maize productivity but the use
of adaptation measures may reverse such impacts.

1. Introduction insecurity, poor social and institutional infrastructure, exasperated by

variability and extreme events due to climate change are major
In Ethiopia, about 66% of the areas fall under semi-arid and arid contributing factors to poverty and malnutrition.
climate zones and it comprises most of the lowlands [1]. Most areas in Maize (Zea mays) is a leading cereal crop grown worldwide [2]. It is
the study region are situated under this climate zone. The farming sys­ also an important crop in Ethiopia with significant area coverage. Maize
tem in the study region is mixed farming cereals being the primary is the second in area coverage after teff (Eragrostis teff) and the first in
crops. Soils are highly degraded and agricultural productivity is very total production [3]. Maize contributes significant importance to human
low. Rainfall is highly variable, and droughts are frequent. The majority diets [4] with the highest percentage of calorie intake [5]. It is the most
of the communities live a subsistence lifestyle that leaves them with widely consumed staple food crop widely grown in sub-Saharan Africa
limited alternative sources of income and farmers use existing technol­ (SSA) and covers more than 33 million hectares each year [6]. Maize
ogies and practices. The rainfed areas are the hotspots of poverty, covers nearly 17% of the 200 million hectares of cultivated land in
malnutrition and food insecurity. Severe land degradation, water Sub-Saharan Africa and consumed by people with varying food

* Corresponding author.
E-mail addresses: fikruc19@gmail.com (F.C. Chekole), ademmohammed346@gmail.com (A. Mohammed Ahmed).

https://doi.org/10.1016/j.jafr.2022.100480
Received 4 August 2022; Received in revised form 7 December 2022; Accepted 14 December 2022
Available online 18 December 2022
2666-1543/© 2022 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-
nc-nd/4.0/).
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

preferences and socioeconomic statuses [6]. Studies indicated that the environmental condition before they are used for different applications.
global demand for maize may increase in 2050by 50% [4]. Thus, con­ The calibration and evaluation of the model help to confirm whether the
sumption for the human diet may majorly increase in developing model simulates real situations. However, the CERES-Maize model has
countries due to the high population growth rate [7]. not been tested and used in the study region. Thus, this study was
Maize yield in most developing countries is very low compared to initiated with the objectives to (1) calibrate and evaluate the
other regions with average yield of 1.7tha− 1, 1.5tha− 1, and 1.1tha− 1 in CERES-maize model for simulation of phenology, growth and yield of
West Africa, East Africa and Southern Africa, respectively. In Ethiopia, maize (2) project the magnitude of climate changes (temperature and
the yield is greater than 3t ha− 1 with a significant contribution to the rainfall) in near future (2020–2049) and mid-term (2050–2079 (3)
total crop productivity [8] but the current national yield is 2.95tha− 1 predict the impact of projected climate change on maize production and
which is far below the world average (5.66tha− 1) [9]. The low pro­ to explore management options that sustain maize production in the
ductivity is due to several factors such as droughts, decline in soil study region.
fertility, heat stress, poor agronomic practices, limited use of inputs,
insufficient access to technology, lack of credit, low seed quality, dis­ 2. Materials and methods
eases, insects and weeds [10,11].
At present time, Ethiopia is constrained by increased temperatures 2.1. Description of the study area
and prolonged droughts. The increase in temperature and rainfall vari­
ability have been adversely affected mainly the agricultural sector due The study was conducted in Harbu district located in the eastern part
to loss of arable land, shifting in agro-ecological zones, altered crop of the Amhara region, Ethiopia during the 2019 main cropping season. It
growing cycles, and increased incidence of insects, weeds and diseases is located at a distance of 350 km from the capital Addis Ababa in the
[12]. A simulation study using crop models indicated that the average north (Fig. 1) with an altitude of 1450 m above sea level (masl) with a
maize yield will likely decrease in Ethiopia in the range of 5%–33% by latitude of 10◦ 55′ 00′′ N and a longitude of 39◦ 47′ 0′′ E. The mean annual
2050 and by about 46% in the 2080s [13] though the magnitude of yield maximum and mean minimum air temperatures are 27.7 ◦ C and 12.5 ◦ C,
reduction depends on the degree of climate change [14]. The respectively. The study area receives 1098 mm of mean annual rainfall
Inter-governmental Panel on Climate Change [15] projected an increase but it is highly erratic and unpredictable [26]. The rainfall is bimodal
in future temperatures, droughts, floods, desertification, and extreme with a long rainy season locally known as Kiremt extending from June to
weather conditions which will negatively affect agricultural production September whereas the short rainy season known as Belg extends from
in many maize-growing areas. Developing countries will likely be more March to May. Soils are dominated by Vertisol [1]. The farming system is
affected due to the rapid population growth [16]. Thus, the increase in crop-livestock production system. The dominant field crops are sorghum
future temperatures above the optimum range may significantly reduce (Sorghum bicolor L.), teff (Eragrostis teff), maize (Zea mays), chickpea
maize yield (Lobell and Field, 2007; [17]. Based on the prediction result, (Cicer arietinum), and mung bean (Vigna radiata). The location and ag­
both maize gain yield and biomass yield may decrease under future roecology of the study area are indicated in Fig. 1 below.
climates [18,19]. However, the expected effects of future climate change
on maize productivity may vary locally, nationally and regionally. In 2.2. Field experiments
Ethiopia in general and in the study region in particular, there are very
few studies conducted regarding the impact of future climate change on The study was conducted in the 2019 main cropping season which
maize productivity and identification of promising adaptive options. As received 925 mm rainfall during the growing season. Well-adapted
a result, there are limited promising adaptive options available to deal Melkasa-2 maize cultivar was used as a test crop. For calibrating the
with the adverse impacts of climate change and climate variability on model, the crop was sown on 10 July in a 10 m*10 m plot size replicated
maize crop. Some studies, however, showed that the use of fertilizers, three times. Two seeds per hill were planted with a spacing of 75 cm
change in sowing dates and supplemental irrigation are some of the between rows and 25 cm within rows to ensure germination for good
options to deal with climate change impacts [20]. He et al. [18] used the stands of the variety and thinned to one plant 12 days after the crop
DSSAT model to evaluate the yield response of maize to different fer­ emergence. Locally recommended blended fertilizer (NPS) which con­
tilizer levels and obtained promising results. Studies also showed that tains 18.9% N, 37.7% P2O5 and 6.95% S) and urea fertilizer (46% N)
crop rotation is beneficial for improving soil physical quality, nutrient were used. All the blended fertilizer was applied to the soil at the time of
availability and soil microbial diversity resulting in high yield and low sowing at a rate of 100kgha− 1 whereas the urea was applied in split at a
environmental risk [21]. [22] showed that the inclusion of legume crops rate of 50kgha− 1 (the half at the sowing of the crop and the remaining
in the crop rotation system was useful for mitigating the effects of half at the knee height stage of the crop). All other agronomic practices
climate change. were applied as required. Crop pests (insects, diseases and weeds) were
At present, computer models play a valuable contribution to better properly controlled throughout the growing period.
investigating crop responses to environmental factors and also predict­ Major phenological events such as days to 50% anthesis and days to
ing yield under different environments and management scenarios. physiological maturity were recorded from five representative plants
Models greatly facilitate the task of optimizing crop growth and deriving which were tagged in each plot. Physiological maturity was defined as
recommendations. Models can also predict the potential impact of the time when the tip of the seeds turned dark brown. Grain yield was
climate change on crop production, long-term soil carbon sequestration recorded at physiological maturity from the net central three rows with
and carbon stock of a landscape. Crop models can provide management a net area of 2.25 m * 3.9 m excluding plants from either end of the rows.
scenarios for adapting crops to climate change. The Decision Support Aboveground dry biomass yield and grain yield were measured after
System for Agrotechnology Transfer (DSSAT) is a software package that oven drying the samples at 80 ◦ C for 48 h. Time series data on above­
integrates the effects of crop phenotypes, soils, weather, and crop ground biomass yield and leaf area index were measured every 10 days
management and has been used to simulate different experiments. The from five plants randomly selected from each plot. Leaf area at 50%
DSSAT has been used to manage soil fertility, irrigation, yield gaps, silking was measured by multiplying leaf length by maximum leaf width
genotype by environment interactions, predict the impact of climate and was adjusted by a correction factor of 0.75 (0.75 * leaf length *
change on crops and select adaptive management options [23]. The maximum leaf width) as suggested by Ref. [27]. Thus, the Leaf area
CERES-maize model is one of the crop models imbedded in DSSAT index (LAI) was calculated by dividing the leaf area by the sampled
package. The model can predict the impact of climate change on maize ground area. For the evaluation of the crop model, phenological and
production and also explore adaptation options [24,25]. However, all yield data were obtained from field experiments conducted under ran­
the crop models need to be calibrated and evaluated under a given domized complete block design (RCBD) in the study region.

