Improving Crop Growth Simulation in the Hydrologic Model DRAINMOD to Simulate Corn Yields in Subsurface Drained Landscapes Ranvir Singh University of Western Sydney

Agricultural and Biosystems Engineering Agricultural and Biosystems Engineering Conference Proceedings and Presentations

6-2008 Improving Crop Growth Simulation in the Hydrologic Model DRAINMOD to Simulate Corn Yields in Subsurface Drained Landscapes Ranvir Singh University of Western Sydney

Matthew .J Helmers Iowa State University, [email protected]

Follow this and additional works at: http://lib.dr.iastate.edu/abe_eng_conf Part of the Agriculture Commons, and the Bioresource and Agricultural Engineering Commons The ompc lete bibliographic information for this item can be found at http://lib.dr.iastate.edu/ abe_eng_conf/263. For information on how to cite this item, please visit http://lib.dr.iastate.edu/ howtocite.html.

This Conference Proceeding is brought to you for free and open access by the Agricultural and Biosystems Engineering at Iowa State University Digital Repository. It has been accepted for inclusion in Agricultural and Biosystems Engineering Conference Proceedings and Presentations by an authorized administrator of Iowa State University Digital Repository. For more information, please contact [email protected]. Improving Crop Growth Simulation in the Hydrologic Model DRAINMOD to Simulate Corn Yields in Subsurface Drained Landscapes

Abstract The primary goal of drainage research has been shifting from maximizing crop production to environmental impacts with the increasing concern related to the transport of nutrients, specifically nitrate-nitrogen (NO3-N) with subsurface drainage water from agricultural lands. It is becoming important to evaluate the impact of drainage design and its management not only on crop production but also on nutrients, primarily nitrogen transformation and transport from agricultural lands. The yh drologic models DRAINMOD and DRAINMOD-NII simulate subsurface drainage design and management, and its impact on N-transformation and transport from artificially drained soils. However, both models use a simplified yield reduction approach to simulate the crop growth, its production and impact on hydrology and nutrient dynamics of subsurface drainage landscapes. The bjo ective of this study, therefore, was to integrate a deterministic crop model CERES Maize into DRAINMOD as an option for the simulation of detailed corn growth and development accounting for weather and soil water dynamics in field conditions. The integrated DRAINMOD and CERES Maize model was evaluated by comparing the simulations with observations from three (Readlyn; Kenyon and Floyd) soil plots located at the Iowa State University Northeast Research Center, Nashua (IA). The preliminary model results are encouraging showing a good correspondence with the observed soil water content, subsurface drainage and crop yields during the years from 1990 to 1992. The integration of detailed crop models into DRAINMOD would capitalize on their strengths, and enhance the capability of modelling the subsurface drainage systems.

Keywords CERES Maize; Subsurface drainage; Drainage water management; Midwest United States

Disciplines Agriculture | Bioresource and Agricultural Engineering

Comments This is an ASABE Meeting Presentation, Paper No. 083571.

This conference proceeding is available at Iowa State University Digital Repository: http://lib.dr.iastate.edu/abe_eng_conf/263

An ASABE Meeting Presentation

Paper Number: 083571

Improving Crop Growth Simulation in the Hydrologic Model DRAINMOD to Simulate Corn Yields in Subsurface Drained Landscapes Ranvir Singh, PhD CRC For Irrigation Futures, University Of Western Sydney, NSW AUSTRALIA. Matthew J. Helmers, PhD Agricultural and Biosystems Engineering, Iowa State University, Ames (IA) 50011.

Written for presentation at the 2008 ASABE Annual International Meeting Sponsored by ASABE Rhode Island Convention Center Providence, Rhode Island June 29 – July 2, 2008 Abstract. The primary goal of drainage research has been shifting from maximizing crop production to environmental impacts with the increasing concern related to the transport of nutrients, specifically nitrate-nitrogen (NO3-N) with subsurface drainage water from agricultural lands. It is becoming important to evaluate the impact of drainage design and its management not only on crop production but also on nutrients, primarily nitrogen transformation and transport from agricultural lands. The hydrologic models DRAINMOD and DRAINMOD-NII simulate subsurface drainage design and management, and its impact on N-transformation and transport from artificially drained soils. However, both models use a simplified yield reduction approach to simulate the crop growth, its production and impact on hydrology and nutrient dynamics of subsurface drainage landscapes. The objective of this study, therefore, was to integrate a deterministic crop model CERES Maize into DRAINMOD as an option for the simulation of detailed corn growth and development accounting for weather and soil water dynamics in field conditions. The integrated DRAINMOD and CERES Maize model was evaluated by comparing the simulations with observations from three (Readlyn; Kenyon and Floyd) soil plots located at the Iowa State University Northeast Research Center, Nashua (IA). The preliminary model results are encouraging showing a good correspondence with the observed soil water content, subsurface drainage and crop yields during the years from 1990 to 1992. The integration of detailed crop models into DRAINMOD would capitalize on their strengths, and enhance the capability of modelling the subsurface drainage systems.

The authors are solely responsible for the content of this technical presentation. The technical presentation does not necessarily reflect the official position of the American Society of Agricultural and Biological Engineers (ASABE), and its printing and distribution does not constitute an endorsement of views which may be expressed. Technical presentations are not subject to the formal peer review process by ASABE editorial committees; therefore, they are not to be presented as refereed publications. Citation of this work should state that it is from an ASABE meeting paper. EXAMPLE: Author's Last Name, Initials. 2008. Title of Presentation. ASABE Paper No. 08----. St. Joseph, Mich.: ASABE. For information about securing permission to reprint or reproduce a technical presentation, please contact ASABE at [email protected] or 269-429-0300 (2950 Niles Road, St. Joseph, MI 49085-9659 USA).

