potential carbon dioxide emission reductions from avoided grassland conversion in the northern great plains

For all soil types across 14 counties in North and South Dakota, we used the DAYCENT model calibrated to the study area to quantify the difference in CO2and N2O emissions under a croppin

grassland conversion in the northern Great Plains

MARISSAAHLERING,1,  JOSEPHFARGIONE,1ANDWILLIAMPARTON2 1

The Nature Conservancy, 1101 West River Parkway, Suite 200, Minneapolis, Minnesota 55415 USA

2

Natural Resource Ecology Laboratory, Colorado State University, NESB B233, Fort Collins, Colorado 80523 USA

Citation: Ahlering, M., J Fargione, and W Parton 2016 Potential carbon dioxide emission reductions from avoided grassland conversion in the northern Great Plains Ecosphere 7(12):e01625 10.1002/ecs2.1625

Abstract Protection of lands threatened with conversion to agriculture can reduce carbon emissions Until recently, most climate change mitigation incentive programs for avoided conversion have focused on forested ecosystems We applied the Avoided Conversion of Grasslands and Shrublands v.1.0 (ACoGS) methodology now available through the American Carbon Registry to a threatened region of grasslands in the northern Great Plains For all soil types across 14 counties in North and South Dakota, we used the DAYCENT model calibrated to the study area to quantify the difference in CO2and N2O emissions under

a cropping and a protection scenario, and we used formulas in the ACoGS methodology to calculate CH4

emissions from enteric fermentation under the protection scenario We mapped the resulting GHG emis-sions across the entire project area Emisemis-sions averaged 51.6 tCO2e/ha over 20 years, and with a 31% reduction for leakage and uncertainty from the ACoGS methodology, carbon offsets averaged 35.6 tCO2e/

ha over 20 years Protection of 10% of the 2.1 million unprotected ha in the project area with the highest emissions would reduce emissions by 11.7 million tCO2e over 20 years (11% of the total emissions from all unprotected grassland) and avoid a social cost of $430 million worth of CO2emissions These results sug-gest that carbon offsets generated from avoided conversion of grasslands can meaningfully contribute to climate mitigation and grassland conservation objectives

Key words: carbon offsets; DAYCENT; grassland conversion; greenhouse gas emissions; Prairie Pothole Region.

Received 4 October 2016; accepted 26 October 2016 Corresponding Editor: Nichole N Barger.

Copyright: © 2016 Ahlering et al This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited   E-mail: mahlering@tnc.org

INTRODUCTION

The demand for food, feed, and energy is

expected to increase in response to a projected

global population of 9.15 billion by 2050

(Alexan-dratos and Bruinsma 2012) and biofuels policy

(Searchinger et al 2009, Lark et al 2015) This

demand can be met by both agricultural

inten-sification through higher yields and

extensifica-tion through conversion of grasslands and

forests to cropland (Johnson et al 2014) The

con-version of natural systems to cropland, however,

may degrade other ecosystem services provided

by these habitats, such as carbon storage, soil

retention, and recreation (Gascoigne et al 2011,

Yahdjian et al 2015) For example, it is estimated that agriculture as a whole contributes 30–35% of the global greenhouse gas (GHG) emissions (Foley et al 2011)

Globally, rangeland habitat covers over 50% of the land surface with just under half in grassland

or savannah (Estell et al 2012) Because of the productivity of grassland soils, approximately 70% of the world’s grasslands have already been converted to agriculture (Foley et al 2011) North American grasslands are no exception (Samson et al 2004), where conversion is still continuing at a rapid pace In particular, in the Prairie Pothole Region (PPR) of North America, conversion of grassland to row-crop agriculture

in recent years has occurred at rates similar to

the highest observed rates of tropical

deforesta-tion in the Brazilian Amazon (Wright and

Wim-berly 2013)

Grassland protection and restoration are key

conservation strategies, and an important policy

mechanism to restore grassland has been the

Con-servation Reserve Program (CRP) Funding for

this program has been reduced, and CRP

enroll-ment has declined by over 10 million acres since

its peak of 37 million acres in 2007 (Stubbs 2014)

Further economic incentives may be necessary to

maintain grasslands One potential option is a

car-bon offset program Carcar-bon offsets have been

available for reduced deforestation and

degrada-tion (REDD) projects in forested ecosystems for a

few years (Gibbs et al 2007, Olander et al 2008)

