**2.2 Spatial nitrogen studies**

Experiments were conducted in 12 different fields located in central Iowa from 2000 to 2002 (Table 1). These fields varied in size from 15 to 130 ha. Each of the field experiments was similar with strips arranged in the field with different N rates and in three of the study fields, variable planting rates were also used as a treatment variable. The strips were a minimum of 50 m wide covering at least 60 rows of corn. The arrangement of the N treatments was randomized across the field and strips were treated as replicates. Nitrogen was applied as UAN (Urea and Ammonium Nitrate) as a preplant treatment in all fields and incorporated into the soil. Rates of N application were determined by using

Spatial Patterns of Water and Nitrogen Response Within Corn Production Fields 79

2000/Carroll 1 70, 116, 162, 210 72,000 plants/ha

2000/Carroll 2 70, 116, 162, 210 72,000 plants/ha

2000/Sac 70, 116, 162, 210 72,000 plants/ha

2000/Shelby 70, 116, 162, 210 72,000 plants/ha

2000/Story 1 52, 100, 145 72,000 plants/ha

2000/Story 2 52, 100, 145, 191 72,000 plants/ha

2001/Story 1 133, 180 75,000 plants/ha

2001/Story 2 51, 102, 144 75,000 plants/ha

2000/Story 3 57, 126, 173 75,000 plants/ha

2002/Calhoun East 78, 134, 190 79,000 plants/ha

2002/Dallas South 78, 134, 190 79,000 plants/ha

Table 1. Nitrogen rates and associated agronomic parameters for the field experiments from

Soil map units were extracted from the soil survey data and detailed topographic data were collected with GPS equipment for each field. Soil nutrient content data were collected from each field for the upper 1 m of the profile for a minimum of 30 locations within each field prior to the growing season. Yield data were obtained from each field using yield monitors on the producers combine and these were calibrated prior to harvest. Yields were corrected

Remote sensing observations for each field were obtained from an aircraft based hyperspectral unit with 35 wavebands (Table 2). Details on the wavebands and radiometric resolution are provided in Table 2 for each waveband. Data were obtained on clear days at four times during the growing season in May, July, August, and September. These times were selected to obtain data after planting and before there was sufficient growth to affect the soil background (May), maximum vegetative growth (July), mid-grain fill (August), and near physiological maturity (September). The pixel sizes were nominally 2 x 2 m across the study sites. These data were collected with GPS signals to provide a location of each pixel and these were georeferenced to the yield data and other parameters collected from each field with a positional accuracy of 5 m. All data were georeferenced and all data were aggregated into 8 x 8 m pixels for further analysis. This size was selected because of the width of harvest equipment for the yield

2000 to 2002.

monitor data.

to 15.5% grain moisture content.

2002/Coon Rapids 67,134,201 57,000, 69,200, 81,500 plants/ha

**Year/Field Nitrogen Rate (kg/ha) Agronomic Variables** 

soil test records from the producer's fields to determine the typical rate applied and then using 0.5 and 1.5 times that rate in the other strips. All production practices were conducted by the producer during the course of the experiment. In one experiment, planting rates were evaluated and randomly assigned as blocks across the field with N rates as subplots within a block. Planting rates were also randomized across the field. In each field the same corn hybrid was planted; however, hybrids varied among fields and years. For each field there were overlays of soil type, elevation collected from Real Time Kinetic GPS equipment to create 1 m contours, and N application rate. Soils within fields were a mixture of Canisteo (Fine-loamy, mixed (calcareous), mesic Typic Haplaquolls), Clarion (Fine-loamy, mixed, mesic Typic Hapludolls), Nicollett (Fine-loamy, mixed, mesic Aquic Hapludolls), Okoboji (Fine, montmorillonitic, mesic Cumulic Haplaquolls), and Webster (Fine-loamy, mixed, mesic Typic Haplaquolls) soils. An example of the field layout for this experiment is shown in Fig. 1.

Fig. 1. Experimental layout for field-scale N management study conducted across central Iowa from 1999 to 2002.

soil test records from the producer's fields to determine the typical rate applied and then using 0.5 and 1.5 times that rate in the other strips. All production practices were conducted by the producer during the course of the experiment. In one experiment, planting rates were evaluated and randomly assigned as blocks across the field with N rates as subplots within a block. Planting rates were also randomized across the field. In each field the same corn hybrid was planted; however, hybrids varied among fields and years. For each field there were overlays of soil type, elevation collected from Real Time Kinetic GPS equipment to create 1 m contours, and N application rate. Soils within fields were a mixture of Canisteo (Fine-loamy, mixed (calcareous), mesic Typic Haplaquolls), Clarion (Fine-loamy, mixed, mesic Typic Hapludolls), Nicollett (Fine-loamy, mixed, mesic Aquic Hapludolls), Okoboji (Fine, montmorillonitic, mesic Cumulic Haplaquolls), and Webster (Fine-loamy, mixed, mesic Typic Haplaquolls) soils. An example of the field

Fig. 1. Experimental layout for field-scale N management study conducted across central

layout for this experiment is shown in Fig. 1.

