wutl‘pls Illllllllllllllllllllllllil'llllll'lllllilll 3 1293 01691 4479 This is to certify that the thesis entitled MODELING MAIZE (Zea mays L.) LEAF DEVELOPMENT AND APEX TEMPERATURE UNDER DIFFERENT THERMAL ENVIRONMENTS presented by Marta Graciela Vinocur has been accepted towards fulfillment of the requirements for Mmelegree in Crop and Soil Sciences \ \ I U Major professor Date 1997 0-7639 MS U is an Affirmative Action/Equal Opportunity Institution LIBRARY Universlty Mlchlgan State PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINE-3 return on or before date due. MTE DUE DATE DUE DATE DUE ma mus-nu MODELING MAIZE (Zea mays L.) LEAF DEVELOPMENT AND APEX TEMPERATURE UNDER DIFFERENT THERMAL ENVIRONMENTS By Marta Graciela Vinocur A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Crop and Soil Sciences 1997 ABSTRACT MODELING MAIZE (Zea mays L.) LEAF DEVELOPMENT AND APEX TEMPERATURE UNDER DIFFERENT THERMAL ENVIRONMENTS By Marta Graciela Vinocur Accurate prediction of leaf appearance rate is required in maize (Zea mays L.) simulation models to estimate leaf area development, biomass and yield. Plant temperature is closely related to the development rate, but the air temperature record used to estimate plant temperature is often biased during the early stages of maize development because the growing plant parts are below the soil surface. A field study was used to compare measured soil, air and apex temperatures with maize leaf appearance rates. Seasonal variation in leaf tip appearance rates was observed for four sowing dates spaced about one month apart during 1996. Solar radiation and temperature of the air, apex and soil (0.01 m, 0.03 m and 0.05 m depths) were recorded on half-hourly intervals. Apex temperature was found to be close to the soil ‘ temperature at 0.03 m or 0.05 m when the apex was below the surface. When the apex was above the surface, its temperature was close to the air temperature. The phyllochron (degree- days between leaf appearance events) was found to be higher (52.4°C/leaf tip) than values used in most existing maize models. A functional model was developed to estimate mean daily apex temperature using inputs of daily maximum and minimum air temperatures and solar radiation. The model was tested using an independent weather data set. The resulting estimates had a root mean square error (RMSE) of 131°C per day and mean bias error (MBE) of -0.06°C respectively. After stem elongation pushed the apex above the soil surface, mean air temperature was close enough to the mean apex temperature to assume that they were equal. DEDICATION I affectionately dedicate this work to my husband, Eduardo, my son ,Pablo and my daughter, Lucia, for their love and unconditional support; to my parents for teaching me that knowledge and hard work are tools for my future; to Dr. Roberto Seiler for his encouragement and inspiration. iii ACKNOWLEDGMENTS I wish to express my deepest gratitude to Dr. Joe T. Ritchie, Homer Nowlin Chair Professor, my advisor, for his sincere concern about my education and his diligence in directing my studies. I am also grateful for the assistance of my guidance committee members Dr. Jeffrey Andresen, Department of Geography and Dr. Kenneth Pofl‘, Department of Horticulture and Botany. Their suggestions and insights made completion of this work possible. I am also thankful to Mr. Brian Baer, program analyst for the Homer Nowlin Chair Group at Michigan State University (MSU) for his computer assistance. I want to express my gratitude to the entire Homer Nowlin Chair Group at MSU whose members came fi'om all continents as students and visiting scholars for their help and friendship. The financial assistance of the Fulbright Commission for this Master degree program is gratefully acknowledge. I am also thankful to Universidad Nacional de Rio Cuarto for financial support and for granting me the leave time to undertake graduate studies at MSU. Appreciation is also expressed to the members of the Latin American Community for their friendship and support during my stay at MSU. iv TABLE OF CONTENTS LIST OF TABLES ..................................................... vii LIST OF FIGURES .................................................... viii Chapter 1: INTRODUCTION .............................................. 1 Objectives ................................................ 10 Bibliography .............................................. 12 Chapter 2: AIR, SOIL AND APEX TEMPERATURES INFLUENCES ON MAIZE (Zea mays L.) LEAF DEVELOPMENT ................................... 17 ABSTRACT ............................................... 17 Introduction ............................................... 18 Materials and Methods ....................................... 23 Results and Discussion ...................................... 26 Conclusions ............................................... 31 Bibliography .............................................. 52 Chapter 3: ESTIMATION OF MAIZE (Zea mays L.) APEX TEMPERATURE ..... 55 ABSTRACT ............................................... 55 Introduction ................ . ............................... 56 Materials and Methods ....................................... 59 Model description .................................... 59 Experimental data .................................... 60 Results and Discussion ...................................... 61 Conclusions ............................................... 64 Bibliography .............................................. 70 APPENDIX Description of the soil ....................................... 73 V Daily Weather Data for East Lansing, MI, 1996 ................... 75 Daily Weather Data - First Sowing Date ......................... 78 Daily Weather Data - Second Sowing Date ....................... 79 Daily Weather Data - Third Sowing Date ........................ 80 Daily Weather Data - Fourth Sowing Date ....................... 82 LIST OF REFERENCES ................................................. 83 vi LIST OF TABLES Table l: Accumulated thermal time (°Cd) using apex, air and soil temperatures from emergence until the ninth leaf-tip appeared for three different sowing dates. S 1, S , 3 and S 5 mean soil thermal time calculated using soil temperatures at 0.01, 0.03 and 0.05 m depths, respectively. ......................................... 27 Table 2: Minimum (Min), maximum (Max) and average air, apex and soil temperatures at three depths on July 18 1996 (Solar radiation 10.01 MI rn'2 ) and July 20 1996 (Solar radiation 29.79 MI m'z). S 1, S 3 and S 5 mean soil temperature at 0.01, 0.03 and 0.05 m depths, respectively ......................................... 28 Table 3: Correlation coefficients (r),slope (b), phyllochron (Phyl) (° Cd/leaf-tip) and standard error of the slope (leaf-tips/° Cd) (STD), for leaf-tip number predictions from the third leaf-tip calculated using thermal time (11') obtained fi'om air, apex and soil temperatures at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S S) depths for four different sowing .................................................. 33 Table 4: Summary of linear regressions, descriptive statistical parameters and measures of model performance between observed and predicted apex daylight (ID),apex night time (17V) and mean daily apex temperatures (TM) for year 1996, at East Lansing, MI. RMSE (root mean square error), % Err (percentage of error), MBE (mean bias error), E (estimated mean), O (observed mean), a (y-axis intercept), b (slope), N (number of observations), and r2 (coefficient of determination). RMSE, MBE, E and O are in °C. ..................................................... 63 vii LIST OF FIGURES Figure 1: Seasonal variation in thermal time accumulation from emergence until the appearance of the ninth leaf tip for the first sowing using air and soil temperatures with a base temperature of 8 °C. Soil temperatures are at 0.01 m (S l), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. ................................. 34 Figure 2: Seasonal variation in thermal time accumulation fiom emergence until the last leaf tip emerged for the second sowing using air, soil and apex temperatures with a base temperature of 8°C. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. .......................................... 35 Figure 3: Seasonal variation in thermal time accumulation fiom emergence until the last leaf tip emerged for the third sowing using air, soil and apex temperatures with a base temperature of 8°C. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. .......................................... 36 Figure 4: Seasonal variation in thermal time accumulation fi'om emergence until the last leaf tip emerged for the fourth sowing using air, soil and apex temperatures with a base temperature of 8°C. Soil temperatures are at 0.01 m (S l), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. .......................................... 37 Figure 5: Seasonal variation in daily average air and soil temperatures from emergence until the ninth leaf tip appeared for the first sowing. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. ...................... 38 Figure 6: Seasonal variation in daily average air, soil and apex temperatures from emergence until the final leaf tip appeared for the second sowing. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. ...... 39 Figure 7: Seasonal variation in daily average air, soil and apex temperatures from emergence until the last leaf tip appeared for the third sowing. Soil temperatures are at 0.01 m (S l), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. ......... 40 Figure 8: Seasonal variation in daily average air, soil and apex temperatures fi'om emergence until the sixth leaf tip appeared for the fourth sowing. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. ...... 41 viii Figure 9: Half-hourly variation of apex, soil and air temperatures for a sunny (July 20, 1996) (a) and a cloudy (July 18, 1996) (b) day during the second sowing. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. ...... 42 Figure 10: Total number of leaf tips as a fimction of thermal time calculated using air and soil temperatures from the third leaf-tip and for the first sowing. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. Each point is the mean of observations done in 10 plants. .......................... 43 Figure 11: Total number of leaf tips as a function of thermal time calculated using soil, air and apex temperatures from the third leaf tip and for the second sowing. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. Each point is the mean of observations done in 10 plants. ................. 44 Figure 12: Total number of leaf tips as a function of thermal time calculated using soil, apex and air temperatures, fi'om the third leaf tip and for the third sowing. Soil temperaturesareat0.01m(S 1),0.03 m(S 3)and0.05 m(S 5)deptbs, respectively. Each point is the mean of observations done m 10 plants. ................. 45 Figure 13: Total number of leaf tips as a function of thermal time calculated using soil., apex and air temperatures fi'om the third leaf tip and for the fourth sowing. Soil temperatures are at 0.01 m (S 1), 0.03 m (S 3) and 0.05 m (S 5) depths, respectively. Each point is the mean of observations done in 10 plants .................. 46 Figure 14: Total number of leaf tips as a function of thermal time calculated using soil temperature at 0.05 m depth (First sowing) and apex temperature for the other three sowing. Slope = 0.01906 leaves/ degree-day, Phyllochron = 52.4 degree-days/leaf tip, r = 0.996 ..................................................... 47 Figure 15: Total number of leaf tips as a function of thermal time calculated using soil temperature at 0.01 m depth. Slope = 0.01776 leaves/ degree-day, Phyllochron = 56.3 degree-days/leaf tip, r = 0.996 ....................................... 48 Figure 16: Total number of leaf tips as a function of thermal time calculated using soil temperature at 0.03 m depth. Slope = 0.01827 leaves/ degree-day, Phyllochron = 54.7 degree-days/leaf tip, r = 0.997 ....................................... 49 Figure 17: Total number of leaf tips as a function of thermal time calculated using soil temperature at 0.05 m depth. Slope = 0.01846 leaves/ degree-day, Phyllochron = 54.2 degree-days/leaf tip, r = 0.997 ....................................... 50 Figure 18: Total number of leaf tips as a function of thermal time calculated using air temperature. Slope = 0.02226 leaves/ degree—day, Phyllochron = 44.9 degree- days/leaf tip, r = 0.996 ............................................. 51 ix Figure 19: Linear regression analysis between A1 (°C) coemcient and Solar Radiation (MJ rn'2 day") for the data used in developed the model (Set 1). Data are from East Lansing, MI, 1996. ................................................ 65 ' Figure 20: Linear regression analysis between estimated and measured mean daylight apex temperatures for the independent data set (Set 2) with data from East Lansing, MI, 1996 ............................................................ 66 Figure 21: Linear regression analysis between estimated and measured mean night apex temperatures for the independent data set (Set 2) with data from East Lansing, MI, 1996 ............................................................ 67 Figure 22: Linear regression analysis between estimated and measured mean daily apex temperatures for the independent data set (Set 2) with data from East Lansing, MI, 1996 ............................................................ 68 Figure 23: Cumulative differences between mean daily estimated apex temperature and mean daily air temperature (calculated as the average of the daily minimum and maximum air temperature) as a fitnction of the day of the year, until the maize crop reach V7 stage, for East Lansing, MI, 1996. ............................ 69 Chapter 1 INTRODUCTION A search for an understanding of the basic relationships between plants and their environment has kept man occupied for centuries. The role of simulation in this search has attempted to bring all factors involved in the system together in order that we may evaluate various cause and effect relationships. Several crop models, with different levels of detail, have been developed to simulate crop growth and development on a daily basis (e.g., Acock et a], 1983; Wilkerson et a1, 1985; Ritchie et al, 1985; Jones and Kiniry, 1986; Boote, et a1, 1989; etc). Such models, due to their mechanistic or functional basis, have been used to address different types of problems. Crop models predict crop yields sown anywhere at anytime, and have proven to be useful tools for decision making by farmers, researchers and policy makers (Ritchie, 1986; Singh, 1989) at field, farm, regional, and national levels (Thornton, 1991; Thornton et al. , 1991). Crop simulation models should consider" plant growth and development as separate processes. Growth and development are affected by different environmental variables (Ritchie and NeSrnith, 1991) and have different sensitivity to water and nitrogen deficits and excesses (Ritchie, 1991). Modeling crop development is critical in order to predict crop productivity. The ability to estimate the stage of crop development is important forplanning management decisions such as timing of irrigation, fertilizer, herbicide or insecticide 2 application; for determining the availability of machinery for cultural practices; for adjusting planting dates so that flowering happens at optimum time to facilitate cross pollination in hybrid seed production; for growing plants with varying maturity dates to ease harvesting at intervals suitable to commercial canning operations; for defining which variety can be grown commercially in a specific area (Shaykewich, 1995) and for selecting a hybrid that will have a reasonable chance of maturing before fiost in the case of late planting or double cropping management systems. According to Wang (1960), one of the first studies related to crop development was done by Reamur (1735) who suggested that the time required for plants to complete a phase of their development could be more accurately estimated from the sum of daily air temperature than fi'om the sum of calendar days. He also stated that the sum of daily mean air temperature was nearly constant to reach a given maturity stage for any plant Following the work of Reamur, several methods of temperature summations to predict phenological stages have been developed (Gilmore and Rogers, 1958; Cross and Zuber, 1972; Brown, 1975; Tollenaar et al, 1979; Coelho and Dale, 1980). Different terms have been used to address this concept: degree days (°C d), day-degrees, heat units, heat sums, thermal units, and growing degree days although thermal time has been recommended for being used in a general terminology for temperature summation methods (Gallagher, 1979). Wang (1960) criticized thermal time methodology, partly because plants do not respond to air temperature in the same way during various development stages. Response differences are due to differences in a minimum threshold temperature for various physiological processes and differences in the location of the response within the plant. 