Agronomy Journal Grow Your Career With ASA
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Published in Agron J 100:328-336 (2008)
DOI: 10.2134/agrojnl2007.0145
© 2008 American Society of Agronomy
677 S. Segoe Rd., Madison, WI 53711 USA
This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kucharik, C. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Kucharik, C. J.
Agricola
Right arrow Articles by Kucharik, C. J.
Related Collections
Right arrow Agroclimatology
Right arrow Crop Growth and Development
Right arrow Global Change
Right arrow Maize
Right arrow Maize Management

AGROCLIMATOLOGY

Contribution of Planting Date Trends to Increased Maize Yields in the Central United States

Christopher J. Kucharik*

Center for Sustainability and the Global Environment (SAGE), The Nelson Institute for Environmental Studies, 1710 University Avenue, Univ. of Wisconsin, Madison, WI 53726

* Corresponding author (kucharik{at}wisc.edu).


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Early planting of maize (Zea mays L.) allows for longer-season hybrids to be used in cool temperate regions. Given that a multidecadal trend toward earlier planting has been occurring across the Corn Belt, it was hypothesized that this shift has supported a portion of recent yield increases. The objectives were to quantify relationships among state level monthly climate variables, maize yields, and planting dates, and to investigate whether multidecadal trends of earlier planting contributed to rising yields during 1979 to 2005 in 12 central U.S. states. Year-to-year changes (i.e., first differences) of predictor variables (monthly mean temperature and precipitation and planting date) and yields were calculated, and multiple linear regression was used to estimate the effect of planting date trends on maize yield increases. In six of the 12 states, a significant relationship (P < 0.05) existed between first differences of planting dates and yields. Multiple linear regression suggested that the management change has potentially contributed between 19 and 53% of the state level yield increases in Nebraska, South Dakota, Minnesota, Iowa, Wisconsin, and Michigan. Yield increases between 0.06 and 0.14 Mg ha–1 were attributed to each additional day of earlier planting, which likely reflects a gradual adoption of longer-season hybrids. Thus, if these earlier planting trends were to suddenly abate, a falloff in annual yield increases may follow in several Corn Belt states. Maize production in northern U.S. states appears to have benefited more significantly from earlier planting due to a shorter growing season in contrast to more southern locations.

Abbreviations: DOY, day of year • GDD, growing degree days • USDA, United States Department of Agriculture • NASS, National Agricultural Statistics Service


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.

Received for publication April 18, 2007.
   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
MAIZE PRODUCTION is dependent on soil fertility, precipitation, temperature regimes, inputs such as fertilizers and irrigation (Ramankutty et al., 2002), and management decisions such as planting date and hybrid selection. This makes the study of connections between climate change, land management, and crop productivity valuable to investigate future food security. Several studies have revealed the effects of climate change on agricultural yields, highlighting future vulnerability if warming trends continue (Andresen et al., 2001; Hu and Buyanovsky, 2003; Lobell and Asner, 2003; Peng et al., 2004; Hu et al., 2005; Tao et al., 2006; Lobell and Field, 2007). However, improvements and changes in technology and farm management may serve as adaptive measures that may partially or completely offset the unfavorable effects of climate change on crops.

Previous research has documented how historical yield increases have resulted from improved technology and hybrids, better agronomic management practices, and increased inputs (Duvick, 1989; French and Headley, 1989; Duvick, 1992; Duvick and Cassman, 1999; Mann, 1999; Brown, 2005; Kucharik and Ramankutty, 2005; Stewart et al., 2005). Other studies have demonstrated correlations between monthly climate and crop yields (Thompson, 1969, 1986, 1988). However, there have been few quantitative studies that have apportioned the contribution of changing management to the magnitude of the historical U.S. yield trend in recent years (Cardwell, 1982; Duvick, 1989; Kaylen and Koroma, 1991; Duvick and Cassman, 1999; Lauer et al., 1999; Lobell and Asner, 2003), or how climate might be impacting management decisions. Rarely have climate influences on planting dates been demonstrated, or planting date relationships to actual yields been correlated in a regional context. This knowledge gap makes it difficult to hypothesize about the future vulnerability of agriculture to climate change. More studies are needed that quantify the specific effects of land management changes on yield, and climate effects on both management decisions and productivity.

