不同灌溉方法对冬小麦农田小环境及根系分布的影响

Journal of Intergrative Agriculture Advance Online Publication 214 Doi: 1.116/S295-3119(14)6927-8 Effects of Different Irrigation Methods on Micro-environments and Root Distribution in Winter Wheat Fields LV Guo-hua 1, Song Ji-qing 1, Bai Wen-bo 1, Wu Yong-feng 1, Liu Yuan 1, and Kang Yao-hu 2 1 State Key Engineering Laboratory of Crops Efficient Water Use and Drought Mitigation, Institute of Environment and Sustainable Development in Agriculture, Chinese Academy of Agricultural Sciences, Beijing, 181, P.R.China 2 Key Laboratory of Water Cycle and Related Land Surface Processes, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Beijing 111, P.R.China Abstract The irrigation method used in winter wheat fields affects micro-environment factors, such as relative humidity (RH) within canopy, soil temperature, topsoil bulk density, soil matric potential, and soil nutrients, and these changes may affect plant root growth. An experiment was carried out to explore the effects of irrigation method on micro-environments and root distribution in a winter wheat field in the 27 28 and 28 29 growing seasons. The results showed that border irrigation (), sprinkler irrigation (), and surface drip irrigation () had no significant effects on soil temperature. Topsoil bulk density, RH within the canopy, soil available N distribution, and soil matric potential were significantly affected by the three treatments. The change in soil matric potential was the key reason for the altered root profile distribution patterns. Additionally, more fine roots were produced in the treatment when soil water content was low and topsoil bulk density was high. Root growth was most stimulated in the top soil layers and inhibited in the deep layers in the treatment, followed by and. This was because of the different water application frequencies. As a result, the root profile distribution differed, depending on the irrigation method used. The root distribution pattern changes could be described by the power level variation in the exponential function. A good knowledge of root distribution patterns is important when attempting to model water and nutrient movements and when studying soil plant interactions. Key words: border irrigation; root profile distribution; sprinkler irrigation; surface drip irrigation; field microenvironment; winter wheat 不同灌溉方法对冬小麦农田小环境及根系分布的影响 吕国华 1, 宋吉青 1, 白文波 1, 武永峰 1, 刘园 1 2, 康跃虎 1 中国农业科学院农业环境与可持续发展研究所, 作物高效用水与抗灾减损国家工程实验室, 中国北京, 181 2 中国科学院地理科学与资源研究所, 陆地水循环及地表过程重点实验室, 中国北京,111 LV Guo-hua, Tel: +86-1-82165, E-mail: lvguohua@caas.cn Correspondence Kang Yao-hu, Tel: +86-1-82165, E-mail: kangyh@igsnrr.ac.cn

摘要 : 灌溉方法能够影响农田小环境, 如冠层内相对湿度 土壤温度 表层土壤容重 土壤基质势 土壤养分等, 这些因素都能影响根系生长 为了研究灌溉方法对农田小环境及根系分布的影响, 分别在 27-28 和 28-29 年冬小麦生长季开展了田间试验研究 研究结果表明地面灌溉 喷灌和地表滴灌对土壤温度没有显著影响 ; 但是, 对表层土壤容重 冠层内相对湿度 土壤有效氮分布和土壤基质势影响显著 土壤基质势的变化是导致根系分布发生变化的主要原因 而且, 由于表层容重高 土壤含水量低, 地面灌溉处理产生了更多的细根 ; 受灌溉频率的影响, 表层根系生长受到促进, 而底层根系生长受到抑制 ; 其中, 滴灌处理最为显著, 其次为喷灌和地面灌溉 随着灌溉方法的改变, 剖面根系分布也发生了变化, 这种变化可以用指数函数的幂的变化来表征 在研究土壤 - 作物体系中, 了解根系分布的变化对于模拟土壤水分和养分运动具有重要的意义 关键词 地面灌溉 ; 根系剖面分布 ; 喷灌 ; 地表滴灌 ; 农田小环境 ; 冬小麦 1. INTRODUCTION Each plant species has its own root growth characteristics, which can be substantially modified by the plant s environment (Bathke et al. 1992). The physical, chemical, and microbiological conditions throughout the soil profile, together with the plant microclimate, should be considered when investigating plant root growth characteristics. Soil physical properties, such as particle size distribution, bulk density, soil aeration, soil water potential, soil strength or penetration resistance, and temperature, can all affect root growth temporally and spatially (Merrill and Rawling 1979; Unger and Kaspar 1994). Zimmerman and Kardos (1961) researched the effect of bulk density on root growth. Their results showed that the greater the soil bulk density, the lower the root weight. Soil water content or water potential also had considerable effects on root growth. Adventitious roots formed, but did not elongate when the soil water content around the crown node was low. However, adventitious root growth appeared normal when the water potential was at or above 15 bars (Ferguson and Boatwright 1968). High levels of available water can lead to relatively low below-ground biomass (Fabião et al. 1995) and

