Simulation of future global warming scenarios in rice paddies with an open-field warming facility
- 8.5k Downloads
To simulate expected future global warming, hexagonal arrays of infrared heaters have previously been used to warm open-field canopies of upland crops such as wheat. Through the use of concrete-anchored posts, improved software, overhead wires, extensive grounding, and monitoring with a thermal camera, the technology was safely and reliably extended to paddy rice fields. The system maintained canopy temperature increases within 0.5°C of daytime and nighttime set-point differences of 1.3 and 2.7°C 67% of the time.
KeywordsEcosystem warming climate change canopy temperature global change infrared heating plant-climate interactions rice Yangtze River valley
List of Abbreviations
Free-air Temperature Enhancement
High daytime temperature
High daytime and nighttime temperature
High nighttime temperature
High temperature with 12 heaters
Intergovernmental Panel on Climate Change
(IRH-warmed Tc) - (Reference Tc)
- ΔTc D
(Daytime IRH-warmed Tc) - (Reference Tc)
- ΔTc N
(Nighttime IRH-warmed Tc) - (Reference Tc)
- ΔTc DN
(Daytime and nighttime IRH-warmed Tc) - (Reference Tc).
Rice (Oryza sativa L.) is a staple food for more than 3 billion humans, mainly in Asia. Its demand will increase because the world's population is expected to reach 9.1 billion by year 2050, including 5.5 billion in Asian countries . Despite a high degree of yield increase, mainly due to Green Revolution technologies, average rice yields show a high magnitude of instability depending on weather and monsoon anomalies . Climate change in addition to high population growth is pressuring the thin margin between supply and demand of rice . Most of our insights regarding impacts of climate change consider observed and predicted changes in temperature over the next century . By the end of this century, global surface air temperature is expected to have increased by 1.1-6.4°C relative to the average temperature during 1980-1999. The best estimate for the expected air temperature increase for Intergovernmental Panel on Climate Change (IPCC) low population growth scenario (B1) is 1.8°C with likely range of 1.1°C to 2.9°C, while the best estimate for a high growth scenario (A2) is 3.4°C with likely range of 2.0 to 5.4°C .
There is substantial spatiotemporal, seasonal, and inter-annual variability in the warming trend. A faster increase in nighttime temperature than daytime temperature is reported for Jiangsu province . Most of the studies conducted to investigate the effect of temperature on different aspects of plants are based on diel mean air temperature, assuming there is no differential influence of nighttime and daytime temperatures . However, the negative impact of high nighttime temperature on rice production is greater than that of daytime or daily mean temperatures . Peng et al., found a 10% decrease in rice grain yield for each 1°C rise in nighttime minimum temperature, while yield was not significantly affected by a rise in daytime maximum temperature . Recently, a trend of increasing differential between daytime and nighttime temperatures has been observed in the literature, with more focus on higher nighttime temperature [9, 10, 11, 12].
Substantial efforts have been made to simulate the effects of predicted global warming on rice through a variety of closed or partially open warming facilities based on chambers of various designs. However, there are several chamber artifacts which can alter experimental and/or environmental parameters for plants grown in such "closed or partially closed" or controlled environment facilities [13, 14, 15]. These parameters are either impossible or difficult to control, depending upon their own gradient and interaction with each others. They include light quantity and quality, wind speed, relative humidity, long-wave radiation, evapotranspiration rates, and CO2 concentration. Often growth mediums (soil nutrient status, pot/container size) provided in the controlled environment systems rarely match with field conditions . Consequently abnormal root growth, development, and functioning may result, and they can alter plant size and morphology as well compared to those grown in the field . Open-top chambers (OTCs) have been frequently used to study the effects of different temperature regimes on rice [16, 17, 18]. The maximum temperature differences using this technique occur during full sunlight conditions. Although OTCs can heat both air and soil, sunny days are necessary for the heating, and the majority of warming is achieved only during daytime .
A promising alternative to warming with chambers and their associated artifacts is the use of infrared heaters (IRH) over open-field plots. The use of IRH to study the response of ecosystems to global warming was started in mid 1990s [20, 21, 22, 23]. The temperature rise of a rice canopy through IRH warming is essentially the same as the warming provided by radiant heating from the sun and sky because it directly heats the canopy. The air in and above the canopy is subsequently warmed by convective sensible heat exchange with the canopy (and cooled by latent heat exchange). If the "constant temperature rise" mode of operation is used, as we did herein, the warming by IRH can be directly related to degree of canopy warming expected through global warming [13, 24, 25].