2
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

Fig. 1. Map showing locations of Harbu district and Amhara region in Ethiopia with the traditional agroecology classification (based on altitude).

2.3. Description of the DSSAT package collected from a soil profile near the experimental site at depth of 1.5 m
and analysed following standard procedures. The samples were also
The DSSAT (V4.7.5) was used for simulating the impact of projected analysed for soil drained upper limit, soil lower limit, saturated water
climate change on maize productivity and to explore management op­ content, stage1 soil evaporation coefficient, and runoff curve number.
tions. The DSSAT package consists of a plant module, a soil module with The whole profile drainage rate coefficients were initially estimated
soil water submodule, soil organic carbon, nitrogen sub modules, soil- using soil texture, soil organic matter content, and soil bulk density by
plant-atmosphere interface module, a weather module and manage­ inputting them into the soil file creation utility program of the DSSAT
ment module. The crop modules are the CERES, the CROPGRO, the software.
CROPSIM and the SUBSTOR modules which can simulate the yield,
development and growth of many different crops [28]. Crop models can 2.4.3. Weather data
simulate yield, growth and development based on the characteristics of Daily weather data for the entire growing season are required for
the simulated crop. calibrating and evaluating the model. Daily maximum and minimum air
The CERES-maize has been tested and used by many researchers temperatures (◦ C), daily total precipitation (mm), and daily total solar
around the world for various applications. The CERES is a family crop- radiation (M J M− 2 Day− 1) were obtained from the nearest weather
soil-climate computer model at the core of computer the DSSAT [29]. station at Kombolcha approximately 15 km from the experimental field.
The DSSAT package integrates crop models to asses yield, resource use The WeatherMan utility program in the DSSAT model was used to
and risk associated with different crop production practices. To use convert the sunshine hours to solar radiation (M J M− 2 Day− 1). Histor­
DSSAT for management decisions, the CERES-maize needs to be cali­ ical daily climate data (1980–2009) for rainfall, maximum and mini­
brated in the new environment. The set of genetic coefficients for the mum temperatures, and daily sunshine hours were obtained from the
CERES-maize model (Hoogenboom et al., 2010) are duration in degree National Meteorological Agency (NMA) in Addis Ababa, Ethiopia. The
days from emergence to the juvenile stage (P1), photoperiod sensitivity daily data gaps were corrected with a monthly bias-corrected version of
coefficient (P2), duration in growing degree days from silking to phys­ the closest grid point of the AgMERRA data set following the Agricul­
iological maturity (P5), maximum possible number of grains per head tural Model Intercomparison and Improvement Project (AgMIP) pro­
(G2), potential kernel growth rate during the linear kernel filling phase, tocols [31,32]. AgMIP is a worldwide cooperative effort that links
the maximum kernel growth rate in mg/(kernel/d)and duration in de­ climate, crop, and socioeconomic modelling to produce improved
gree days for a leaf tip to emerge (Phylochron interval (PHINT)) [28]. modelling capacity and better local, regional, and global integrated
The minimum data needed for model calibration and evaluation and assessment of climate change impacts on the agricultural sector [33].
further details on the model process have been presented by Refs. [28, Future daily climate data (rainfall, maximum temperature, minimum
30]. temperature, and solar radiation) for the 2030s (2020–2049) and 2060s
(2050–2079) were obtained from the 17 CMIP5 GCM run under RCP4.5
and RCP8.5 scenarios (Table 1). RCPs are the greenhouse gas concen­
2.4. Model inputs tration trajectories adopted by the IPCC for its fifth assessment [46] and
are well described in the IPCC [46]. The RCP4.5 scenario represents
2.4.1. Crop management data GHG concentrations with increasing speed until the forcing is 4.5 Wm-2
Crop models require management information like sowing date, in the year 2100. This is a moderate emission scenario of concentration
sowing depth, plant spacing, simulation start date, cultivar type, soil rise. The RCP8.5 scenario meant GHG concentrations rise until the
type, fertilizer management, fertilizer type, time and depth of applica­ forcing is 8.5 W m− 2 in the year 2100. This is a high emission or strong
tion. These data were input into the model during the simulations. forcing scenario. In the present study, RCP4.5 and RCP8.5 were used
Recently recommended blended fertilizer (NPS) and nitrogen in the because RCP4.5 represents the moderate scenario whereas RCP8.5
form of urea were used at rates of 100 kg ha− 1 and 50kgha− 1, represents the worst scenario. The RCP2.6 scenario was not included in
respectively. this study because this is a highly ambitious scenario and may not reflect
the situation of the future climate.
2.4.2. Soil data We used the MarkismGCM software to downscale the outputs of the
Important soil data required by the model are soil texture, soil pH, GCM into local conditions. The software uses a climate record for the
organic carbon, total nitrogen, available phosphorous, exchangeable location based on latitude, longitude and elevation. Thus, the monthly
cations, electrical conductivity, and bulk density. These data were

