A number of social and economic factors drive CO₂ and other GHG emissions associated with agriculture (see Table 5.1), including dietary preferences and traditions; domestic and global commodity markets; federal incentives for conservation programs; and technical capabilities for production, processing, and storage in different geographic regions. For example, policies and economic factors that influence bioenergy and biofuel feedstock production systems have diverse direct and indirect impacts on the carbon cycle as discussed later in this chapter and in Ch. 3: Energy Systems. A biofuel’s carbon footprint depends on the feedstock and its associated management as well as the efficiency of the eventual energy produced from the feedstock. Changes in the management of these social and economic factors can affect soil carbon sequestration and storage and agricultural GHG emissions. Another driver of changes in agricultural production systems is consumer demand for types of food (e.g., meat versus dairy versus vegetable) and provenance of food (e.g., grass-fed, organic, and local). Such influences can have both negative and positive effects on the carbon cycle in direct and indirect ways (see Box. 5.1, Food Waste and Carbon). Decision support tools have been developed over the last decade to address agricultural impacts on climate and environmental drivers that play a role in the carbon cycle (for examples, see Ch.18: Carbon Cycle Science in Support of Decision Making).

Agricultural land carbon storage and loss are the net result of multiple fluxes including plant photosynthetic uptake (i.e., atmospheric CO₂ capture by plants), ecosystem respiratory loss (i.e., carbon released as CO₂ from plants and soil organisms), harvested biomass removal either by grazing or cutting, input from additional feeds, enteric methane (CH₄) production by livestock, and the return of manure by grazing animals or addition of manure or other carbon-rich fertilizer amendments to agricultural lands.

5.3.1 Perennial Systems

The most extensive perennial systems in North America are grasslands, pasture, and hayed lands (see Ch. 10: Grasslands). Other perennial crops (i.e., crops growing and harvested over multiple years) of regional importance include tree crops (mostly fruit and nuts) and vineyards. Because many perennial fruit, nut, and vegetable systems generally are intensively managed, the type of management—such as cover crops and intercropping, irrigation and tillage, fertilizer use, and intensity of cultural activities—largely determines the carbon balance of these production systems. Additionally, biofuel feedstock crops, including perennial grasses and short-rotation woody crops, occupy a very small percentage of agricultural land area, but they have the potential to either sequester carbon or create a carbon debt, depending on the system and land use that the system replaced (e.g., Adler et al., 2007, 2012; Mladenoff et al., 2016). Although differences in net carbon and GHG balance do exist, perennial bioenergy crops generally increase soil carbon in lands converted from annual crops because belowground carbon allocation (to roots) increases once the crops are established, even though the biomass is harvested for energy (Anderson-Teixeira et al., 2013; Valdez et al., 2017). However, managing perennials as biofuel crops often requires additional nitrogenous fertilizer, which can increase nitrous oxide (N₂O) emissions and reduce the associated mitigation potential (Johnson and Barbour 2016; see Ch. 3: Energy Systems).

Perennial systems avoid the 4- to 8-month fallow period common among many annual row-crop systems (Drinkwater and Snapp 2007); therefore, perennial plants can use the sun’s energy to drive photosynthesis outside the typical growing season (Baker and Griffis 2005), contributing to increased soil carbon sequestration as compared to annual systems (Sainju et al., 2014). In agricultural systems dominated by perennial plants, photosynthesis generally, but not always, exceeds ecosystem respiration, so on balance these ecosystems remove more CO₂ from the atmosphere than they contribute each year (Gilmanov et al., 2010). The total net amount of CO₂ exchanged between perennial systems and the atmosphere varies among regions, with net carbon loss occurring most often in drought-prone and desert systems (Liebig et al., 2012). In grazed ecosystems, better management practices, such as prescribed grazing, adaptive multipaddock grazing, improved grass species and introduction of legumes, fertilization, and irrigation, generally will increase soil carbon sequestration (Conant et al., 2001; Teague et al., 2013). Estimates of the potential for U.S. pasture and hayed lands to sequester carbon (with improved management) vary, ranging from near 0 to 3 or more megagrams of carbon (Mg C) per hectare per year, with reasonable mean values of up to about 0.5 Mg C per hectare per year (Conant et al., 2001).

When productivity increases in agricultural systems, land managers frequently remove more aboveground biomass. In some cases, this increase in carbon removal by harvesting offsets the amount of carbon that would otherwise be sequestered, but the main driver of soil carbon sequestration is the production of belowground biomass that is not removed from the field. As a result, increased forage productivity often is associated with increased soil carbon sequestration (Allard et al., 2007; Ammann et al., 2007; Cong et al., 2014; Skinner and Dell 2016) because increased aboveground biomass normally is associated with increased belowground biomass. Initial conditions and ecosystem characteristics influence carbon sequestration potential. Depleted soils likely will accumulate additional carbon, whereas soils in which carbon inputs and outputs are roughly equal will show no change or perhaps a net loss of carbon over time (Smith 2004). Grazed pastures typically sequester more soil carbon than hayed land (Franzluebbers and Stuedemann 2009; Franzluebbers et al., 2000; Senapati et al., 2014) because cutting can cause a greater initial reduction and slower recovery in photosynthetic uptake of carbon than grazing (Skinner and Goslee 2016). Perennial root systems also become active early and remain active late in the growing season and thus can take up and use reactive nitrogen before it is lost from the system. The capture and efficient use of nitrogen (e.g., nitrate and ammonia applied at the correct time and rates) can avoid nitrogen losses. As a result, N₂O emissions for perennial systems are typically much lower than those for annual systems (Ma et al., 2000; Qin et al., 2004; Robertson and Vitousek 2009).

5.3.2 Annual Systems

As with perennial systems, carbon storage or loss in annually cropped lands is the net result of inputs from unharvested plant residue (especially below ground); root exudation and turnover; organic matter deposition; soil amendments such as manure; and losses from respiration, residue, leaching, soil organic matter mineralization (decomposition), and harvested biomass removal. In turn, these input and output pathways respond to previous and current land use, soil properties (e.g., soil type and depth), climate, and other environmental factors. Typically, annual cropping systems are managed intensively; as such, their associated carbon stocks are closely related to land management choices (e.g., tillage, crop and crop rotation, residue management, fertilizer and nutrient inputs, extent and efficiency of drainage, and irrigation and use of cover crops) and the duration of those practices.

