Carbon cycling in cultivated land and its global significance
Post on 06-Jul-2016
Embed Size (px)
Global Change Biology (1998) 4, 131141
Carbon cycling in cultivated land and its globalsignificance
G R E G O R Y A . B U YA N O V S K Y and G E O R G E H . WA G N E RUniversity of Missouri-Columbia, Soil and Atmospheric Science Department, 144 Mumford Hall, Columbia, MO 65211, USA
Long-term data from Sanborn Field, one of the oldest experimental fields in the USA,were used to determine the direction of soil organic carbon (SOC) dynamics incultivated land. Changes in agriculture in the last 50 years including introduction ofmore productive varieties, wide scale use of mineral fertilizers and reduced tillagecaused increases in total net annual production (TNAP), yields and SOC content. TNAPof winter wheat more than doubled during the last century, rising from 2.02.5 to56 Mg ha1 of carbon, TNAP of corn rose from 34 to 9.511.0 Mg ha1 of carbon.Amounts of carbon returned annually with crop residues increased even more drastically,from less than 1 Mg ha1 in the beginning of the century to 33.5 Mg ha1 for wheatand 56 Mg ha1 for corn in the 90s. These amounts increased in a higher proportionbecause in the early 50s removal of postharvest residues from the field was discontinued.SOC during the first half of the century, when carbon input was low, was mineralizedat a high rate: 89 and 114 g m2 y1 under untreated wheat and corn, respectively.Application of manure decreased losses by half, but still the SOC balance remainednegative. Since 1950, the direction of the carbon dynamics has reversed: soil underwheat monocrop (with mineral fertilizer) accumulated carbon at a rate about 50 g m2 y1, three year rotation (corn/wheat/clover) with manure and nitrogen applicationssequestered 150 g m2 y1 of carbon. Applying conservative estimates of carbon sequestra-tion documented on Sanborn Field to the wheat and corn production area in the USA,suggests that carbon losses to the atmosphere from these soils were decreased by atleast 32 Tg annually during the last 4050 years. Our computations prove that cultivatedsoils under proper management exercise a positive influence in the current imbalancein the global carbon budget.
Keywords: carbon sequestration, crop residue, cultivated land, global carbon balance, net annualproduction
Received 24 October 1996; revised version accepted 11 March 1997
Carbon flow through cultivated lands has never beenstudied to the same extent and detail as that in the nativeecosystems. It is widely accepted that conversion ofnative land, be it prairie or forest, into a cultivated systemcauses precipitous degradation of the soil organic matter(SOM). Typically 2040% of the native SOM is lost whenvirgin lands are converted to agriculture (Schlesinger1986; Mann 1986; Detwiler 1986; Cole et al. 1989). It isgenerally assumed that over time cultivated soils reacha new equilibrium at a lower level of organic carbon(Haas et al. 1957; Hobbs & Brown 1965; Unger 1968;Mann 1985).
Correspondence: Dr Gregory Buyanovsky, fax 1 1/573-884-4960,e-mail firstname.lastname@example.org
1998 Blackwell Science Ltd. 131
Post-harvest residues are the sole source of carbon toreplenish soil organic matter decomposing as a result ofcultivation. Linear relationships between carbon inputswith residues and soil organic matter levels have beenestablished in several field experiments (Rassmussenet al. 1980; Cole et al. 1993; Rasmussen & Parton 1994).However, long-term observations of organic carbondynamics in cultivated soils combined with data on totalproductivity of crops are extremely rare. Because of thenon-nutrient status of carbon, its flow rate and storagecharacteristics are almost never included in agronomicstudies. Uncertainties in the carbon balance of agriculturallands prevent proper assessment of their role in globalcarbon balance, thus precluding accurate global carbon
132 G . A . B U YA N O V S K Y & G . H . WA G N E R
balance sheets, and, hence, approximations of the poten-tial for sequestration of carbon in cultivated soils areimprecise.
