- Email: [email protected]
Potential tree and soil carbon storage in a major historical floodplain forest with disrupted ecological function
Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
Contents lists available at ScienceDirect
Perspectives in Plant Ecology, Evolution and Systematics journal homepage: www.elsevier.com/locate/ppees
Research article
Potential tree and soil carbon storage in a major historical floodplain forest with disrupted ecological function Brice B. Hanberry a,∗ , John M. Kabrick b , Hong S. He a a b
Department of Forestry, University of Missouri, 203 Natural Resources Building, Columbia, MO 65211, USA USDA Forest Service, Northern Research Station, University of Missouri, 202 Natural Resources Building, Columbia, MO 65211, USA
a r t i c l e
i n f o
Article history: Received 16 April 2014 Received in revised form 11 December 2014 Accepted 18 December 2014 Available online 26 December 2014 Keywords: Agriculture Climate change Ecosystem services Flooding Reforestation Restoration
a b s t r a c t Floodplain forests are extremely productive for agriculture and historical floodplain forests have been converted to prime agricultural land throughout the world, resulting in disruption of ecosystem functioning. Given that flooding may increase with climate change and reforestation will increase resiliency to climate change, we tested whether reforested floodplains also have great potential to store carbon and the effects of even modest increases in forested acreage on carbon storage. To calculate potential aboveground biomass in the Lower Mississippi River Alluvial Valley (LMAV) of the United States, we determined current and historical tree biomass used density estimates and diameter distributions from tree surveys and relationships between diameter and biomass from current forests. To calculate potential soil organic carbon if the landscape was forested, we used soil organic matter from soil surveys of the agricultural landscape, and multiplied the carbon by a factors of 1.25, 1.5, and 1.75 based on published reports of soil carbon increases due to afforestation. Our results showed that area-weighted mean biomass density (trees ≥12.7 cm in diameter) for historical forests was 300 Mg/ha, ranging from 228 Mg/ha to 332 Mg/ha by ecological subsection, based on the most conservative diameter distribution. Mean biomass density for current forests was 97 Mg/ha, ranging from 92 Mg/ha to 111 Mg/ha. Mean carbon density for agricultural soils was 96 Mg/ha, whereas combined tree and soil carbon densities varied from 169 Mg/ha to 317 Mg/ha; soil carbon accounted for 0.5–0.7 of total carbon density. Historical forested carbon storage in the Missouri LMAV was about 234 TgC, with the most conservative diameter distribution and assuming 80% forest coverage. Current forested carbon storage in the Missouri LMAV is about 2% of historical storage, at 5 TgC in 30,000 ha of forests, but may reach 23 TgC if forested extent almost triples, with the addition of 50,000 ha of marginal agricultural land, and carbon storage increases in trees and soil. The entire LMAV currently stores 97 TgC in forests and reasonable carbon storage for the entire LMAV may be about 335 TgC, based on increased carbon storage and reforestation of 600,000 ha of marginal agricultural land, which would double the current forested extent. Although 335 TgC storage for the LMAV is only about 1.5 times greater than historical carbon storage of the Missouri LMAV, doubling the forested extent will increase other ecosystem functions, including carbon storage, flood abatement, and reduction of fertilizer pollution in the Gulf of Mexico. © 2014 Geobotanisches Institut ETH, Stiftung Ruebel. Published by Elsevier GmbH. All rights reserved.
Introduction Floodplain forests have been converted to agriculture and other land uses globally (Gore and Shields, 1995; Zedler and Kercher, 2005). Accompanied by reduction in floodplain forests, loss of ecological function interrupts ecosystem services including flood abatement, biodiversity support, water quality improvement, and
∗ Corresponding author. Tel.: +1 5738755341. E-mail address: [email protected] (B.B. Hanberry).
carbon management (Zedler and Kercher, 2005). Both carbon and nutrient capture (i.e., sequestration) and retention (i.e., storage) are limited without the presence of long term vegetation ground cover. For example, agricultural fertilizers move through floodplains to coastal waters, creating dead zones as algal blooms deplete oxygen (Turner and Rabalais, 2003; Zedler and Kercher, 2005). Increased long term biomass through reforestation of floodplain forests will contribute to carbon and nutrient storage along with other ecosystem services. Carbon storage currently is unrealized in historical floodplain forests. The Mississippi River Basin drains six major watersheds
http://dx.doi.org/10.1016/j.ppees.2014.12.002 1433-8319/© 2014 Geobotanisches Institut ETH, Stiftung Ruebel. Published by Elsevier GmbH. All rights reserved.
18
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
Heavy precipitation events will disrupt current infrastructure in place to control rivers, resulting in frequent flooding on agricultural lands on historical floodplains (Pryor et al., 2014). Ecosystem services of flood and erosion control and carbon storage provided by floodplain forests may become more valuable than agricultural lands vulnerable to flooding and that consequently will become less productive under climate change. Increased restoration of floodplain forests will increase resiliency to extreme events of climate change (Groffman et al., 2014). In light of increased flooding expected under climate change, reforestation for a variety of ecosystem services may become a more viable land use in former floodplain forests. Mature, unharvested forests in floodplain forests are rare at the stand scale and non-existent at a landscape scale. Potential carbon storage of mature floodplain forests therefore is unknown for the LMAV and important for evaluating to benefits of land use conversion. Our aim was to test the (1) potential for the LMAV to store carbon in the major carbon pools of aboveground tree biomass and belowground soil carbon and (2) effects of even modest increases in forested acreage on above- and belowground carbon storage. We used historical tree surveys from the LMAV in Missouri to quantify and map forest aboveground biomass for four ecological subsections (Ecomap, 1993; Figs. 1 and 2, mean area = 1.45 million ha, SD = 1.20 million ha). We then calculated forested carbon storage potential for the Missouri LMAV and generalized our results to the entire LMAV. Under land use change from forest to crop and potentially restoration back to forest, we provide tree and soil carbon accounting valuable for management scenarios in a 10.4 million ha landscape and a method for soil carbon accounting applicable to other regions.
Fig. 1. The Lower Mississippi River Valley: the four ecological subsections in Missouri (outlined in white) cross state boundaries to form the entire White and Black River Alluvial Plains ecological section (shaded dark gray). Forests (shaded black) are present in areas with greater elevation.
that cover about 40% of the continental United States. One of the watersheds, the Lower Mississippi River Alluvial Valley (LMAV; Fig. 1) once contained the greatest area of floodplain forests. By 1978, only 2 million ha (Schoenholtz et al., 2001) of the historical 10.4 million ha forested extent remained (areal extent excludes open water; Fry et al., 2011). Currently, cooperative partnerships are in place to replant trees in portions of the LMAV, primarily within marginal agricultural lands that have hydric soils with poor drainage (King and Keeland, 1999; Frey et al., 2010). In addition to unrealized carbon storage, floodplain forests may have greater potential to store carbon than other ecosystem types, particularly compared to surrounding upland forests in temperate zones (Suchenwirth et al., 2012). Deep alluvial soils are recognized for crop productivity, even though tree productivity has been less well-documented (Shoch et al., 2009). Although we are unaware of historical accounts of trees in the LMAV, directly north of LMAV, multiple trees within one ha had diameters of 2–5 m, exceeding current state record trees for species (in the Wabash River Valley; Jackson, 2006). Decisions about land use in floodplains may become increasingly important as climate changes (Brown et al., 2014; Groffman et al., 2014). Temperatures in the midwestern United States may increase 3–5 ◦ C by the end of the century (Pryor et al., 2014). Although longer growing seasons and rising carbon dioxide levels initially may increase yields of some crops, climate change eventually may decrease agricultural productivity due to wet springs, anomalous frosts, drought, and heat stress, particularly during pollination and reproductive development (Pryor et al., 2014).
Methods Study area Missouri’s LMAV landscape is about 1 million ha excluding open water, nearly 10% of the LMAV (Ecomap, 1993; Fry et al., 2011; Fig. 1). About 80% of land cover is cultivated crops, 4% is pasture or hay, 2.5% is forest, and 5% is forested wetlands. Crowley’s Ridge, one of four ecological subsections (Ecomap, 1993) in the Missouri LMAV, is a loess-covered upland that rises above flatter alluvial plains and accounts for most of the landscape’s pasture and forest. Pasture and forest comprise 24% and 14%, respectively, of land cover in Crowley’s Ridge in Missouri. Roughly 80% of soils in the Missouri LMAV have limitations, primarily (66% of total soil) due to poor drainage. About 130,000 ha are classified as not prime farmland (including land that needs to be protected from flooding; Soil Survey Geographic Database, Natural Resources Conservation Service, http://soildatamart.nrcs.usda.gov), of which about 47,900 ha is cultivated crops and 13,800 ha is forest. Tree surveys and density estimates The United States General Land Office (GLO) was established in 1812 to survey, map, and sell land for settlement. The GLO surveys divided area into square townships measuring 9.6 km × 9.6 km, which were divided further into 36–1.6 km2 sections. Surveyors selected two to four bearing trees at the intersection of section lines and midpoints between section corners. For each selected tree, surveyors recorded species, diameter, distance, and bearing to survey point. The GLO surveys contain bias because surveyors selected trees at survey points, resulting in non-random trees. Furthermore, selected trees were of moderate diameter, to increase longevity as section markers. We excluded trees with diameters
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
19
Fig. 2. Estimated current soil carbon density (Mg/ha) by ecological subsection in the predominantly agricultural Lower Mississippi River Valley (panel A). Potential increases in soil and tree carbon density (Mg/ha) by ecological subsection if agricultural lands are converted to forested lands and managed as young (panel B) or mature forests (panel C).