Keywords. CERES Maize; Subsurface drainage; Drainage water management; Midwest United States

Introduction In recent decades researchers have been devoting much effort to developing hydrologic models to analyze environmental and water use problems. These efforts are justifiable by the cost and time effective application of hydrologic models to understand the cause-effect relationships between hydrological and biophysical variables in the soil-water-atmosphere-plant continuum. They are useful for understanding crop water requirements, crop growth, drainage design, and transformation and transport of nutrients, solutes and pesticides. Examples of such models are DRAINMOD (Skaggs, 1978), RZWQM (Ahuja et al., 1999), ADAPT (Agricultural Drainage and Pesticide Transport Model) (Alexander, 1988), SWAP (Soil Water Atmosphere Plant) (Van Dam et al., 1997), HYDRUS (Simunek et al., 2005), and DSSAT (Jones et al., 2003), etc. With the included theoretical and practical knowledge, each model has its strength for simulation of certain processes such as RZWQM for water quality, ADAPT for drainage, nutrient cycle and pesticide transport, HYDRUS and SWAP for unsaturated soil water flow and DSSAT for crop- water-nutrient interactions. The deterministic hydrologic model DRAINMOD was preliminary developed to simulate subsurface drainage design and its effect on trafficability and crop production in humid regions including Midwestern United States (Skaggs, 1978). It predicts surface runoff, infiltration, evapotranspiration, subsurface drainage and seepage (vertical and lateral) from subsurface drained landscapes. The rates of infiltration, evapotranspiration, drainage, and distribution of soil water in the soil profile are computed by the approximate methods, which have been tested and validated for a range of soil and boundary conditions (Skaggs, 1980). With low input requirements DRAINMOD has been very helpful to understand the hydrology of subsurface drainage systems, and thereby to optimize the drainage design and management for maximum crop production benefits (Skaggs and Chescheir, 1999). However, the primary goal of drainage research has been shifting from maximizing crop production to environmental impacts with the increasing concern on transport of nutrient (NO3-N) with subsurface drainage water from agricultural lands (Baker et al., 1975; Gilliam, 1987; Skaggs et al., 1994). Therefore, Breve et al. (1997) developed DRAINMOD-N to simulate N dynamics and evaluate the impact of drainage design and management on N-transformation and its transport from artificially drained soils. Later the simplified N cycle of DRAINMOD-N was improved in DRAINMOD-NII to include detailed N cycling and carbon cycling (Youssef et al., 2005). The driving hydrological variables such as infiltration, subsurface drainage, and surface runoff for DRAINMOD-NII are predicted by the original hydrologic model DRAINMOD. In addition to the hydrological variables, DRAINMOD-NII also requires the crop yield parameters such as potential crop yield, harvest index and root-to-shoot ratio to simulate N-transformation and transport from artificially drained soils (Youssef et al., 2005). However, the crop growth in DRAINMOD is simulated through a simple yield reduction approach (relative yield) accounting for excess water, drought and delayed planting stress during the crop growth season. In field conditions, the physiological growth of crops is affected by weather variability in terms of solar radiation and air temperature, and soil water and nutrient regimes. The simple yield reduction approach based on the defined crop representation in DRAINMOD and DRAINMOD-NII does not simulate any interaction between the crop growth and weather, and water and nutrient dynamics in the soil-water-atmosphere-plant continuum. There is the potential that this could be improved using a deterministic crop model to establish the feedback mechanism between the physiological crop growth, weather variability, and soil water and nutrient conditions in field conditions. One such deterministic field scale crop model CERES Maize has been developed to simulate corn growth depending upon weather, and soil water and N-dynamics in the plant-soil environment (Jones and Kiniry, 1986). In CERES Maize model the

potential dry matter production is simulated as a function of the incoming solar radiation and leaf area index, and is reduced to actual dry matter production by stress factors accounting temperature, water and nutrient stress on crop production. The hydrologic model DRAINMOD has the potential strength for simulation of water regime under subsurface drained conditions, while CERES Maize has strength for simulation of detailed crop growth under variable weather and soil water conditions. It could be expected that sharing the information between both models in a systematic fashion may help to capture the complexity of natural processes in subsurface drained landscapes. Therefore, this study was conducted with the objective of integrating CERES Maize into DRAINMOD to simulate subsurface drainage and corn production in subsurface drained landscapes. The integrated DRAINMOD and CERES Maize, referred as the DMCM model hereafter, was evaluated by comparing the simulations with the field observations at the Iowa State University Northeast Research Center, Nashua (IA).

Materials and Methods

Integration of DRAINMOD and CERES Maize The deterministic crop model CERES Maize was integrated as an option for detailed corn growth simulations into DRAINMOD capable of simulating subsurface drainage systems (Fig. 1). Facilitated by the same programming language (FORTRAN) of both models, this integration was performed at the source code level by compiling CERES Maize algorithms into the main architectural structure of DRAINMOD program. Both models interact with each other on a daily time step and establish a feedback mechanism to account for the effects of soil water regime on crop growth and vice-versa. For instance DRAINMOD supplies simulated daily soil water contents and potential evapotranspiration to CERES Maize for the simulation of water limited crop growth, and in return CERES Maize supplies daily effective rooting depth required by DRAINMOD to simulate the daily water balance in the soil profile. The key link in the integration of CERES Maize and DRAINMOD is the simulation of soil water contents in the soil profile by DRAINMOD, and then its transfer to CERES Maize for the simulation of crop production (Fig. 1). The soil water distribution in DRAINMOD assumes three zones: a dry, wet and saturation zone (Fig. 2). A dry zone in the soil profile is created when the evapotranspiration demand is higher than the upward water movement rate and water is removed from the root zone down to lower limit water content (wilting point) (Skaggs, 1980). The effective rooting depth (supplied by CERES Maize) in the model defines the root zone from which water can be removed to supply evapotranspiration demand. The dry zone, therefore, can extend equally to the root zone. The soil water content in the dry zone is equal to the lower limit water content (wilting point) plus the upward water flux in the dry zone (Fig. 2). The upward water movement in the dry zone is determined as a function of the water table depth given as an input to the model. A wet zone extends from the water table to the root zone or possibly through the root zone to the soil surface if a dry zone does not exist. The soil water content in the wet zone is assumed to be in hydrostatic equilibrium with the water table depth independent of whether water is removed by evapotranspiration or subsurface drainage from the soil profile. This assumption generally holds for conditions in which the D-F assumptions are valid and the water table is relatively shallow (Skaggs, 1980). With this assumption, the soil water content in the wet zone could be calculated using the model input soil water characteristics curve relating soil water content to pressure head corresponding to the water table depth in the soil profile. For instance, the soil water content at 0.25 m above water table in the wet zone corresponds to the soil water held in the soil at a –0.25 m pressure head. The sum of wet and dry zone depths gives the water table depth on a time step. A saturation zone exists below the water table depth, and its soil water content is assumed to be equal to the saturation limit of the soil profile (Fig. 2).