The amount of energy required to achieve a specified increase in canopy temperature by IRH is influenced by canopy conductance in response to soil moisture conditions, light intensity, temperature, humidity, and wind speed . Less energy is required when the stomates are closed, such as occurs under water stress or at night. However, with warmer leaves, higher vapor pressure occurs in the sub-stomatal cavities in the infrared-warmed canopies, which can create unrealistic vapor pressure gradients between the inside of the leaves and air [25, 26]. This problem can be minimized through the introduction of supplemental irrigation for upland crops [13, 24, 26]. However, it is not required for paddy fields, which are already grown under flooded conditions, so depletion of soil moisture was not an issue in the results reported herein.
Selection of the infrared heaters IRH is important so that thermal radiant properties are suitable for the plant warming application i.e., no significant radiation less than 850 nm and high emissivity . Since most of the commercially available IRH are manufactured for indoor use only, measures are required to make them water-proof and safe for reliable outdoor use. In addition, the heaters need to be deployed in configurations that provide uniform distribution of the warming across the field plots. Kimball et al. (2008) arranged infrared heaters in hexagonal fashion and achieved highly repeatable and uniform distribution of warming on the canopy of wheat and Sudan grass . This updated model of IRH had few ecosystem limitations and little disturbance in ecosystem . Rice has additional challenges to adopt such a technique, as it is grown in standing water, and numerous high capacity electrical cables are hazards.
A facility is required to provide warming of field plots to fully assess the likely effects of different scenarios of global warming on future rice production. Infrared heater arrays with the modifications introduced in this study can safely produce open-field warming of rice canopy temperatures like those expected with climate change scenarios in the paddy fields. Our goal was to construct such an open-field warming facility for paddy fields like those in the Lower Reaches of the Yangtze River Valley (the oldest niche of rice production) and elsewhere to simulate differential daytime and nighttime warming under low (B1) and high (A2) population growth scenarios. Danyang FATE (Free-air Temperature Enhancement) facility will help us to predict the fate of rice production in its oldest niche, i.e., the Yangtze River valley.
Result and discussion
Infrared Heater (IRH) description
Infrared Heater Array
Three steel posts, each of 3 m length, were inserted vertically into the paddy soil to a depth of 0.5 m, making an equilateral triangle of each side 450 cm (Figure 1B, G). A suspension cable (1/8 inch steel cable) was fixed around the posts (Figure 1D, F). Each post was supported by another diagonal steel cable, tied to a 120 cm long stake in the soil (Figure 1C, E, F). At grain filling stage, six IRHs were hung from the suspension cable at a height of 120 cm above the top of the canopy (0.8 times the radius of array), covering a 3-m diameter circular soil area (7.1 m2) to from a hexagonal FATE array (Figure 1G). IRHs were tilted at 45° from horizontal and ± 30° from the cable axes to provide uniform distribution of infrared radiation in the plot area (Figure 1D). Each array was equipped with six IRHs (total capacity 6000 W), infrared thermometer (IRT), dimmer, and an electrical distribution unit protected in a weather resistant enclosure (Model ENC12/14, Campbell Sci. Inc., Logan, UT, USA).
A "dummy" array of same size, shape and structure, except with dummy heaters (with the same color of reflecting housing), was also erected (Figure 1A). Both IRH and dummy arrays were deployed in the center of square plots (6 m × 6 m), allowing a buffer strip of at least 6 m at their perimeter. The shading of six IRH from nadir was about 2%. However, because the heaters were deployed around the perimeter of the circular plots, only half of the IRH would shade the plots at any specific time, so the amount of shading over the plots was about 1% .
Using a PID controller subroutine obtained from Campbell Scientific (Campbell Sci., Inc., Logan, UT, USA), Kimball (2005) developed a program to control infrared heaters using the components listed in the previous paragraph . This program was modified to suit the configuration of our particular heating system and warming treatments (for Exp. I, II & III). Scaling of the PID signals was introduced for smoother control of the IRHs so that there was less oscillation and acting like an ON/OFF controller compared to the original program. A car battery with charger was used as a source of 12 V power for the datalogger (Figure 2D, F). The set point differences between the heated and the corresponding reference plots were 1.3°C during daytime and 2.7°C during nighttime to achieve warming averages resembling the lower and upper limits of the B1 scenario of global warming predicted by IPCC (Table SPM.3, page 13) .