3 Thus, using the same base temperature for all stages and a measured temperature that may not be the temperature of the site where developmental processes occur may result in prediction errors. Solar radiation, wind, vapor pressure deficit, rainfall, etc., also influence plant development but they are not taken into account by thermal time methodologies. In addition, these methodologies do not consider the efl'ects of extreme day or night temperatures, inter-diumal temperature changes (defined as the difference between the maximum temperature of one day and the minimum temperature of the following day) and the difference between day and night temperatures (Wang, 1960). Ritchie and NeSmith (1991) described different thermal time calculations using distinct base temperatures for several phenological stages of the crop. They also explained some possible sources of errors in thermal time calculations: error due to the time of manual recording of maximum and minimum temperatures in standard weather stations, differences in the value of the mean temperature based on the method of calculation (simple average of the daily maximum and minimum or average of hourly or less than hourly mean values), and. error due to the location of the weather station related to the crop and position of the instrument related to the place where development is occurring. Geiger (1971) also described the errors associated with thermal time calculations when the temperatme data are taken fi'om climatological stations near to sites of prediction. He stated that such calculations may not reflect the altered microclimate of the crop canopy caused by crop surface roughness and exposure differences between the standard climatological station grass surface and the field crop. Other studies also noted inaccuracies in the temperature data due to inconsistent time of observation of daily maximum and minimum temperatures (Mitchell, 1958; Baker, 1975; Schaal and Newman, 1976; Schaal and Dale, 1977), and systematic biases when mean 4 temperatures are calculated fiom simple maximum-minimum data (Hortik and Arnold, 1965; Robertson, 1968). Maize (Zea mays L) development research in recent years has been focused on understanding the primary role of temperature on development and on the improvement of the thermal time approach. Most maize phenological modeling is based on the concept of thermal time or growing degree-days (Splinter, 1974; Duncan, 1975; Coelho and Dale, 1980). Some variations of this concept have been introduced with the incorporation of photoperiod (Coligado and Brown, 1975), or by varying base temperature (Jones and Kiniry, 1986; Kiniry, 1991; Stapper and Arkin, 1980), which have improved the ability to predict development although none of the changes have eliminated thermal time. Thus temperature remains the primary factor driving maize development. Thermal time is usually calculated from air temperature, while the specific location on the plant where temperature influences development is in the developing point, the zone Where plant cell division and expansion is occurring (Ritchie, 1991). During the early growth stages of a maize plant, the zones of cell division and expansion are slightly below the soil surface. Under these conditions the development rates (leaf initiation, leaf appearance, or reproductive initiation rates) are more closely associated with temperature near the soil sm'face than with the air temperature (Beauchamp and Lathwell, 1967; Cooper and Law, 1978; Duburcq et al. , 1983; Hesketh and Dale, 1987). Models that use air temperature to predict canopy development implicitly assume that air temperature and temperature of the cell expansion zone are equal (Cellier, et al., 1993). These researchers stated that the assumed air-plant temperature similarities are usually acceptable when the canopy is fully developed because a large part of solar radiation is then dissipated into latent 5 heat, which induces low temperature differences between the air and the vegetation. However, in the early growth period, the plant is not large enough to significantly affect the energy exchanges between the soil and the atmosphere. In such cases most of the incident solar radiation is dissipated into sensible heat, resulting in a potentially large difference between air at the site of measurement and the near soil surface temperature. Temperatures at the cell expansion zone will be more extreme (Duncan et al., 1973), particularly in the daytime, so that the crop growing point can experience a significantly higher mean temperature than is recorded by the standard air temperature. This will result in an underestimation of the thermal time by a model using standard air temperatures. Previous experiments underscore the importance of the point where temperature is measured. Beauchamp and Lathwell (1967), in greenhouse studies of the effect of the root- zone constant temperatures on the early development of maize (based on leaf appearance rates), determined that for any growth-stage interval, the number of days required to reach that stage increased with decreasing root-zone temperature. Their data also revealed that the influence of the root-zone temperature in determining interval length was relatively greater once the plants had passed the 2-leaf stage and persisted only until the 6-leaf stage. 'Ihus root-zone temperatures regulated maize development only during the period of leaf initiation. After that, air temperature assumed increasing importance as the meristem is moving away from the root zone temperature. In another greenhouse study of the relationship between young maize stalk internal near meristem temperature (measured with a thermistor probe) and aerial and root zone temperature, Beauchamp and Torrance (1969) found a predominating influence of root zone temperature on the temperature of tissues in the apical region of maize plant shoots. They explained that when the soil temperature is lower than 6 the air temperature, the temperature of the apical meristem located 0.02 m to 0.03 m above the soil surface tends to remain 1°C to 3°C higher than soil temperature. Afterwards, in a controlled environment experiment, Watts (1972) showed that when meristem temperature was modified but shoot and root temperature were kept constant, rapid changes in maize leaves extension rates occurred. Subsequently, in a field experiment with different soil covers in order to induce soil temperature differences, Watts (1973) reported a close relationship between mean daily soil temperatures at 0.05 m depth and rates of leaf expansion and leaf appearance in maize. Barlow et al., (1977) noted that the rate of leaf elongation decreased with lower soil temperature due to the effect of lowering the temperature of the shoot apical meristem region in young maize plants. Coelho and Dale (1980) and Hanway (1982) concluded that soil temperatures strongly affect the rate of maize growth until the sixth leaf-stage when the growing point emerges above ground level. Cutforth and Shaykewich (1989) determined that the duration of the planting to emergence interval of maize was predominantly controlled by soil temperature (measured at 0.05 m depth) but that the duration fiom emergence to stem elongation was significantly related to air temperature and not to soil temperature. Although, their results were in contrast to the findings of other researchers cited above, Cutforth and Shaykewich (1989) explained the difi‘erences considering that the criterion used in their study was stem elongation and not leaf appearance rate or leaf extension. Previous studies have demonstrated that maize leaf development responds more to soil temperatm'e (Walker, 1969) than to air temperature until the sixth leaf-stage (Beauchamp and Lathwell, 1967; Watts, 1973). However, most models or prediction equations use air temperature as the basis for the thermal indices since it is more readily available than soil 7 temperature. Since most of the developmental processes can be observed at the apex level and the apex can be considered as an early image of the future plant, use of the actual apex temperature is most desirable. Measurements of soil temperature are often unavailable. Thus, soil temperature models have been developed for a variety of purposes and with varying degree of complexity and data requirements. They range in approach fi'om the more empirical and statistical models (e.g., Cruse et al., 1980; Gupta et al., 1981; Meikle and Treadway, 1979, 1982) to the more physical and deterministic models (e.g., Hanks, et al., 1971; Shroeder et al.,l978; Horton and Chung, 1991). Although statistical models are simple to construct and use, they are ofien site specific and require a large data base for developing the empirical coefficients. On the other hand, soil temperature models based on physical processes (radiative energy balance and sensible, latent, and grormd-conductive heat energy fluxes) like those fiom Bucham (1982), Sasamori (1970) or Schieldge et al., (1982) need much input data (e. g., solar radiation, radiation balance or vapor pressure) which are generally not available at a sufficiently dense time and spatial scale. Ten Berge (1990) and Horton and Chung (1991) developed models to predict bare soil temperature which are physically based but required 30 minutes interval data of solar radiation, vapor pressure, wind speed and rainfall or daily global radiation, maximum and minimum air temperature, average wind speed and total rainfall, respectively. Luo et al. , (1992) constructed a model using principles of energy balance and soil heat transfer which realistically simulated soil temperature with variable crop cover and soil water content but also required many inputs. Several other models have been developed to estimate soil temperature fiom air temperature (e.g., Hasfurther and Burman, 1974; Toy et al. , 1978; Gupta el al. , 1983; Dwyer et al., 1990). 