START

DRAINMOD

READ SOIL AND DRAINAGE SYSTEM INPUT, and SIMULATION CONTROL CERES Maize First day Yes of READ MONTHLY WEATHER DATA READ CROP INPUT, and month? SIMULATION CONTROL

No Calculate Potential Evapotranspiration Yes Simulate crop and Soil Moisture for Crop Model? growth for JDAY CERES Maize No Simulate Water Calculate Effective Root Balance for JDAY Depth for DRAINMOD

No Simulation End?

Yes

WRITE OUTPUT

STOP

Figure 1. Flow chart of the integrated DRAINMOD and CERES Maize model.

θ θ LL Soil Water Content SAT

Upward water Dry zone flux (UP) SW = θLL+ UP (h)

Wet zone

SW = θ (h) Soil depth (cm) Soil depth

Water table Saturation zone

SW = θSAT

Figure 2. Schematic representation of soil water distribution in the soil profile.

The soil water contents in the soil profile simulated by DRAINMOD (Fig. 1) are supplied to CERES Maize for simulation of the crop production. CERES Maize simulates corn development, its growth rate and final yield using the radiation use efficiency approach. The potential daily biomass accumulation per plant (PCARB) is estimated as a function of incoming solar radiation and leaf area index as follows (Jones and Kiniry, 1986):

PAR[1− e−Kc LAI ] PCARB = RUE (1) PLTPOP where, PAR is the incoming photosynthetic active radiation (wavelength 400-700 nm) (MJ m-2 d- 1 -1 ); RUE is the radiation use efficiency (g MJ ) of the canopy; Kc is the extinction coefficient of the canopy (-); PLTPOP is the plant population (No. m-2); and LAI is the leaf area index (-). The potential daily biomass accumulation is the maximum level of biomass accumulation which could be achieved only under optimal growing conditions. However, there are often suboptimal growing conditions in field situation, and the potential daily biomass accumulation is reduced to suboptimal daily biomass accumulation accounting for temperature, water and nutrient stress, and presence of any pest and disease. In this study, the potential daily biomass accumulation was reduced to suboptimal daily biomass accumulation (CARB) accounting for only temperature (PRFT) and water stress (TURFAC and SWFAC) (Fig. 3).

Optimum Temperature No Stress 1.0 1.0 SWFAC

) (reduce photosynthesis)

s A

H t

F e i c

) e r g

a r u h t T t

S a T

R e e

r p m d

(

u n

m p

r e e a

o r T r

t a

C r

c t u

p w a r t o e F

F m L

S ( TURFAC (reduce plant expansion) 0.0 0.0 High Stress TBASE TOP1 TOP2 TMAX 1.0 1.5 Temperature TRWUP / EP (a) (b) o Figure 3. Crop stress factors as function of (a) temperature (PRFT) and (b) water (TURFAC, SWFAC) conditions (adapted from the presentations at 2006 Training Program on DSSAT version 4.0, Griffin, Georgia).

The photosynthesis reduction factor (PRFT) is governed by the daily average temperature TAVGD= 0.25*Tmin+0.75*Tmax, where Tmin and Tmax are the daily minimum and maximum daily temperature, respectively. The photosynthesis reduction factor (PRFT) is less than optimal (<1) under both low (TOP2) conditions, and become zero if temperature is lower than TBASE or higher than TMAX (Fig. 3a). The water stress is accounted through two reduction factors: one for the photosynthesis (SWFAC) and other for the plant expansion (TURFAC) (Fig. 3b). These factors are based on the ratio of potential root water uptake (TRWUP) and potential transpiration (EP0) demand for crop growth. The potential transpiration is separated from the potential evapotranspiration (supplied by DRAINMOD) using the leaf area index and co-efficient of light extinction. The potential root water uptake is the integration of potential root water uptake from each layer depending on soil water availability supplied by DRAINMOD and rooting depth and its distribution in the soil profile. If the soil layers

are saturated more than the minimum volume required for supplying oxygen to roots for optimum growth, the potential root water uptake is reduced accounting the excess water stress conditions. The produced water and temperature limited daily biomass accumulation provide energy for the maintenance and growth respiration of plant organs, and increase the dry matter of crop. The daily dry mass production is partitioned into different crop organs roots, leaves, stems and storage (ear) organs. This partitioning is a function of crop development (phonological) stage. Using the degree-days approach, CERES Maize simulates six crop phonological stages: emergence; tassel initiation; silking; grain filling; and maturity. The produced dry mass is partitioned initially between leaves and stem (from emergence to silking), and then between ear and grain growth (from silking to grain filling). The remaining dry mass is allocated to the root growth. However, if dry mass available for root growth is below a minimum threshold, the allocation to up surface plant organs (grain, leaves and stem) is reduced to compensate the minimum level of root growth. Further descriptions of crop development and growth process included in CERES Maize are given in Jones and Kiniry (1986) and Jones et al. (2003).