REDD methodologies have been published by

certifiers, and carbon credits for protection of

forested systems are traded on the voluntary

mar-ket both within the United States and abroad

Until recently, an approved methodology did not

exist to certify carbon offsets from avoided

con-version of grassland systems, but with the

publi-cation of a methodology by the American Carbon

Registry (Dell et al 2013), the menu of potential

economic incentives for private landowners to

maintain their rangelands for ranching instead of

row-crop agriculture has increased

This is thefirst study to quantify the potential

GHG emission reductions for the avoided

con-version of both restored and native grassland at

a large scale We chose a particularly at-risk

region in the northern Great Plains, the central

PPR, because of its significance to grassland

biodiversity (Doherty et al 2015) and the high

threat of conversion (Wright and Wimberly 2013,

Lark et al 2015) We applied the existing

Avoided Conversion of Grassland and

Shrub-lands to Row Crop Agriculture v1.0

methodol-ogy (Dell et al 2013) to the soil types across the

study area to quantify the climate benefits of

grassland protection in this region We used the

existing and well-tested DAYCENT model

cali-brated for the study area to quantify potential

emissions (Hartman et al 2011, Del Grosso et al

2016) In the PPR, extensive work has been done

to quantify carbon sequestration rates for

restored wetlands (Gleason et al 2005, 2008) and

others have quantified sequestration of restored

grasslands at a local scale (Phillips et al 2015)

This study focuses exclusively on upland grass-land habitats, incorporates other sources of GHG emissions such as N2O and CH4,and maps emis-sions at a regional scale

METHODS

Study area

We focused on 14 counties in central North Dakota and South Dakota: Hyde, Hand, Faulk, Edmunds, and McPherson in South Dakota and McIntosh, Emmons, Logan, Stutsman, Kidder, Burleigh, Sheridan, Mclean, and McHenry in North Dakota These counties align with the Mis-souri Coteau landform, a prominent feature in the PPR that is important habitat for waterfowl, water-bird, shorewater-bird, and grassland songbird popula-tions (Doherty et al 2015) We targeted this area because of its biodiversity significance and high rate of grassland to cropland conversion (Wright and Wimberly 2013, Lark et al 2015)

DAYCENT modeling

To calculate potential carbon emissions, we included the following greenhouse gas (GHG) pools: above- and belowground live biomass and soil organic carbon We also included net nitrous oxide (N2O) flux and methane (CH4) flux from enteric fermentation by cattle We used the DAY-CENT model to calculate the potential avoided

CO2and N2O emissions for each soil type in the

14 county regions (Hartman et al 2011) We cali-brated the model output using data from the region The soil carbon output was validated using long-term datasets from North and South Dakota (Haas and Evans 1957) and data from cultivated and uncultivated sites in the Great Plains (Burke et al 1989) The N2Oflux output was calibrated with data from long-term datasets (Haas and Evans 1957), and the DAYCENT model has been shown to correctly simulate N2O fluxes for different types of agricultural systems (Del Grosso et al 2008a, b) Finally, aboveground plant productivity was calibrated using a dataset from North and South Dakota (Sala et al 1988) Because soil carbon can vary dramatically by soil type, we ran DAYCENT for each soil type across the project area, parameterizing the model with county-level weather and latitude and lon-gitude values For climate, we used repeated VEMAP and DayMet interpolated daily weather