There are concerns about a current imbalance in theglobal carbon budget. The difference between a 0 netflux and a release from tropical deforestation is 1.52.0 Pg y1 (Houghton 1995). Assumptions have beenmade that the carbon which cannot be accounted for isaccumulating in temperate ecosystems of the NorthernHemisphere (Tans et al. 1990), mainly in forests. Houghton(1995) argues, however, that an accumulation of thismagnitude in forests is unlikely. As for croplands andgrasslands, an increase in carbon content of such magni-tude, in his opinion, would be too obvious to gounnoticed. Taking into account total amount of carbonin agricultural ecosystems (111142 Pg) (Schlesinger 1984;Buringh 1984), a yearly accumulation of 12 Pg C wouldcorrespond to an increase of about 1%. For a soil with2.5% of organic carbon such an increase would beunnoticeable for at least 4050 years. Soil survey oroccasional observations could not reveal changes of thismagnitude due to differences in understanding of whatconstitutes SOM, the high variability in its content, andshortcomings of analytical techniques. Only long-termobservations with permanent sampling sites, tightly con-trolled management and documented inputs and outputscan register small increase occurring during a spanof decades.
Sanborn Field, one of the oldest experimental fields inthe United States, uniquely meets the requirements forstudies of this kind. The Field has been maintainedthroughout the last 110 years. Despite shortcomings ofthe initial layout and irregular sampling of soil and crops,the field presents an invaluable asset for studies of carbonflux. Records on management, fertilization, yields of grainand forage (above-ground biomass), as well as the resultsof some analyses of historic soil samples are available.Particularly applicable are Sanborn Field records relatedto numerous fertility experiments conducted during thelast century, which allow an assessment of carbondynamics during a 100 1 year period since cultivationwas commenced on the field (Upchurch et al. 1985). Inaddition, important complementary studies using thisfield have focused, in recent years, on carbon cycling.This has opened the possibility of using old records incombination with data from sophisticated recentmeasurements (Buyanovsky & Wagner 1986, 1987;Buyanovsky et al. 1987; Balesdent et al. 1988).
We evaluated experimental material collected duringthis long period and transposed this into carbon balancecharacteristics for cultivated fields representative ofMidwest agriculture. Combined with national crop statist-ics these evaluations were used to assess possible carbon
1998 Blackwell Science Ltd., Global Change Biology, 4, 131141
sequestration in U.S. croplands during postwar agricul-tural practices.
Sanborn Field Data Set
Sanborn Field, on the campus of the University ofMissouri-Columbia, was established in 1888 with rotationand manure treatments on 39 plots, each 30 3 10 m insize, separated by grass borders 1.5 m wide. The soil isa Mexico silt loam (fine montmorillonitic, mesic, UdollicOchraqualf) developed in thin loess deposits overlyingglacial till. The surface layer contains 2.52.9% organicmatter. Mean annual air temperature of the region is13 C, with maximum monthly average in July (26 C)and minimum in January ( 1.5 C). Mean annualprecipitation is 973 mm, with potential evapo-transpiration of 790 mm. The soil has an argillic horizon(Bt), which causes perching and lateral flow of waterabove. The soil is typical for the American Midwestclaypan area, and there are about 20 million ha ofagricultural land in central part of the USA with similaredaphic characteristics used for intensive grain produc-tion. Annual carbon circulation for this region canapproach hundreds of millions of tons.
Sanborn Field plots have been cropped and managedunder specified guidelines simulating regular farmingpractices since the inception of the field (Upchurch et al.1985; Buyanovsky et al. 1990; Brown 1994). Initially, ninecropping practices, using corn, oats, winter wheat, redclover and timothy were used in the experiment. Someplots were under continuous single crops and otherinvolved rotations. The continuum of management prac-ticed on numerous plots of Sanborn Field reflects thehistory of agriculture in the central region of the USA.
For the analyses reported herein we have used datafrom treatment plots most of which have maintainedtheir integrity throughout the whole period (Table 1).Periodically, some revisions were made in the experi-mental plan of the field, but for the work reported hereonly one change in management was of real importance,that of discontinuation from 1950 onward of the practiceof collecting above-ground residues from plots cultivatedto corn and wheat. Forages (timothy, alfalfa, bromegrass)have always been managed with forage harvested andremoved from the plot.