Field Measurements and Model Input Parameters The integrated DMCM model was to be tested with field measurements recorded from the experimental field plots located at the Iowa State University Northeast Research Center, Nashua (IA). Total research area is 15 ha consisting a total of 36 experimental plots of 0.40 ha plot size (Kanwar, 1991). Predominantly soils of the experimental plots are Readlyn (fine-loamy, mixed, mesic, Aquic Hapludolls), Kenyon (fine-loamy, mixed, mesic, Typic Hapludolls) and Floyd (fine-loamy, mixed, mesic, Aquic Hapludolls). These soils are moderately to somewhat poorly drained with an average slope of 0 to 4 percent, but generally less than 2 percent. Tile drains are installed along the long dimension through the center of each plot and on the borders with a drain spacing of 28.5 m and drain depth of 1.20 m. Subsurface tile drains installed at plot borders help to prevent lateral subsurface flow from adjacent plots. Therefore, the lateral seepage was not included into the simulations, while vertical seepage was included by defining an impermeable (restrictive) layer of 250 cm thickness with a hydraulic (vertical) conductivity of 0.00022 cm hr-1 and a peizometric head of 250 cm in the aquifer. The depth to the impermeable layer was assumed at a soil depth of 390 cm (Sanoja et al., 1990). Table 1 summarizes the input parameters of DRAINMOD to characterize the installed subsurface drainage system at the experimental plots.

Table 1. Subsurface drainage system for the experimental field plots located at the Iowa State University Northeast Research Center, Nashua (IA).

Drainage System Parameter Value Unit Drain depth 120 cm Drain spacing 2850 cm Effective radius of drains 2 cm Drainage coefficient 3.5 cm day-1 Maximum surface storage 1.25 cm Kirkham's depth for flow to drains 0.25 cm Depth to impermeable (restrictive) layer 390 cm Thickness of impermeable (restrictive) layer 250 cm Hydraulic Cond. of impermeable (restrictive) layer 0.00022 cm hr-1 Piezometric head in the aquifer 250 cm

Field measurements at the experimental plots have been conducted for various combinations of tillage, crop rotation and nitrogen application rates. The experimental plots have been monitored for subsurface drainage, NO3-N concentrations, soil water content, agronomic activities and crop production (Ahmed, 1996; Garrison et al., 1999). In this study, three experimental plots (14, 25 and 31) were selected with different soil types (Readlyn; Kenyon and Floyd, respectively) cultivated with continuous corn, no tillage and an application of 202 kg N ha-1 y-1 from 1990 to 1992 years. The soil profile of 0-390 cm at the selected plots was divided into six to seven soil layers depending on the collected soil information (Table 2). As soil input DRAINMOD requires the relationships between soil water content versus pressure head (soil water characteristic curve), water table depth versus volume drained, water table depth versus upward flux, and Green-Ampt parameters for the soil profile. These relationships were generated by a soil utility 3 program in DRAINMOD using the soil hydraulic parameters: the saturated water content θs (cm -3 3 -3 -1 cm ), the residual water content θr (cm cm ), factors α (cm ) and n (-), an empirical pore -1 tortuosity/connectivity parameter λ (-), and the saturated hydraulic conductivity Ksat (cm d ) (Van Genuchten, 1980; Mualem, 1976). The required soil hydraulic parameters were predicted by a pedotransfer function model, ROSETTA (Schaap et al. 2001) using the available soil texture and bulk density data (Table 2). One of the soil input parameters for DRAINMOD, the lateral -1 saturated hydraulic conductivity LKsat (cm hr ) was calibrated by Singh et al. (1996) for the Readlyn; Kenyon and Floyd soils of the selected experimental plots (Table 2). The deterministic crop model CERES Maize also requires the soil water limits: Lower Limit (LL), Drain Upper Limit (DUL) and Saturation limit (SAT) for simulating the water stress on crop production through the potential root water uptake (Fig. 3b). These soil water limits were set according to the predicted soil water characteristic curve for each soil layer as follows: Lower Limit (LL) equals to the residual water content; Drain Upper Limit (DUL) equals to the field capacity (at 33 kPa); and Saturation limit (SAT) equals to the saturated water content (at 0 kPa) (Table 2).

Table 2. Soil properties of selected Readlyn, Kenyon and Floyd soil plots at the Iowa State University Northeast Research Center, Nashua (IA) (adapted Singh et al., 1996).

Soil type/ Soil depth Bulk density*Soil texture (% by vol.)* LL** DUL** SAT** LKsat*** Horizon No. (cm) (g cm-3) Silt Clay (cm3 cm-3)(cm3 cm-3)(cm3 cm-3)(cm hr-1) Readlyn 1 0-20 1.34 43 26 0.08 0.27 0.43 3.1 2 20-30 1.45 43 26 0.07 0.25 0.41 3.1 3 30-43 1.45 38 25 0.07 0.25 0.41 3.1 4 43-54 1.50 38 25 0.07 0.24 0.39 3.1 5 54-68 1.60 24 21 0.06 0.21 0.37 3.1 6 68-89 1.65 28 26 0.06 0.23 0.36 3.1 7 89-390 1.70 28 26 0.06 0.23 0.35 3.1 Kenyon 1 0-20 1.36 42 20 0.09 0.30 0.47 3.1 2 20-41 1.53 34 25 0.07 0.23 0.39 3.1 3 41-50 1.55 32 26 0.07 0.24 0.39 3.1 4 50-69 1.60 30 27 0.07 0.24 0.38 3.1 5 69-89 1.65 28 28 0.06 0.24 0.37 3.1 6 89-123 1.70 31 25 0.06 0.22 0.35 3.1 7 123-390 1.75 31 25 0.05 0.22 0.33 3.1 Floyd 1 0-43 1.29 44 26 0.08 0.28 0.45 3.2 2 43-58 1.40 42 26 0.07 0.26 0.42 3.2 3 58-85 1.45 22 24 0.07 0.23 0.42 3.2 4 85-115 1.58 29 24 0.06 0.23 0.38 3.2 5 115-153 1.70 40 25 0.06 0.22 0.35 3.2 6 153-390 1.75 40 25 0.06 0.22 0.33 3.2 * Experimentally measured by Singh (1994); ** LL is the Lower Limit, DUL is the Drained Upper Limit, and SAT is the Saturation limit derived by ROSETTA (Schaap et al. 2001); *** Lateral Saturated Hydraulic Conductivity calibrated by Singh et al. (1996).