3
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

Table 1 2.5. Procedure for model calibration

The 17 CMIP5 global circulation models (GCMs) used in the study.
Models Institutions Resolutions Latitude x The CERES-maize model in DSSAT (V 4.7.5) was used to predict the
longitude phenology, growth and yield of maize. The model was calibrated using
BCC-CSM 1.1 Beijing Climate Centre, China 2.8125 × 2.8125 data collected from a field experiment. Model calibration is a process
[34] Meteorological Administration whereby crop genetic coefficients are adjusted until simulated values
BCC-CSM 1.1 Beijing Climate Centre, China 2.8125 × 2.8125 agree well with the observed values. The genetic coefficients of the
(m) [34] Meteorological Administration melkasa-2 maize cultivar were estimated manually by model iteration
CSIRO-Mk Commonwealth Scientific and 1.875 × 1.875
3.6.0 [35] Industrial Research Organization and
until close matches were obtained between simulated and observed
the Queensland Climate Change Centre anthesis, physiological maturity, grain yield, aboveground biomass
of Excellence yield, and leaf area index. During the calibration, the crop genetic co­
FIO-ESM [36] The First Institute of Oceanography, 2.812 × 2.812 efficients that determine the phenology of the crop were first adjusted
SOA, China
followed by coefficients that determine growth and finally coefficients
GFDL-CM 3 Geophysical Fluid Dynamics Laboratory 2.0 × 2.5
[37] that determine yield and yield components. The performance of the
GFDL-ESM 2G Geophysical Fluid Dynamics Laboratory 2.0 × 2.5 model during the calibration was tested using coefficient of determi­
[38] nation (R2), index of agreement (d) [49] and root mean square error
GFDL-ESM2M Geophysical Fluid Dynamics Laboratory 2.0 × 2.5 (RMSE) [50]. The adjusted coefficients were used in the subsequent
[38]
GISS-E2-H [39] NASA Goddard Institute for Space 2.0 × 2.5
evaluation and applications of the model.
Studies
GISS-E2-R NASA Goddard Institute for Space 2.0 × 2.5
2.6. Procedure for model evaluation
[39]. Studies
HadGEM2-ES Met Office Hadley Centre 1.2414 × 1.875
[40]. Model evaluation meant comparing the simulated outputs with the
IPSL-CM5A-LR Institute Pierre-Simon Laplace 1.875 × 3.75 observed values. Model evaluation is essential procedure to evaluate the
[41]. capability of the calibrated model to predict the real situation. Data for
IPSL-CM5A- Institute Pierre-Simon Laplace 1.2587 × 2.5
the model evaluation were obtained from yield trial experiments con­
MR [41]
MIROC-ESM Atmosphere and Ocean Research 2.8125 × 2.8125 ducted in the study region. The crop parameters used for model evalu­
[42]. Institute (The University of Tokyo), ation were days to anthesis, days to physiological maturity, grain yield
National Institute for Environmental and aboveground biomass yield. Finally, the performance of the crop
Studies, and Japan Agency for Marine- model was tested statistically using root mean square Error (RMSE),
Earth Science and Technology
MIROC-ESM- Atmosphere and Ocean Research 2.8125 × 2.8125
Willmot‟s Index of agreement (d), coefficient of determination (R2) and
CHEM [42] Institute (The University of Tokyo), normalized root mean square error (NRMSE).
National Institute for Environmental √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅
Studies, and Japan Agency for Marine-
∑i 2
i=1 (Pi − Oi )
Earth Science and Technology RMSE =
n
MIROC5 [43] Japan Agency for Marine-Earth Science 1.4063 × 1.4063
and Technology,
Atmosphere and Ocean Research Where RMSE is the root mean square error, Pi and Oi are the simulated
Institute (The University of Tokyo), and and observed values, respectively, and n is the number of observations.
National Institute for Environmental ⎡ ⎤
Studies ∑ n
(Pi − Oi)2
MRI-CGCM3 Meteorological Research Institute 1.125 × 1.125 ⎢ i=1
⎥
[44] d=1 − ⎢⎣∑n
⎥ 0 < d < 1
⎦
2
Nor-ESM1-M Norwegian Climate Centre 1.875 × 2.5 (|Pi − O|) + (|oi − O|)
[45]
i=0

Where d is the index of agreement. Oi and Pi are observed and predicted

values of rainfall, daily average temperature, and daily average diurnal values, respectively, for the ith data pair. O is mean of the observed.
temperature variations were important records. The MarkSim also RMSE
included the temporal phase angle. The rotation eliminated timing dif­ NRMSE =
N
x 100
ferences in climate events. Thus, the analysis was performed on stan­
dardized climate data. The climate record was rotated to a standard date Where NRMSE is the normalized root mean square error expressed as a
using 12 points fast Fourier transform, on the first phase angle calculated percentage. N is the mean of the observed variable.
using both rainfall and temperature [47]. In MarkSim, almost all oper­
ations are performed in the rotated data space. The WorldClim V1.3 2.7. Method of analysis of climate change impact on maize production
climate database was used to interpolate the climate at the required
point. It uses historical weather data from several databases and used The calibrated CERES-maize model was used to assess the impact of
thin plate smoothing with a fixed lapse rate by employing the program projected climate change on maize production in the 2030s and 2060s.
ANUSPLIN. Bicubic interpolation was used over a kernel of the nearest The daily climate data used by the model were maximum and minimum
sixteen GCM cells on a 1 × 1 grid of GCM differentials. These were temperatures, solar radiation and rainfall. Well adapted maize cultivar
calculated from polynomials fitted to each GCM result and were used to (melkasa-2) was used as a test crop. The simulation was initiated on 10
return the values for any year or RCP. Thus, the ensemble of 17 GCMs July when the soil profile was at the upper limit of soil water and the
was calculated directly from the polynomial coefficients for each GCM. simulation was performed under rainfall conditions. Therefore, the
The estimated GCM differential values were then added to the rotated model was run for the baseline period (1980–2009) and for future
records which is unintelligent downscaling to local monthly climate climate periods (2030s and 2060s) under RCP4.5 and RCP8.5 scenarios
values [48]. The software was used stochastically downscale the daily under the typical crop managements and soil conditions of the study
weather sequences. area. Thus, the model was run and the average yield was recorded for
each climate period. The output values from the model were used for
comparison as follows.

