Climate Change and Managed Ecosystems - Chapter 8 pot

Periods of initial growth, regrowth, grazing, dormancy, and transition between dor-mancy and growth could be identified from daily net CO2 exchange in both seedling and first production

Dioxide Exchange over a Temperate, Short-Season Grassland: Transition

from Cereal to

Perennial Forage

V.S Baron, D.G Young, W.A Dugas, P.C

Mielnick, C La Bine, R.H Skinner, and J Casson

CONTENTS

8.1 Introduction 164

8.2 Methods 165

8.2.1 Net Ecosystem CO2 Exchange 165

8.2.2 Soil CO2 Flux 166

8.3 Results and Discussion 166

8.3.1 Climate 166

8.3.2 BREB CO2 Flux 166

8.3.2.1 Initial Growth Period 167

8.3.2.2 Regrowth 168

8.3.2.3 Grazing 168

8.3.2.4 Dormant Period 168

8.3.2.5 Spring 168

8.3.2.6 Diurnal CO2 Flux 169

8.3.3 Soil Respiration 170

8.3.4 Ecosystem Sink or Source 172

8.4 Conclusion 172

Acknowledgment 173

References 173

8.1 INTRODUCTION

North American grasslands may be part of a terrestrial carbon (C) sink.1–3 The terrestrial or “missing” sink, which includes cropland and forests, may play a role

in offsetting CO2 emissions generated from fossil fuel use and global changes in land management (e.g., deforestation) through C sequestration in soils, vegetation, and residues.1,2,4 Fan et al.5 estimated the northern boundary of the terrestrial sink

to be approximately 51° N lat

The potential sink size of grasslands may be significant because they cover a large area.4 World grasslands represent 32% of global vegetation on an area basis.1

Estimated area of U.S pasture and rangeland is 51 and 161 million ha, respectively, compared to cropland at 155 million ha Area of public grazing land in the U.S is

124 million ha.4 Canadian pastureland was estimated at 10 million ha, of which 9 million ha is located in western Canada.6–9 Entz et al.10 estimated an area of 44 million ha of range in the Northern Great Plains region of Manitoba, Saskatchewan, Alberta, North Dakota, South Dakota, and Montana There are approximately 6.5 million ha of native rangeland mostly south of 52° N lat in the semi-arid region of the Prairie Provinces.11

Rangelands have relatively low potential C sequestration rates per hectare, but cover large areas of the North American continent and, thus, could act as a large potential C sink Improved pastures have larger potential rates of C sequestration as they are located in humid and subhumid regions and receive greater management inputs; most are managed below production potential,4,12 but are also subject to greater removal of biomass-C as conserved ruminant feed

Most pastureland in the U.S is located east of 98° W long.,12 where annual precipitation balances or exceeds evapotranspiration.13 In Canada, the majority of pasture is located on the black and gray wooded soils in the subhumid Aspen Parkland and Boreal Transition zone of Western Canada8 and in all regions of eastern Canada.6–9

Alberta contains the largest area of pasture of the Prairie Provinces at 2.2 million ha

as well as 6.6 million ha of rangeland consisting of naturalized and native species.14

Micrometeorological studies carried out by the USDA-Agricultural Research Ser-vice Rangeland Carbon Dioxide Project15 documented ecosystem CO2 balance for various grasslands.16–23 These studies provided ecological insights into fundamental processes that affect C sequestration All CO2 uptake occurs as a result of photosyn-thesis during the growing season Uptake is countered by respiration, resulting in CO2 emission from the crop canopy, root, and soil microbial degradation of litter and crop organic residues.24,25 Bremer et al.25 estimated that CO2 respired from a tall-grass prairie ecosystem in Kansas was three to four times that accumulated in new biomass during the season; a fraction of the CO2 respired by the system is re-assimilated

Most of this research, involving aspects of grassland CO2 flux, was conducted during the growing season in the Great Plains region While several short-term studies indicate that grassland ecosystems act as small sinks, there are only a few studies that evaluated ecosystems annually Climate, season, species, phenology, and management (e.g., grazing, fire, etc.) influence both CO2 uptake and ecosystem respiration.16,22,24–26

In particular, drought20,24 and the dormant season19,20 are periods of net CO2 loss Thus,

in areas with short growing seasons, intermittent drought, and long dormant periods such as the Aspen Parkland region,27 C sequestration may be limited