In addition to the soil input parameters, CERES Maize requires crop growth parameters mainly grouped as cultivars, ecotype and species parameters. The species parameters define the characteristics of corn including the radiation and CO2 parameters governing the photosynthesis, temperature effects of photosynthesis and grain filling, initial emergence condition, seed growth parameters, and root growth parameters such as minimum volume required for supplying oxygen for optimum root growth and maximum root water uptake. The ecotype parameters define the radiation use efficiency, light extinction coefficient, base and optimum temperature for crop development during vegetative and reproductive stage, and thermal time development (Table 3). The cultivar parameters specify the development characteristics of a specific maize cultivar at the experimental plots at the Iowa State University Northeast Research Center, Nashua (IA).

Table 3. Crop parameters for simulation of corn at the Iowa State University Northeast Research Center, Nashua (IA).

General parameters Crop window (Julian day) 105 - 335 julian day Plant population 7.2 plants m-2 Row spacing 61 cm Ecotype parameters* Cultivar parameters** RUE 4.2 g MJ-1 P1 200.0 oC -1 Kc 0.85 - P2 0.75 hr TBASE 8 oC P5 800.0 oC TOPT 34 oC G2 850.0 kernels plant-1 ROPT 34 oC G3 8.5 mg kernel-1 day-1 P2O 12.5 hrs PHINT 49.0 oC GDDE 6 oC cm-1 o DSGFT 170 C * RUE Radiation use efficiency, gram plant dry matter per million jule (MJ) of PAR; Kc Light extinction coefficient; TBASE Base temperature below which no development occurs; TOPT Temperature at which maximum development rate occurs during vegetative stages; ROPT Temperature at which maximum development rate occurs for reproductive stages; P2O Daylength below which daylength does not affect development rate; GDDE Growing degree days per cm seed depth required for emergence; and DSGFT Growing degree days from silking to effective grain filling period. ** P1 Growing degree days from seedling emergence to end of juvenile phase; P2 Photoperiod sensitivity coefficient; P5 Growing degree days from silking to physiological maturity; G2 Potential kernel number; G3 Potential kernel growth rate; and PHINT Growing degree days required for a new leaf.

The integrated DMCM model also requires daily weather input values of precipitation, air temperature, incoming total solar radiation and wind speed. All the weather variables were available except wind speed, which was used as average of 86.4 km d-1. The precipitation, and maximum and minimum temperature were recorded at the experimental site (Iowa State University Northeast Research Center), while total solar radiation was used from a weather station located in Ames, IA (~140 km southwest of the experimental site). Figure 4 reproduces the recorded daily precipitation, temperature and solar radiation during the study period from 1990 to 1992. Generally the daily mean temperature is below 0 oC in the months of November (Julian day 328) to the start of March (Julian day 62), and then rises in the months of June, July and August (from 152 to 243 Julian days). The total daily solar radiation varied from 0.50 to 29.76 MJ m-2 with low values (< 10 MJ m-2) in the months from November (Julian day 305) to mid of February (Julian day 45). The annual precipitation varied from 72 cm in 1992 to 105 cm in 1990 with an average of 91 cm over the years from 1990 to 1992. Most of the precipitation (70% of the total) occurred during the months of March through August (from 60 to 243 Julian days) with a maximum average monthly precipitation of 19.4 cm in July. DRAINMOD requires hourly precipitation for its simulation processes. This was obtained by distributing the daily observed precipitation uniformly over 6 hours (from 1600 to 2200 hrs) during the day of recorded precipitation. Also, DRAINMOD requires monthly ET factors to correct the Thornthwaite ET calculations (Thornthwaite, 1948) for site-specific conditions (Skaggs, 1980). The monthly ET factors used during the simulations are as follows: 1.3 (Jan- Feb), 0.6 (Mar-

May), 0.8 (Jun), 0.9 (Jul-Aug), 1.2 (Sept) and 1.3 (Oct-Dec). Similar monthly ET factors were used by Sands et al. (2003) for south-central Minnesota, and by Singh et al. (2006) for north central Iowa. In addition, the soil temperature parameters were the same as summarized by Luo et al. (2000; 2001) for Minnesota. )

40 -2 30 25 C)

o 20 20 0 15 10 -20 5 Temperature (

-40 m (MJ radiation Solar 0 1 61 121 181 241 301 361 1 61 121 181 241 301 361 (a) Julian day (b) Julian day

20 120 105 97 100 15 80 72 10 60 5 40 20 Precipitation (cm) Precipitation (cm) 0 0

Jul. 1990 1991 1992 Apr. May Jan. Jun. Mar. Oct. Feb. Nov. Aug. Dec. Sept. (d) Year (C) Months Figure 4. Weather variables (a) air temperature, (b) solar radiation, and (c & d) precipitation recorded during the years from 1990 to 1992.

Results and Discussion

Comparison of DMCM model with field measurements The performance of a modeling effort is judged by whether simulated outputs sufficiently represent the system being simulated or not. In this study, the integrated DMCM model was evaluated by simulating three selected experimental plots with different soils (Readlyn, Kenyon and Floyd) cultivated with no tillage and continuous corn at the Iowa State University’s Northeast Research Center, Nashua (IA). The performance of the DMCM model was evaluated by comparing the simulated and observed soil water content, subsurface drainage and corn yield at the selected experimental plots from 1990 to 1992. The soil water content was observed two to four times during a year corresponding to before and after planting and harvesting, and at different depths from 0-10, 10-20, 20-30, 30-45, 45-60, 60-90 and 90-120 cm (Ahmed, 1996). The observed dry weight basis soil water content (% dwb) was converted to the volumetric basis (cm3 cm-3) by multiplying with the bulk density in Table 2. The mean absolute error (MAE) and mean bias error (MBE) (Willmott, 1982) were calculated to quantify the differences between the observed and simulated soil water contents (cm3 cm-3) for Readlyn, Kenyon and Floyd plots. The MAE over different soil depths and observations dates