The relationship between one important management decision, planting date, and yield potential has been previously documented by agronomists (Gupta, 1985; Swanson and Wilhelm, 1996; Epplin and Peeper, 1998; Lauer et al., 1999; Nielsen et al., 2002). For example, 8% of the yield increase in Minnesota during the 1930 to 1979 time period was credited to a shift toward earlier planting by 10 d (Cardwell, 1982), which was approximately equal to a yield increase of 0.031 Mg ha–1 for every additional day that was added to the growth period. Because earlier planting increases the length of time that plants can take advantage of favorable growing conditions and accumulate biomass, the highest yields generally result where the growing season is longest and soil moisture is nonlimiting (Kucharik, 2006). Earlier planting increases the likelihood that higher yield potential (e.g., late maturing or long season) maize hybrids will reach maturity before killing frost arrives in the fall, and that flowering will take place before midsummer heat stress is most likely to occur (Duvick, 1989).

Kucharik (2006) reported that a multidecadal trend toward earlier maize planting has occurred concurrently with maize yield increases across the U.S. Corn Belt from 1979 through 2005 (Fig. 1 ), and is presumed to be an extension of a much longer trend dating back to around 1930 (Duvick, 1989). During the 2001 to 2005 timeframe, the initiation of maize planting began as early as the first week of April in Missouri and Kentucky, while the states of Wisconsin, Minnesota, and South Dakota experienced the latest start to planting, with seed sowing commencing during the last few days in April (Table 1 ). Kucharik (2006) showed that maize planting is therefore occurring 2 wk earlier than it was in the late 1970s, with an overall trend for the entire Corn Belt of –0.48 d yr–1 earlier (Fig. 1), and concluded that factors other than climate change (e.g., springtime warming) were the driving force behind these large-scale management changes. These factors included a shift toward conservation tillage and plowing in the fall, and several technological factors including the development of hybrids with a higher tolerance to suboptimal growing conditions (e.g., cold, wet soils) or seeds coated with temperature activated polymers (Gesch and Archer, 2005), increased resistance to diseases and pests, and improved equipment (planter) functioning (Lauer, 2001; Bruns and Abbas, 2006). Regardless of the exact causes that have supported earlier maize planting, this trend could have supported a switch to hybrids suited for a longer growing season, which could contribute to yield gains (Swanson and Wilhelm, 1996; Lauer et al., 1999; Nielsen et al., 2002), or it could lengthen the time window that farmers had to complete their planting before significant weather delays would decrease end-of-season yields, particularly in northern Corn Belt states. Figure 2 illustrates the wide range of average growing degree days (GDD; base 10°C) accumulated from 1April to 30 September each year (1979–2005) across the Corn Belt, substantiating the idea that a considerable range in management practices (e.g., hybrid selection and planting date) is necessary to adapt to different growing conditions and maximize production, particularly in northern locations that are more temperature limited.


Figure 1
View larger version (19K):
[in this window]
[in a new window]

  		Fig. 1. Multidecadal trends in Corn Belt average maize planting date (day of year that 10% of planting is completed) and maize yield. The regional averages are calculated using an area-weighted approach based on the harvested area of maize in each state (SD, NE, KS, MN, IA, MO, WI, IL, MI, IN, OH, KY) and its contribution to the regional total.

 

View this table:
[in this window]
[in a new window]

  		Table 1. Summary of the average maize 10% planted date for 2001-2005 (standard deviation in parentheses), the average maize yields about 1979 and 2005, the total absolute yield difference between average yields in 1979 and 2005, and the percent yield increase during the 1979 to 2005 period relative to state level yields about 1979. The average state level corn yields calculated for about 1979 and 2005 were derived by fitting a linear trend to annual yield values and calculating the expected yield value along the trendline for those two particular years.