reduced root initiation as the depth increases (Torreano and Morris 1998) because adequate water can be accessed by surface roots. In contrast, water deficit stress resulted in a greater proportion of the roots growing deeper into the soil (Bai and Li 23; Benjamin and Nielsen 26). Soil temperature affects the growth of root system components, initiation and branching, orientation and direction of growth, and root turnover. Kaspar and Bland (1992) showed that as the temperature increased, roots grew faster and reached a maximum growth rate at about 3 C for maize (Zea mays) and pecan (Carya illinoinensis), after which the rate began to decrease. Kar et al. (1976) investigated the effects of variations in temperature regime on rice root growth in association with other soil physical properties. The results indicated that high temperature increased root degeneration, but the high bulk density of the sandy loam soil decreased root degeneration. In addition, soil chemical properties also affect root development. Nielsen et al. (196) found that phosphorus fertilizer could increase the root growth of oats (Avena sativa). Nitrogen fertilization has been found to increase root weights at all soil moisture levels (Kmoch et al. 1957). However, Jackson and Bloom (199) suggested that tomato root systems showed lower responses to fertilizer N because of changes in the root distribution pattern. The influence of atmospheric humidity on root growth has also been studied (Breazeale and Mcgeorge 1953) and the results showed that root elongation significantly increased when the atmosphere changed from arid to humid. The results also indicated that sprinkler irrigation, which simulates a rain, may stimulate more root growth than conventional methods of irrigation in semiarid districts. In irrigated fields, each irrigation method used has different effects on the microclimate and soil environment. Tolk et al. (1995), Liu and Kang (26) and Cavero et al. (29) showed that the air temperature and vapor pressure deficit were lower in a sprinkler irrigated field than in a surface irrigated field. Other studies showed that the topsoil in sprinkler irrigated fields was looser than in surface irrigated fields (Sun 26). Soil water dynamics are also differentially affected by the type of irrigation method used. Frequent irrigations by micro-irrigation can help to maintain higher average soil water contents than conventional methods when using the same amount of water (Rawlins and Roats 1975). Furthermore, Wang et al. (1997) and Home et al. (22) found that surface irrigation appeared to leach chemicals more rapidly than both drip and sprinkler irrigation, which resulted in different nutrient distribution patterns. Therefore, the type of irrigation method used affects a plant s environment, which would lead to variations in root growth and profile distribution. Root water uptake patterns under traditional border irrigation () (a kind of surface irrigation), sprinkler irrigation (), and surface drip irrigation () have been investigated previously (Lv et al. 21). In this study, we investigated soil and micro-environmental characteristics under winter wheat, such as relative humidity within the canopy, soil temperature, topsoil bulk density, soil matric potential, and soil available nutrients, and their effect on root distributions when subjected to three different irrigation methods. In order to make accurate comparisons, the same amount of

water was applied under each irrigation treatment. The objectives of this study were to research the micro-environments and root distribution pattern changes for winter wheat when treated with the different irrigation methods. A good knowledge of how the root profile distribution is affected by irrigation methods is very important when attempting to model how water and nutrient dynamics are influenced by irrigation. This information should lead to improvements in irrigation scheduling. 2. RESULTS AND DISCUSON 2.1 Precipitation and irrigation Precipitation values from the jointing stage (174 DAS [days after sowing]) to harvesting (249 DAS) in 28 and 29 are shown in Fig. 1. In 28, total precipitation was 12.3 mm, there were 21 rainfall events, and precipitation was relative uniformly distributed. In 29, total precipitation was 98.4 mm, there were 13 rainfall events, but precipitation was scarce from the flowering stage (24 DAS) to the grain-filling stage (239 DAS). In addition, from 174 to 249 DAS, the average air temperature in 29 was 19. C, which was higher than in 28 (17.7 C ) (Fig. 1); and the average air relative humidity (RH) in 29 was 59.9%, which was lower than in 28 (71.3%). Therefore, it was wet and rainy in 28, compared to 29. The first irrigation in each treatment was carried out at the jointing stage in 28 (177 DAS) and 29 (174 DAS). The number of water applications varied with irrigation method and growing season. There were two and three irrigations for (Fig. 2-a), three and five for (Fig. 2-b), and 22 and 43 for in 28 (Fig. 2-c) and 29 (Fig. 2-d), respectively. The total irrigation volume in 29 was 242.7 mm, which was nearly twice the volume (121.3 mm) applied in 28. This was mainly due to the long rainfall deficit period and the relatively dry weather in 29. Irrigation was terminated midway through the grain-filling stage in both 28 and 29. 2.2 Bulk density of the topsoil Before sowing the winter wheat, the experimental field was ploughed with a rotary cultivator to a depth of 2 cm. The topsoil was loose and bulk density was about 1.1 g cm 3. However, after a long period of natural soil sedimentation in the winter, the bulk density in the topsoil had increased to about 1.2 g cm 3 by the turn-green stage. The volume of water applied and the water application method should affect the structure of topsoil (Sun, 26). At harvest, the topsoil bulk density sample means for each treatment were 1.35 g cm 3 for, 1.26 g cm 3 for, and 1.22 g cm 3 for. The higher bulk density in the topsoil could affect root growth and lead to the production of more fine roots. In the treatment, the water ponding depth above the soil surface would produce a variable pressure head, which could lead to a higher bulk density. Previous research that investigated the system found that the water droplets produced by the sprinklers broke down aggregates and compacted the thin