<12.7 cm to maintain a consistent diameter threshold. There were 25,343 trees of this size recorded during 1817–1860. The USDA Forest Service Forest Inventory and Analysis Program (FIA) conducts forest inventories in long term plots that are surveyed during a five year cycle. Because there were few plots in Missouri’s LMAV, we used plots from the entire White and Black River Alluvial Plains ecological section (Ecomap, 1993; Fig. 1), which crossed state boundaries. We used the latest cycles from Arkansas during 2006–2010 (63% of trees), Mississippi during 2009–2011 (20% of trees), and Missouri during 2004–2008 (8% of trees), as well as four other states (FIA DataMart, www.fia.fs.fed.us/tools-data). We selected trees with diameters ≥12.7 cm and selected plots that contained at least two trees and were 100% forestland, which FIA defines as land at least 0.4 ha in size and 37 m wide with at least 10% cover by live trees of any
size, “including land that formerly had such tree cover and that will continue to have forest use”. This selection included 6330 trees. In order to quantify historical biomass, we used simulations to estimate biomass by ecological subsection based on density estimates and diameter distributions (see section below for description of simulations; also Rhemtulla et al., 2009). Historical survey points contained only 2–4 trees per plot, so it was not possible to estimate biomass at each plot. For current biomass, the FIA program provides tree biomass based on allometric equations. Biomass (biomass of tree bole, stump, and top excluding foliage) can be summed per plot and expanded to a hectare. To make certain that simulations were accurate, we also simulated FIA biomass using the same method as for GLO biomass. We estimated current tree density by summation of the number of trees per plot and expanded to a per hectare basis.
20
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
We first estimated historical density for each ecological subsection (Fig. 1) with the Morista estimator (Morisita, 1957) for (1) survey points with two trees and (2) survey points with three trees and the nearest three trees for survey points with four trees (diameters ≥12.7 cm; Hanberry et al., 2011). We then produced a low and high value based on adjustment for potential spatial patterning (clustered or regular patterns; Hanberry et al., 2011). We corrected these values for surveyor bias. Surveyors potentially did not select the nearest trees to each survey point, which would produce a distance rank of 1. Selection of more distant trees will result in underestimated densities. We adjusted density estimates using a rank-based method to estimate a low value, assuming selected trees had a mean rank of 1.4, and a mean value, assuming selected trees had a mean rank of 1.8. Using a complementary method to correct for non-random ratios of quadrants and azimuth and differences between frequencies of species and diameter of trees selected at survey points and trees recorded along survey lines, we calculated density estimates for another mean value and a high value (Hanberry et al., 2012a). We then determined a weighted average of the two mean values from the two complementary methods and retained the low value from the rank-based method and high value from the bias method. Simulated biomass estimation We estimated biomass using simulations of densities and diameter distributions, similarly to Rhemtulla et al. (2009), for both GLO and FIA surveys. We developed parameters for probability distribution functions of historical and current tree diameters by ecological subsection truncated at 12.7 cm using lognormal (which describe one-storied, even-aged forests; Podlaski 2010) exponential distributions (for uneven-aged forests), Weibull and gamma distributions (for structurally heterogeneous forests; Proc Severity, SAS software, version 9.1, Cary, NC). We simulated 10,000 stands of one hectare using random diameters generated from diameter distribution parameters truncated at 12.7 cm and random densities generated from density estimates and uncertainty (R Runuran, R Development Core team, 2012; J. Leydold & W. Hörmann, http://statmath.wu.ac.at/unuran, http://cran.rproject.org/web/packages/Runuran/index.html). We used allometric equations by ecological subsection from FIA surveys based on a simple regression of the relationship between (the log transformation of) biomass and (the log transformation of) diameter to quantify the biomass of each tree. Because allometric equations were developed for trees grown in forests under current land use and management and not available for trees grown under different conditions, we maintained a constant allometric relationship for each subsection. We then summarized the biomass for each stand and determined mean biomass and standard deviation per hectare. We used mean biomass per ecological subsection from FIA surveys, calculated by summarizing and expanding biomass at FIA plots, to determine mean absolute difference between simulated current biomass for each distribution and mean biomass, thereby testing the accuracy of our simulations. We identified the distribution for historical biomass that had the least difference from simulated FIA biomass of the same distribution as the most conservative fit for comparisons. Nevertheless, we provided the maximum value for each ecological subsection from other distributions. Soil carbon storage We used soils surveys from Soil Survey Geographic (SSURGO) Database (Natural Resources Conservation Service,
http://soildatamart.nrcs.usda.gov) to calculate soil organic matter for each ecological subsection. We used a conversion factor of 0.58 to convert soil organic matter to soil organic matter (Guo and Gifford, 2002). Because the LMAV is predominantly agricultural, soil surveys were from predominantly agricultural soils. Based on literature of soil carbon accumulation from reforestation of agricultural land, soil carbon is approximately 1.5 times greater after afforestation (Guo and Gifford, 2002; Clark and Johnson 2011; Indorante et al., 2013). Furthermore, this soil carbon level appears to be maximized after time such that soils in older forests are not able to store any more carbon in the soil (Hoover et al., 2012). In addition to the adjustment factor of 1.5, we also produced a low estimate for newly restored young forests using an adjustment of 1.25 and a high estimate of 1.75 for soils capable of carrying greater carbon capacity (Guo and Gifford, 2002). Potential forested carbon storage The amount of forest land cover during the survey period of 1817–1860 is unknown. Only about 5% of the landscape currently is wetlands, but before drainage for agriculture, standing water may have reduced forested extent. Nevertheless, for 97% of the landscape, surveyors were able to record trees along the systematic survey grid. In addition to water and flooding, fire and wind disturbance also would decrease the forested extent. The LMAV was used by Native Americans (e.g., the Cahokia settlement near current East St. Louis), but by the 1800s, Native American population density had decreased. As for European Americans, at least while accessible upland forests were available and harvesting technology was less developed, wet soils deterred harvest of floodplain forests. Therefore, we provided a range of forested extents from 60% to 100% to represent the historical landscape along with percentage of current forested land and reforestation scenarios for marginal agricultural land, approximately two to three times current forested extent. We then calculated total forested carbon storage in combined trees and soil for young forests and mature forests, using a conversion factor of 0.5 to convert biomass to carbon. We extrapolated to the entire LMAV based on forested extent and a range of calculated carbon densities. Results To verify the accuracy of the forest biomass simulations made with different distributions, comparisons to FIA biomass estimates were made. Mean absolute difference between simulated biomass of current forests and mean biomass estimates from FIA plots by ecological subsection was 3.16 Mg/ha for the gamma distribution, 3.35 Mg/ha for the Weibull distribution, 4.42 Mg/ha exponential distribution, and 12.41 Mg/ha for the lognormal distribution. Mean absolute difference between simulated biomass estimates for historical forests and simulated biomass of current forests by ecological subsection was 184 Mg/ha for the Weibull distribution and 208 Mg/ha for the lognormal distribution. The other two distributions had differences of about 370 Mg/ha. To remain consistent in comparisons, we selected the overall most conservative distribution, the Weibull distribution. Area-weighted mean biomass for historical forests averaged 300 Mg/ha, ranging from 228 Mg/ha to 332 Mg/ha by ecological subsection (trees ≥12.7 cm and the Weibull distribution; Table 1). Area-weighted mean biomass for current forests averaged 97 Mg/ha, ranging from 92 Mg/ha to 111 Mg/ha by ecological subsection. Historical biomass was greater per ha than current biomass by a factor of three. Using the least conservative distribution for simulated historical biomass, historical biomass was greater per ha than current biomass by a factor of six (area-weighted).