Studies to date suggest that annually cropped mineral soils in the United States sequester a small amount of carbon, but carbon emissions from cropped organic soils and a number of other farm management practices largely offset this benefit (Del Grosso and Baranski 2016; U.S. EPA 2016; see Figure 5.2). Cropped organic soils (e.g., Histosols) comprise only a small portion (<1%) of overall U.S. cropland, but these organic soils can be a large source of atmospheric carbon on a per area basis. This carbon loss occurs because cropped organic soils commonly result from draining wetlands, which greatly enhances decomposition rates in these high-carbon soils that, historically, have been under water and relatively safe from decomposition. Reversion of these drained and cropped organic soils to wetlands or flooded rice production slows the soil carbon losses but also can result in increased CH₄ and N₂O emissions, implying that water management can play a key role in the net carbon and GHG balances (Bird et al., 2003; Deverel et al., 2016; Oikawa et al., 2017). However, N₂O does not necessarily increase with land-use conversion to paddy rice because there is evidence of N₂O uptake by recently converted upland crops to flooded rice (Ye and Horwath 2016). Other practices that tend to lead to carbon loss include leaving land fallow without vegetation, growing low-residue crops (e.g., cotton), and plowing intensively (USDA 2014). Conversely, several practices may increase soil carbon stocks, such as including hay and grass in annual crop rotations, growing cover crops, maintaining plant cover, reducing the fallow (vegetation-free) period by increasing cropping intensity especially on marginal land as encouraged by CRP, and possibly reducing tillage intensity (USDA 2014). This increase in soil carbon stocks can vary by ecosystem but is particularly prevalent where these practices are used on soils previously depleted of their original carbon stores.

Compared to perennial crops, annual crop systems tend to have higher nitrogen losses, including N₂O emissions. In addition, nitrogen fertilizer additions generally lead to increased CH₄ emissions and decreased CH₄ oxidation from soils, particularly under anoxic conditions or flooded soil systems such as rice (Liu and Greaver 2009).

5.3.3 Livestock Systems

The North American livestock sector currently represents a significant source of GHG emissions, generating CO₂, CH₄, and N₂O throughout the production process. Livestock contributions to GHG emissions occur either directly (e.g., from enteric fermentation and manure management) or indirectly (e.g., from feed-production activities and conversion of forest into pasture or feed crops).

Enteric Fermentation

Methane and CO₂ are natural end-products of microbial fermentation of carbohydrates and, to a lesser extent, amino acids in the rumen of ruminant animals and the hindgut of all farm animals. Methane is produced in strictly anaerobic conditions by highly specialized methanogenic microbes. In ruminants, the vast majority of enteric CH₄ production occurs in the rumen (i.e., the largest compartment of the ruminants’ complex stomach); rectal emissions account for about 3% of total enteric CH₄ emissions (Hristov et al., 2013b). Methanogenic microbes inhabit the digestive system of many monogastric and nonruminant herbivore animals (Jensen 1996). In these species, CH₄ is formed by processes like those occurring in the rumen and is similarly increased by intake of fibrous feeds. Summarizing published data, Jensen (1996) estimated that a 100-kg pig produces about 4.3% of the daily CH₄ emissions of a 500-kg cow. Nonruminant herbivore animals such as horses consume primarily fibrous feeds and emit greater amounts of CH₄ than nonruminant species that consume primarily nonfibrous diets, but a horse’s CH₄ production per unit of body weight is still significantly less than that of ruminants. Wild animals, specifically ruminants (e.g., bison, elk, and deer), also emit CH₄ from enteric fermentation in their complex stomachs or the lower gut. The current contribution of wild ruminants to global GHG emissions is relatively low (Hristov 2012).

The U.S. Environmental Protection Agency (EPA) reports that CH₄ emissions from enteric fermentation and manure management amounted to about 232.8 teragrams (Tg) per year CO₂e (functionally equivalent to 63.5 Tg C) in 2015, with an additional 17.7 Tg per year CO₂e (4.8 Tg C) as N₂O emitted from manure management (U.S. EPA 2018). Combined, these emissions represented 3.8% of total U.S. GHG emissions. About 97% of the enteric fermentation and 57% of the CH₄ emissions from manure management were from beef and dairy cattle; 78% of the N₂O emissions from manure management also were attributed to beef and dairy cattle. These estimates are derived from a “bottom-up” approach that begins with estimates of emissions on a per-animal basis and multiplies those estimates over total relevant numbers of animals. “Top-down” approaches, based on measurements of changes in GHG concentrations over large areas and inferences about the sources of those changes, yield different estimates for CH₄ emissions. Combining satellite data and modeling, several studies proposed that livestock emissions may range from 40% to 90% greater than EPA estimates (Miller et al., 2013; Wecht et al., 2014). There is more uncertainty in predicting CH₄ emissions from manure, partially because these emissions depend heavily on the particular manure handling system and temperature. The sources of discrepancy between the top-down and bottom-up approaches need to be identified to derive accurate estimates for both total and livestock CH₄ emissions in North America (NASEM 2018).

There is no disagreement, however, that cattle are a significant source of CH₄ emissions. Based on U.S. EPA (2018) estimates, CH₄ emissions from cattle make up 25.9% of total U.S. CH₄ emissions if only enteric emissions are counted, or 36.2% if emissions from manure management are included. In a national life cycle assessment of fluid milk, 72% of GHG emissions associated with milk production occurred on the farm, with 25% being from enteric CH₄ fermentation. The remaining 28% was associated with processing, packaging, distribution, retail, and consumers (Thoma et al., 2013). A similar life cycle assessment of beef indicates that 87% of GHG emissions associated with beef are from cattle production, with only 13% resulting from post-farm processes (Asem-Hiablie et al., 2018). Similar to ruminants, animal production is the main contributor of GHG emissions in the swine industry. A life cycle assessment of the U.S. pork industry (Thoma et al., 2011) reported the following breakdown of emission contributions for each stage of the production cycle: 9.6%, sow barn (including feed and manure management); 52.5%, nursery-to-finish (including feed and manure handling); 6.9%, processing (including 5.6% for processing and 1.3% for packaging); 7.5%, retail (e.g., electricity and refrigerants); and 23.5%, the consumer (e.g., refrigeration, cooking, and CH₄ from food waste in landfills). Major sources of GHG emissions in the poultry industry differ depending on the type of production. For broilers (i.e., meat-producing birds), feed production contributes 78% of the emissions; direct on-farm energy use, 8%; post-farm processing and transport of meat, 7%; and manure storage and processing, 6%. For layers (i.e., egg-producing birds), feed production contributes 69% of emissions; direct on-farm energy use, 4%; post-farm processing and transport, 6%; and manure storage and processing, 20% (MacLeod et al., 2013).