 

Figure 2
View larger version (27K):
[in this window]
[in a new window]

  		Fig. 2. Average number of growing degree-days (base 10°C) accumulated between 1 April and 30 September (inclusive), over the 1979 to 2005 time period. The calculation used the daily average temperature from the National Centers for Environmental Prediction (NCEP) Reanalysis gridded climate data, available from the National Oceanic and Atmospheric Administration (NOAA) CIRES Climate Diagnostics Center in Boulder, CO (http://www.cdc.noaa.gov; verified 19 Dec. 2007).

 
Average maize yield increases over the past several decades have ranged from 1.5 Mg ha–1 in Kansas to greater than 3.5 Mg ha–1 in Minnesota, Iowa, and South Dakota (Table 1). These trends represented an increase in maize yields ranging from 21% to 108% relative to their average values several decades ago (Table 1). Table 1 illustrates the wide variation in the percent yield increases across the Corn Belt, where the overall range in average state level maize yields has decreased from 3.7 Mg ha–1 (i.e., the difference between 7.3 and 3.6 Mg ha–1 in Kansas and South Dakota, respectively, at the beginning of the period) to 2.6 Mg ha–1 (i.e., the difference between 10.0 and 7.4 Mg ha–1 in Iowa and South Dakota, respectively, at the end of the period).

These recent trends raise an important question of whether a long-term trend toward earlier planting have contributed to yield gains given the known importance of early seed sowing to end-of-season maize yields? If we are able to quantify the impacts of changes in farm management on yield gains, then we may better understand the impact that future changes in management might have, and whether these could help curtail negative impacts of climate change on crop yields through adaptive measures. The research presented here tested the hypothesis that a trend toward earlier planting of maize from 1979 to 2005 has contributed significantly to the linear increase observed in maize yields in the central United States over the same time period. Because it is nearly impossible to isolate the impact of planting date changes on yield trends when genetic advances have simultaneously improved resistance of seedlings to less than optimal soil conditions in the early spring, thereby allowing for earlier planting, we suggest the results here represent a composite response of those associated improvements in hybrids. Datasets of maize planting dates, yields, and monthly mean temperature and precipitation– all at the state level– were combined through statistical analyses to identify key relationships between climate, planting dates, and yields, and to assess whether recent climate trends, planting date trends, or a combination of both have contributed to increases in maize yields.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
This study used a dataset of state level maize planting dates for 12 central U.S. states (IL, IN, IA, KS, KY, MI, MO, MN, NE, OH, SD, and WI) for the 1979 to 2005 time period (Kucharik, 2006). Weekly maize planting progress data (expressed as a percentage of planting that is completed), available through the USDA-NASS (available at www.nass.usda.gov; verified 19 Dec. 2007), were previously analyzed to produce a continuous daily record of planting progress. The date of occurrence for planting progress at the 10% threshold of completion was subjected to linear regression analysis to quantify statistical trends for each individual state over the 27-yr study period (Kucharik, 2006). The key data used in the current study were the day-of-year (DOY) that 10% of planting was completed, designated as the planting date, and the statistical trends of this for each state as reported in Kucharik (2006). The 10% planted date was designated as the lowest planting progress threshold that signified an initiation of planting. This threshold was used because the weekly USDA crop progress reports do not begin citing the percentage of planting completed until values are typically between 5 and 10%. State level maize yield data obtained from USDA-NASS were also compiled for the 1979 to 2005 time period. State level monthly climate data (mean temperature and total precipitation) for the months of March through August were obtained from the National Climatic Data Center (NCDC) (available at www7.ncdc.noaa.gov/CDO/CDODivisionalSelect.jsp; verified 19 Dec. 2007).

