Response of Crop Water Requirement and Yield of Irrigated Rice to Elevated Temperature in Bangladesh

In the changing climatic condition, temperature is the most vulnerable parameter and is projecting a trend of increase in the future. Crop growth and development process depend largely on air temperature. This study aims to determine the role of increasing air temperature in yield, crop water requirement (CWR), and other agronomic parameters of irrigated rice. Ceres-rice model associated in the Decision Support System for Agrotechnology Transfer (DSSAT) was used in 15 diﬀerent locations of Bangladesh. Grain yield, growth duration, and crop water requirement of widely cultivated irrigated rice (Boro rice) variety BRRI dhan28 were analysed in normal temperature and elevated air temperature by 1 ° C, 2 ° C, 3 ° C, and 4 ° C. The result revealed detrimental eﬀect of elevated temperature on growth duration and grain yield. The estimated highest growth duration reduction of 30 days was found in Moulvibazar for 4 ° C temperature rise. The grain yield reduction was projected by 0–17%, 16–35%, 31–49%, and 39–61% from the normal condition if the seasonal mean temperature increased by 1 ° C, 2 ° C, 3 ° C, and 4 ° C, respectively. The country average crop water requirement was found to be 405mm of which the highest 445 mm and the lowest 358mm were recorded in Moulvibazar and Chandpur, respectively. The study revealed that the country average rice CWR reduced by 5%, 8%, 12%, and 17% over the normal condition for 1 ° C, 2 ° C, 3 ° C, and 4 ° C rising temperature, respectively. For 1 ° C temperature rise, BRRI dhan28 life span shortened by 6.4 days, grain yield reduced by 695kg, and estimated CWR decreased by 14mm. The projected declining CWR indicated that irrigated rice will require less irrigation water, but it will cause considerable yield loss under elevated temperature. Though elevated temperature will save huge irrigation water used in country-wide Boro rice cultivation, the crop developers need to introduce new heat-tolerant cultivar to minimize yield loss.


Introduction
Air temperature is the elementary environmental factor that regulates plant growth and development process. Both seasonal and annual potential evapotranspiration (ET 0 ) were linearly affected by maximum and minimum air temperature [1]. Crop life span mainly depends on accumulated growing degree days (GDD) and is derived from mean daily temperature over base temperature of crop. Besides, grain yield of crop largely depends on air temperature. Intergovernmental panel for climate change (IPCC) projected rising temperature by 1.5°C for RCP 4.5 and 2°C for RCP 8.5 by the end of the 21 st century, relative to the average from the year 1850 to 1900 [2]. A historical (1990-2011) analysis showed increasing trend of daily average temperature in Northwest hydrological region of Bangladesh [3].
In recent days, climate change has been a worldwide concern. Bangladesh is facing climate change challenges for adequate cultivable land, irrigation water availability in dry season, groundwater level depletion, salinity intrusion in river and lands, excessive drought, erratic rainfall distribution, and so forth [4][5][6][7]. Around the world, researchers are continuously trying to understand the effect of climate change on irrigated crop cultivation [8][9][10][11][12]. Various studies predicted a considerable increase in irrigation water demand (10-25 percent) in rice cultivation in different climate change scenarios [8,[13][14][15][16][17][18][19]. However, this high demand for irrigation water due to climate change will most likely bring water shortage or water availability crisis in north-western and southern part of Bangladesh. Already in north-western part of Bangladesh, groundwater level is declining rapidly due to excessive withdrawal for irrigation in dry season and to mitigate the drought [20]. Bangladeshi districts of Rajshahi, Pabna, Rangpur, Bogura, and Dinajpur are experiencing fast groundwater level depletion and huge shortage of irrigation water due to lack of adequate rainfall recharge [21]. In southern part of Bangladesh, in both tidal nonsaline and saline regions, the major rivers like Boleshwar, Burishwar, Tetulia, and Biskhali are the main sources of irrigation water. However, the availability of irrigation water is declining due to increase in river and adjacent canal water salinity [22,23].
Bangladesh crop agriculture is considerably dominated by rice cultivation. Considering all rice growing season in Bangladesh, dry season rice (Boro rice) has the maximum share (53 percent) in total rice production [24]. As a major crop in dry season, Boro rice is fully dependent on irrigation, mostly from the month of January to May, because Boro rice gets a small portion of total annual rainfall in Bangladesh between January and May [7]. One of the major issues for Bangladesh agriculture is the variation in crop water requirement due to climate change in dry season. IPCC [25] projected that climate change will cause temperature increase and uneven rainfall distribution pattern in Bangladesh. Elevated temperature will change evapotranspiration as well as rice growing period length. High temperature will create the increased demand of evapotranspiration. Excess irrigation water supply in the rice field will be required to mitigate that demand. Consequently, rice plant physiology will be changed due to increased temperature and it will shorten the rice growth duration. e short growth span will reduce the number of irrigation days in rice field. However, rice crop is sensitive to temperature during its critical stages and increased temperature above the optimum may reduce yield drastically [26]. Maniruzzaman et al. [24] noticed yield loss and reduced growth duration of irrigated rice (Boro) in elevated air temperature over normal temperature in Bangladesh.
So, a comprehensive analysis of crop water requirement of rice in case of rising temperature, including potential evapotranspiration and irrigated rice growth duration, is necessary to understand the impacts of elevated temperature on irrigated rice cultivation for future planning and management of Bangladesh water resources. In this regard, this study was undertaken to determine the effect of elevated air temperature on crop water requirement and grain yield of irrigated rice.