Construction of Danyang-FATE facility
T-shaped steel hangers were attached to the top of a vertical post of each array to carry electrical cables and avoid any contact of the cables with irrigation water (Figure 4E). Electrical distribution units with control switches for each array were secured in weather resistant enclosures (ENC 12/14, from Campbell Sci. Inc. UT, USA) and attached to one of the posts in each array (Figure 2I). Electrical power from the control room was connected to the electrical supply units for each IRH array, which then divided and supplied electricity to all six heaters of each array (Figure 2).
Warming treatments were imposed between August 31 and October 14 (2009) and September 2 and October 13 (2010). The set point differences between the heated and corresponding reference plots were the same as described in Exp-I (1.3°C during daytime and 2.7°C during nighttime) using the same equipment as described for Exp-I, but with modifications as described above. However, two additional treatments, HNT (High Nighttime Temperature, + 2.7°C CK) and HDT (High Daytime Temperature, +1.3°C CK) were applied in addition to HDNT (High Daytime and Nighttime Temperature, +1.3°C/+2.7°C CK) using the same equipment as described for Exp-I. Each treatment was applied in triplicate (four treatments, three replications, 12 arrays) and averages from two experiment years (2009 and 2010) are used in this paper.
Strict safety measures were taken to avoid electric hazards at the FATE site (Exp-II and III). All electrical equipment in the control room and the IRHs (each array separately) were connected using 2-mm wire to copper-clad steel (with 1inc diameter and 1.23-m length) grounding rods driven into the soil. In each array, IRHs were tightly fastened to the supporting posts to provide additional grounding. All points of potential electric hazards were regularly checked manually, and thermal images of all electric supply and control units, datalogging equipment, and dimmers were captured using a Thermacam (Figure 8). Differences in surface temperatures of electric cables and dimmers for working and non-working conditions are clear from the thermographs. Burning of an electric cable occurred in 2009 (Figure 8A, B) due to over-heating. All respective cables were immediately replaced with reinforced insulation to avoid this problem again, and at the end of the 2009 experiment, all cables (4-mm thick, braided, 25 A) were replaced with cables of higher electric capacity (8-mm thick, solid). In addition to the T hangers on a vertical post of each array, four posts with T-shaped hangers were anchored in the buffer strip with concrete. Electrical cables from these additional hangers were connected to the electrical supply unit of each array.
Data acquisition and thermal images
Like Exp. I and II, temperature signals from the IRTs were measured by the datalogger, and 15-min averages were stored in its database and output for Exp-III as well. The corrections in IRT canopy temperatures (from the heated plots) that were made to adjust for radiation from the IRHs that was reflected from vegetation  were doubled to account for the doubled number of heaters. Based on sensed temperatures, signals were transmitted through current/voltage output module (SDM-CV04) to the dimmers, which ultimately regulated the IRHs through regulating power supply to each array, according to the target set-point and treatment to ensure stable warming. For all three experiments described here, daily sunrise and sunset timings were automatically determined for each day by the datalogger through latitude and longitude of experiment site. Stored data was acquired from the datalogger using Loggernet 3 (Campbell Sci. Inc. UT, USA) regularly (either twice or thrice a week) during the experiments. Diel, daytime, nighttime hourly averages of reference and heated plots, and respective temperature differences (ΔT) were calculated from 15-minute averages. Canopy thermal images were taken using a ThermaCAM camera (Model P25, FLIR systems, Boston, USA), as described in our earlier experiments . In 2010, a weather station (WatchDog Model 2700, Spectrum Technologies, East-Plainfield, IL, USA) was installed at FATE site (about 200 m from the arrays) to obtain ancillary weather data, including solar radiation, air temperature, wind speed and direction at 2 m above the soil, rainfall, and relative humidity.
Performance of infrared heating arrays
Simulation of B1 Scenario
Simple correlation analysis showed that wind speed (U, m s-1) had a significant influence on the diurnal cycle of canopy warming (ΔTc) with a correlation coefficient of -0.93 (HDNT). Temporal variability of IRHs and other systems under windy conditions was consistent with previously reported studies [13, 26, 27, 30]. Decreased effectiveness of IRH under high wind velocity reported for rice, under controlled greenhouse conditions cannot perfectly resemble the field conditions, where wind speed greatly varies and has a diurnal pattern . Moreover, Mohammad and Tarpley (2009) had larger IRHs to increase only nighttime temperature, when field conditions generally experience lower wind speed . As expected, maximum warming was achieved during nighttime, when conditions are generally calm and stomata are closed. The diurnal warming patterns that can be achieved with infrared warming systems like presented herein and their relation with wind speed has previously been reported [13, 27]. Generally, during some of the time, high wind speeds prevent target levels of warming to be attained. Using larger heaters would increase the percentage of time that target warming can be achieved, but of course electrical power requirements and operating expense would increase. Power requirements could be reduced by using heaters with a larger characteristic dimension, which imparts an increased "radiometric" efficiency [26, 31]. Judicious arrangement of heaters in a honeycomb pattern could improve "geometric" efficiency, especially at larger plot scales .