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23 Table 1 Historical (GLO) and current (FIA) live aboveground biomass (Mg/ha; trees ≥12.7 cm), based on the Weibell distribution, and maximum mean values based on alternative distributions. Ecological subsection
Mississippi River Crowley’s Ridge White and Black Rivers St. Francis River
GLO
FIA
GLO
Table 3 Total forested carbon storage (TgC) for the Missouri LMAV and entire LMAV, varying by forested extent and carbon density (Mg/ha) of a range of forested carbon densities. % Forested
Mean
SD
Mean
SD
Max mean
SD
300 228 332 310
77 62 76 65
92 94 111 104
18 17 28 20
759 304 480 529
284 103 156 147
Area-weighted mean soil carbon storage of the current agricultural landscape of the entire LMAV averaged 96 Mg/ha, ranging from 71 Mg/ha to 223 Mg/ha by ecological subsection (Table 2; Fig. 2). Adjustment (of 1.5) for a forested cover produced an areaweighted mean soil carbon storage of 143 Mg/ha, ranging from 106 Mg/ha to 334 Mg/ha by ecological subsection. The low adjustment (of 1.25) area-weighted mean was 120 Mg/ha and the high adjustment was 167 Mg/ha. We combined results with aboveground biomass density, after conversion to carbon density. Young forests that have accrued little additional soil carbon or above-ground biomass have a mean combined soil and tree carbon density of 169 Mg/ha, ranging from 136 Mg/ha to 327 Mg/ha by ecological subsection (Table 2; Fig. 2). Mature forests with moderate increases in soil organic carbon have a mean combined carbon density of 293 Mg/ha, ranging from 221 Mg/ha to 484 Mg/ha by ecological subsection. In young forests, soil carbon represents 71% of forest carbon (ranging from 65% to 85%) whereas in mature forests with moderate organic soils, soil carbon represents 49% of forest carbon (ranging from 44% to 69%). The historical potential in the Missouri LMAV for forested carbon storage in trees and soil was approximately 234 TgC (1 Tg = 1,000,000 Mg). This estimate is based on 80% forested forest coverage, the most conservative simulated estimate of historical tree biomass, and mean forested soil values (Table 3). Current forested carbon storage was approximately 5 TgC, or 2% of historical carbon storage, based on 30,000 ha of forested land. The restoration potential from about 50,000 ha of marginal agricultural land currently in crops combined would result in total forested carbon storage of 23 TgC, an increase of about five-fold, after trees have matured and soil has developed moderate carbon capacity. We can extrapolate these results to the entire 10.4 million ha LMAV, which currently is about 5.5% forested, excluding woody wetlands that is larger proportion of the entire LMAV due to
21
Total carbon storage Reforested carbon density Low
Missouri LMAV 100 90 80 70 60 8 3 LMAV 100 90 80 70 60 17 11 6
Mean
High
169 152 135 118 101 14 5
293 264 234 205 176 23 9
317 285 254 222 190 25 10
1758 1582 1406 1230 1055 299 193 97
3047 2742 2438 2133 1828 518 335 168
3297 2967 2637 2308 1978 560 363 181
major rivers (about 22% of the landscape; Ecomap, 1993; Fry et al., 2011). The LMAV currently stores 97 TgC in forests. Given time for trees to accrue and store more biomass and expanding restoration efforts to add another 600,000 ha and double the forested extent to 1.15 million ha will increase storage during a century to 335 TgC. Historically, given 80% forested extent and assuming the most conservative diameter distribution and mean soil adjustment and that the non-forested extent that soils contained roughly the same amount of carbon as agricultural soils, the LMAV stored 2438 TgC. Discussion We provided an accounting of the major above- and belowground carbon pools in the 1 million ha Missouri LMAV and extrapolated results to the 10.4 million LMAV under changing land use. The LMAV currently is in intensive crop production due to fertile alluvial soils; however, these crops may be most vulnerable to climate change, particularly flooding (Brown et al., 2014; Groffman et al., 2014; Pryor et al., 2014). In addition to marginal soils currently in use for agriculture, more soils will become prone to flooding and less valuable for crop production. Crops that are restored to forest will provide an outlet for floodwaters and store
Table 2 Range in carbon density (Mg/ha) potentially stored in soil and trees after restoration of agricultural land in the MMAV. Total carbon represents combined soil and tree carbon; low values are for young forests with low soil carbon whereas mean and high values are for mature forests with mean or high soil carbon values, respectively. Ecological subsection
Mississippi River Crowley’s Ridge White and Black Rivers St. Francis River Arkansas Alluvial Plain Arkansas Grand Prairie Atchafalaya Bastrop Ridge Baton Rouge Terrace Macon Ridge Opelousas Ridge Red River Southern Mississippi River Teche Terrace Area-weighted mean
Agricultural
Reforested soil carbon
Reforested tree carbon
Reforested total carbon
Low
Mean
High
Young
Mature
Low
Mean
High
85 71 86 84 109 91 223 90 92 79 113 113 96 159
107 89 108 105 136 114 278 113 115 99 141 141 120 198
128 106 129 125 163 137 334 135 138 119 169 169 144 238
149 124 151 146 190 160 389 158 161 138 197 198 168 278
46 47 55 52 49 49 49 49 49 49 49 49 49 49
150 114 166 155 150 150 150 150 150 150 150 150 150 150
153 136 163 156 184 163 327 161 163 147 189 190 169 247
278 221 295 280 313 287 484 285 288 269 319 319 294 388
300 238 317 301 340 310 539 308 311 288 347 348 318 428
96
120
143
167
49
150
169
293
317
22
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
greater carbon. Carbon density will increase by about two-fold if managed as young forests compared to agricultural lands and three-fold if managed as mature forests (169–317 Mg/ha carbon density on forest lands compared to 96 Mg/ha carbon density on agricultural lands; Fig. 2). The carbon accounting may be used as an input into scenarios to examine when the benefits of ecosystem services may exceed the economic returns of crop production. Current forests in the LMAV are young and simulated aboveground biomass estimates for the northern LMAV of 97 Mg/ha are comparable with the entire LMAV (97 Mg/ha; summarization and expansion of FIA plots; B. Hanberry, University of Missouri, unpublished data) and other, non-alluvial, forests in the eastern United States. Our biomass estimates of about 300 Mg/ha by ecological subsection based on the most conservative diameter distribution for mature floodplain forests concur with Rheinhardt et al.’s (2012) estimate of 311 mg/ha for riparian forests >50 years in North Carolina. Older forests, however, can store more biomass than younger forests and do not appear to reach a steady state in carbon dynamics, notwithstanding classical assumptions (Carey et al., 2001; Pan et al., 2011). Therefore, it would not be unusual if old growth forests in one of the nation’s largest and most productive watersheds had greater biomass than mature second growth along headwater streams. Continued aboveground biomass storage in fertile sites may lead to achievement of 500–600 Mg/ha of the less conservative diameter distributions. Soil carbon averaged 70% of the combined tree and soil carbon in young forests and 50% of the combined carbon in mature forests with moderate ability to hold organic matter. These values matched well with pools in other temperate forest systems. Soil carbon alone typically represents about half of forest carbon (Turner et al., 1995), and perhaps greater in hydric soils of floodplain forests that have limited carbon cycling due to poor aeration (Pregitzer and Euskirchen, 2004). Aboveground tree biomass excluding foliage may account for 33% to 65% of the total carbon budget (Turner et al., 1995; Rheinhardt et al., 2012). We accounted for the two major carbon pools but not for smaller diameter trees or other carbon components including foliage, non-tree vegetation, roots, snags and coarse woody debris. Tree roots may represent about 3% of aboveground tree biomass and tree foliage <0.5% (allometric equations provided by FIA DataMart, www.fia.fs.fed.us/tools-data), but dead wood may represent >10% of aboveground biomass (Rhemtulla et al., 2009). Total soil and tree carbon storage in LMAV forests is about 4% of historical storage, which demonstrates the total potential for carbon storage that has been lost through land use and explains the failure in ecosystem function of the LMAV. Restoration to historical levels is unrealistic, but increases in total carbon storage by three- to five-fold will increase the region’s contribution to carbon storage. In the LMAV and elsewhere, reforested land has a great potential to store carbon and modest increases in forested acreage increase carbon storage. Conversion of marginal agricultural land that are prone to flooding already is occurring in floodplains, particularly in the LMAV, where there is heavy enrollment in programs such as the Conservation Reserve Program, included in the Farm Security and Rural Investment Act (i.e., Farm Bill; www.nrcs.usda.gov/ wps/portal/nrcs/main/national/programs). However, contracts for these programs often are short-term and land use will vary with best economic outcomes. That is, enrolled land provides temporary ecosystem services including wildlife habitat, recreation, improved water quality, erosion control, floodwater outlets, and esthetic purposes, but not necessarily long term development of forests along with permanent support of ecosystem services. Nevertheless, given millions of hectares of marginal agricultural land, public lands such as National Wildlife Refuges as well as some