varied from 0.02 to 0.11 cm3 cm-3 with an overall average of 0.04, 0.05 and 0.08 cm3 cm-3 for the Readlyn, Kenyon and Floyd soil plots, respectively. The MAE gives the estimate of average absolute error in the units of observed and simulated values of a variable. However, it does not give the nature of error i.e. over- or underestimation. The mean bias error (MBE) gives the nature of error with a positive sign indicating the overestimation, i.e. simulated values are higher than the observed values. The MBE over different soil depths and observations dates ranged from -0.06 to 0.11 cm3 cm-3 with an overall average of 0.03, 0.01 and 0.08 cm3 cm-3 for the Readlyn, Kenyon and Floyd soil plots, respectively. When value of MBE becomes equal to MAE then the simulations are consistently over- or underestimated depending on the positive or negative sign of the MBE value. The ratio of average MBE and MAE values varied from 0.11 to 0.95 being higher for the Floyd soil plot. These numbers indicate the overestimation of soil water contents for the Floyd soil plot. Using the predicted soil moisture in the top 90 cm, the soil water storage in this zone can be computed and compared to the measured soil water storage from this same zone (Fig. 5) in the soil profile. Again, Figure 5 indicates the overestimation of water storage in the upper 90 cm of soil profile, especially for the Floyd soil plot. The average MAE and MBE values were equal to 7.01 cm for the water storage in the upper 90 cm of the Floyd soil profile. This confirms a consistent overestimation of soil water contents for the Floyd soil plot. However, the soil water contents were simulated best for the Kenyon soil plot with average values of MAE and MBE being equal to 2.38 and 0.52 cm, respectively for the water storage in the upper 90 cm of the soil profile.

40 Readlyn 35 30 25 20 Observed Simulated

40 Kenyon 35 30 25 20 Floyd 40 35 30 Water Storage (cm) in upper 90 cm soil profile cm 90 in upper Water(cm) Storage 25 20 1990001 1990091 1990181 1990271 1990361 1991086 1991176 1991266 1991356 1992081 1992171 1992261 1992351 Year_Julian day Figure 5. Observed and simulated water storage in the upper 90 cm of soil profile of the Readlyn, Kenyon and Floyd soil plots during the simulation period from 1990 to 1992. The performance of DMCM model was further evaluated by comparing the observed and simulated subsurface drainage from the Readlyn, Kenyon and Floyd soil plots. The subsurface

drain flows were recorded for each plot on alternate days from April through November during each year (Kanwar et al., 1996). However, only on rare occasions was any winter subsurface drainage observed, and little winter subsurface drainage in this area is expected (Randall, 2004). The cumulative subsurface drainage over the years from 1990 to 1992 was simulated 11% higher for the Kenyon soil plot (Co-efficient of Mass Residual, CRM = 0.11), 7% higher for the Readlyn soil plot (CRM=0.07) and 6% lower for the Floyd soil plot (CRM=-0.06) (Table 4). While the difference between the observed and simulated subsurface drainage are not too large, the similar values of MBE and MAE revealed a consistent overestimation of subsurface drainage for the Kenyon soil plot, and underestimation for the Floyd soil plot (Table 4). The underestimation of subsurface drainage may be the reason of the simulation of higher soil water contents for the Floyd soil plot (Fig. 5). Such discrepancies may be improved by the site-specific calibration and validation of the model by adjusting soil parameters to fit the observed and simulated soil water content and subsurface drainage.

Table. 4. Observed and simulated subsurface drainage (cm) for the selected Readlyn, Kenyon and Floyd soil plots during the simulation period from 1990 to 1992.

Soil Type/ Readlyn Kenyon Floyd Year Observed Simulated Observed Simulated Observed Simulated 1990 28 28 27 29 27 27 1991 35 35 30 36 36 35 1992 17 22 21 22 26 22 Cumulative 79 85 78 87 89 84 MAE 2.2 2.8 1.7 MBE 1.8 2.8 -1.7 CRM 0.07 0.11 -0.06

The corn yields were simulated 9 to 30% higher than the observed corn yields for different plots during the simulation years from 1990 to 1992. Figure 6 reproduces the observed and simulated corn yields for the Readlyn soil plot where the MAE and MBE values were equal about to 1.75 ton ha-1. Following the similar trend the MAE between the observed and simulated corn yields was 1.77 and 1.10 ton ha-1 for the Kenyon and Floyd soil plot, respectively. Though the corn yields for the Readlyn soil plot were simulated consistently higher than the observed yields they followed the general trend of the observations (Fig. 6). It is expected that these corn yield simulations could be improved through site specific calibration and validation since the model used default crop parameters (Table 3).

14 Observed Simulated 12 ) -1 10

4 Corn Yield (ton ha 2

0 1990 1991 1992 Year

Figure 6. Observed and simulated crop yield (ton ha-1) for the selected Readlyn soil plots during the simulation period from 1990 to 1992.

Comparison of DRAINMOD and DMCM models The integrated DMCM model was further evaluated by comparing the simulated variables with original DRAINMOD model for the Readlyn soil (Table 2) using a weather dataset for 45 years (from 1961 to 2005). This long-term weather dataset included the precipitation and air temperature recorded at the Charles City (IA) (~17 km away from the experimental site) and the total solar radiation in the Ames (IA) (~140 km away from the experimental site). There were no significant differences in the hydrology of the Readlyn soil plot simulated by the DMCM and DRAINMOD model. This was expected since the DMCM model used the same algorithms for the simulation of hydrology as incorporated into the DRAINMOD model (Skaggs, 1980). However, we did expect differences in the simulation of crop yields by the DMCM model which accounts for the effect of weather variability in terms of solar radiation and temperature during the crop growing season (Eq. 1 and Fig. 3a). The potential corn yield simulated by the DMCM model varied from 10.0 to 17.0 ton ha-1 with an average of 14.1 ton ha-1 for Iowa’s conditions (Fig. 7). Accounting for the water stress on the photosynthesis and plant expansion (Fig. 3b) the DMCM model reduced the simulated potential corn yields to actual corn yields which varied from 8.8 to 16.4 ton ha-1 with an average of 13.7 ton ha-1 for the Readlyn soil plot (Fig. 7).