Since 1981, we have conducted on Sanborn Field severalsmall-scale field experiments with major regional crops,among them winter wheat (Triticum aestivum L.) and corn(Zea mays) (Buyanovsky & Wagner 1986, 1987, 1997a,1997b; Buyanovsky et al. 1986, 1987, 1994). The experi-ments which employed 14C labelling technique have beendesigned to assess total net annual production of thecrops which includes grain, above-ground biomass at thetime of harvest, and roots (measured several times during
C U L T I VA T E D L A N D A N D G L O B A L C A R B O N 133
Table 1 Cropping systems of Sanborn Field used for analysis of soil organic carbon dynamics
System Treatment Years under thetreatment
Continuous wheat Manure (13.4 Mg ha1) 100Full mineral fertilizer 100None 100
Continuous corn Manure (13.4 Mg ha1) 100Full treatment 50None 100
Continuous timothy Manure (13.4 Mg ha1) 100None 100
Corn/Wheat/Clover Manure (13.4 Mg ha1) 60Manure (13.4 Mg ha1) 1 N (37 kg ha1 under wheat, 112 kg ha1 under corn) 40
a growing season). This information has been used toestimate the relationship between different parts of a crop(grain/TNAP, shoot/root ratio, etc.), and, subsequently, tocalculate the total carbon input under different cropsduring the 1001 year period of the large-scaleexperiments.
Measurements of soil organic carbon have been takenwith intervals of 1030 years, by dry combustion in apurified stream of oxygen (Nelson & Sommers 1982). Torecalculate organic carbon on soil mass we used detailedbulk density measurements from the 60s and 80s.
Total net annual production (TNAP) TNAP of cropscultivated on Sanborn Field have been heavily impactedby many factors, among which the major ones are man-agement, weather and variety. With factors other thanweather progressively improving, TNAP increased(Table 2). The only notable exceptions are plots withoutamendments.
Application of mineral fertilizers or manure to winterwheat provided a very slow increase in TNAP duringthe first 60 years. During that period, net productionincreased from 2 to 3 Mg C ha1 y1. During thefollowing 40 years, with modernization of cultivationpractices, introduction of genetically improved varietiesand with residues added back, each hectare of wheataccumulated 45 Mg C ha1 y1. Rotation with manureapplied annually supported high productivity during thewhole 100-year period. However, when manure was usedwith mineral fertilizers its effect on TNAP was negligible.
Manured corn during the first 50 years produced anaverage of about 3.2 Mg C ha1 y1, as compared with2.2 Mg C ha1 y1 for the untreated plot. Corn responseto manure was lower than that for wheat, which doubledTNAP under the effect of manure. In a 3-year rotation(corn, wheat, clover) with manure, annual carbon accu-mulation increased but not very significantly (from about
1998 Blackwell Science Ltd., Global Change Biology, 4, 131141
3.3 to 4.2 Mg C ha1 y1) during first 60 years. Applicationof nitrogen was necessary to increase TNAP of corn to910 Mg C ha1 y1.
TNAP of unfertilized plots practically did not changeduring 100 years. Slightly higher TNAPs of untreatedwheat and corn were observed after 1950, when thepractice of collecting residues was abandoned. On aver-age, nonfertilized wheat accumulated about 11.5 Mg Cha1 y1 during the growing season and corn accumulated1.72.2 Mg C ha1 y1.
Naturally, TNAP indirectly mirrored the increase ingrain yields. A general progressive yield increase wasobserved from the inception of the experiment, butduring the first 50 years, grain productivity offertilized wheat increased from 0.91 to 1.51.6 Mg ha1
y1 and in the following 50 years it more than doubledto 4.2 Mg ha1 y1. Productivity of corn in a 3-yearrotation with 13.4 Mg ha1 manure and 112 kg ha1 offertilizer nitrogen increased by 25% during the first50 years and doubled during the later period.
The early period increases are probably attributableto improved varieties. During the latter period furtherprogress in plant breeding would have accelerated thepositive effect on yield by linking this breeding effort toselection toward fertilizer response and due to betterweed control and improved cultivation practices. In theearly years, plant breeders of wheat improved harvestability of grain by selection of stronger stems and someyears ago they successfully decreased straw length. Theratios of grain yield to vegetative biomass have narrowedonly slightly, however (Buyanovsky & Wagner 1997a).
Carbon input to soils Sanborn Field was established onan area of tallgrass prairie with a plant cover characterizedby several dominant warm season grasses including bigbluestem (Andropogon gerardi Vitman), little bluestem(Schizacharium scoparium Nash), prairie drop seed (Sporo-bolus heterolepis [A. Gray] A. Gray), and Indian grass(Sorghastrum nutans [L.] Nash). Kucera (1987) estimated
134 G . A . B U YA N O V S K Y & G . H . WA G N E R