Study Area.
e study area covered fifteen locations representing the whole Bangladesh (Figure 1). e country has diversified climate around the year and soil condition. e climate of Bangladesh can be categorized into three distinct seasons, the winter (November to February), the hot summer premonsoon (March to May), and the monsoon (June to October) [27]. December to May is the suitable period for Boro rice cultivation. e maximum air temperature during Boro season varied from 28.4 to 31.6°C with the highest in Rajshahi and the lowest in Rangpur region (Table 1). e minimum air temperature ranged from 16.8°C in Rangpur and Moulvibazar to 19.6°C in Patuakhali region. e highest normal rainfall during Boro season was observed in Moulvibazar (770 mm) and the lowest rainfall was in Rajshahi (291 mm).

DSSAT Model.
is study utilized CERES-Rice model incorporated in DSSAT 4.7. e CERES-Rice model accounted weather, soil properties, and management practices on crop performance [28]. e model simulates agronomic yield, yield contributing parameters, nitrogen level, soil-water regimes, hydrological parameters, and so forth under varying managements and environments. It also simulates interaction effects among the varying managements and environments.
is study adopted Penman-Monteith method [29] to calculated reference crop evapotranspiration, soil conservation service method for infiltration, leaf photosynthesis response curve for photosynthesis, and Ritchie water balance for hydrology.

Data Collection and Model Parameterization.
Historical (1981 to 2017) daily weather data of maximum and minimum temperature, relative humidity, wind speed, sunshine hour, and rainfall of fifteen locations were collected from Bangladesh Meteorological Department (BMD). Data of soil physical properties including particle size distribution, bulk density, chemical properties like soil organic carbon and nitrogen, and so forth were collected from Soil Resource Development Institute (SRDI) and are presented in Table 2. All these data were used as input to the model. Daily weather files and soil files were created in the DSSAT model. Major agronomic management inputs to the model are presented in Table 3. For the current study, planting date of 1 December 2015 and seedling age of 45 days were used for all 15 locations in model simulation. Nitrogen fertilizer (in the form of urea) was applied at 15 days after transplanting (DAT), 30 DAT, and 45 DAT according to BRRI recommendation. When standing water disappeared from the field surface, irrigation was applied at the rate of 50 mm.