Conversion of cropland to grassland and intensification of grassland management are key methods to enhance rate of C sequestration in soils.4,12 Studies about the ecosystem CO2 exchange process are needed to assess the Aspen Parkland region for potential as a C sink Our objective was to evaluate net ecosystem CO2 exchange, within season and annually, during the establishment or seedling year and during the first production year of a forage stand in transition from a cereal to forage sequence to be used for pasture in a mixed farm crop rotation

8.2 METHODS

Net CO2 ecosystem exchange measurements were made on a black chernozemic sandy loam soil in transition from cereal to perennial pasture, at Lacombe, Alberta, Canada (52° 26′ N: 113° 45′ W) The site had been in a cereal-forage rotation over the past 20 years Farming practices prevalent in the region were used — silage and hay were removed in the seedling and first production years and grazing of the forage species mixture occurred in September or October For the seedling year the

field was sown (May 15, 2002) to barley (Hordeum vulgare L.), a nurse crop, and under sown to a mixture of meadow bromegrass (Bromus riparius Rehm.) and alfalfa (Medicaigo sativa L.) The nurse crop was removed as silage on August 1, 2002 In

2002, the seedling forage stand was allowed to regrow until fall when it was grazed severely (4.2 Animal Unit Months, AUM, over 19 days) between September 27 and October 15 For the first production year, in 2003, the field was harvested as hay on July 16, allowed to regrow, and was grazed lightly (0.63 AUM over 14 days) between September 2 and September 15 Hereafter reference to heavy and light grazing are stocking rates of 4.2 and 0.63 AUM ha–1, respectively, over 19 and 14 days during the seedling and first production years, respectively Each year 100 kg ha–1 N was applied in the spring

A tower with Bowen ratio/energy balance (BREB) instrumentation (Model 023/CO2 Bowen ratio system, Campbell Scientific, Inc., Logan, UT) was placed on a 2% northwest-facing slope after seeding in spring 2002 Borders of the field provided

a minimum 200-m fetch from the tower Bowen ratios were calculated as described previously.16,18,19,28 Measurements began on May 26, 2002 When the BREB method was not valid for calculating turbulent diffusivity, because of differences in the sign

of the sensible heat flux and the temperature/humidity gradient, it was calculated using wind speed, atmospheric stability, and canopy height.16 This occurred 12% of the time Daily net CO2 flux measurements between May 15 (seeding date) and May

26 were estimated from the average daily flux for the first week of measurements Carbon dioxide and water vapor concentration gradients were measured at 1.4 and 2.4 m above the soil surface with an infrared gas analyzer (Model 6262, LICOR, Inc., Lincoln, NE) Methods were described previously by Frank and Dugas18 and Frank et al.19

The seedling year includes year-round CO2 exchange data from May 15, 2002

to May 14, 2003 and the first production year includes data from May 15, 2003 to May 14, 2004

Soil respiration measurements were made with a vented closed system using a LICOR 6200 portable infrared gas analyzer fitted with a soil respiration chamber (LICOR 6000-09) with a volume of 1000.3 cm3 and a diameter of 10.3 cm allowing

an area of exposed soil of 71.5 cm2 Rings were inserted 25 mm into bare soil in three positions around the BREB tower and were kept free of vegetation Flux measurements were made by placing the chamber over the rings for periods of 3 to

5 min Measurements were made at approximately 2-week intervals during the growing season, spring and fall from 1100 to 1300 hours, but not in winter BREB measurements are presented as a daily total of net CO2 flux(g m2/d–1) Soil respiration measurements are presented in the same units, but are averaged over

at least three subsamples Net CO2 uptake into the ecosystem is indicated by a positive sign and net efflux by a negative sign

Linear regression was used to relate midday soil respiration (independent variable)

to BREB nighttime CO2 flux (dependent variable) A significant linear regression

coefficient (P ≤ 0.05) was indicative of a relationship between the fluxes of different

origins The coefficient of determination (R2) was used to quantify the extent to which the variation in BREB nighttime CO2 flux was explained by soil respiration

8.3 RESULTS AND DISCUSSION

The seedling year season could be described as a dry spring and summer with a moist fall (Figure 8.1); April to October precipitation was 277 mm and the long-term average for this period was 363 mm The first production year season had a wet spring with a dry summer; April to October precipitation was 262 mm Tem-peratures during the June to August period were generally above average Mean winter temperatures (October to March) were above average for the seedling year and average to below average for the first production year Spring time air temper-atures were average

