Additionally, diversion of corn to ethanol production essentially reduces food supply, which may lead farmers worldwide to convert natural land to new cropland in order to compensate for the diverted grain . This indirect LUC would generate the same net effects as that of direct LUC as discussed above, although its estimation is much more complicated for the difficulty and uncertainty involved in quantifying the impacts of US bio-fuel policies on global land and agricultural commodity markets . Studies continue to explore the effect of indirect LUC of bio-fuels with refined modeling methodologies , improved understanding of agricultural and food systems around the world , and extended assessment to non-GHG impacts . Yet, there is another consequence of corn ethanol expansion to which relatively less attention has been paid. In conjunction with rising corn prices, substantial land cover shift from cotton to corn has been observed—particularly between 2005 and 2009—through both direct expansion of corn into cotton and indirect expansion of corn into soybean, then of soybean into cotton . This observation is supported by farm-level data, which reveal that as growing corn became more profitable, some farmers reacted by reducing cotton land for growing corn . Furthermore, the National Agricultural Statistics Service Cropland Data Layer provides high-resolution maps derived from satellite imagery clearly demonstrating that land shifts from cotton to corn occurred in several states . Overall, between 2006 and 2009 when corn prices increased substantially relative to cotton prices , cotton area harvested reduced by 40 % , while corn area in the cotton growing states expanded by 1.3 million ha . Despite the potential large-scale land shift from cotton to corn,black plastic nursery pots there have been few studies on associated environmental impacts. Here, we address this knowledge gap.
We note that corn displacing cotton was only part of the complex land use dynamics in the past “ethanol decade” that involved also land shift from, for example, soybeans and hay to corn, cotton to soybeans, and natural vegetations to corn . The reason we focus only on cotton to corn here is that environmental impacts of land shift between cotton to corn, both high-input crops, are less clear than that between relatively low-input crops and high-input crops . In a recent study, Wallander et al. stated that “When acreage shifts from one high-input crop to another , however, ethanol induced changes may be negligible or could even reduce environmental externalities.” In this study, we seek to test the validity of this statement, focusing on regional environmental issues along with a growing body of literature on the non-GHG consequences of bio-fuels expansion . A land shift from one crop to the other can alter both direct, or on-site, and indirect, or offsite, environmental effects. For example, increased use of nitrogen fertilizers as a result of the land shift not only can elevate N related emissions such as NOx and N runoff but also requires more energy and material inputs in the process of fertilizer production. The system boundary of the study, therefore, was drawn to cover both direct and indirect emissions. In particular, we paid a special attention to direct environmental emissions from crop production given their significance relative to indirect emissions . We calculated indirect emissions embodied in input materials that take place along supply chains, using the Ecoinvent database . In our data compilation, we placed an emphasis on the crop growth and agricultural input structures at the state level, as previous studies showed that national, average data may fall short in capturing the environmental impacts of crop production at a regional level . This is because agricultural systems display high degrees of variability across regions in terms of input structure due primarily to differences in geography, weather patterns, soil type, and management practices .
Also, data on major agricultural inputs such as fertilizers and pesticides collected by the US Department of Agriculture are only available at the state level . The reference year of this study is 2005 given that cotton area experienced a substantial decline between 2005 and 2009. Major inputs in crop growth include fertilizers, pesticides, energies, and irrigation water. We obtained relevant state-level data from several USDA surveys and censuses reflecting cotton and corn farming practices around 2005 and then compiled a set of state-specific inventories. Not all inputs data, however, are available for every state that grows cotton and corn. The USDA Farm and Ranch Irrigation survey, for example, includes more states than surveys of energy and agrichemical use. Nevertheless, the states for which all inputs data are available capture the majority of US cotton and corn production. Specifically, the inventories we compiled cover 19 corn growing states, which account for 95 % of domestic corn production in 2005, and 9 cotton growing states, which account for 88 % of domestic cotton production in 2005. Due to use of agricultural inputs like fertilizers and pesticides, crop production contributes to an array of environmental impacts from acidification, eutrophication, water scarcity to human and ecological toxicity . To best capture these impacts associated with US cotton and corn growth, we estimated all potential onsite environmental emissions based on various databases, models, and literature . The emissions data compiled cover >100 different substances, the majority of which are pesticides and volatile organic compound emissions. Numerical information on all emission factors used in this study can be found in the Table S1–S6 . After compiling emissions data for cotton and corn, we evaluated their environmental impacts using characterization factors from life cycle impact assessment . Reflecting the relative significance of an emission or resource, characterization factors are used to aggregate emission results, usually including a large number of different substances, into a dozen of impact category scores that enable better comparison between alternatives . In this study, we focused on regional environmental aspects of cotton and corn, and based on our previous study , we selected eight impact categories to which cotton and corn production potentially contribute.
