.¢.....!..u .....:.IO\. . ..-».......;....... m 5/ v: ...y..... THESE! m i\\\\\\\\ \\\\\ 7\\\\\\\\\\\\\\ 3001293 r ' LIBRARY Michigan State University \ I‘ This is to certify that the dissertation entitled FUNCTIONAL RELATIONS OF ROOT DISTRIBUTIONS AND FLUX OF WATER AND NITRATE presented by Robert Martin Aiken has been accepted towards fulfillment of the requirements for 1311.1); degreein Crop and $01] Sciences Date £7042! é /J;73\ MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. DATE DUE DATE DUE DATE DUE J9T_1§9ort:||n——— ”I DEC 0 5 2009 h Li —I__l:—I MSU Is An Affirmative Action/Equal Opportunity Institution chS-DJ FQNCTIONAL RELATIONS OF ROOT DISTRIBUTIONS WITH THE FLUX AND UPTAKE OF WATER AND NITRATE BY Robert Martin Aiken A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Crop and Soil Sciences 1992 ABSTRACT FUNCTIONAL RELATIONS OF ROOT DISTRIBUTIONS WITH THE FLUX AND UPTAKE OF WATER AND NITRATE BY Robert Martin Aiken Nitrate leaching reflects poor nutrient retention, poses a hazard to public health and a challenge to solute transport theory. Field observations and numerical simulation of soil-plant interactions are integrated to identify sensitivity of simulated nitrate leaching to errors in predicted root function. Seasonal changes in maize root distributions, canopy development, and gradients in soil water content and carbon dioxide partial pressures were quantified in 1990 and 1991 during water deficits in a field lysimeter under an irrigated rain shelter. Seasonal changes in these parameters and in soil mineral and leached N were determined during 1991 in four field lysimeters subjected to conventional or no-till crop culture. Horizontal and vertical gradients in root intersections with 0.05 (I.D.) x 1.4 m polybutyrate tubes corresponded with transient deficits in plant water supply, and subsequent root proliferation during mid-vegetative growth stages. A horizontal complement to the vertical rooting front is characterized by exponential distributions of inter-root distances under a row crop. Geostatistical measures of clustered root distributions indicate spatial correlation up to 0.45 m at anthesis. Failure to consider depth-dependent gradients in root xylem potential most likely accounts for systematic bias in soil water' depletion. predicted by a simplified solution to a cylindrical model of root water uptake.‘ Soil plus root respiration is related to vertical root distributions, but vertical gradients in C02 and C02 flux fail to satisfy conditions for the steady state assumption. Increased N03-N retention in conventional till soil, relative to no-till soil indicates solute partitioning among mobile and immobile regions of soil water may be modified by historic tillage effects. Deviations in N03-N concentrations of leachate from seasonal trends coincide with extreme high or low drainage flux conditions, invalidating assumptions of homogeneous pore velocities and solute concentrations. Simulated N03-N leaching rates are sensitive to errors in predicted infiltration, canopy dimensions and drainage below the root zone, but are insensitive to reductions in maximum root length density. Managing soil-plant interactions for optimal productivity and solute retention requires accurate simulation of system behaviour when regulation shifts from atmospheric boundary conditions to soil system transport and transformation processes. to Ann ammmum A work such as this reflects the efforts of many workers. The vision 'of Frank D'Itri and Boyd Ellis led to the financial support from the USDA National Needs Graduate Fellowships Program (Grant #88-38420-3834) that enabled my post graduate study here. My graduate committee, Drs. Boyd Ellis, Jim Flore, Ken Poff and Joe Ritchie gave excellent support throughout this process. The quiet and consistent support of my spouse, Ann Miner, and the loving patience of our children, George and Stuart, provided the foundation for completing my commitment to this program. Dr. Ritchie graciously offered the field lysimeter under the rain shelter for data collection that provides much substance in this work. Abelardo and Liliana Nunez- Barrios gave me an early start at this lysimeter. Scott NeSmith’s early root observations provided a chance to learn about root spatial structure. I thank Reimar Carlesso for our struggles with neutron probe calibration. He; Mike Robertson; and Huang, Bihu helped with data collection when I had to be away. Jeff Hamelink's and Trish Sweanor's diligent and dedicated efforts account for much of the nitrate leaching, soil water and carbon dioxide data for the Agroecosystem field lysimeters at Kellogg Biological V Station. Dan ZRoper"worked ‘with. soil incubation studies. Sandy Halstead, manager of Phil Robertson's lab bore with my humor and hosted much of the lab analyses. Marty Rosek was always ready with a sound lesson in geomorphology. Jim Bronson and crew kept field operations in .line. Keith Paustian and Phil Robertson provided able on-the-spot advice. The model evaluation reported here would not be possible without the diligence of Brian Baer and co-workers. Their maintenance of a complex simulation program enables work such as reported here. Ed Martin, Jimmy Chou, Evangelyn Alocilja, Brad Johnson, Frederick Dadoun, Jon Lizazo, Aries Gerakis, and Reimar Carlesso all helped me learn the subtleties of CERES simulation. A special thanks are in order for my co-workers in Soil Biophysics. John Ferguson; Rose Soanes; Huang, Bihu; Zhou, Feng; Chuck Belanger; Stephanie Murphy; and Victor Vitorella all helped to lighten the load of a busy student. Your perspectives and advice will be remembered. Sue Ellen Johnson's Sustainable Ag discussion group helped keep me on track with broader implications of this work. Stimulating discussions with Dana Barclay, Mary Ann Bruns, Michel Cavigelli, Lisa Huberty, Dan Dettweiler, Gaye Burpee, Tim Lynam and Ernesto Franco-Vizcaino helped broaden my perspective on soils. The MSU faculty somehow manage to find time for students vexed by some problem. Drs. Fran Pierce, George vi Merva, Kay Gross, Ray Kunze, Sharon Anderson, Dick Harwood, Dave Harris, and Lee Jacobs provided a sounding board when in need. Phil Robertson helped ease my mind about geostatistic questions. Finally, I'd like to thank my major advisor, Al Smucker, for the open door and ready conversations that nurtured much of the insight gained from this experience. His standards of excellence and his gracious acceptance of what a student has to offer will continue to inspire my efforts in science and in life. vii PREFACE Stewardship of the earth is a theme passed down in Western and Eastern traditions. We are given a vision of the good steward in the Christian text " . . . to give them their portion of food at the proper time." Luke 12:42. And from the Taoist text Those who esteem the word as self will be committed to the world Those who love the world as self will be entrusted with the world Tao Te Ching, 13 Each of us must confront the age-old questions "How do we esteem the world? To whom or what do we give service?" This work seeks to clarify the limits of our understanding of complex systems. Such pursuit of knowledge is frequently justified by expectations of beneficial applications. Our ability to fulfill these expectations is a measure of our science, our social institutions, and our spirit. TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES INTRODUCTION CHAPTER 1 NITRATE LEACHING AND ROOT FUNCTION INTRODUCTION SYSTEMS THEORY AND SOLUTE TRANSPORT SOIL NITROGEN AND PLANT CARBON ALLOCATION IMPLEMENTATION CHAPTER 2 SEASONAL TRENDS IN MAIZE ROOT DISTRIBUTIONS INTRODUCTION METHODS AND MATERIALS Crop culture and biomass distribution Lysimeter instrumentation Soil water Canopy development Root distribution Geostatistic analysis RESULTS Seasonal trends Root spatial distribution Root separation distances DISCUSSION ix xi xiii 11 14 18 22 22 23 27 29 35 35 42 49 51 CHAPTER 3 REMOTE SENSING OF MAIZE ROOT FUNCTION INTRODUCTION METHODS AND MATERIALS Crop culture Lysimeter instrumentation Evaporative demand Crop development Soil water depletion C02 gradients and net 002 flux Analysis of root function RESULTS Canopy and root development, and evaporation Soil water depletion Carbon dioxide evolution DISCUSSION ¢2CHAPTER 4 ROOT FUNCTION AND NITRATE LEACHING INTRODUCTION METHODS AND MATERIALS Crop culture Lysimeter instrumentation Weather Soil conditions Common scenario analysis RESULTS Crop development and soil C02 Soil water ‘ Soil nitrate N leaching losses DISCUSSION SUMMARY AND CONCLUSIONS REFERENCES APPENDIX 56 58 53 58 so 62 . 65 57 69 7o 70 ‘76 81 85 140 146 154 \\ LIST OF TABLES Table 2.1 Maize cultural practices for lysimeter in rain shelter containing Spinks sand at Kellogg Biological Station Table 2.2 Day of year for Maize phenologic development for lysimeter in rain shelter containing Spinks sand at Kellogg Biological Station Table 2.3 Semivariance analyses of detrended data for maize root spatial structure within a Spinks sand under a rain shelter at the Kellogg Biological Station Table 3.1. Maize cultural practices for lysimeter in rain shelter containing a Spinks sand at the Kellogg Biological Station. Table 3.2 Day of year for Maize phenologic development for lysimeter in rain shelter containing Spinks sand at Kellogg Biological Station Table 4.1 Lysimeter horizons and NO3-N status on day 123, 1991 of Kalamazoo loam soil at Kellogg Biological Station. Rye biomass removed on day 135, 1991. xi 24 24 46 59 59 93 Table 4.2 Soil NO3-N distribution on day 122, 1991 108 in field lysimeters with non-disturbed Kalamzaoo soil at Kellogg Biological Station, as determined by suction lysimeters. Soil NO3-N status of AP horizons estimated by leaching observed in Bt horizons after day 122. xii LIST OF FIGURES Figure 1.1 Flow chart of the functional relations of 17 root distributions as addressed in Chapter 2 (2), root function in Chapter 3 (3), and flux of water and nitrate in Chapter 4 (4); which illustrate the hierarchical structure of root analysis employed in this work. Figure 2.1 Instrumentation for repeated measurements 25 of root distributions and soil water in a non-weighing lysimeter under an irrigated rain shelter. Figure 2.2 Distributions of maize roots with respect 37 to soil depth (A) 1986, water deficits during the reproductive phase, (B) 1990, water deficits during the vegetative phase, (C) 1991, water deficits during the vegetative and reproductive phases. Figure 2.3 Seasonal trends in maize root development 38 (A) 1986, water deficits during the reproductive phase, (B) 1990, water deficits during the vegetative phase, (C) 1991, water deficits during the vegetative and reproductive phases. xiii Figure 2.4 (A) Maize rooting front detected by soil 40 water depletion and MR root intersections and (B) available water stored in rooted soil and in soil underlying the root zone for water deficits during 1991 vegetative and reproductive phases. Figure 2.5 Maize root proliferation rates for water 41 deficits during vegetative and reproductive phases in 1991 (A) 0.5 to 0.9 m depths, (B) 1.07 to 1.6 m depths. Figure 2.6 Maize root distributions in a Spinks sand 43 derived from exponential trend and semivariograms for 1991 water deficit during the vegetative and reproductive phases (A) Eight leaf growth stage (Day 194), (B) Anthesis growth stage (Day 217), (C) Grain fill growth stage (Day 238). Figure 2.7 Maize root distributions in a Spinks sand 44 derived from exponential trends and semivariograms for 1986 water deficit during the reproductive phase (A) Eight leaf growth stage (Day 191), (B) Anthesis growth stage (Day 212), (C) Grain fill growth stage (Day 238). Figure 2.8 Maize root distributions in a Spinks sand 45 derived from exponential trend and semivariograms for 1990 water deficit during the vegetative phase (A) Eight leaf growth stage (Day 191), (B) Anthesis growth stage (Day 212). xiv Figure 2.9 Cumulative freqency distribution of 50 distances separating maize root observations at 0.5 m and 0.72 m depths in a Spinks sand during periods of water deficits in 1991. Root observations were taken at 194, 217, and 238 days, which corresponded to 14, 37, and 58 days following the first intersection (DAI) of roots with MR tubes. Figure 2.10 Distribution of maize root separation 55 distances (RSD) in a Spinks sand, computed from maize root distributions observed at anthesis growth state (day 217) for 1991 water deficit during the vegetative phase. RSD computed as 2*(n'RN)'°'5, assuming uniform root distribution within 0.05 x 0.05 m cells in the soil profile. Figure 3.1 Instrumentation for repeated measurements ' 61 of root distributions, soil water, and soil atmosphere in a non-weighing lysimeter under an irrigated rain shelter. Figure 3.2 Maize canopy development on a Spinks sand 71 for water deficits during vegetative phase in 1990, (A) green leaf area index, (B) top ligule height. Maize canopy development for water deficit periods during vegetative and reproductive phases in 1991, (C) green leaf area index, (D) top ligule height. Dashed lines are 90% confidence intervals. Figure 3.3 Seasonal trends in maize root-soil water 73 interactions on a Spinks sand in 1991 for water deficit during vegetative and reproductive phases: (A) root development (B) available soil water (C) soil water depletion (D) predictive accuracy of a cylindrical root water uptake model. Figure 3.4 The ratio of mean daily evapotranspiration 75 (ETm) to mean daily potential evaporation (E0) is illustrated for maize subject to water deficits during vegetative and reproductive phases in 1991. ETm was determined by soil water depletion (N = 56 observations for each of two similar profiles). E0 is daily potential evaporation, computed from a modified Priestly-Taylor equation. LCI and UCI are 90% lower and upper confidence intervals for daily soil water depletion, integrated over sampling intervals. Figure 3.5 Comparison of soil water depletion, observed 77 by neutron thermalization (SWDO), with depletion due to root water uptake predicted by a simplified solution to a cylindrical model of root water uptake (SWDp). Observations under“maize in a Spinks sand for water deficits during vegetative stage in 1990. (A) Soil profile distribution of available water, (B) Soil profile distributions of root number, (C) Soil water depletion, predicted and observed for days 205-208, (D) Soil water xvi depletion, predicted and observed for days 208-214. Upper and lower 90% confidence limits are computed for N=8 observations. Figure 3.6 Seasonal trends in C02 source strength for 82- water deficit during the vegetative phase for maize on a Spinks sand in 1990 (A) seasonal trends in the soil profile C02 gradients, (B) seasonal trends in soil plus root C02 source strength, (C) seasonal trends in root distribution. Figure 3.7 Seasonal trends in C02 source strength for ' 83 water deficit during the vegetative and reproductive phases for maize on a Spinks sand in 1991 (A) seasonal trends in the soil profile C02 gradients, (B) seasonal trends in soil plus root C02 source strength, (C) seasonal trends in root distribution. Figure 4.1 Diagram of field plots established at the 90 Kellogg Biological Station in 1986 to investigate N supply and tillage effects on soil-plant interactions. Figure 4.2 Instrumentation ports for non-destructive 91 sampling of root distributions, soil water, soil solutes, and soil atmosphere above and below soil horizon interfaces in non-disturbed field lysimeters on a Kalamazoo loam soil. xvii Figure 4.3 Initial volumetric soil water (SWC) profile 104 on day 122, 1991, determined by TDR; lower limit to maize root water extraction (LL), determined by neutron thermalization; and drained upper limit (DUL), determined by TDR, for a Kalamazoo soil at the Kellogg Biological Station: (A) No-tillage (NT6) or (B) conventional tillage (CT2). Figure 4.4 Residual soil [N03-N] on day 122, 1991 for 107 field lysimeters containing non-disturbed Kalamazoo loam soil at the Kellogg Biological Station (A) no-tillage and (B) conventional tillage. Actual depths of each soil horizon are listed in Table 4.1. Figure 4.5 Root distributions in field plots 110 associated with non-disturbed lysimeters containing Kalamazoo loam soil at Kellogg Biological Station, and corresponding to fertility status of the lysimeters (A) no-tillage (NT) with no fertilizer, (B) moldboard plow and disc tillage (CT) with historic fertilizer supply. Figure 4.6 Maize canopy development in non-disturbed 112 field lysimeters subject to either no-tillage (A) green leaf area index, (B) top ligule height; or conventional tillage (C) green leaf area, (D) top ligule height. Solid lines are means of actual measurements, dashed lines represent 90% upper and lower confidence intervals. xviii Figure 4.7 Seasonal trends in vertical distribution of 114 soil C02 partial pressures in non-disturbed field lysimeters containing Kalamazoo loam soil at Kellogg Biological Station, subject to either no-tillage (A) NT6, (B) NT9; or conventional tillage (C) CT 2, (D) CT 13. Figure 4.8 Seasonal trends in soil water distribution 116 in non-disturbed field lysimeters subject to either no- tillage (A) NT 6, (B) NT 9; or conventional tillage (C) CT 2, (D) CT 13. Figure 4.9 The distribution of rain used to simulate 118 soil-plant effects on NO3-N leaching in field lysimeter on a Kalamazoo loam soil at Kellogg Biological Station. Records obtained from National Weather Service reporting station two Km west of lysimeters (days 123-153); and from data logger located with 200 m of lysimeters (days 154-275). Time-to- ponding (TTP) estimates of infiltration under CT based on soil water intake rates. Figure 4.10 Seasonal trends in predicted and observed 120 soil water content (SWC) for upper layer of Bt horizon of Kalamazoo soil under (A) no-tillage (lysimeter NT 6), or (B) conventional tillage (lysimeter CT 2) at the Kellogg Biological Station. Simulation runs used time-to-ponding (TTP) or curve number (CN) methods of estimating infiltration and RLDmax values of 5 or 2. xix Figure 4.11 Seasonal trends in soil NO3-N distributions 123 in non-disturbed lysimeter profiles subjected to either no-tillage (A) NT6, (B) NT9; or conventional tillage (C) CT2, (D) CT 13. Note difference in scale among figures. Figure 4.12 Seasonal trends in soil [NO3-N] observed 125 and simulated by CERES-Maize for lysimeter CT 2, conventional tillage treatment with high initial residual NO3-N. Simulated runs using time-to-ponding (TTP) and curve number (CN) methods of predicting infiltration, and RLDw values of 5 or 2 cm cm'3. Figure 4.13 Cumulative water and nitrate efflux 2.0 m 128 below the soil surface observed or simulated by CERES-Maize for non-disturbed field lysimeters subjected to either no tillage (A) cumulative drainage, (B) cumulative NO3-N leaching; or conventionnal tillage (C) cumulative drainage, (D) cumulative NO3-N leaching. Simulated runs using time to ponding and curve number methods of predicting infiltration, and RLDmEx values of 5 or 2 cm cm'3. XX Figure 4.14 Distribution of leacheate NO3-N 130 concentrations with respect to cumulative outflow volume from non-disturbed field lysimeters subjected to either no-tillage (A) NT 6, (B) NT 9; or conventional tillage (C) CT 2, (D) CT 13. xxi Nitrate leaching is of practical and theoretical interest. Loss of nitrate below the root zone poses a public health hazard and evinces poor nutrient retention. Nitrate leaching occurs throughout agricultural regions with sandy soils such as western Michigan (Ellis, 1988) and the karst region of Iowa and Minnesota (Hallberg, 1987). Surface waters are contaminated when sediments, phosphorus, nitrates and pesticides are carried off in surface runoff (Baker, 1988). Management of nitrate leaching is enhanced by knowledge of factors determining synchrony of N supply and plant uptake. A role for simulation in design and management of biological systems is demonstrated by applications to irrigation and bioreactor technology. Effective simulation of complex systems accurately predicts. state and output parameters of value to decision-makers for a bounded set of input conditions and system parameters (Manetsch and Park, 1987) . Nutrient retention in managed ecosystems is enhanced when knowledge of management and environmental effects on alternative fates of nutrients can be directly interpreted by decision-makers. Predicting transport of nitrate is confounded by heterogenous distribution of water and N in time and space, A 1 2 multiple transformations of soil N, and complex interactions among soil, plant, and weather factors. General solutions to equations relating nitrate leaching to management and environmental inputs require simplifying assumptions or detailed specification of the soil-plant-atmosphere system. Solutions to the 1N leaching' problem. provide insight to similar problems where biological functicn is directly and systematically related to transformations and transport of environmental toxins. . Nitrate leaching potential is largely determined by the synchrony of nitrate supply (nitrification yand fertilization) with root nitrate uptake (Robertson and Smucker, 1988). In. practice, fertilizer N is directly amenable to management. Mineralization of N in organic matter is subject to substrate quality, thermal and xeric constraints (Paul and Clark, 1989). Soil supply and root uptake of N during exponential vegetative growth is an important determinant of whole plant specific growth rate, as root and shoot growth are alternative sinks for assimilates (Kachi and Rorison, 1989). Root:shoot signals can regulate ‘transpiration and plant growth rates in response to soil dessication (Davies and Zhang, 1991) and soil hardness (Masle and Passioura, 1987). Whole plant growth and development is also modified by root proliferation in soil zones locally enriched in N (Granato and Raper, 1989); and by modification of nitrate reductase activity. Indeed, the specific activity of root uptake is a fundamental parameter required to simulate optimal allocation of C and N to plant organs for varying levels of radiation and C02 (Hilbert et al., 1991). Nitrate leaching potential is reduced and productivity maintained when nitrification rates coincide with root N uptake to maintain optimal shoot N concentration (Hilbert, 1990). Models of plant growth (Thornley, 1972; Thornley, 1991; Agren and Ingestad, 1987) and ecological succession (Tilman, 1988) are based on the premise of optimal C allocation with respect to soil water and nutrients. Partitioning of assimilates to root, leaf, stem and reproductive organs is optimal when relative growth rates are maximized for a given supply of water, nutrients, and radiation. Soil supply of water and nutrients are simulated in a family of models oriented to agroecosystems (Jones et al., 1986; Addiscott and Wagenet, 1985). Valid models of plant growth and soil supply of water and nutrients require accurate knowledge of root distribution and function. Root function, as a sorptive sink for water and nutrients, is an important element in the soil water and N balance, and as a determinant of plant growth and development. The classic models of root function assume uniform root distribution (Gardner, 1960; Barber and Silberbush, 1984); though spatial heterogeneity in root distributions is well known (Ogata et al., 1960) with corresponding effects on root function (Gardner 1964). Soil structure effects on root distribution and subsequent