18.0 Potential yield

) Actual yield

-1 16.0

14.0

12.0

10.0 Corn Yield (ton ha (ton Corn Yield 8.0

1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 1995 1997 1999 2001 2003 2005 Figure 7. Predicted potential and actual corn yield (ton ha-1) simulated by the integrated DMCM model for the Readlyn soil plot over the 45 years from 1961 to 2005.

Figure 8 compares the predicted corn yields based on DRAINMOD and DMCM simulations for the Readlyn soil plot over the 45 years from 1961 to 2005. The relative yields (actual/potential) simulated by DRAINMOD could be translated to actual corn yields by two methods, (a) the relative yields multiplied with the average potential corn yield (Fig. 8a), and (b) the relative yields multiplied with the annual potential yields (Fig. 8b). The potential corn yields were simulated by the integrated DMCM model (Fig. 7).

18.0 18.0

16.0 16.0 ) -1 14.0 14.0

12.0 12.0

10.0 10.0 Corn Yield ha (ton 8.0 8.0 (a) Relative Yield * Average Potential Yield (b) Relative Yield * Annual Potential Yield (DRAINMOD Model) (DRAINMOD Model) 6.0 6.0 Actual Yield (DMCM Model) Actual Yield (DMCM Model)

4.0 4.0 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005 1961 1965 1969 1973 1977 1981 1985 1989 1993 1997 2001 2005 Figure 8. Predicted actual corn yields based on DRAINMOD and DMCM model simulations for the Readlyn soil plot over the 45 years from 1961 to 2005.

Both methods “a” and “b” produced the same average actual corn yield of 14.0 ton ha-1 which was comparable with the average actual corn yield of 13.7 ton ha-1 simulated by the DMCM model. The method “a” (relative yield * average potential yield) did result into a very small variation in the actual corn yields (range: 13.5-14.1 ton ha-1, standard deviation: 0.2 ton ha-1) as compared to the actual corn yields simulated by the DMCM model resulting into a standard deviation of 1.4 ton ha-1 (range: 8.8-16.4 ton ha-1) (Fig. 8a). The method “b” (relative yield * annual potential yield) however resulted a similar variation as the actual corn yields simulated by the DMCM model (Fig. 8b). This is due to the fact that method “b” used the annual potential yields simulated accounting for the effect of variable weather conditions in terms of solar radiation and temperature on crop production.

The crop growth affects the hydrology of the system through root water uptake and evapotranspiration, and is also expected to influence the N-transformation and transport through plant nitrogen uptake from the soil profile. A simple yield reduction approach such as method “a” might have some limitations for the simulation of N-transformation and transport processes of subsurface drainage systems. The potential production level of a crop is determined by the characteristics and genetics of the plant and could be achieved only under optimal growing conditions. However, the potential yield of a crop is also defined by weather conditions such as temperature and solar radiation received in a region (Lövenstein et al., 1992). The integrated DMCM model accounts for the weather effects on the crop production through CERES Maize (Eq. 1 and Fig. 3a), and thus capitalizes on the strength of both models DRAINMOD for subsurface drainage hydrology and CERES Maize for corn growth and development to increase the ability of simulating subsurface drained landscapes.

Conclusion The primary goal of drainage research has been shifting from maximizing crop production to environmental impact with the increasing concern related to the transport of nutrients, specifically nitrate-nitrogen (NO3-N) with subsurface drainage water from agricultural lands. It is becoming important to evaluate the impact of drainage design and its management not only on crop production but also on nutrients (nitrogen) transformation and transport from agricultural lands. The hydrologic models DRAINMOD and DRAINMOD-NII simulate subsurface drainage design and management, and its impact on N-transformation and transport from artificially drained soils. The crop growth in DRAINMOD and DRAINMOD-NII is simulated through a simple yield reduction approach (relative yield) accounting for excess water, drought and delayed planting stress during the crop growth season. However, the physiological growth of crops is also affected by weather variability in terms of solar radiation and air temperature. In this study, the simple crop representation in DRAINMOD was expanded by integrating the deterministic crop model CERES Maize as an option for the simulation of detailed corn growth and development accounting for weather and soil water dynamics in the soil-water-plant- atmosphere continuum. The integrated DRAINMOD and CERES Maize, referred as the DMCM model, was evaluated by comparing the simulations with the observations from three (Readlyn; Kenyon and Floyd) soil plots located at the Iowa State University Northeast Research Center, Nashua (IA). The preliminary model results are encouraging showing a good correspondence with the observed soil water content, subsurface drainage and crop yields during the years from 1990 to 1992. The integration of detailed crop models in DRAINMOD could enhance the capability of modelling the subsurface drainage systems. Future development efforts should continue to integrate the DSSAT crop models (Jones et al., 2003) into the DRAINMOD to expand on the simple crop representation.

References

Ahmed, S.I. 1996. Effect of N-management practices on residual NO3-N under corn-soybean and continuous corn rotation. MS Thesis. Ames, Iowa: Iowa State University, Department of Agricultural and Biosystems Engineering. Ahuja, L.R., J.D. Hanson, K.W. Rojas, and M.J. Shaffer (eds). 1999. The Root Zone Water Quality Model. Water Resources Publications LLC. Highlands Ranch, CO. Alexander, C. 1988. ADAPT - A model to simulate pesticide movement into drain tiles. MS thesis. Columbus, Ohio: Ohio State University, Department of Agricultural Engineering.