lands owned by private stakeholders are being actively managed for long term reforestation. For example, the Wetlands Reserve Program has an option for a permanent easement agreement to restore land to forest or wetlands. However, time is necessary for forests to approach 300 Mg/ha of biomass and longer still for trees to attain diameters of 40 cm that contribute to great biomass and resemble trees of historical floodplain forests (Hanberry et al., 2012b). Harvesting using sound silvicultural practices can provide biofuels, fiber, and other forest products that contribute to long term carbon storage while enhancing tree growth. Agroforestry plantations of fast-growing species may supply a more economical alternative for private land owners, although with the single objective of financial return, tree crops may not be competitive against conventional agricultural crops even on marginal soils (Frey et al., 2010). Earlysuccessional species, in particular sweetgum but also cottonwood and willows (Hanberry et al., 2012b), grew in areas disturbed by frequent flooding and provide an ecological precedent for plantations, but fossil fuel inputs, short plantation rotations, and short-term storage in pulpwood or biofuels would offset carbon storage benefits. Harvest, drainage, agriculture, flood control, and altered hydrology have affected the Lower Mississippi River Alluvial Valley, similarly to floodplain forests world-wide. The LMAV currently is only 5% forested, and despite the presence of forested wetlands on 20% of the landscape, is not able to sustain ecosystem functioning due to intensive agricultural use on 60% of the landscape. As one consequence, nutrient inputs from agricultural fertilizers are causing hypoxia and dead zones in the Gulf of Mexico. Because of major ecological and economic consequences of deforestation, the LMAV may be an ideal landscape to manage for carbon storage under climate change due to the vulnerability of floodplains to flooding. Reforestation programs are in place to restore floodplain forest ecosystems and services such as carbon management, improved water quality, wildlife habitat, and floodwater retention, but climate change will increase the frequency, severity, and extent of flooding and it is important to make land use changes now to mitigate future flooding. Forests represent such a small percent of the landscape that addition of about 600,000 ha would double the forested extent (Fry et al., 2011). A 1.15 million ha forested extent is about 10% of the LMAV, a little larger than the LMAV in Missouri. Thus, realistic future carbon storage for the entire LMAV of 335 TgC would exceed the historical carbon storage of the Missouri LMAV. Acknowledgements We thank anonymous reviewers for their suggestions that improved the manuscript. This project was funded by the USDA Forest Service Northern Research Station and Eastern Region. Additional funds were provided by the Department of Interior USGS Northeast Climate Science Center. The contents of this paper are solely the responsibility of the authors and do not necessarily represent the views of the United States Government. References Brown, D.G., Polsky, C., Bolstad, P., Brody, S.D., Hulse, D., Kroh, R., Loveland, T.R., Thomson, A., 2014. Ch. 13: Land use and land cover change. In: Melillo, J.M., Richmond, T.C., Yohe, G.W. (Eds.), Climate Change Impacts in the United States: The Third National Climate Assessment. U.S. Global Change Research Program, pp. 318–332. Carey, E.V., Sala, A., Keane, R., Callaway, R.M., 2001. Are old forests underestimated as global carbon sinks? Global Change Biol. 7, 339–344. Clark, J.D., Johnson, A.H., 2011. Carbon and nitrogen accumulation in postagricultural forest soils of western New England. Soil Sci. Soc. Am. J. 75, 1530–1542. Ecomap, 1993. National Hierarchical Framework of Ecological Units. USDA Forest Service, Washington, DC.
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23 Frey, G.E., Mercer, D.E., Cubbage, F.W., Abt, R.C., 2010. Economic potential of agroforestry and forestry in the Lower Mississippi Alluvial Valley with incentive programs and carbon payments. South. J. Appl. For. 34, 176–185. Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., Wickham, J., 2011. Completion of the 2006 National Land Cover Database for the conterminous United States. PE&RS 77, 858–864. Gore, J.A., Shields Jr., F.D., 1995. Can large rivers be restored? BioScience 45, 142–152. Groffman, P.M., Kareiva, P., Carter, S., Grimm, N.B., Lawler, J., Mack, M., Matzek, V., Tallis, H., 2014. Ch. 8: Ecosystems, biodiversity, and ecosystem services. In: Melillo, J.M., Richmond, T.C., Yohe, G.W. (Eds.), Climate Change Impacts in the United States: The Third National Climate Assessment. U.S. Global Change Research Program, pp. 195–219. Guo, L.B., Gifford, R.M., 2002. Soil carbon stocks and land use change: a meta analysis. Global Change Biol. 8, 345–360. Hanberry, B.B., Fraver, S., He, H.S., Yang, J., Dey, D.C., Palik, B.J., 2011. Spatial pattern corrections and sample sizes for forest density estimates of historical tree surveys. Landsc. Ecol. 26, 59–68. Hanberry, B.B., Yang, J., Kabrick, J.M., He, H.S., 2012a. Adjusting forest density estimates for surveyor bias in historical tree surveys. Am. Midl. Nat. 167, 285–306. Hanberry, B.B., Kabrick, J.M., He, H.S., Palik, B.J., 2012b. Historical trajectories and restoration strategies for the Mississippi River Alluvial Valley. For. Ecol. Manage. 280, 103–111. Hoover, C.M., Leak, W.B., Keel, B.G., 2012. Benchmark carbon stocks from old-growth forests in northern New England. USA. For. Ecol. Manage. 266, 108–114. Indorante, S.J., Kabrick, J.M., Lee, B.D., Matta, J.M., 2013. Quantifying soil profile change caused by land use in central Missouri loess hillslopes. Soil Sci. Soc. Am. J. 78, 225–237. Jackson, M.T., 2006. Forest communities and tree species of the Lower Wabash River Basin. Proc. Indiana Acad. Sci. 115, 94–102. King, S.L., Keeland, B.D., 1999. Evaluation of reforestation in the Lower Mississippi River Alluvial Valley. Restor. Ecol. 7, 348–359. Morisita, M., 1957. A new method for the estimation of density by the spacing method, applicable to non-randomly distributed populations. Seiri Seitai 7, 134–144.
23
Pan, Y., Chen, J.M., Birdsey, R., McCullough, K., He, L., Deng, F., 2011. Age structure and disturbance legacy of North American forests. Biogeosciences 8, 715–732. Pregitzer, K.S., Euskirchen, E.S., 2004. Carbon cycling and storage in world forests: biome patterns related to forest age. Global Change Biol. 10, 2052– 2077. Pryor, S.C., Scavia, D., Downer, C., Gaden, M., Iverson, L., Nordstrom, R., Patz, J., Robertson, G.P., 2014. Ch. 18: Midwest. In: Melillo, J.M., Richmond, T.C., Yohe, G.W. (Eds.), Climate Change Impacts in the United States: The Third National Climate Assessment. U.S. Global Change Research Program, pp. 418– 440. R Development Core team, 2012. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Rheinhardt, R., Brinson, M., Meyer, G., Miller, K., 2012. Integrating forest biomass and distance from channel to develop an indicator of riparian condition. Ecol. Indic. 23, 46–55. Rhemtulla, J.M., Mladenoff, D.J., Clayton, M.K., 2009. Historical forest baselines reveal potential for continued carbon sequestration. PNAS 106, 6082–6087. Schoenholtz, S.H., James, J.P., Kaminski, R.M., Leopold, B.D., Ezell, A.W., 2001. Afforestation of bottomland hardwoods in the Lower Mississippi Alluvial Valley: status and trends. Wetlands 21, 602–613. Shoch, D.T., Kaster, G., Hohl, A., Souter, R., 2009. Carbon storage of bottomland hardwood afforestation in the Lower Mississippi Valley, USA. Wetlands 29, 535–542. Suchenwirth, L., Förster, M., Cierjacks, A., Lang, F., Kleinschmit, B., 2012. Knowledgebased classification of remote sensing data for the estimation of below- and above-ground organic carbon stocks in riparian forests. Wetl. Ecol. Manage. 20, 151–163. Turner, D.P., Koerper, G.J., Harmon, M.E., Lee, J.J., 1995. A carbon budget for forests of the conterminous United States. Ecol. Appl. 5, 421–436. Turner, R.E., Rabalais, N.N., 2003. Linking landscape and water quality in the Mississippi River Basin for 200 years. BioScience 53, 563–572. Zedler, J.B., Kercher, S., 2005. Wetland resources: status, trends, ecosystem services, and restorability. Annu. Rev. Environ. Resour. 30, 39–74.