8 As Potter and Williams (1994) emphasized, the desirable characteristics of soil temperature models for practical long-term simulations models should be: (i) minimal inputs, because the simulations often have only daily weather station data as input; (ii) high operational speed, because the simulations are for many years; (iii) sensitivity to management operations that may vary the crop biomass or residue on the soil surface; and (iv) reasonable robustness over a wide range of soil and climatic conditions. The models described above represent a broad spectrum of soil temperature models that have been developed. On the other hand, to my knowledge, only one model based on physical processes (Cellier et al., 1993) was designed to model maize apex temperature under field conditions when the leaf area index of the crop is lower than 0.5 and the apex is about 0.03 m above the soil surface. The model estimates the apex temperature for both day- time and night-time averages from hourly values of solar radiation, wind speed, air temperature and humidity. As the empirical coefficients obtained for the model are climate dependent, a separate calibration may be required to apply it to difiemnt environments. When Cellier et al., (1993) measured meristem temperatures using thin thermocouples, they found average differences between meristem and air temperatures up to +5 ° C during day- light hours with higher differences on sunny days than in overcast ones (4-6°C compared with l-2° C respectively). Higgins et al., (1964) proposed that leaf development was a valid index for estimating plant response to environmental conditions on a short term basis. Since leaf development involves leaf differentiation and leaf growth, the rate of leaf appearance provides an easily discernible index of plant-part differentiation without plant destruction. New leaf appearance events occur many times and at a predictable rate during the plant life 9 cycle. From the three ways in which leaf appearance has been described for maize (primordia, tip and collar), tip appearance is the simplest, nondestructive and almost linear throughout mostly of the vegetative cycle. Leafcollar appearance rates have been described as a linear function of the mean daily temperature when maize is grown at constant temperature over the 16°C to 28°C temperature range until V12 stage and in the absence of moisture or nutrient stress (W arrington and Kanemasu, 1983). However, they found a curvilinear relationship when maize is grown under differential day/night temperature regimes. Earlier work of Tollenaar et al, (1979) and Thiagarajah and Hunt (1979) found similar results for the rate of leaf tip appearance for a range of temperatures between 12°C and 26°C. The inverse of the slope of the thermal time-leaf tip appearance curve is called the phyllochron. The phyllochron or rate of leaf appearance , is defined as the time between the appearance of successive leaves on a shoot and is usually expressed in units of thermal time per leaf (McMaster and Wilhelm, 1995). The phyllochron provides a convenient method to describe plant vegetative development and aids in understanding and modeling crop development. As stated above, during the early developmental stages of maize, the apex of the plant is 0.01 or 0.02 m below the soil surface. During those times , the soil temperatme should be a better indicator of the apex temperature than air temperature. It is logical to expect that if we wish to be able to simulate maize development under different soil management scenarios and different sites, we need to estimate plant responses to soil temperatures resulting fiom these scenarios. If we choose to use near surface soil temperatures in estimating thermal time, we need to collect the data and determine the functional relationship between soil 10 temperature at that level, the apex temperature and air temperature. Such information is usually not available and is the main focus of this thesis. Objectives Maize is one of the most economically important plants grown in North America. Cool conditions after sowing due to environmental factors or different soil tillage practices may delay maize development by decreasing soil temperatures and may increase the risk of fiost terminating grain filling. On the other hand, warmer conditions due to environmental or management practices that increase plant development rates could shorten the crop cycle and reduce yields. Between these two extremes, a wide variety of environmental conditions can modify maize development. Cm'rent models can not provide a precise determination of maize phenological stages mainly because of the uncertainties of the plant-air temperatrue differences. It is unrealistic to expect to have apex temperature data available. Soil temperature data to use as an estimate of apex temperature are not usually measmed in many sites either. Most sites where plant development predictions are needed will have only air temperature data because the cost and time that soil and apex temperatures require is not available. If maize simulation models will be continue in use as management tools, in risk assessment and for predicting crop yield, a more accurate determination of the crop phenological stages is required. Bearing the above factors in mind, the specific objectives of this research are: ll 1 - To determine the functional relationship between air temperature measured with standard instrumentation at screen level (1.5 m height) and apex temperature under difi‘erent meteorological conditions. 2 - To determine the functional relationships between apex temperature and soil temperature at the different soil depths. ' 3 - To model maize leaf tip appearance rate using soil and apex temperatures and compare these results with those from air temperature. 4 - To model maize development using the functional relationships developed between air and apex temperature and / or soil and apex temperature in order to get a generalized approach to the observed data. Bibliography Acock, B., V.R. Reddy, F .D. Whisler, D.N. Baker, J.M. McKinion, H.F Hodges, and KJ Boote. 1983. The soybean crop simulator GLYCIM: Model Documentation 1982. Report N°2, United States Department of Energy, Carbon Dioxide Research Division, Ofice of Energy Research, Washington, DC. Baker, D.G. 1975. Effect of observation time on mean temperature estimates. Bull. Am. Meteorol. Soc., 14:471-476 Barlow, E.W.R., L.Boersma, and J .L.Young. 1977. Photosynthesis, transpiration, and leaf elongation in corn seedlings at suboptimal soil temperatures. Agron. J ., 69(1):95-108 Beauchamp, E.G., and D. J. Lathwell. 1967 . Root-zone temperature effects on the early development of maize. Plant and Soil, 26 (2): 224-234 Beauchamp, E.G., and J .K.Torrance. 1969. Temperature gradients within young maize plant stalks as influenced by aerial and root zone temperatures. Plant and Soil, 30 (2): 241-251 Boote, K.J., J.W. Jones, G.G. Wilkerson, and SS. Jagtap. 1989. PNUTGRO V1.02: Peanut crop growth simulation model. User’s guide. Florida Agricultural Experiment station Journal N°8420. University of Florida, Gainesville. Brown, D.M. 1975. Heat units for corn in Southern Ontario. Ontario Min. of Agric. and Food, Canada Fact sheet AGDEX 111/31, Order N 75-077. Bucham, GD, 1982. Predicting bare soil temperaturesl. Theory and models for multi-day mean diurnal variation. Journal of Soil Sci., 33: 185-197 Cellier, P., F. Ruget, M. Chartier and R. Bonhomme. 1993. Estimating the temperature of a maize apex during early growth stages. Agric. For. Meteorol., 63:35-54 Coelho, D. T. and RF. Dale. 1980. An energy-crop growth variable and temperature fimction for predicting corn grth and development: Planting to silking. Agron. J .,72:503-510 Coligado, M. C., and D.M. Brown. 1975. A bio-photo-thermal model to predict tassel initiation time in corn (Zea Mays L.). Agric. Meteorol., 15:11-31 12 l3 C00per, P.J.M., and LR. Law. 1978. Enhanced soil temperature during very early growth and its association with maize development and yield in the Highlands of Kenya. J. Agric. Sci. (Cambridge), 89:569-577 Cross, H2. and MS. Zuber. 1972. Prediction of flowering dates in maize based on difi‘erent methods of estimating thermal units. Agron. J., 64:351-355 Cruse, R.M., D.R. Linden, J.K. Radke and W.E. Larson. 1980. A model to predict tillage efi’ect on soil temperature. Soil Sci. Soc. Am. J ., 44:378-383 Cutforth, H.W., and CF. Shaykewich. 1989. Relationship of development rates of com fi'om plant to silking to air and soil temperature and to accumulated thermal units in a prairie environment. Can. J. Plant Sci., 69: 121-132 Duburcq, 1.3., R. Bonhomme, and M. Derieux. 1983. Durée des phases vegetative et reproductrice chez le mais. Influence du genotype et du milieu. Agronomic, 3:941-946 Dtmcan, W.G., D.R. Davis, and W.H. Chapman. 