4
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

Predicted value − Baseline value Table 2

Change (%) = *100
Bseline value The calibrated maize genetic coefficient for the variety melkassa-2 within the
DSSAT model.
Parameters Description Genetic
2.8. Method of analysis of adaptation options for maize production coefficients

P1 Thermal time from seedling emergence to the end 245

Three maize cultivars with different maturity duration (short, me­ of the juvenile phase (expressed in degree days
dium and long duration) and four levels of nitrogen ((0, 46, 92, above a base temperature of 8o C) during which the
138kgha− 1) were tested as adaptative options against the adverse plant is not responsive to changes in photoperiod.
P2 Extent to which development (expressed as days) is 0.60
impact of climate changes in the 2030s and 2050s under both RCP4.5
delayed for each hour increase in photoperiod
and RCP8.5 scenarios. Virtual cultivars that incorporated various traits above the longest photoperiod at which
of the standard cultivar (melkasa-2) were developed (1) the standard development proceeds at a maximum rate (which
cultivar (no change), (2) cultivar with 10% shorter maturity duration is considered to be 12.5 h).
P5 Thermal time from silking to physiological 850.0
and (3) cultivar with 10% longer maturity duration. To develop a short
maturity (expressed in degree days above a base
duration cultivar, the genetic coefficients of the standard cultivar (mel­ temperature of 8o C).
kasa-2) that determine seedling emergence to the end of the juvenile G2 Maximum possible number of kernels per plant. 780.0
phase (P1), thermal time from the end of the juvenile stage to tassel G3 Kernel filling rate during the linear grain filling 8.50
initiation (P2) and thermal time from silking to physiological maturity stage and under optimum conditions (mg day − 1).
PHINT Phylochron interval; the interval in thermal time 48.0
(P5) were decreased by 10% whereas for developing a longer duration
(degree days) between successive leaf tip
cultivar, those coefficients were increased by 10%. Therefore, here after appearances.
the cultivars are designated as SMC (short maturing cultivar), SC
(standard cultivar) and LMC (long maturing cultivar). In the simulation,
individual analysis was performed to predict the responses of maize to
Table 3
cultivars change and nitrogen levels. Comparison between simulated and observed anthesis, physiological maturity,
grain yield, aboveground biomass yield and leaf area index of melkasa -2 cultivar
during the model calibration at Harbu, Ethiopia.
2.9. Statistical analysis
Crop parameters Simulated Measured RMSE nRMSE
(%)
Analysis of variance (ANOVA) technique was employed to compare
Anthesis (days) 62 61 1 1.60
the impact of climate change on maize production and also to compare Physiological maturity (days) 141 144 3 2.00
management scenarios. Thus, the output data from the model was sta­ Grain yield at harvest (kg ha− 1) 5612 5500 112 2.00
tistically analysed based on randomized complete block design (RCBD). Aboveground biomass yield (kg 16201 17015 814 4.80
In the simulations, years were considered as replications (blocks) ha− 1)
LAI (maximum) 3.350 3.900 0.55 14.10
because the maize yield in one year under a given treatment was not
affected by another year. Prior year carry-over of soil water was not
considered. Since each simulation year had unpredictable weather
conditions, formal randomization of simulation years was not needed
[26]. The analysis was performed using a statistical analysis system [51]
software and treatments means were compared using the least signifi­
cant test (LSD) at 5% and 1% probability levels. Description statistics
(%) was also employed for comparing mean values.

3. Results and discussion

3.1. Results of model calibration

The CERES-Maize model uses six genetic coefficients for the simu­
lation of phenology, growth and yield. The descriptions of the genetic
coefficients and the calibrated values are indicated in Table 2 below.
The comparisons between the observed and simulated anthesis,
physiological maturity, grain yield, above-ground biomass yield and leaf
area index for the cultivar melkasa-2 are provided in Table 3 and Fig. 2.
The result showed that the RMSE values for anthesis and physiological
maturity were 1 and 2 days with nRMSE values of 1.60% and 2%,
respectively (Table 3). The RMSE values for grain yield and above­
ground biomass yield were 112kgha− 1 and 814kgha− 1 with nRMSE
values of 2% and 4.8%, respectively (Table 3). From the time series data,
the goodness of fit (R2) was 0.84 for leaf area index and 0.93 for
aboveground biomass yield. The nRMSE values were 14.10% and 4.8%
with d-index values of 93% and 94%, respectively (Fig. 2). The RMSE
value for leaf area index and aboveground biomass yield were 0.57 and
765, respectively (Fig. 2). The overall results of the model calibration
phase indicated that there were strong agreements between the Fig. 2. Comparison between simulated and measured leaf area index (A) and
observed and the simulated values indicating the genetic coefficients aboveground biomass yield (B) of melkasa-2 maize cultivar during model cali­
were adjusted properly. However, it was further evaluated with inde­ bration at Harbu wereda, Ethiopia.
pendent set of data and the result is discussed under section 3.3.

5
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

3.2. Results of model evaluation 3.3. Impact of projected climate change on maize phenology

The results of the model evaluation for anthesis, physiological The projected climate change impact on phenology of maize cultivar
maturity, aboveground biomass yield and grain yield of the melkasa-2 melkasa-2 is depicted in Fig. 4. The simulation result showed that
maize cultivar are provided in Fig. 3. The result of statistical analysis anthesis and physiological maturity of the cultivar may significantly
showed that the goodness of fit (R2) was 75% for anthesis, 64% for decrease (P < 0.05) in the 2030s and 2060s under both RCP4.5 and
aboveground biomass yield, 82% for grain yield and 68% for physio­ RCP8.5 scenarios relative to the base period (1980–2009). The result
logical maturity. The d-index values were 93% for anthesis, 68% for also indicated that the impact varied with the RCP scenarios. The
aboveground yield, 88% for grain yield and 93% for physiological highest reduction in anthesis (9 days) and physiological maturity (11
maturity (Fig. 3). The agreement between the simulated and observed days) were under RCP8.5 in 2060s (Fig. 4) whereas the lowest reduction
values among some of the parameters were not strong. However, the of anthesis (4 days) and physiological maturity (7 days) were under
overall results showed that the performance of the model was good in RCP4.5 in the 2030s (Fig. 4). The highest reductions in anthesis and
simulating phenology, growth and yield of maize. Thus, it was palusable physiological maturity in the 2060s under RCP8.5 scenario could be
to use the model to assess the impact of projected climate on maize associated to the highest increase in temperature for this scenario which
production and to explore adaptation options for sustainable maize accelerated the maize growth and development. This imply that pro­
production in the study region. jected climate change particularly the increase in temperature could
adversely affect maize productivity in the study region. The change in
anthesis and physiological maturity dates directly influence grain yield
and biomass yield of maize by reducing the time required for grain
filling and biomass accumulation periods. In agreement with this result
Liu et al. (2013) reported that high temperatures in future climates could
accelerate phenological stages and reduce the duration of the maize
growth period. The study by Ref. [52] also indicated that future climate
could advance the anthesis and maturity dates of maize which leads to
yield reduction. The study by Ref. [53] showed that high temperatures
in future climates may lead to quick accumulation of heat from planting
to anthesis and the crop flower earlier.