Simulation of A2 Scenario
Similar to Exp-II, warming effects of IRH on canopy temperature (Tc) in Exp-III was uniformly distributed (Figure 11C, F). Greater warming was achieved during nighttime, which is similar to the historical trend observed for the Jiangsu province, where the FATE site is located (Figure 12D, H) . Daytime and nighttime warming averages were 2.81°C and 5.40°C respectively, with a diel average of 4.12°C. This attained warming closely resembles that expected in the A2 scenario (2.0-5.4°C). The diurnal cycle of ΔTc (canopy temperature between heated and reference plots) was negatively correlated with wind speed (path coefficient of -0.65) (Figure 12H). Daytime temperature differentials were within 0.5°C of the daytime 2.7°C target 68% of the time and nighttime 5.7°C target temperature 55% of the time (Figure 13B). As with Exp-II, some of the variability near dawn and dusk was due to the transition periods between daytime and nighttime warming, which is in agreement with earlier reports [13, 29].
This paper describes the first ever deployment of infrared heaters over rice in open paddy fields in order to simulate the effects of global warming scenarios on rice canopy temperatures. Hexagonal arrays of the infrared heaters in combination with infrared thermometers, dimmers, current/voltage output modules and an automatic control system fulfilled the requirements for an appropriate ecosystem warming system for paddy fields. The system safely provided uniform controlled warming over the plots, and it provided reproducible results under natural open-field conditions of wind and light. Safety features were added including sealing and water-proofing the heater assemblies, setting the support posts in concrete, proper grounding of all electrical equipment, and stringing all electrical cables via overhead supports. This warming system can be used to conduct ecosystem warming experiments for evaluating plant responses to different daytime and/or nighttime temperature increases as predicted by low and high greenhouse-gas emission scenarios. The system was able to track the set-point temperature differences adequately with about 67% of the observations falling within ± 0.5°C of target set-points. However, the efficiency and effectiveness of the IRH heating system was reduced at high wind speed. Using larger and/or more efficient heaters would improve the performance percentage, but electrical power costs would increase. At nighttime, when wind speeds were lower and when stomata were closed, greater warming occurred compared to daytime. These qualities of the FATE facility makes it suitable to simulate global warming scenarios and will be helpful to predict vulnerability of rice specifically in the Yangtze River valley and in general throughout the world.
The present research was part of a project to construct a FATE facility to study the response of a rice-wheat cropping system to expected global warming in the Lower Reaches of the Yangtze River Valley, China. Three experiments were conducted during the summers of 2008-2010 at Jiangning Experimental Station of Nanjing Agricultural University, Nanjing (118°30 E, 31°50 N) and Danyang FATE facility, Danyang, Jiangsu, China (119°27' E, 31°54 N). Site details are described in our previous studies [28, 32]. Experiment-I (Exp-I), was conducted in 2008 at Jiangning experimental station to assess the feasibility of modified IRH arrays, similar to those designed by Kimball et al., (2008), for paddy rice fields . Experiment-II (Exp-II) and Experiment III (Exp-III) were conducted at Danyang FATE facility to simulate B1 and A2 emission scenarios. Exp-II was conducted for two years (summer 2009 and 2010), while Exp-III (with 12 heaters per array rather than the original 6) was conducted for one year (summer 2010).
In Exp-I, japonica rice (Oryza sativa L. cv. Zhendao88) was grown in the paddy fields, and for Exp-II and Exp-III, two indica hybrids, Shanyou63 (heat resistant) and Teyou559 (heat susceptible) were used. Sensitivity of these cultivars to high temperature has already been reported and tested in our previous experiments [28, 33, 34]. These cultivars have similar growth behavior and stature. Each array was divided into two equal parts (north to south), and cultivars were randomly assigned to one half or the other of the arrays. One-month-old seedlings were transplanted in the last week of June each year. Local recommendations for fertilizer and irrigation were followed. Warming treatments were applied at the grain filling stage (Exp-I & II, from August 31 through October 14 in 2009 and September 2 through October 13 in 2010). For Exp-III, they were applied during reproductive (August 12 through September 1, 2010) and grain filling (September 2 through October 13 in 2010) stages.