Contents lists available at ScienceDirect
Perspectives in Plant Ecology, Evolution and Systematics journal homepage: www.elsevier.com/locate/ppees
Research article
Potential tree and soil carbon storage in a major historical floodplain forest with disrupted ecological function Brice B. Hanberry a,∗ , John M. Kabrick b , Hong S. He a a b
Department of Forestry, University of Missouri, 203 Natural Resources Building, Columbia, MO 65211, USA USDA Forest Service, Northern Research Station, University of Missouri, 202 Natural Resources Building, Columbia, MO 65211, USA
a r t i c l e
i n f o
Article history: Received 16 April 2014 Received in revised form 11 December 2014 Accepted 18 December 2014 Available online 26 December 2014 Keywords: Agriculture Climate change Ecosystem services Flooding Reforestation Restoration
a b s t r a c t Floodplain forests are extremely productive for agriculture and historical floodplain forests have been converted to prime agricultural land throughout the world, resulting in disruption of ecosystem functioning. Given that flooding may increase with climate change and reforestation will increase resiliency to climate change, we tested whether reforested floodplains also have great potential to store carbon and the effects of even modest increases in forested acreage on carbon storage. To calculate potential aboveground biomass in the Lower Mississippi River Alluvial Valley (LMAV) of the United States, we determined current and historical tree biomass used density estimates and diameter distributions from tree surveys and relationships between diameter and biomass from current forests. To calculate potential soil organic carbon if the landscape was forested, we used soil organic matter from soil surveys of the agricultural landscape, and multiplied the carbon by a factors of 1.25, 1.5, and 1.75 based on published reports of soil carbon increases due to afforestation. Our results showed that area-weighted mean biomass density (trees ≥12.7 cm in diameter) for historical forests was 300 Mg/ha, ranging from 228 Mg/ha to 332 Mg/ha by ecological subsection, based on the most conservative diameter distribution. Mean biomass density for current forests was 97 Mg/ha, ranging from 92 Mg/ha to 111 Mg/ha. Mean carbon density for agricultural soils was 96 Mg/ha, whereas combined tree and soil carbon densities varied from 169 Mg/ha to 317 Mg/ha; soil carbon accounted for 0.5–0.7 of total carbon density. Historical forested carbon storage in the Missouri LMAV was about 234 TgC, with the most conservative diameter distribution and assuming 80% forest coverage. Current forested carbon storage in the Missouri LMAV is about 2% of historical storage, at 5 TgC in 30,000 ha of forests, but may reach 23 TgC if forested extent almost triples, with the addition of 50,000 ha of marginal agricultural land, and carbon storage increases in trees and soil. The entire LMAV currently stores 97 TgC in forests and reasonable carbon storage for the entire LMAV may be about 335 TgC, based on increased carbon storage and reforestation of 600,000 ha of marginal agricultural land, which would double the current forested extent. Although 335 TgC storage for the LMAV is only about 1.5 times greater than historical carbon storage of the Missouri LMAV, doubling the forested extent will increase other ecosystem functions, including carbon storage, flood abatement, and reduction of fertilizer pollution in the Gulf of Mexico. © 2014 Geobotanisches Institut ETH, Stiftung Ruebel. Published by Elsevier GmbH. All rights reserved.
Introduction Floodplain forests have been converted to agriculture and other land uses globally (Gore and Shields, 1995; Zedler and Kercher, 2005). Accompanied by reduction in floodplain forests, loss of ecological function interrupts ecosystem services including flood abatement, biodiversity support, water quality improvement, and
∗ Corresponding author. Tel.: +1 5738755341. E-mail address: [email protected] (B.B. Hanberry).
carbon management (Zedler and Kercher, 2005). Both carbon and nutrient capture (i.e., sequestration) and retention (i.e., storage) are limited without the presence of long term vegetation ground cover. For example, agricultural fertilizers move through floodplains to coastal waters, creating dead zones as algal blooms deplete oxygen (Turner and Rabalais, 2003; Zedler and Kercher, 2005). Increased long term biomass through reforestation of floodplain forests will contribute to carbon and nutrient storage along with other ecosystem services. Carbon storage currently is unrealized in historical floodplain forests. The Mississippi River Basin drains six major watersheds
http://dx.doi.org/10.1016/j.ppees.2014.12.002 1433-8319/© 2014 Geobotanisches Institut ETH, Stiftung Ruebel. Published by Elsevier GmbH. All rights reserved.
18
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
Heavy precipitation events will disrupt current infrastructure in place to control rivers, resulting in frequent flooding on agricultural lands on historical floodplains (Pryor et al., 2014). Ecosystem services of flood and erosion control and carbon storage provided by floodplain forests may become more valuable than agricultural lands vulnerable to flooding and that consequently will become less productive under climate change. Increased restoration of floodplain forests will increase resiliency to extreme events of climate change (Groffman et al., 2014). In light of increased flooding expected under climate change, reforestation for a variety of ecosystem services may become a more viable land use in former floodplain forests. Mature, unharvested forests in floodplain forests are rare at the stand scale and non-existent at a landscape scale. Potential carbon storage of mature floodplain forests therefore is unknown for the LMAV and important for evaluating to benefits of land use conversion. Our aim was to test the (1) potential for the LMAV to store carbon in the major carbon pools of aboveground tree biomass and belowground soil carbon and (2) effects of even modest increases in forested acreage on above- and belowground carbon storage. We used historical tree surveys from the LMAV in Missouri to quantify and map forest aboveground biomass for four ecological subsections (Ecomap, 1993; Figs. 1 and 2, mean area = 1.45 million ha, SD = 1.20 million ha). We then calculated forested carbon storage potential for the Missouri LMAV and generalized our results to the entire LMAV. Under land use change from forest to crop and potentially restoration back to forest, we provide tree and soil carbon accounting valuable for management scenarios in a 10.4 million ha landscape and a method for soil carbon accounting applicable to other regions.
Fig. 1. The Lower Mississippi River Valley: the four ecological subsections in Missouri (outlined in white) cross state boundaries to form the entire White and Black River Alluvial Plains ecological section (shaded dark gray). Forests (shaded black) are present in areas with greater elevation.
that cover about 40% of the continental United States. One of the watersheds, the Lower Mississippi River Alluvial Valley (LMAV; Fig. 1) once contained the greatest area of floodplain forests. By 1978, only 2 million ha (Schoenholtz et al., 2001) of the historical 10.4 million ha forested extent remained (areal extent excludes open water; Fry et al., 2011). Currently, cooperative partnerships are in place to replant trees in portions of the LMAV, primarily within marginal agricultural lands that have hydric soils with poor drainage (King and Keeland, 1999; Frey et al., 2010). In addition to unrealized carbon storage, floodplain forests may have greater potential to store carbon than other ecosystem types, particularly compared to surrounding upland forests in temperate zones (Suchenwirth et al., 2012). Deep alluvial soils are recognized for crop productivity, even though tree productivity has been less well-documented (Shoch et al., 2009). Although we are unaware of historical accounts of trees in the LMAV, directly north of LMAV, multiple trees within one ha had diameters of 2–5 m, exceeding current state record trees for species (in the Wabash River Valley; Jackson, 2006). Decisions about land use in floodplains may become increasingly important as climate changes (Brown et al., 2014; Groffman et al., 2014). Temperatures in the midwestern United States may increase 3–5 ◦ C by the end of the century (Pryor et al., 2014). Although longer growing seasons and rising carbon dioxide levels initially may increase yields of some crops, climate change eventually may decrease agricultural productivity due to wet springs, anomalous frosts, drought, and heat stress, particularly during pollination and reproductive development (Pryor et al., 2014).
Methods Study area Missouri’s LMAV landscape is about 1 million ha excluding open water, nearly 10% of the LMAV (Ecomap, 1993; Fry et al., 2011; Fig. 1). About 80% of land cover is cultivated crops, 4% is pasture or hay, 2.5% is forest, and 5% is forested wetlands. Crowley’s Ridge, one of four ecological subsections (Ecomap, 1993) in the Missouri LMAV, is a loess-covered upland that rises above flatter alluvial plains and accounts for most of the landscape’s pasture and forest. Pasture and forest comprise 24% and 14%, respectively, of land cover in Crowley’s Ridge in Missouri. Roughly 80% of soils in the Missouri LMAV have limitations, primarily (66% of total soil) due to poor drainage. About 130,000 ha are classified as not prime farmland (including land that needs to be protected from flooding; Soil Survey Geographic Database, Natural Resources Conservation Service, http://soildatamart.nrcs.usda.gov), of which about 47,900 ha is cultivated crops and 13,800 ha is forest. Tree surveys and density estimates The United States General Land Office (GLO) was established in 1812 to survey, map, and sell land for settlement. The GLO surveys divided area into square townships measuring 9.6 km × 9.6 km, which were divided further into 36–1.6 km2 sections. Surveyors selected two to four bearing trees at the intersection of section lines and midpoints between section corners. For each selected tree, surveyors recorded species, diameter, distance, and bearing to survey point. The GLO surveys contain bias because surveyors selected trees at survey points, resulting in non-random trees. Furthermore, selected trees were of moderate diameter, to increase longevity as section markers. We excluded trees with diameters
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
19
Fig. 2. Estimated current soil carbon density (Mg/ha) by ecological subsection in the predominantly agricultural Lower Mississippi River Valley (panel A). Potential increases in soil and tree carbon density (Mg/ha) by ecological subsection if agricultural lands are converted to forested lands and managed as young (panel B) or mature forests (panel C).