Periods of initial growth, regrowth, grazing, dormancy, and transition between dor-mancy and growth could be identified from daily net CO2 exchange in both seedling and first production years (Figure 8.1) From May 15 to September 30 (entire growth period), net uptake occurred on only 65 and 61% of the 140 days for seedling and first production years, respectively (Figure 8.1A and B) Daily flux values were highly variable due to effects of variation in radiation, temperature, and precipitation

on plant and soil processes Occasional spikes of soil CO2 efflux often occurred immediately after rainfall events that flush CO2 from soil pores.24,29

8.3.2.1 Initial Growth Period

For both seedling and first production year, maximum daily BREB CO2 flux coin-cided with periods of initial growth in June and July (Figure 8.1A and B) From May 15 until Aug 1, 2002 seedling year average daily net flux was 6.7 g m2 d–1

FIGURE 8.1 Net daily CO2 flux and precipitation for (A) May 15, 2002 to May 14, 2003 (seedling year) and (B) May 15, 2003 to May 14, 2004 (first production year) for meadow bromegrass–alfalfa.

A

B

Seedling Year 2002-03

-2 d

-2 d

40

30

20

10

0

-10

-20

-30

-40

30

20

10

0

-10

-20

-30

-40

100

90

80

70

60

50

40

30

20

10

0

90

80

70

60

50

40

30

20

10

0

1361 66 1962 26 2562 86 3163 46 10 40 70 1001 30

May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May

First Production Year 2003-04

and from May 15 until July 16, 2003 the average flux in the first production year was 1.7 g m2 d–1

8.3.2.2 Regrowth

Harvest was followed immediately by a period of net CO2 loss that lasted 25 days

in the seedling year and 14 days in the first production year McGinn and King30

observed 12 days of ecosystem respiration after cutting alfalfa during midsummer

in Ontario Mean daily CO2 flux during these periods was –4.9 and –5.5 g m2 d–1

during seedling and first production year, respectively

Once a plant canopy was reestablished, net CO2 uptake into the ecosystem occurred well into September of both years In the seedling year, August and September rainfall resulted in regrowth of forage, but the perennial stand had to establish a canopy after the barley was removed in spite of dry early summer conditions Regrowth was slow in the first production year due to drought, but generally net CO2 uptake occurred (Figure 8.1B) Average net CO2 flux after harvest until September 30 was –0.75 and 0.41 g m2 d–1 during seedling and first production years, respectively Kim et al.24 observed net CO2 release at a rate of –3 g m2 d–1

from a tall-grass prairie ecosystem in Kansas during drought conditions

8.3.2.3 Grazing

A season-end grazing removed most of the live vegetative material from the seedling year stand during early October A very light grazing in September of the first production year left residual biomass for light interception During hard grazing (September 27 to October 30) in the seedling year an average net CO2 release of –7.9 g m2 d–1 was observed During the light grazing (August 27 to September 15)

in first production year an average net CO2 loss of –0.8 g m2 d–1 was observed Fall residual dry matter was less than 50 and approximately 300 g m–2 in seedling and first production years, respectively

8.3.2.4 Dormant Period

Almost no growth occurred during the dormant period, October 1 to March 31, although 6 and 17 days of net uptake occurred in the seedling and first production years, respectively, mostly in October Over the October 1 to March 31 period, average daily CO2 release was –3.6 and –1.6 g m2 d–1 during seedling and first production years, respectively Relatively large losses from the system occurred during October, November, and March of the seedling year Over the entire dormant period the range in CO2 flux was 1.5 to –14.7 g m2 d–1 in the seedling year (Figure 8.1A) and 7.2 to –8.2 g m2 d–1 in the first production year (Figure 8.1B)

8.3.2.5 Spring

The daily net CO2 loss from the ecosystem in March of the seedling year (Figure 8.1A) lasting until late May 2003 (Figure 8.1B) was large From April 1 until May

15 there were only 4 days of net uptake This is partly due to the lack of leaf area

as a result of heavy grazing the previous fall Thus, net CO2 loss occurred from respiration required to generate new leaf material just as it did after cutting Pasture dry matter yields on April 18, May 1, and May 15 were 0.0, 100, and 500 kg ha–1, respectively, in 2003 (seedling year) By contrast there was net uptake during 80%

of the days between April 1 and May 14, 2004 for the first production year (Figure 8.1B) Dry matter yield on May 15, 2004 was 1600 kg ha–1 The CO2 flux averaged –12.7 and 3.3 g m2 d–1 from April 1 to May 14 during the seedling year (2003) and first production year (2004), respectively