Model Calibration and Validation.
e tested cultivar BRRI dhan28 is short duration variety of Boro season developed by Bangladesh Rice Research Institute (BRRI). It usually takes 145 days to complete the life cycle [30]. e cultivar coefficient of BRRI dhan28 was taken from Maniruzzaman et al. [31]. e genetic coefficient values of default cultivar BR14 were selected from the DSSAT genotype list (Table 4). e sensitive parameters (Table 4) of the tested   International Journal of Agronomy cultivars were adjusted considering the observed grain yield and growth duration. e model was calibrated with known grain yield and growth duration of "time of transplanting" experiment conducted at BRRI farm, Gazipur, during Boro, 2014-15. e genetic coefficient was calibrated considering the observed and simulated grain yield of the cultivar. e model validation used known values of grain yield and growth duration of BRRI dhan28 from a multilocation trail conducted in the same year. For calibration and validation, the model predicted results were compared with field observed values. e model performance was analysed with normalized root mean square error (nRMSE), prediction error (P e ), coefficient of determination (R 2 ), and degree of agreement (d).
e equations are given as follows: (1) e coefficient of determination (R 2 ) ranges from 0 to 1 in which values close to 1 indicate a good agreement and values greater than 0.5 are considered acceptable in watershed simulations [32]. In case of nRMSE, a simulation can be considered excellent if nRMSE is smaller than 10%, good if between 10 and 20%, fair if between 20 and 30%, and poor if larger than 30% [33]. e magnitude of P e (%) is expected to be less than 15%, indicating the good performance of the model [34]. e values of d range between 0 and 1, where 0 indicates no agreement and 1 indicates a perfect agreement between predicted and observed data [35].

Simulation of Growth Duration, Crop Water Requirement, Grain Yield, and Water Productivity.
e model simulated grain yield and agronomic parameters for 15 individual study locations for the base year Boro, 2014-15, following normal condition. To do this, multilocation trial crop management files were created using the respective soil and weather files of the stations. Impact of elevated air temperature on yield and phenology was determined using environment modification tools by adding 1°C, 2°C, 3°C, and 4°C with observed maximum and minimum temperatures of 2014-2015 of each study station. e actual water productivity (ET c based) was calculated considering grain yield and crop evapotranspiration using the following equation:  International Journal of Agronomy where water productivity is in kg m −3 , grain yield is in kg ha −1 , and ET c is in mm.

Model Validation.
e CERES-Rice model was validated with simulated and observed data of grain yield and growth duration of multilocation trial during Boro, 2014-15 (Table 6). Excellent model validation was found with nRMSE values of 2.5 and 2.9, R 2 values of 0.67 and 0.8, and P e values of 0.7% and 13%, for grain yield and for growth duration, respectively.

3.3.
Temporal Distribution of Reference Crop Evapotranspiration. Reference crop evapotranspiration (ET o ) is the evapotranspiration under the given meteorological conditions from a reference surface (hypothetical grass reference crop, i.e., green, well-watered grass of uniform height, actively growing and completely shading the ground with an assumed crop height of 0.12 m, a fixed surface resistance of 70 s m −1 , and an albedo of 0.23) [29]. It provides an idea about atmospheric water demand and actual crop water requirement in a given meteorological condition. In Gazipur, ET o showed rising trend from January and reached its peak in the month of May (Figure 2). After May, it tended to decrease over time and reached the minimum during December. is was the combined effect of mean temperature, bright sunshine hours, wind speed, and relative humidity. Increased ET o occurred during March to May due to lower relative humidity and higher mean temperature. Less bright sunshine hours and high humidity caused the reduction in monthly ET o in rainy season (June to October). On the contrary, in the other months (November to February), with low temperature, less wind speed, and reduced sunshine hours due to foggy weather, ET o reached its minimum [36]. Seasonal (January to May, i.e., Boro season) ET o showed the increasing trend over incremental temperature. Seasonal ET o increased by 15, 30, 45, and 58 mm for 1°C, 2°C, 3°C, and 4°C temperature rise, respectively. Figure 3(a) analyses the crop growth duration response to mean seasonal temperature. It shows significant relationship between the two parameters. Growth duration of Boro rice decreased by 6.4 days per 1°C temperature rise. Maniruzzaman et al. [24] showed similar findings of 7.45 days' growth duration reduction of BRRI dhan28 for 1°C temperature rise. Mahmud [37] also reported 7-8 days' growth duration reduction per 1°C temperature rise for BR23 cultivar. Spatial variation showed the highest growth duration of 155 days of BRRI dhan28 in Dinajpur with the lowest seasonal mean temperature of 21.1°C (Figure 4). e lowest growth duration of 131 days was observed in Chottogram, since it had the highest seasonal mean temperature of 23.3°C. e country average growth duration of the tested cultivar was found to be 144 days with the mean temperature of 22°C. Among the locations, the simulated growth duration reduced to 147, 140, 133, and 128 days if temperature increased by 1°C, 2°C, 3°C, and 4°C from the base temperature in Dinajpur. Among the locations, the maximum growth duration reduction was reported in Moulvibazar. e growth duration was decreased by 11, 18, 23, and 30 days for 1°C, 2°C, 3°C, and 4°C temperature rise, respectively. On the other hand, the minimum growth duration reduction was found in Patuakhali. e growth duration was decreased by 5, 10, 15, and 20 days for 1°C, 2°C, 3°C, and 4°C temperature rise, respectively. Rising temperature accelerates the crop growth and development process, resulting in shortening the growth duration of irrigated rice [38]. However, growth duration contributes directly to the yield, but shorter life span due to rising temperature may affect the production in the future [24].