function, constraining the volume of soil explored by individual roots, is a subject of current research (Passioura, 1991; Jones, 1983) . Knowledge of root distributions in soil permits microanalysis of soil water depletion zones around individual roots (Lafolie et al., 1991), strengthening our understanding of management effects on soil N. Simulation of root function in natural and managed ecosystems complements knowledge gained by controlled experimentation. The former approach accounts for systematic variation by quantifying causal relations with systems analysis techniques (Manetsch and Park, 1987) . The latter approach identifies causal relationships by partitioning treatment and experimental sources of variation into elements of statistical models, subject to probabal istic interpretation . Indeed , fundamental biophysical phenomena, such as thermal modification of nodal root growth trajectories may underly agronomic treatment effects (Tardieu and Pellerin, 1991) . Knowledge gained by either approach can be transferred: directly by validated simulation models, or indirectly, via interpretations of valid principles. Empirical tests of either simulation or experimental approaches to understanding root systems require accurate quantification of root networks and associated activities. Effective methods should quantify changes in the distribution of roots, water, and nutrients in specified 5 soil volumes over specified time intervals. Such methods should be non-invasive, submit to acceptable data reduction techniques, and yield parameters that are pertinent to relevant scientific principles. .Advances in microvideo technology (Upchurch and Ritchie, 1983; Ferguson and Smucker, 1989) are field-scale techniques that meet these requirements, in situ. The following work addresses the hypothesis that minirhizotron techniques enable empirical tests of root distributions and activities that are relevent to water management. Analysis of root function is structured to address- a hierarchy of questions regarding root distribution, function, and relation to water management in Chapter 1. Spatial analysis of maize root distributions, with implications for root function and plant growth under water deficits, is reported in Chapter 2. Maize root sink strength for water and source strength for C02 in a field lysimeter are related to root distributions, soil and atmospheric conditions in Chapter 3. Finally, the relative importance of root distributions to errors in simulated N leaching is reported for a set of field lysimeters in Chapter 4. CHAPTER 1 NITRATE LEACHING AND ROOT FUNCTION 12139229119! Root function is systematically related to nitrate leaching. Flux of nitrate beyond the root zone depends on several physiochemical states of the soil profile just prior to an infiltration event. Root functions addressed in this study include: 1) distributed sink for water and N 2) determinants of plant growth and future demand for water and N 3 ) substrates for decomposition and net N mineralization, 4) soil formation factors, modifying soil structure and transport properties. These root functions principally affect the state of the soil profile, conditioning nitrate leaching potential rather than actual nitrate flux. It follows that analysis of root effects on N leaching requires consideration of soil-plant interactions in a systematic context. Scientific principles supporting a systems framework for soil-plant management of N are reviewed in the remainder of this chapter. Structured analysis of a N management system includes identification of performance criteria, functional relations of canopy and root architecture, and use of state equations to simulate mass flow. Systems theory indicates the output [Y(t)] of a system is related to input [X(t)] by the transfer function [G(t)] (Manetsch and Park, 1987) Y(t) = G(t) * X(t) [1.1] Thus, knowledge of the transfer function permits prediction of the state of the system at any time for a set of inputs having defined limits. A system is well defined when the transfer function can be written in the form of a state equation 1(t + 1) = Mt) * 1(t) + B(t) * Mt) [1-2] where A and B are matrices defining both system and environmental parameters which influence the system state output parameters, and 1(t) and x(t) are vectors and linear algebra operations apply. A valid state equation for a system describes the condition of the system at all times . when specified initial and boundary conditions are known. The relative significance of root function to nitrate leaching can be quantified by implementing the solute conservation equation for specific soil-plant systems. The equation of state for soil nitrate derives from the solute conservation equation, presented for vertical flow in a dual (mobile, immobile) phase water model (Jury et al., 1991). an,“ de dsz deNm em-u + em---- = D----2 - ------ - re [13} dt dt dz dz where am is the mobile phase of soil water, with corresponding solute concentration Nm; aim is the immobile phase of soil water, with corresponding solute concentratiOn Nim- D is the effective mean hydrodynamic dispersion coefficient, Jw is water flux, rs is solute reaction rate, 2 is soil depth and t is time. Change in solute concentration for these phases results from hydrodynamic dispersion, mass flow, and solute reaction. Sorptive processes can be readily considered by an additional term on the right side of equation (1.3]. Exchange of solutes between mobile and immobile water phases, analogous to macro and micropores, is defined by time-varying concentration gradients and a rate factor (a). eim SE22 = “(Nm ' Nim) [1.4] dt Soil supply and root uptake of nitrate are included in the solute reaction term. rs = rw*Nm + rrn + rmn [1-5] where rw is root water uptake, rrn is root active uptake of Nm, and rmn is nitrification rate. Nitrification follows net. mineralization of organic matter, which. generally' proceeds when. the C:N ratio of substrate falls below 25:1 (Paul and Clark, 1989). Root N uptake is defined by Michaelis-Menton kinetics (active 9 uptake) and solute concentration in the transpiration stream (passive uptake). The water conservation equation--including infiltration, redistribution, evapotranspiration, and drainage components--is implicit in the solute conservation equation. dam de --- = ---- - rw [1.6] dt dz Root water uptake (rw) relates to the capacity of soil water to meet evaporative demand of a plant canopy, as represented by conductivity and water potential gradients across the root-soil interface (Campbell, 1985). rw = ----------------------- [1.7] (l-n) ln(r2*RLD) where kr is root hydraulic conductivity, r is root xylem water potential, k8 is soil hydraulic conductivity, 3 is soil water potential, RLD is root length density, 2 is soil depth, n is a power of the hydraulic conductivity function, and r is root radius. A simplified solution for root water uptake, assuming a constant root-soil water potential gradient, and a general unsaturated conductivity function was derived by Ritchie (1985). 0.00264 exp (62(9v - LL)) rw = """""""""""""""" [1.8] 6.68 - ln RLD 10 where 9v is volumetric soil water content, and LL is the lower limit of volumetric soil water extractable by a specified crop for a specific soil. Equations [1.7] and [1.8] assume uniform root distributions, with cylindrical flow of soil water to the root surface (Gardner, 1960). Water flow is proportional to hydraulic gradients and cylindrical resistance, by analogy, to Ohms law. But root distributions are not uniform (Tardieu et al., 1988b), root:soil contact is not uniform (Kooistra et al., 1992), as required for cylindrical geometry, and water uptake in lower regions of the soil profile are lower than predicted from root length density (Gardner, 1991) . Knowledge of root distribution can help modify predicted effects of canopy evaporative demand on the soil water balance. Root function, root distribution and soil structure are intimately related (Passioura, 1991; Tardieu et al., 1988c). Large massive units of compacted soil, which restrict root penetration, maintain sharp hydraulic gradients between the interiors and surfaces of the pad when root penetration is restricted to clusters around these units (Amato, 1991) . Root networks also contribute to the formation of soil aggregates (Russell, 1977; Wang et al., 1986). Root elongation can promote continuity of soil macropores or biopores which promote the transport of water and solutes. These preferential flow paths (Ahuja et al., 1991) ; have 11 been reported for structured soil under maize (Warner and Young, 1991). Thus root networks interact with soil structural features to modify root function and soil transport properties. Transport of nitrate beyond the root zone results from interactions of factors determining the soil water and N balance. Root distributions and activities are important parameters for solving soil water, N, C, and energy balances. The relationship of root function, a distributed sink for soil water and N, with nitrate leaching can be tightly coupled, and necessitates the analyses of root distribution in time and space. Root effects on whole plant growth and development are of particular interest as changing root and canopy distributions determine plant 'demand' for soil water and N. N P O T O Theories of C partitioning in response to varying supply of growth factors, such as water and N, solve plant growth equations for partitioning coefficients that maximize relative growth rates (Hilbert, 1990). A general plant growth equation (Agren and Ingestad, 1987) relates primary productivity to shoot photosynthesis. dW --- = k as is w = Ps ' [1.91 dt where W is whole plant biomass, k is assimilate conversion 12 to biomass, as is specific shoot photosynthetic activity, fs is partitioning coefficient for shoot, and P8 is whole plant photosynthesis. Primary productivity is directly related to root activity (water uptake, or transpiration) under non- stress conditions, when adjusted for vapor pressure deficit (Tanner and Sinclair, 1983). P8 = --; ----- [1.10] where k' is normalized transpiration efficiency, T is transpiration, e* is saturated vapor pressure, and e is ambient ‘vapor’ (pressure. This relation indicates photosynthesis is affected by transpiration rates, as modified by relative Ihumidity conditions. But ‘we know maximal leaf photosynthesis rates approach a linear function of leaf [N] for C3 plants (van Keulen et al., 1989). Also, the fraction of assimilates partitioned to roots and shoots varies with plant nutritional status and soil N supply. Thus we must extend the analysis of plant growth to consider effects of soil [N] on root uptake and tissue [N] effects on assimilation and allocation rates. The effects of soil N supply on assimilation rates and partitioning fractions to root and shoot organs can be related to specific root uptake of N and specific shoot photosynthesis (Hilbert, 1990). Theory relating soil [NO3-N] and specific activity of root and shoot organs to C allocation requires simplifying assumptions, but indicates the effect of soil N supply on C 13 allocation to root and shoot growth. Biomass accumulation during vegetative growth is approximated by an exponential function when root and leaf organs are primary sinks for assimilates. When root and shoot tissue N concentrations (Np) are identical, the fraction of assimilate allocated to shoot (f8) is a function of specific root uptake of N (or) and specific shoot photosynthetic activity (as) (Thornley, 1972). f5 = .............. [1.123] where root and shoot are the sole sinks for assimilate, by difference we obtain the partitioning fraction for root (fr) fr = 1 - -------------- [1-12b] This assimilate partitioning relation indicates shoot growth is favored by high root uptake of soil N. As low soil [N] reduces root N uptake, more assimilates are partitioned to. roots. Analogous arguments can be made for plant growth responses to soil water availability‘ and evaporative demand. That is, shoot growth is favored when soil water supply is adequate; but root growth dominates under soil water deficits (Smucker et al., 1991). Predicted C allocation patterns under varying soil supply of water and N correspond to compensatory root and shoot growth . 14 frequently observed under controlled and field conditions (Russell, 1977). A root growth model' accounting for proliferation, senescence and extensive growth is presented in Hillel and Talpaz (1976). RLDj = (RLDj'l P t) - (121.03"1 D t) - (RLDj'l E t) [1.13] where RLDj is root length density at time 'j', RLDj"1 is root length density at time j-1. P is root proliferation in a soil layer, D is root death, and E is root elongation from soil layers above. P and E are modified by exponential declining functions of soil water. Quantifying root growth and function in ‘ relation to whole plant development is hypothesized to improve the accuracy of models simulating plant effects on soil N transformation and transport. INELEEENIAIIQH Solutions to the solute conservation equation [1.3]. (Jury et al., 1991) generally follow two approaches: 1) analytic solutions given simplifying assumptions 2) numerical solutions given system specification. Analytic solutions provide insight to general features of system performance including limits to system stability. Numerical solutions offer more realistic simulation of soil- plant conditions, validated by comparison with observed system behavior. Emphasis is given to the latter approach' 15 as a general method for testing hypotheses about system structure. Finite element solutions to solute transformation and transport problems can be implemented on mini-computers with much greater flexibility in specifying boundary conditions and system parameters than analytic solutions (Campbell, 1985). Numerical solutions to the solute conservation equation are conveniently structured into sub-components, or modules. When each module can be described by a governing state equation, the requisite boundary conditions, system and state parameters are clearly identified. Validation of structured numerical solutions can proceed at the module level, simplifying the process of error diagnosis by separating large complex models into smaller component parts. This step reduces the number of potential error sources, thereby aiding the "debugging" process of error diagnosis. The validation approach involves evaluation of the deviation of predicted parameters from actual observations. Modifications of system structure (as opposed to system parameters) may be justified when this deviation is reduced. A modular structure is a means of fracturing complex problems into simpler, coherent components for analysis, then knitting the interacting components together for the integrated solution. Root function is related to NO3-N leaching as an explicit term in the solute conservation equation and as a partial determinant of plant demand for water and N. The 16 range of conditions where N leaching is sensitive to root function can be analyzed. when soil-plant-atmosphere interactions are accurately' defined. To this end, the remaining chapters are directed. The solute conservation equation [1.3] suggests a hierarchical structure for analysis of root function and N leaching (Figure 1.1) . Knowledge of root distribution in space and time (equation [1.13]) is required for simulating root function, and is analyzed in Chapter 2. Root water uptake (equation [1.7]) is a necessary term in the soil water balance (equation [1.5]) which is evaluated in Chapter 3. Interaction of the soil water and N balance (equation [1.3]), illustrating the relation of root function to N loss bywleachingimisleyeluated in Chapter 4-' £312 17 J. - water Flux :”". Roo 2.2.x. dIstrIthlon Jch = F [Lch RnI} Lc - Lr - Rn - 1 . Jc - carbon Flux Valldatlon data Sotl - Plant - Atn. slnulatlon canopy dlst. root dlst. radlatlon InFIItratIon Alt. root Perfornance‘ varlables Figure 1.1 Flow chart of the functional relations of root distributions as addressed in Chapter 2 (2), in Chapter 3 (3), and flux of water and nitrate in Chapter 4 (4); which illustrate the hierarchical structure of root nodules Canopy (depth. tlnel Voter (depth. tune) NItrote (depth. tine) Leached N (tine) 2 error a b C Tine analysis employed in this work. root function SEASONAL TRENDS IN NAIEE ROOT DISTRIBUTIONS IEIBQDEQIIQN ‘Knowledge of root distributions in time and space is necessary for accurate predictions of root source/sink strength, regulation of transpiration and growth, and modification of soil structure and solute transport processes. The fraction of assimilate allocated to roots is an important determinant of relative growth rate and interspecific interactions such as nutrient acquisition and canopy light exclusion (Tilman, 1988) . Distributions of root and soil resistances to water and ion flow are-modified by variation in root clustering (Bruckler et al., 1991), diameter (Barber' and. Silberbush, 1984), and. suberization (Russell, 1977). Optimizing soil and water management and gaining inference of interspecific interactions are strengthened by knowledge of factors regulating root system development (Klepper et al., 1983). Root morphology appears to be controlled by the interactions among root tip meristems, localized soil environments, and transport of nutrients, assimilates, and growth regulators among root and shoot organs (Kuiper, 1987). Root proliferation and elongation rates are known to 18 19 respond to temperature, anoxia, mechanical impedence, water deficits, plant phenology, and the relative nutrient status of both plants and soil (Russell, 1977). These root growth responses include changes in diameter, elongation rates, branching frequencies, and permeability. Soil thermal and hydric conditions are known to modify the geometry of root distributions (Allmaras and Nelson, 1973; Kuchenbuch and Barber, 1988) . Higher inter-row soil temperatures stimulated root proliferation and branding into the warmer inter-row region in a growing season with lower soil termperatures, but not in a growing season with higher soil temperatures (Allmaras and Nelson, 1973) . Fortin and Poff (1990) demonstrated. a positive thermotropic growth response of maize roots. Curvature of root growth towards a thermal gradient were not related to passive thermal growth effects, which would favor growth away from warmer conditions. Tardieu and Pellerin (1991) found that mean soil temperature, during the 100 degree days after maize nodal root appearance, could account for differences in root- growth trajectory associated. with differences in sowing date, mulching, experimental site location and year treatments. Clearly, the horizontal component of root growth trajectories can be modified by soil thermal conditions. Root. and shoot. biomass are functionally linked. as alternative sinks for assimilate. The relative fraction of photosynthates allocated to roots and~shoots is subject to f1 he mi di ar. (10 19 am as. 1‘9; 195 20 genetic regulation, and environmental modification. Genetic advances in wheat are partially attributed to reduced carbon allocation to root organs, without change in water extraction, rooting depth, or water use (Siddique et a1. , 1990). Increased harvest index and water use efficiency of grain for modern wheat varieties were attributed to lower root:shoot ratios, relative to older wheat varieties. Soil water deficits are known to increase the fraction of assimilate partitioned to root organs (Russell, 1977, Zhang and Davies, 1991); though extended water deficits reduce leaf expansion, photosynthetic capacity and total root biomass relative to water sufficiency. Environmental factors such as soil structural characteristics can also modify root distributions (Passioura, 1991). Vertical distributions of root length ’density are frequently approximated by exponential functions of soil depth. This distribution results from the geotropic growth habit and lateral formation of root systems developing under minimal stress in homogeneous soils. The distribution of distance separating roots, an important determinant of ion and water absorption, is skewed, with exponential to log- normal distributions (Logsdon and Allmaras, 1991; Tardieu, 1988b) . Interactions of- clustered roots, with large gaps among clusters are neglected by models of root function that assume water and solutes traverse uniform distances to reach the root surfaces (Passioura, 1991; Lafolie et al., 1991). The geometric structure of root networks reflect the 21 plasticity of their growth responses to heterogenous soil conditions. Exponential distributions of distances separating roots provide positive evidence of clustered root distributions; for many roots within the cluster are separated by small distances, while large distances separate the clusters of roots. Root clusters can also be diagnosed by geostatistic techniques, such as spatial correlation. Positive spatial correlation indicates that the presence or absence of a root at a given location is positively correlated with root distribution in the locality. The absence of spatial correlation indicates that the presence of a root provides only random information about root distributions in the locality (Peck, 1983). Indeed, the presumption that root intersections at a plane of observation are correlated with similar root distributions in the bulk soil is a variation on the theme of spatial correlation. Non-destructive, repeated measurements of root spatial distribution can characterize seasonal changes in root clustering and inter-root distances. By sampling identical soil volumes over time, the statistical dilemma of relating spatially variable sampling locations is avoided. Indeed, characterizing the spatial structure of root distributions can yield insight to biophysical implications of root function (Lafolie et al., 1991). Geostatistic analysis also provides optimal spatial interpolation, for mapping spatial 22 distributions with known variance structure (Warrick et al., 1986). Seasonal and spatial trends in root distributions can be quantified by repeated measure of root intersections with minirhizotron [access tubes (Smucker, 1990). Concurrent determination of canopy development and soil water depletion permit synchronous analyses of root and shoot growth activity. The following research was conducted to characterize the spatial structure of root distributions and interpret the biophysical consequences. METHOD AND MATERIALS WW Maize (Zea mays L., hybrid Pioneer 3573) grown in a 1.4 x 1.4 x 1.8 m non-weighing lysimeter (Figure 2.1) under a rain shelter (NeSmith and Ritchie, 1992) was subjected to extended water deficits during the reproductive stages of development in 1986, and the vegetative growth stages in 1990 (Table 2.1, Ritchie et al., in preparation). A third data set was collected during’ water deficits extending through the vegetative and reproductive growth stages in 1991 (Table 2.1). A Spinks sand (Sandy, mixed, mesic Psammentic Hapludalfs) was packed into the lysimeter and associated 4.6 x 6.2 m field plot beneath the automated rain shelter. The bulk density and corresponding depth intervals of soil layers are 1.3 Mg m'3, 0.0 to 0.25 m; 1.27 Mg m'3, 23 0.25 to 0.3 m; 1.46 Mg m’3, 0.8 to 1.4 m; 1.49 Mg m'3, 1.4 to 1.8 m. The lysimeter was bordered by crap on all sides. Cultural practices are summarized in Table 2.1. Wise Polybutyrate minirhizotron (MR) tubes (0.05 x 1.4 m) were installed perpendicular to crop rows and parallel to the soil surface (Figure 2.1). Duplicate tubes at each depth provided access for micro-video cameras (Ferguson and Smucker, 1989) and neutron probe (Gardner, 1986) . Upper surfaces of the MR tubes are 0.50, 0.72, 0.90, 1.07, 1.27, 1.43 and 1.60 m below the soil surface. Paired stainless steel rods installed as vertical wave guides for time domain reflectometry (TDR, Topp et al., 1982) were installed in the surface 0.15 and 0.3 m of the soil at locations 0.3, 0.6, 0.9, and 1.2 m from the east edge of the lysimeter during 1991. No TDR wave guides were installed in 1986 nor in 1990. £211.13t1r Soil water depletion was determined by neutron thermalization at three to seven day intervals following initiation of water stress in 1990 and 1991. The neutron probe was inserted in two vertical tubes centered in the crop row and interrow in 1990, or in the horizontal MR tubes, in 1991 (Figure 2.1). Volumetric soil water contents were determined by computing the ratio of counts observed 24 Table 2.1 Maize cultural practices for lysimeter in rain shelter containing Spinks sand at Kellogg Biological Station Parameter 1990 1991 Planting date (Day of Year) 144 £3§8bVE 1" Water deficit initiation 179 179 Water deficit termination 220 -- 'Plant population (Plants mfz) 7.14 9.18 Nitrogen (g N mIZ) 20.0 12.0 Phosphorus (g P mIZ) 18.0 6.0 Potassium (g K m72) 6.0 6.0 Table 2.2 Day of year for Maize phenologic development for lysimeter in rain shelter containing Spinks sand at Kellogg Biological Station Phenological stage 1990 1991 Emergence A 152 156?’ 