Diurnal CO2 fluxes were representative of the contrasting periods within years (Figure 8.2) The 20 min average CO2 fluxes averaged over a 14-day period for winter (Figure 8.2A) showed small CO2 flux loss during day and night in early January for both seedling and first production years During spring (Figure 8.2B),

CO2 fluxes were negative during day and night for the seedling year, reflecting high soil respiration rates (Table 8.1) and lack of a crop canopy, following intense grazing during the fall of 2002 By contrast, spring diurnal CO2 fluxes during the first production year resembled those of the initial growth period (Figure 8.2C), although the peak values were smaller For the examples given (Figure 8.2B and C), average daytime CO2 flux for the first production year-spring was approximately 39% of the

FIGURE 8.2 The 20-minute average CO2 flux for 24-h periods of seedling and first produc-tion year (A) Winter, averaged over 14 days, (B) spring, averaged over 14 days, (C) initial growth, averaged over 7 days prior to harvest, and (D) regrowth, averaged over 7 days after harvest.

A

0.8

0.6

0.4

0.2

0.0

–0.2

–0.4

–0.6

–0.8

–2 s

B

0.8

0.6

0.4

0.2

0.0

–0.2

–0.4

–0.6

–0.8

C 0.8 0.6 0.4 0.2 0.0 –0.2 –0.4 –0.6 –0.8

–2 s

D 0.8 0.6 0.4 0.2 0.0 –0.2 –0.4 –0.6 –0.8

Seedling Year Production Year

Winter, Jan 2–15 1 wk Preharvest Initial Growth

Spring, April 1–14 1 wk Postharvest Regrowth

0 600 1200 1800 2400

Time of Day (MST – h)

0 600 1200 1800 2400

Time of Day (MST – h)

average of both years’ initial growth daytime flux; spring nighttime flux (Figure 8.2B) was 53% of the nighttime CO2 flux for the initial growth period (Figure 8.2C) Diurnal CO2 flux for the initial growing period (averaged over 7 days prior to harvest) were comparable to those from other studies (e.g., Frank et al.19 and Sims and Bradford20) for periods of active growth Diurnal CO2 flux for the regrowth period (averaged over a 7-day period immediately after harvest) showed negative daytime and nighttime flux for the seedling year (Figure 8.2C) Regrowth daytime CO2 flux for the first production year averaged near zero The low daytime CO2 flux for both years reflected a small leaf area during regrowth However, average nighttime regrowth CO2 flux for both years was approximately 82% of those observed during initial growth, explaining the tendency for relatively large net losses from the eco-system during regrowth periods of both years A decline in afternoon CO2 flux rates for seedling and first production year initial growth (Figure 8.2C) and first production year regrowth (Figure 8.2D) was likely due to tissue water deficit, resulting in stomatal closure, as suggested in a similar example by Sims and Bradford.20 The growing seasons used in this study were drier than normal

Our measurements of respiration are the sum of root and soil microbial respiration.25

During the night, the BREB system measures aboveground (including canopy and litter) and belowground respiration As expected, nighttime BREB CO2 flux was highest during periods of maximum plant growth Averaged over initial growth periods BREB nighttime respiration averaged –4.96 and –8.58 CO2 flux g m2 d–1 for seedling (May 15 to August 1) and first production years (May 15 to July 16), respectively, compared to –2.5 and –2.0 g m2 d–1, respectively, during dormant periods

TABLE 8.1

2002 to 2004 from Seedling Year and First Production Year Stands of Meadow Bromegrass–Alfalfa Mixtures at Lacombe, Alberta

2002

2003

2004

Mean soil CO2 flux varied depending upon on time of year and stage of crop development (Table 8.1) Soil CO2 flux during the growing season (May 16 to September 30) was approximately three times greater than the dormant season (November 1 to March 31), but growing season fluxes were more variable Spring