Temperature Effect on Growth Duration and Grain Yield.
Elevated temperature affected directly the grain yield of tested variety BRRI dhan28. A highly significant (R 2 � 0.86, p < 0.000) relation was found between temperature and grain yield. It showed that the grain yield of irrigated rice reduced by 695 kg ha −1 (11.9%) per 1°C seasonal mean temperature rise (Figure 3(b)). Maniruzzaman et al. [24] reported 4.77% yield reduction for BRRI dhan28, whereas Mahmud [37] reported 9.7 to 22.7% decreased yield per 1°C temperature rise. Increase in global mean temperature would reduce the global rice yield by an average of 3.2% [39]. e response of grain yield at normal and elevated air temperature is analysed and presented in Table 7. e highest simulated grain yield at normal temperature observed was 6346 kg ha −1 in Moulvibazar, whereas the lowest grain yield was 4893 kg ha −1 in Chandpur. e dissimilarity of grain yield could be explained by spatial variation of growth duration, as well as climate and soil properties. Since life span becomes shorter with the increasing temperature, it affects directly the grain yield. e highest 17% yield reduction was predicted in Mymensingh and Cumilla and no yield reduction was projected for Rangpur and Dinajpur regions with the 1°C temperature rise. A little yield loss (<5%) was monitored in Gazipur and Moulvibazar region with the same scenarios. In case of 2°C temperature rise, 16 to 35% yield reduction was estimated, of which the highest reduction was found in Cumilla and the lowest in Rangpur region. e yield reduction ranged from 31 to 49% and from 39 to 61% from the normal condition if the seasonal mean temperature increased by 3°C and 4°C, respectively. In another study, Rani and Maragatham [40] showed 13.3% and 23% grain yield reduction for 2°C and 4°C temperature rise International Journal of Agronomy 5 compared to ambient temperature of dry season rice in Tamil Nadu, India.

Crop Water Requirement of Irrigated Rice Affected by
Elevated Temperature. Crop water requirement depends on climatic condition, types of crops, stage of crops, and life span of crop. Since large spatiotemporal variation (Table 8) was observed in Bangladesh, it affected directly the growth duration as well as seasonal crop water requirement of irrigated rice. Country average of 405 mm CWR was observed in Bangladesh during Boro, 2015, and the maximum 445 mm and the minimum 358 mm water demands were estimated in Moulvibazar and Chandpur, respectively. Shahid [7] showed 423 mm to 483 mm potential CWR of rice in north-western region of Bangladesh. Hossain et al. [36] showed an estimation of average 505 mm rice evapotranspiration of BRRI dhan28 during Boro season having a growth duration of 150 days in north-western hydrological region of Bangladesh. Rising temperature showed the decreasing CWR over normal weather condition in all locations of Bangladesh. is was due to the cumulative effect of increasing evapotranspiration and decreasing crop growth duration. e country average of 5% CWR reduced from normal condition, and the biggest change (−7%) was observed in Dinajpur. e minimum deviation (−3%) was recorded in Faridpur, Chandpur, and Gazipur regions when mean temperature increased by 1°C. Similarly, the projected country average CWR reduction ranged from 6 to 12%, from 10 to 16% and from 13 to 23% for 2°C, 3°C, and 4°C elevated temperature, respectively. e maximum CWR reduction was found in Moulvibazar at 2°C temperature rise, whereas Bogura region showed the maximum sensitivity at 3°C and 4°C rising scenarios.
is study identified that CWR was the resultant of seasonal mean temperature and growth duration. e high temperature increased the CWR but reduced the life span of crop, resulting in less water requirement. Seasonal CWR reduced by 14 mm per 1°C temperature rise in Bangladesh      Dinajpur  5579  0  18  31  39  Rangpur  6150  0  16  33  42  Bogura  5330  10  25  35  51  Rajshahi  5837  8  26  35  44  Jashore  6011  16  27  40  51  Satkhira  6191  13  33  41  51  Mymensingh  5843  19  33  42  56  Cumilla  6251  19  35  49  57  Barishal  5886  16  34  41  51  Chottogram  5252  16  34  43  51  Moulvibazar  6346  5  24  39  51  Faridpur  6286  17  30  43  53  Patuakhali  5969  16  33  44  51  Chandpur  4893  16  31  48  61  Gazipur  6118  3  17  33 41     International Journal of Agronomy ( Figure 5). Acharjee et al. [4] also found a similar result and indicated that shorter growth duration due to higher temperature decreased potential crop water requirement of Boro rice. Higher crop water requirement demanded more irrigation water for crop production, since Boro rice grows completely in irrigated condition. Crop fulfils its water demand from either rainfall or irrigation. But Boro rice is typically in winter (December-January) and the crop lasts until premonsoon (April-May) season in Bangladesh. Dry season rice has more irrigation demand than the rice grown in other seasons. Reduced CWR will necessitate less irrigation requirement. Acharjee et al. [4] estimated 6.5-10.9% reduced potential crop water requirement during Boro season for RCP 4.5 and 8.5, respectively, for 2050s. ey also reported 8.3% and 17.6% decreased crop water requirement by 2080s for RCP 4.5 and 8.5, respectively. is estimated reduction in CWR will reduce the irrigation cost as well as total cost of rice production. It also decreases fuel consumption used in irrigation pumps, which will minimize greenhouse gas emission to the environment. In Bangladesh, 76% of irrigation water comes from groundwater source [41]. us, minimized CWR will reduce groundwater abstraction.