surface layers (Ragab 1983; Tarchitzky et al. 1984; Adeoye 1986). However, in this experiment, the leaf area index of the winter wheat was larger than 2 m 2 m 2, which helped reduce soil compaction. Furthermore, in the treatment, the drip emitters were close to the soil surface, which meant that there was little damage to the topsoil structure. 2.3 Soil matric potential Fig. 3 shows the soil matric potential (SMP) values at 2, 4, 6, and 9 cm depths throughout the 28 experimental period. Irrigation was carried out from 177 DAS and terminated at 223 DAS. During the irrigation period, the SMP values for the and treatments at all depths were almost 33 cm, which showed that there was sufficient soil water for winter wheat growth. Additionally, the pan evaporation results showed that the irrigation volume could be reduced to increase the water use efficiency. However, further research is needed to optimize the irrigation scheduling. The treatment SMP values were the lowest of the three treatments at every depth from 24 to 224 DAS. This was caused by long term irrigation and rainfall deficits. The lowest SMP top soil values in the treatment may stimulate root growth in the deep layers. After irrigation ceased, the SMP at 2 and 4 cm depth decreased more sharply in the treatment than in the or treatments (Figs. 3-a and 3-b). The lowest value was found in the treatment at 6 cm depth. The SMP values at 2, 4, 6, and 9 cm depths in 29 are shown in Fig. 4. Irrigation began at 174 DAS and terminated at 228 DAS. During the irrigation period, the SMP values at each depth in the treatment were relatively constant and were higher than 3 cm. They were also the highest SMP values of all three treatments for most of the experimental period. In the and treatments, SMP fluctuated with water application, particularly at 2, 4, and 6 cm depths. The irrigation frequency was higher in the treatment, so the SMP was higher than in the treatment, even at 9 cm depth, After irrigation ceased, the SMP trend was similar to 28. Although the same amount of water was applied in each treatment, the higher application frequency in the treatment produced higher SMPs throughout the soil profile during the experimental period. However, when irrigation stopped, the treatment with the highest irrigation frequency () showed the fastest SMP decrease in the top soil. These changes in SMP could greatly affect root growth and profile distribution. 2.4 Soil temperature As shown in Fig. 5, the high density of contour lines for the 4 cm depth shows that the soil temperature change was considerable in this layer, changing from.8 C to 29.2 C in 28, and from.1 C to 31.9 C in 29. At 4 1 cm depth, the soil temperature change was relatively small, being 8 C to 16 C in 28 and from 1 C to 2 C in 29. In 28, the soil temperatures at all depths in each treatment showed no obvious differences. From 17 to 18 DAS, the measured average soil temperature in the deep soil was only 8 C,

which was due to the cold weather in the winter. The topsoil temperatures rose quickly when spring began and air temperatures started to rise. In the deep layers, the slow rise in temperature was due to the high winter wheat LAI, which blocked direct solar radiation. In addition, water movement played a key role in the transfer and diffusion of heat. The heat in the topsoil could be transferred to the deep layers by moving water, and the heat in the irrigation water (the ground water temperature ranged from 14 C to 16 C ) may also contribute to the temperature rise in the deep soil. In 29, the average soil temperature at the soil surface for the treatment was 19.5 C, followed by 18.6 C for the and 16.9 C for the treatments. The low soil water content in the treatment led to the lowest soil thermal capacity and the highest temperature values. However, there were no obvious differences among the three treatments at other depths in the soil profile, which was similar to the soil temperature profile distribution in 28. Yoshida and Eguchi (1989) researched the effect of different root temperatures on water uptake in intact cucumber roots. The results showed a higher water uptake rate when the root temperature was greater than 16 C, but there was no significant increase as the temperature rose. Therefore, the soil temperature variation in the different treatments probably has little influence on root growth. 2.5 Soil nutrient distributions Before the experiments, all the plots at the field site contained the same levels of nutrients. Figs. 6- a and 6-b show the available nutrient distributions at the beginning and at harvest. Soil available N is more mobile in the soil solution, and is more susceptible to leaching by water movement. Thus, soil available N showed greater variability along the soil profile at harvest than soil available P (Figs. 6-c and 6-d) and K (Figs. 6-e and 6-f). Between and 4 cm depth, soil available N was relatively high at more than 4 mg kg 1. The differences among the three treatments were due to N uptake by winter wheat. At 4 6 cm depth, soil available N ranged from 2 to 4 mg kg 1, and the highest levels were found in the treatment, which might be due to leaching. The differences in N at each soil depth were small in 28 (Fig. 6-a), compared to the soil available N distribution in 29, because of the wet weather and the smaller irrigation volumes applied to the soils. Soil available P was relatively high at 2 cm depth, but lower in the deep soil layers (Figs. 6-c and 6-d). There were no differences in nutrient contents between the deeper layers for all three treatments. There were also no significant differences among the three treatments for soil available K in 28 and 29 (Figs. 6-e and 6-f). This may be due to the soil s buffering capacity against potassium release and adsorption. 2.6 Relative humidity within the canopy Fig. 7-a shows that there were no significant differences in RH within the canopy among the three treatments in 28, and the RH means during the experimental period were 89.5% for the treatment, 89.7% for the treatment, and 89.7% for the treatment. The high RH was due to