Manure Management

Manure can be a major source of GHG emissions, depending on the type of livestock. For ruminants, manure emissions normally are less than those from enteric production, but for nonruminants, manure is the major source of GHG emissions. Microbial activity breaks down organic carbon in manure, releasing both CH₄ and CO₂, and the amount of each produced is related to oxygen availability. Much of the carbon in manure eventually ends up in the atmosphere in one of these two forms, and because CH₄ is a more powerful GHG than CO₂, converting this biogenic carbon to CO₂ would be beneficial.

Methane emissions from all manure produced and handled in the United States were estimated to be 66.3 Tg CO₂e in 2015 (U.S. EPA 2018). These emissions occur in the housing facility, during long-term storage, and during field application (see Table 5.2). The housing facility usually is a relatively small source. Manure lying on a barn floor or open-lot surface is exposed to aerobic conditions where CH₄ emissions are low (IPCC 2006; USDA-ARS 2016). Manure deposited by grazing animals also is exposed to aerobic conditions, with CH₄ emissions similar to those from a barn floor or open lot. When manure in the housing facility is allowed to accumulate in a bedded pack up to a meter deep, anaerobic conditions develop, leading to greater CH₄ emissions (IPCC 2006).

Long-term storage normally is the major source of carbon emissions from manure (see Table 5.2). Liquid or slurry manure typically is stored for 4 to 6 months prior to cropland application. During storage, anaerobic conditions are maintained in which CH₄ formation and emission rates are largely controlled by manure temperature (IPCC 2006; USDA-ARS 2016). Longer storage periods will produce greater emissions. Manure solids can float to the surface, particularly in slurry manure, where a crust is formed. This natural crust can reduce storage CH₄ emissions by 30% to 40% (IPCC 2006; USDA-ARS 2016). Solid manure may be stored up to several months in a stack with or without active composting. This type of storage maintains more aerobic conditions, which reduce CH₄ emissions.

Following storage, manure typically is applied to cropland as a nutrient source for plant growth. During unloading from storage and field application, any CH₄ remaining in the manure is released. These emissions are small compared to those from other sources. Following application of the manure spread onto the soil in a thin layer, aerobic conditions suppress further CH₄ production. Manure also may be incorporated into the soil so that any CH₄ produced is oxidized and consumed (Le Mer and Roger 2001). Thus, optimizing the timing, quantity, and incorporation of manure applications with plant productivity and growth patterns and needs can reduce the associated CH₄ and N₂O emissions.

5.4.1 Trends in Acres Cultivated, Soil Carbon, and Overall Emissions

The First State of the Carbon Cycle Report (CCSP 2007) showed total agricultural and grazing lands in North America (e.g., cropland, pasture, rangeland, shrub lands, and arid lands) accounting for 17% of global terrestrial carbon stocks. Most of this carbon pool existed within soils; less than 5% resided in cropland vegetation. More recent data estimate that the annual U.S. soil carbon sequestration rate decreased between 1990 and 2013, primarily due to changes in land use and variability in weather patterns. Worth noting are the large interannual fluctuations in the size of the mineral soil CO₂ sink (USDA 2016). The major non-CO₂ emissions from U.S. agriculture in 2013 were N₂O from cropped and grazed soils (44% of U.S N₂O emissions) and enteric CH₄ from livestock (28% of U.S. CH₄ emissions). In 2015, the major non-CO₂ emissions from U.S. agriculture were N₂O from agricultural soil management (52% of all agricultural emissions, or 4.4% of all U.S. GHG emissions) and enteric CH₄ from livestock (29% of agricultural emissions, or 2.5% of all U.S. GHG emissions). Combined with forestry, the agricultural sector contributed to a total net carbon sequestration of 270 Tg CO₂e in 2013 (USDA 2016), while total agricultural GHG emissions (excluding land use, land-use change, and forestry activities) amounted to 567 Tg CO₂e in 2015 (U.S. EPA 2018).

Agricultural GHG emissions in North America were 706 Tg CO₂e in 2014 and 2015 (Table 5.1), including 567 Tg CO₂e in the United States (excluding emissions from land use, land-use change, and forestry; U.S. EPA 2018), 59.0 Tg CO₂e in Canada, and 79.9 Tg CO₂e in Mexico (Table 5.1). Agricultural non-CO₂ emissions were primarily N₂O from cropped and grazed soils and CH₄ from enteric fermentation in livestock. In 2014 and 2015, North America’s major sources and annual rates of GHG emissions (in CO₂e) included: agricultural soil management (318.0 Tg), enteric fermentation (234.8 Tg), manure management (117.7 Tg), and rice cultivation (12.5 Tg; Table 5.1). Trends that drive North American GHG emissions from agriculture include changes in five areas: 1) the amount of nitrogen fertilizer applied, which correlates with land area planted in corn, cotton, and wheat (USDA 2016); 2) the number of ruminants, especially beef cattle and dairy cows because they produce large quantities of enteric and manure CH₄; 3) trends in human diet choices, which drive changes in land use, numbers of livestock, and volumes of inputs like fertilizer; 4) area of agricultural land opened by clearing forest, which converts large amounts of carbon in plants and soils to CO₂; and 5) the amount of food wasted, which leads to CH₄ emissions from landfills and also drives additional production with associated GHG emissions (e.g., Hall et al., 2009). Overall, actively managed agricultural lands have a strong capacity to reduce GHG emissions to the atmosphere and take up and store carbon. Varying management options thus could lead to substantial reductions in emitted CO₂ and CH₄ and sequester significant amounts of carbon.

According to the U.S. 2012 Agricultural Census, 370 million ha were classified as farmland (see Table 5.3). Such lands declined by 3.1 million ha between 2007 and 2012 (USDA-NASS 2012). Out of the converted croplands, 18% changed to nonagricultural uses (e.g., urban growth and transportation); another 3% reverted to forest; and the remaining 79% were used for other types of agricultural land, primarily pastures (USDA-NRCS 2015). The conversion of farmland to other uses appears to have slowed compared with the period from 2002 to 2007, when greater than 9.6 million ha of farmland were converted to other uses (USDA-NASS 2012). In 2012, 19% of the total 786.8 million ha in the contiguous 48 states, Hawaiʻi, Puerto Rico, and the U.S. Virgin Islands was classified as cropland, 1% as CRP, 6% as pastureland, and 21% as rangeland (USDA-NRCS 2015).