files from 1895 to 2003 (Kittel et al 1997,

Thorn-ton et al 1997) For each county, we extracted

every unique combination of percentage sand,

silt, and clay values from the soil taxa occurring

in that county according to the SSURGO dataset,

which resulted in 888 unique combinations The

soil data from SSURGO are mapped as polygons

across the study area, and DAYCENT uses the

soil data to model carbon in gC/m2and nitrogen

in gN/m2 To avoid outliers, errors, and missing

data in the SSURGO dataset, bulk density was

derived from the SSURGO texture values (Saxton

et al 1986) For each soil type, we ran three

DAY-CENT simulations: one to establish an historic

baseline and equilibrium, a second to simulate

the most likely cropping scenario had it been

converted, and a third to simulate the

continua-tion of the current native condicontinua-tion For all

simu-lations of the native grassland, we used a mixed

grass prairie that included nitrogen fixers and

where the grass component was a 50/50

break-down of warm and cool season grasses As a

starting point for further simulations, we ran the

mixed grass prairie through the DAYCENT

model for 5000 years with a moderate impact of

grazing This provided a historic baseline with

carbon pools in equilibrium We used the output

of this simulation as the starting point for the

second two simulations

We obtained information on the dominant

crop rotation, conversion practices, and

subse-quent tillage practices for recently converted

land from Natural Resources Conservation

Ser-vice (NRCS) District Conservation officers for

both South Dakota and North Dakota We used

this information to parameterize our simulations

for the conversion to cropping scenario Using

the starting values from the baseline simulation

that brought the carbon pools to equilibrium, we

ran the cropping scenario for 150 years Thefirst

50 years was a continuation of the native mixed

grass system with moderate grazing pressure In

year 51, an end of growing season haying

fol-lowed by an herbicide treatment was applied

The remaining years alternated a corn and

soy-bean rotation with conventional plowing

occur-ring every year prior to planting, as our NRCS

consultations indicated were still the dominant

practices Fertilizer was applied to the corn

rota-tions twice, once at the time offirst cultivation at

a rate of 5 gN/m2 and once at the time of

planting at a rate of 6.6 gN/m2 Fertilizer was not added to the soybean plantings because soy-beans are nitrogen fixers and do not require nitrogen additions The same cultivation meth-ods were used for both corn and soybean, and the grain was harvested at the end of every growing season for both crops To simulate the continuation of the native condition under the avoided conversion scenario, we used the start-ing values from the historic baseline scenario and extended the native condition scenario of moderate grazing for 150 years For both of the extend simulations (i.e., native conditions and the cropping scenario), we used repeated weather data for each county from 1973 to 2003 (Thornton et al 2014) From each scenario, we used the annual output variables for total sys-tem carbon, aboveground production, and N2O flux to calculate the CO2 and N2O emissions Belowground and aboveground CO2 emissions were calculated separately for each scenario Belowground carbon was calculated as the dif-ference between total system carbon and above-ground carbon

The DAYCENT model is sensitive to the weather data used to run the simulations Large fluctuations in temperature and precipitation can cause large annual variation in the amount of carbon lost or gained Therefore, wefit an emis-sion curve to the modeled data to reduce the annual variability in emissions due to weather inputs in any given year (Appendix S1) Emis-sions are based on changes in the belowground carbon pool For the year after conversion, the relevant change in the belowground carbon pool

is the change from unconverted grassland To avoid undue influence of climate variables on our estimate, we averaged our estimate of carbon

in unconverted grassland over a 100-year time-frame to capture the majority of the climate vari-ability in our weatherfile, rather than the carbon level from the single year prior to conversion For subsequent years, the relevant change in the belowground carbon pool is the year-on-year temporal change in the cropped scenario As the belowground carbon pool in the cropland reaches a new equilibrium, the pool stops chang-ing and emissions decline to zero An exponen-tial loss model best fit the loss of belowground

CO2 from the cropping simulations This fitted model for each soil type in each county was then

used to calculate emissions on an annual basis

for the 100 years following conversion

Nitrous oxide emissions are reported from the

simulations as an annual flux for both the

cropped and grassland scenarios The offsets are

based on the difference between these annual

fluxes Therefore, beginning with the first year of

cultivation, we subtracted each year of the

grass-land scenario from the same year of the cropping

scenario to obtain the annual N2O emissions for

avoided conversion We used the difference in

N2Oflux over time to fit a linear model that

aver-aged over temporal fluctuations due to weather

variability

The aboveground biomass pool changes only

once, immediately upon conversion to annual

cropland We calculated the average daily

above-ground live carbon for the 100-year project

period for both the grassland and cropland

sce-narios and subtracted the cropping production

from the grassland production The result was

added to thefirst year of emissions from the total

system carbon and the nitrous oxide

The expected land use for grasslands not

con-verted to cropland in this region is cattle

produc-tion Therefore, methane emissions from the

cattle were deducted from the emissions accrued

from the avoided conversion We calculated the

methane emissions produced through enteric

fer-mentation from cattle using the equations in the

American Carbon Registry’s Methodology for

the Avoided Conversion of Grasslands and

Shrublands (ACoGS; Dell et al 2013) Methane

emissions are based on the number and type of

livestock, the number of days spent grazing in

each year, the enteric fermentation emission

fac-tor of methane per head per day, and the global

warming potential for methane We used the

rec-ommended values from the ACoGS

methodol-ogy for the enteric fermentation emission factor

of non-dairy cattle (53 kg CH4head 1
yr 1) and

for the global warming potential of methane

(Hongmin et al 2006)

To estimate the methane emissions for thefive

focal counties in north central South Dakota, we

used a publication distributed by South Dakota

State University’s Extension office to obtain

rec-ommended stocking rates (Mousel 2013:35) Cattle

are the dominant livestock type in the focal area,

and Mousel (2013) provides stocking

recommen-dations for 1400 lb cattle across the state For

1400 lb cows, two different recommended stock-ing rate categories occur across thefive focal coun-ties, 1.33 and 1.01 cows
ha 1
month 1 Because stocking rate recommendations incorporate both number of cows and length of time spent grazing, these numbers can be used to calculate emissions Using the above recommendations, we calculated the methane emissions from enteric fermentation

to be 0.123 or 0.094 tCO2e
ha 1
yr 1, depending

on the county (Table 1)