the wet and rainy weather in 28. Furthermore, RH in the canopy was higher than air RH measured by the standard weather station. The same phenomenon occurred in 29, but the RH within the canopy showed some differences, especially from 24 DAS to 228 DAS. These were caused by the irrigation and precipitation deficits. The RH mean within the canopy for the treatment was 85.6%, which was significantly higher than the 8.2% for the and 72.7% for the treatments (P<.5). In 28, RH within the canopy remained high and was nearly 9% in every treatment during the experimental period. In 29, there were large RH differences between treatments. Previous research showed that root elongation significantly increased when the atmosphere changed from arid to humid (Breazeale and Mcgeorge, 1953). Therefore, root growth might be stimulated under the high RH conditions found in the and treatments. 2.7 Root distribution Fig. 8 (a, b, c, d) shows the root distributions at the turn-green, jointing, flowering, and grainfilling stages in 28. By the turn-green stage, the roots had penetrated down to 7 cm depth and there were no significant differences among the treatments (Fig. 8-a). By the jointing stage, at 2 cm depth, RLD was significantly higher in the treatment (P<.5), followed by the and treatments (Fig. 8-b); at 2 6 cm depth, the treatment had the highest RLD; and at 6 1 cm depth, had the highest RLD. By the flowering stage, RLD increased throughout the profile, but the distribution pattern was similar to that at the jointing stage. At 2 cm depth, RLD was significantly higher in the treatment than in the and treatments (P<.5); and at 2 1 cm depth, the treatment was the highest, followed by the and treatments (Fig. 8-c). The differences in the RH, soil temperature, and soil nutrient profile distributions in 28 were small. Therefore SMP might be the key reason for the different RLDs. The high soil matric potential in the treatment could encourage root growth in the top layer, but is not conducive to root growth in the deeper layers (Torreano and Morris 1998). However, the low soil matric potential in the topsoil (Figs. 5-a and 5-b) may have stimulated root growth in the deeper layers (Rowse 1974; Kätterer et al. 1993; Benjamin and Nielsen 26). As a result, root profile distribution patterns varied between treatments because of the difference in irrigation frequency. Midway through the grain-filling stage, RLDs decreased sharply in the treatment, which coincided with the fastest decrease in topsoil SMP (Figs. 5-a and 5-b). At the same time, RLD in the deeper layers increased slightly in the treatment. In addition, RLD declined in each treatment because of the decrease in physiological water requirements during the grain filling period, which resulted in similar RLDs throughout the profile. Root distributions at the different growth stages in 29 are shown in Fig. 8 (A, B, C, D). Profile root distribution patterns for each treatment in 29 were similar to the corresponding stages in 28. The RLD during the flowering stage at 2 cm depth was highest in the treatment (P<.5), followed by the and treatments (Fig. 8-C). This may be due to the large

difference in SMPs among the three treatments in 29. The low RLDs at 4 6 cm depth, which was lowest in the treatment, might be due to the low soil available N in this layer. Therefore, soil available N also had an important impact on root growth among the three treatments. The root weight to length ratio indicates root thickness and the greater the value, the thicker the root, and vice versa. Fig. 9 shows the ratio profile distribution for each treatment. The ratio was significantly lower at 1 and 4 1 cm depth in the treatment, and was significantly higher in the treatment (P<.5). The increase in fine roots might be due to the water deficit and the high bulk density in the 1 cm layer. The fine roots might be better at extracting water at low soil water contents. In contrast, the high soil water contents in the treatment led to relatively thick roots. At 2 4 cm depth, the ratios for all treatments decreased, which might be due to the large bulk density in this soil layer, which may prevent roots from penetrating the soil. At 4 1 cm depth, the high water content in the treatment led to more thick roots than in the and treatments. In conclusion, root distribution pattern and root thickness were differentially affected by the three irrigation methods. A good knowledge of root distribution will help researchers to simulate soil water dynamics. Table 2 shows the root distribution fitting equation for,, and in 28 and 29. According to the fitting equation, the RLD profile distribution could be expressed by an exponential function. The RLD at the soil surface varied with the winter wheat growth stage and irrigation method. Furthermore, within the same treatment, the power level of the function showed little variation. The power level varied from.26 to.35 in the treatment, from.36 to.43 in the treatment, and from.45 to.53 in the treatment. The power level means for the,, and treatments were.3,.4, and.5, respectively. Therefore, the change in root distribution pattern caused by the irrigation methods could be described by the power level variation in the exponential function. 3. CONCLUON Irrigation method had a great effect on the micro-environments under winter wheat. The RH within the canopy, soil temperature, topsoil bulk density, soil nutrients, and soil matric potential profile distributions were all affected by the irrigation method used. The topsoil bulk density was significantly affected by the treatment, but was less affected by the and treatments. Soil temperature in the profile did not significantly differ among the three treatments. During the irrigation period, soil matric potential was highest in the treatment and fluctuated slightly. This was due to the higher water application frequency, compared to the and treatments. Atmospheric relative humidity within the canopy showed no significant differences among the three irrigation treatments in the relatively wet year of 28. However, 29 was a dry year, so the relative humidity mean within the canopy for the treatment was 85.6% and was significantly higher than the values of 8.2% in the and 72.7% in the treatments. The difference in atmospheric relative humidity in the canopy may affect root growth. However, the soil matric