Similar to these trends in North America, global GHG emissions from large ruminants, such as beef and dairy cattle, are about seven times greater than emissions from swine or poultry (Gerber et al., 2012). Dairy production systems, however, are considerably more efficient than beef systems. As an example, Eshel et al. (2014) estimated, using a full life cycle assessment, that GHG emissions per ­human-edible megacalorie (MCal) were 9.6 kg CO₂e for beef versus 2 for pork, 1.71 for poultry, and 1.85 for dairy. Similarly, GHG emissions per kg of human-edible protein were 214 kg CO₂e for beef, 42 for pork, 20 for poultry, and 32 for dairy (Eshel et al., 2014).

U.S. cattle inventories have fluctuated during the last several decades from a peak of over 130 million heads (both beef and dairy) in the 1970s to a low of 88.5 million in 2014. Cattle numbers increased to 89 million in 2015 and an estimated 92 million in 2016 (USDA-NASS 2016). According to the 2016 inventory, there were 30.3 million beef cows, 9.3 million dairy cows, 19.8 million heifers weighing 227 kg or more, 16.3 million steers at 227 kg or more, 14 million calves under 227 kg, and 2.1 million bulls. Beef and dairy cows, because of their high feed consumption and higher-fiber diets, are the largest emitters of enteric CH₄, producing about 95 and 146 kg CH₄ per head per year, respectively; emissions from feedlot cattle fed high-grain diets are considerably less at 43 kg per year per head (U.S. EPA 2018). Increased cattle productivity has resulted in increased feed efficiency and decreased enteric CH₄ emission intensity (i.e., CH₄ emitted per unit of milk or meat). As an example, the estimated CH₄ emission intensity for the U.S. dairy herd has decreased from 31 g per kg milk in 1924 to 14 g per kg in 2015 (Global Research Alliance on Agricultural Greenhouse Gases 2015).

Cattle inventories in Canada have fluctuated annually, but long-term trends are relatively stable—about 12 million heads in January 2016, down slightly from a peak in 2005 (Statistics Canada 2016). Beef cattle account for more than 80% of these animals. In recent decades, improvements in management efficiency have led to a decline in GHG emissions per unit of livestock product. For example, estimated emissions per kilogram of liveweight beef leaving the farm declined from 14 kg CO₂e in 1981 to 12 kg CO₂e in 2011 (Legesse et al., 2016).

U.S. beef consumption has been declining steadily over the past decade (see Figure 5.3) while consumption of dairy products has been increasing (see Figure 5.4). The previously mentioned life cycle assessment analyses that found greater carbon efficiency of dairy versus beef suggest that this trend should translate to lower emissions from the livestock sector. Most of the beef and veal consumed in the United States was domestically produced (about 86% in 2015; 18.6% of imported beef was from Canada), while about 9.6% of beef produced in the United States in 2015 was exported to other countries. Fluid milk consumption per capita has been decreasing—from about 89 kg per year in 2000 to 71 kg per year in 2015, while consumption of cheese, butter, and yogurt, most of which is domestically produced, has been steadily increasing. As in the United States, per capita consumption of livestock products in Canada also has declined in recent decades. For example, beef and fluid milk consumption decreased from 39 kg of beef per capita in 1980 to 24 kg in 2015 (Agriculture and Agri-Food Canada 2016) and from 90 liters of fluid milk per capita in 1996 to 71 liters in 2015 (Government of Canada 2016).

The strong influence of these carbon-intensive food consumption patterns on the global carbon cycle highlights the challenge of assigning emissions to a particular country. As mentioned previously, 2.5% of beef consumed in the United States is imported from Canada. Most inventories assign these emissions to the country where production occurs, but a main lever that could influence GHG emissions associated with this production rests, in this case, with the United States, because demand is a strong driver of supply and production.

5.4.2 Climate Change Effects and Feedbacks on Carbon

Climate change, including changes in temperature, precipitation, and the frequency of extreme events, could alter the productivity of agricultural systems through its effects on plant and animal growth as well as carbon sequestration and storage by influencing soil respiration and plant allocation to soil carbon. Climate change also could have an indirect effect on enteric CH₄ emissions (i.e., from ruminant animals) and directly influence manure and soil-derived CH₄ emissions through temperature increases. The effect on enteric emissions is through increased or decreased feed (i.e., dry matter) intake; projected increased ambient temperatures can decrease dry matter intake and thus proportionally reduce enteric CH₄ emissions. As an example, the average maximum temperature for the northeastern United States is projected to increase 6.5°C by 2100 (projected by Representative Concentration Pathway 8.5, a high-emissions scenario). This temperature increase is expected to decrease dry matter intake of dairy cows in the region by an additional 0.9 kg per day due to heat stress (Hristov et al., 2017a). This decreased intake will amount to a reduction in daily enteric CH₄ emissions of about 17 g per cow. If this reduction is extrapolated over 365 days and 1.4 million cows in the northeastern United States, the increased temperature will lead to a decrease in enteric CH₄ emissions from dairy cows of about 8.7 metric tons per year, but the net effect on CO₂e per kg of product depends on the effect of temperature on productivity. In contrast, increased temperatures are expected to increase manure CH₄ emissions. The microbial decomposition of manure, producing CH₄, is sensitive to temperature, so the projected climate changes suggest an increase in emissions of about 4% by midcentury and 8% by 2100 (Rotz et al., 2016).

Climate change effects on soil carbon sequestration will involve a balancing act between the impacts of elevated CO₂, higher temperatures, and either increasing or decreasing precipitation depending on the region under consideration. Elevated CO₂ and increased precipitation are expected to increase carbon inputs into systems and increase their potential to sequester carbon, whereas higher temperatures are expected to increase ecosystem respiration. Also, yields of major crops (corn, soybeans, wheat, and rice) are predicted to decline as global temperature increases (Zhao et al., 2017). Reduced precipitation or soil moisture along with the drying effects of warming would be expected to decrease plant production and carbon inputs in most upland systems. In unmanaged ecosystems, limited nitrogen availability could constrain the positive effects of elevated CO₂ on plant growth (Norby et al., 2010; Thornton et al., 2007), although in managed pasture and hayland systems, fertilization would be expected to overcome such constraints. Tubiello et al. (2007) suggested that the balance between competing pressures would result in greater crop yields in temperate regions compared with those in semiarid and tropical regions. However, several analyses suggest that increased atmospheric CO₂ will increase soil CO₂ respiration by almost as much as the stimulation of inputs, resulting in little net change in soil carbon pools (Dieleman et al., 2012; Todd-Brown et al., 2014; van Groenigen et al., 2014). Because the potential effects of climate on soil carbon sequestration could be relatively small in most North American agricultural systems, at least compared with the large changes expected in the Arctic (Todd-Brown et al., 2014; see Ch. 11: Arctic and Boreal Carbon), management is projected to have a greater effect on carbon sequestration than will changes in climate (Álvaro-Fuentes and Paustian 2011; Lugato and Berti 2008).