1973. Developmental temperatures in corn. Proc. Soil and Crop Sci. Soc. F 1a., 32:59-63 Duncan, W.G. 1975. SIMAIZ: a model simulating growth and yield in corn. In D.N. Baker et al (eds). The application of systems methods to crop production. Proc. Symp., Mississippi State, MS. 7-8 June 1973. Mississippi Agric. For. Exp. Stn., Mississippi State, MS Dwyer, L.M., H.N. Hayhoe, and J .L.B. Culley. 1990. Prediction of soil temperature fiom air temperature for estimating corn emergence. Can. J. Plant Sci., 70: 619-628 Gallaher, E.C. 1979.Field studies of cereal leaf growth. J. Exp. Bot., 30(117):625-636 Geiger, R. 1971.The climate near the ground. Harvard University Press, Cambridge, MA, 611 pp. Gilmore, EC. and J. S. Rogers. 1958. Heat units as a method of measuring maturity in corn. Agron. J., 50:611-615 Gupta, S.C., J .K. Radke, and W.E. Larson. 1981. Predicting temperatures of hm and residue covered soils with and without a corn crop. Soil Sci. Soc. Am. J ., 45: 405-412 Gupta, S.C., W.E. Larson, and DR. Linden. 1983. Tillage and surface residue efi‘ects on soil upper boundary temperatures. Soil Sci. Soc. Am. J ., 47:1212-1218. Hanks, R.J., D.D. Austin, and WT Ondrechen. 1971. Soil temperature estimation by a numerical method. Soil Sci. Soc. Am. Proc., 35:665-667. l4 Hanway, J.J. 1982. How a corn plant develops. Iowa State Coop. Ext. Serv. Spec. Rep. 48 (rev.feb.l982) Hasfurther, V.R, and RD. Burman. 1974. Soil temperatme modeling using air temperature as a driving mechanisms. Trans. ASAE, 17:78-81 Hesketh, J .D., and RP. Dale. 1987. Data for plant growth modeling: an evaluation. p 57-71. In Plant growth modeling for resource management. Vol.1. CRC Press Inc., Boca Raton, FL .Higgins, J .J ., J .R.Hawn, and E.J.Koch. 1964. Leafdevelopment: Index of plant response to environmental factors. Agron.J., 56:489-492 Hordk, H.J. and CY. Arnold. 1965. Temperature and the rate of development of sweet corn. Proc. Am. Soc. Hortic. Sci., 87:303-312 Horton, R., and S.Chung. 1991. Soil heat flux. In J Hanks and J .T. Ritchie (eds) Modeling plant and soil systems. Agronomy Monographs N° 31 . American Society of Agronomy, Madison, WI. Jones, C.A., and J .R. Kiniry. 1986. CERES-Maize: a simulation model of maize growth and development. Texas A&M Univ., College Station (Diss. Abstr. 85-14258) Kiniry, IR 1991. Maize phasic development In Hanks J and J .T. Ritchie ( Eds.) Modeling plant and soil systems. Agronomy Monographs N° 31. American Society of Agronomy, Madison, WI. Luo, Y., R.S. Loomis, and TC Hsiao. 1992. Simulation of soil temperature in crops. Agric. Forest Meteorology, 61: 23-38 McMaster, GS, and W.W. Wilhelm. 1995. Comparison of equations for predicting the phyllocrhon of wheat. Crop Sci., 35:30-36 Meikle, R.W., and TR. Treadway. 1979. A mathematical method for estimating soil temperature. Soil Sci., 129: 226-236 Meikle, R.W., and TR. Treadway. 1982. A mathematical method for estimating soil temperature in France. Soil Sci., 133: 58-60 Mitchell, J .M., Jr. 1958. The effect of changing observation time on mean temperature. Bull. Am. Meteorol. Soc., 39:83-89 Potter, K.N., and J .R. Williams. 1994. Predicting daily mean soil temperatures in the EPIC simulation model. Agron. J., 86: 1006-1011. 15 Reamur, RA.F.1735.Thermometric observations made at Paris during the year 1735, compared to those made below the equator on the Isle of Mauritius, at Algiers and on a few American islands. Acad. Sci. Memoirs Acad. Sci. Paris.545 Ritchie, J .T., D.C. Godwin, and S. Otter-Nacke. 1985. CERES-Wheat. A simulation model of wheat growth and development. Texas A&M University Press, College Station. Ritchie, J .T. 1986. Using computerized crop models for management decisions. Pages 27-41 . In Proc. International DLG-Congress for Computer Technology. May 1986. Hannover, Fed. Rep. of Germany. Ritchie, J .T and D.S. NeSmith. 1991. Temperature and crop development. Pages 5-29 In Modeling plant and soil systems. R.J.Hanks and J .T.Ritchie, (Eds). Agronomy Monographs N° 31. American Society of Agronomy, Madison, WI. Ritchie, J .T. 1991. Specifications of the ideal model for predicting crop yields. In R.C. Muchow and J .A. Bellany (Eds). Climatic risk in crop production: Models and management for the semi-arid tropics and subtropics. Proc. Intnl. Symposium, St. Lucia, Brisbane, Queensland, Australia. July 2-6, 1990. C.A.B. International, Wallingford, U.K Robertson, G.W. 1968. A biometeorological time scale for a cereal crop involving day and night temperatures and photoperiod. Int. J. Biometeorol., 12:191-223 Sasamori, T. 1970. A numerical study of atmospheric and soil boundary layer. J. of Atrn. Sci., 27:1122-1137 Schaal, LA. and J .E.Newman. 1976. Biased reading can alter heat units. Am. Soc. Agron. Crop. Soil., 28:7-10 Schaal, LA, and RF. Dale. 1977. Time of observation temperature bias and climate change. J. of Applied Meteorology, 16:215-222. Schieldge, J .P., A.B. Kahle, and RE. Alley. 1982. A numerical simulation soil temperature and moisture variation for a bare field. Soil Sci., 133:197-207 Schroeder, C., D.W. Michele, V.W. Pooch, and W.R. Teague. 1978. Modeling heat moisture flow in soils. Simulation (Nov): 173-179 Shaykewich, CF. 1995. An appraisal of cereal crop phenology modeling. Can. J. Plant Sci., 75:329-341 Singh, U. 1989. Introduction to modeling. Report for training program on computer simulation for crop growth and fertilizer responses. May 15-26, 1989. IFDC, Muscle Shoals, AL. l6 Splinter, W.E. 1974. Modeling of plant growth for yield prediction. Agric. Meteorol., 14:243-253. Stapper, M., and G.F Arkin. 1980. CORNF : a dynamic growth and development model for maize (Zea mays L.). Texas Agric. Exp. Stn. Res. Ctr. Program and model documentation 80-2. Ten Berge, H.F.M. 1990. Heat and water transfer in bare topsoil and the lower atmosphere. Center for Agricultural Publishing and Documentation (Pudoc), Wageningen. Thiagarajah, MR, and LA. Hunt. 1982. Effects of temperature on leaf growth in corn (Zea mays). Can. J. Bot., 60:1647-1652. Thornton, PK. 1991. Application of crop simulation models in agricultural research and development in the tropics and subtropics. Paper Series P-15. IFDC, Muscle Shoals, AL Thornton, P.K., J .B. Dent, and Z. Basci. 1991. A framework for crop simulation model applications. Agric. Syst., 37 :307—320 Tollenaar, M., T.B.Daynard, and RB. Hunter. 1979. Effect of temperature on rate of leaf appearance and flowering in maize. Crop. Sci., 19:363-366 Toy, T.J., A.J. Kuhaida Jr, and BE. Munson. 1978. The prediction of mean monthly air temperature. Soil Sci., 126:96-104 Walker, J .M. 1969. One-degree increments in soil temperature affect maize seedling behavior. Soil Sci. Soc. Am. Proc., 33:729-736 Wang, J .Y. 1960.A critique of the heat unit approach to plant response studies. Ecology 41 :785-789 Warrington, I.J., and ET. Kanemasu. 1983. Corn growth response to temperature and photoperiod II. Leaf-initiation and leaf-appearance rates. Agron. J ., 75:755-761. Watts, W.R. 1972. Leaf extension in Zea mays. Journal of Experimental Bot., 23-76:7l3- 721. Watts, W.R. 1973. Soil temperature and leaf expansion in Zea Mays. Expl. Agric., 9: 1-8 Wilkerson, G.G., J .W. Jones, K.J. Boote, K.T. Ingram, and J.W. Mishoe. 1983. Modeling soybean growth for crop management. Trans. of ASAE, 26:63-73 Chapter 2 AIR, SOIL AND APEX TEMPERATURES INFLUENCES ON MAIZE (Zea mays L.) LEAF DEVELOPMENT ABSTRACT Accurate prediction of leaf appearance rate is required in maize (Zea mays L.) simulation models to estimate canopy development, and ultimately maize yield. Most maize simulation models use air temperature for thermal time calculations to predict leaf appearance rate although nearby soil temperature is more closely related to the growing apex temperature than air temperature during early stages of development. A field experiment was conducted in 1996 at East Lansing, Michigan, on a Capac loam soil, to determine the effect of soil, air and apex temperatures on maize development and to evaluate their utility in the improvement of leaf developmental predictions. Maize leaf tip and leaf number were observed on four different sowing dates. Solar radiation and temperature of the air, apex and soil (0.01 m, 0.03 m and 0.05 m depths) were recorded on half-hourly intervals. Measured apex temperature was close to the soil temperature at 0.03 m or 0.05 m when the apex was below the surface or slightly above, otherwise the apex temperature was more related to the air temperature. An average bias of about 1.6 °Cd (degree-days) was found when air thermal time was used instead of apex thermal time. The phyllochron was found to be higher (52.4°Cd/leaf tip) that values actually used in most maize models, a possible reason of over-prediction 'of leaf development rates. 