3.4. Impact of projected climate change on maize grain and aboveground

biomass yields

The impact of projected climate change on grain and aboveground

biomass yields of maize are depicted in Fig. 5. The simulated grain and
aboveground biomass yields for the baseline period were 2918kgha− 1
and 8071kgha− 1, respectively (data not shown). Results showed that
both maize grain and aboveground biomass yields may decrease in the
2030s and 2060s (Fig. 5). Similar to the impact on the phenology, the
highest reduction in grain yield (19.1%) and aboveground biomass yield
(6.6%) were under RCP8.5 in 2060s (Fig. 5) whereas the lowest reduc­
tion in grain (12.9%) and aboveground biomass yield (4.6%) were under
RCP4.5 scenario in the 2030s. The magnitude of reduction both in grain
and aboveground biomass yields varied by RCPs and the climate periods
that could be associated with variations in temperature and rainfall. The
overall results showed that decreasing trends in maize grain yield and
aboveground biomass yields under both RCPs in the 2030s and 2060s
(Fig. 5). The highest reduction in grain yield and aboveground biomass
yields in the 2050s under RCP8.5 scenario could be associated with the
highest increase in temperature which may accelerate the phenology of
the crop. The studies by Moradi et al. [54], and Xu et al. [55], also
indicated that high temperature accelerated crop growth stages and led
to grain yield reduction. The study by Ref. [56] also showed that pre­
dicted maize yield decreased by about 20% in the Central Rift Valley,
Ethiopia. Thus, we concluded that projected climate change in the 2030s
and 2060s could adversely affect maize productivity in the study region
but such adverse effects may be reduced by developing suitable adap­
tation options.
The increase in temperature in future climate may reduce water use
of the crop by accelerating water loss from the soil and plant. Water
availability may also be affected by runoff that might be caused by
intense rainfall and reduced water infilteration (Fig. 6). Runoff may
have negative effect on water use and nutrients availability by the crop.
Thus, we concluded that water deficit caused by high temperature and
low water infilteration rate may causes maize yield reduction under the
Fig. 3. Comparison between simulated and measured anthesis (A), physiolog­ projected climate conditions of the study region. Thus, adaptive man­
ical maturity (B), grain yield (C) and aboveground biomass yield (D) of melkasa- agement practices are required to sustain maize productivity in this
2 maize cultivar during the model evaluation at Harbu, Ethiopia. region [57]. suggested that the major causes for the reduction in yield of

6
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

Fig. 4. Change in anthesis and physiological maturity (days) of melkasa-2 maize cultivar in 2030s and 2060s under RCP4.5 and RCP8.5 scenarios in relative to the
base period (1980–2009) at Harbu district, Ethiopia.

Fig. 5. Change in grain yield and biomass yield (%) of melkasa-2 maize cultivar in 2030s and 2060s under RCP4.5 and RCP8.5 scenarios in relative the base period
(1980–2009) at Harbu district, Ethiopia.

Fig. 6. Effect of projected climate change on maize seasonal transpiration (mm) and runoff (mm) in the baseline period, 2030s and 2060s under RCP4.5 and RCP8.5
scenarios at Harbu Wereda, Ethiopia.

maize could be global warming that reduce water availability for crops. by water deficit and outbreak of diseases, insects, and weeds [59]. The
A study by Lv, S. et al. (2014) reported that both potential and attainable majority of maize production in the study region is rainfed and the yield
yield of maize crop decreased by 8.0% from 1961 to 2009 mainly due to may significantly decrease under future climate conditions of the 2030s
climate change [58]. reported that global warming is harmful to most and 2060s.
crops. The negative effect of climate change on crops may be aggravated

7
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

3.5. Adaptation options for maize under the projected climate changes compared to short maturing cultivars with high yield potential [63].
The simulation results also revealed that maize grain yield was
Maize cultivars with different maturity duration (short, medium, and significantly (p < 0.01) increased due to the application of nitrogen
long) and four levels of nitrogen (0, 46, 92 and 138kgha− 1) were eval­ fertilizer in the base period, in the 2030s and 2060s under both RCPs
uated as adaptation options to the impact of climate change on maize (Table 4). Results showed that increasing nitrogen from 0kgha1 to
production. Results indicated that changes in cultivars significantly 184kgha− 1 increased maize grain yield in the present climate and under
reduced the impact of projected climate change on maize both in the the projected climates in the 2030s and 2060s under both RCPs
2030s and 2060s under both RCP scenarios. Based on the simulation (Table 4). However, economically optimum maize grain yield was
result, the long duration cultivar did not show any grain yield variation simulated due to the application of 92kgNha− 1 under both RCP sce­
in respect to the standard cultivar however it increased grain yield by narios and the climate periods. The lowest grain yield was simulated
10.8% in the 2030s under RCP4.5, by 11% in the 2030s under RCP8.5, from treatment that received no nitrogen fertilizer at all under both
by 11% in 2060s under RCP4.5 and by 15.5% in the 2060s under RCPs and climate periods (Table 4). We concluded that nitrogen is most
RCP8.5. However, the short duration maize cultivar significantly limiting nutrient for maize production in the study region. Nitrogen is an
decreased grain yield by 15.8% (base period), by 13.2% in the 2030 essential macro nutrient that must be supplied for maize to sustain its
under RCP4.5, by 14.6% in 2030 ubder RCP8.5, by14.6% in 2060 under productivity. Thus, crop production packages in the study region should
RCP4.5 and by 15.4% in 2060s under RCP8.5in (Fig. 7). Previous studies give focus to nitrogen fertilizer for maize production. The increase in
also showed that short duration varieties may not perform well under yield due to nitrogen application could be associated to greater nitrogen
projected climate change rather long duration varieties may perform and water use efficiency of the maize crop under the changed climate
well. The long duration cultivar did not show significant gran yield conditions. Therefore, maize producers in the study region could apply
variation in the base period but it increased grain yield by 10.8% under 92kgNha− 1 for maize crop to obtain reseanable yield. However, the high
RCP4.5 in 2030, by 11% under RCP8.5 in 2030, by 11% under RCP4.5 in demand in nitrogen application by maize under future climate may be
2060 and by 15.5 under RCP8.5 in 2060 (Fig. 7). The increase in grain of associated to more need of other nutrients like P,K and micronutrients as
the long maturing cultivar might be associated with the capability of the well as more water demand to cope up with the removal with biomass
cultivar to maintain its photosynthetic ability during the grain filling increased by enhanced N application. To solve such constraints, it could
period under the high temperature conditions. The study indicated that be important to apply more nutrients from nutrients sources such as
high temperature in future climate could adversely affect the growth compost and manure that could improve water holding as well as nu­
cycle of the short maturing cultivar more than the long duration cultivar. trients retention capacities of the soil. Increasing more P and K con­
The Study by Liu et al. [60] also showed that late maturing maize cul­ taining fertilizers are also important. The use of irrigation where
tivars at low latitude areas may maintain longer growing cycle. The possible may be important source of water to solve the problem of water
warming climate may increase thermal resources which may allow the deficit under future climate condition. The study by Mohammad et al.
crop to take advantage of the extra heat. However, the elevated tem­ (2010) also indicated that optimum nitrogen increased yield and yield
perature in future climate may accelerate various aspects of crop components of maize under the projected climate change. The result
metabolism and changes the rate of physiological processes. Therefore, showed that increasing nitrogen rate by 60kgha− 1 increased maize yield
crops exposed to high temperatures above the optimum may lead to by 78%–89% across different climate change scenarios in Ethiopia [56].
significant yield reduction. The reduction of crop growth stages under The present study indicated that maize production in low input systems
future climate due to the high temperature may lead to seed number could be modified optimum nitrogen management combined with
reduction, accelerate seed growth rate, reduce seed filling duration, and adapted maize cultivars.
reduce yield. Thus, long maturing maize cultivars may be suggested as
suitable cultivars in the study region under the projected climate change 4. Conclusion
conditions. The study by Ref. [61] revealed that high grain yield in
maize genotypes could be directly attributed to maintaining the Maize (Zea mays L.) is an important cereal crop with significant area
photosynthetic rater during the grain filling period. The study by coverage in Ethiopia. The yield in Ethiopia is below the genetic potential
Ref. [62] also indicated that long duration maize cultivars showed high of the crop. Climate variability and climate change have been suggested
water use efficiency that favored high assimilation and high grain yield to be important constraints for maize. Thus, this study was conducted
whereas the short duration cultivars showed lower water use efficiency with the objectives to (1) evaluate performance of the CERES-maize
that led to low yield. Studies also showed that long maturing cultivars of model in DSSAT for simulating phenology, growth and yield of maize
millet and sorghum were more resilient to future climate conditions as (2) assess impact of projected climate change on maize production and