<12.7 cm to maintain a consistent diameter threshold. There were 25,343 trees of this size recorded during 1817–1860. The USDA Forest Service Forest Inventory and Analysis Program (FIA) conducts forest inventories in long term plots that are surveyed during a five year cycle. Because there were few plots in Missouri’s LMAV, we used plots from the entire White and Black River Alluvial Plains ecological section (Ecomap, 1993; Fig. 1), which crossed state boundaries. We used the latest cycles from Arkansas during 2006–2010 (63% of trees), Mississippi during 2009–2011 (20% of trees), and Missouri during 2004–2008 (8% of trees), as well as four other states (FIA DataMart, www.fia.fs.fed.us/tools-data). We selected trees with diameters ≥12.7 cm and selected plots that contained at least two trees and were 100% forestland, which FIA defines as land at least 0.4 ha in size and 37 m wide with at least 10% cover by live trees of any
size, “including land that formerly had such tree cover and that will continue to have forest use”. This selection included 6330 trees. In order to quantify historical biomass, we used simulations to estimate biomass by ecological subsection based on density estimates and diameter distributions (see section below for description of simulations; also Rhemtulla et al., 2009). Historical survey points contained only 2–4 trees per plot, so it was not possible to estimate biomass at each plot. For current biomass, the FIA program provides tree biomass based on allometric equations. Biomass (biomass of tree bole, stump, and top excluding foliage) can be summed per plot and expanded to a hectare. To make certain that simulations were accurate, we also simulated FIA biomass using the same method as for GLO biomass. We estimated current tree density by summation of the number of trees per plot and expanded to a per hectare basis.
20
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
We first estimated historical density for each ecological subsection (Fig. 1) with the Morista estimator (Morisita, 1957) for (1) survey points with two trees and (2) survey points with three trees and the nearest three trees for survey points with four trees (diameters ≥12.7 cm; Hanberry et al., 2011). We then produced a low and high value based on adjustment for potential spatial patterning (clustered or regular patterns; Hanberry et al., 2011). We corrected these values for surveyor bias. Surveyors potentially did not select the nearest trees to each survey point, which would produce a distance rank of 1. Selection of more distant trees will result in underestimated densities. We adjusted density estimates using a rank-based method to estimate a low value, assuming selected trees had a mean rank of 1.4, and a mean value, assuming selected trees had a mean rank of 1.8. Using a complementary method to correct for non-random ratios of quadrants and azimuth and differences between frequencies of species and diameter of trees selected at survey points and trees recorded along survey lines, we calculated density estimates for another mean value and a high value (Hanberry et al., 2012a). We then determined a weighted average of the two mean values from the two complementary methods and retained the low value from the rank-based method and high value from the bias method. Simulated biomass estimation We estimated biomass using simulations of densities and diameter distributions, similarly to Rhemtulla et al. (2009), for both GLO and FIA surveys. We developed parameters for probability distribution functions of historical and current tree diameters by ecological subsection truncated at 12.7 cm using lognormal (which describe one-storied, even-aged forests; Podlaski 2010) exponential distributions (for uneven-aged forests), Weibull and gamma distributions (for structurally heterogeneous forests; Proc Severity, SAS software, version 9.1, Cary, NC). We simulated 10,000 stands of one hectare using random diameters generated from diameter distribution parameters truncated at 12.7 cm and random densities generated from density estimates and uncertainty (R Runuran, R Development Core team, 2012; J. Leydold & W. Hörmann, http://statmath.wu.ac.at/unuran, http://cran.rproject.org/web/packages/Runuran/index.html). We used allometric equations by ecological subsection from FIA surveys based on a simple regression of the relationship between (the log transformation of) biomass and (the log transformation of) diameter to quantify the biomass of each tree. Because allometric equations were developed for trees grown in forests under current land use and management and not available for trees grown under different conditions, we maintained a constant allometric relationship for each subsection. We then summarized the biomass for each stand and determined mean biomass and standard deviation per hectare. We used mean biomass per ecological subsection from FIA surveys, calculated by summarizing and expanding biomass at FIA plots, to determine mean absolute difference between simulated current biomass for each distribution and mean biomass, thereby testing the accuracy of our simulations. We identified the distribution for historical biomass that had the least difference from simulated FIA biomass of the same distribution as the most conservative fit for comparisons. Nevertheless, we provided the maximum value for each ecological subsection from other distributions. Soil carbon storage We used soils surveys from Soil Survey Geographic (SSURGO) Database (Natural Resources Conservation Service,
http://soildatamart.nrcs.usda.gov) to calculate soil organic matter for each ecological subsection. We used a conversion factor of 0.58 to convert soil organic matter to soil organic matter (Guo and Gifford, 2002). Because the LMAV is predominantly agricultural, soil surveys were from predominantly agricultural soils. Based on literature of soil carbon accumulation from reforestation of agricultural land, soil carbon is approximately 1.5 times greater after afforestation (Guo and Gifford, 2002; Clark and Johnson 2011; Indorante et al., 2013). Furthermore, this soil carbon level appears to be maximized after time such that soils in older forests are not able to store any more carbon in the soil (Hoover et al., 2012). In addition to the adjustment factor of 1.5, we also produced a low estimate for newly restored young forests using an adjustment of 1.25 and a high estimate of 1.75 for soils capable of carrying greater carbon capacity (Guo and Gifford, 2002). Potential forested carbon storage The amount of forest land cover during the survey period of 1817–1860 is unknown. Only about 5% of the landscape currently is wetlands, but before drainage for agriculture, standing water may have reduced forested extent. Nevertheless, for 97% of the landscape, surveyors were able to record trees along the systematic survey grid. In addition to water and flooding, fire and wind disturbance also would decrease the forested extent. The LMAV was used by Native Americans (e.g., the Cahokia settlement near current East St. Louis), but by the 1800s, Native American population density had decreased. As for European Americans, at least while accessible upland forests were available and harvesting technology was less developed, wet soils deterred harvest of floodplain forests. Therefore, we provided a range of forested extents from 60% to 100% to represent the historical landscape along with percentage of current forested land and reforestation scenarios for marginal agricultural land, approximately two to three times current forested extent. We then calculated total forested carbon storage in combined trees and soil for young forests and mature forests, using a conversion factor of 0.5 to convert biomass to carbon. We extrapolated to the entire LMAV based on forested extent and a range of calculated carbon densities. Results To verify the accuracy of the forest biomass simulations made with different distributions, comparisons to FIA biomass estimates were made. Mean absolute difference between simulated biomass of current forests and mean biomass estimates from FIA plots by ecological subsection was 3.16 Mg/ha for the gamma distribution, 3.35 Mg/ha for the Weibull distribution, 4.42 Mg/ha exponential distribution, and 12.41 Mg/ha for the lognormal distribution. Mean absolute difference between simulated biomass estimates for historical forests and simulated biomass of current forests by ecological subsection was 184 Mg/ha for the Weibull distribution and 208 Mg/ha for the lognormal distribution. The other two distributions had differences of about 370 Mg/ha. To remain consistent in comparisons, we selected the overall most conservative distribution, the Weibull distribution. Area-weighted mean biomass for historical forests averaged 300 Mg/ha, ranging from 228 Mg/ha to 332 Mg/ha by ecological subsection (trees ≥12.7 cm and the Weibull distribution; Table 1). Area-weighted mean biomass for current forests averaged 97 Mg/ha, ranging from 92 Mg/ha to 111 Mg/ha by ecological subsection. Historical biomass was greater per ha than current biomass by a factor of three. Using the least conservative distribution for simulated historical biomass, historical biomass was greater per ha than current biomass by a factor of six (area-weighted).