Wind speed (U) data derived from the weather station were also averaged to get daytime, nighttime, and diel means. Canopy temperatures (Tc) from Exp-II and Exp-III were averaged to determine mean daytime (06:00-18:00), nighttime (18:00-06:00), and diel temperatures. Data for reference and IRH-warmed plots were then expressed as deviations from ambient control (ΔTc). Data were averaged across all days within the experimental periods to determine mean deviation from control plot (ΔTc) for daytime (ΔTc D) and nighttime (ΔTc N). To investigate the dependence of achieved canopy warming on wind speed, we determined simple correlations of ΔTc with wind speed at different temporal scales.
The authors thank the Higher Education Commission (HEC) of Pakistan for funding a scholarship to MIAR during his studies for the Ph.D. degree. This work was funded in part by grants from Special fund for Agro-scientific Research in the Public Interest: Impact of climate change on agricultural production of China, (Project No. 200903003), The Fundamental Research Funds for Central Universities of China (Project No. KYT201005) and The National Key Technology Support Program of China (Project No. 2011BAD16B14). We would also like to thank Mr. Yongjun Yao and his team (Beijing Truwel Instruments, Beijing, China) for technical assistance for modifications in the initial site design; Dr. Waseem Raza, Muhammad Ammar Khan (Nanjing Agricultural University, Nanjing, PR China) and Dr. Zulfiqar Ali (University of Agriculture, Faisalabad, Pakistan & Jiangsu Academy of Agricultural Sciences, Nanjing, PR China) for revising an early draft of this manuscript; Guangbin Wei, Rashid Mehmood Rana, Hiufeng Ning, Jiangfang Qiao and Ding Chengjiang (Nanjing Agricultural University, Nanjing, PR China) for their help during construction of the FATE facility. The authors greatly appreciate the consistent logistic and technical help from the staff and officials at Jiangning Experimental station, Nanjing and Danyang FATE site, and Danyang Agricultural and Forestry Bureau.
- 2.Krishnan P, Ramakrishnan B, Reddy KR, Reddy VR: High-Temperature Effects on Rice Growth, Yield, and Grain Quality. Advances in Agronomy. Edited by: Donald LS. 2011, Academic Press, 111: 87-206.Google Scholar
- 3.Fischer KS: Challenges for rice reseach in Asia. Sustainability of rice in the global food system. Edited by: Dowling NG, Greenfield SM, Fischer KS. 1998, International Rice Research Institute, 95-98.Google Scholar
- 5.IPCC: Summary for Policymakers. In Climate Change 2007. The Physical Science Basis Contribution of Working Group I to the Fourth Assessment Report of the Intergovernmental Panel on Climate Change. Edited by: Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, Tignor M, Miller HL. 2007, Cambridge University Press, Cambridge, United Kingdom and New York, NY, USAGoogle Scholar
- 6.Yuan CH, Tang JP: Regional Climatic response of Jiangsu to global warming. J Nanjing Univ. 2007, 43: 655-69. (In Chinese with English abstract)Google Scholar
- 10.Prieto P, Penuelas J, Llusia J, Asensio D, Estiarte M: Effects of long-term experimental night-time warming and drought on photosynthesis, Fv/Fm and stomatal conductance in the dominant species of a Mediterranean shrubland. Acta Physiologiae Plantarum. 2009, 31: 729-39. 10.1007/s11738-009-0285-4.CrossRefGoogle Scholar
- 14.Hall AE: 2001, Crop responses to environment: CRC PressGoogle Scholar
- 15.Kimball B, Pinter P, Wall GW, Garcia R, LaMorte R, Jak PMC, Frumau KFA, Vugts HF: Comparisons of responses of vegetation to elevated carbon dioxide in free-air and open-top chamber facilities. Advances in Carbon Dioxide Effects Research ASA Special Publication 61. Edited by: Allen Jr. LH, Kirkham MB, Olszyk DM, Whitman CE. 1997, American Society of Agronomy, Crop Science Society of America, Soil Science Society of America Madison, WI, 61: 113-30.Google Scholar
- 26.Kimball B: Theory and performance of an infrared heater for ecosystem warming. Global Change Biol. 2005, 11: 2041-56.Google Scholar
- 31.Kimball BA, Conley MM, Lewin KF: Performance and energy costs associated with scaling infrared heater arrays for warming field plots from 1 to 100 m. Theor Appl Climatol.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.