Fig. 7. Change in grain yield (%) of short and long maturing maize cultivars in relative to the standard cultivar (Melkasa-2) in the baseline period, 2030s and 2060s
under RCP4.5 and RCP8.5 scenarios at Harbu, Ethiopia. LMC, long maturing cultivar: SMC, short maturing cultivar.

8
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

Table 4
Effect of nitrogen (N) on maize grain yield in the baseline, 2030s and 2060 under RCP4.5 and RCP8.5 scenarios at Harbu district, Ethiopia.
N rates (kg ha− 1) Grain yield (kg ha− 1)

Base period (1980–2009) 2030s (RCP4.5) 2030s (RCP8.5) 2060s (RCP4.5) 2060s (RCP8.5)

0 2466.3 2169.7 2173.6 2173.6 2062.0

46 4506.5 3867.4 3894.5 3894.5 3673.2
92 5446.4 4617.6 4572.4 4572.4 4321.1
138 5447.3 4634 4589.5 4589.5 4320.7
184 5448.9 4643.5 4588.4 4588.4 4320.8

LSD (p<5%) 228.5 189.85 192.49 192.49 178.18

CV (%) 9.58 9.31 9.5 9.8 9.34

Where, LSD, least significant difference at 5% probability; CV, coefficient of variation.

(3) explore adaptation measures to the impact of climate change. Data availability
First, the CERES-maize model was calibrated and evaluated in te
study refion by using climate, soil and crop data. Projected climate Data will be made available on request.
change impact and adaptive options were studied using the calibrated.
The daily baseline climate data for the period 1980–2009 was modified References
by the projected daily climate data for the period 2030s and 2060s under
RCP4.5 and RCP8.5 climate scenarios. Three maize cultivars (short, [1] M. Adem, T. Tamado, S. Piara, M. Adamu, S. Ali, Identifying best crop management
practices for chickpea (Cicer arietinum L.) in Northeastern Ethiopia under climate
medium and long durations) and four levels of nitrogen (0, 46, 92, 138 change condition, Agric. Water Manag. 194 (2017) 68–77, 2017.
kgha− 1) were evaluated as adaptive options. [2] B. Shiferaw, B.M. Prasanna, J. Hellin, M. Bänziger, Crops that feed the world 6.
Results of model calibration and evaluation showed that the crop Past successes and future challenges to the role played by maize in global food
security, Food Secur. 3 (2011) 307–327.
genetic coefficients were reasonably estimated. The results of climate [3] CSA, Agricultural Sample Survey Report on Area and Production for Major Crops
change impact analysis showed that days to anthesis, days to physio­ (Private Peasant Holdings Meher Season) for 2007/08, The Federal Democratic
logical maturity, grain and aboveground biomass yields of maize may Republic of Ethiopia, Addis Ababa, Ethiopia, 2015. Statistical Bulletin 278.
[4] A. Ignaciuk, D. Mason-D’Croz, Modeling Adaptation to Climate Change in
decrease in the 2030s and in 2060s. The mean grain yield in the region
Agriculture, OECD Publishing, 2014. OECD Food, Agriculture and Fisheries Papers,
may decrease by 14.15% and 17.2% in the 2030s and in 2060s, No. 70.
respectively. The result of adaptive options showed that changes in [5] E.T. Nuss, S.A. Tanumihardjo, Quality protein maize for Africa: closing the protein
inadequacy gap in vulnerable populations, Adv. Nutr. 2 (2011) 217–222.
cultivars and increasing nitrogen may increase grain yield of maize. We
[6] FAOSTAT, Food and Agriculture Organization of the United Nations 2015 .FAO
concluded that climate change could negatively impact maize produc­ Production Year Book, FAO, Rome, 2015. Available at, http://apps.fao.org.
tivity in future climate of the study region. However, maize yield can be [7] FAO, How to Feed the World in 2050. Issue Brief from the High-Level Expert Forum
increased using adaptive measures. However, this study has cerian Held in Rome, 12–13 October, 2009 (Rome, Italy).
[8] M. Smale, D. Byerlee, T. Jayne, Maizerevolution in Sub-saharan Africa. Policy
limitations as it showed only farm level impact of future climate on Research Working Paper 5659, World Bank, Washington DC, 2011.
maize productivity. Thus, future studies should consider regional anal­ [9] USDA, World Agricultural Production. Foreign Agricultural Service. Circular Series
ysis of climate change impact on maze productivity. WAP 6–16 June 2016. Approved by the World Agricultural Outlook Board, 2016.
[10] A. Taffesse, P. Dorosh, S. Asrat, Crop Production in Ethiopia Regional Patterns and
Trends. Ethiopia Strategy Support Program II, 2011.
Funding [11] T. Erkossa, F. Itanna, K. Stahr, Indexing soil quality: a new paradigm in soil science
research, Aust. J. Soil Res. 45 (2007) 129–137.
[12] P.W. Heisey, K.D. Rubenstein, Using crop genetic resources to help agriculture
Not applicable. adapt to climate change: economics and policy, Econ. Info. Bullet. Number (2015)
139.
Conflict of interest/competing interest [13] A. Muluneh, B. Birhanu, L. Stroosnijder, W. Bewket, S. Keesstra, Impact of
predicted changes in rainfall and atmospheric carbon dioxide on maize and wheat
yields in the Central Rift Valley of Ethiopia, J. Reg. Environ. Change (2015),
There are no conflicts of interest among the authors regarding this https://doi.org/10.1007/s10113-014-0685-x.
manuscript. [14] G. Nelson, M. Rosegrant, Koo, Climate Change: Impact on Agriculture and Costs of
Adaptation, 2010 (IFPRI Food Policy Report).
[15] IPCC, Climate Change: Impacts, Adaptation and, Vulnerability. Summary for Policy
Availability of data and material Makers, 2009. http://64www.ipcc.ch/publications and data/ar4/wg2/en/spm.
html.
[16] A. Solomon, A. Sabrina, C. Manuela, Climate change impact on Ethiopian small-
Not applicable. holder production efficiency, in: Fifth IAERE Annual Conference 16-17 February
2017, Rome, 2017.
Code availability [17] J.L. Hatfield, K.J. Boote, B.A. Kimball, D.W. Wolfe, D.R. Ort, C. Izaurralde, et al.,
Agriculture, in: P. Backlund, A. Janetos, D. Schimel, M. Walsh (Eds.), The Effects of
Climate Change on Agriculture, Land Resources, Water Resources, and Biodiversity
Not applicable. in the United States .U.S. Climate Change Science Program and the Subcommittee
on Global Change Resources, Washington, D.C, 2008, pp. 21–74.
[18] He, et al., Climate change impacts on crop yield, soil water balance and nitrate
Authors’ contribution leaching in the semiarid and humid regions of Canada, PLoS One 13 (11) (2018),
0207370.
The author Fikru Chekole was involved in executing the field [19] W. Smith, et al., Towards an improved methodology for modeling climate change
impacts on cropping systems in cool climates, Sci. Total Environ. 728 (2020),
experiment, data collection and data analysis, and also writing the 138845, https://doi.org/10.1016/j.scito tenv.2020.138845, 2020.
manuscript. Dr Adem Mohammed was involved in the data analysis and [20] E. Bryan, T.T. Deressa, G.A. Gbetibouo, C. Ringler, Adaptation to climate change in
writing of the manuscript. Ethiopia and South Africa: options and constraints, Environ. Sci. Pol. 12 (2009)
413–426.
[21] W.D. Reynolds, C.F. Drury, X.M. Yang, C.S. Tan, J.Y. Yang, Impacts of 48 years of
Declaration of competing interest consistent cropping, fertilization and land management on the physical quality of a
clay loam soil, Can. J. Soil Sci. 94 (2014) 403–419, 2014.
[22] Y. Ma, et al., Modeling the impact of crop rotation with legume on nitrous oxide
The authors declare that there is no conflict of interest among the emissions from rain-fed agricultural systems in Australia under alternative future
authors regarding this manuscript. climate scenarios, Sci. Total Environ. 630 (2018) 1544–1552.