3 ‘ 5 / Eight leaf 191 194‘ 1"/ I"! :1 J Anthesis 212, 217 Grain £111 230” ' 238 25 Vertlcal Neutron Crag row / iIITIIIW E- \ J j __j I I IIIIIIII. ./7 TDR wove fl / guides I I I /’/ 51 I'll” fl / """ oi J ‘ / atmosphere [”[lil I! 95 9‘? access / :"H tube . ;; 14 ° / CD 0 ‘:.::3 N» a " :E g {3 / Neutron 1,3.- g / thermal tzottion T J 3: / sphere 0.50 C) C) // A flinirhtzotron 0.73 "9‘ O 6/ occeee tube __SL 0.90 ->!. O O . 1.07 -fi (3 C) 1.27 —> O O M 1 43' -9- he ~°¢ - x. o o y 1.60 -9 Q o o O _ o .2!" lies 1.0‘\ 1\115 0’ :2 South 1 ' 4” North as Prorlle ProFtle A Figure 2. 1 Instrumentation for repeated measurements of root distributions and soil water in a non-weighing lysimeter under an irrigated rain shelter. 26 at 0.15 m intervals in vertical or horizontal access tubes, to standard counts. This count ratio is related to volumetric soil water content by a field calibration procedure. The apparent dielectric constant, as determined by TDR, is related to soil water content by the Topp equation (Topp et al., 1982) ev = -0.053 + 0.0282*Ka -0.0005*Ka2 + 0.0000043*Ka3 [2.1] where Ka = (ct/L)2, t = (B-A)/(Vp*c), c is the propogation of an electromagnetic wave in free space (30 cm/nsec). L is the length of the wave guides inserted into the soil. VP is the experimentally determined propagation velocity of the cable. B is the distance of the reflected pulse from the pulse generator, taken as the tangent made by the zero and positive. slopes traced in the waveform. A is the distance separating the pulse generator from the junction of the TDR cable and wave guides, taken as the point preceeding the large negative slope in the waveform. Spatial and analytic components of soil water variability are computed as the standard deviation of replicate observations at identical depths. Functional soil properties describing the capacity of the soil to supply water can relate soil water content and root distributions to water deficits (Ritchie, 1985). The drained upper limit (DUL) water content was determined by constant soil water distribution following saturation, when 27 active roots were absent. The lower limit (LL) of volumetric soil water extractable by maize was determined after 73 days of drought stress, when the soil water distribution became static even though active roots were present. Available water (AWi) for soil layer i was computed as the difference between volumetric soil water content (evi) and the lower limit (LLi) AWi - 6V1 - LLi [2.2] The quantity of available water stored in a soil layer (SAWi) is the product of available water and layer thickness (dz). The quantity of available water stored in the rooted soil profile (SAWr) is the sum of stored water in all soil layers within the rooting depth, where zr is maximum depth of rooting. i=zr saver: I sawi [2.3] The quantity of available water stored below the active root zone (SAWV) is the sum of stored water in all layers from the rooting depth to the lysimeter base (zb). i=zb Sva= z: sawi [2.4] i=2r 933222.912212232n5 Canopy structure *was. determined for three representative plants from each of two rows. Date of leaf maturation and senescence, mature leaf length and width, 28 height to top mature ligule, and reproductive development phases were parameters selected to quantify green leaf area (GLA), and biomass of leaf and stem. Green leaf area is computed as a time (t) dependent function for each plant: i=m GLA(t) = 2 (Li x wi x 0.75) [2.5] i=s where L1 and W1 are the length and width of the "ith" leaf, m is the top mature leaf number, and 8 refers to the top senescent leaf number. Confidence intervals about the canopy green leaf area index (GLAI(t)) and height to top mature ligule (H(t)) are computed from GLA(t) and plant population. Root intersections with MR tubes were recorded six to ten times during the growing seasons by microvideo cameras (Circon in 1986 and 1990; and Bartz Technology Co. 650 Aurora Ave. Santa Barbara CA 93109 in 1991) . Video image dimensions for the Circon camera are 12 x 18 mm, for the Bartz camera, 13.5 x 18 mm. The longer frame dimension is perpendicular to the MR tube. Active roots were identified by high light reflectanCe, opaque appearance and structural integrity. Senescent roots observed prior to crop establishment were excluded from root counts as were translucent roots and roots with low reflectance. Root number (RN) values were determined by computing the number of root counts (N) per unit area (A) of the soil-MR tube interface, and converted to root length density (RLD) by the following relation, suggested by Upchurch and Ritchie (1983) 29 RLD = Nd/Ad [2.6] where d is the MR tube outer diameter. This relation assumes the mean length of roots intersecting the volume occupied by the MR tube would be equal to the outer diameter of the MR tube, if the soil were not displaced by the MR tube (Upchurch and Ritchie, 1983). MM The rooting front, e.g. initial root arrival at a given depth, was determined by the reduction of stored soil water at rates exceeding 0.05 mm day-1 for a 0.15 111 soil layer, or by the initial root observation at the soil-MR tube interface. Root proliferation (Rp) during a sampling interval was computed for each MR tube as the rate of change in RN with respect to time (t) - Rp = dRN/dt [2.7] Vertical distributions of mean root intersections observed for each horizontal MR tube were analyzed as a logrithmic function of $011 depth: RN = a + b*log(z) [2.8] where RN is root intersections cm"2 , z is soil depth (m), a and b are empirical coefficients. Vertical trends in root distribution, quantified by Equation [2.8] were used to adjust root observations for geostatistical analysis (described below). Seasonal trends in root distributions were quantified by multiple linear regression of RN on simple and interacting effects of day of year (DOY), (DOY)2, 30 and log(z). A second order model of seasonal trends in root distributions was used, reflecting declining root proliferation rates during reproductive growth stages after rapid root development during vegetative growth. The fraction of soil contained in gaps between roots, e.g. distances separating root observations, was determined for the upper MR tube. These root separation distances (RSD) were quantified relative to the time elapsed since the initial root arrival at the MR tube, e.g. days after interception (DAI) . Distributions of distances separating roots were computed as relative frequency functions of distance separating roots at the MR tube surface: Nrsd‘i) Frsd A0 [2.13b] where gamma(h) is the semivariance computed as a function of distance (h) separating paired observations, Co is the nugget variance, C is the structural variance, and A0 is the range of spatial correlation. Interpolation 'of non-random root distributions is presented by summing positional trends, as described by equation [2.8], and kriged residual errors, based on the theoretical model of the semivariogram. Root length density is predicted for blocks of soil (0.05 x 0.05 m) by adding RN’ predicted by the positional trend to the kriged value obtained from detrended data. The (0.05 x 35 0.05 m) dimensions of interpolated soil blocks provide a convenient estimation unit, intermediate between the minimum observation unit (0.012 m) and scale of soil water sampling (0.15 111). RESULTS 8 t do Root distributions of maize declined exponentially with depth in the Spinks sand for each of the three years (Figure 2.2A-C). Stratified root proliferation at 1.3 and 1.6 m depths, but not at 1.1 and 1.45 m depths modified the exponential distribution during reproductive development for drought stress conditions. Seasonal trends in root development are quantified by multiple regression of RN on simple and interacting effects of soil depth and day of year (Figures 2.3A-C). Extensive growth of the rooting front and root proliferation in certain soil layers are related to soil water depletion in harizons at greater soil depths. The rate of root vertical extensive growth into soil layers defines the rate of movement of the rooting front. The vertical position of the rooting front defines the lower boundary of rooted soil, used to quantify water stored in the rooted soil profile (SAWr, equation [2.3], and in soil underlying the rooted zone (SAWV, equation [2 . 4]) . Root proliferation, resulting from extensive growth of initiated root laterals, occurs in soil layers above the vertical 36 position of the rooting front. Root water uptake capacity for a soil layer is modified by the number and geometric distribution of root laterals intersecting that layer. Thus, the vertical position of the rooting front, and the geometric distribution of roots proliferating above the rooting front modify the quantity of water available to plants, and water uptake capacity of plant root systems. Root arrival at 1.1 m depth, was observed by the soil water depletion method (detected by neutron thermalization) 15 days prior to root intersection with the MR tubes for the 1991 season (Figure 2.48). The greater sensitivity to root arrival by the soil water depletion method is attributed to the larger cross sectional area (0.15 mx1.4 m) sensed relative to the microvideo camera (18 mm x 1.4 m). Since few roots are required to initiate detectable water removal (Passioura, 1991), the probability of detecting root-arrival should be proportional to the cross section of the soil surface detected, and subject to detection limits of the sampling system. This result is consistent with reports that soil water depletion occurs at slightly greater depths than observation of grain sorghum roots by triplicated soil core samples (42 mm i.d.) at depths exceeding 1.0 m (Robertson et al., in press). The rooting front progressed at a rate of 29 mm/day, from day 183 to day 210, in 1991 (Figure 2.4A), coinciding with soil water depletion in upper soil layers and increasing depth of water uptake (Figure 2.4B). For example, 37 0.8 . . . a 0.8 . - . . ; RN(212)- 0.241.1.010°Io¢(z) 111-0334 A . , ~ RN(226)= 03090310041312) 1150391 " ““9””9‘3'0'5‘2 W") “3‘03“ RN(191)=-0.043-1.119‘|ox(21 R'-0.993 ° “Nmz’=°3°2'°-‘°"‘°t(') " '“3‘ A 0 6 I .4 A 0 6 " - .3. 0 .3 0 E N E x \ \ O 2 \_ \ ‘3 \ \ \. \ o . Z 0'4 E \'\. u ‘ 2 0.4 *' \ \ -' V V~\ v \ :2 5 \ ‘\.‘ II (I) \ \ O O i \ \ \ ‘ x. 0 D 226 8 \\ O \ ‘‘‘‘‘‘ .y o \\ 01 \ \ T°~-~. m x \ o 0.2 r- . \ a 0.2 1:. O \\‘ q o \‘e onqmz 4“ Day 212 ‘ \ . ‘ Ex * ° 0 ‘s Day 191 1 1 I f J- 0 1 1 x J: :3 n 0.5 0.75 1 1.25 1.5 0.5 0.75 1 1.25 1.5 8011 Depth (m) Soil Depth (111) 0.8 . . . . x RN(194)-0.010-0.312‘Iog(z) R’-o.965 C o “mu-01:2 4936-1631:) R’sOJIB 11 - wm31-0.422-0.562-16;(z) R’so.291 A 0.6 r . ‘ '1 1w. . - C \ \‘x. U \ \.\ It ' \ \. O \ “\. E 0.4 i. \\ “.‘.“\ Day 238 w \ ~~~~~~ s ‘ . \ c r o \ \ m 0.2 '- 3 ’ \ \ \ . .1 . 9 Day 217 ‘ x \ 0 1 \\ Du] ° 0 1 ‘ r 11 1 0.5 0.75 l 1.25 1.5 Soil Depth (111) Figure 2.2 Distributions of maize roots with respect to soil depth (A) 1986, water deficits during the reproductive phase, (B) 1990, water deficits during the vegetative phase, (C) 1991, water deficits during the vegetative 'and reproductive phases. Roots (No. 01114) .0 \l LII 38 y—n Roots (No. cm-Z) Roots (No. cm'z) O 0: 8 1& 9.9. 9‘0 0 '2 4 £032.24” 050.75 Oflbe 1.25 1”va RN = 4.46 + 0.645: + 0.0000221»),z - 0.009$zlog(Day) 112 = 0.77 Figure 2.3 Seasonal trends in maize root development (A) 1986, water deficits during the reproductive phase, (B) 1990, water deficits during the vegetative phase, (C) 1991, water deficits during the vegetative and reproductive phases. 39 on day 199, 52 mm of stored available water were available in the 1.3 m of soil containing roots,corresponding to 50% of the maximum water storage capacity (IBSNAT, 1985). Loss of leaf turgor was observed at this phase of vegetative growth; this period corresponds with ' maximum root proliferation at the 0.5 m depth (Figure 2.5A). By day 210, root extension (indicated by soil water depletion, Figure 2.4A) to a depth of 1.5 m coincided with root proliferation at 0.7 m (Figure 2.5A) and 1.6 m (Figure 2.5B). Loss of leaf turgor and leaf rolling were observed on days 213 and 217 despite root water uptake from depths of 1.6 m prior to day 213, and 92 mm of available water stored in the rooted zone. These data suggest water supply to shoots was limited by root distribution in soil layers containing available water. The visual evidence of plant water stress while root were presence in soil layers ‘with available soil water contradicts the interpretation of SAWr (equation [2.3] as a measure of the quantity of water available to plants. The presence of roots in water-bearing soil layers was not sufficient to prevent loss of turgor, observed on days 199, 213 and 217. We infer plant water status was modified by the horizontal gradients in distribution of roots and water in the soil. Thus, transient plant water deficits can result from non-uniform distribution of root growth in soil layers during a drying cycle. The remainder of this chapter is devoted to methods for quantifying spatial variability in root distributions. 4O 0 w r A i i Soil water depletion ‘ A I 0 5 : 0 MR root intersection ‘ é . '5 a. > 4) O > 1 z: 1 ' i O . m p . ’ I Rooting Front ‘ 1.5 ~ " q . 1 f! L 1 180 210 240 270 Day of Year 10 - 1 - . E ’ B 8, 3 . . 52 1 co 3 6 . . 2 Rooted soil .0 J .2 '3 4- 1 . > r < I B 2 _ Below : 8 Rooted soil ‘. " 63 1 ' I 0 . ' - A 1 - - 180 210 240 270 Day of Year Figure 2.4 (A) Maize rooting front detected by soil water depletion and MR root intersections and (B) available water stored in rooted soil and in soil underlying the root zone for water deficits during 1991 vegetative and reproductive phases. 41 ”—7 0.04 >5 (O '0 {5, 0.03 " E Z 0.02 ~ .2 ‘5 15 (101 :5: ‘5 a? 0 . . . 180 210 240 Day of Year 0.04 . 9 0 w P O p—e I Root Proliferation (ARN cm-2 day-‘) Q 8 9 180 210 240 Day of Year Figure 2.5 Maize root proliferation rates for water deficits during vegetative and reproductive phases in 1991 (A) 0.5 to 0.9 m depths, (B) 1.07 to 1.6 m depths. 42 W The spatial distribution of roots can be quantified by 'kriging'. Application of this interpolation procedure involves computing a weighted average of root number (RN) observations from the 16 MR video grams which are nearest each 0.05 x 0.05 111 cell in the soil profile observed via horizontal MR tubes. Weighting factors are derived from a theoretical semivariogram (e.g. equation [2.13a and b]), obtained by geostatistical analysis. Vertical trends in RN, quantified by equation [2.12], are added to kriged values derived from detrended MR observations to obtain the maps of root distributions presented in Figures 2.6 - 2.8. Horizontal and vertical gradients in RN are clearly indicated for mid-vegetative growth at eight leaf stage, Figure 2.6A. Spatial patterns of the maize root system subjected to water deficit conditions, were dominated by root system accumulations below the rows. Initial rootextension during the eight leaf stage occurred under crop rows, resulting in few root clusters in the inter-row region which was largely unexplored by roots. Root system development increased RN under crop rows by anthesis, and expansion into the region between rows, reducing the volume of soil not explored by roots. Clustered root distributions beneath the rows persisted through the. grain fill growth stage, although additional soil exploration by the expanding root system and root senescence resulted in more uniform soil exploration with time. 43 - 1? - 't’ >’ Roots (N0. cm'z) . - F . 9.1 is;\\&\ b V' g ' 1' a: Roots (No. cm‘z) I. 9 'I 9; - 0 ,1:§~:.2:._.':§::.:o,~' ; 1:47;: ”$95691 ‘ 1.1 ,. I . 905“ 30“ r. . 1.. . ‘ \l/I 1 1111‘ ’ 1' .‘i . (’7 0.5-1 ‘ ‘ ,2; “1 V", I, ‘ 1 ‘ .Ubflvfit 111:” Q °J 11;? [139’11‘1/"1 " ll ’ S 1 m1 01] be 1.5 0 \Y'og‘u D \ Figure 2.6 Maize root distributions in a Spinks sand derived from exponential trend and semivariograms for 1991 water deficit during the vegetative and reproductive phases (A) Eight leaf growth stage (Day 194), (B) Anthesis growth stage (Day 217), (C) Grain fill growth stage (Day 238). 44 B 2 4‘ 'E 5. f; 15 11111111 ~ ‘ 11 1'11 1' E as: 05 ’ )“V/ 0 ood\hi n; . . fi‘ .&WLEMWZQ;5 ogpfigéflwfik 1511/1 1// JW MW )1 1, Mir/ii) Roots (N0. cm'z) Figure 2.7 Maize root distributions in a Spinks sand derived from exponential trends and semivariograms for 1986 water deficit during the reproductive phase (A) Eight leaf growth stage (Day 191), (B) Anthesis growth stage (Day 212), (C) Grain fill growth stage (Day 238). -2 Roots (No. cm ) "‘ N ..'. " l." .\ o - . ‘/ N 6. 3.6: g .“S.\ s\\ '. Q‘. ..‘ 0‘ .(."..' . .‘ Q . C .. ‘ . x O H . s . H‘. .n ‘ § , " ‘.‘ ‘3 $33307 .9 ‘5"! H o o... ‘ ‘ "I .0 O. .0.“ ‘t' ‘ .' H ‘ ' | . . 0‘ 0'. I . ~ ‘. O "'1: 1.1:? . .0 . a... ‘ g... b. Roots (No. cm‘z) Figure 2.8 Maize root distributions in a Spinks sand derived from exponential trend and semivariograms for 1990 water deficit during the vegetative phase (A)‘ Eight leaf growth stage (Day 191), (B) Anthesis growth stage (Day 212). 46 The spatial structure of detrended root observations exhibit clear seasonal characteristics that are consistent among water stress conditions (Table 2.3). Variability of root distributions—-estimated by structural variance-- increased during root proliferation, associated with plant developmental phases. The trend of increased spatial heterogeneity is clearly indicated by an eight-fold increase in structural variance from the eighth leaf growth stage through grain fill growth stages during the 1991 season. Table 2.3 Semivariance analyses of detrended data for maize root spatial structure within a Spinks sand under a rain shelter at the Kellogg Biological Station Growth Theoretical ----- Variance ----- Coef. of Stage Model Nugget Struct. N/S Range Det. (R2) 1986 Eight Leaf Linear 0.142 -- -- -- 0.091 Anthesis Spherical 0.232 0.282 0.82 0.45 0.256 Grain Fill Spherical 0.274 0.362 0.76 0.31 0.679 1990 Eight Leaf Exponent. 0.021 0.034 0.62 0.33 0.325 Anthesis Spherical 0.091 0.156 0.58 0.31 0.584 1991 Eight Leaf Spherical 0.018 0.029 0.62 0.79 0.312 Anthesis Exponent. 0.086 0.139 0.62 0.30 0.358 Grain Fill Exponent. 0.121 0.243 0.50 0.21 0.779 N/S is the ratio of Nugget Variance to Structural Variance. 47 Structural variance also increased with plant developmental phases in the 1986 and 1990 seasons. The positive correlation of structural variance with plant and root developmental phases is consistent with field observations of seasonal increases in spatial variability of root distributions. Increased root spatial heterogeneity is matched by seasonal decreases in the range of spatial correlation. The range of spatial correlation extended to 0.79 m at the eight leaf growth stage in 1991. This distance roughly corresponds to the distance separating root clusters under crop rows, and the absence of roots under the furrow (Figure 2.6A). By time of anthesis, the range of spatial correlation decreased to 0.3 m, a distance corresponding to half the row spacing distance--equivolent to soil zones below crop row and furrow. The ratio of nugget to structural variance, another indicator of spatial correlation, decreased slightly from anthesis to grain fill growth stages in all three years (Table 2.3). This positive indication of spatial correlation corroborates visual evidence of clustered root distributions (Figures 2.6-2.8). This result indicates the assumprion of uniform root distribution is no more valid after-root proliferation than earlier phases of root system development. Models pertaining to soil water and nutrient managment can gain accuracy in simulated root function by considering effects of clustered root distributions on uptake of water 48 and solutes. Root quantification under field conditions can be improved by ensuring sampling distances exceed the range of spatial correlation expected for the respective root development phase. The structural variance reported here can be interpreted as spatially independent estimates of variance in RN at three stages of phenological development. Such estimates are useful for determining the number of sampling and experimental units required for detecting experimental treatment effects (Steel and Torrie, 1980 pp. 164-166). Diagnosis of vertical trends in mean RN during vegetative growth is confirmed by lower RN values with respect to soil depth (Figures 2.6A, 2.7A, and 2.8A). The magnitude and spatial heterogeneity of RN increased during early and late reproductive growth (Figures 2.63-c, 2.7B-c, and 2.88) . This seasonal trend towards increasing heterogeneity is attributed to the sensitivity of root proliferation to locally variable soil conditions. Geostatistical analysis confirms our expectation of clustered root distributions, for spatial correlation, a diagnostic for clustered distributions (Vieira et al., 1983), is indicated by nugget to structural variance ratios of less than 1. «The range of spatial correlation observed at reproductive growth stages, 0.2 to 0.4 m (Table 2.3), indicates the the distance from any given, or reference video frame where the probability of root intersection approaches a random. distribution. Within this range of 49 spatial correlation, the frequency of root intersections are expected to be more similar to the value observed in a reference frame. Thus, video frames with root intersections are likely to be neighbored by frames that also record root intersections. The corollary also holds: video frames adjacent to a frame devoid of roots are also likely to lack root intersections. Confirmation of horizontal and vertical root clustering indicates that the range of distances which influence the diffusive resistance to water and solute flow to root surfaces is not uniform, as frequently assumed in solutions to cylindrical flow models of root water uptake (Gardner, 1960, Ritchie, 1985; Passioura, 1991). These and additional concepts regarding the application of heterogenous root distributions to simulations predicting nutrient and water uptake require rethinking of the source codes used for planning soil and water management strategies. 