17 18 Introduction Crop simulation models should consider plant growth and development as separate processes. Growth and development are affected by different environmental variables (Ritchie and NeSmith, 1991) and have different sensitivity to water and nitrogen deficits and excesses (Ritchie, 1991). Most maize phenological modeling is based on the concept of thermal time (IT) or growing degree-days (GDD) (Splinter, 1974; Duncan, 1975; Coelho and Dale, 1980). Some variations of this concept have been introduced with the incorporation of photoperiod (Coligado and Brown, 1975) or by varying base temperature (Jones and Kiniry, 1986; Kiniry, 1991; Stapper and Arkin, 1980), which have improved the ability to predict development although none of the changes have eliminated thermal time. Thus, temperature remains the primary factor driving maize development. Wang (1960) criticized thermal time methodology, partly because plants do not respond to air temperature in the same way during various development stages. Response differences are due to the differences in minimum threshold temperatures for various physiological processes and to differences in the location of the response within the plant. Thus, using the same base temperature for all stages and a measured temperature that may not be the temperature of the site where developmental processes occur may result in prediction errors. Solar radiation, wind, vapor pressure deficit, rainfall, etc., also influence plant development but they are not taken into account by thermal time methodologies. In addition, these methodologies do not consider the effects of extreme day or night temperatures, inter-diumal temperature changes (defined as the difference between the 19 maximum temperature of one day and the minimum temperature of the following day) and the difference between day and night temperatures (Wang, 1960). For thermal time to be appropriate as a predictor of plant development, there should be a linear relationship between the development rate and plant temperatm'e over a well- defined range of temperatures; the daily temperature should not fall below the base temperature or exceed an upper threshold temperature for a significant part of the day; and the developing plant portion should have the same mean temperature as the temperature being used in the summation (Ritchie and NeSmith, 1991). Thermal time is usually calculated fi'om air temperature, while the specific location on the plant where temperature influences development is in the developing point, the zone where plant cell division and expansion is occtu'ring (Ritchie, 1991). During the early grth stages of a maize plant, the zones of cell division and expansion are slightly below the soil surface. Under these conditions the development rates (leaf initiation, leaf appearance, or reproductive initiation rates) are more closely associated with temperature near the soil surface than with the air temperature (Beauchamp and Lathwell, 1967; Cooper and Law, 1978; Duburcq et al., 1983; Hesketh and Dale, 1987). Models that use air temperature to predict canopy development implicitly assume that air temperature and temperature of the cell expansion zone are equal (Cellier, et al., 1993). These researchers stated that the assumed air-plant temperature similarities are usually acceptable when the canopy is fully developed because a large part of solar radiation is then dissipated into latent heat, which induces low temperature differences between the air and the vegetation. However, in the early grth period, the plant is not large enough to significantly affect the energy exchanges between the soil and the atmosphere. In such cases most of the incident 20 solar radiation is dissipated into sensible heat, resulting in a potentially large difference between air at the site of measurement and the near soil surface temperature. Temperatm'es at the cell expansion zone will be more extreme (Duncan et al., 1973), particularly in the daytime, so that the crop growing point can experience a significantly higher mean temperature than is recorded by the standard air temperature. This will result in an underestimation of the thermal time by a model using standard air temperatures. On the other hand, during periods when soils are cooler than air temperatures, as often is the case during the early spring, the estimated thermal time required for development based on air temperatures would be greater than that based on soil or apex temperattu'es (Hesketh and Warrington, 1989). Cellier et al., (1993) found day-light average differences between air and apex temperature up to +5 ° C, with higher difi‘erences on sunny days than in overcast ones (4-6°C compared with 1-2°C respectively) during the early stages of development of maize plants. They measured maize apex temperature by inserting thin thermocouples in the plant at two heights (0 and 0.03 m) above the soil surface. Cellier et al., (1993) designed a model based on physical processes, to estimate the apex temperature for both day-time and night time averages fiom hourly data of solar radiation, wind speed, air temperature and humidity. As the empirical coefficients obtained for the model are climate dependent, a separate calibration may be required to apply it to different environments. J eppson and Crookston (1986) found that apex temperature was closely associated to soil temperature. They suggested that plants were able to dissipate all the plant-intercepted heat from the sun and that any apex heating came indirectly fiom the soil. The differences found between air and 21 meristem temperature support the evidence of biases when air mm is used in thermal time calculations instead of apex or near surface temperatures. Higgins et al., (1964) proposed that leaf development was a valid index for estimating plant response to environmental conditions on a short term basis. Since leaf development involves leaf difi‘erentiation and leaf growth, the rate of leaf appearance provides an easily discernible index of plant-part differentiation without plant destruction. New leaf appearance events occur many times and at a predictable rate during the plant life cycle. From the three ways in which leaf appearance has been described for maize (primordia, tip and collar), leaf tip appearance is the simplest, nondestructive and almost linear throughout mostly of the vegetative cycle. Leaf-collar appearance rates have been described as a linear function of the mean daily temperature when maize is grown at constant temperature over the 16°C to 28°C temperature range until V12 stage and in the absence of moisture or nutrient stress (Warrington and Kanemasu, 1983). However, they found a curvilinear relationship when maize are grown under differential day/night temperature regimes. Earlier work of Tollenaar et al, (1979) and Thiagarajah and Hunt (1979) found similar results for the rate of leaf tip appearance for a range of temperatures between 12°C and 26°C. Dwyer and Stewart (1986) showed a high correlation of leaf stages (collar visible) with three different thermal indices: GDD, Maize Heat Unit Index and Night Temperature Index. They suggested that air temperature alone could adequately account for variability in the time of appearance of each mature leaf as soil type and year did not contribute significantly to that variability. The slope of the curve of leaf tip appearance versus thermal time for the data of Tollenar et al. (1979) was approximately 0.0265 leaves/°Cd (leaves per degree-day) between 22 8°C and 34°C (Ritchie and NeSmith, 1991). The inverse of the slope of the thermal time- leaf tip appearance curve is called the phyllochron. The phyllochron or rate of leaf appearance, is defined as the time between the appearance of successive leaves on a shoot and is usually expressed in units of thermal time per leaf (McMaster and Wilhelm, 1995). The phyllochron provides a convenient method to describe plant vegetative development and aids in understanding and modeling crop development. Ritchie and NeSmith (1991) stated that measurements of the phyllochron were in the narrow range of 38° to 45°Cd/leaf tip appearance when a base temperature of 8°C was used in the thermal time calculation. In other environmental studies, the thermal requirement per tip varied fiom 33 to 42° Cd and the base temperature from 6 to 9°C (Zur et al., 1989; Tollenaar et aI. , 1979). Field studies at different elevations in Kenya resulted in 41°Cd per leaf tip and a 9°C temperature base (Cooper, 1979). Picard et al.,(1985) reported 35 to 43°Cd per tip using a base temperature of 6°C. Assuming a base temperature of 8°C, the thermal requirement varied ficm 37 to 42°Cd per leaf in three field studies involving numerous hybrids at Urbana fi'om 1985 to 1987 (J .D. Hesketh, unpublished data, cited in Hesketh and Warrington, 1989). In all these fields studies estimates were not corrected for soil temperature effects. During the early developmental stages of maize, the apex of the plant is 0.01 or 0.02 m below the soil surface. During those times, soil temperature should be a better indicator of the apex temperature than air temperature. Although previous studies (Cellier et al. , 1993; Ben-Haj Salah and Tardieu, 1996) showed a difference between apex and air temperatm'es during short periods of the maize life cycle, the