potential was the key reason for the change in root profile distribution patterns among the microenvironments present in the three treatments. Soil available N also played a role in root distribution because of the soil losses due to water leaching and root uptake. The water deficit and high bulk density in the topsoil stimulated the production of fine roots in the treatment, which improved the crop s ability to extract water at low water contents. The thickest roots were found in the treatment, followed by the treatment. Root growth was stimulated in the topsoil layer and inhibited in the deeper layers in the treatment because of the high soil water contents caused by the increased water application frequency compared to the and treatments. As a result, the root distribution pattern differed for each irrigation method. The change in the root profile distribution pattern caused by the irrigation methods could be described by the power level variation in the exponential function. The variation in root profile distributions induced by the different irrigation methods will provide valuable information that could be used to improve agricultural practices. 4. MATERIALS AND METHODS 4.1 Experimental site Field experiments on winter wheat (Triticum aestivum L.) were carried out from the turn-green stage to harvesting during the 28 and 29 growth seasons at Tongzhou Experimental Base for Water-Saving Irrigation Research, Institute of Geographic Sciences and Natural Resources Research, Chinese Academy of Sciences, Tongzhou district, Beijing, China (39 36 N, 116 48 E; 2 m above sea level). The mean annual temperature is 11.3 C. Mean annual precipitation at the experimental site is 55 mm, only about 3% of which is received during the winter wheat season; spring and early summer in particular are normally quite dry. The water table depth is about 1 m, and the ground water temperature is about 14 16 C. The soil is predominately silt and loam and very deep. The soil hydraulic parameters are shown in Table 1. The precipitation data were collected from automatic weather stations installed close to the experimental plots. 4.2 Experimental design The experiment consisted of three treatments, each replicated four times. The 12 plots were arranged in a randomized block design, separated from adjacent plots by 3 m wide isolation strips. Each plot measured 8 m 8 m. Winter wheat, cv Jingdong No. 8, was sown at a rate of 3 kg ha 1 in 15 cm wide rows. Before sowing, the land was deep ploughed to 2 cm depth by a rotary cultivator. Each plot was given a basal dose of ammonium phosphate (3 kg ha 1 ), which was ploughed in, followed by a topdressing of urea (3 kg ha 1 ) when the crop was irrigated after it had begun growing again in the spring.

The three irrigation treatments were,, and, and each treatment had a separate pipeline. A water meter was installed to accurately control the amount of water applied. The system was equipped with a 63 mm PVC hydraulic pipeline and movable hose tubes. The inlet discharge was about 8 m 3 h 1. The system comprised of four impulse sprinklers (Elgo 8B2, Israel), which were placed at the four corners of each plot, and were mounted on 12 cm high risers. The spray angle was adjusted to 9. The spray range was about 8 m. Irrigation intensity was approximately 1 mm h 1, and the measured Christiansen irrigation uniformity coefficient (CU) was about 89.6%. The system consisted of a drip tape placed at a distance of 3 cm from the plants, which was midway between two adjacent rows. Emitter spacing was 2 cm and the flow rate was 2.7 L h 1. The first irrigation in each treatment was carried out at the jointing stage and the irrigation volume was determined by the difference between the measured mean volumetric SWC and the feidl capacity (FC) between 8 cm depth. Then a method based on pan evaporation (Liu and Kang 27) was adopted to determine the daily actual evapotranspiration (ET). The formula used was: when H.9 H max, ET K C E pan (1) H K C.13 1.17 (2) H max when H >.9 H max and LAI 1.75, K 1.2 (3) C where ET is the measured crop evapotranspiration (mm); K c is the crop coefficient, H is the plant height (m); H max is the maximum plant height (m) (the maximum plant height varies depending on the winter wheat cultivar grown; this information can be found in the seed specification); H max is equal to 9 cm in the experiment); and LAI is the leaf area index. In the treatment, the winter wheat was irrigated once a day according to the ET estimated from the previous day s data, except on rainy days. In the treatment, the winter wheat was irrigated around every 15 days, and in the treatment, irrigation was carried out around every 3 days, i.e. at the turn-green, flowering, and grain-filling stages. The irrigation volume was equal to the total supplied under drip irrigation to the and treatments. 4.3 Evaporation from the pan A standard pan (Model ADM, China meteorological administration, 23), 2 cm in diameter, was placed above the winter wheat canopy (the bottom of the pan was just above the top of the canopy) and was raised as the plants grew taller. Pan evaporation (E pan ) was measured at 8: h daily from the beginning of the treatments until harvest.