0 t 1 During the early phases of root proliferation, root separation distances (RSD) are large at soil depths from 0.5 to 0.72 m (Figure 2.9). Two weeks after the first intersection of roots with MR tubes (14 DAI) 30% of the MR transects was contained in RSD lengths of less than 0.03 m, 34% of the soil transects ‘was included in. RSD lengths ranging from 0.03 to 0.18 cm, and RSD lengths exceeding 0.18 m comprised the remaining fraction of the transects. Five 50 o/ ‘ / 371m ,/ ' I/ d o/‘ -l o'/' '/ /. 141w 10 15 20 25 Root Separation Distance (cm) Cumulative Fraction of Soil Transect Figure 2.9 Cumulative freqency distribution of distances separating maize root observations at 0.5 m and 0.72 m depths in a Spinks sand during periods of water deficits in 1991. Root observations were taken at 194, 217, and 238 days, which corresponded to 14, 37, and 58 days following the first intersection (DAI) of roots with MR tubes. 51 weeks after root interception (37 DAI), all RSD lengths were less than 0.12 m with 70% of the soil containing roots within 0.02 m of each other. Eight weeks after root interception (58 DAI), 90% of the soil contained. roots within 0.03 m of each other. We conclude that the distribution of distances separating roots is exponential during initial root emergence at a horizontal plane, with degree of skew decreasing following root proliferation and extensive growth. DISCUSSION The declining exponential distribution of roots through the soil profile is consistent with a simple conceptual model of root system development. dz/dt = k1 [2.146] dx/dt = k2 [2.14b] dRN/dt = k3 [2.14c] where dz/dt is the mean vertical displacement of a root tip, dx/dt is mean horizontal displacement, dRN/dt is mean rate of root proliferation, k1, k2 and k3 are constants, or variates subject to genetic and environmental modification. Thus, the notion of a rooting front is extended horizontally and includes a description of-root proliferation sequences. RN decreases with. depth. at. a given time interval, and increases with time, for each depth. The distribution of root separation distances (RSD) is conditioned by a mean horizontal displacement, resulting from geotropic root 52 trajectories, planes of weakness within the soil matrix, and weather conditions. Time dependent changes in the root proliferation front are characterized by increasing frequencies of root intersection (with the horizontal plane, and decreasing frequencies of inter-root distances for a given soil depth. The horizontal dimension of root proliferation defines the horizontal limit of the volume of rooted soil. Localized soil water depletion at the root proliferation front can result in transient water stress, altering stomatal resistance, c assimilation and allocation patterns, root initiation frequencies and formation of suberin layers in aging roots. The geometry of root proliferation, and corresponding function, is constrained by plant stem distributions, root branching patterns, and environmental modification of root tip growth trajectories. Uniform root trajectories would result in constant horizontal displacement of vertical laterals relative to the parent root, but the distribution of distances among laterals would be non-uniform for roots' with identical trajectories but varying positions of initiation along the parent root. Since initial roots in rooting proliferation are expected to -develop directly below the stem, initial distribution of roots should correspond with clusters under stems and distances between clusters dimensionally similar to distances separating plant stems i.e. row spacing. This initially results in a skewed distribution of inter-root 53 distances. As root proliferation proceeds, the population of roots at a given depth increases and gaps between roots decrease. This conceptual model of root system development suggests the fraction of soil contained in large gaps between roots is greatest at the earliest stages of root proliferation and agrees with values measured in this study, (Figure 2.9). Mean inter-root distances decrease with root proliferation and horizontal displacement, especially in the more shallow soils. Distribution of soil contained in classes of root separation distances (RSD) is skewed during the initial phase of root elongation, which could be described by an exponential function of decreasing RSD but approaching a normal distribution subsequent to greater root proliferation. This shift in the distribution of RSD is consistent with Erlang distributions, applied to simulation of time delay processes (Manetsch, 1976). As the frequency of roots increases, the distribution of RSD can also shift from exponential towards normal distributions. The skewed distribution of roots, counterpart to skewed distribution of distances separating roots, also follows from a geotropic model of root system distribution. As roots proliferate, distances between pairs of roots become more uniform. It follows that simplifying assumptions of homogeneous root distribution are most erroneous during intermediate phases of root proliferation. Failure to consider the distribution of distances separating spatially 54 variable roots (Figure 2.10), will confound estimates of water and nutrient uptake, including diffusion of immobile ions to root surfaces, and interactions among roots. A conceptual model of the horizontal displacement of a root lateral from its parent branch can be derived from a combination of the genetic control of root morphogenesis, which is modified by soil environments and the geotropic tendencies of root growth. The growth trajectory of a root lateral is a product of elongation rate and curvature towards gravity. In the simple case of a homogeneous soil and a horizontal lateral root bud orientation, RSD increase with elongation rates and decrease with curvature rates. Plants that 'optimize' RSD, relative to water and ion absorption, are likely to synchronize lateral initiation with soil conditions resulting in 'optimal' growth trajectories of root tips. Simulations of root function are improved when algorithms correspond to conceptual advances in analysis of root function. Gardner (1991) developed the concept of 'water uptake front', which corresponds to the root proliferation phase of development discussed here. Passioura (1985) derived a simplified root water uptake relation, lumping root distribution and hydraulic conductivity terms into time constant describing exponential decline in soil water depletion. Soil structural heterogeneities and corresponding root clusters can alter the time constant for root water uptake by an order of magnitude (Passioura 1991). 55 structure, and associated soil water depletion by analysis of MR techniques can be used to evaluate these relations, leading to optimization of soil-plant interactions. 254 Root Separation Distance (cm) Figure 2.10 Distribution of maize (root separation distances (RSD) in a Spinks sand, computed from maize root distributions observed at anthesis growth state (day 217) for 1991 water deficit during the vegetative phase. RSD computed as 2*(n'RN)'°'5, assuming uniform root distribution within 0.05 x 0.05 m cells in the soil profile. CHAPTERS REMOTE 8.8186 01' £001! MOTION 18139229219! Applications of root functions to management systems are constrained by our knowledge of heterogenous root distributions. Spatial and temporal variation in root networks restrict many inferences of root function gained by destructive root sampling methods. In situ and simultaneous analysis of root distribution and function can improve our understanding of soil-plant interactions. Root networks function as sorptive sinks for water and nutrients, regulate plant water use, communicate with shoots, and may contribute to growth optimization. Soil carbon dioxide (C02) evolution is related to root growth and maintenance respiration and to microbial decomposition of organic carbon (Hall et al., 1990). Root respiration can amount to 40% of asSimilate partitioned to roots (Martin, 1987), and account for significant errors in micrometerological estimates of photosynthesis. Physical models of root function frequently assume uniform root diameter and inter-root distances, though these parameters are log-normally distributed (Tardieu, 1988b; Logsdon and Allmaras, 1991). Non-uniform distributions of roots, water, 56 5'7 and nutrients violate assumptions required for models of radial transport of water and ions to single roots (Passioura, 1991; Luxmore and Stolzy, 1987). Simultaneous, in situ, and repeated measurements of root distribution and function are necessary to quantify the effects of root morphology on root function. Minirhizotron (MR) imaging technology permits cost-effective, non- destructive, repeated field measurements of root and soil morphology (Upchurch and Ritchie, 1983) . MR root observations are systematically' related to conventional, destructive measurements of root length density for soil depths below 15 cm when access tube orientation is greater than 45° relative to vertical (Upchurch and Taylor, 1990). Quantifying the relation of MR root observations to root function can strengthen field analysis of soil-plant interactions, especially when coordinated with measurements of additional parameters. We hypothesize that biophysical models of root function can relate soil water depletion and 002 partial pressure gradients to root distributions derived from root intersections with MR tubes. Water uptake is a critical root function that is readily determined by neutron thermalization when water infiltration and redistribution are absent. Carbon dioxide source strength within the soil profile can be {derived from [C02] and soil water (6v) distributions by the flux gradient method (de Jong et al., 19745; Campbell, 1985). This study was conducted to 58 determine the functional relation of MR root observations, neutron probe estimates of soil water and microtube gas sampling of soil C02 with sink strength for water and source strength for C02. ' among m umnms 922231113311! Maize (Zea mays, L. hybrid Pioneer 3573) grown in a 1.4 x 1.4 x 1.8 m non-weighing lysimeter (Figure 3.1) under a rain shelter was subjected to water deficits during vegetative growth in 1990 and during vegetative and reproductive growth in 1991. A Spinks sand (Sandy, mixed, mesic Psammentic Hapludalfs) was packed into the lysimeter and the associated 4.6 x 6.2 m field plot. The bulk density and corresponding depth intervals of soil layers are 1.3 mg m‘3, 0.0 to 0.25 m; 1.27 Mg m'3, 0.25 to 0.8 m; 1.46 Mg m’3, 0.8 to 1.4 m; 1.49 Mg m'3, 1.4 to 1.8 m. Crop culture is summarized in Table 3.1 and water supply is illustrated in Figure 3.33. W Polybutyrate tubes (0.05 x 1.4 m) were installed perpendicular to crop rows, parallel to the soil surface, with upper surfaces at depths of 0.50, 0.72, 0.90, 1.07, 1.27, 1.43, and 1.60 m. Duplicate HR tubes at each depth provided access for recording root intersections at soil- tube interfaces with a microvideo camera (Ferguson and 59 Table 3.1. Maize cultural practices for lysimeter in rain shelter containing a Spinks sand at the Kellogg Biological Station. Parameter 1990 1991 Planting Date (Day of year) 144 156 Emergence Date (Day of year) 152 163 Plant Population (plants m'z) 7.14 9.18 Nitrogen (g N m'z) 20.0 12.0 Phosphorus (g P m'z) 18.0 6.0 Potassium (g R m'z) 6.0 6.0 Table 3.2 Day of year for Maize phenologic development for lysimeter in rain shelter containing Spinks sand at Kellogg Biological Station Phenological stage 1990 1991 Emergence 152 156 Eight leaf 191 194 Anthesis 212 217 Grain fill 230 '238 60 Smucker, 1989) and sensing soil water with neutron probe (Gardner, 1986). The horizontal minirhizotron (MR) tubes were offset by 0.15 m such that a minimum of 0.3 m separated vertically adjacent MR tubes. Aluminum access tubes installed vertically in the crop row and furrow provided additional access for neutron probe determinations of soil water in 1990. Paired stainless steel rods installed vertically to 0.15 and 0.3 m soil depths served as parallel wave guides for determination of soil water content by time domain reflectometry (TDR, Il'opp, 1986) in 1991. Teflon capillary tubing, 0.5 mm dia. and 0.3, 0.6, 0.9, and 1.2 m in length, were fitted with septa and attached to the MR access tubes, providing access to the soil atmosphere in the regions visible by the MR tubes and microvideo camera. Replicate soil atmosphere microtubes were also installed vertically to depths of 0.03, 0.08, 0.15 and 0.3 m at positions 0.3, 0.6, 0.9 and 1.2 m from the east wall of the lysimeter. A weather data system (LI-12008, LI-COR, PO Box 4425, Lincoln, NE 68504), located within 20 m of the lysimeter, recorded daily global radiation, maximum and minimum temperature, and precipitation. W Potential evaporative demand (E0) is computed daily by a modified Priestly-Taylor equation (Ritchie, 1985) as a function of global radiation (R9), daily ambient temperature extremes (Tmaxr Tmin’v leaf area index (LAI) and soil 61 Vertlcol Neutron Crop row\/.i IIIIITV/ / I I I ' TDRdweve ////[l ll Ii r1 ///7 gut ee ’// 2<' L" I.' ' " fl /..-- atmosphere III I III II / Q Lube ’sz 3;;g; *x{”»*' x” w 9 l H1d=0 I‘°S . 3 /Neutron E / (D C) ///,R\ ’ Alnlrhlzotron 0.72 ‘9 C) 65/ ecceee tube ._jZ_ . ‘9 . 0 90 C) C) _) I.27 -9 C) C) o 1.43 —> ‘ a9 g0 o o \. .~ g? I a? I QF;» /EH .2 .35 1.0 1J5 *3‘ 1 4n South ' North as Profile Prorlle “ Figure 3.1 Instrumentation for repeated measurements of root distributions, soil water, and soil atmosphere in a non-weighing lysimeter under an irrigated rain shelter. 62 albedo. Mean E0 corresponding to soil water sampling intervals is computed from daily E0. Daily radiation, used in computation of E0, is adjusted for extended periods of rain shelter closure, when required for maintenance. The adjustment procedure is based on the assumption that diurnal radiation is distributed over a 14 h period as a sine function of daylight hours, Rn' = Rn I sine 2*pi(t-5)/28 at {3.11 where Rn' is adjusted net radiation, Rn is net radiation, t is time of day (EST), sunrise occurs at 05:00 EST and sunset occurs at 19:00 EST. Integrating and evaluating at time of opening (to) and closure (to) gives the equation Rn'= Rn*[(cos 2*pi(to-5)/28) - (cos 2*pi(tc-5)/28)] [3.2] Potential evaporative demand computed from weather data is used to verify the accuracy of evapotranspiration determined by the soil water balance method. QIQR_§£!2122!225 Canopy structure was determined for three representative plants from each of two rows. Date of leaf maturation and senescence, mature leaf length and width, height to top mature ligule, and reproductive development phases are used to quantify green leaf area (GLA). 63 Green leaf area is computed as a time (t) dependent function for each plant GLA(t) = Li x W1 x 0.75 [3.3] 01MB where Li and W1 are the length and width of the "ith" leaf, m is the top mature leaf number, and s refers to the top senescent leaf number. Confidence intervals (Steel and Torrie, 1980) about the canopy green leaf area index (GLAI(t)) and height to top mature ligule (H(t)) are computed from GLA(t) and plant populations. Root intersections with MR tubes were recorded five times, by a Circon microvideo camera in 1990, and eight times, by a Bartz microvideo camera in 1991. Frame dimensions for the Circon camera are 12 x 18 mm, for the Bartz camera, 13.5 x 18 mm. The longer frame dimension is perpendicular to the MR tube. Active roots were manually identified by their high light reflectance, opaque appearance, structural integrity and diameter exceeding 75 um. Senescent roots observed prior to crop establishment were excluded from root counts as were translucent roots and roots with low reflectance. Root number (RN) values were determined by computing root counts (N) per unit area (A) of the soil-MR tube interface, and were equated to root length density (RLD) by the following relation (Upchurch and Ritchie, 1983) ' RLD = Nd/Ad [3.4] where d is the outer diameter of the MR tube. This relation assumes the mean length of roots intersecting the volume 64 occupied by the MR tube would be equal to the outer diameter of the MR tube, if the soil were not displaced by the MR tube (Upchurch and Ritchie, 1983). The rooting front, e.g. initial root arrival at a given depth, is determined by reduction of stored soil water at rates exceeding 0.05 mm day"1 for a 0.15 m soil layer, or by the initial root observation at the soil-MR tube interface. Root proliferation (Rp) during a sampling interval is computed for each MR tube as the rate of change in RN with respect to time RP = dRN/dt [3.5] Vertical distributions of mean root intersections observed for each horizontal MR tube are analyzed as a logrithmic function of soil depth RNZ = a + b*log(z) [3.6] where z is soil depth, a and b are empirical coefficients. This relation reflects the exponential distribution of roots that results from branching patterns. Root distribution, in the horizontal direction perpendicular to crop row, is interpolated among MR observations in the 0.5 to 1.7 m soil profile by a geostatistical model and a block 'krige' interpolation routine (GS+, Gamma Design, Box 201, Plainwell, MI 49080). Kriged values are computed as a weighted average of 16 nearest observations. The weighting function is based on the spatial structure. of ‘variance ‘within the region of interest. Spatial structure is quantified by semivariograms, 65 providing a theoretical basis for optimal interpolation among observation points, with a defined error structure (Warrick et al., 1986). Root distributions in 0.05 x 0.05 m cells throughout a 1.3 x 1.2 m region are interpolated from root intersections with MR tubes. Deviations from vertical trends were obtained for each observation along the MR tube by subtracting the trend value from the observed value. Spatial correlation in deviations from the vertical trends is quantified by semivariance analysis. A theoretical model, describing semivariance as a function of distance separating observations, is fit to the semivariogram data. Root length density values, corresponding to 0.05 x 0.05 m ‘blocks 'within ‘the soil profile, are obtained. by adding 'kriged' interpolated values, derived from detrended data, to the vertical trend, quantified by equation [3.6]. nummwmmu Soil water distribution was determined by neutron thermalization at three 'to seven day intervals following initiation of water stress. The neutron probe was inserted in vertical tubes centered in the crop row and interrow in 1990 and in the horizontal MR tubes. Only the horizontal MR tubes were used for neutron thermalization detection in 1991. The ratio of counts observed at 0.15 m intervals, to standard counts is related to volumetric soil water content by a field calibration procedure (Gardner, 1986). The apparent dielectric. constant, determined by TDR is related 66 to soil water by the Topp equation (Topp et al., 1982) 9v = -0.053 + 0.0282*Ka - 0.0005 * Km2 + 0.00000431er3 [3.7] where K, = (ct/L)2, t -- (B-A)/(Vp*c), c is the propogation of an electromagnetic wave in free space (30 cm s'9) , L is the distance the wave guides are extended into the soil. Vb is the experimentally determined propogation velocity of the cable. 8 is the distance of the reflected pulse from the pulse generator, taken as the tangent made by the zero and positive slopes traced in the waveform. A is the distance separating the pulse generator from the junction of the TDR cable and wave guides, taken as the point preceeding the large negative slope in the waveform. Spatial and analytic components of soil water variability are computed as the standard deviation of replicate observations at identical depths. The drained upper limit (DUL) soil water content is determined by constant soil water distribution following saturation, when active roots are absent. The lower limit (LL) of water extractable by maize was determined after 73 days of water deficits when the soil water distribution became satic even though active roots were present. Soil water available to maize crop (AW) is computed by subtracting LL for a horizon from volumetric soil water content for that layer. The quantity of available water stored in a soil layer '1' is the product of AW and layer 67 thickness (D1). The quantity of available water stored in the rooted soil profile (SAWr) is the sum of stored water in all soil layers within the rooting depth, where zr is maximum depth of rooting, as determined by the rooting front. . i=zr . SAWr - E SAWi [3.8] The quantity of available water stored below the active root zone (SAWV) is the sum of stored water in all layers from the rooting depth to the lysimeter base (zb). i=zb Sva= 2 saw [3.9] i=zr Stored soil water depletion for layer 'i' (SWDi) is computed as the change in stored water (dSWi) per unit time (dt), and equated to root sink strength (rw) when infiltration, drainage and soil water redistribution terms of the soil water balance equation are zero. swni =‘ dSWi/dt [3.10] Mean daily total evaporation (Etm) is computed as the sum of SWDi for each layer. Analytic error and horizontal ‘ variability in SWDL is computed as the standard deviation of four replicate TDR differences, or eight replicate neutron probe differences for each soil layer. mzngm Samples of the soil atmosphere, collected after unidirectional vacuum flushing the dead space of. the 68 collection system three times, were stored in 1.0 ml syringes sealed by inserting needles in neoprene stoppers. The C0; partial pressure was determined within three hours of collection-by a modified infrared gas analyzer (IRGA), using N; as a carrier gas (Schumacher and Smucker, 1985) . The output signal of the Beckman Model 865 IRGA (Beckman Instruments, Fullerton, CA) was integrated. by' a Hewlett Packard 1040a instrument (Hewlett Packard, Dallas TX), and related to C02 partial pressure by calibration with a minimum of five external standards, prepared prior to each sampling period. The C02 partial pressure for each soil depth was computed as the mean of four samples, taken along each MR tube, at each depth. The standard deviation of four observations includes spatial variability in horizontal distribution of source strength and analytical error. Carbon dioxide source strength (rc(i)) and change in stored C02 (dCOZ/dt) for soil depth '1' was computed from the mass balance of C02 at each sampling node. rc(i) - eg’dCOZ/dt = Jc(i) - Jc(i-1) [3.11] computed from C0; flux (Jc(i)) above and below each sampling node, where 89 is air-filled porosity. Jc(i) = -------------------------- [3.12] z(i+1) - z(i) where D is gas diffusivity; [C02(i+1)] is the partial pressure of C02 at soil depth 'i+1'; [C02(i)] is the partial 69 pressure of C02 at soil depth 'i'; and z(i+1) and z(i) are corresponding soil depths. Gas diffusivity is computed as an exponential function of air-filled porosity (89) (Campbell, 1985) 0 = no'e(eg) - [3.13] where Do is the binary diffusion coefficient for C02 (1.39 x 10'5 m2 s-l), 6(99) = 0.9*egz°3, 99 = e - 8v, and e is total porosity. Analxaig_ef_regt_fnasfign Functional relations of root distributions to water sink strength and C02 source strength are determined by error analysis. Available water and RLD are related to soil water depletion by solving for root water uptake equation (rw) as described by Gardner, (1960), .41IK(8) (w, - Va) rw = ------------------ [3.14] ln (oz/r2) where K(8) is unsaturated hydraulic conductivity, *3 is root water potential, V, is water potential of the bulk soil, c is the radius of a cylinder of soil from which water diffuses to the root, and r is root radius. Ritchie (1985) derived a simplified solution to the equation [3.14], assuming: 7O K(6) = 10"5 exp (62(9v - LL) (r-s) =21cmwater C 3 (n Run-0.5 r = 0.2 mm substituting these relations into equation [3.14] yields 0.00264 exp (62(6v - LL) Q: = - ........................ [3.15a] rw = RLD'qr [3.15b] where 8v is volumetric soil water content, and LL is the lower limit for available soil water. Deviation of predicted root water uptake (rw) from observed soil water depletion (SWD) is computed by l Predictive Error = - 2 w/(rW - 8WD); [3.16] N Coefficients in Equation [3.14] are used as published. RESULTS Qan221_aa4_z221_42zglenagnti_ang_2zaneratiga Canopy height proceeded at similar rates in 1990 and 1991 (Figure 3.2), though planting dates differed by 12 days. Mature leaf areas differed however, as the average leaf area from day 180 to 220 was 40% greater during the 1991 growing season. Greater development of leaf areas during the 1991 season appeared to be due to a greater plant 71 Mature Leaf Area (m2 ml) (“1) mflzon alnfin d01 0 0 2? D _, E. .............. ~ - a: - 15 -8 5 --------------- r as 075' 2 5.. < 4 l 0 “5 :1: o 2. -‘ "8- 2 - :4 ‘2' mm! « 0.5 a 2 HM“: -- 90mm! 1 l 0 180 210 240 270 Day of Year Figure 3.2 Maize canopy development on a Spinks sand for water deficits during vegetative phase in 1990, (A) green leaf area index, (8) top ligule height. Maize canopy development for water deficit periods during vegetative and reproductive phases in 1991, (C) green leaf area index, (D) top ligule height. Dashed lines are 90% confidence intervals. '72 population (Table 3.1), and greater leaf dimensions. Rapid leaf senescence after day 220 in 1991 (Figure 3.2C) may be attributed to the prolonged water deficits occurring (Aiken, 1992, pp. 36-41) during this period of plant growth. Available water stored in the root zone was less than 50% of potential water holding capacity after day 199. Soil water depletion coincided with canopy loss of turgor observed on day 199 and leaf rolling observed on days 213 and 217. Root accumulations during 1991 are evident above soil horizon interfaces at 0.72 and 1.27 m depths (Figure 3.3A). Discontinuities in soil texture and associated hydruaulic properties can promote root proliferation when water and nutrients accumulate above horizon interfaces (Smucker and Roberson (1989). Root accumulations at 1.6 m are attributed to the artificial boundary effects of the lysimeter base, which retained water above the drained upper limit. Root proliferation also corresponded with water deficits and with the ratio of observed and predicted daily evaporation (Figure 3.4). An exponential decline in soil water is apparent in the upper 0.5 m prior to loss of leaf turgor observed on day 199. This evidence of deficient water supply to the canopy coincided with a green leaf area 2 index of 2 m m"2 (Figure 3.2C), RN values of less than 0.1 cm cm'3 (Figure 3.3A) which were restricted to the upper 1.3 3 3 of available water in m of soil, and less than 0.03 cm cm” the upper 0.75 m soil layers (Figure 3.38). Deficient water supply to canopy is confirmed by mean daily evaporation 73 A B Ems» \\\\ 20:; 8 . .\\ T: I 3306* $\\\\\\\ “015 a W 3 0:4 . v / 0 4t A w—"i’é’ “ 2i? 7 3' O AvailableWte P a a S 7.....". o _ M Observed Uptake (cm day') Predicted Uptake (cm day‘) c Figure 3.3 Seasonal trends in maize root-soil water interactions on a Spinks sand in 1991 for water deficit during vegetative and reproductive phases: (A) root development (B) available soil water (C) soil water depletion (D) predictive accuracy of a cylindrical root water uptake model. Note shift in axes for available water (B), required to display seasonal trends in soil water depletion. 74 (Etm) values, computed from soil water depletion over sampling intervals, which are 38% lower than mean daily evaporation potential (E0), computed from weather data by a modified Priestly-Taylor equation over the 196-205 day interval (Figure 3.4) . Subsequent root proliferation throughout the 0.5 to 1.65 m profile (Aiken, 1992, pp. 36- 41) illustrates compensatory growth, increasing the water supply capacity of the root network. The pattern of deficient plant water supply and subsequent root proliferation recured, with leaf rolling observed on day 213 and 217. Mean daily evaporation was 57% lower than potential evaporation during the 213-217 day interval (Figure 3.4) and roots proliferated at 0.5, 1.27, and 1.6 m depths (Aiken, 1992, pp. 36-41). These results indicate transient deficits in plant water supply can result from localized water depletion in rooted soil, though roots are growing into wet soil. Thus, detecting the vertical rooting front failed as an index of plant-available water, for 9 cm of "available" water were stored in "rooted" soil. (Aiken, 1992, p. 40) when the ratio of Etm/Eo fell below 0.45. Mean daily evaporation (Etm) exceeded potential evaporation (Eo)- during the 217-224 day interval (Figure 3.4) . This result is attributed to biased estimates of green leaf area index (GLAI, Figure 3.2C), attaining maximum values at this period. If GLAI was 2.5 m2 m"2 rather than the reported 5 m2 m'z, E0 would increase by 56% due to. '75 Relative Evaporation (ETob, E04) 2io 240 270 Day of Year Figure 3.4 The ratio of mean daily evapotranspiration (ETm) to mean daily potential evaporation (E0) is illustrated for maize subject to water deficits_ during vegetative and reproductive phases in 1991. PIT,n was determined by soil water depletion (N = 56 observations for each of two similar profiles). E0 is daily potential evaporation, computed from a modified Priestly-Taylor equation. LCI and UCI are 90% lower and upper confidence intervals for daily soil water depletion, integrated over sampling intervals. 76 effects of decreased albedo. The magnitude of this probable bias corresponds to the magnitude that observed evaporation (Em) exceeds predicted evaporation (E0) (Figure 3.4) . Thus, conclusions that observed evaporation exceeded atmospheric potential evaporation is probably invalid as potential evaporation is likely underestimated due to biased estimates of soil plus canopy albedo. vat Derivations of the single root water uptake model are frequently used to solve the uniform cylindrical flow of water to a single root (Gardner, 1960; Ritchie, 1985) These formulae are used to relate soil water depletion to RLD and AW, observed at seasonal and spatial scales. Soil water depletion rates, observed at three to seven day intervals (SWDO) , are compared with root water uptake, predicted by equation [3.15a and b] (SWDP). Predicted values are based on root intersections with horizontal MR tubes, and soil water detected by neutron thermalization in the same MR tubes. Measured soil water depletion rates decreased as a linear function of soil depths, from 0.75 to 1.45 m prior to anthesis in 1990 (Figure 3.5C and D). Predicted soil water depletion tended to increase with soil depth, reflecting vertical trends in the profile distribution of soil water and roots (Figure 3.5A and B). Predictive error was nearly equivalent to the magnitude of soil water depletion for successive sampling intervals. These results suggest a 77 0.1 v v - * ~fi a e ~ - 6‘ - DaysZOS-208 A _ g 0.08 - — o.,-szos-214 ; ‘ ’ 7° '7 1 O E t Days 205-208 0 ° 8 T: 0.06 " ‘> — Days 208-214 A 2 I Z 5; 9 ’ O 2 004 1 B n c}: .5! . g 0.02 ~ ~_ ‘< , n . o e 4 4+7 A :: ..... c -,,i + A 0 C 1. D 0.006 - -* ~0.006 O 0 Predicted I 3 ‘7 Predicted ' Observed U 1 .5 O 3 0 Observed 0.004 r I § (Map {.mo (1110) nonoldaq 1919M [[03 I U 0.002 ” I: I I .. . V v I . Soil Water Depletion (cm3 cm-3 day-l) 1 1- i i i i L h p i I i 0‘“ 0:5 1 I 0 ”0:5“ 1 1.5 Soil Depth (111) Soil Depth (m) Figure 3.5 Comparison of soil water depletion, observed by neutron thermalization (SWDO) , with depletion due to root water uptake predicted by a simplified solution to a cylindrical model of root water uptake (SWDp). Observations under maize in a Spinks sand for water deficits during vegetative stage in 1990. (A) Soil profile distribution of available water, (B) Soil profile distributions of root number, (C) Soil water depletion, predicted and observed for days 205-208, (D) Soil water depletion, predicted and observed for days 208-214. Upper and lower 90% confidence limits are computed for N=8 observations. 78 systematic bias. in. equation [3.15a and b]. Biases in estimates of root water uptake could result from systematic errors in measure of available water (AW) and root length density (RLD, equated to RN by equation [3.41) . Systemic bias in AW determination is unlikely as the neutron thermization conditions were identical at all depths, and neutron probe calibration results indicate linear response of the range observed (Gardner, 1986). However, MR intersections may underestimate RLD during initial root extension into soil layers, as the rooting front is detected by soil water depletion prior to root intersection (Aiken, 1992, p. 36). But underestimating RLD would result in underestimating SWD, while the contrary was observed (Figure 3.5C and D). Thus, systematic bias in measurement of AW is unlikely, and effects of a likely biased in estimates of RLD are opposite in direction to the bias in predicted SWD. Systematic bias may result from invalid assumptions of soil and plant hydraulic properties. Differences in radial conductivity in suberized and unsuberized dermal layers of conductive and sorptive regions of roots (Russell, 1977) are neglected in equations [3.15a and b], and could account for biased predictions. But consideration of differences in radial conductivities would tend to increase the bias in predicted SWD, for younger roots at greater soil depths are expected to have a higher proportion of sorptive regions than older roots. higher in the soil profile. Soil 79 unsaturated hydraulic conductivity (1((8) , equation [3.15a]) is sensitive to the value of the exponential coefficient, doubling in value with a 10% increase in the exponent when AW is 0.12 cm3 cm'3. But errors in the me) function can only account for the magnitude, not direction, of predictive error, for a smaller coefficient would still predict increasing water uptake for the high AW conditions observed at greater soil depths. Systematic bias in soil and root components of radial resistance to root water uptake fail to account for the direction of bias in predicted SWD. The assumption of uniform hydraulic gradient between root and soil is valid only if root axial resistance to water flow and gradients in root xylem water potential are negligable. Experimental evidence (Yamauchi et al. , in press; Hainsworth and Aylmore, 1989) contradicts this assumption. Axial resistance to water flow in roots would increase the hydraulic gradient required for water flow through xylem vessels. Thus, xylem water potential would be less negative with respect to rooting depth, with corresponding reductions in radial root-soil hydraulic gradients with respect to rooting depth. Depth-dependent effects of frictional resistance along xylem vessels accounted for water uptake patterns observed for cotton and soybean (Klepper et al. , 1983) . We attribute systematic bias in predicted water uptake to effects of axial resistance on radial root-soil hydraulic gradients that are not considered in equation [3.15a and b]. These effects 80 could be approximated by making the soil-root hydaulic gradient a linear declining function of soil depth. Seasonal trends in predicted SWD (Figure 3.3D) and soil water depletion (Figure 3.3C) observed in 1991 confirm diagnosis of systematic bias observed in 1990 (Figure 3.5). More water was taken up from layers above 1.0 m.than from layers below’ this. depth. prior' to depletion at day 200 (Figure 3.3C). This is attributed, in part, to the timing of root extension and proliferation into wet soil (Figure 3.3A). Recall, this coincides ‘with ‘visual evidence of deficient plant water supply, observed on days 199, 213 and 217 (Aiken, 1992, pp. 36-39). In contrast to field observations, predicted SWD underestimates the proportion of water uptake at more shallow depths and overestimates the proportion of uptake deeper in the soil profile. Thus, systematic bias in predictions of the vertical distribution of root water uptake persisted throughout an extended drying cycle for maize roots growing into wet soil. Error in predicted root water uptake can be attributed. to changes in RN during root proliferation phases, when they are not accounted for in equations [3.15a and b]. Robertson et al. (in press) determined that root water uptake during early phases of grain sorghum root proliferation was overestimated by an exponential decline model of soil water depletion, taken from Passioura (1983), that assumes constant RLD. 81 WW Seasonal trends in C02 evolution observed in 1990 are also related to root development (Figure 3.6A and 3.6C) . Net C02 flux, including source strength and transient storage effects, was computed from flux gradient data using Equation [3.11]; and is concentrated in the upper 0.5 m of soil, where root observations by MR were restricted by the deeper locations of MR tubes (Figure 3.1) . Carbon dioxide accumulation in the 0.5 to 0.9 m layer at day. 195 corresponds with increased source strength (Figure 3.68) and root proliferation (Figure 3.6C) at this depth and time. Accumulation of C02 below 0.75 m to 1.6 111 during day 210-220 corresponds to a shift in the proliferation of roots to these soil layers. A flush of respiration on day 220 (Figure 3.6B) coincided with an initial wetting event when the water deficit period was ended by irrigation. Seasonal trends in C02 evolution observed in 1991, Figure 3.7, indicate similar increasing C02 gradients with respect to soil depth, as observed in 1990, Figure 3.6; though C02 accumulations below 0.7 m were twice that observed in 1990 (Figure 3.6A). The steepest gradient in the partial pressure of C02, Figure 3.7A, occurred in the 0.5 to 0.7 m soil layer in 1991, but between 0.15 and 0.5 m layers in 1990. Relatively greater C02 accumulations and a shift in maximal C02 gradients in 1991 to lower soil depths, relative to 1990 indicate greater root plus soil respiration at lower soil depths. This difference is attributed to 82 Soil [C02] (3 m4) Net C01 Flux (g 111'2 day“) Roots (N0. cm-z) ‘. ’4 \ K 1 O O ‘ .0?‘?;!f.j - - Figure 3.6 Seasonal trends in net C02 flux for water deficit during. the vegetative phase for maize “on a Spinks sand in 1990 (A) seasonal trends in the soil profile C02 gradients, (B) seasonal trends in soil plus root C02 source strength, (C) seasonal trends in root distribution. Soil [C02] (3 111-3) 83 30‘ I ' i . l ‘ / a [Mil/«‘1‘ 0}’A’///I/ ll/I‘III 8‘ 130 .' W‘J/' 15 :25 Day :11'0 H-."if!’/w¢b pa 0 ~l 01 Roots (N0. cm‘z) 9 o D.’ a... . I .9 l .03 c .r v o 5. 0'. O O . «9.9.2.2.. ‘ Figure 3.7 Seasonal trends in net C02 flux for water deficit during the vegetative and reproductive phases for maize on a Spinks sand in 1991 (A) seasonal trends in the soil profile C02 gradients, (B) seasonal trends in soil plus root C02 source strength, (C) seasonal trends in root distribution. 84 greater soil microbial decomposition of senescent roots remaining from the previous growing season. Decomposition was delayed in 1991, but not in 1990, by dry soil conditions the previous fall and winter. Reduced C02 accumulations, observed prior to day 217 and after day 260, coincide with reduced soil water depletion (Figure 3.3C) . Presumably, reduced transpiration corresponds to reduced photosynthesis, consistent with reports of a close linkage of root respiration with canopy assimilation (Hall et al., 1990; Martin, 1987). The non-linearity of C02 gradients and corresponding non-linear distributions of C02 flux indicate assumptions of steady state C02 flux conditions may be invalid. This inference follows from inspection of the steady state condition, derived from the mass balance for C02 (equation 3.11): Under steady state conditions, gradients in C02 flux can only be attributed to re, the net reactions involving C02 (primarily generation by respiration). The persistent large negative values for net C02 flux at 0.08 to 0.15 m depths are not likely to result from soil reaction with CO; or plant root uptake of C02. More likely, negative net flux reflects C02 accumulation in these layers as C02 respired by roots and soil organisms in lower soil layers diffuse through soil pores towards the atmospheric sink. Thus, 85 regions of C02 source strength; occurring near the soil surface, at 0.75 m and 1.5 m; are bounded by regions of net C02 accumulation at time of sampling (just prior to solar noon). DISCUSSION A simplified solution to the single root water uptake model results in systemtic, depth-dependent bias in predicted root water uptake. Failure to consider depth dependent gradients in root xylem potential, arising from root axial resistance to flow, most likely accounts for the positive bias, with respect to soil depth, in predicted root water uptake. Vertical gradients in soil water depletion are consistent with nonuniform uptake observed by Hainsworth and Aylmore (1989); and with nonlinearities in root water flux relations predicted by Dalton et al. (1975). Spatial and temporal variations in soil and root hydraulic properties can contribute to errors in predicting the quantity and distribution of root water uptake in the soil profile For example, water potential gradients between roots and the soil are expected to fluctuate as stomatal resistance adjusts to fluctuating evaporative demand, radiative heat loading, etc. Soil dessication can reduce the root-soil water contact, and alter root radial resistance to water flow due to senescence or suberization of root tissue. These effects are not considered, however, in predicted root water uptake (equation [3.15]). 86 Horizontal gradients in root distributions, and clustered root distributions are expected to result in corresponding gradients in water and nutrient uptake (Aiken, 1992 p. 39) . Failure to account for these gradients can reduce sensitivity of crop growth models to transient plant water deficits that can alter transpiration, assimilation, and C allocation patterns. Horizontal gradients in root distributions, affecting soil water uptake and transport are particularly relevant to ridge and no till systems where repeated root exploration of the same soil volumes beneath crop rows accelerates the depletion of immobile nutrients; and enhances the innoculum potential for root pathogens. Modification of root water uptake models should take into account. root axial resistance to ‘water flow, spatial root geometries, seasonal changes in root hydraulic properties, horizontal gradients in the various soil conditions and related transport parameters. The regularities in root system development implicit in thermotropism (Fortin and Poff, 1990) , multi-order branching (Rose, 1983) , and phasic development (Klepper, 1987; Ritchie and NeSmith, 1991) suggest opportunities for predicting root system development. Integrating these regularities in root development with concepts of compensatory growth (Russell, 1977) suggests the need for a renewed direction for analysis of soil stress effects on plant growth and development (Luxmore and Stolzy, 1987). ROOT IUNCTION AND NITRATE LNRONING UNDER IRIS! IEIBQDEQIIQH Nitrate leaching represents a public health hazard and the low retention of an essential nutrient. Leaching losses are minimized when root uptake is synchronous with soil nitrate supply. Root uptake of N in the transpiration stream is an explicit sink in the conservation equation for solutes. Root effects on whole plant growth and development also modify N leaching potential by conditioning root and canopy architecture, and subsequent demand for water and N. Relating the effects of root system development and subsequent activity to nitrate leaching requires knowledge of factors regulating soil [N03] and water flux, which determine N leaching rates at the lower boundary of rooted soil. Simulations of the soil N balance, including leaching losses, are hampered by inadequate knowledge of the distribution and activity of plant root systems. Knowledge of root responses to environmental factors are based on young plants cultured under controlled environments (Klepper et al. 1983). Simulations of root geometries in field conditions typically assume horizontal homogeneity and prescribed vertical distributions, subject to environmental constraints. Accurate specification of the‘ functional 87 88 equilibria between root and shoot activities and growth (Brown and Scott, 1984) may improve simulation sensitivity to root function. Numerical simulations of complex systems are verified when solutions conform to fundamental physical laws, e.g. conservation of energy and matter, and accurately represent the conceptual model of system structure (Manetsch and Park, 1987) . Simulations are validated when system state and output parameters are accurately predicted. Simulation of root effects on N leaching can be validated by evaluating the deviation of predicted water, C, and N distributions from those observed under field conditions. This study was conducted to determine the sensitivity of N leaching predictions to errors within root system distributions, as simulated by CERES-Maize (Jones et al., 1986) . We hypothesize that the predictive accuracy of simulated N leaching, by CERES-Maize, is sensitive to errors in simulated root morphogenesis. Model predictions of soil and plant C, water, and N state and output parameters are compared with field observations of lysimeters with conventional or no-till crop cultures. W W Maize (Zea mays, L. hybrid Pioneer 3704) was grown in duplicate 1.2 x 2.1 x .1.8 m field lysimeters under conventional till (CT) or no-till (NT) crop culture, with no 89 fertilizer additions in 1991. A previous corn crop, established in .July, 1990 ‘was followed by a rye crop, established in October, 1991. The rye was killed on day 131, 1991 at the vegetative stage with Round-up (0.06 L m-2 of 1:8 Round-up:water solution). The rye shoots were clipped and removed on day 135, 1991. Residual nitrate from application on day 212, 1990 was estimated to be 500 Kg N Ha-l and greater. Moldboard plowing was simulated by spading the soil to a depth of 0.2 m on day 140, 1991. Three rows of corn were hand-planted in each lysimeter on day 140 which was concurrent with the associated field plots, Figure 4.1. Plants were thinned to a population of 8.33 plants m-2, seven plants per row. Nonuniform germination required thinning and transplanting in the lysimeters at the three- leaf stage for seven plants in NT6, two plants in NT9, and 15 plants in CT13. The canopy over the lysimeter was continuous with the field canopy in 1991, with the exception of a 1 m. gap over each access chamber adjoining each lysimeter. WW Lysimeters were installed and instrumentation was installed in early to mid 1990, as described in Figure 4.2 and Table 4.1. Methods used to determine the state of crop and soil parameters are described in later sections. Methods used to simulate weather effects on crop and soil interactions are described in the final section. 9O Historically Titled Never Tllled l"‘]' . 