Rice Water Productivity Affected by Elevated Temperature.
Generally, actual water productivity (WP ET ) decreased with the increase in mean air temperature ( Figure 6). is was the result of reduced grain yield and ET c . ough both ET c and grain yield were reducing in elevated temperature, decreased grain yield played the driving role of less WP ET in all the study locations. However, WP ET of Dinajpur, Rangpur, and Moulvibazar increased in 1°C temperature rise condition compared to the normal condition. is happened because the grain yield in 1°C temperature rise was similar to normal temperature. Dinajpur, Rangpur, and Moulvibazar are situated in cold tempered region. Due to this, the growth duration of the tested cultivar in 1°C temperature rise remained above the average growth duration (145 days).

Carbon Dioxide (CO 2 ) Concentration Effect on Yield.
Mean surface temperature of Bangladesh is projected to rise by 2.0°C to 4.0°C according to Representative Concentrate Pathways (RCP) by the end of 2100 compared to the base years of 1986-2005 [2]. Carbon dioxide concentration will be increased as well [38]. e current study showed that the elevated temperature decreased grain yield and growth duration of the cultivar. However, increase in CO 2 concentration compensates the grain yield lost by temperature rise [24]. Increased air CO 2 level enhances photosynthesis and reduced respiration processes of crop and improves crop water use efficiency. Hence, the process contributes to crop yield [42,43].

Conclusion
Elevated air temperature affected life span and grain yield of irrigated crop, but it reduced the seasonal crop water requirement that results in less irrigation demand for dry season rice cultivation. Elevated temperature shortened the crop growth duration and reduced the crop water requirement. In response, the reduced crop water requirement will decrease the irrigation demand and, consequently, the number of irrigation days. e combined effect of less irrigation demand and irrigation days will save huge amount of irrigation water currently required for dry season Boro rice cultivation in Bangladesh. However, the reduction in crop growing period due to temperature rise will cause substantial yield reduction of current rice cultivars. It would be a near-future challenge for crop developers to introduce a new type of cultivar which can tolerate the effect of elevated temperature with no significant yield loss. At present, Bangladesh is the 4th rice-growing country around the world and it contributes 7% of total rice production in the world [44]. Although many countries grow rice, Bangladesh is the most vulnerable country to climate change, especially temperature rise. Besides, water availability or scarcity is one of the crucial factors for dry season rice cultivation. us, in the present situation, there is dire need to investigate in detail the future rice water demand along with the rising temperature. is will help to plan irrigation strategies with limited water resources.

Data Availability
All data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.