Several statistical approaches (all using the SAS-JMP statistical discovery software package) (SAS Institute, 2002) tested for the presence of significant relationships and correlations between monthly climate variables, and annual planting dates and yields. First, simple linear regression models were used to illustrate the general relationships between key variables such as yield, planting date, and monthly climate. While an earlier initiation of planting could support higher yields by expanding the potential growth period for plants, the actual impact of planting date trends on yields can only be investigated after the effects of other factors not directly associated with planting are accounted for and are removed from the analysis (Nicholls, 1997). This is complicated further because climate has the potential to influence both planting date and yields simultaneously. Therefore, a previously cited approach was used to detrend datasets and evaluate correlations between the time series of planting date, yield, and climate. This method calculated annual year-to-year changes in each variable, referred to as calculation of the first differences (Nicholls, 1997; Peterson et al., 1998; Free et al., 2004; Lobell et al., 2005; Lobell and Field, 2007) in each state's time series of planting dates, yields and all monthly climate variables. The first differences methodology simply calculates a time series of the numerical differences of each variable in successive years (e.g., the value of a variable for the previous year is subtracted from the value in the current year), creating a dataset in which long-term trends are effectively removed to determine year-to-year variability (Nicholls, 1997). Multiple linear regression analysis was then used to determine the relationship between first differences of planting dates and yields in the context of considering other climate factors. Interactive terms that represented the effect of springtime weather (e.g., temperature and precipitation for the months of March through May) on planting dates and yields were also included in the multiple linear regression approach. The coefficients from these regression relationships can then be used to estimate how longer-term trends in planting dates would have impacted yield trends (Nicholls, 1997). For individual monthly climate data in each state, as well as the state maize yields, linear regression analysis was used to determine the slope of any long-term trends and the level of statistical significance (or P value).


   RESULTS AND DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Using a simple regression model incorporating raw state level data (i.e., trend not removed), the correlation between state level planting dates and yields for all states from 1979 to 2005 suggests that an increase of approximately 0.066 Mg ha–1 in maize yield resulted with each day of earlier planting (Fig. 3 ). Furthermore, mean April temperature appears to explain 45% of the variation in annual planting dates across the region from 1979 to 2005 (Fig. 4 ). However, determining the true effect of planting date trends on yield increases requires a more detailed analysis after removal of the long-term trends.


Figure 3
View larger version (27K):
[in this window]
[in a new window]

  		Fig. 3. Scatter plot of all annual state average maize yields and the 10% corn planted dates (day of year) for the 1979 to 2005 period, with best-fit linear regression line.

 

Figure 4
View larger version (27K):
[in this window]
[in a new window]

  		Fig. 4. Scatter plot of all annual state average 10% maize planted dates (day of year) and state average April temperatures for the 1979 to 2005 period, with best-fit linear regression line.

 
Climate vs. Yields and Planting Dates
In 11 of 12 states (all except Missouri) analyzed, a significant (P < 0.05) relationship between first differences of planting date and April precipitation was present, highlighting a delayed planting response attributed to above average April rainfall. When all of the state level data were pooled in a separate statistical regression analysis (Fig. 5 ), year-to-year changes in April rainfall appear to explain 31% of the year-to-year variability in maize planting across the Corn Belt. Furthermore, planting is delayed by approximately 1 d for every 10 mm increase in total April precipitation (Fig. 5). Additionally, 9 of the 12 states had a significant relationship (P < 0.05) between first differences of April temperature and planting initiation. Therefore, weather conditions during the month of April are a primary determinant of recent and current planting behavior across the region.


Figure 5
View larger version (31K):
[in this window]
[in a new window]

  		Fig. 5. Scatter plot of first differences of all annual state level April precipitation and 10% maize planted dates (day of year) for the 1979 to 2005 period, with the best-fit linear regression line.

 
All 12 states showed a significant relationship between first differences of yields and August temperatures. The overall relationship for all 12 states (Fig. 6 ) suggests that years with the highest year-to-year yield increases were found when August temperatures decreased, which is in agreement with previous studies (Thompson, 1988; Hu and Buyanovsky, 2003). This pooled-data regression analysis suggests that for every additional degree (°C) increase in August mean temperature, yields decreased by 0.39 Mg ha–1 (Fig. 6) and that 40% of the variability in maize yields at the state level were explained by August temperature fluctuations, whereby cooler August temperatures led to higher yields. This is likely attributed to decreased heat stress during grain fill, lower nighttime respiratory losses of carbon and water, or a slower transition through the last phenological stages allowing for more grain biomass to accumulate. Eight of the 12 states showed a significant relationship between first differences of yield and July temperatures, and higher than average precipitation in July supported above average yields in 8 of the 12 states at the 95% confidence level. This supports other studies that have quantified the relationship between monthly summertime climate variables and yields across the Corn Belt (Thompson, 1986 and 1988; Kaylen and Koroma, 1991; Lobell and Asner, 2003). While this discussion highlights some of the most important relationships between monthly climate, planting dates, and yields, it was not the intent here to provide a detailed account of the impact of climate on yields given higher spatial resolution yield (e.g., county level) and climate data is available (Lobell and Asner, 2003), which are likely to contribute to more robust results. Nonetheless, this study does illustrate some of the more general relationships that occur over large regions.