9
F.C. Chekole and A. Mohammed Ahmed Journal of Agriculture and Food Research 11 (2023) 100480

[23] Bhupinderdhir, Crop productivity in changing climate chapter, Sustain. Agricult. [45] A. Kirkevag, T. Iversen, O. Seland, J.B. Debernard, T. Storelvmo, J.E. Kristjansson,
Rev. 27 (2018) 213–241, https://doi.org/10.1007/978-3-319-75190-0-8. Aerosol-cloud-climate interactions in the climate model CAMO slo, Tellus 60 (3)
[24] J.W. White, G. Hoogenboom, B.A. Kimball, G.W. Wall, Methodologies for (2008) 492–512.
simulating impacts of climate change on crop production, Field Crop. Res. 124 [46] IPCC, Summary for policymakers, in: T.F. Stocker, D. Qin, M. Plattner, J. Allen,
(2011) 357–368. A. Boschung, Y. Nauels, V. Bex, P.M. Midgley (Eds.), Climate Change 2013:The
[25] Fu Tongcheng, K.O. Jonghan, W. Gerard, Wall, J. Paul, J.R. Pinter, A. Bruce, Physical Science Basis. Contribution of Working Group I to the Fifth Assessment
J. Kimball Michael, Ottman Han-Yong Kim, Simulation of climate change impacts Report of the Intergovernmental Panel on Climate Change, Cambridge University
on grain sorghum production grownunder free air CO2 enrichment, Int. Agrophys. Press, Cambridge, United Kingdom and New York, NY,USA, 2013.
30 (2016) 311–322, https://doi.org/10.1515/intag-2016-0007. [47] J.W. Jones, G. Hoogenboom, C.H. Porter, K.J. Boote, W.D. Batchelor, L.A. Hunt, P.
[26] M. Adem, T. Tamado, S. Piara, K. Driba, M. Adamu, Modeling climate change W. Wilkens, U. Singh, A.J. Gijsman, J.T. Ritchie, DSSAT cropping system model,
impact on chickpea production and adaptation options in the semi-arid North- Eur. J. Agron. 18 (2003) 235–265.
Eastern Ethiopia, J. Agricult. Environ. Int. Develop. 110 (2) (2016) 377–395, [48] R.L. Wilby, J. Troni, Y. Biot, L. Tedd, B.C. Hewitson, D.M. Smith, R.T. Sutton, A
https://doi.org/10.12895/jaeid.20162.510. review of climate risk information for adaptation and development planning, Int. J.
[27] C. Francis, A.J.N. Rutger, A.F.E. Palmer, A rapid method for plant leaf area Climatol. 29 (2009) 1193–1215.
estimation in maize (Zea mays L), Crop Sci. 9 (1969) 537–539. [49] C.J. Willmott, G.S. Akleson, R.E. Davis, J.J. Feddema, K.M. Klink, D.R. Legates,
[28] G. Hoogenboom, J.W. Jones, C.H. Porter, P.W. Wilkens, K.J. Boote, L.A. Hunt, G. J. O‟Donnell, C.M. Rowe, Statistic for the evaluation and comparison of models, J.
Y. Tsuji (Eds.), Decision Support System for Agrotechnology Transfer Version 4.5, Geophys. Res. 90 (5) (1985) 8995–9005.
vol. 1, University of Hawaii, Honolulu, HI, 2010. Overview. [50] K.M. Loague, R.E. Green, Statistical and graphical methods for evaluating solute
[29] IBSNAT, Decision Support System for Agrotechnology Transfer (DSSAT). User transport models: overview and application, J. Contam. Hydrol. 7 (1991) 51–73.
Guide, University of Hawaii, Honolulu. HI, 1994. Version 3. [51] SAS (Statistical Analysis System) Institute. Inc, The SAS System for Windows TM,
[30] G. Hoogenboom, J.W. Jones, P.C.S. Traore, K.J. Boote, In the book, in: J. Kihara, et 2008.
al. (Eds.), Improving Soil Fertility Recommendations in Africa Using the Decision [52] D. Xiao, Y. Qi, Y. Shen, F. Tao, J.P. Moiwo, J. Liu, et al., Impact of warming climate
Support System for Agrotechnology Transfer (DSSAT), Springer Science + Business and cultivar change on maize phenology in the last three decades in North China
Media B.V, Dordrecht, 2012, https://doi.org/10.1007/978-94-007-2960-52, 2012. Plain, Theor. Appl. Climatol. (2015), https://doi.org/10.1007/s00704-015-1450-x.
[31] AgMIP, Guide for Running AgMIP Climate Scenario Generation Tools with R in [53] S.B. Baviskar, et al., Heat units and heat unit efficiency influenced by environment
Windows, AgMIP, New York, 2013. effect on yield and dry matter of rabi sorghum, Int. J. Chem. Stud. 5 (3) (2017)
[32] AgMIP, The Coordinated Climate-Crop Modeling Project C3MP: an Initiative of the 395–398.
Agricultural Model Intercomparison and Improvement Project. C3MP Protocols [54] R. Moradi, A. Koocheki, M.N. Mahallati, H. Mansoori, Adaptation strategies for
and Procedures, AgMIP, New York, 2013. maize cultivation under climate change in Iran: irrigation and planting date
[33] C. Rosenzweig, A.C. Ruane, J.M. Winter, K.J. Boote, C. Porter, J. Jones, S. Wasseng, management, Mitig. Adapt. Strategies Glob. Change (2) (2012) 265–284, https://
J.L. Hatfield, P. Thorburn, J.M. Antle, G.C. Nelson, S. Janssen, B. Basso, F. Ewert, doi.org/10.1007/s11027-012-9410-6.
D. Wallach, G. Baigorria, The agricultural model intercomparison and [55] C.C. Xu, W.X. Wu, Q.S. Ge, Y. Zhou, Y.M. Lin, Y.M. Li, Simulating climate change