The 2018 EPA inventory (U.S. EPA 2018) attributed 567 Tg CO₂e to the agricultural sector for 2015 (excluding emissions related to land use, land-use change, and forestry activities), accounting for 8.5% of total U.S. emissions.¹ This proportion reflects a small increase since 1990, primarily due to increased CH₄ emissions from manure management. Nitrous oxide emissions from agricultural soil management were the largest sources of GHGs at 295 Tg CO₂e, and these emissions, largely due to synthetic nitrogen fertilizer applications, accounted for 77.7% of all U.S. N₂O emissions. Other sources primarily included enteric fermentation (166.5 Tg CO₂e), manure management (66.3 Tg CO₂e and 17.7 Tg CO₂e as CH₄ and N₂O, respectively), rice cultivation (12.3 Tg CO₂e), field burning (0.4 Tg CO₂e), and CO₂ emissions from urea fertilization and liming (4.9 and 3.8 Tg CO₂e, respectively). Within the enteric fermentation emissions, beef cattle accounted for 70.9% and dairy cattle 25.6%. Worth noting is that these numbers have been relatively stable since 1990 even though production of beef and dairy products has increased. Agricultural croplands remaining as cropland in the United States (i.e., not converted to or from other land uses) represent a small sink sequestering an estimated 0.1% of the CO_2e removed from the atmosphere by land use, land-use change, and forestry activities (U.S. EPA 2018). As noted previously, agricultural practices that remove CO₂ from the atmosphere include conversion from cropland to permanent pastures or hay production, reduction in acreage managed with summer fallow, adoption of conservation tillage practices, and increased applications of manure or sewage sludge. Overall, SOC increases in croplands remaining cropland and croplands converted to grasslands collectively offset losses caused by recent conversions of long-term grassland to cropland (U.S. EPA 2015, 2016, 2018; see also Ch. 12: Soils, Section 12.5.1).

In Canada, agricultural soils (55.2 million ha) contain about 4.1 petagrams (Pg) C (0- to 30-cm soil depth) and 5.5 Pg C (0- to 100-cm soil depth), as calculated from the Canadian Soil Information Service National Soil Database and reported in Ch. 12: Soils. As of 2013, Canadian agricultural land removed 11 Tg CO₂ per year, which would counter about 2% of the total Canadian national GHG emissions (ECCC 2018). The reduction was attributed to decreased summer fallow and increased adoption of no-till practices in Canadian prairies. However, this value is starting to decline (e.g., down from 13 Tg CO₂ in 2005) because changes in SOC stocks and fluxes tend to approach equilibrium at some point after a change in conditions.

5.5.1 Impact of Management Practices

Croplands

Most cropland carbon stocks are in the soil and reflect management history and practices that increase or decrease soil carbon stocks. Integration of practices that can increase soil carbon stocks include 1) maintaining land cover with vegetation (e.g., use of deep-rooted perennials, elimination of summer fallow, and inclusion of cover crops in annual systems); 2) protecting the soil from erosion (e.g., reduced or no tillage and residue cover); and 3) improving nutrient management (Srinivasarao et al., 2015; Swan et al., 2015). The magnitude and longevity of carbon stock changes have strong environmental and regional differences that are subject to subsequent changes in management practices. Conversely, practices that convert lands from perennial systems, such as converting retired or other lands to row crops, consistently show release of stored carbon back to the atmosphere (Gelfand et al., 2011; Huang et al., 2002). Other management practices with the potential to release stored carbon are inadequate return of crop residues (e.g., Blanco-Canqui and Lal 2009), aggressive tillage (Conant et al., 2007), over application of nitrogen fertilizer, and burning of crop residue (Robertson and Grace 2004; Wang et al., 2011).

The timescale for carbon storage in soils is a critical factor for GHG mitigation. Numerous estimates of the rates and potential magnitude of long-term soil carbon accumulation, storage, and sequestration related to management have been reviewed and presented (e.g., Minasny et al., 2017; Paustian et al., 2016; Sperow 2016; Stockmann et al., 2013; Swan et al., 2015). Management practices that increase carbon inputs include planting high-residue crops and returning crop biomass to the soil; minimizing or eliminating summer fallow (particularly bare fallow); adding cover crops to reduce winter fallow; extending and intensifying cropping rotations (e.g., double-cropping or relay cropping and adding forage perennials); retiring marginal lands to perennials; and adding perennials in buffer strips, field borders, filter strips, grassed waterways, vegetative barriers, and herbaceous wind barriers (e.g., Mosier et al., 2006; Paustian et al., 2016; Sainju et al., 2010; Sperow 2016). Swan et al. (2015) estimated carbon storage rates of 0.42 to 0.95 Mg C per hectare per year among conservation practices that shift to perennials (e.g., retiring marginal land or planting perennials as barriers or borders), while inclusion of cover crops was estimated to accrue 0.15 to 0.27 Mg C per hectare per year. Practices that eliminate summer fallow can increase SOC directly by increasing carbon input or modifying microclimate (i.e., temperature and water), a practice that can decrease mineralization rates by reducing temperature and water content (Halvorson et al., 2002; Sainju et al., 2015).