The same procedure for calculating methane emissions from enteric fermentation was used for North Dakota Manske (2004) reports recom-mended stocking rates across North Dakota based

on three ecological divisions: drift prairie, Mis-souri Coteau, and West River All of the North Dakota counties in this project fall primarily within the Missouri Coteau, with the exception of McHenry County, which is primarily drift prairie For each county, we used the recommended stocking rates for the appropriate landform on good condition upland prairies To be consistent with the recommendations made for South Dakota, we used the stocking rates given for

1400 lb cows For the Missouri Coteau counties, methane emissions from enteric fermentation are 0.101 tCO2e
ha 1
yr 1, and for the drift prairie, methane emissions are 0.119 tCO2e
ha 1
yr 1

(Table 1)

Table 1 Stocking rate recommendations for 1400 lb cows in cows
ha 1
month 1 by county for the 14 focal counties in central North Dakota and South Dakota (Manske 2004, Mousel 2013) and associated enteric methane emissions, expressed as metric tons

of carbon dioxide equivalents

County

Stocking rate (cows
ha 1
month 1

)

Methane emissions (tCO 2 e
ha 1
yr 1

)

Data analysis

Once the total emissions were calculated for

each soil type in each county, we cross-walked

the emissions back to the soil types in the spatial

SSURGO data layer to map the emissions across

the 14 county regions Because carbon offsets are

only available for unprotected lands at risk of

conversion, we created an unprotected grassland

layer for the project area using ArcGIS Desktop

10.1 (ESRI, Redlands, California, USA) We

cre-ated a current grassland and shrubland layer

from the 2014 CropScape data layer (USDA

National Agricultural Statistics Service Cropland

Data Layer 2014) including the following

land-cover classes: grassland/pasture, herbaceous

wet-lands, other hay/non-alfalfa, and shrubland

Grasslands in this region are commonly

inter-spersed with wetlands, but wetland soils are not

eligible for carbon credits under the ACoGS

methodology We used the National Wetlands

Inventory data layer to remove any remaining

wetlands from the grassland layer (U.S Fish and

Wildlife Service 2010) Finally, we used the

national Protected Areas Database (USGS GAP

2012) and the National Conservation Easement

Database (NCED 2011) to remove already

pro-tected lands from the current grassland layer

We summarized total emissions and emissions

from the 10% of unprotected hectares with the

highest and lowest emissions In reality,

individ-ual pastures will include a range of soil types

and emission values for a given pasture would

be calculated for each soil type and weighted by

the percentage of that soil type in the pasture

Because we do not have access to parcel

bound-aries, we calculate emissions based on soil types

in the highest and lowest categories to report the

variability in the dataset across the project area

ACoGS projects are focused on grasslands under

imminent risk of conversion to cropland Risk of

conversion is driven by the value of the land as

cropland vs pastureland To evaluate lands at

risk based on cash value, we used survey data

from the U.S Department of Agriculture (USDA)

to compare rental rates of cropland to

pasture-land by county across our study area (http://quic

kstats.nass.usda.gov/results/58B27A06-F574-315B

-A854-9BF568F17652#7878272B-A9F3-3BC2-960D

-5F03B7DF4826) Lands with a value as cropland

at least twice that of pastureland were considered

to be under imminent risk Using a 3-year average

(2012–2014), the value of cropland was at least double the value of pastureland for all counties except McPherson, SD The value of cropland in McPherson county was 93% greater than pasture-land; avoided conversion from this county could still be certified, but with a discount applied to the estimated avoided conversion

Within each county, some lands are unsuitable for cropping and so are not at risk of conversion

We identified and excluded these lands from our analyses To identify these lands, we used the Natural Resource Conservation Service’s Land Capability Classes (LCCs) in the SSURGO data-base to characterize cropping suitability The LCC is an index with integer values that range from 1 to 8 To determine which LCCs were at risk of conversion, we used empirical data on recent conversion in our project area We used the CropScape data (USDA NASS CDL 2014) to evaluate how much grassland was lost across the project area between 2010 and 2014 for each LCC We also converted the tCO2e emissions to a social cost of carbon from published values (EPA 2013c) We used the Environmental Protection Agency’s (EPA) Interagency Working Group on Social Cost of Carbon recommended value of $37 per tCO2under a discount rate of 3.0% in 2015

We report all emissions for thefirst 20-year per-iod of the DAYCENT modeling because this is the minimum project term for ACoGS projects and the period over which the majority of the carbon emissions accrue (Dell et al 2013)

RESULTS

The calibration datasets were well simulated

by the model The model showed a 30% loss of soil carbon after 40 years, which is consistent with research in this region (Haas and Evans 1957) and meta-analyses (Davidson and Acker-man 1993), and results showed that soils high in clay content had lower loss of carbon than sandy soils, which is consistent with empirical data reviewed by Burke et al (1989) The modeled