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23 Table 1 Historical (GLO) and current (FIA) live aboveground biomass (Mg/ha; trees ≥12.7 cm), based on the Weibell distribution, and maximum mean values based on alternative distributions. Ecological subsection
Mississippi River Crowley’s Ridge White and Black Rivers St. Francis River
GLO
FIA
GLO
Table 3 Total forested carbon storage (TgC) for the Missouri LMAV and entire LMAV, varying by forested extent and carbon density (Mg/ha) of a range of forested carbon densities. % Forested
Mean
SD
Mean
SD
Max mean
SD
300 228 332 310
77 62 76 65
92 94 111 104
18 17 28 20
759 304 480 529
284 103 156 147
Area-weighted mean soil carbon storage of the current agricultural landscape of the entire LMAV averaged 96 Mg/ha, ranging from 71 Mg/ha to 223 Mg/ha by ecological subsection (Table 2; Fig. 2). Adjustment (of 1.5) for a forested cover produced an areaweighted mean soil carbon storage of 143 Mg/ha, ranging from 106 Mg/ha to 334 Mg/ha by ecological subsection. The low adjustment (of 1.25) area-weighted mean was 120 Mg/ha and the high adjustment was 167 Mg/ha. We combined results with aboveground biomass density, after conversion to carbon density. Young forests that have accrued little additional soil carbon or above-ground biomass have a mean combined soil and tree carbon density of 169 Mg/ha, ranging from 136 Mg/ha to 327 Mg/ha by ecological subsection (Table 2; Fig. 2). Mature forests with moderate increases in soil organic carbon have a mean combined carbon density of 293 Mg/ha, ranging from 221 Mg/ha to 484 Mg/ha by ecological subsection. In young forests, soil carbon represents 71% of forest carbon (ranging from 65% to 85%) whereas in mature forests with moderate organic soils, soil carbon represents 49% of forest carbon (ranging from 44% to 69%). The historical potential in the Missouri LMAV for forested carbon storage in trees and soil was approximately 234 TgC (1 Tg = 1,000,000 Mg). This estimate is based on 80% forested forest coverage, the most conservative simulated estimate of historical tree biomass, and mean forested soil values (Table 3). Current forested carbon storage was approximately 5 TgC, or 2% of historical carbon storage, based on 30,000 ha of forested land. The restoration potential from about 50,000 ha of marginal agricultural land currently in crops combined would result in total forested carbon storage of 23 TgC, an increase of about five-fold, after trees have matured and soil has developed moderate carbon capacity. We can extrapolate these results to the entire 10.4 million ha LMAV, which currently is about 5.5% forested, excluding woody wetlands that is larger proportion of the entire LMAV due to
21
Total carbon storage Reforested carbon density Low
Missouri LMAV 100 90 80 70 60 8 3 LMAV 100 90 80 70 60 17 11 6
Mean
High
169 152 135 118 101 14 5
293 264 234 205 176 23 9
317 285 254 222 190 25 10
1758 1582 1406 1230 1055 299 193 97
3047 2742 2438 2133 1828 518 335 168
3297 2967 2637 2308 1978 560 363 181
major rivers (about 22% of the landscape; Ecomap, 1993; Fry et al., 2011). The LMAV currently stores 97 TgC in forests. Given time for trees to accrue and store more biomass and expanding restoration efforts to add another 600,000 ha and double the forested extent to 1.15 million ha will increase storage during a century to 335 TgC. Historically, given 80% forested extent and assuming the most conservative diameter distribution and mean soil adjustment and that the non-forested extent that soils contained roughly the same amount of carbon as agricultural soils, the LMAV stored 2438 TgC. Discussion We provided an accounting of the major above- and belowground carbon pools in the 1 million ha Missouri LMAV and extrapolated results to the 10.4 million LMAV under changing land use. The LMAV currently is in intensive crop production due to fertile alluvial soils; however, these crops may be most vulnerable to climate change, particularly flooding (Brown et al., 2014; Groffman et al., 2014; Pryor et al., 2014). In addition to marginal soils currently in use for agriculture, more soils will become prone to flooding and less valuable for crop production. Crops that are restored to forest will provide an outlet for floodwaters and store
Table 2 Range in carbon density (Mg/ha) potentially stored in soil and trees after restoration of agricultural land in the MMAV. Total carbon represents combined soil and tree carbon; low values are for young forests with low soil carbon whereas mean and high values are for mature forests with mean or high soil carbon values, respectively. Ecological subsection
Mississippi River Crowley’s Ridge White and Black Rivers St. Francis River Arkansas Alluvial Plain Arkansas Grand Prairie Atchafalaya Bastrop Ridge Baton Rouge Terrace Macon Ridge Opelousas Ridge Red River Southern Mississippi River Teche Terrace Area-weighted mean
Agricultural
Reforested soil carbon
Reforested tree carbon
Reforested total carbon
Low
Mean
High
Young
Mature
Low
Mean
High
85 71 86 84 109 91 223 90 92 79 113 113 96 159
107 89 108 105 136 114 278 113 115 99 141 141 120 198
128 106 129 125 163 137 334 135 138 119 169 169 144 238
149 124 151 146 190 160 389 158 161 138 197 198 168 278
46 47 55 52 49 49 49 49 49 49 49 49 49 49
150 114 166 155 150 150 150 150 150 150 150 150 150 150
153 136 163 156 184 163 327 161 163 147 189 190 169 247
278 221 295 280 313 287 484 285 288 269 319 319 294 388
300 238 317 301 340 310 539 308 311 288 347 348 318 428
96
120
143
167
49
150
169
293
317
22
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23
greater carbon. Carbon density will increase by about two-fold if managed as young forests compared to agricultural lands and three-fold if managed as mature forests (169–317 Mg/ha carbon density on forest lands compared to 96 Mg/ha carbon density on agricultural lands; Fig. 2). The carbon accounting may be used as an input into scenarios to examine when the benefits of ecosystem services may exceed the economic returns of crop production. Current forests in the LMAV are young and simulated aboveground biomass estimates for the northern LMAV of 97 Mg/ha are comparable with the entire LMAV (97 Mg/ha; summarization and expansion of FIA plots; B. Hanberry, University of Missouri, unpublished data) and other, non-alluvial, forests in the eastern United States. Our biomass estimates of about 300 Mg/ha by ecological subsection based on the most conservative diameter distribution for mature floodplain forests concur with Rheinhardt et al.’s (2012) estimate of 311 mg/ha for riparian forests >50 years in North Carolina. Older forests, however, can store more biomass than younger forests and do not appear to reach a steady state in carbon dynamics, notwithstanding classical assumptions (Carey et al., 2001; Pan et al., 2011). Therefore, it would not be unusual if old growth forests in one of the nation’s largest and most productive watersheds had greater biomass than mature second growth along headwater streams. Continued aboveground biomass storage in fertile sites may lead to achievement of 500–600 Mg/ha of the less conservative diameter distributions. Soil carbon averaged 70% of the combined tree and soil carbon in young forests and 50% of the combined carbon in mature forests with moderate ability to hold organic matter. These values matched well with pools in other temperate forest systems. Soil carbon alone typically represents about half of forest carbon (Turner et al., 1995), and perhaps greater in hydric soils of floodplain forests that have limited carbon cycling due to poor aeration (Pregitzer and Euskirchen, 2004). Aboveground tree biomass excluding foliage may account for 33% to 65% of the total carbon budget (Turner et al., 1995; Rheinhardt et al., 2012). We accounted for the two major carbon pools but not for smaller diameter trees or other carbon components including foliage, non-tree vegetation, roots, snags and coarse woody debris. Tree roots may represent about 3% of aboveground tree biomass and tree foliage <0.5% (allometric equations provided by FIA DataMart, www.fia.fs.fed.us/tools-data), but dead wood may represent >10% of aboveground biomass (Rhemtulla et al., 2009). Total soil and tree carbon storage in LMAV forests is about 4% of historical storage, which demonstrates the total potential for carbon storage that has been lost through land use and explains the failure in ecosystem function of the LMAV. Restoration to historical levels is unrealistic, but increases in total carbon storage by three- to five-fold will increase the region’s contribution to carbon storage. In the LMAV and elsewhere, reforested land has a great potential to store carbon and modest increases in forested acreage increase carbon storage. Conversion of marginal agricultural land that are prone to flooding already is occurring in floodplains, particularly in the LMAV, where there is heavy enrollment in programs such as the Conservation Reserve Program, included in the Farm Security and Rural Investment Act (i.e., Farm Bill; www.nrcs.usda.gov/ wps/portal/nrcs/main/national/programs). However, contracts for these programs often are short-term and land use will vary with best economic outcomes. That is, enrolled land provides temporary ecosystem services including wildlife habitat, recreation, improved water quality, erosion control, floodwater outlets, and esthetic purposes, but not necessarily long term development of forests along with permanent support of ecosystem services. Nevertheless, given millions of hectares of marginal agricultural land, public lands such as National Wildlife Refuges as well as some