Figure 6
View larger version (30K):
[in this window]
[in a new window]

  		Fig. 6. Scatter plot of first differences of all annual state level August temperatures and maize yields for the 1979 to 2005 period.

 
Planting Dates vs. Yields
Six states (IA, MI, MN, NE, SD, and WI) had a significant (P < 0.05) relationship between the first differences of planting dates and yields, suggesting that annual maize yields were significantly affected by the date of seed sowing across a large portion of the region (Fig. 7 ). The six states were predominantly found on the northern and western periphery of the Corn Belt, leaving out portions of the Corn Belt that typically have larger growing season degree-day accumulations such as Kansas, Missouri, Illinois, Kentucky, and Indiana (Fig. 2). The northern states might logically be places where earlier planting would have had the largest impact toward helping increase maize yields, given they are more temperature limited than regions further to the south (Gupta, 1985; Lauer et al., 1999; Darby and Lauer, 2002). In many of these states (IA, MI, NE, OH, SD, and WI), there was a significant relationship between first differences of maize yield and April and May temperature and precipitation, which coincided with the relationship of planting date with the same climate variables in those states. This result further substantiates the important connection between springtime weather conditions, planting date, and an influence on yields in any particular year, and support the claim that springtime climate effects can confound the apparent impact of earlier planting on yields.


Figure 7
View larger version (27K):
[in this window]
[in a new window]

  		Fig. 7. Scatter plots of first differences of state level 10% maize planted dates ({Delta} Planting date) and {Delta} Yield for the 1979 to 2005 period. Best-fit linear regression lines are plotted in cases where the relationship was significant at P < 0.05.

 
Table 2 highlights the multiple linear regression models for climate, planting data, and maize yields for the six states where first differences of planting dates were significantly related to first differences of maize yields. Predictor variables were chosen using both stepwise forward and backward multiple linear regression in the JMP statistical package. Akaike's Information Criterion (AIC), adjusted R2 quantities, and individual P values were used as general guides in choosing the most important predictors to explaining yield variability. Akaike's Information Criterion provides an estimate of the expected, relative distance between the fitted model and the unknown true mechanism that generated the data and accounts for the trade-off between the number of parameters in the model and the model fit. Interaction terms between planting date and all springtime temperature and precipitation data were also added to the multivariate linear regression models. In all states except Nebraska and Kansas, the years of 1988 and 1989 were removed from the analysis after a separate test determined that these data points were outliers. This might be expected given the catastrophic drought that occurred in 1988, whereby yield losses were substantial and were in no way correlated with planting dates that year. Because yields in 1988 were very low, the year-to-year change from 1988 to 1989 is magnified or exaggerated also, thereby causing relationships between yield and planting date changes to be unrepresentative in 1989. It is likely that the effects of irrigation in Nebraska and Kansas helped support higher yields and less variability in those years of 1988 to 1989, and therefore were not identified as outliers. All of the state level regression models contained both springtime and summertime climate variables, along with at least one interaction term (Table 2). However, none of the interaction terms were a significant factor in any of the overall equations. In general, the adjusted R2 values for models ranged from a low of 0.31 for South Dakota, to a high of 0.73 in Wisconsin (Table 2). The results in Table 2 show one general and expected pattern relating climate and yields. Late summertime temperatures generally have a very significant impact to maize yields regardless of geographic location. For example, these results suggest that higher than expected temperatures in July and August generally cause declines in yields. The picture is more mixed for springtime variables.