Iowa from 1999 to 2002.


Table 1. Nitrogen rates and associated agronomic parameters for the field experiments from 2000 to 2002.

Soil map units were extracted from the soil survey data and detailed topographic data were collected with GPS equipment for each field. Soil nutrient content data were collected from each field for the upper 1 m of the profile for a minimum of 30 locations within each field prior to the growing season. Yield data were obtained from each field using yield monitors on the producers combine and these were calibrated prior to harvest. Yields were corrected to 15.5% grain moisture content.

Remote sensing observations for each field were obtained from an aircraft based hyperspectral unit with 35 wavebands (Table 2). Details on the wavebands and radiometric resolution are provided in Table 2 for each waveband. Data were obtained on clear days at four times during the growing season in May, July, August, and September. These times were selected to obtain data after planting and before there was sufficient growth to affect the soil background (May), maximum vegetative growth (July), mid-grain fill (August), and near physiological maturity (September). The pixel sizes were nominally 2 x 2 m across the study sites. These data were collected with GPS signals to provide a location of each pixel and these were georeferenced to the yield data and other parameters collected from each field with a positional accuracy of 5 m. All data were georeferenced and all data were aggregated into 8 x 8 m pixels for further analysis. This size was selected because of the width of harvest equipment for the yield monitor data.

profile.

Water Use (mm)

0

with two nitrogen application rates.

100

200

300

400

500

600

shown in Fig. 2.

Spatial Patterns of Water and Nitrogen Response Within Corn Production Fields 81

in water holding capacity of nearly 100 mm in the upper 1 m for these two soil profiles. By extension, this creates a difference in the amount of soil water available to the growing crop during the season. In this comparison, Clarion soil with the lower amount of N applied showed the larger amount of crop water use during the season and ultimately showed the larger grain yield at the end of the growing season compared to high rates of N application. In typical growing seasons in central Iowa, there is adequate soil water available for plant growth early in the growing season and it is not until the onset of the reproductive period in which crop water use rates increase and precipitation amounts begin to diminish and to meet atmospheric demand there is a reliance on the amount of water stored within the soil

A recent study by Hatfield et al. (2007) revealed large differences in the daily and seasonal amounts of crop water use across corn and soybean fields in central Iowa. They found the primary reasons for these differences were due spatial variation in precipitation and spatial variation in soils across various fields. These results confirm the observations collected from multiple soils within the same field. The observations collected from several different studies across multiple years reconfirm the observations

> Clarion Spring N (100 kg/ha) Webster Spring N (100 kg/ha) Clarion Fall N (200 kg/ha) Webster Fall N (200 kg/ha)

Day of Year 100 120 140 160 180 200 220 240 260 280

Fig. 2. Crop water use patterns during the 2000 growing season for corn on two soil types

An extension beyond the crop water use patterns is the development of an assessment of water use efficiency. Water use efficiency is expressed as the amount of grain yield relative to the seasonal total of transpiration. In this case, transpiration was determined by removing the soil water evaporation component from the total crop water use amounts. In this analysis, water use efficiency was calculated for the 150 kg ha-1 N application rate. What is

Corn Water Use 2000


Table 2. Wavebands and bandwidth for the AISA hyperspectral data collected over the field sites in 2000 to 2002.

Data were analyzed on the individual N strips within the field and across the field. Vegetative indices were computed from the reflectance data obtained from the aircraft data and correlated to strip yield and field level yield observations. Correlation and regression analysis were conducted between N rate and corn yield for each individual field and across all fields for each year of the study. T-tests among means were conducted on differences between soils within a field strip while analysis of variance was conducted on N rates across a field with an interaction term based on the soil type by N rate comparison. These analyses were made using ANOVA and GLM models in SAS (SAS 2009). Spatial analysis was conducted for each field using GS+ version 5.1 to determine the spatial relationships among yield and the vegetative indices across different fields. Autocorrelation values were computed across the field using the field location points as coordinates to compute the range and sill for yield and the red/green index.