effect on maize development was not explained. Based on the utility of the leaf tip appearance rate and temperature relationship 23 to predict maize development, it is logical to expect that if we want to simulate maize development under different soil management scenarios and different sites, we need to estimate plant responses to soil temperatures resulting fiom these scenarios. An experiment therefore was designed to determine the effect of soil, air and apex temperatures on maize (Zea mays L.) development and to evaluate their utility in the improvement of maize phenology prediction. Materials and Methods A field experiment was conducted in 1996 at the Michigan State University Research Farm, East Lansing, Michigan (42° 78' N, 84° 60' W), on a Capac loam soil (Fine-loamy, Mixed, Mesic, Aerie Ochraqualf). More details of the soil description are given in the Appendix). Maize (hybrid ’Pioneer 3572’) was manually planted at a plant density of 6.5 plants rn'2 and at 0.05 m depth on four different dates: 29 May, 30 June, 2 August and 29 August. The seeds were sowed approximately one month apart to expose the crop to different thermal environments. Each adjacent experimental unit was 45 m2, with five rows 15 m in length and 0.75 m apart except for the fourth sowing which had only one row 15 m in length. All plots were moldboard plowed in the fall and received secondary tillage prior to planting in the spring. Plots were raked before planting to level the surface. Starter fertilizer was applied at sowing at a rate of 45-45-45 kg ha'1 (N-P-K). Weed plants were hand removed during the growing season. An irrigation of 25 mm was applied on 24 July. 24 Soil temperatures were measured for each sowing date, at depths of 0.01 m, 0.03 m and 0.05 m every five seconds and halfhome average values were stored in a LI-COR 1000 data logger (LI-COR, Inc, Lincoln, NE). Air temperature at screen level (1.5 m height) and solar radiation were recorded on each plot at the same fiequency. A second data logger, Campbell CR 10 (Campbell Scientific, Inc, Logan, Utah) was used to complement the first one when measurements in the different sowing were overlapped. Temperature in the region around the apex (called from now apex temperature) was measured using thin copper-constantan thermocouples needles of 0.008" diameter (Omega HYP-0-33-1-T-G-120-SMP-M, Mini Hypodermic thermocouple probe, Omega Engineering, Stamford, CT). Each needle was thermally insulated with silicone and covered with shrinkable tubing, leaving only 0.005 m uncovered at its end where the thermocouple junction was placed. Data were recorded every 5 seconds and averaged every 30 minutes. They were stored in a data logger (Campbell CR 10 or LI-COR 1000, with 1000-10 Thermocouple Terminal Block). The temperature sensor was changed to a new plant almost daily to avoid error in temperature values due to the possible effect of damaged tissue where the sensor was inserted. Maximum and minimum soil, air, and apex temperatures were recorded at the same fi-equency. Rainfall data were collected at the nearest weather station (MSU Horticultural Research Station, East Lansing). Data were aggregated over the day, and the daily maximum, minimum, and average values were calculated fiom the half hourly maximum, minimum and average readings. Ten consecutive plants away from the plots’ borders were marked at the beginning of the growing season. The date when 50% of the plants reached the different stages of development as defined by Ritchie and Hanway, (1982) was observed. Vegetative plant 25 development was studied using appearance rate of leaf tips. Fully expanded leaves (FL) were recorded when the collar appeared and total number of leaves (T'LN) was determined when the tip appeared. The number of tips on each plant was observed three times per week. Thermal time (TT) was calculated using soil temperature at each measured depth, apex temperature and air temperature for each of the four sowing fi'om emergence until the last tip appeared. Thermal time was defined as: 77:2;(f- Tb) where: I" is daily mean air, soil or apex temperature and Th is the base temperature at which development stops. The base temperature used was 8°C (Ritchie and NeSmith, 1991). The daily mean was calculated by averaging half hourly mean temperature values for the different variables. When the daily average temperature was below T, , no value was added to the summation. Thermal time calculated with the different temperatures was compared to determine which provided the best estimation of tip appearance rate and vegetative development. The phyllochron (thermal time required for each tip to appear) was calculated for each one of thermal time-leaf tip relationships determined using air, soil and apex temperattues. Because needle thermocouples were not available until the second sowing, apex temperature was not measured for the first sowing. Some failures in the recording of the apex temperature at the beginning of the third sowing provided an incomplete record. The missing data of the third sowing were estimated using regression formula developed with the available apex and soil temperature data for that sowing. 26 Results and Discussion Thermal time calculation fiom daily average air, apex and soil temperatures at the different depths, for each sowing and from crop emergence are presented in Figures 1 to 4 for the four sowing dates. Accumulation of thermal time was faster for soil or apex temperatures than air temperatures for all the sowing dates because of higher average temperatures in the soil and apex. Apex thermal time was closer to soil thermal time calculated using soil temperatures at 0.03 m or 0.05 m depths for the second and third sowing, and to soil temperature at 0.01 m depth for the fourth sowing. These results demonstrate that apex temperature is closer to the temperature of the soil at the depth where the growing point is situated until the apex emerged above the soil surface when it becomes to be more affected by air temperature. The apical meristem remained below ground until V5 stage (Ritchie and Hanway, 1982) and reached the soil surface a few days prior to V6 stage when the total number of leaf tips was 9 for the first three sowing dates. No data for this stage were available for the fourth sowing because plants were killed by a fiost at V4 stage and with 6 leaf tip. The second and third sowing reached V6 stage at about the same amount of thermal time and days (25 days) while the first sowing required more days (27 days) and more thermal time (Table 1). Differences in the thermal time required to reach different development stages are due to variations in the patterns of soil, apex and air temperatures observed during the four sowing. Soil and air temperatures did not differ consistently until June 23 (Day of the year 17 5) during the first sowing probably because of wet weather and high soil water content which decreased soil temperatures (Figure 5). Daily average differences between apex and 27 air temperature of up to +5°C on sunny days and up to +2°C on overcast days were found for the last three sowing when the apex was below the soil surface or slightly above (Figures 6, 7 and 8). Smaller fluctuations between air and apex temperatures were found when stem elongation moved the apex above the soil surface, thus making it more afl'ected by the air temperatme. These'results supported previous similar findings by Cellier et al, (1993). Soil and apex temperatures followed the same pattern and had similar values while the apex was below the soil surface (Figures 6, 7 and 8) which also agreed with Duncan et al., (1973) conclusions. Table 1: Accumulated thermal time (°Cd) using apex, air and soil temperatures fiom emergence until the ninth leaf-tip appeared for three different sowing dates. S 1, S 3 and S 5 mean soil thermal time calculated using soil temperatures at 0.01, 0.03 and 0.05 m depths, respectively. Sowing Air S 1 S 3 S 5 Apex First 373.7 447.7 441.3 432.1 Second 327.2 433.6 425.6 418.9 411.8 Third 335.4 430.2 414.3 406.9 407.1 Hourly series of all measured temperatures are shown in Figure 9 for a sunny and a cloudy day for the crop sowed on the second date. Higher differences between apex and air temperatures are evident after sunrise, increasing during the day with a maximum around noon (almost 10°C difference) and decreasing in the late aftemoon (Figure 9 a). A similar pattern was observed for soil temperature at 0.01 m depth while a delay in the time of occurrence of maximum and minimum temperatures was shown by the other two soil depths due to the soil buffering effect (Table 2). Smaller differences between all temperatures are 28 evident on overcast days (Figure 9 b) because of the decreased amount of soil heating associated with less direct solar radiation on the soil surface. Even on cloudy days however, the apex and soil temperatures were slightly higher than air temperatures (Table 2). Table 2: Minimum (Min), maximum (Max) and average air, apex and soil