4 a 12 16 18 20 22 24 CT.NF NT.NF NT.F CT.F Rel NT,NF Rel NT.NF 3 7 11 ' 15 17 19 23 NT.NF CT.F CT.NF NT,F NT .NF Rel Rel film 15:11 2 6 10 14 qfl FF CT.NF NT,NF ~11? CT.F e on £92215 01' Conventional Tlllage 1 s 9 ' '13 .NT 'No-tlllaae CT.F - NT'F "T'NF CT'NE F . F ertillzed mi Not-tertlllzed Rel Reterence (unplanted) I‘- 27111 -e-I Figure 4.1 Diagram of field plots established at the Kellogg Biological Station in 1986 to investigate N supply and tillage effects on soil-plant interactions in lysimeters located in plots 2, 6, 9, and 13. 91 x. x X' xxx xx)! X .x- x .xr" x , X .X .X ,r ,X .x ' 3-x-x xflxx 3"", 71.12 xx. .x-‘x' -m'fI' £10 1.» Alt-xx- / . PLANT Ruvs AP _1~*"' ................................................................................ B t A MINIRHIZUTRUNS // VAVEGUIDES ” // + Cit x o ................. -“ ~' ................................... X X 00 ZBt 203 THERMDCUUPLE_\\. WC: ”I + cfi x o .............. " x 0 °\ sou. SULUTIUN C + SAMPLERS TEFLUN SCREEN ' S //I I 21 l\ \ m , A 229 33 Figure 4 . 2 Instrumentation ports for non-destructive sampling of root distributions, soil water, soil solutes, and soil atmosphere above and below soil horizon interfaces in non-disturbed field lysimeters on a Kalamazoo loam soil. 92 Field lysimeters were installed in field plots established at the Kellogg Biological Station (KBS) in 1986 to investigate N supply and tillage effects on soil-plant interactions (Robertson and Smucker, 1988). The non-weighing lysimeters were located five meters from the boundary of the 27 x 40 111 field plots by excavating around a 1.2 x 2.1 x 1.8 m pedon. A stainless steel lysimeter chamber was inserted over the nondisturbed portion of the excavation, and driven around the pedon. The lysimeter-pedon assemblage was inverted and a base and 0.43 m extension were welded onto the assemblage, and filled with sand from the soil parent material. A nylon screen separated a layer of pea- sized gravel which was covered by a stainless steel base sloped to the center drain (Figure 4.2) . The lysimeter was uprighted and returned to the original field site. The surrounding soil was restored to the excavation site. Soil samples and mapping of horizon boundaries provided detailed information of soil physical and chemical properties for each lysimeter. Instrumentation ports for the nondestructive sampling of root distributions, soil water content, soil solutes, and soil atmosphere were installed in clusters 0.02 m above and below soil horizon interfaces, directly below the center row. Polybutyrate minirhizotron (MR) tubes (0.06 x 1.2 10) parallel to the center crop row and the soil surface provided access for recording root intersections (Ferguson and Smucker, 1989) by a color microvideo camera (Bartz 93 Technology Co., 650. Aurora Ave. Santa Barbara, CA 93109). Teflon capillary tubing (0.5 mm ID x 2 m), fitted with septa and taped to the MR access tubes provided access to the soil atmosphere at 0.25, 0.50, 0.75 and 1.00 m behind the lysimeter wall. Hypodermic needles Table 4. 1 Lysimeter horizons and NO3-N status on day 123, 1991 of Kalamazoo loam soil at Kellogg Biological Station. Rye biomass removed on day 135, 1991. Soil layer CT2 CT13 NT6 NT9 Ap 0-25 0-23 0-21 0-21 E - 24-30 21-30 21-30 Bt 25-53 30-64 30-56 30-48 ZBtZB 53-73 64-84 56-66 48-55 28t2C - - 66-83 - 28t3 - - 83-107 55-78 3E\Bt 73- '84- 107- 78- Residual N 1443 346 183 I 450 Kg N ha-1 Rye Biomass 4730 4440 4080 4760 Kg ha-1 (0.5 mm ID) or stainless steel capillary tubing (0.5 mm ID.) , fitted with similar septa, were installed in the vertical direction in the topsoil with' sampling depths at 0.03, 0.07 and 0.15 m. Paired 0.01 x 0.3 m stainless steel 94 .rods served as parallel wave guides for time domain reflectrometry (TDR) determination of soil water (Topp et al., 1982). The TDR rods were oriented horizontally and parallel with crop row (Figure 4.2). The soil solution was sampled by suction lysimeters (Soil Moisture, P.O. Box 30025, Santa Barbara CA 93105) composed of ceramic cups (0.03 111 OD) attached to 0.4 m cylinder and inserted 0.3 111 into the soil. Each lysimeter drained at the base of the soil profile. Continuous drainage emptied into a 58 L collection vessel, enabling determination of instantaneous and cumulative outflows of water and corresponding samples of solute concentrations. miner. Daily global radiation (R9), precipitation (P), and temperature maxima (Tmax) and minima (Tmin) were recorded from day 154 through day 275 by a weather data system (LI- 12008, LI-COR, Lincoln, NE 68504) located within 200 m of the lys imeters . Dai ly quantity and duration of precipitation, Tmaxr and Tmin were also observed at a National Weather Service (NWS) reporting station two Km west of the lysimeters. NWS weather data is combined with R9 data, recorded at the KBS Pond Lab, to characterize weather conditions at the lysimeters when on-site data were not available. Potential evaporative demand (E0) is computed daily by a modified Priestly-Taylor equation (Ritchie, 1985) as a function of R9, Tmaxr Tminr leaf area, and soil albedo. 95 W Canopy development was determined on 21 sampling dates for five plants within the population of each lysimeter. Date of leaf maturation and senescence, mature leaf length and width, height to top mature leaf ligule, and reproductive development phases were determined to quantify canopy structure. Green leaf area (GLA) was computed for each plant at each sampling period 111 GLA = 2‘. (L1 x wi x 0.75) [4.1] 8 where Li and W1 are the length and width of the "ith" leaf, 0.75 is a shape factor, m is the top mature leaf number, and s refers to the top senescent leaf number. Confidence intervals (Steel and Torrie, 1980) about the canopy green leaf area index (GLAI) and height to top mature ligule (H) are computed from GLA and plant populations. Variable plant development required adjusting individual plant leaf area and stem height according to the relative frequency of small and large plants in the sampled and whole lysimeter populations. Plant height at day 200, determined for all plants in the lysimeter, served as the basis for adjusting the weights of individual plant measurements. The relative frequency of plants in six height classes, bounded by one or two standard deviations 96 from the mean was determined for the population of plants in each lysimeter, and for plants selected for detailed sampling. weighting factors for mean canopy leaf area and plant height are computed as wi " fpilflsi [4.2] where w; is the weighting factor for plants in the "ith" size class, fpi is the frequency of plants in the lysimeter population (N=21) for the "ith" size class, f“ is the frequency of plants in the lysimeter sample population (N=5) for the "ith" size class. Weighted mean canopy leaf area is computed by GLAI = ------------ [4.3] where "1 represents the sum of the relative frequency weighting factors, GLAi represents the sum of mature green leaf area, m is the top mature leaf number, and s is the top senescent leaf number for the time of computation. Root intersections with MR tubes were recorded five times by a microvideo camera in 1991. Frame dimensions are 13.5 x 18 mm, with the longer dimension perpendicular to the MR tube. Active roots were identified by high light reflectance, opaque appearance, structural integrity and diameters exceeding 75 um. Senescent roots, observed prior to crop establishment, were excluded from root counts as were translucent roots and roots with low reflectance. Root 97 number values are determined by computing root counts (N) per unit area (A) of the soil-MR tube interface, and are equated to root length density (RLD) by the following relation, suggested by Upchurch and Ritchie (1983) RLD = Nd/Ad .[4.4] where d is the MR tube outer diameter. This relation assumes the mean length of roots intersecting the volume occupied by the MR tube would be equal to the outer diameter of the MR tube, if the soil were not displaced by the MR tube (Upchurch and Ritchie, 1983) . Root proliferation (Rp) during a sampling interval is computed for each MR tube as the rate of change in RN with respect to time (t) Rp = RN/ t [4-5] Vertical distributions of mean root intersections observed for each horizontal MR tube are analyzed as a logarithmic function of soil depth RLD = a + b*log(z) [4.6] where z is soil depth, a and b are empirical coefficients determined by the software CricketGraph (MacIntosh, Redmond WA 98073). 8211.2224131225 Distributions of soil water content and C02 partial pressures were determined bi-weekly, from day 140 through day 275. Drainage and leachate [NO3-N] determinations began 98 on day 88, and continued each week for one year. Extraction of the soil solution, and subsequent analysis coincided with soil water and C02 sampling, but were discontinued when water failed to flow into individual suction lysimeters due to dehydration of adjacent soil. Soil water content was determined at sampling ports by time domain refleCtometry (TDR). TDR is based on the principle that an electromagnetic wave, transmitted along parallel wave guides, is damped in a systematic manner that is proportional to the surrounding dielectric field. As the dielectric constant is nearly 30 times greater than mineral soil constituents, TDR observations are highly correlated with volumetric soil water content. A Tektronix 1502C cable tester served as signal source and analyzer. The apparent dielectric constant, determined by TDR (Topp et al., 1982), is related to soil water content by 0v = -0.053 + 0.0282*1 i 0.1: ‘\\\§\\;\\§;\:{k‘}|‘ u g 0.1: l'§§§;‘\&’"m g .325 a; “Film '1 ll‘ilmmwi :3... ,5 'IIITI "‘ L’“ . 801709121221) 1 120 $33, 0‘16“ 01! %Z) 1 at?! ‘1 A /\D =3 5 ~» . :5 1‘ K \ :1 ~ ~ ’5 \\ g ‘ \\\v\\\\§ g §§€'T"‘ L ' 2 111111 “I“ “ ..M I 120 18032124" Figure 4.8 Seasonal trends in soil water distribution in .non-disturbed field. lysimeters. subject to either no- tillage (A) NT 6, (B) NT 9; or conventional tillage (C) CT 2, (0) CT 13. 117 Greater root water uptake in the AP horizon for CT relative to NT could contribute to tillage effects on soil water depletion. Inference of reduced soil water recharge in the upper layer of the Bt horizon for CT is corroborated by field observations of surface soil dispersion and crusting-- indicating reduced infiltration rates; evidence of rill erosion on CT lysimeters and field plots--indicating runoff; and early failure of suction lysimeters in the upper Bt of CT lysimeters horizons relative to NT lysimeters and field plots--indicating soil water depletion. We infer tillage increased surface runoff, reduced infiltration and restricted subsequent water flux into and below the root zone. Differential infiltration due to preferential flow paths that can be associated with no-tillage (Kanwar, 1991), and to surface crusts developing under CT would be greater for rains of high intensity, exceeding water intake capacity of CT but not NT. Such an event occurred on day 182 when 51 mm were received within a six h period; and again on day 203 when 77 mm occurred in 9 h (Figure 4.9). It is likely these, and similar subsequent rains exceeded the water intake capacity of CT soils, restricting soil water recharge relative to NT. 118 3° 1 Observed Rainfall (mm) G .g 0 S 80 Ié 8 CT Infiltration (TTP) H O... 0 150 180 210 240 270 Day of Year Figure 4.9 The distribution of rain used to simulate soil-plant effects on Neg-N leaching in field lysimeter on a Kalamazoo loam soil at Kellogg Biological Station. Records obtained from National Weather Service reporting station two Km west of lysimeters (days 123-153); and from data logger located with 200 m of lysimeters (days 154-275). Time-to- ponding (TTP) estimates of infiltration under CT based on soil water intake rates. 119 Simulated soil water content (SWC) values are compared 'with. field observations for the upper layer of the Bt horizon (0.25 m depth, the TDR measurement nearest to the soil surface). Soil water recharge was simulated by two different methods. The standard procedure in CERES-Maize relies on proportionate loss of rainfall to runoff; quantified by a curve number (CN) derived by the Soil Conservation Service for soil texture, drainage, and slope characteristics (Ritchie and Crum, 1989) . An alternative method under development (personnal communication, J.T. Ritchie) is based on the time-to-ponding (TTP) analysis of infiltration; surface water flux conditions exceeding soil intake capacity are defined by nonlinear functions of soil hydraulic characteristics. Soil water was simulated by both method of computing infiltration and two root growth conditions: mm, a parameter defining maximum root length density was assigned either the default value of 5.0 or 2.0 cm cm’3. Thus, interacting effects of infiltration and RLD on predictive accuracy are evaluated. Simulated SWC was generally higher than field observations for NT6 (Figure 4.10A) . Virtually identical results were obtained for NT lysimeters when either curve number (CN) or time-to-ponding (TTP) methods were used to partition rainfall to infiltration, reflecting similarity in simulated infiltration rates. Restricting maximum RLD (RLDmax) to 2.0 cm cm-3 had no effect on simulated SWC. 120 0.4 ’ 55‘ : o sar E 0.35 r g 0 3 ._ 4 DUL- ; E 0 . ‘a’ 0.25 - O D g 0.2 a Q”. 3 . r: 0.15 : O . V1 1 0.1 ............. 120 .150 180 210 240 270 Day of Year 0.4, . ' . . E B .m 0.35 E . C SWC 0.3 0.25 l 0.2 :- 0.15 : Soil Water Content (01.13 cm'3) 0.1 Day of Year Figure 4.10 Seasonal trends in predicted and observed soil water content (SWC) for upper layer of Bt horizon of Kalamazoo soil under (A) no-tillage (lysimeter NT 6), or (B) conventional tillage (lysimeter CT 2) at the Kellogg Biological Station. Simulation runs used time-to-ponding (TTP) or curve number (CN) methods of estimating infiltration and RLDfiax values of 5.0 or 2.0 cm cm'3. 121 Soil water content under CT lysimeters was underestimated by CN and TTP methods from day 150 (crop emergence) through day 170 (Figure 4. 10B) . The TTP method of determining infiltration underestimated SWC after day 170. Soil water contents simulated by the TTP method were parallel to but lower than observed soil water. The TTP method results in reduced soil water recharge and increased surface runoff for precipitation events when rainfall intensity exceeds the soil intake capacity due to surface crusting. Restricting RLDmax to 2.0 cm cm-3 resulted in slightly higher SWC when simulated by the TTP method, but not the CN method after day 180 (Figure 4.108). This result is consistent with the logic of total soil water depletion limited by evaporative demand at high SWC, but by root and soil interactions at low SWC. The numerical value of SWC predicted by CN method was close to observed SWC for the period from 180 to 210 days,, though fluctuations in simulated SWC did not correspond to observed wetting and drying cycles after day 220. The time_ trend of SWC predicted by the TTP method .of partitioning precipitation into runoff and infiltration fractions was nearly parallel to observed SWC. Accounting for surface barriers to infiltration under high rain intensities by the TTP method reduced the biased estimate of SWC for CT plots when surface crusts reduced soil water intake capacity. The TTP method offers potential to improve accuracy in simulated soil water dynamics . 122 8911.8131319 Residual soil nitrate differed among all soil profiles in both magnitude and distribution (Table 4.3, Figure 4.4). Uncertainty about the quantity, of initial N application rates, on day 212, 1990, limit comparisons of tillage effects on N03-N leaching to distributions within soil layers; though the quantity of N03-N stored in soil horizons indicate initial application rates exceeded 400 to 1700 Kg N ha-l. The fraction of residual N03-N retained in Ap and Bt horizons under CT was 0.61 (CT 2) and 0.49 (CT 13), but only 0.10 (NT 6) and 0.002 (NT 9) under NT. It is interesting to note that N03-N accumulations occurred at the~ lower boundaries of the Bt and 2Bt horizons for the CT lysimeters (Figure 4.4) . Characteristics of these horizons include clay accumulations, and, textural changes at horizon boundaries. Systematic differences on NO3-N retention by surface horizons are associated with tillage treatments and may account for the reports of root accumulations at this horizon interface (Robertson and Smucker, 1989). ‘Seasonal trends in the soil N03-N profile are depicted in Figure 4.11. Soil N03-N was consistently lower for NT, when compared to CT, reflecting minimal retention by AP and Bt horizons, leaching losses and subsequent uptake by rye and maize crops in the NT. Lack of residual NO3-N at maize planting (day 156) for NT 9 indicated plant N uptake was generally restricted by net N mineralization rates. An exception to this rule occurred from day 180 to 200, when 123 Soil Nitrate (mg N Kg;I Soil) 52' Soil Nitrate (mg N K34 Soil) 5‘ “:~‘¢‘ So- 05 75 ' 12015°o(19" ”Depth? 1 90 93" I”) 9 C :9 \D 5’, 20° 58 4o \ E0100; \\\\\\\ ' E 2° A \ v E 1 x “ 8 ‘y\ \\ 5 5°‘ \‘\\\§\\ .5. ‘° ’ \Q:\\\ i 01 |\\\\\\I}I|III L210 2 0 \\‘I}‘|II L 210 8 0.25 II" 180 :3 0.25 \ ”I 180 s . 0.5 120 150 61:91 30. 0.5 075 ,I 120 ”01’! 01113er :1: l 90 90‘! o Deplb (m) 1 9° 9"! 0 Figure 4.11 Seasonal trends in soil N03-N distributions in non-disturbed lysimeter profiles subjected to either no tillage (A) NT6, (B) NT9; or conventional tillage (C) CT2, (D) CT 13. Note difference in scale among figures. 124 ' soil N03-N accumulated (Figure 4.11B), suggesting that net N mineralization exceeded crop uptake. Rapid depletion of soil N in NT6 (Figure 4.11A), during the interval of time between day 120 to 150 illustrates effective sequestering of mineral N by the rye crop. Data collected from NT 9 indicate similar' trends during this period, but are incomplete due to equipment failure in the upper Bt suction lysimeter, and not presented. Seasonal trends in soil N03-N for CT lysimeters reflect historic and applied tillage modifications of soil water (Figure 4.8) , with corresponding effects on solute transport and transformation processes. Residual soil [N03- N] on day 122 in CT 2 were very high, by agronomic stands, but moderate for CT 13 (Table 4.3, Figure 4.5B). Soil NO3-N in the upper Bt horizon of CT 2 increased by 60 Kg ha-l within two weeks following tillage (day 141 Figures 4.11C, 4.12) . The concentrations of N03-N in Ap and B1; horizons indicate high application and retention rates of N03-N for this lysimeter. Soil N03-N depletion in CT 13 from days 90 through 150 is attributed to rye uptake and leaching losses. Continued soil NO3+N losses after day 180 corresponded with rapid maize crop vegetative growth after the eight leaf stage. Simulated soil N03+N for upper and lower boundaries of the Bt horizon are compared with observations at lysimeter CT2 (Figure 4.12) where initial concentrations were 86 mg N Kg-l (upper Bt) and 152 mg N Kg-l (lower Bt) Table 2.2. 125 Simulated soil NO3-N for upper and lower boundaries of the Bt horizon are compared with observations at lysimeter CT2 (Figure 4.12) where initial concentrations were 86 mg N Kg-l (upper 81:) and 152 mg N Kg'l (lower Bt) Table 2.2. Increased [NO3-N] in the upper Bt after day 142 followed soil tillage (day 140) and 15 mm of precipitation; this indicates the influx of soil N03-N subsequent to soil disturbance. A drop in [N03-N] in the lower Bt after day 133 coincided with soil water depletion, and is attributed to rye root uptake of water and N as SWC was less than the 200 j T 1 I y... M O v. 8.8. 17-12. I" v Soil Nitrate (mg N Kg;l Soil) 8 50 — 0115.0 . 0 0112.0 , -- 11115.0 _ v 11'on o i a 1 _i - 1 - A 1 i a 1 a i 120 150 180 210 240 270 Day of Year Figure 4.12 Seasonal trends in soil [N03-N] observed by suction lysimeters and simulated by CERES-Maize for lysimeter CT 2, conventional tillage treatment with high initial residual NO3-N. Simulated runs using time-to-ponding (TTP) and curve number (CN) methods of predicting infiltration, and RLD“, values of 5 or 2 cm cm‘j. 126 drained upper limit (DUL). Data on soil [NO3-N] from suction lysimeters are not available after day 184 due to failure of soil water extraction. Simulated soil [N03-N] was lower than actual observations. The influx of soil N03-N after day 142 was not reflected in simulated N transport processes. The TTP method of partitioning rainfall into infiltration resulted in higher soil N levels after day 190 relative to the CN method. This result is attributed to reduced infiltration and leaching for the TTP method as confirmed by lower SWC for TTP estimates of infiltration after day 180 (Figure 4.108). Reducing RLDmax from 5.0 to 2.0 cm cm-3, slightly altered predictions of soil N03-N, due to effects on water and N extraction (Figure 4.12). Seasonal trends in soil N03€N distributions reflected net effects of soil water transport, root uptake, and microbial transformations. Residual N034N distributions in Bt horizons were moderate (CT 13) or very high (CT 2) for CT lysimeters but low (NT 6) to depleted (NT 9) for NT lysimeters, reflecting differences in solute retention induced by tillage effects. An increase in No3+N following tillage observed in CT 2 suggests solutes may be retained in regions of soil water that are resistant to leaching, unless disturbed by tillage. Simulation of seasonal trends in soil No3-N is sensitive to method of predicting infiltration, and slightly sensitive to. RLDw when soil water approached the lower limit of plant extractable SWC. 127 W Transport of N03-N to the BE\Bt horizon, observed in NT lysimeter, requires sufficient infiltration to displace solute bearing water held in upper horizons. The concentration of N03-N in water drained below the root zone combines with drainage flux rate to determine nitrate leaching rates. Flux of water and N out the lysimeter are reported in this section, with consideration for soil and root effects on flux rates. Lysimeter drainage occurred prior to day 150 and day 300 (Figure 4.13), corresponding to soil water depletion by root uptake and evaporation. Drainage flux from NT lysimeters exceeded that of the CT lysimeter, and persisted through the maize growing season. Reduced drainage under CT relative to NT is consistent with evidence of reduced infiltration under tillage and reduced soil water recharge (Figures 4.8 and 4.9). Distributions of roots affect. drainage by modifying the distribution of soil water extractions, and the. magnitude of water extracted when soil supply limits canopy evaporation. Patterns of soil water extraction determine the distribution of soil water recharge that is required prior to drainage. Warner and Young (1991) demonstrated that root channels may provide preferential flow paths for applied water, due to stem flow and macropore flow along the root and soil interface. § 1‘}. c 8' Cumulative Drainage (L ml) 3'. o M O O 8 N M O Cumulative Drainage (L m!) 128 t—oa—N 888 VI 0 A 1 A -20 -10 1204 180 A 240 300 A 360 115 Day of Year 120 180 240 3 00360115 Day of Year 30 20 1'5 10 15 (em N~‘0N 01191191191 N 91011111111110 (z-m N"0N 3) 09110111 N 91111111111110 Figure 4.13 Cumulative water and nitrate efflux 2.0 m below the soil surface observed or simulated by CERES-Maize for non-disturbed field lysimeters subjected to either no- (A) (B) leaching; or conventionnal tillage (C) cumulative drainage, tillage cumulative drainage, cumulative NO3-N (D) cumulative NO3-N leaching. Simulated runs using time- to-ponding and curve number methods of predicting infiltration, and RLDmx values of 5 or 2 cm cm‘3. 