2003 (April 1 to September 15) soil respiration losses were highest, supporting net BREB losses for the seedling year during April and early May (Figure 8.1A) Losses during spring 2004 (first production year) were about half the CO2 flux values of spring 2003 Similar results were shown by De Jong et al.31 in southern Saskatchewan, who concluded that high soil respiration rates following a drought were due to wetting and drying cycles, which stimulated soil microflora to accelerate respiration rate Dormant season daily mean soil CO2 flux were in agreement with Kim et al.24 Frank and Dugas18 determined daily average soil flux to be –1.7 g CO2

m–2 from measurements made throughout fall and winter in North Dakota Linear regression analyses between soil respiration CO2 flux rates and BREB night time fluxes indicated that soil CO2 fluxes did not predict BREB nighttime fluxes well (Table 8.2) Frank et al.22 found that a single flux rate taken at midday overestimated the average of five sequential measurements taken at 3-hour intervals

by 9% However, they22 determined that single, midday soil CO2 flux rates were most representative of daily soil flux CO2 compared to soil flux rates taken at other times of the day

TABLE 8.2

Linear Regression Relationship between Midday Soil (independent variable)

Bromegrass–Alfalfa Stand during Seedling Year and First Production Year Combined at Lacombe, Alberta

Entire Growing Season (May 15 to Sept 30)

Initial Growth Period c

Re-growth Period c

Entire Dormant Period (Oct 1 to March 31)

Spring Period (April 1 to May 15)

a Probability of a significant regression.

b RMSE is root of the mean square error of the linear regression.

c Initial growth period for 2002–2003 is May 15 to Aug 1 and for 2003–2004 is May 15 to July 16; regrowth period for 2002–2003 is Aug 2 to Sept 30 and 2003–2004 is 17 July 17 to Sept 30.

During the growing season, soil respiration did not explain variation in ecosystem

BREB respiration well, as indicated by low R2 values and regression equations with low probability of significance (Table 8.2) A large proportion of ecosystem respiration

is derived from canopy dark respiration when dry matter yield is close to maximum.32

However, during the regrowth phase, which included periods of net loss after harvest (Figure 8.1A and B) and drought, soil respiration explained 46% of the variation in BREB nighttime CO2 flux During the dormant period (essentially October and November measurements) soil CO2 flux explained 56% of the BREB nighttime CO2 flux During the spring period there was a transition from dormancy to growth; both plant and soil microorganism metabolism should be high, but the crop leaf area is relatively low compared to mid-June Over the spring period soil respiration explained 39% of the variability in BREB nighttime respiration (Table 8.2) Soil CO2 flux should underestimate CO2 flux from ecosystem respiration as it does not account for litter decomposition and dark respiration of live vegetative material

During the seedling year there was an average annual net daily CO2 flux of –2.0 g m–2

d–1for a net loss from the ecosystem of 730 g CO2 m–2 yr–1 During first production year the average annual net daily CO2 flux was –0.01 g m–2 d–1for a net annual loss of 3.65

g CO2 m–2 yr–1 Year to year variation in annual CO2 flux is likely to be high Long-term annual ecosystem equilibrium should be bounded by a standard deviation of CO2 flux In most cases enough years of data have not been collected to determine what this statistical boundary might be Studies such as the current one are often not replicated

or numbers of replicates are very small Statistical rigor improves with the number of years involved in the study This may not be economically possible, so it is important

to be aware that small annual net losses might be indicative of an equilibrium state Dugas et al.16 estimated that annual fluxes of 183 to 293 g CO2 m–2 yr–1 at a Temple,

TX site as in “approximate equilibrium.” Sims and Singh33 as cited by Frank et al.19

estimated that the CO2 budget of native grasslands should be near equilibrium Losses in the current study indicate that on the basis of annual BREB CO2 flux data that the ecosystem acted as a CO2 source during the seedling year and that it was close to equilibrium during the first production year However, approximately 530 g

m–2 yr–1 of dry matter was removed as silage and during grazing in the seedling year and approximately 450 g m–2 yr–1 was removed in the first production year, equating

to an additional 805 and 684 g CO2 m–2 of C loss from the ecosystem besides that determined by BREB CO2 flux measurements Thus, in both years the forage stand acted as a CO2 source The removal of C from the ecosystem as harvested dry matter does not complete the C accounting because return of materials as manure after feeding

or as feces during grazing may bring a fraction of the original crop-C back to the ecosystem

8.4 CONCLUSION

More years of annual measurement of CO2 flux are required to determine precise patterns for CO2-dynamics on cropland pasture in mixed cropping systems This