4.4 Air temperature and humidity Air temperature and relative humidity (RH) at the experimental station were measured by an automatic standard weather station (Davis Vantage Pro2, USA). Within the canopy of the winter wheat, the RH was calculated using the wet and dry bulb temperatures obtained from thermometers that were housed in a ventilated radiation shield. RH was calculated as follows: e( t) RH 1 (4) e ( t) where RH is relative humidity (%); e(t) is the actual vapor pressure at the dry bulb temperature in the canopy t (kpa); and e (t) is saturate vapor pressure at t (kpa). Actual vapor pressure, e(t), was calculated as follows: e t) e( t ) A P ( t t ) (5) ( w w where A is the psychrometric coefficient (.667 1 3 C 1 ) and P is the atmospheric pressure (kpa) at the experimental station; t and t w are the temperatures of the dry bulb and wet bulbs ( C ), respectively; and e(t w ) is saturated vapor pressure at temperature t w (kpa), as determined by Allen et al. (1998) using the following formula: 17.27t w e( tw).618exp( ) tw 237.3 (6) We determined e (t) from Eq. (6) by substituting the dry bulb temperature (t) for t w. The sensors for the dry and wet bulbs were installed in the middle of the canopy in each plot. The data were recorded every 2 minutes. 4.5 Soil temperature Soil temperature was measured by thermal resistors that were made of materials whose resistance varies as a function of temperature, and these were installed at seven depths:, 5, 1, 15, 2, 4, and 8 cm. The electric current was recorded every 2 minutes. Before measuring soil temperature, good correlations between the electric current and the adjusted temperature were obtained by using a thermostatic water bath. 4.6 Bulk density Soil bulk density at the soil surface was measured twice a year by the volumetric ring (VR) method before the treatments began and at the harvest stage. There were nine duplications and the measurements were taken 2 m apart in the middle of the plots in order to account for the effects of spatial variation on bulk density. 4.7 Soil matric potential Four tensiometers were installed in the middle of each treatment block to measure the soil matric potential at 2, 4, 6, and 9 cm depths. The tensiometers were arranged in two rows that were 6 cm apart. The matric potential of the soil was measured daily at 8: h throughout the experimental period.

4.8 Root length density Soil cores were removed by a 1 cm long auger with an internal diameter of 5 cm. They were then used to measure root length density (RLD) in the field. The cores were washed and the roots collected. Root samples were taken within rows and midway between the rows in each plot. The sampling depth went down to 1 cm, and a sample was taken every 1 cm. Samples were collected at four growth stage points. These were the turn-green, jointing, flowering, and grainfilling stages. Roots collected from each soil layer in the field were scanned with a scanner and analyzed using ArcGis (Version 9., ESRI), from which total root lengths were determined (Zheng et al. 24). RLD was calculated from the relationship between the total root lengths and the volume of the soil core. Acknowledgements This study is part of the work of Project (5119214, 311174 and 5139211) supported by the National Science Foundation and National Key Technology Research and Development Program (211BAD32B) and Basic Scientific Research Foundation of National non-profit Scientific Institute of China (BSRF2133). The authors thank the anonymous reviewers and editors for their insightful comments. References Adeoye K B. 1986. Physical changes induced by rainfall in the surface layer of an Alfisol, northern Nigeria. Geoderma, 39, 59-66. Allen R G, Pereira L S, Raes D, and Smith M. 1998. Crop evapotranspiration: guidelines for computing crop water requirements, FAO Irrigation and Drainage Paper 56. United Nations Food and Agriculture Organization, Rome. Bai W M, Li L H. 23. Effect of irrigation methods and quota on root water uptake and biomass of alfalfa in the Wulanbuhe sandy region of China. Agricultural Water Management, 62, 139-148. Bathke G R, Cassel D K, Hargrove W L, and Porter P M. 1992. Modification of soil physical properties and root growth response. Soil Science, 154(4), 316-329. Benjamin J G, Nielsen D C. 26. Water deficit effects on root distribution of soybean, field pea and chickpea. Field Crops Reseach, 97, 248-253. Breazeale L E, Mcgeorge W T. 1953. Influence of atmospheric humidity on root growth. Soil Science, 76(5), 361-366. Cavero J, Medina E T, Pug M, Martinez-Cob A. 29. Sprinkler irrigation changes maize canopy microclimate and crop water status, transpiration, and temperature. Agronomy Journal, 11, 854-864. Fabião A, Madeira M, Steen E, Kätterer T, Ribeiro C, Araújo C. 1995. Development of root biomass in an Eucalyptus globulus plantation under different water and nutrient regimes. Plant and soil, 168-169, 215-223.