View this table:
[in this window]
[in a new window]

  		Table 2. Results of multiple linear regression models between first differences of 10% maize planting date (days), maize yields (Mg ha–1), and monthly climate (mm or °C). Statistics are only reported for states that had a significant relationship (P < 0.05) between first differences of planting date and yield.

 
Previous research has already documented that long-term trends in the initiation of maize planting across the central United States has become earlier on average by about 0.48 d yr–1 during the 1979 to 2005 period, ranging from –0.82 d yr–1 in Missouri to no significant trend in Ohio (–0.23 d yr–1) (Kucharik, 2006). These results are shown in Table 3 , along with the state level maize yield trends over the same time period that ranged between 0.06 Mg ha–1 yr–1 in Kansas to a high of 0.15 Mg ha–1 yr–1 in Minnesota. The result of the multiple linear regressions suggested that for every day that planting was delayed, maize yields decreased between –0.06 Mg ha–1 d–1 in Wisconsin to –0.14 Mg ha–1 d–1 in Iowa (Tables 2 and 3). The overall relationship resulting when all state level data were pooled in a linear regression model suggests that an annual yield decline of 0.066 Mg ha–1 occurs for each day of later planting (Fig. 8 ). This is similar to the response derived from comparing raw planting date and yield values that were not detrended (Fig. 3). All of these results compare very well with field trials conducted in the Midwest that have shown that maize yields decrease between 0.04 and 0.33 Mg ha–1 d–1 for every day that planting is delayed past the optimal window (Lauer et al., 1999), although our estimated response appears to be on the conservative side. However, a conservative estimate is not surprising considering our response is derived using quantity averages (e.g., yield and planting date) over large geographic regions; this undoubtedly smoothes out some of the spatial variability across each state that occurs because of varied weather and management choices each year.


View this table:
[in this window]
[in a new window]

  		Table 3. Summary of statistical trends in the state level 10% planted date from 1979 to 2005 (from Kucharik, 2006), the linear trend of annual yield increases and total maize yield increase in each state between 1979 and 2005, the result of multiple regression models testing for a relationship between first differences of maize yield and planting dates, and the estimated contribution of earlier planting to yield trends expressed in Mg ha–1 and as a percentage of the overall yield increases.

 

Figure 8
View larger version (29K):
[in this window]
[in a new window]

  		Fig. 8. Scatter plot showing first differences of all annual state level 10% maize planted dates (day of year) and maize yields for the 1979 to 2005 period, with best-fit linear regression line.

 
Analysis of Long-Term Climate Trends
Each monthly climate timeseries for each state was analyzed (mean temperature and precipitation for March through August) to detect cases where (1) significant (P < 0.1) long-term trends in monthly climate were observed, and (2) trends were present for months and states where a significant relationship existed between first differences of climate variables, and planting dates and yields. Linear regression suggested that only 8 out of the possible 144 mo of total climate records (i.e., a matrix of 12 states by 6 mo by two climate variables) had a significant trend at the P < 0.1 level, and only two showed a trend at the P < 0.05 probability level. In the states of Iowa, South Dakota, Illinois, Kansas, Kentucky, and Missouri, there were no significant trends in any of the analyzed monthly climate variables over the 1979 to 2005 time period. In Indiana and Ohio, significant trends in April mean temperature were noted at the P < 0.1 level, but because there were not any significant relationships in first differences of either yield or planting date with these climate variables (Table 2), they were deemed to have not contributed significantly to trends in planting date or yield. In Nebraska, a significant trend in March precipitation (P = 0.068) was noted, but first differences of this variable were not significantly correlated with first differences in planting date or yield. In Minnesota, a significant long-term trend of increasing precipitation in May (P = 0.043) was noted, but was not identified as having a significant correlation with year-to-year changes in planting dates or yield (Table 2). In Wisconsin, a significant (P = 0.036) trend of decreasing August precipitation was found over the 27-yr period, but first differences of this variable were not significantly correlated with year-to-year changes in planting date and yields, respectively (Table 2).