N2Oflux rates also matched empirical data well, with less than a 10% difference between simu-lated and observed values (Haas and Evans 1957) Finally, the model predicted annual grass production in the range of 100–110 gC/m2, which

is within the range reported by Sala et al (1988), and the simulated mean corn and wheat yields

also matched observed values from 1940 to 2003

(Hartman et al 2011)

Across the project area, emissions on different

soil types ranged between 36.8 and 63.3 tCO2e/ha

over 20 years assuming constant tillage and

crop-ping scenarios (mean = 51.6 tCO2e/ha, SD= 2.9,

median= 51.7 tCO2e/ha) Belowground carbon,

including both plant biomass and soil carbon, is

by far the greatest predicted source of emissions

with emissions from N2O a distant second

(Table 2) Model results for aboveground biomass

production for corn and soybean crops are quite

high and on average greater than the

above-ground biomass production for native mixed

grass prairie (Table 2) This difference in

produc-tivity must be subtracted from the offsets accrued

from the belowground and N2O emissions

Fur-thermore, when grasslands in this project are

pro-tected from conversion to cropland, the most

common use of these grasslands is as rangeland

or pasture for cattle grazing Cattle are not

gener-ally present on the site once it has been converted

to corn and soybeans (although some grazing of residue is possible, we conservatively do not esti-mate emissions from enteric fermentation on crop residue) Therefore, the methane emissions from enteric fermentation of the cattle in the avoided conversion scenario must also be subtracted from the belowground and N2O offsets (Table 2) The 14 county project area covers 5,130,527 ha that includes 2,080,095 ha of unprotected grass-land (Fig 1) If all of the unprotected grassgrass-land was converted to annual row crop, it would release 106,879,912 metric tons of CO2e into the atmo-sphere The variability in per ha CO2e emissions is largely stratified by latitude and longitude across the project area with the counties in South Dakota and the southeastern portion of the North Dakota project area having the highest per ha emissions (Fig 1) In the DAYCENT model, county-level weather data and latitude and longitude were used and probably account for most of the county-level boundary differences that appear across the project area Regardless, the variability in emis-sions between soil types across the entire project area is not dramatic Compared to the 10% of ha with the highest emissions (11,741,194 tCO2e), the 10% of ha with the lowest emissions is only 20% lower (9,387,896 tCO2e)

Within the project area, there was a 14% loss of unprotected grassland between 2010 and 2014

Of this unprotected grassland, 80% of it was in LCCs 1–4 and 84% of the conversion occurred on this land; 15% of the unprotected grassland was

in LCCs 5–6 and 15% of the conversion occurred

on this land; 5% of the unprotected grassland was in LCCs 7–8, but only 1% of the conversion occurred on this land Because they were con-verted in proportion to their occurrence on the landscape, this indicates that LCCs 1–6 are considered suitable for cropping and are at risk

of conversion, while LCCs 7–8 are generally unsuitable for cropping Protection of all land in LCCs 1 through 4 would reduce emissions by 85,167,883 tCO2e over 20 years, and protection

of all land in LCCs 5 and 6 would reduce emis-sions by an additional 16,148,568 tCO2e over

20 years

DISCUSSION

Our model results demonstrate that protecting the carbon stored in grasslands can meaningfully

Table 2 Difference between the conversion and

pro-tection scenario in metric tons of carbon dioxide

equivalents (CO2e) per ha over the 20-year project

period by source averaged across all soil types for

each county

County Belowground† Aboveground‡ N 2 O CH 4 §

North Dakota

Burleigh 47.64 1.024 7.969 2.03

Emmons 42.30 1.086 6.437 2.03

Kidder 44.97 0.983 7.366 2.03

McHenry 44.75 0.839 6.835 2.37

McIntosh 48.06 1.164 7.944 2.03

McLean 44.43 0.986 8.033 2.03

Sheridan 46.21 0.979 7.811 2.03

Stutsman 46.26 1.242 8.932 2.03

South Dakota

Edmunds 49.42 1.357 9.936 1.88

McPherson 47.15 1.204 8.451 2.47

† Includes belowground biomass and soil organic carbon.

‡ This value is a one-time loss over the 20-year project

period Values are negative because the annual aboveground

crop biomass production is greater than the annual native

mixed grass prairie production.

§ Values are negative because cattle are only present in the

avoided conversion scenario and the methane emissions must

be subtracted from the carbon emission savings.

Fig 1 Potential greenhouse gas emissions in metric tons of CO2e per hectare over a 20-year project period for all unprotected grassland across The Nature Conservancy (TNC) project area

contribute to reducing global GHG emissions.