Ferguson H, Boatwright G O. 1968. Effects of Environmental Factors on the Development of the Crown Node and Adventitious Roots of Winter Wheat (Triticum aestivum). Agronomy Journal, 6: 258-26. Jackson L E, and Bloom A J. 199. Root distribution in relation to soil nitrogen availability in field-grown tomatoes. Plant and Soil, 128, 115-126. Home P G, Panda R K, and Kar S. 22. Effect of method and scheduling of irrigation on water and nitrogen use efficiencies of Okra (Abelmoschus esculentus). Agricultural Water Management, 55, 159-17. Kar S, Varade S B, Subramanyam T K, Ghildyal B P. 1976. Soil physical conditions affecting rice root growth: bulk density and submerged soil temperature regime effects. Agronomy Journal, 68, 23-26. Kaspar T C, Bland W L. 1992. Soil temperature and root growth. Soil Science, 154(4), 29-299. Kmoch H G, Ramig R E, Fox R L, Koehler F E. 1957. Root development of winter wheat as influenced by soil moisture and nitrogen fertilization. Agronomy Journal, 49, 2-26. KÄtterer T, Hansson A, Andrén O. 1993. Wheat root biomass and nitrogen dynamics effects of daily irrigation and fertilization. Plant Soil, 151, 21-3. Liu H J, Kang Y H. 26. Effect of sprinkler irrigation on microclimate in the winter wheat field in the North China Plain. Agricultural Water Management, 84, 3-19. Liu H J, Kang Y H. 27. Sprinkler irrigation scheduling of winter wheat in the North China Plain using a 2 cm standard pan. Irrigation Science, 25, 149-159. Lv G H, Kang Y H, Li L, Wan S Q. 21. Effect of irrigation methods on root development and profile soil water uptake in winter wheat. Irrigation Science, 28: 387-398. Merrill S D, Rawlins S L. 1979. Distribution and growth of sorghum roots in response to irrigation frequency. Agronomy Journal, 71, 738-745. Nielsen K F, Halstead R L, Maclean A J, Holmes R M, Bourget S J. 196. The influence of temperature on the growth and mineral composition of oats. Canada Journal of Soil Science, 4, 255-264. Ragab R A. 1983. The effect of sprinkler intensity and energy of falling drops on soil surface sealing. Soil Science, 136(2), 117-123. Rawlins S L, Roats P A C. 1975. Prospects for high-frequency irrigation. Science, 188, 64-61. Rowse H R. 1974. The effect of irrigation on the length, weight, and diameter of lettuce roots. Plant and Soil, 4, 381-391. Sun Z Q. 26. Research on the effect of sprinkler irrigation on soil structure and the characteristic of water and nitrogen distribution. Ph. D thesis of Graduate University of Chinese Academy of Sciences, Beijing, China. (in Chinese) Tarchitzky J, Banin A, Morin J, Chen Y. 1984. Nature, formation and effects of soil crusts formed by water drop impact. Geoderma, 33, 135-155. Tolk J A, Howell T A, Steiner J L, Krieg D R. 1995. Role of transpiration suppression by evaporation of intercepted water in improving irrigation efficiency. Irrigation Science, 16, 89-95. Torreano S J, Morris L A. 1998. Loblolly pine root growth and distribution under water stress. Soil Science Society of American Journal, 62, 818-827. Unger P W, Kaspar T C. 1994. Soil compaction and root growth: a review. Agronomy Journal, 86, 759-766.

Wang D, Yates S R, Simunek J, and van Genuchten M Th. 1997. Solute transport in simulated conductivity fields under different irrigations. Transactions of the American society of Agricultural Engineers, 123(5), 336-343. Yoshida S, Eguchi H. 1989. Effect of root temperature on gas ex-change and water uptake in intact roots of cucumber plants (Cucumis sativus L.) in hydroponics. Biotronics, 18, 15 21. Zheng C H, Kang Y H, Yao S M, Yan C Z, Sun Z Q. 24. Method of root analysis using GIS technology. Transaction of the Chinese Society of Agricultural Engineering, 2(1), 181-183. (in Chinese) Zimmerman R P, Kardos L T. 1961. Effect of bulk density on root growth. Soil Science, 91(4), 28-288.

1 Table 1 Primary soil properties of experimental site 2 3 4 5 6 7 Soil depth Soil type BD OM TN TP ph θ r θ F θ s K s cm g cm -3 g kg -1 g kg -1 g kg -1 cm 3 cm -3 cm day -1-25 Sandy loam 1.39 8.69.6 1.61 7.61.69.286.395 16.53 25-5 Sandy loam 1.59 6.14.43 1.4 7.53.74.38.361 13.11 5-1 Sandy loam 1.48 5.92.38 1.13 7.62.94.291.386 21.37 1-115 clay 1.4 6.53.39 1.2 7.81.19.373.491 7.14 115-16 Loamy clay 1.48 3.18.19 1.3 8.6.11.366.489 9.37 BD: bulk density, measured by undistributed soil core method; OM: organic matter, measured by Potassium Bichromate Titrimetric Method; TN: total nitrogen, determined by Kjeldahl method (Bremner and Mulvaney, 1982); TP: total phosphorus, determined by perchloric acid digestion (Olsen and Sommers, 1982); ph (1:2.5 soil/water): measured by ph meter;θr: the residual water contents (cm 3 cm 3 );θ s : the saturated water contents (cm 3 cm 3 ); θ F : the field capacity (cm 3 cm 3 ); K s : saturated hydraulic conductivity (cm d 1 )

8 9 Table 2 Fitting equation of root length density profile distribution under different treatments in 28 and 29 Fitting equation ( RLD f () exp( a L) ) Year Treatment Jointing stage Flowering stage f () a R 2 f () a R 2 2.74.29.96 3.96.3.97 28 3.36.37.98 4.55.37.97 3.91.45.96 4.86.47.97 3.62.35.97 3.98.26.94 29 3.81.43.96 4.92.36.97 4.95.53.95 6.95.49.94 RLD: root length density (cm 3 cm -3 ); f (): RLD at the soil surface; L: soil depth (cm); a: fitted parameter