The results of the aforementioned statistical analysis suggest that climate change was not a significant driver to either planting date trends or yield trends during the 1979 to 2005 period at the state level. While it was previously discussed in Kucharik (2006) that long-term trends in planting date did not appear to be attributed to warmer springtime weather conditions, that analysis used an independent climate dataset and different measures of determining whether trends in springtime weather conditions were driving earlier planting trends. Thus, there is strong evidence that planting dates and their long-term trends are indeed driven by other changes in agronomic practices and technological advances across the central United States (Kucharik, 2006). However, the apparent lack of climate being a significant driver to yield trends over the central United States is in contrast to the report of Lobell and Asner (2003) that showed how increasing yield trends from 1982 to 1998 over the United States were supported by a trend toward cooler summertime (June–August) temperatures. Nevertheless, in a follow-up study at the global scale, Lobell and Field (2007) noted that the choice of period of study was somewhat influential on determining the temperature effects on maize yield trends, stating that increased global warming after 1998 has likely tempered the impacts of a cooler climate period that enhanced U.S maize yields from 1982 to 1998. Additionally, there are two other potential reasons for these discrepancies between studies: (i) the length of data record was different between the two studies, and (ii) the resolution and source of climate information was different in these studies. The Lobell and Asner (2003) study used 0.5° gridded climate and crop yield information in their analyses, which would have a higher likelihood of showing more variability in a spatial context. State-level climate data likely reduces the probability of picking up more subtle long-term trends, but nonetheless is still useful because both crop yields and planting dates have also been aggregated to the state level.

The main shortcoming of the current approach is that climate data at the state level applies to a large area covering the entire state, but not all of the observations (and regions) that contributed to state level climate averages are planting and growing maize. Thus, the contribution of climate to yield trends in the current approach might be more difficult to ascertain. Furthermore, given that irrigation is used extensively across Kansas and Nebraska, the results may potentially be skewed across that region. Nonetheless, given that significant relationships existed between first differences of monthly climate variables and planting dates and yields that were consistent with previous studies suggests that state level data is still valuable in this type of analysis. Performing an analysis with higher resolution management data in the future, possibly made available through combining census data and satellite observations, could also reduce the uncertainty in future predictions. Future work will use both higher resolution climate and management data coupled with an agroecosystem model to further investigate climate-yield-management interactions.

Contribution of Planting Date Trends to Yield Increases
The contribution of planting date trends to yield increases was computed for the states of Iowa, Michigan, Minnesota, Nebraska, South Dakota, and Wisconsin (Table 3). These states had a significant (P < 0.05) relationship between first differences of planting dates and yields (Fig. 7), and year-to-year differences in planting dates contributed significantly as a predictor variable in the multivariate linear regression analysis. For other states, it was determined through separate multiple regression that trends toward earlier planting had negligible (i.e., not significant) impacts to yield trends (Table 3). The impact of earlier planting on yield trends was computed by multiplying the yield response due to planting date by the trend in planting dates over the 1979 to 2005 period (Table 3). These calculations showed that the estimated increase in yields at the state level due to earlier planting ranged from a high of 1.86 Mg ha–1 in Iowa to a low of 0.58 Mg ha–1 in Wisconsin. These values were then converted to a percent increase when considering the total change in yields over the time period (Table 3). This analysis showed that earlier planting trends contributed between 19% (Minnesota) and 53% (Iowa) to the total yield trends in six states, but was insignificant in the remaining states analyzed.


   CONCLUSIONS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
The results presented here illustrate how dependent agricultural production across the Corn Belt is to the timing of seasonal weather events, which has also been previously documented in many regional analyses (Thompson, 1988; Hu and Buyanovsky, 2003; Lobell and Asner, 2003). April precipitation is a dominant climate factor driving soil moisture and trafficability, affecting maize-planting progress across a large portion of the region (Fig. 5), and August temperature (Fig. 6) and July precipitation are both significantly correlated with year-to-year changes in yield. Future projections of climate change for these time periods will prove to be extremely useful in estimating the future impacts of global change on maize yields across the U.S. Corn Belt. Attention should be directed toward future seasonal weather changes and their effects at the weekly to monthly timescale, and their subsequent impacts on management choices and productivity.