This is important because the carbon stored in

these grasslands has a high risk of loss to the

atmosphere under current conversion rates

(Wright and Wimberly 2013, Lark et al 2015) By

targeting areas with the greatest potential for

GHG emissions if converted, protecting only 10%

of the currently unprotected grasslands would

avoid 11.7 tCO2e emissions, which is equivalent

to taking 2.5 million passenger cars off the road

for a year (EPA 2013a, b, FHWA 2013) and would

avoid a societal cost of $430 million (EPA 2013c)

Protection of grasslands in this project area would

also provide numerous other ecosystem services,

regionally and continentally Specifically, the PPR

supports the highest population densities of

breed-ing waterfowl in all of North America (Zimpfer

et al 2013), which is crucial for the over 90 million

people in the United States that engage in wildlife

recreation, including hunting (U.S Department of

the Interior 2011, Copper et al 2015), and supports

an annual industry of $145 billion (U.S

Depart-ment of the Interior 2011)

The CO2e emissions reported here reflect total

avoided GHG emissions if protection of at-risk

grassland occurs However, as in all systems,

there is leakage and uncertainty of protection that

needs to be accounted for To ensure that the

carbon offsets traded in the markets are not

over-estimated, the ACoGS methodology applies

deductions for leakage and a buffer pool

deduc-tion for uncertainty of protecdeduc-tion permanence

Under the ACoGS methodology, the offsets in

this project area would be subject to a 20%

mar-ket leakage deduction (Dell et al 2013) The

ACoGS methodology also requires 11% of the

credits to be placed in a buffer pool to mitigate

the risk of “reversals” in which sequestered

car-bon ends up being released, for example, if a

pro-tected grassland was illegally plowed (VCS

2010) Although ten percentage of offsets

previ-ously contributed to the buffer pool are returned

everyfive years if there are no reversals, the offset

numbers presented here conservatively ignore

any potential recovery of credits in the buffer

pool Combining these two adjustments, we

arrive at a 31% deduction to our model output,

which yields average per ha offsets of 35.6 tCO2e

over 20 years and a total of 9.4 million tCO2e

over 20 years for the 10% of hectares with the

highest emissions

Emission quantification is an essential step to support the development of carbon offset pro-jects as an economic incentive to avoid grassland conversion Currently, the primary tool for grass-land protection in the PPR is perpetual ease-ments The rising price of corn in recent years has increased land values, cropland rental rates, grassland conversion rates, and subsequently the cost of perpetual easements from $482/ha in 1998

to $1,922/ha in 2012 (Walker et al 2013) This ris-ing cost of easements highlights the need for additional funding sources, such as carbon off-sets Financing carbon projects also requires the use of capital that is donated or provided with

no or low interest rates because the payments for perpetual easements must be made up front, while the carbon offsets are issued over ensuing decades (periodically, after each verification) Once certified, carbon offsets can be sold on the voluntary market The average offset in 2014 sold for about $3.8/t (Forest Trends Ecosystem Marketplace 2015), although avoided conversion projects with significant co-benefits can fetch sub-stantially higher prices Using the average carbon offsets modeled here brings this to $137/ha for a 20-year period, which is not enough to cover easement costs that as of 2012 are over $1,235/ha

in both North and South Dakota (Walker et al 2013)

Carbon offsets need not be enough on their own

to secure grasslands, but can be supplemental Additional lands are protected from conversion when carbon offset funding partners with existing conservation easement efforts by providing sup-plementary funding that (1) targets easements to those properties most at risk of conversion and highest in carbon offset potential and (2) generates new funding that is used for additional grassland protection Such an approach could bring in millions of dollars of new funding for conserva-tion easements, targeted to those grasslands most

at risk of conversion and with the greatest poten-tial for reducing GHG emissions The spapoten-tially mapped and modeled carbon emissions reported here provide the information necessary to embark

on such a supplemental funding program for grassland easements Carbon offsets from avoided conversion can help protect at-risk grasslands, reduce GHG emissions, and produce positive out-comes for biodiversity and ranchers in a manner similar to REDD+ projects in forested ecosystems

We thank the Paulson Family Fund for providing

funding to complete this work and the USDA

Conser-vation InnoConser-vation Grant titled “Ducks Unlimited

Avoided Grassland Conversion Carbon Project” for

funding the development of the methodology that

directed the GHG emission quantification We thank

Ducks Unlimited for providing joint funding along

with The Nature Conservancy for the DAYCENT

model validation Finally, we thank Robin Kelly for

assistance throughout the analysis process and

Wil-liam Gascoigne for constructive comments on the

manuscript

LITERATURECITED

Alexandratos, N., and J Bruinsma 2012 World

agri-culture towards 2030/2050: the 2012 revision ESA

Working Paper No 12-03 FAO, Rome, Italy

Burke, I C., C M Yonker, W J Parton, C V Cole,

K Flach, and D S Schimel 1989 Texture, climate,

and cultivation effects on soil organic matter

con-text in U.S grassland soils Soil Science Society of

America Journal 53:800–805

Copper, C., L Larson, A Dayer, R Stedman, and

D Decker 2015 Are wildlife recreationists

conser-vationists? Linking hunting, birdwatching, and

pro-environmental behavior Journal of Wildlife

Management 79:446–457

Davidson, E A., and I L Ackerman 1993 Changes in

soil carbon inventories following cultivation of

previously untilled soils Biogeochemistry 20:

161–193

Del Grosso, S J., H T Gollany, and M Reyes-Fox

2016 Simulating soil organic carbon stock changes

in agroecosystems using CQESTER, DAYCENT,

and IPCC Tier 1 methods Advances in

Agricul-tural System Modeling 6:89–110

Del Grosso, S J., W J Parton, D S Ojima, C A

Keough, T H Riley, and A R Mosier 2008a

DAY-CENT simulated effects of land use and climate on

county level N loss vectors in the USA Pages 571–

595 in R F Follett and J L Hatfield, editors

Nitro-gen in the environment: sources, problems, and

management Second edition Elsevier Science,

Dordrecht, The Netherlands

Del Grosso, S J., T Wirth, S M Ogle, and W J Parton

2008b Estimating agricultural nitrous oxide

emis-sions EOS, Transactions, American Geophysical

Union 89:EO510001

Dell, R., M A Ahlering, J Fargione, P Wiesberg,

D Diaz, A Rood, S DeGryze, and B D Sharma

2013 Methodology for avoided conversion of

grassland and shrublands to crop production

Version 1.0 American Carbon Registry, Arlington, Virginia, USA

Doherty, K E., J S Evans, J Walker, J H Devries, and

D W Howerter 2015 Building the foundation for international conservation planning for breeding ducks across the U.S and Canadian border PLoS ONE 10:e0116735

Environmental Protection Agency (EPA) 2013a Inven-tory of U.S greenhouse gas emissions and sinks: 1990–2011 Annex 6 (Additional Information), Table A-275 U.S EPA #430-R-13-001 U.S Environmen-tal Protection Agency, Washington, D.C., USA Environmental Protection Agency (EPA) 2013b Inven-tory of U.S greenhouse gas emissions and sinks: 1990–2011 Chapter 3 (Energy), Tables 3-12, 3-13, and 3-14 U.S EPA #430-R-13-001 U.S Environ-mental Protection Agency, Washington, D.C., USA Environmental Protection Agency (EPA) 2013c Tech-nical update of the social cost of carbon for regula-tory impact analysis United States Government, Washington, D.C., USA

Estell, R E., K M Havstad, A F Cibils, E L Fredrickson, D M Anderson, T S Schrader, and

D K James 2012 Increasing shrub use by live-stock in a world with less grass Rangeland Ecol-ogy & Management 65:553–562

Federal Highway Administration (FHWA) 2013 High-way statistics 2011 Table VM-1 Office of HighHigh-way Policy Information, Federal Highway Administra-tion, Washington, D.C., USA

Foley, J A., et al 2011 Solutions for a cultivated planet Nature 478:337–342

Forest Trends Ecosystem Marketplace 2015 State of the voluntary carbon market 2015 Forest Trends Ecosystem Marketplace, Washington, D.C., USA Gascoigne, W R., D Hoag, L Koontz, B A Tangen,

T L Shaffer, and R A Gleason 2011 Valuing ecosystem and economic services across land-use scenarios in the Prairie Pothole Region of the Dakotas, USA Ecological Economics 70:1715–1725 Gibbs, H K., S Brown, J O Niles, and J A Foley

2007 Monitoring and estimating tropical forest car-bon stocks: making REDD a reality Environmental Research Letters 2:045023

Gleason, R A., N H Euliss Jr., R L McDougal, K E Kermes, and E N Steadman 2005 Potential of restored prairie wetlands in the glaciated North American prairie to sequester atmospheric carbon Paper 92 USGS Northern Prairie Wildlife Research Center, Jamestown, North Dakota, USA

Gleason, R A., M K Laubhan, and N H Euliss Jr

2008 Ecosystem services derived from wetland conservation practices in the United States Prairie Pothole Region with an emphasis on the U.S Department of Agriculture Conservation Reserve

and Wetlands Reserve Programs U.S Department

of the Interior, U.S Geological Survey, Reston,

Virginia, USA

Haas, H J., and C E Evans 1957 Nitrogen and

car-bon changes in Great Plains soils as influenced by

cropping and soil treatments Technical Bulletin

1164 USDA, Washington, D.C., USA

Hartman, M D., E R Merchant, W J Parton, M P

Gutmann, S M Lutz, and S A Williams 2011

Impact of historical land-use changes on

green-house gas exchange in the U.S Great Plains,

1883-2003 Ecological Applications 21:1105–1119

Hongmin, D., J Mangino, T A McAllister, J L

Hatfield, D E Johnson, K R Lassey, M Aparecida

de Lima, and A Romanovskaya 2006 Emissions

from livestock and manure management Pages

10.1–10.87 in H S Eggleston, L Buendia, K Miwa,

T Ngara, and K Tanabe, editors Guidelines for

national greenhouse gas inventories Volume 4:

Agriculture, forestry and other land uses

Intergov-ernmental Panel on Climate Change, IGES,

Hayama, Japan

Johnson, J A., C F Runge, B Senauer, J Foley, and

S Polasky 2014 Global agriculture and carbon

trade-offs Proceedings of the National Academy of

Sciences USA 111:12342–12347

Kittel, T G F., J A Royle, C Daly, N A Rosenbloom,

W P Gibson, H H Fisher, D S Schimel, L M

Berliner, and VEMAP2 Participants 1997 A

grid-ded historical (1895-1993) bioclimate dataset for

the conterminous United States Pages 219–222 in

Proceedings of the 10th Conference on Applied

Climatology, 20–24 October 1997, Reno, NV

Amer-ican Meteorological Society, Boston, Massachusetts,

USA

Lark, T J., J M Salmon, and H K Gibbs 2015

Crop-land expansion outpaces agricultural and biofuel

policies in the United States Environmental

Research Letters 10:044003

Manske, L L 2004 Generalized average stocking

rates North Dakota State University Extension,

Dickinson Research Extension Center, Dickinson,

North Dakota, USA http://www.ag.ndsu.edu/arc

hive/dickinso/research/2003/range03r.htm

Mousel, E M 2013 Rancher’s guide to managed

grazing in South Dakota Library of Congress

Cata-loging-in-Publication Data, USA http://issuu.com/

emousel/docs/ranchers_guide_to_grazing_v4

NCED 2011 National Conservation Easement

Data-base June 6, 2011 http://www.conservationeaseme

nt.us/

Olander, L P., H K Gibbs, M Steininger, J J

Swen-son, and B C Murray 2008 Reference scenarios

for deforestation and forest degradation in support

of REDD: a review of data and methods Environ-mental Research Letters 3:025011

Phillips, R L., M R Eken, and M S West 2015 Soil organic carbon beneath croplands and re-established grasslands in the North Dakota Prairie Pothole Region Environmental Management 55:1191–1199 Sala, O E., W J Parton, L A Joyce, and W K Lauenroth 1988 Primary production of the central grassland region of the United States Ecology 69:40–45

Samson, F B., F L Knopf, and W Ostlie 2004 Great Plains ecosystems: past, present and future Wild-life Society Bulletin 32:6–15

Saxton, K E., W J Rawls, J S Romberger, and R I Papendick 1986 Estimating generalized soil-water characteristics from texture Soil Science Society of America Journal 50:1031–1036

Searchinger, T., et al 2009 Fixing a critical climate accounting error Science 326:527–528

Stubbs, M 2014 Conservation Reserve Program (CRP): status and issues 7-5700 R42783 Congres-sional Research Service, Washington, D.C., USA Thornton, P E., S W Running, and M A White 1997 Generating surfaces of daily meteorological vari-ables over large regions of complex terrain Journal

of Hydrology 190:214–251

Thornton, P E., M M Thornton, B W Mayer,

N Wilhelmi, Y Wei, R Devarakonda, and R B Cook 2014 Daymet: daily surface weather data on a 1-km grid for North America, Version 2 Oak Ridge National Laboratory Distributed Active Archive Cen-ter, Oak Ridge, Tennessee, USA http://daac.ornl.gov United State Geological Survey Gap Analysis Program (USGS GAP) November 2012 Protected areas database of the United States (PADUS) Version 1.3 Combined Feature Class

United States Department of Agriculture National Agricultural Statistics Service Cropland Data Layer (USDA NASS CDL) 2014 Published crop-specific data layer USDA-NASS, Washington, D.C., USA http://nassgeodata.gmu.edu/CropScape/

U.S Department of the Interior, U.S Fish and Wildlife Service, U.S Department of Commerce, and U.S Census Bureau 2011 National survey of fishing, hunting and wildlife-associated recreation FHW/ 11-NAT (RV) Washington, D.C., USA

U.S Fish and Wildlife Service 2010 National Wet-lands Inventory U.S Department of the Interior, Fish and Wildlife Service, Washington, D.C., USA http://www.fws.gov/wetlands/

Verified Carbon Standard (VCS) 2010 AFOLU Non-Permanence Risk Tool: VCS Version 3 http://www v-c-s.org/sites/v-c-s.org/files/AFOLU%20Non-Per manence%20Risk%20Tool,%20v3.2.pdf