These impact categories are acidification, eutrophication, smog formation, freshwater ecotoxicity, and water use as well as human health cancer, non-cancer, and respiratory effects. Characterization factors for all categories except water use are taken from the Tool for the Reduction and Assessment of Chemical and Other Environmental Impacts developed for the USA by the EPA . Characterization factors for water use were based on the ReCiPe model . Note that TRACI 2.0, compared with its original version , has incorporated the recently developed USEtox model for the ecotoxicty, human health cancer, and non-cancer impact categories .For comparison between the two crops, results are organized on the basis of per hectare produced. Figure 2.1 shows the average environmental impacts, weighted by state area harvested, of corn relative to that of cotton in 2005 in the USA. For most impact categories, corn and cotton per hectare show roughly similar environmental impacts, with relative magnitude ranging from 1.4 for acidification and 0.9 for human health cancer. For freshwater ecotoxicity, however, corn shows about one third of impact by cotton per hectare, and corn’s water use is less than half that of cotton. Above all,greenhouse pot most of the environmental impacts associated with cotton and corn production are due to on-site environmental emissions rather than that embodied in input materials like fertilizers and pesticides. Their acidification effect is due in large part to application of nitrogen fertilizers and diesel combustion . Although N intensity of corn is much larger than that of cotton , corn farming uses much less diesel . Overall, the acidification impact of corn per hectare is 1.4 times that of cotton. The same can be said about smog formation. Not surprisingly, the two crops’ eutrophication impact is caused mainly by use of N and phosphate fertilizers. Although corn has higher nutrient application intensities than cotton, its average N and P leaching and runoff rates are lower ; thus, the two crops have a comparable eutrophication impact. Water use by cotton and corn comes primarily from irrigation: about 400 m3 is applied per hectare corn produced as opposed to 940 m3 applied per hectare cotton produced. Freshwater ecotoxicity for both crops is due in large part to pesticide use, and cotton per hectare has a freshwater ecotoxicity about three times that of corn. This is partly because pesticide application intensity of cotton is approximately twice as much as that of corn . Also, many of the pesticides such as cyfluthrin, lamb dacyhalothrin, and cypermethrin used in cotton growth generally show higher toxicity-related characterization factors than the major ones used in corn growth. The two crops’ potential human health respiratory impacts are comparable, although that of cotton is slightly higher. The respiratory effect is mainly caused by diesel combustion, application of N fertilizers, and emissions embodied in P fertilizers. Human health cancer and non-cancer impacts of corn per hectare are slightly larger than that of cotton.
Heavy metals contained in phosphate constitute the major contributor to both crops’ non-cancer effect, but use of acephate, an insecticide, is also another important source of non-cancer impact for cotton. This is why corn’s relative magnitude of non-cancer effect is not as large as that of phosphate application intensity . The two crops’ potential human health cancer impact is due to a number of factors including diesel combustion and heavy metals brought about by phosphate as well as the cancer impact embodied in fertilizers. The results above indicate that corn and cotton grown per hectare in the USA on average generate roughly comparable impacts for most of the impact categories except for water use and freshwater ecotoxicity, where cotton shows lower impacts. The results seem consistent with the view of a recent USDA study , “When acreage shifts from one high-input crop to another , however, ethanolinduced changes may be negligible or could even reduce environmental externalities.” We argue that, however, the average results as shown in Fig. 2.1 are inadequate to capture the net environmental impacts associated with land cover change from cotton to corn that took place in the USA. First, Fig. 2.1 is largely a portrait of corn and cotton growth in different regions and, weighted by state crop area, mainly represents the major crop-growing states where respective crops are likely the most suitable to grow. But, when land shifts from cotton to corn growth, it happens in cotton-growing areas in the South. Lands in these areas can be by and large considered marginal lands for corn in both geographic and economic senses as they are generally less suitable for corn growth than the Corn Belt.The range of spatial variation in cotton growth is two to threefold for acidification, smog formation, eutrophication, human health non-cancer, and respiratory effects and four to sixfold for freshwater ecotoxicity and human health cancer effect. The range of spatial variation in corn growth is about two to threefold for acidification, smog formation, human health cancer, non-cancer, and respiratory effects and fourfold for eutrophication. Water use can vary by orders of magnitude for both crops as some states use little irrigation water while some rely heavily on irrigation . In short, the results for average corn and cotton as reflected in Fig. 2.1 fall short of representing the environmental performance of marginal corn in cotton-growing states and, therefore, should not be used for evaluating environmental impacts of land use change from cotton to corn or vice versa. Comparing Fig. 2.2 with Fig. 2.1 reveals that corn and cotton growths in 2005 at the state level can be quite different from the average situation. Land shift from cotton to corn in Georgia and Texas would likely aggravate all of the impact categories except freshwater ecotoxicity. For North Carolina, however, the land shift would increase water withdraw and aggravate eutrophication impact, but would not cause substantial changes to human health effects. For TX, land shift from cotton to corn would especially aggravate acidification and smog formation impacts. This is because TX, as the major producer of cotton in the US, applies far less nutrients per hectare cotton produced than per hectare corn produced there .