Precipitation (mm) Air temperature ( o C ) 6 5 28P 28AT 29P 29AT 3 25 4 2 3 15 2 1 1 5 174 184 194 24 214 224 234 244 1 11 12 Days after sow ing (DAS) Fig.1 Precipitation, and air temperature in 28 and 29 winter wheat seasons

Irrigation quota (mm) Irrigation quota (mm) Irrigation quota (mm) Irrigation quota (mm) 14 12 1 8 6 4 2 8 174 184 194 24 214 224 234 244 7 28 6 29 b 5 Sprinkler irrigation () 4 3 2 1 6174 194 214 234 5 4 3 2 1 28 29 Border irrigation () 28 Surface drip irrigation () a c 7 6 d 5 29 4 Surface drip irrigation () 3 2 1 174 184 194 24 214 224 234 244 13 14 15 Days after sow ing (DAS) Fig.2 Irrigation quota under different irrigation methods in 28 and 29 winter wheat seasons

Soil matric potential (cm) Days after sow ing (DAS) 174-1 184 194 24 214 224 234 244-2 -3-4 -5-6 2 cm a -7-271.9-1 -258. -26.4-217.7-483.6-479.8-488.6-512.6-2 -3-4 -5-6 -7-2 -3-4 -5-6 -7 4 cm 174-1 184 194 24 214 224 234 244-2 -3 6 cm 174 184 194 24 214 224 234 244-1 b c 16 17 18-4 -5-6 9 cm d -7 Fig.3 Varation in soil matric potential at different depths during the irrigation in 28 winter wheat season under different irrigation method (the dotted line means the termination of irrigation)

Soil matric potential (cm) 19 Days after sow ing (DAS) 174 184 194 25 215 226 236-1 -2-3 -4-5 2 cm a -6 174 184 194 25 215 226 236-1 -2-3 -4-5 4 cm b -6 174-1 184 194 25 215 226 236-2 -3-4 -5-6 -7 6 cm c -8 174 184 194 25 215 226 236-1 -2 2 21 22 23 24-3 -4-5 9 cm d -6 Fig.4 Varation in soil matric potential at different depths during the irrigation in 29 winter wheat season under different irrigation method (the dotted line means the termination of irrigation)

Days after sowing (DAS) 25 26 27 28 29 d ( ) e ( ) f ( ) Fig.5 Temporal and spatial distribution of soil temperature during the winter wheat season in 28 (a, b, c)and 29 (d, e, f)

Depth Depth Depth 2 4 Soil available -1 ) 2 4 6 8 2 4 Soil available -1 ) 2 4 6 8 6 8 28 a 6 8 29 b 1 1 12 14 16 2 4 Initial dist Soil available -1 ) 5 1 15 2 25 12 14 16 2 4 Initial dis Soil available -1 ) 5 1 15 2 25 6 8 28 c 6 8 29 d 1 1 12 14 16 2 4 Initial dis Soil available -1 ) 2 7 12 17 12 14 16 2 4 Initial dis Soil available -1 ) K (mg kg 2 7 12 17 6 8 28 e 6 8 29 f 1 1 3 31 32 12 12 14 Initial dis 14 Initial distribution 16 16 Fig.6 Mean soil available nutrients (N, P, and K) profile distribution at the sowing (initial distribution) and harvest of winter wheat under different irrigation methods

Relative humidity (%) Relative humidity (%) 33 34 35 36 1 9 8 7 6 5 4 3 2 1 1 9 8 7 6 5 4 3 2 1 Standard w eather station 174 184 194 24 214 224 234 244 Standard w eather station 174 184 194 24 214 224 234 244 Days after sowing Fig.7 Varation in relative humidity within the canopy and on the standard weather station during the winter wheat season in 28 (a) and 29 (b) a b

Depth(cm) Depth(cm) Depth (cm) Depth (cm) Root length density (cm cm -3 ) 1 2 3 4 5 6 Root length density (cm cm -3 ) 1 2 3 4 5 6 2 4 6 a 2 4 6 A 8 Turn-green stage (28, 16 DAS) 8 Turn-green stage (29, 162 DAS) 1 1 2 4 6 8 1 b Jointing stage (28, 188 DAS) 1 2 3 4 5 6 2 4 6 B 8 Jointing stage (29, 19 DAS) 1 1 2 3 4 5 6 2 4 6 8 1 1 2 3 4 5 6 c Flow ering stage (28, 212 DAS) 2 4 6 C 8 Flow ering stage (29, 217 DAS) 1 1 2 3 4 5 6 2 2 37 38 39 4 6 8 1 d Grain-filling stage (28, 238 DAS) 4 6 8 1 D Grain-filling stage (29, 238 DAS) Fig.8 Root length density profile distribution of winter wheat 28 and 29 under different irrigation methods

Depth (cm) 4 Root w eight to length ratio (mg cm -1 )..1.2.3.4 1 2 3 4 5 6 7 8 9 41 42 43 44 1 Fig.9 Root weight to length ratio profile distribution of winter wheat under different irrigation methods 25