The current trend toward earlier maize planting dates appears to have contributed between 19 and 53% to recent yield gains in several states in the northern and western portions of the Corn Belt (Nebraska, Iowa, South Dakota, Minnesota, Wisconsin, and Michigan), the most likely geographic locations that would be able to take advantage of earlier planting and expect to see a significant contribution to increasing yields. Because delayed planting has been shown to lead to yield declines, it might be expected that a long-term trend of earlier maize planting across the Corn Belt has played an important role in supporting yield increases since the late 1970s. A secondary management change or adaptation coupled to this earlier planting phenomenon has likely been a shift toward selecting and planting higher yield potential hybrids in these areas so they can take advantage of the resulting growing season extension. It is also possible that earlier maturing hybrids, which have lower yield potentials and are planted more extensively across the northern and western periphery of the Corn Belt states, have produced slightly higher yields over the years because they also have experienced a longer period of potential growth and decreasing risk of killing frost before physiological maturity is attained. However, we emphasize that other genetic advances have simultaneously improved the resistance of seedlings to less than optimal soil conditions in the early spring, thereby allowing for the earlier planting to take place. This makes it difficult to completely isolate the impact of earlier planting on yield increases, and thus we suggest the results here represent a composite response of those simultaneous improvements in hybrids.

Through these management changes, farmers have increased the likelihood that they can complete all of their planting before yield declines could result due to large delays in seed sowing. Another potential benefit is that the plant may be progressing through later key phenological stages (i.e., silking or tasseling) at an earlier point during the summer when water stress and problems with pollination and disease are less likely to occur. While climate does not appear to be a significant driving force behind these planting trends, nor to the larger scale yield trends, the latter finding is currently in opposition to other studies (Lobell and Asner, 2003; Lobell and Field, 2007). This aspect suggests that an even more detailed analysis with finer scale management data will be needed in the future to isolate the specific effects of management and climate on yield trends.

Even though planting date trends did not significantly contribute to yield trends in the states of Kansas, Missouri, Illinois, Indiana, Ohio, and Kentucky, contributions may still exist at the county or crop reporting district level, or may become apparent at the state level over a longer time period. However, there should be some concern that if planting date trends eventually cease to exist in any location, then a slowdown in annual yield gains may begin to materialize as the period of active growth can no longer be extended, unless significant future warming occurs during the fall, specifically during the months of September and October. As Kucharik (2006) acknowledged, farmers will eventually be limited by frozen soils and wintertime conditions that inhibit field operations altogether in the springtime. The duration of plant growth is a key determinant to yield potential, and it may become more difficult in the future to increase that period of time as the trend toward earlier planting reaches a plateau.


   ACKNOWLEDGMENTS
 
We extend gratitude to the associate editor and three anonymous reviewers for providing detailed comments and suggestions that greatly improved the manuscript. The work was supported by two sources: (1) a NASA Interdisciplinary Science grant, and (2) by the Department of Energy under Award Number DE-FC02-06ER64158, through the National Center for Climate Change Research (NICCR). This report was prepared as an account of work sponsored by an agency (DOE) of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability of responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof.

All rights reserved. No part of this periodical may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 




This article has been cited by other articles:


Home page
 Plant Physiol.Home page
M. D. Edgerton
Increasing Crop Productivity to Meet Global Needs for Feed, Food, and Fuel
Plant Physiology, January 1, 2009; 149(1): 7 - 13.
[Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Kucharik, C. J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Kucharik, C. J.
Agricola
Right arrow Articles by Kucharik, C. J.
Related Collections
Right arrow Agroclimatology
Right arrow Crop Growth and Development
Right arrow Global Change
Right arrow Maize
Right arrow Maize Management

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Crop Science Vadose Zone Journal
Journal of Natural Resources
and Life Sciences Education		 Soil Science Society of America Journal
Journal of Plant Registrations Journal of
Environmental Quality		 The Plant Genome