Numerous publications have reported that ­no-tillage practices store more carbon in soil than those using conventional tillage (e.g., Paustian et al., 2016; Sperow 2016; West and Post 2002). Conversely, others have disputed this claim, especially when including soil carbon measurements deeper than 30 cm (e.g., Baker et al., 2007; Luo et al., 2010; Powlson et al., 2014; Ugarte et al., 2014). No-tillage and other conservation practices were developed to control soil erosion, and this co-benefit is well established. Erosion removes soil carbon from farm fields and relocates that carbon to other parts of the landscape; the amount of this transported carbon that is sequestered in sediments compared to the amount converted to CO₂ or CH₄ is difficult to estimate (Doetterl et al., 2016). In Ch. 12: Soils, the role of soil erosion is discussed in greater detail and suggests that burial of eroded carbon constitutes a small sink. Comparing SOC sequestration rates from a system managed without tillage to a system with tillage results in negative, neutral, and positive rates of SOC sequestration: 1) 27 ± 19 Mg SOC per hectare per year, (n = 49; Liebig et al., 2005), 2) 0.40 ± 61 Mg SOC per hectare per year (n = 44; Johnson et al., 2005), or 0.45 ± 0.04 Mg SOC per hectare per year (n = 147; Franzluebbers 2010). Likewise, studies using eddy covariance techniques report divergent responses to tillage. For example, Bernacchi et al. (2005) demonstrated that no-tillage agriculture on clay-rich soil built SOC, whereas others (Baker and Griffis 2005; Chi et al., 2016; Verma et al., 2005) used gas exchange techniques to suggest conservation or no-tillage systems were near carbon neutral. In another review, Collins et al. (2012) found that carbon sequestration rates varied from no measurable increase (Staben et al., 1997) to 4 Mg C per hectare per year (Lee et al., 2007), varying with depth monitored, study duration, fertilizer formulation, and location. Several rationales have been postulated for this variability. If sampling depth is shallower than the tillage depth, the apparent change in SOC may be an artifact of sampling depth (Baker et al., 2007) or caused by residue redistribution (Staricka et al., 1991) and vertical stratification of soil carbon (Luo et al., 2010). Meta-analyses by Luo et al. (2010) and Ugarte et al. (2014) suggest that other factors contributing to variability in SOC sequestration include climatic and soil properties interacting with management factors (e.g., cropping frequency, crop rotation diversity, nitrogen, and drainage) along with impacts on rooting depth and above- and belowground biomass, as well as soil heterogeneity and the long time frames required to find a definitive increase or decrease in SOC. Collectively, the evidence indicates that adoption of no tillage may store more carbon, especially in the soil surface, compared to storage with conventional tillage. However, conclusively measuring short-term changes is difficult because of soil heterogeneity and slow rates of change (also discussed in Ch. 12: Soils). In particular, increased N₂O or CH₄ emissions have been shown to occur for as many as 10 years after no-till adoption (Six et al., 2004), though this effect is greater and more consistent in medium to poorly drained soils (Rochette 2008). Thus, quantifying GHG mitigation by management also must account for changes in N₂O and CH₄, which can occur coincidently with changes in soil carbon storage (VandenBygaart 2016).

From a carbon emissions perspective, biofuels have received a great deal of attention because of their potential to produce a more carbon neutral liquid fuel relative to fossil fuels. Biofuels from annual crops currently supply about 5% of U.S. energy use, mostly from corn grain ethanol (~36% of the corn grain harvest) and soy biodiesel (~25 % of the soybean harvest; USDA 2017b). Although the potential for reduced GHG emissions with biofuels is compelling, some life cycle assessment analyses suggest that corn grain ethanol has marginally lower (or even greater) GHG emissions compared with those from fossil fuels (e.g., Del Grosso et al., 2014; Fargione et al., 2008). However, more recent studies suggest that currently available technologies can achieve greater GHG reductions of 27% to 43% compared to gasoline when assessed on an energy equivalent basis (Canter et al., 2015; Flugge et al., 2017). Reasons for reduced net GHG intensity for grain- and oil-based biofuels include improved crop-management practices and diminished emissions from land-use change because most of the yield gap from diverting food and feed crops to biofuel feedstocks has been met by increasing per-unit area yields, taking into account the benefits of co-products (e.g., using dried distiller grains for livestock feed) and implementing more efficient feedstock conversion technologies (Flugge et al., 2017). Typically, cellulosic biomass conversion technologies are considered too expensive to compete with liquid fuels derived from other sources (Winchester and Reilly 2015), but innovations at all levels are advancing conversion technology. The impact of cellulosic biofuels on the carbon cycle (Fulton et al., 2015) will depend on ensuring that appropriate mitigation strategies are followed during feedstock choice (perennial or annual) and cultivation (e.g., related to soil carbon stock changes [Blanco-Canqui 2013; Johnson et al., 2012, 2014; Qin et al., 2015]), transportation, and conversion to biofuels (U.S. DOE 2016).

Co-Benefits of Conservation Management

Many common conservation practices improve soil aeration, aggregate stability, and nutrient reserves, while modulating temperature and water and increasing microbial activity and diversity. As a result, soil under some conservation-management regimes can be more resilient to climate variability and more productive (Lal 2015; Lehman et al., 2015). For example, adoption of practices that can conserve soil carbon (e.g., perennial crops, cover crops, and no tillage) may reverse the effects of tillage-intense systems associated with environmental and soil degradation (Mazzoncini et al., 2011). Plant material maintained on the soil surface improves soil physical properties (e.g., Johnson et al., 2016), nutrient availability, and microbial biomass and activity (Feng et al., 2003; Weyers et al., 2013). These improvements result in enhanced soil and water quality and soil productivity (Franzluebbers 2008). Cover crops improve soil health by increasing microbial diversity, biomass, and activity (Bronick and Lal 2005; Lehman et al., 2012, 2015; Schutter and Dick 2002); they also improve soil aggregation, water retention, and nutrient cycling (Blanco-Canqui et al., 2013; Drinkwater et al., 1998; Kladivko et al., 2014; Liebig et al., 2005; Sainju et al., 2006). Thus, there are management practices that simultaneously benefit a number of soil health and carbon storage attributes.

5.5.2 Emissions Reduction

Livestock

Enteric fermentation and manure management represent 44% of the 2015 agricultural GHG emissions in the United States (U.S. EPA 2018) and 36% and 58% of the agricultural emissions in Canada and Mexico, respectively (FAOSTAT 2017). Of the total U.S. GHG emissions in 2015, however, emissions from enteric fermentation and manure management made up only 3.8% (U.S. EPA 2018). Methane mitigation practices for livestock include practices related to reducing emissions from enteric fermentation (i.e., cattle) and manure management (i.e., cattle and swine) as discussed by Hristov et al. (2013b) and Herrero et al. (2016). Increasing forage digestibility and digestible forage intake generally will reduce CH₄ emissions from rumen fermentation (and stored manure) when scaled per unit of animal product. Enteric CH₄ emissions may be reduced when corn silage replaces grass silage in the diet. Legume silages also may have an advantage over grass silage because of their lower fiber content and the additional benefit of reducing or replacing inorganic nitrogen fertilizer use. Dietary lipids are effective in reducing enteric CH₄ emissions, but the applicability of this practice will depend on its cost and effects on feed intake, production, and milk composition in dairy cows. Inclusion of concentrate feeds in the diet of ruminants likely will decrease enteric CH₄ emissions per unit of animal product, particularly when the inclusion is above 40% of dry matter intake.