Baker, J.L., K.L. Campbell, H.P. Johnson, and J.J. Hanway. 1975. Nitrate, phosphorus and sulfate in subsurface drainage water. J. Environ. Qual. 4: 406-412. Breve, M.A., R.W. Skaggs, J.E. Parsons, and J.W. Gilliam. 1997. DRAINMOD-N, A nitrogen model for artificially drained soils. Trans. of the ASAE 40(4): 1067-1075. Garrison, M.V., W.D. Batchelor, R.S. Kanwar, and J.T. Ritchie. 1999. Evaluation of the CERES- MAIZE water and nitrogen balances under tile-drained conditions. Agricultural Systems 62: 189-200. Gilliam, J.W. 1987. Drainage water quality and the environment. Keynote address. ASAE Publication No. 787. St. Joseph, Mich.: ASABE. Jones, C.A., and J.R. Kiniry. 1986. CERES-Maize: A simulation model of maize growth and development. Texas A&M University Press. College Station, Texas. Jones, J.W., G. Hoogenboom, C.H. Porter, K.J. Boote, W.D. Batchelor, L.A. Hunt, P.W. Wilkens, U. Singh, A.J. Gijsman, and J.T. Ritchie. 2003. The DSSAT cropping system model. European Journal of Agronomy 18 (3-4): 235-265. Kanwar, R.S. 1991. Long term tillage effects on the quality of subsurface drainage and shallow water. In Proc. of the Conf. on Environmentally Sound Agriculture, April 16-18, 1991, Orlando (FL). St. Joseph, Mich.: ASABE. Lövenstein, H.M, R. Rabbinge, and H. van Keulen. 1992. World Food Production. Textbook 2: Biophysical factors in agricultural production. ISBN 90-358-1111-1. Heerlen, The Netherlands: Open university. Luo, W., R.W. Skaggs, and G.M. Chescheir. 2000. DRAINMOD modifications for cold conditions. Trans. ASAE 43(6): 1569-1582. Luo, W., R.W. Skaggs, A. Madani, S. Cizikci, and A. Mavi. 2001. Predicting field hydrology in cold conditions with DRAINMOD. Trans. ASAE 44(4): 825-834. Mualem, Y. 1976. A new model for predicting the hydraulic conductivity of unsaturated porous media. Water Resour. Res. 12: 513-522. Randall, G. 2004. Subsurface drain flow characteristics during a 15 year period in Minnesota. ASAE Publication Number 701P0304, St. Joseph, MI: ASABE. Sands, G.R., C.X. Jin, A. Mendez, B. Basin, P. Wotzka, and P. Gowda. 2003. Comparing the subsurface drainage flow prediction of the DRAINMOD and ADAPT models for a cold climate. Trans. ASAE 46(3): 645-656. Sanoja, J., R.S. Kanwar, and S.W. Melvin. 1990. Comparison of simulated (DRAINMOD) and measured tile outflow and water table elevations from two field sites in Iowa. Trans. ASAE 33(3): 827-833. Schaap, M.G., F.J. Leij, and M.Th. van Genuchten. 2001. Rosetta: A computer program for estimating soil hydraulic parameters with hierarchical pedotransfer functions. J. Hydrol. 251:163-176. Simunek, J., M.Th. van Genuchten, and M. Sejna. 2005. The HYDRUS-1D Software Package for Simulating the Movement of Water, Heat, and Multiple Solutes in Variably Saturated Media, Version 3.0, HYDRUS Software Series 1. Riverside, California: University of California, Department of Environmental Sciences. Singh, P. 1994. Modification of Root Zone Water Quality Model (RZWQM) to simulate the tillage effects on subsurface drain flows and NO3-N movement. PhD. Thesis (Diss. Abstr. 94- 24259). Ames, Iowa: Iowa State University, Department of Agricultural and Biosystems Engineering. Singh, P., R.S. Kanwar, K.E. Johnsen, and L.R. Ahuja. 1996. Calibration and evaluation of subsurface drainage component of RZWQM V.2.5. J. Environ. Qual. 25: 56-63.

Singh, R., M.J. Helmers, and Z. Qi. 2006. Calibration and validation of DRAINMOD to design subsurface drainage systems for Iowa’s tile landscapes. Agricultural Water Management 85: 221-232. Skaggs, R.W. 1978. A water management model for shallow water table soils. Technical report No. 134. Raleigh, North Carolina: University of North Carolina, NC State University, Water Resources Research Institute. Skaggs, R.W. 1980. DRAINMOD reference report. Methods for design and evaluation of drainage-water management systems for soils with high water tables. USDA-SCS, South National Technical Center, Fort Worth, Texas. Skaggs, R.W., M.A. Breve, and J.W. Gilliam. 1994. Hydrologic and water quality impacts of agricultural drainage. Crit. Rev. Environ. Sci. Technol. 24: 1-32. Skaggs, R.W., and G.M. Chescheir. 1999. Application of drainage simulation models. In: Skaggs, R.W. and J. van Schilfgaarde (eds.), Agricultural Drainage. SSSA, Madison, WI. Thornthwaite, C.W. 1948. An approach toward a rational classification of climate. Geographical Review 38: 55-94. Van Dam, J.C., J. Huygen, J.G. Wesseling, R.A. Feddes, P. Kabat, P.E.V. van Walsum, P. Groenendijk, and C.A.van Diepen. 1997. Theory of SWAP version 2.0. Simulation of water flow, solute transport and plant growth in the Soil-Water-Atmosphere-Plant environment. Report 71, Sub department of Water Resources, Wageningen University. Technical document 45, Alterra Green World Research. Wageningen, The Netherlands. Van Genuchten, M. Th. 1980. A closed from equation for predicting the hydraulic conductivity of unsaturated soils. Soil Sci. Am. J. 44: 892-898. Willmott, C.J. 1982. Some comments on the evaluation of model performance. Bull. Amer. Meteor. Soc. 63: 1309-1313. Youssef, M.A., R.W. Skaggs, G.M. Chescheir, and J.W. Gilliam. 2005. The Nitrogen simulation model, DRAINMOD-N II. Trans. ASAE 48(2): 611-626.