lands owned by private stakeholders are being actively managed for long term reforestation. For example, the Wetlands Reserve Program has an option for a permanent easement agreement to restore land to forest or wetlands. However, time is necessary for forests to approach 300 Mg/ha of biomass and longer still for trees to attain diameters of 40 cm that contribute to great biomass and resemble trees of historical floodplain forests (Hanberry et al., 2012b). Harvesting using sound silvicultural practices can provide biofuels, fiber, and other forest products that contribute to long term carbon storage while enhancing tree growth. Agroforestry plantations of fast-growing species may supply a more economical alternative for private land owners, although with the single objective of financial return, tree crops may not be competitive against conventional agricultural crops even on marginal soils (Frey et al., 2010). Earlysuccessional species, in particular sweetgum but also cottonwood and willows (Hanberry et al., 2012b), grew in areas disturbed by frequent flooding and provide an ecological precedent for plantations, but fossil fuel inputs, short plantation rotations, and short-term storage in pulpwood or biofuels would offset carbon storage benefits. Harvest, drainage, agriculture, flood control, and altered hydrology have affected the Lower Mississippi River Alluvial Valley, similarly to floodplain forests world-wide. The LMAV currently is only 5% forested, and despite the presence of forested wetlands on 20% of the landscape, is not able to sustain ecosystem functioning due to intensive agricultural use on 60% of the landscape. As one consequence, nutrient inputs from agricultural fertilizers are causing hypoxia and dead zones in the Gulf of Mexico. Because of major ecological and economic consequences of deforestation, the LMAV may be an ideal landscape to manage for carbon storage under climate change due to the vulnerability of floodplains to flooding. Reforestation programs are in place to restore floodplain forest ecosystems and services such as carbon management, improved water quality, wildlife habitat, and floodwater retention, but climate change will increase the frequency, severity, and extent of flooding and it is important to make land use changes now to mitigate future flooding. Forests represent such a small percent of the landscape that addition of about 600,000 ha would double the forested extent (Fry et al., 2011). A 1.15 million ha forested extent is about 10% of the LMAV, a little larger than the LMAV in Missouri. Thus, realistic future carbon storage for the entire LMAV of 335 TgC would exceed the historical carbon storage of the Missouri LMAV. Acknowledgements We thank anonymous reviewers for their suggestions that improved the manuscript. This project was funded by the USDA Forest Service Northern Research Station and Eastern Region. Additional funds were provided by the Department of Interior USGS Northeast Climate Science Center. The contents of this paper are solely the responsibility of the authors and do not necessarily represent the views of the United States Government. References Brown, D.G., Polsky, C., Bolstad, P., Brody, S.D., Hulse, D., Kroh, R., Loveland, T.R., Thomson, A., 2014. Ch. 13: Land use and land cover change. In: Melillo, J.M., Richmond, T.C., Yohe, G.W. (Eds.), Climate Change Impacts in the United States: The Third National Climate Assessment. U.S. Global Change Research Program, pp. 318–332. Carey, E.V., Sala, A., Keane, R., Callaway, R.M., 2001. Are old forests underestimated as global carbon sinks? Global Change Biol. 7, 339–344. Clark, J.D., Johnson, A.H., 2011. Carbon and nitrogen accumulation in postagricultural forest soils of western New England. Soil Sci. Soc. Am. J. 75, 1530–1542. Ecomap, 1993. National Hierarchical Framework of Ecological Units. USDA Forest Service, Washington, DC.
B.B. Hanberry et al. / Perspectives in Plant Ecology, Evolution and Systematics 17 (2015) 17–23 Frey, G.E., Mercer, D.E., Cubbage, F.W., Abt, R.C., 2010. Economic potential of agroforestry and forestry in the Lower Mississippi Alluvial Valley with incentive programs and carbon payments. South. J. Appl. For. 34, 176–185. Fry, J., Xian, G., Jin, S., Dewitz, J., Homer, C., Yang, L., Barnes, C., Herold, N., Wickham, J., 2011. Completion of the 2006 National Land Cover Database for the conterminous United States. PE&RS 77, 858–864. Gore, J.A., Shields Jr., F.D., 1995. Can large rivers be restored? BioScience 45, 142–152. Groffman, P.M., Kareiva, P., Carter, S., Grimm, N.B., Lawler, J., Mack, M., Matzek, V., Tallis, H., 2014. Ch. 8: Ecosystems, biodiversity, and ecosystem services. In: Melillo, J.M., Richmond, T.C., Yohe, G.W. (Eds.), Climate Change Impacts in the United States: The Third National Climate Assessment. U.S. Global Change Research Program, pp. 195–219. Guo, L.B., Gifford, R.M., 2002. Soil carbon stocks and land use change: a meta analysis. Global Change Biol. 8, 345–360. Hanberry, B.B., Fraver, S., He, H.S., Yang, J., Dey, D.C., Palik, B.J., 2011. Spatial pattern corrections and sample sizes for forest density estimates of historical tree surveys. Landsc. Ecol. 26, 59–68. Hanberry, B.B., Yang, J., Kabrick, J.M., He, H.S., 2012a. Adjusting forest density estimates for surveyor bias in historical tree surveys. Am. Midl. Nat. 167, 285–306. Hanberry, B.B., Kabrick, J.M., He, H.S., Palik, B.J., 2012b. Historical trajectories and restoration strategies for the Mississippi River Alluvial Valley. For. Ecol. Manage. 280, 103–111. Hoover, C.M., Leak, W.B., Keel, B.G., 2012. Benchmark carbon stocks from old-growth forests in northern New England. USA. For. Ecol. Manage. 266, 108–114. Indorante, S.J., Kabrick, J.M., Lee, B.D., Matta, J.M., 2013. Quantifying soil profile change caused by land use in central Missouri loess hillslopes. Soil Sci. Soc. Am. J. 78, 225–237. Jackson, M.T., 2006. Forest communities and tree species of the Lower Wabash River Basin. Proc. Indiana Acad. Sci. 115, 94–102. King, S.L., Keeland, B.D., 1999. Evaluation of reforestation in the Lower Mississippi River Alluvial Valley. Restor. Ecol. 7, 348–359. Morisita, M., 1957. A new method for the estimation of density by the spacing method, applicable to non-randomly distributed populations. Seiri Seitai 7, 134–144.
23
Pan, Y., Chen, J.M., Birdsey, R., McCullough, K., He, L., Deng, F., 2011. Age structure and disturbance legacy of North American forests. Biogeosciences 8, 715–732. Pregitzer, K.S., Euskirchen, E.S., 2004. Carbon cycling and storage in world forests: biome patterns related to forest age. Global Change Biol. 10, 2052– 2077. Pryor, S.C., Scavia, D., Downer, C., Gaden, M., Iverson, L., Nordstrom, R., Patz, J., Robertson, G.P., 2014. Ch. 18: Midwest. In: Melillo, J.M., Richmond, T.C., Yohe, G.W. (Eds.), Climate Change Impacts in the United States: The Third National Climate Assessment. U.S. Global Change Research Program, pp. 418– 440. R Development Core team, 2012. R: A Language and Environment for Statistical Computing. R Foundation for Statistical Computing, Vienna, Austria. Rheinhardt, R., Brinson, M., Meyer, G., Miller, K., 2012. Integrating forest biomass and distance from channel to develop an indicator of riparian condition. Ecol. Indic. 23, 46–55. Rhemtulla, J.M., Mladenoff, D.J., Clayton, M.K., 2009. Historical forest baselines reveal potential for continued carbon sequestration. PNAS 106, 6082–6087. Schoenholtz, S.H., James, J.P., Kaminski, R.M., Leopold, B.D., Ezell, A.W., 2001. Afforestation of bottomland hardwoods in the Lower Mississippi Alluvial Valley: status and trends. Wetlands 21, 602–613. Shoch, D.T., Kaster, G., Hohl, A., Souter, R., 2009. Carbon storage of bottomland hardwood afforestation in the Lower Mississippi Valley, USA. Wetlands 29, 535–542. Suchenwirth, L., Förster, M., Cierjacks, A., Lang, F., Kleinschmit, B., 2012. Knowledgebased classification of remote sensing data for the estimation of below- and above-ground organic carbon stocks in riparian forests. Wetl. Ecol. Manage. 20, 151–163. Turner, D.P., Koerper, G.J., Harmon, M.E., Lee, J.J., 1995. A carbon budget for forests of the conterminous United States. Ecol. Appl. 5, 421–436. Turner, R.E., Rabalais, N.N., 2003. Linking landscape and water quality in the Mississippi River Basin for 200 years. BioScience 53, 563–572. Zedler, J.B., Kercher, S., 2005. Wetland resources: status, trends, ecosystem services, and restorability. Annu. Rev. Environ. Resour. 30, 39–74.