improvement project (AgMIP): protocols and pilot studies, J. Agric. Meteorol. 170 impacts and potential adaptations on rice yields in the Sichuan Basin, China, Mitig.
(2013) 166–182. Adapt. Strategies Glob. Change 22 (4) (2017) 565–594, https://doi.org/10.1007/
[34] T. Wu, A mass-flux cumulus parameterization scheme for large scale models: s11027-015-9688-2.
description and test with observations, Clim. Dynam. 38 (2012) 725–744. [56] B.T. Kassie, S. Asseng, R.P. Rotter, H. Hengsdijk, A.C. Ruane, M.K. Van Ittersum,
[35] M.A. Collier, S.J. Jeffrey, L.D. Rotstayn, K. Wong, S.M. Dravitzki, C. Moeseneder, Exploring climate change impacts and adaptation options for maize production in
C. Hamalainen, J.I. Suppiah, J. Antony, A. Zein, M. Atif, The CSIROMk3.6.0 the Central Rift Valley of Ethiopia using different climate change scenarios and
Atmosphere-Ocean GCM: participation in CMIP5 and data publication, in: 19th crop models, Clim. Change 129 (2015) 145–158, https://doi.org/10.1007/s10584-
International Congress on Modelling and Simulation, Perth, Australia, 12–16 014-1322-x.
December 2011, 2011. http://mssanz.org.au/modsim. [57] E. Guo, et al., Assessing spatiotemporal variation of drought and its impact on
[36] Z. Song, F. Qiao, Y. Song, Response of the equatorial basin-wide SST to wave maize yield in Northeast China, J. Hydrol. 553 (2017) 231–247.
mixing in a climate model: an amendment to tropical bias, J. Geophys. Res. 117 [58] S. Seo, R. Mendelsohn, M. Munasinghe, Climate change impacts Sivakumar 1992.
(2012) C00J26. Climate change and implications for agriculture in Niger, Clim. Change 20 (2005)
[37] L. Donner, The dynamical core, physical parameterizations, and basic simulation 297–312.
characteristics of the atmospheric component AM3 of the GFDL Global Coupled [59] I. Niang, O. Ruppel, M. Abdrabo, A. Essel, C. Lennard, J. Padgham, P. Urquhart,
Model CM3, J. Clim. 24 (13) (2011). Africa, in: Climate Change 2014: Impacts, Adaptation and Vulnerability.
[38] J.P. Dunne, GFDL‟s ESM2 global coupled climate–carbon earth system models. Part Contribution of Working Group II to the Fifth Assessment Report of the
I: physical formulation and baseline simulation characteristics, J. Clim. 25 (2012) Intergovernmental Panel on Climate Change, Cambridge University Press, 2014.
6646–6665. [60] Z. Liu, K.G. Hubbard, X. Lin, X. Yang, Negative effects of climate warming on maize
[39] G.A. Schmidt, et al., Present day atmospheric simulations using GISS Model E: yield are reversed by the changing of sowing date and cultivar selection in
comparison to in-situ, satellite and reanalysis data, J. Clim. 19 (2006) 153–192. Northeast China, Global Change Biol. 19 (2013) 3481–3492.
[40] W.J. Collins, N. Bellouin, M. Doutriaux-Boucher, N. Gedney, P. Halloran, T. Hinton, [61] A.K. Borrell, G.L. Hammer, A.C.L. Douglas, Does maintaining green leaf area in
J. Hughes, C.D. Jones, M. Joshi, S. Liddicoat, G. Martin, F. O‟Connor, J. Rae, sorghum improve yield under drought? I. Leaf growth and senescence, Crop Sci. 40
C. Senior, S. Sitch, I. Totterdell, A. Wiltshire, S. Woodward, Development and (2000) 1026–1037.
evaluation of an Earth-System model-HadGEM2, Geosci. Model Dev. (GMD) 4 (4) [62] G. Zelalem, H. Azama, F. Masresha, GietahunY, Growth, water status,
(2011) 1051–1075, https://doi.org/10.5194/gmd-4-1051-2011. www.geosci-mod physiological, biochemical and yield response of stay green sorghum (sorghum
el-dev.net/4/1051/2011/. bicolor (L.) moench) varieties-A field trial under drought prone area in Amhara
[41] J.L. Dufresne, Climate change projections using the IPSL-CM5 earth system model: regional state, Ethiopia, J. Agron. 14 (2015) 188–202, https://doi.org/10.3923/
from CMIP3 to CMIP5, Clim. Dynam. (2013) 1–43. ja.2015.188.202.
[42] S. Watanabe, S. Emori, M. Kimoto, MIROC-ESM2010: model description and basic [63] B. Sultan, P. Roudier, P. Quirion, A. Alhassane, B. Muller, M. Dingkuhn, P. Ciais,
results of CMIP5-20c3m experiments, Geosci. Model Dev. (GMD) 4 (4) (2011) M. Guimberteau, S. Traore, C. Baron, Assessing climate change impacts on
845–872. sorghum and millet yields in the sudanian and SahelianSavannas of West Africa,
[43] M. Watanabe, Improved climate simulation by MIROC5: mean states, variability, Environ. Res. Lett. 8 (2013), https://doi.org/10.1088/1748-9326/8/1/014040.
and climate sensitivity, J. Clim. 23 (2010) 6312–6335. Article ID: 014040.
[44] S. Yukimoto, A new global climate model of Meteorological Research Institute:
MRI-CGCM3 – model description and basic performance, J. Meteorol. Soc. Jpn. 90a
(2012) 23–64.

10