A number of feed additives, such as nitrates, also can effectively decrease enteric CH₄ emissions in ruminants. Because these additives can be toxic to the animals, proper adaptation is critical. However, nitrates may slightly increase N₂O emissions, which decreases their overall mitigating effect by 10% to 15% (Petersen et al., 2015). Through their effect on feed efficiency, ionophores are likely to have a moderate CH₄-mitigating effect in ruminants fed high-grain or grain-forage diets. Some direct-fed microbial products, such as live yeast or yeast culture, might have a moderate CH₄-mitigating effect by increasing animal productivity and feed efficiency, but the effect is expected to be inconsistent. Vaccines against rumen methanogens may offer mitigation opportunities in the future, but the extent of CH₄ reduction appears small, and adaptation and persistence of the effect are unknown. A recently discovered enteric CH₄ inhibitor, 3-nitrooxypropanol, has shown promising results with both beef and dairy cattle. Under ­industry-relevant conditions, the inhibitor persistently decreased enteric CH₄ emissions by 30% in dairy cows, without negatively affecting animal productivity (Hristov et al., 2015). Similar or even greater mitigation potential has been reported for beef cattle (Romero-Perez et al., 2015). If its effectiveness is proven in long-term studies, this mitigation practice could lead to a substantial reduction of enteric CH₄ emissions from the ruminant livestock sector.

Animal management also can have an impact on the intensity (i.e., emissions per unit of animal product) of CH₄ emissions from livestock systems. For example, increasing animal productivity through genetic selection for feed efficiency can be an effective strategy for reducing CH₄ emission intensity. Other management practices for significantly decreasing total GHG emissions in beef and other meat production systems include reducing age at slaughter of finished cattle and the number of days that animals consume feed in the feedlot. Improved animal health, reduced mortality and morbidity, and improved reproductive performance also can increase herd productivity and reduce GHG emission intensity in livestock production (Hristov et al., 2013a).

Several practices are known to reduce CH₄ emissions from manure but cannot be considered in isolation of other GHG sources and pollutants such as N₂O and ammonia (NH₃). Practices such as the use of solid manure storage and composting can reduce CH₄ emissions, but N₂O and NH₃ emissions will increase, and the end result may not be a reduction in overall GHG emissions. Mitigation of carbon emissions also may have tradeoffs with other pollutants including other gaseous emissions, nutrient leaching to groundwater, and nutrient runoff to surface waters. For example, eliminating long-term manure storage can greatly reduce CH₄ emissions, but daily spreading of manure throughout the year can cause greater nutrient runoff. Mitigation strategies must be considered from a whole-farm perspective to ensure a net environmental benefit (Montes et al., 2013).

Potential CH₄ mitigation strategies include manure solids separation, aeration, acidification, biofiltration, composting, and anaerobic digestion (Montes et al., 2013). Removal of solids from liquid manure reduces available carbon for methanogenesis, and composting or storing the solids in a stack under more aerobic conditions reduces total CH₄ emissions. For long-term manure storage, covers likely will become mandatory to reduce NH₃, CH₄, and N₂O emissions. Semipermeable covers such as the natural crust on slurry manure or added floating materials such as straw, wood chips, expanded clay pellets, and some types of plastic can reduce CH₄ and NH₃ emissions from storage by 30% to 80%, but they also may increase N₂O emissions. Greater reductions and perhaps near elimination of emissions can be achieved by sealing the cover and using a flare to convert the accumulated CH₄ to CO₂. Anaerobic digesters also can be used to enhance CH₄ production, capturing the produced biogas and using it on the farm to heat water and generate electricity. Extracting the carbon from manure reduces storage emissions, and the reduction in purchased gas and electricity provides other off-farm environmental benefits. Composting solid manure in aerated windrows can greatly reduce CH₄ emissions, but this processing will increase NH₃ and N₂O emissions (Montes et al., 2013).

Experimental processes of acidification and biofiltration show potential for reducing CH₄ emissions if practical and economical systems can be developed (Montes et al., 2013). Decreasing the pH of manure reduces NH₃ and CH₄ emissions, but the cost of the acid, safety in handling, and difficulty in maintaining the low pH all deter its use. Biofiltration can extract CH₄ from ventilation air in barns, but the large size and cost preclude adoption. Biofilters also may create N₂O emissions, offsetting some of the carbon reduction benefits.

Rice Production

Rice emits four to five times more CH₄ and N₂O to the atmosphere (Linquist et al., 2012) and uses two to three times more water per kg than other cereals (Bouman et al., 2007; Tuong et al., 2005). Sustainably oriented production practices have been developed with the goal of mitigating the environmental impact of rice and improving the economic benefits through reductions in production costs. These practices include the irrigation management practice of alternate wetting and drying (AWD) or intermittent flooding, whereby the soil surface is allowed to dry for several days to a week before rewetting in midseason. This practice can be repeated up to five times during the growing season without reducing harvest yield. The concurrent re-oxygenation of the soil layer keeps CH₄ emissions low, and studies have shown that water-saving irrigation methods such as AWD reduce net CH₄ emissions produced under water-saturated conditions (Linquist et al., 2015; Rogers et al., 2013). Even one 6-day, midseason drainage event, temporarily reducing anaerobic soil conditions, can reduce post-drainage CH₄ emissions by 64% with no evident effect on yield (Sigren et al., 1997). This practice also has the co-benefit of reducing grain arsenic concentrations because it changes the soil reduction-oxidation (redox) potential (Linquist et al., 2015). Other irrigation techniques that reduce the inundated soil period also will reduce the CH₄ emissions from rice paddies. These methods include the use of drill-seeding rather than water-seeding or transplanting rice (Pittelkow et al., 2014) and carry the additional benefit of reducing the pumping requirements of irrigation water; thus, they will reduce GHG production associated with the energy use of burning fossil fuels—whether through diesel or indirectly through electricity generation. The reduced pumping benefits are particularly true in rice production regions of the Midsouth that are distinct from those in California, where irrigation needs are met from gravity-fed reservoirs draining the Sierra Nevada mountains. However, for any CH₄-reducing rice production regime, care must be taken to keep N₂O emissions low. As indicated, rates of N₂O emissions are particularly sensitive to inputs from nitrogen fertilization, fallow-season field conditions, and midseason or season-end drainage events (Pittelkow et al., 2013). In many cases, both CH₄ and N₂O are released in any drainage event, with end-of-season drainage transferring 10% of seasonal CH₄ and 27% of seasonal N₂O to the atmosphere as entrapped gases are released from the soil.