temperatures at three depths on July 18 1996 (Solar radiation 10.01 MJ rn'2 ) and July 20 1996 (Solar radiation 29.79 MJ m'z). S 1, S 3 and S 5 mean soil temperature at 0.01, 0.03 and 0.05 m depths, respectively Day July 18 July 20 Temperature Max Min Average Max Min Average Air 29.5 19.3 23.2 25.1 9.7 17.2 Apex 31.4 20.1 24.1 34.8 12.5 22.7 S 1 31.7 20.3 24.2 37.5 13.1 23.9 S 3 ' 28.8 21.0 24.0 33.5 14.9 23.4 S 5 28.1 21.2 24.0 33.1 15.7 23.6 Differences in air and apex temperature patterns affected thermal time calculations and suggested that apex temperature should be used instead of air temperature in the determination of maize leaf development. When apex temperature is not available, soil temperature at 0.03 m or 0.05 m depths appear to be adequate substitutes. To characterize maize leaf development, the relationship between total number of leaf tips and thermal time calculated with the different temperatures was studied. Figures 10,11, 12 and 13 show these relationships for the different sowing dates fiom emergence of the third leaf tip. The third tip was chosen as the lowest one because the first and second tips appeared at a considerably faster rate. Leaf development was highly correlated to thermal 29 time calculated with the measured air, soil and apex temperatm'es for all sowing dates (T able 3). The number of leaf tips considered in the analysis was difi‘erent for each sowing making comparisons of the results between the different sowing dates difi'rcult. For the first and fourth sowing, the number of tips was nine and six respectively. The second and third sowing tips number were sixteen and fifteen, respectively. Differences in thermal time calculated using soil and air temperatures were determined dtuing the first stages of crop development (Figures 2 and 3) when soil and air temperatures showed larger differences (Figures 6 and 7) and were maintained for the remainder of the growing season. Smaller differences between soil and air temperatures characterized the first and fourth sowing (Figure 5 and 8). Lower correlation of apex thermal time for the third sowing could be related to the use of estimated values of apex temperature at the beginning of that sowing due to failure of the recording system. Fortin and Pierce (1991) and Dadoun (1993) suggested the use of soil temperature in thermal time calculations when the apex is below the surface and air temperature afterwards although their research was carried out with different types of mulch applied to the soil surface. Although correlation coefficients between leaf tip number and thermal time calculated with the air, apex or soil temperatures recorded in each plot were higher in all cases, differences arise when the phyllochron is considered. The second and third sowing, within the third to sixteen or fifteen leaf tips respectively, showed a phyllochron calculated with air thermal time between the ranges found by previous researchers with the same base temperature (Ritchie and NeSmith, 1991; J .D. Hesketh, unpublished data, cited in Hesketh and Warrington, 1989). The first and fourth sowing, with fewer leaf tips (from the third to the ninth or sixth leaf tips respectively) indicated higher phyllochron. Cooler temperatmes 30 characterized the beginning of the first sowing and most of the fourth sowing, which may have affected maize leaf development through delaying leaf tip appearance rate. Landi and Crosbie (1982) found that short periods of cold stress prior to full emergence of the fifth leaf reduced leaf emergence in some hybrids. Symptoms of cold stress may be also expected at temperatures of 10 - 12° C (Taylor and Rowley, 1971). Signs of cold stress were evident at the end of the third and fourth sowing when air and apex temperatures were around 15 °C or lower. An error in the prediction of maize leaf development of about 7° Cd per phyllochron is evident if air thermal time is used instead of apex thermal time. Thus, models that estimate maize leaf development based on air thermal time are likely to over-predict the number of leaf tips. For the second and third sowing, the difference between the meastued number of . leaf tips and the predicted using air thermal time instead of apex thermal time is almost 2 leaf tips. When all data are aggregated (Figure 14) using apex thermal time for the last three sowing dates and soil thermal time calculated with S 5 for the first sowing, the average phyllochron is 52.4 degree-days for all but the first two tips. Although a linear relationship between leaf tip appearance and thermal time is shown, a deviation from the linear pattern is observed during some portions of the second and third sowing from the point where the number of tips is around nine or above. At that time the apex is beginning to move above the soil surface level due to rapid shoot elongation. After that a closer relationship to air temperature was observed. A slight water stress which occurred fiom the seven to the eight tip could have affected the rate of leaf tip appearance in the second sowing and induced this change in the rate pattern. Irrigation was applied when the second sowing had 8 tips. 31 Muchow and Carberry (1989) found that a reduction in the rate of leaf-collar appearance occurred when the crop experienced water stress dining the early growth stages (approximately from tassel initiation to anthesis). Re-watering corrected the situation and the rate was increased for the leaf collar appearance as compared to the non-stressed plots (Muchow and Carberry, 1989). For the third sowing, a period of air temperatmes around 15°C or less began when the crop had approximately 11 leaf tips (F igme 7). The effect of low temperatures on leaf development was described earlier which may explain the decrease in the leaf tip appearance rate. There was no apparent reason for the increase in the rate of leaf tip appearance for leaves 14 and above observed for the second and third sowing although it may be related to the stage of development as the final leaf tips are appearing for the third sowing. When the combined data analysis was carried out with soil thermal time (S 1, S 3 and S 5) (Figure 15, 16 and 17 respectively) or air (Figure 18) instead ofapex thermal time, the average phyllochrons have different values which support the necessity of choosing the right temperature to accurate predict leaf development. Conclusions This experiment demonstrated that there was a consistent bias between apex temperature and air temperature during early growth stages of maize crop development under different thermal environments. These findings are in agreement with previous studies of Cellier et al. , (1993) and Ben-Haj Salah and Tardieu, (1996). Soil temperatme at 0.03 m or 0.05 m depths proved to be a better indicator of the temperature that is affecting early 32 developmental processes than air temperature because soil temperatures were quite close to the apex temperature when the apex was near the soil surface. When the crop had more than six full developed leaves or nine leaf tips and the apex was above the soil surface, apex temperatures were closer to air temperatures. This study demonstrated that the temperature measured in air, apex or soil, when accumulated through most of the season had a high correlation with leaf development. The apex temperature was more consistently correlated between the sowing dates. As a result, crop models which use air temperattu'e in thermal time calculation are over-predicting the total number of leaf tips and maize leaf development A higher phyllochron than used in most maize crop models was identified, indicating the necessity of changing its value from about 40°Cd per phyllochron to about 52°Cd in order to adequate predict maize leaf development based on the bias in thermal time found in this study. 33 Table 3: Correlation coefficients (r),slope (b), phyllochron (Phyl) (° Cd/leaf-tip) and standard error of the slope (leaf-tips/° Cd) (STD), for leaf-tip number predictions fi'om the third leaf-tip calculated using thermal time (11') obtained fi'om air, apex and soil temperatures at 0.01 m (S l), 0.03 m (S 3) and 0.05 m (S 5) depths for four different sowing Sowing TT b Phyl r STD First (until leaf-tip 17 Air 0.0191 52.36 0.999 0.0002 until leaf-tip 9) Air 0.0195 51.28 0.999 0.0004 S 1 0.0162 61.73 0.997 0.0005 S 3 0.0164 60.98 0.998 0.0004 S 5 0.0167 59.88 0.998 0.0004 Second (until leaf-tip 16) Air 0.0218 45.87 0.998 0.0003 S 1 0.0185 54.05 0.998 0.0003 S 3 0.0187 53.42 0.998 0.0003 S 5 0.0188 53.16 0.998 0.0003 Apex 0.0198 50.53 0.997 0.0003 Third (until leaf-tip 15) Air 0.0219 45.57 0.995 0.0005 S 1 0.0172 58.14 0.994 0.0004 S 3 0.0178 56.07 0.995 0.0004 S 5 0.0178 56.07 0.996 0.0003 Apex 0.0190 52.60 0.992 0.0005 Fourth (until leaf-tip 6) Air 0.0201 49.75 0.997 0.0006 S 1 0.0162 61.73 0.996 0.0006 S 3 0.0156 64.10 0.996 0.0006 S 5 0.0147 68.03 0.996 0.0005 Apex 0.0180 55.56 0.997 0.0006 34 450- 1858 350. ,0 ,. 300- 9 250- 200- 150- 100- 13°8 50‘ 0 Thermal time (Degrees-days) <3 OI