A gronomy J our n al • Volume 110 , I ssue 1 • 2 018 1 T he goal of an N recommendation system is to accurately estimate the gap between the N provided by the soil and the N required by the plant. Accurately estimating this gap depends on the ability of the recommendation system to accurately estimate fi eld or subfi eld specifi c economically optimal nitrogen rates (EONR). Current recommendation systems are not as accurate as needed to provide consistently reliable estimates of N needs across years at the fi eld or subfi eld scale. Uncontrollable factors like temperature, rainfall timing, intensity and amount, and interactions of temperature and rainfall with factors such as N source, timing and placement, plant genetics, and soil characteristics combine to make N rate recommendations for an individual fi eld or rates for subfi elds a process guided as much by science as by the best professional judgement of farmers and farm advisors.Substantial evidence has accumulated that EONRs can vary widely across fi elds, within fi elds and over years in the same fi eld for a wide range of crops and geographies. Examples ABSTRACTNitrogen fi xation by the Haber-Bosch process has more than doubled the amount of fi xed N on Earth, signifi cantly infl uencing the global N cycle. Much of this fi xed N is made into N fertilizer that is used to produce nearly half of the world's food. Too much of the N fertilizer pollutes air and water when it is lost from agroecosystems through volatilization, denitrifi cation, leaching, and runoff . Most of the N fertilizer used in the United States is applied to corn (Zea mays L.), and the profi tability and environmental footprint of corn production is directly tied to N fertilizer applications. Accurately predicting the amount of N needed by corn, however, has proven to be challenging because of the eff ects of rainfall, temperature, and interactions with soil properties on the N cycle. For this reason, improving N recommendations is critical for profi table corn production and for reducing N losses to the environment. Th e objectives of this paper were to review current methods for estimating N needs of corn by: (i) reviewing fundamental background information about how N recommendations are created; (ii) evaluating the performance, strengths, and limitations of systems and tools used for making N fertilizer recommendations; (iii) discussing how adaptive management principles and methods can improve recommendations; and (iv) providing a framework for improving N fertilizer rate recommendations.

N to surface water is primarily by subsurface flow of nitrate (Schilling, 2002; Steinheimer et al., 1998), partic-Applying only as much N fertilizer as is needed by a crop has ecoularly when N fertilizer has been applied at rates exnomic and environmental benefits. Understanding variability in need for N fertilizer within individual fields is necessary to guide approaches ceeding crop needs (Burwell et al., 1976).to meeting crop needs while minimizing N inputs and losses. Our ob-Small-plot research has shown that experiments in jective was to characterize the spatial variability of corn (Zea mays L.) different production corn fields can differ substantially N need in production corn fields. Eight experiments were conducted in their need for N fertilizer (Bundy and Andraski, 1995; in three major soil areas (Mississippi Delta alluvial, deep loess, clay- Schmitt and Randall, 1994). Need for N fertilizer may pan) over 3 yr. Treatments were field-length strips of discrete N rates also vary widely over large fields (Malzer et al., 1996; from 0 to 280 kg N ha Ϫ1 . Yield data were partitioned into 20-m Mamo et al., 2003) though very little research has been increments, and a quadratic-plateau function was used to describe published addressing this issue. Attempts to predict the yield response to N rate for each 20-m section. Economically optimal amount of N fertilizer needed have met with limited N fertilizer rate (EONR) was very different between fields and was success in humid regions (Kitchen and Goulding, 2001). also highly variable within fields. Median EONR for individual fields ranged from 63 to 208 kg N ha Ϫ1 , indicating a need to manage NThe dominant practice for agricultural producers is to fertilizer differently for different fields. In seven of the eight fields, apply the same rate of N fertilizer over whole fields and a uniform N application at the median EONR would cause more than even whole farms. In fields with spatially variable N needs, half of the field to be over-or underfertilized by at least 34 kg N this practice leads to frequent mismatches between N ha Ϫ1 . Coarse patterns of spatial variability in EONR were observed fertilizer rate and crop N need. Overapplication is more in some fields, but fine and complex patterns were also observed in frequent since producers have an economic incentive most fields. This suggests that the use of a few appropriate manageto err more frequently in that direction: The cost of ment zones per field might produce some benefits but that N manageunneeded N fertilizer in areas of overapplication is less ment systems using spatially dense information have potential for than the cost of lost yield potential in areas of undergreater benefits. Our results suggest that further attempts to develop application. systems for predicting and addressing spatially variable N needs are justified in these production environments.

Nitrogen fertilizer is a fundamental input for production of corn (Zea mays L.) that can move to ground and surface waters when overapplied. Previous research has shown that chlorophyll meter (CM) readings can indicate N stress in corn, but has not addressed whether the amount of N needed can be predicted by CM readings. Our objective was to evaluate whether CM readings can predict corn N need and yield response to N. Sixty-six N rate experiments were conducted in seven north-central states over a 4-yr period. Linear regression was used to relate absolute and relative CM readings over a range of growth stages to economically optimal N rate (EONR) and yield response to N applied at growth stage V7 or earlier. Chlorophyll meter readings at all growth stages from V5 to R5 were significantly related (P , 0.0001 in 22 of 24 models, P , 0.01 in 2 models) to EONR and yield response to N. Relationships were stronger for relative than for than absolute CM readings, and also were stronger when the corn had received no N fertilizer at planting. Coefficients of determination ranged from 0.53 to 0.76 for relative CM reading as a predictor of EONR or yield response to N, and were lower for the V5 to V9 stage than for later stages. Earlier research has indicated that measurements with this level of predictive accuracy can produce N rate recommendations that are more profitable than current N management practices. Our findings suggest that CM readings (and potentially other measures of corn color) are quantitatively related to early-season EONR and yield response to N over a wide range of environments with enough accuracy to be helpful in making management decisions.

Nitrogen fertilizer is typically applied to corn (Zea mays L.) shortly before planting, but there are several reasons why later N applications may be of interest: to spread work away from the busy planting season, to avoid the frequent wet field conditions in spring, to reduce or remedy in‐season N loss in wet years, or to allow use of in‐season diagnostic tools. One of the obstacles to the use of later N applications is the fear that irreversible yield loss will occur due to N stress. Our objective was to evaluate the yield impact of delaying N applications until the late vegetative growth stages and as far as silking. We conducted a total of 28 experiments with timing of a single N application as the experimental treatment. We found little or no evidence of irreversible yield loss when N applications were delayed as late as stage V11, even when N stress was highly visible. There was weak evidence of minor yield loss (about 3%) when N applications were delayed until stage V12 to V16. Only 3 of the 28 experiments had N applications later than V16—all were at silking and relative yields were 0.71, 0.89, and 0.95. Though full yield was not achieved when N applications were delayed until silking, yield was still highly responsive to N application at this stage—yield response exceeded 2.2 Mg ha−1 in all three experiments.