129 The seasonal pattern of drainage prior to maize crop uptake of soil water is reflected in simulated drainage of both the CT and NT lysimeters (Figure 4.13). The simulation model indicates that drainage ceases during crop growth and underestimates drainage prior to crop water uptake. The model predicts soil drainage resumes at day 240 for both NT lysimeters, nearly two months prior to observed drainage for this treatment (Figure 4.13A) . Reducing RLDm from 5.0 to 2.0 cm cm-3 did not significantly modify drainage rates. Seasonal patterns in NO3-N leaching were nearly parallel with those observed in drainage flux, reflecting similar outflow concentrations among the lysimeters. This result illustrates the sensitivity of N03-N leaching predictions to errors in predicted soil water drainage. The relatively greater N leaching losses from NT 9 result from higher’ outflow’ concentrations, nearly ‘twice that of the other lysimeters on day 122 (Figure 4.14). Inspection of cumulative leachate indicates a trend to decreasing leachate concentrations for NT lysimeters. This diminution of outflow concentration indicates the pulse of solute had already flowed through the soil prior to initial sampling on day 99, 1991 (Figure 4.13), assuming convective dispersion retarding solute flow (Jury et al., 1991). In contrast, outflow N03-N concentrations of CT lysimeters remained constant or increased slightly. This result confirms the "perched" profile distribution of soil N for CT2 (Figure 4.48) indicating the pulse of solute had not leached prior to 130 350 - . . e . l . . . . 300» 7 8:82:81?” A .. T 181.128.1811 B 250 - 200 ~ 150 . 501 Nitrate Concentration (mg NO,-N L") 1 1 1 4 a n a 1 I I 1 I I r v v 1 sample D — Integrated - Point sample " 300 - — Integrated sample .. - Point sample 200’ Nitrate Concentration (mg NO,-N L-l) 0 50 100 150 200 250 0 50 100 150 200 250 300 Cumulative Leachate (L 111-1) Cumulative Leachate (L m4) Figure 4.14 Distribution of leacheate N03-N concentrations with respect to cumulative outflow volume from non-disturbed field lysimeters subjected to either no—tillage (A) NT 6, (B) NT 9; or conventional tillage (C) CT 2, (D) CT 13. 131 initial sampling (day 99) . The fluctuations in outflow concentrations, occurring when cumulative leachate was 75 mm, corresponds with low outflow rates during crop water uptake during the period from days 150 through 300 (Figure 4.13A and 4.13C) . These fluctuations indicate vertical and/or horizontal mixing of solutes is greater at low efflux rates relative to high efflux rates. The contrasting seasonal trend in N03-N concentration in drainage outflow for NT and CT lysimeters corroborate evidence of tillage-induced effects on solute retention. Cumulative precipitation (945 mm from day 213, 1990 through day 121, 1991) and irrigation (55 mm on days 322-325, 1990) following N application (day 212, 1990) was sufficient to exchange the entire pore volume of the soil profile. (Given profile-average saturated water content of 0.36, and profile depth of 2 111, total water filled pore volume of the soil profile is 720 mm.) Thus, N03-N transport via leaching losses can account for ~the N03-N distributions observed under NT. However, retention of 49% to 61% of residual N in upper soil layers under CT (Figure 4.4) indicate solute transport characteristics of this Kalamazoo loam are modified by tillage. Water draining from the lower boundary of each lysimeter was lower for CT than NT lysimeters; though drainage was substantially reduced during crop growth for all lysimeters (Figure 4.13A and C). Soil water recharge 132 and drainage throughout crop growth were higher for NT 6 than NT 9. Nitrate leaching rates result from leacheate [NO3-N] and drainage flux from upper soil horizons. Reduced soil water recharge and increased N03-N retention by AI) and Bt horizons for CT conform to trends in tillage effects on cumulative leacheate [NO3-N] (Figure 4.14). Leacheate [N03- N] declined with cumulative water drainage for both NT lysimeters. This result indicates the peak pulse of solute passed through the 2.0 m profile prior to initial sampling on day 88, 1991. In contrast, leacheate [N03-N] under CT 13 fluctuated about a nearly constant value of 100 mg L-l; and increased from around 75 mg L-l to 150 mg L..1 for CT 2. These results suggest NO3-N transport under CT, with restricted infiltration, may be proportional to solute concentrations in upper soil layers, where NO3-N may be retained, rather than moving as a uniform front through the soil matrix. DISCUSSION Retention of N03-N in Bt horizons under CT but not NT is partially attributed to reduced infiltration and subsequent drainage. These tillage effects on the soil water budget are indicated by reduced soil water recharge (Figure 4.8A-D and Figure 4.9A and B) and cumulative drainage (Figure 4-13). Soil textural differences at horizon boundaries could affect solute retention. Nitrate 133 accumulations at the boundaries of the fine textured Bt horizon and coarse texture 3E/Bt horizon were detected among the lysimeters, though the quantities of NO3-N retained in CT and NT lysimeters differ. Thus soil textural discontinuities can account for N03-N accumulations observed at horizon boundaries; but textural discontinuities are not sufficient to account for greater NO3-N retention in Bt horizons under CT relative to NT. lelage modification of solute partitioning into micropore regions of soil water could interact with infiltration and horizon boundaries to account for the observed N retention. Solute partitioning to a micropore region of soil water is represented in a dual region solute transport model (Jury et al., 1991,); which may help interpret differences in N03-N retention and transport associated with tillage treatments. When water is held in mobile (em) or immobile (aim) regions, the convective dispersion equation (discussed in Chapter 1) can be expanded to partition solutes into mobile (Nb) and immobile (Nim) fractions: dNim de dsz deNm aim---- + 6m--- = n---§ - ------ -ra [4.15] dt dt dz dz where D is mean effective hydrodynamic dispersion, z is soil depth and JV is water flux. Solute exchange between mobile and immobile water regions is governed by concentration gradients and a transfer coefficient (a) 134 doimNim ......... = a(Nm — Him) [4.15] (11: Dual region solute transport theory suggests tillage effects on solute retention could occur especially when water infiltration and redistribution patterns are modified by tillage. Conventional tillage increased the fraction of soil water held in macropores (effective pore diameters greater than 48 um) relative to no till at the same KBS field plots reported in this study (Reinert, 1990) . Thus, more water , may be partitioned into mobile regions for CT. If the solute transfer coefficient (a) of equation [4.16) also increases with tillage, solutes could diffuse more readily into immobile soil water under CT. Given sufficient diffusion time, a substantial fraction of solute could diffuse into water held in the immobile region, proportionate to the concentration gradient between the regions. Solutes would be leached from the immobile region during infiltration events, proportionate to concentration gradients (Nm - Nim) and the residence time of water held in the mobile region. Thus, a dual 'region solute transport model suggests tillage affects N03-N retention by modifying the fraction of water held in mobile regions, and increasing the transfer coefficient for diffusion of solutes into water held in immobile regions. 135 Solute transport in leachate outflow was closely coupled with water flux in the mobile region. Thus, accuracy of solute transport predictions are sensitive to errors in predicted water flux. Factors regulating water flux in the soil matrix can shift relative to atmospheric boundary conditions and the state of soil water and corresponding transport characteristics. Accurate simulation of soil water flow, and associated solute transport, depends on correct specification of system state and boundary conditions, and the governing relationships. Errors in simulated system behaviour can result from failure to detect conditions when regulation of water flow shifts from boundary conditions to system processes. Examples of shifting regulation of water flux can be found in both infiltration and evaporation at the soil surface. The time-to-ponding (TTP) method (Chou, 1990) of partitioning rain into runoff and infiltration components explicitly defines the rain or irrigation flux intensity (boundary condition) that exceeds the soil water transport capacity (system process). Regulation of infiltration shifts from boundary condition (surface flux) to soil system process (soil water transport capacity) when rain intensity exceeds the threshold value. Similarly, regulation of soil surface evaporation shifts from boundary condition (evaporative demand) to soil system process (soil water transport capacity) when a similar threshold condition is reached (Idso et al., 1974) . The principle of regulatory 136 shifts from boundary conditions to system processes can be extended to water evaporation from a plant canopy by considering soil-root interactions as determinants of soil water transport capacity. Unfortunately, atmospheric boundary conditions are quantified more readily than soil-plant interactions. Thus simulation error is likely to increase when regulation of water flux shifts from atmospheric boundary conditions to soil water transport processes. This shift tends to occur as rain intensity exceeds soil water transport capacity, and as evaporative demand exceeds soil water transport capacity. The remaining discussion is devoted to consequences of shifts in regulation of water flux, and simulation error. Errors in simulated N03-N flux below the root zone are attributed to errors in simulated boundary conditions and soil-plant interactions. A principle error in boundary conditions for the soil water balance involves plant canopy development. Water stress effects on canopy development were not detected in simulated plant ’growth for CT lysimeters (Figure 4.6C); N stress effects on canopy development for NT lysimeters were overestimated (Figure 4.6A). These systematic biases altered simulated evaporative demand. Consequently, soil water depletion was overestimated for CT lysimeters and underestimated for NT lysimeters. Underestimating the_ quantity of water required torecharge the soil profile resulted in errors in simulated drainage and NO3-N leaching. Soil water drainage and NO3-N 137 leaching was predicted to occur nearly 60 days prior to actual field observations for NT. This result occured because drainage with matric flow typically occurs only after soil water profile is recharged. CERES-Maize underestimated the quantity of water required to recharge the soil profile, thereby forcasting drainage and NO3-N leaching prior to actual occurance. In this study, soil system processes, rather than atmospheric boundary conditions, provided significant regulation of soil-plant interactions. Soil crusts, forming on the surface of CT lysimeters, restricted infiltration, thus reducing recharge of soil water available for transpiration, relative to NT lysimeters. Nitrogen supply for the NT lysimeters derived primarily fron net N mineralization from organic matter, since mineral N was largely depleted by the prior rye crop. Soil processes modified availability and distribution of water and N, thereby influencing growth and function of root and canopy elements. Simulation of soil-plant interactions was improved when the shift in system regulation from boundary conditions to system processes was explicit. The time trend of SWC was more accurately depicted by the TTP method (defining threshold rain intensities that exceed soil water transport capacity) than by the CN method (assuming proportionate loss of rainfall to runoff). In the same way, simulated SWC was sensitive to the value assigned to maximum root length 138 density (RLDmax) only when simulated soil water depletion was restricted by soil water transport capacity. When SWC approached the lower limit of soil water supply, accuracy of simulated SWC was improved by reducing RLDmax from 5 to 2 cm cm’3, a value more consistent with field observations. We infer further gains in simulation accuracy may result when simulation structure correctly defines conditions when system regulation shifts from boundary conditions to system processes. Previous chapters of this work demonstrate soil water depletion and crop development rates are modified by the non-uniform distribution of roots. Localized soil water depletion and corresponding root regulation of canopy resistance can shift regulation of water flux from atmospheric conditions to soil-plant interactions . Simulation of the soil water balance is expected to gain accuracy when effects of non-uniform distributions of water and roots, corresponding variations in soil-root hydraulic gradients, and subsequent 'root-shoot interactions are incorporated in simulation structures. Soil-plant interactions can be predicted solely from environmental boundary conditions with some degree of accuracy when supply of growth factors (water and nutrients) is abundant. However, supply of growth factors are scarce for the large proportion of farmers with limited resources. Further, the dual objectives of optimizing productivity and nutrient retention implies 'just in time' supply of mobile 139 plant nutrients. Uhder conditions of scarcity, soil system processes, rather than environmental boundary conditions are likely to dominate regulation of soil-plant interactions. Simplified. analysis of root function, e.g. incorporating root distributions into "lumped" coefficients defining a time constant of soil water depletion, may be adequate for simulating the soil water balance under conditions of abundance. But sensitivity to heterogenous ‘ distributions of roots and soil water, and root effects on regulation of transpiration, assimilation and allocation processes may be required for simulation models that can guide water management practices under conditions of scarcity. SUNNLRY AND CONCLUSIONS The architecture or geometric arrangement of root system elements is directly related to root function. The activity of meristematic tissue in root tips represents the fundamental element of root system. development (Porter, 1989) . Initiation and growth trajectories of root tips define the structure of root distributions, and the corresponding regions of rooted soil. Localized depletion of soil water can result in restricted root water uptake prior to root extension into wet soil. Maize canopy structure can thus be altered by transient water deficits during vegetative growth. The architecture of_root systems can be understood as the dynamic modification of root tip growth modules in relation to soil structural features that interact with environmental factors to influence whole plant growth and development. The cylindrical model of water and solute flux to root surfaces is the conceptual foundation of current research in root function. Analytic solutions to cylindrical flow are simplified by assumptions of uniform root distributions and constant soil-root hydraulic gradients. Non-uniform root distributions result from. branching growth habits, root growth responses to heterogenous distributions of soil water and nutrients, and soil structural features. A horizontal 140 141 component to the root proliferation front was detected in minirhi zotron root intersections , with exponential distributions of inter-root distances. Clustered root distributions at anthesis and grain fill growth stages were quantified by spatial correlation up to 0.45 m. Soil water depletion was restricted by the volume of rooted soil during mid-vegetative water deficits. Additional solutions to cylindrical flow models may gain accuracy be re-evaluating assumptions of xylem water potential and cylinder radius boundary conditions, taking into account the heterogenous distributions of roots and soil water. A systematic bias was detected in a simplified solution to the cylindrical model of water flow to root surfaces. Predicted root water uptake, during extended drying cycles, increased with available water content at progressive soil depths. But observed soil water depletion rates decreased with soil depth prior to depletion of water in more shallow soil layers. Predictive error is attributed to the assumption of uniform soil-root hydraulic gradients, which neglects depth-dependent gradients in root xylem potential arising from frictional resistance in xylem vessels. Integrating water uptake of single root segments to the scale of multiple root segments, or root networks requires theory beyond the cylindrical flow model. Analysis of root system behaviour can be improved by considering differential radial conductivity of suberized and non-suberized regions of roots (Russell, 1977), and by considering friction- 142 induced pressure drops in root xylem potential (Klepper et al., 1983). Effects of soil and root structural geometries on soil capacitance, e.g. the change in volumetric water content with respect to a pressure gradient, suggests a fresh approach to root-soil interactions. Soil C02 profile distributions and vertical gradients in flux rates indicate soil plus root respiration violates the assumption of steady state. In situ sampling of C02 source strength requires a minimum sampling frequency of two samples per cycle of periodic fluctuation in C02 source strength (Press at al., 1986, pp.386-7). Analysis of subsequent data requires a model that distinguishes transient changes in C02 partial pressures from C0; source/sink relations. Retention of N03-N in Bt horizons of conventionally tilled, but not no-till soil is partially attributed to tillage effects on surface crusting and continuity of soil pores, resulting' in differential infiltration rates and preferential flow paths. However, the time trend of solute concentrations in leachate indicate outflow concentrations may be proportionate to solutes stored in upper soil layers. A solute transport model distinguishing mobile and immobile regions of soil water is consistent with these observations, suggesting tillage can increase the rate of solute diffusion from water held in mobile and immobile soil regions. Simulation of N leaching losses was sensitive to errors in soil water distribution, and corresponding drainage flux. 143 These errors can result when soil processes regulating infiltration and water redistribution, and soil-plant interactions regulating root water uptake are not correctly specified. Reducing simulated root length density from a maximum of 5.0 to 2.0 cm cm'3 did not alter simulation results, with the exception of root water uptake when soil water approached the lower limit of soil water supply. The time trend of simulated soil ‘water recharge in the Bt horizon was more accurately portrayed by a time-to-ponding method of partitioning rain into runoff and infiltration components. This method explicitly specifies the conditions when regulation of infiltration rates shifts from boundary conditions (rain intensity) to system processes (soil water transport capacity). Simulations of soil-plant interactions can guide water management practices for optimizing productivity and solute retention. A simplified approach to soil-plant interactions can be adequate to predict system behavior when supply of growth factors is abundant and boundary conditions regulate system processes. Analysis of complex soil-plant interactions may be required when supply of growth factors is limiting, for system behavior can be regulated by these interactions, as well as boundary conditions. Expanding our perspective on root function to consider effects on soil structure and transport processes, and on regulation of transpiration, assimilation and allocation of C can extend the validity of soileplant management simulations. 144 CONCLUSIONS 1) The front of root initiation into soil volumes has horizontal as well as vertical components for row crops, with dimensions influenced by root lateral proliferation rates and growth trajectories of root tips. 2) Clustered, non-random root distributions are characterized by spatial correlation up to 0.45 m at anthesis, and by exponential distributions of inter-root distances that decline with root proliferation. 3) Restricted root proliferation can reduce soil water depletion prior to extensive growth into wet soil, thereby modifying transpiration, C assimilation, allocation, and canopy growth processes. 4) Vertical gradients in soil water depletion, not predicted by a simplified solution to a cylindrical model of root water uptake, are attributed to vertical gradients in root xylem potential. 5) Soil distributions of C02 correspond with greater root and soil organic :matter’ accumulations, but ‘violate steady state conditions, based on flux calculations. 6) Greater retention of N03-N in f ine-textured Bt horizons of a layered soil for conventional till, but not no-till treatments suggests tillage reduced barriers to solute diffusion between mobile and immobile regions of soil water. 145 7) Deviations in N03-N concentrations of leachate from seasonal trends coincide with extreme high or low drainage flux conditions, indicating assumptions of homogeneous pore velocities and concentrations are invalid. 8) Simulated N03-N leaching losses were sensitive to errors in predicted soil water associated with infiltration, canopy dimensions, and subsequent drainage below the root zone; but not to reductions in maximal root length density from 5 to 2 cm cm'3, a value more consistent with field observations. 9) The time trend of soil water content in the Bt horizon was more accurately portrayed by a time to ponding method of partitioning rain into runoff and infiltration, explicitly defining the shift in system regulation from atmospheric boundary conditions to soil system processes. 146 REFERENCES Addiscott, T.M. and R.J. Wagenet. 1985. Concepts of solute leaching in soils: A review of modelling approaches. J. of Soil Sci. 36:411-424. Agren, G.I. and T. Ingestad. 1987. Root:shoot ratio as a balance between nitrogen productivity and photosynthesis. Plant, Cell and Environment. 10:579-586. Ahuja, L.R., D.G. DeCoursey, B.B. Barnes, and K.W. Rojas. 1991. Characteristics and importance of preferential macropore transport studied with the ARS root zone water quality model. In: T.J. Gish and A. Shirmohammadi (eds.) Preferential Flow: Proceedings of the National Symposium. ASAE. St. Joseph, MI Aiken, R.M., M.D. Jawson, K. Grahammer, and A.D. Polymenopoulos. 1991. Positional, spatially correlated and random components of variability in carbon dioxide efflux. J. Environ. Qual. 20:301-308. Aiken, R.M. 1990. Functional relations of root distributions with the flux and uptake of water and nitrate. Ph.D. Dissertation. Michigan State University, East Lansing. Allmaras, R.R. and W.W. Nelson. 1971. Corn (Zea mays L.) root configuration as influenced by some row-interrow variants of tillage and straw mulch management. Soil Sci. Soc. Amer. Proc. 35:974-980. Allmaras, R.R. and W.W. Nelson. 1973. Corn root- configuration response to soil temperature and matric suction. Agronomy J. 65:725-730. Amato, M. 1991. Spatial distribution and water uptake of roots in structured growth media. Ph.D. Dissertation. Michigan State University. East Lansing, MI. 158 p. Baker, D.B. 1988. Farm chemicals in surface waters. In Water Quality: A realistic perspective. Barber, S.A. 1971. Effect of tillage practice on corn (Zea mays L.) root distribution and morphology. Agron. J. 63:721-726. 