Between 1960 and 2000, global crop net primary production (NPP) more than doubled, and global cropland area in 2011 was estimated to be 1.3 billion ha (Wolf et al., 2015). Global crop NPP in 2011 was estimated at 5.25 Pg C, of which 2.05 Pg was harvested and respired offsite (Wolf et al., 2015). Global livestock feed intake was 2.42 Pg C, of which 52% was grazed and the rest was either harvested biomass or residue collected from croplands. Global human food intake was 0.57 Pg C in 2011 (Wolf et al., 2015). The global agricultural carbon budget indicates a general increase in NPP, harvested biomass, and movement of carbon among global regions. At the global scale, cereal crops declined and have been replaced primarily with corn, soybean, and oil crops. While total NPP and yield (i.e., biomass per area) have increased in nearly all global regions since 1960, the most pronounced increase has been in southern and eastern Asia where harvested biomass has tripled. Also, cropland NPP in the former Soviet Union significantly declined in 1991, with the level of production recovering around 2010 (Wolf et al., 2015).

Annual crop cultivation and crop burning often is considered carbon neutral (IPCC 2006; U.S. EPA 2018) because biomass is harvested and regrown annually. Although biomass itself is technically carbon neutral, this assumption does not necessarily account for changes in soil carbon that may be associated with production practices, which affect the carbon cycle and net emissions. The impact of non-CO₂ emissions is accounted for in the other categories. The increased global uptake of carbon by croplands influences the annual oscillation of global atmospheric carbon (Gray et al., 2014), as more carbon is taken up and released annually than would occur without extensive global cropland production. The cycling of cropland biomass into soils and the cultivation of soils influence how much of the carbon in crop biomass is respired back to the atmosphere versus remaining in the soil, ultimately determining if a cropping system is a net source or sink.

5.7.1 Inventory Uncertainties

As previously discussed, enteric and manure fermentation are the sources of livestock CH₄ emissions. These two sources are affected by different factors and carry different levels of uncertainties. The U.S. EPA estimated 95% confidence interval lower and upper uncertainty bounds for agricultural GHG emissions at –11% and +18% (CH₄ emissions from enteric fermentation) and –18% and +20% and –16% and +24% (CH₄ and N₂O emissions from manure management, respectively; U.S. EPA 2018). Whereas emissions from enteric fermentation are relatively well studied and predictable, there is larger uncertainty regarding manure CH₄ emissions and net effects of different intensities and types of grazing (see also Ch. 10: Grasslands). Large datasets have established CH₄ emissions from enteric fermentation at 16 to19 g per kg dry matter intake for dairy cows (higher-producing cows have lower emissions per unit of feed intake) to 21 to 22 g per kg dry matter intake for beef cows on pasture (Hristov et al., 2013b). Levels of manure CH₄ emissions, however, largely depend on the type of storage facility, duration of storage, and climate (Montes et al., 2013). Emissions from certain dairy manure systems (e.g., flush systems with settling ponds and anaerobic lagoons) can be higher than estimates used by current inventories. So-called top-down approaches have suggested that livestock CH₄ emissions are considerably greater than EPA inventories. Miller et al. (2013) and Wecht et al. (2014) proposed that livestock CH₄ emissions may be in the range of 12 to 17 Tg per year, which is roughly 30% and 85% greater than EPA’s estimate for 2012 (U.S. EPA 2016). Thus, future research is needed to address these discrepancies and reconcile top-down and bottom-up estimates.

Large uncertainties in GHG emissions from agricultural systems also exist because of their high spatial and temporal variability, measurement methods, cropping systems, management practices, and variations of soil and climatic conditions among regions (Hristov et al., 2017b, 2018). Uncertainty in GHG measurements often exceeds 100% (Parkin and Venterea 2010). Finally, there is considerable uncertainty in soil carbon accumulation and emissions from soils under different conditions and management practices, all of which are complicated by uncertainties about the total amount of land area under different management practices (see Ch. 12: Soils for more information on soil carbon balance).

5.7.2 Modeling and Modeling Uncertainties

Whole-farm models representing all major farm components and processes provide useful tools for integrating emission sources to predict farm-scale GHG emissions (Del Prado et al., 2013). By predicting emission processes and their interactions, models can provide a better understanding of production system emissions and be used to explore how different management decisions could affect GHG emissions. This approach has been used to estimate the carbon footprint of common U.S. dairy production systems at around 1 ± 0.1 kg CO₂e per kg fat- and protein-corrected milk produced, in which about half of these emissions come from enteric CH₄ emissions (Rotz and Thoma 2017). With a similar approach, the carbon footprint of beef cattle production was found to be 18.3 ± 1.7 kg CO₂e per kg carcass weight, with about 60% of emissions in the form of enteric and manure management CH₄ (Rotz et al., 2015).

Uncertainty exists in any measurement or projection of GHG emissions. The uncertainty of farm-scale projections is related to the uncertainty in projecting emissions from individual sources (Chianese et al., 2009). The IPCC (2006) suggested a ±20% uncertainty in predicting both enteric and manure management CH₄ emissions. Through the use of process-based models representing common management strategies for the United States, the uncertainty for predicting enteric emissions may be reduced to ±10%, but uncertainty for manure management likely will remain around ±20% (Chianese et al., 2009). Considering these uncertainties along with those of other agricultural emission sources, total GHG emissions can be determined with an uncertainty of ±10% to ±15%. As process-level models improve, verified with accurate measurements, this uncertainty can be reduced. As with inventories, uncertainties also are great for modeling agricultural carbon fluxes related to soil processes. Improving the modeling of these processes and incorporating them into large-scale carbon flux models will help increase understanding and reduce uncertainties in carbon models for agricultural lands.