146 147 Barber, S.A. and M. Silberbush. 1984. Plant root morphology and nutrient uptake. In: Roots, nutrient and water influx, and plant growth. SSSA-CSSA-ASA. Madison WI. Brown, D.A. and H.D. Scott. 1984. Dependence of crop growth and yield on root development and activity. In Roots, nutrient and water influx, and plant growth. SSSA-CSSA-ASA. Madison, WI. Bruckler, L., F. Lafolie, and F. Tardieu.- Modeling root water potential and soil-root water transport: II. Field comparisons. Soil Sci. Soc. Am. J. 55:1213-1220. Campbell, 6.8. 1985. Soil physics with BASIC: Transport models for soil-plant systems. Elsevier, Amsterdam Chou, Tien-Yin. 1990. Modeling infiltration using time- to-ponding and a storm generator approach. Ph. D. Dissertation. Michigan State University. East Lansing, MI. Cruz, R.T., W.R. Jordan, and M.C. Drew. 1992. Structural changes and associated reduction of hydraulic conductance in roots of Sorghum bicolor L. following exposure to water deficit. Plant Physiol. 99:203-212. Dalton, F.N., P.A.C. Raats, and W.R. Gardner. 1975. Simultaneous uptake of water and solutes by plant roots.) Soil Sci. Soc. Am. J. 67:334- Davies, W.J. and J. Zhang. 1991. Root signals and the regulation of growth and development of plants in drying soils. Ann. Rev. Plant Physiol. 42:55-76. de Jong, E., H.J.V. Schapport and R.B. Macdonald (1974). Carbon dioxide evolution from virgin and cultivated soil as affected by management practices and climate. Can. J. Soil Sci. 54:299-307. Ellis, B.G. 1988. Nitrates in water supplies. In Water Quality: A realistic perspective. Elliot, E.T. 1986. Aggregate structure and carbon, nitrogen and phosphorus in native and cultivated soils. Soil Sci. Soc. Am. J. 50:627-633. Ferguson, J.C. and A.J.M. Smucker. 1989. Modification of the minirhizotron video camera system for measuring spatial-and temporal root dynamics. Soil Sci. Soc. Am. J. 53:1601-1605. Fortin, M-C A. and K. Poff. 1990. Temperature sensing by primary roots of maize. Plant Physiol. 94:367-369. 148 Gardner, W.R. 1960. Dynamic aspects of water availability to plants. Soil Sci. 89:63- Gardner, W.R. 1964. Relation of root distribution to water uptake and availability. Agron. J. 56:41- Gardner, 1986. Soil water measurement. In Methods of soil analysis, Part 1. Physical and mineralogical methods. No. 9. Amer. Soc. of Agron. Madison, WI. Gardner, W.R. 1991. Modeling water uptake by roots. Irrig. Sci. 12:109-114. Granato, T.C. and C.D. Raper, Jr. 1989. Proliferation of maize (Zea mays L.) roots in response to localized supply of nitrate. J. Exp. Bot. 20(211):263-276. Hainsworth, J.M. and L.A.G. Aylmore. 1989. Non-uniform soil water extraction by plant roots. Plant and Soil 113:121-124. Hall, A.J., D.J. Connor, and D.M. Whitfield. 1990. Root respiration during grain filling in sunflower: The effects of water stress. Plant and Soil 121. 57-66. Hallberg, G.R. 1987. Agricultural chemicals in ground water: Extent and implications. Amer. J. of Alt. Agric. Vol. II, No. 1:3-15. Hilbert, D.W. 1990. Optimization of plant root:shoot ratios and internal nitrogen concentration. Annals of Botany. 66:91-99. Hilbert, D.W., A. Larigauderie, and J.R. Reyonds. 1991. The influence of carbon dioxide and daily photon-flux density on optimal leaf nitrogen concentration and root:shoot ratio. Annals of Botany. 68:365-376. Hillel, D. and H. Talpaz. 1976. Effect of root growth parameters on the pattern of soil moisture extraction by non-uniform root systems. Soil Sci. 121:307-312. IBSNAT Technical Report 5. 1985. Decision Support System for AgrotechnOlogy Transfer (DSSAT). International Benchmark Sites Network for Agrotechnology Transfer. Univeristy of Hawaii. Honolulu HI. Idso, S.B., R.J. Reginato, R.D. Jackson, B.A. Kimball, and F.S. Nakayama. 1974. The three stages of drying of a field soil. Soil Sci. Soc. Am. Proc. 38:831-837. Jones, C.A. 1983. Effect of soil texture on critical bulk densities for root growth. Soil Sci. Soc. Am. J. 47:1208-1211. 149 Jones, C.A., J.T. Ritchie, J.R. Kiniry, and D.C. Godwin. 1986. Subroutine structure. In C.A. Jones and J.R. Kiniry (eds.) CERES-Maize: A simulation model of maize growth and development. Texas A&M University Press. College Station, TX. Jury, W.A., W.R. Gardner, W.H. Gardner. 1991. Soil Physics. Fifth edition. John Wiley and Sons, New York. pp 218-242. Kachi, N. and I.H. Rorison. 1989. Optimal partitioning between root and shoot in plants with contrasted growth rates in response to nitrogen availability and temperature. Functional Ecology. 3:549-559. Kanwar. R.S. 1991. Preferential movement of nitrate and herbicides to shallow growndwater as affected by tillage and crop rotation In: T.J. Gish and A. Shirmohammadi (eds.) Preferential Flow: Proceedings of the National Symposium. ASAE. St. Joseph, MI Klepper, B. 1987. Origin, branching and distribution of root systems. In: Gregory, P.J., J.V. Lake, and D.A. Rose (eds.) Root development and function. pp 103-124. Klepper, B., R.W. Rickman, and H.M. Taylor., 1983. Farm management and the function of field crop root systems. Agric. Water Manage. 7:115-141. Kooistra, M.J., D. Schoonderbeek, F.R. Boone, B.W. Veen and M. Van Noordwijk. 1992. Root-soil contact of maize, as measured by a thin-section technique. II Effects of soil compaction. Plant and Soil 139:119-129. Kuchenbuch R.O. and S.A. Barber. Significance of temperature and precipitation for maize root distribution in the field. Plant and Soil.106:9-14. Kuiper, P.J.C. 1987. Response of roots to the physical environment: Goals for future research. In: P.J. Gregory, J.V. Lake, D.A. Rose (eds.) Root development and Function. Lafolie, F., L. Bruckler, and F. Tardieu. 1991. Modeling root water potential and soil-root water transport: I. Model presentation. Soil Sci. Soc. Am. J. 55:1203-1212. Logsdon, S.D. and R.R. Allmaras. 1991. Maize and soybean root clustering as indicated by root mapping. Plant 150 Luxmore, R.J. and L.H. Stolzy. 1987. Modeling belowground processes of roots, the rhizosphere and soil communities. In: K. Wisiol and J.D. Hesketh (eds.) Plant growth modeling for resource management. Vol II. CRC Press, Boca Raton, FL. Manetsch, T.J. 1976. Time-varying distributed delays and their use in aggregative models of large systems. IEEE Transactions on Systems, Man and Cybernetics, Vol. SMC-6, No. 8:547-553. Manetsch, T.J. and G.L. Park. 1987. System analysis and simulation with applications to economic and social systems. Part I - Methodology, modeling and linear system fundamentals. Fifth duplicated edition. Michigan State University. East Lansing, MI. Martin, J. 1987. Carbon flow through the rhizosphere of cereal crops--A review. INTECOL Bulletin 15:17-23. Masle, J. and J.B. Passioura. 1987. The effect of soil strength on the growth of young wheat plants. Aust. J. Plant Physiol. 14:643-656. NeSmith, D.S. and J.T. Ritchie. 1992. Effects of soil water deficits during tassel emergence on development and yield component of maize (Zea mays L.). Field Crops Res. 28:251-256. Ogata, G., L.A. Richards, and W.R. Gardner. 1960. Transpiration of alfalfa determined from soil water content changes. Soil Sci. 89:179 - Passioura, J.B. 1983. Roots and drought resistance. Agricultural Water Management 7:265-280. Passioura, J.B. 1985. Roots and water economy of wheat. In: Wheat growth and modelling. W.Day, R.K. Atkin (ed.) pp. 185-198. Plenum: New York. Passioura, J.B. 1991. Soil structure and plant growth. Aust. J. Soil Res. 29:717-728. Paul, E.A. and F.E. Clark. 1989. Soil microbiology and biochemistry. Academic Press. San Diego. Porter, J.R. 1989. Modules, models and meristems in plant architecture. In: G. Russell, B. Marshall, and P.G. Jarvis (ed.) Plant canopies: Their growth, form and function. pp. 63-84. Cambridge University Press. Cambridge. Press, W.H., B.P. Flannery, S.A. Teukolsky, and W.T. Vetterling. 1986. Numerical Recipies. Cambridge University Press. Cambridge. 151 Protopapas, A.L. and R.L. Bras. 1987. A model for water uptake and development of root systems. (Soil Sci. 144:352- 366. Reinert, D.J. 1990. Soil structural form and stability induced by tillage in a typic hapludalf. Ph.D. Dissertation. Michigan State University, East Lansing. Ritchie, J.T. 1981. Soil water availability. Plant and Soil 58:327-338. Ritchie, J.T. 1985. A user-oriented model of the soil water balance in wheat. In: Wheat Growth and Modelling. W. Day and R.K. Atkin (eds.) Plenum Press, New York. Ritchie, J.T. and J. Crum. 1989. Converting soil survey characterization data into IBSNAT crop model input. p. 155- 167. In: J. Bouma and A.K. Brget (eds.) Land qualities in space and time. Wageningen, The Netherlands. Ritchie, J.T. and D.S. NeSmith. 1991. Temperature and crop development. In: Hanks and Ritchie (eds.) Modeling plant and soil systems. Agronomy 31:5-29. Ritchie, J.T., A. Nunez-Barrios, A.J.M. Smucker, and D.S. NeSmith. Soil water deficits and the root dynamics of maize (Zea Hayes L.): Late growth stages. (In preparation) Robertson, G.P. and A.J.M. Smucker. 1989. Process level interactions in terrestrial ecosystems: An experimental approach. Final report to National Science Foundation. NSFBSR- Robertson, M.J., S. Fukai, M.M. Ludlow, and G.L. Hammer. Water extraction by grain'sorghum in a sub-humid environment. II. Extraction in relation to root growth. Field Crops Research (In press). Rose, D.A. 1983. The description of the growth of root systems. Plant and Soil 75:405-415. . Rosenberg, N.J., B.L. Blad, and 8.8. Verma. 1983. Microclimate the biological environment. 2nd Edition. Wiley Interscience. New York. Russell, R.S. 1977. Plant root systems. Their function and interaction with the soil. McGraw-Hill, London. Schumacher, T.E. and A.J.M. Smucker. 1985. Carbon transport and root respiration of Phaseolus vulgaris subjected to short term localized anoxia. Plant Physiol.. 78:359-364. 152 Siddique, K.H.M., R.K. Belford, and D. Tennant. 1990. Root:shoot ratios of old and modern, tall and semi-dwarf wheats in a mediterranean environment. Plant and Soil 121:89-98. Smucker, A.J.M. 1984. Carbon utilization and losses by plant root systems. In: Roots, nutrient and water influx, and plant growth. pp 27-46. A.S.A. Special Publications No. 49. SSSA-CSSA-ASA Madison, WI. Smucker, A.J.M. 1990. Quantification of root dynamics in agroecological systems. Remote Sensing Reviews. 5(1):237-248. Steel, R.G.D. and J.H. Torrie. 1980. Principles and procedures of statistics. 2nd Edition. McGraw-Hill. New York. Tanner, C.B. and T.R. Sinclair. 1983. Efficient water use in crop production: Research or re-search. pp. 1-26. In: H.M. Taylor, W.R. Jordan, and T.R. Sinclair (ed.) Limitations to efficient water use in crop production. ASA- CSSA-SSSA. Madison, WI. Tardieu, F. 1988a. Analysis of spatial variability of maize root density. I. Effect of wheel compaction on the spatial arrangement of roots. Plant and Soil. 107:259-266. Tardieu, F. 1988b. Analysis of spatial variability of maize root density. II. Distances between roots. Plant and Soil. 107:267-272. ' Tardieu, F. 1988c. Analysis of spatial variability of maize root density. III. Effect of wheel compaction on water extraction. Plant and Soil. 109:257-262. Tardieu, F. and S. Pellerin. 1991. Influence of soil temperature during root appearance on the trajectory of nodal roots of field grown maize. Plant and Soil 131:207- 214. Taylor, H.M. and B. Klepper. 1975. Water uptake by cotton root systems: An examination of assumptions in the single root model. Soil Sci. 120:57- Thornley, J.H.M. 1972. A balanced quantitative model for root:shoot ratios in vegetative plants. Annals of Botany. 36:431-441. Thornley, J.H.M. 1991. A transport-resistance model of forest growth and partitioning. Annals of Botany. 68:211- 226. 153 Tilman, D. 1988. Plant strategies and the dynamics and structure of plant communities. Princeton University Press. Princeton. Topp, G.C., J.L. Davis, and A.P. Annan. 1982. Electromagnetic determination of soil water content using TDR: 1. Applications to wetting fronts and steep gradients. Soil Sci. Soc. Am. J. 46:672-678. Uehara, G. 1985. The International Benchmark Sites Network for Agrotechnology Transfer (IBSNAT). In: W. Day and R.K. Atkin (ed) Wheat growth and modeling. Plenum Press,New York. P. 271-274. Upchurch, D.R. and J.T. Ritchie. 1983. Root observations using a video recording system in mini- rhizotrons. Agron. J. 75:1009-1015. van Keulen, H., J. Goudriaan and N.G. Seligman. 1989. Modelling the effects of nitrogen on canopy development. In:G. Russell, 8. Marshall, and P.G. Jarvis (ed.) Plant canopies: Their growth, form and function. pp. 63-84. Cambridge University Press. Cambridge. van Noordwijk, M., G. Brouwer, and K. Harmanny. Concepts and methods for studying interactions of roots and soil structure. Geoderma. (In press). Vieira, S.R., J.L. Hatfield, D.R. Nielsen, and J.W. Biggar. 1983. Geostatistical theory and application to variability of some agronomical properties. Hilgardia 51(3). Wang, J., J.D. Hesketh, and J.T. Woolley. 1986. Preexisting channels and soybean rooting patterns. Soil Sci. 141:432-437. Warner, G.S. and R.A. Young. 1991. Measurement of preferential flow beneath mature corn. In: T.J. Gish and A. Shirmohammadi (ed.) Preferential Flow: Proceedings of the National Symposium. ASAE. St. Joseph, MI Warrick, A.W., D.B. Myers, and D.R. Nielsen. 1986. Geostatistical methods applied to soil science. p 53-82. In: Methods of soil analysis, Part 1. Physical and mineralogical methods. No. 9. Amer. Soc. of Agron. Madison, WI. Yamauchi, A., H.M. Taylor, D.R. Upchurch, and B.L. McMichael. (In Press) Axial resistance to water flow of intact cotton taproots. APPENDIX Initial and boundary conditions for CERES-Maize v2.1 simulation of soil-plant interactions (IBSNAT, 1985). Weather Input File "MSKB9101.M21 LOC YR DAY Rg Tmax Tmin Ppt TTP MSKB 91 121 12.5 13.0 7.7 0.0 0.0 MSKB 91 122 24.7 13.6 4.1 0.0 0.0 MSKB 91 123 25.6 16.7 1.0 0.0 0.0 MSKB 91 124 20.0 18.7 6.6 0.0 0.0 MSKB 91 125 4.1 14.9 6.8 6.0 6.0 MSKB 91 126 4.3 14.6 5.7 0.7 0.7 MSKB 91 127 21.8 13.9 3.3 0.0 0.0 MSKB 91 128 12.4 16.4 3.1 0.0 0.0 MSKB 91 129 25.0 26.4 7.7 0.0 0.0 MSKB 91 130 24.9 26.5 9.8 0.0 0.0 MSKB 91 131 21.8 28.9 12.0 0.0 0.0 MSKB 91 132 24.3 31.0 15.4 0.0 0.0 MSKB 91 133 20.9 30.9 14.9 0.0 0.0 MSKB 91 134 25.4 29.7 14.0 0.0 0.0 MSKB 91 135 27.8 32.7 10.5 0.0 0.0 MSKB 91 136 20.8 31.7 15.4 0.0 0.0 MSKB 91 137 14.2 24.5 6.7 0.7 0.7 MSKB 91 138 18.5 16.0 6.2 0.0 0.0 MSKB 91 139 27.5 22.2 7.2 16.5 6.5 MSKB 91 140 26.4 27.3 7.9 0.0 0.0 MSKB 91 141 24.5 29.6 10.4 0.0 0.0 MSKB 91 142 19.0 30.3 15.4 0.0 0.0 MSKB 91 143 11.3 27.2 19.2 9.6 9.6 MSKB 91 144 16.9 28.4 19.7 0.5 0.5 MSKB 91 145 9.2 27.1 19.7 2.5 2.5 MSKB 91 146 15.3 26.6 21.1 1.3 1.3 MSKB 91 147 23.8 29.9 20.3 1.0 1.0 MSKB 91 148 24.7 32.3 19.5 0.0 0.0 MSKB 91 149 21.3 33.5 21.8 0.0 0.0 MSKB 91 150 23.5 31.1 19.5 0.0 0.0 MSKB 91 151 24.1 31.6 19.5 18.0 18.0 MSKB 91 152 17.6 31.2 17.4 1.7 1.7 MSKB 91 153 19.0 27.8 17.7 11.4 11.0 MSKB 91 154 24.7 29.0 13.8 0.0 0.0 MSKB 91 155 29.75 28.3 16.1 0.0 0.0 MSKB 91 156 23.97 28.3 9.4 0.0 0.0 MSKB 91 157 27.37 23.8 8.3 0.0 0.0 MSKB 91 158 20.59 26.1 8.8 0.0 0.0 MSKB 91 159 28.63' 26.6 13.3 0.0 0.0 MSKB 91 160 23.97 27.2 12.7 0.0 0.0 MSKB 91 161 18.23 28.8 12.7 0.0 0.0 154 MSKB HSKB MSKB MSKB HSKB MSKB MSKB MSKB MSKB HSKB MSKB MSKB MSKB HSKB HSKB MSKB MSKB MSKB MSKB HSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB NSKB 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 13.26 28.19 29.03 25.70 18.72 17.34 31.11 26.22 29.13 29.05 27.64 4.26 30.57 25.78 26.97 27.70 25.54 23.85 25.35 24.39 18.98 24.98 26.80 23.30 27.10 23.94 17.72 29.21 26.59 23.69 28.49 10.45 9.54 27.39 29.06 25.26 21.65 23.29 26.01 22.87 15.61 16.97 28.27 26.33 26.40 24.20 24.99 20.17 3.94 21.92 25.87 26.85 20.00 8.81 155 28.8 26.1 28.3 28.3 31.6 31.1 29.4 28.3 30.5 30.5 31.6 32.2 32.2 24.4 27.2 30.5 31.1 31.6 31.1 32.2 32.7 30.5 30.5 29.9 29.4 34.9 31.1 29.4 30.5 26.6 27.2 29.4 28.8 27.7 26.6 27.7 29.4 29.4 31.6 32.2 32.2 32.2 31.6 32.2 26.1 24.9 24.4 26.6 26.6 25.5 23.8 28.3 29.4 31.6 15.5 18.3 14.9 11.6 15.5 18.8 18.8 13.3 15.5 16.6 17.2 17.2 13.8 10.5 11.6 16.1 16.1 20.5 20.5 22.2 23.3 17.7 19.4 19.4 18.3 18.8 17.2 22.2 19.9 11.6 14.9 16.1 16.1 16.1 15.5 13.3 14.4 16.1 19.4 21.1 23.3 21.6 21.1 21.6 15.5 14.9 11.1 10.5 14.9 14.9 14.4 12.7 18.8 16.1 1.1 p NDOOOGOOOOOOQUOOOOOOHU01°OOOO‘OObUIl-‘HOOONOOO‘OUOOOOOOWOOOO CO... OOOOOOOOOOOOOOOOOOOOOO0.0000000000000000000000000 UI OOOOOOOOOOOOCOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO 1.1 H ~11—1 1.1 ' ura 010::c>o¢nc>o<3c>o<3u>u<3c>o<3c>01aoaH<3c>o<301ocae-otac>o<3c>u<3c>m<3010<3c>o<3c>m<3c>o<3 00000000000000000o0000000000000...00000000009000.0000o UIOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOUIOOOOOOOOOOOOOOOOOOOOOO MSKB MSKB MSKB NSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB HSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB HSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB MSKB 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 91 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 21.43 23.30 25.32 21.17 1.84 22.85 24.38 25.82 23.16 22.42 21.59 20.35 20.61 11.90 16.84 2.71 21.28 21.28 22.09 23.66 23.19 22.19 21.51 21.46 21.40 15.02 15.34 20.42 20.86 11.83 4.30 18.69 17.95 17.91 18.23 18.36 17.43 12.04 18.38 8.06 8.39 6.14 15.83 7.99 17.79 11.25 10.46 12.35 19.69 6.09 15.35 6.89 3.95 13.47 156 31.6 24.4 23.8 26.1 27.2 26.6 26.1 26.6 27.7 27.7 28.3 28.3 28.3 28.8 28.8 26.6 21.6 24.9 26.1 29.4 28.3 28.8 31.6 31.1 30.5 32.2 32.2 31.1 29.4 25.5 27.7 27.7 26.6 25.5 27.7 29.9 30.5 31.6 29.9 26.6 22.2 22.7 23.8 29.9 29.4 24.4 22.2 16.1 13.3 17.2 17.7 16.6 14.9 13.3 #01) U euoeeooouoouwtooommoooooqooooooooooooooHuuoooooooowooooa OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOON 1.1 8.1-1 buoeeooouooutotooommoooooqooooooooooooooHmuooooooooraooooo O.COO...0.0.0....CO0.00000000000000000000000000000 .OOO OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOUIOOOOOOOOOOOOON 157 MSKB 91 270 15.34 12.7 6.6 0.0 0.0 MSKB 91 271 18.33 13.3 1.1 0.0 0.0 MSKB 91 272 16.34 17.2 -1.1 0.0 0.0 MSKB 91 273 15.33 19.9 -1.6 0.0 0.0 MSKB 91 274 11.75 25.5 7.7 0.0 0.0 Rg is global radiation (Mg m.'2) Tmax is maximum daily temperature (°C) Tmin is minimum daily temperature (°C) Ppt is daily precipitation (mm) TTP is time-to-ponding estimate of infiltration for crusted soil, conventional tillage. This method assumes soil intake of the initial 2 mm of precipitation but runoff for rainfall exceeding 3 (Aiken, 1992, pp 105- 106) 158 Soil Input File "MSKB9101.MZ2 98 CT Lys #2 Kalamazoo loam, mixed, mesic Udic Haplustoll 84.0 8.5 13.5 1. 0.13 7. 0.4' 5. 5. 11. 11. 11. 11. 12. 12. 20. 21. 21. 21. 21. 21. -1. 97 CT Lys #13 Kalamazoo loam, mixed, mesic Udic Haplustol 0.13 7. 0.4 84. 8.5 13.5 1. 5. 6. 11. 10. 10. 10. 10. 11. 11. 20. 20. 20. 20. 20. 19. -1. 0.13 0.13 0.13 0.13 0.13 0.10 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.13 0.13 0.13 0.22 0.19 0.15 0.11 0.09 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.33 0.33 0.33 0.27 0.26 0.24 0.18 0.14 0.14 0.14 0.14 0.14 0.14 0.14 0.33 0.33 0.33 0.31 0.31 0.31 0.27 0.27 0.22 0.11 0.11 0.11 0.11 0.11 0.11 0.41 0.41 0.41 0.42 0.42 0.42 0.37 0.37 0.34 0.34 0.34 0.34 0.34 0.34 0.43 0.43 0.43 0.40 0.40 0.40 0.40 0.38 0.38 0.34 0.34 0.34 0.34 0.34 0.34 0.31 0.31 0.31 0.27 0.27 0.28 0.22 0.15 0.14 0.14 0.14 0.14 0.14 0.14 0.31 0.31 0.31 0.32 0.31 0.31 0.30 0.26 0.20 0.11 0.11 0.11 0.11 0.11 0.11 1.00 1.00 0.81 0.64 0.49 0.36 1.00 1.00 0.81 0.64 0.49 0.36 0.25 0.14 0.05 0.01 00000 1.5 1.5 1.5 1.46 1.46 1.46 1.55 1.52 1.52 1.52 2.8 OOOOOOOOOOOHN 0000000000000 OOOOHNUIQ‘ODDQQ HHHm 2.67E-3 58. 6.68 2.8 oooooeooooo OOHNUQW‘O‘OQG OOOOOOOOOOHN O O 1.1 2.67E-3 58.0 6.68 0.03 1.0 3.0 41.9 3.0 41.9 3.0 41.9 8.4 84.9 5.0 116. 3.8 153. 2.9 94.9 2.1 57.1 2.0 43.6 1.8 37.4 1.5 31.0 1.4 24.8 1.3 18.4 1.2 12.1 0.03 1.0 0.01 ‘.01 0.01 .01 0.01 .01 0.01 .01 0.01 4.9 0.5 12.0 0.5 18.6 0.5 21.6 0.4 17.1 0.1 2.6 1.1 6.7 1.1 6.7 1.1 6.7 1.1 6.7 1.1 6.6 5.0 mmmmmmmmmmmmmax UIUIUIUIUIU'IUIUIUIUIUIUIUIUIUIOH UMU'IUIUIUIUIWWUIWUIUIUI 5 6 6 6 6 6 6 6 6 6 6 6 6 6 6 6 159 96 NT Lys #6 Kalamazoo loam, mixed, mesic Udic Haplustoll 0.13 9. 0.6 78. 8.5 13.5 1. 5. 5. 10. 9. 10. 10. 10. 10. 10. 11. 11. 25. 25. 25. 27. -1. 95 NT Lys #9 Kalamazoo loam, mixed, mesic Udic Haplust 0.14 0.14 0.14 0.18 0.20 0.20 0.20 0.09 0.08 0.05 0.05 0.05 0.05 0.05 0.05 0.31 0.31 0.31 0.35 0.35 0.34 0.34 0.22 0.19 0.14 0.14 0.14 0.14 0.14 0.14 0.42 0.42 0.42 0.40 0.38 0.38 0.38 0.34 0.34 0.34 0.34 0.34 0.34 0.34 0.34 0.31 0.31 0.31 0.32 0.33 0.31 0.31 0.25 0.16 0.14 0.14 0.14 0.14 0.14 0.14 0.13 9. 0.6 78. 8.5 13.5 1. 5. 5. 10. 10. 8. 8. 11. 14. 14. 20. 20. 20. 20. 20. 18. -1. 0.14 0.14 0.14 0.20 0.20 0.20 0.09 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.05 0.31 0.31 0.31 0.35 0.36 0.33 0.22 0.19 0.19 0.14 0.14 0.14 0.14 0.14 0.14 0.41 0.41 0.41 0.40 0.37 0.37 0.36 0.36 0.36 0.34 0.34 0.34 0.34 0.34 0.34 0.31 0.31 0.31 0.32 0.35 0.33 0.22 0.18 0.16 0.14 0.14 0.14 ,0.14 0.14 0.14 1.00 1.00 0.92 0.82 0.73 0.63 0.52 1.00 1.00 0.92 0.82 0.73 0.63 0.52 0.35 0.20 1.49 1.49 1.49 1.54 1.50 1.50 1.50 1.54 1.61 2.67E-3 58. 6.68 2.67E-3 58. 6.68 0.03 1.0 5.0 2.8 0.01 0.04 6.5 2.8 0.01 0.04 6.5 1.7 0.01 0.04 6.5 0.9 0.1 0.04 6.5 0.9 0.01 0.01 6.5 0.9 0.01 0.02 6.5 0.7 0.4 0.03 6.5 0.5 0.3 0.4 0.01 3.7 6.5 0.2 0.01 0.5 6.5 0.1 0.01 0.7 6.5 0.01 0.2 1.7 6.5 0.01 0.7 3.5 6.5 0.01 0.9 5.3 6.5 0.01 1.1 7.1 6.5 011 0.03 1.0 5.0 2.8 0.1 0.06 6.5 2.8 0.1 0.06 6.5 1.7 0.01 0.06 6.5 0.9 0.1 0.06 6.5 0.9 0.01 0.01 6.5 0.9 0.1 0.01 6.5 0.7 0.01 0.01 6.5 0.5 0.01 0.01 6.5 0.4 0.01 0.01 6.5 0.2 0.2 1.3 6.5 0.1 0.6 4.3 6.5 0.01 1.0 7.4 6.5 0.01 1.2 10.5 6.5 0.01 1.7 13.5 6.5 0.01 2.1 16.4 6.5 - .. -... 11.41.... 71111111111.-. . 11.11 1.11:3... . .. 111...: . 11.1-1 . . .lttilt . .1 1 .. I... 111211 . 1 1 -.1 .. . 1.111111 r ...1 I ... t u . . _. . 4.... 1’ I :11 -11 r’.! 001 . ..-.t .-.. . .-l t . .1.) .. - -0.) ...11...t . .... . . . .-.- ... Hilliatt. " .....tfi. . .llt v.11. . .Ilr .. -. . ”21.21.11... . 1-1-1111 .1 1 ..--.. t . 1- -.I. V. a. a! . ; . .. :‘Jejoll..:1-.1J:" -‘ . '15-‘1‘1' . ll. _ l"."';".[“ ‘j',: i‘:rl:: “;.' 1111' 4 1 . a. ‘91 . I. o. ...... ...1 ... . . 11v -.- ..- "$1.41....riilfl. .l). ...: . 5.11.. ....- . . 1.131...- ... -. I c- 1.111.151... . . .I.. 1 . . . 1 ....ll . . : . C . .V’JHI‘4436414 "thrill . . . . v . .... . .. . l . 1 at 41.1..1 .. 1“. v . 0.0.... 1:41: 1.11:. 1.1. Ijéiiitlg' 1 191:1J‘4': 73:05:11. - 1 ..3 n l a I: 11:. 1.9).. 'i :11 ,1. (01...» 1 1 ..... .411 ”v.1 tl‘tl. 131111} :1.. ' .- 11111111 4191101 11014 . lo'.ll.. olt‘ 1‘ .5 .-.-5 .gtiiti.‘ 1.1-1: . . 1 Rea-11.. iiitief 1 .r‘u. envtot.lt 11140:. 04111 . 1. . cl. 11.1.4.1. (.....llia. .4 ..o . .. 11.1.1.1:6: 1 . ... . . . . . .... . a .14 6.4.9.13. 0"“.- 61:..101l1': ..pnrm..1.fih . 2 .1 .al. .azeaeuull. "Mann... «9.1118 1 .. .... . ...:5II]!|II‘-..O. 14.4141411440 nit 1.40-1.‘ t. u. ... 41.11.10.121! ... .. .‘T . 2' .401 . . 144-0., v 110.71 .. 3. .... 1|». .vt 6.9.341.- . . .. . 3.! .. .111 04.10.! 10.1114 19431144421134 «121.. 11 111.11. .1... 11.1.1. 11...”.14. new“... .1 av . L11 1.1.! 2.1411... 1 147.! 541.1%qu .... 31.. 4| . t I! n . . 10 l .010... .11...“ ".01.... v “Nation-.143?! Ill . 1.. 1. .11 ill. v .. ... 1.1a 3.11.1!!!)118 u I .4 1. a. 104.16.. vvlflo .I..i§4440.'0 - . ... 121.! .1: . ......t 11......“ 311.14.“... 9.1311... .... "a