Skip to main content
NCBI home page
iScience logoLink to iScience
. 2021 Mar 31;24(5):102382. doi: 10.1016/j.isci.2021.102382

The role of short-term and long-term water level and wave variability in coastal carbon budgets

Katherine N Braun 1, Ethan J Theuerkauf 2,3,∗
PMCID: PMC8091053  PMID: 33997674

Summary

We investigated soil organic carbon dynamics at three freshwater coastal sites in the Laurentian Great Lakes using a simple carbon budget box model. Long-term carbon budgets (1939–2018) were developed using aerial photography and then compared to short-term carbon export (2018–2019) developed using drone data. This study puts forth a refined coastal carbon budget model that advances previous model iterations by: (1) examining spatial variability in carbon budgets, (2) including a temporally dynamic carbon inventory term, and (3) updating the erosional term. Half of the initial carbon stock of the combined sites was lost in the 80-year study period, which is severely imbalanced with the age of those coastal habitats (400–2000 cal years BP). Major periods of carbon loss corresponded to periods of elevated water level. Short-term loss of carbon during 2018–2019 corresponded to northeasterly extreme wave events during a period of above-average water level.

Subject areas: Regional Geology, Soil Science, Erosion, Environmental Management

Graphical abstract

graphic file with name fx1.jpg

Open in a new tab

Highlights

  • •

    A coastal carbon budget model was refined to account for spatiotemporal heterogeneity

  • •

    Half the soil organic carbon stored at the study sites was exported in 80 years

  • •

    Carbon loss occurs during decadal periods of water level rise

  • •

    High wave events and/or elevated water level cause carbon loss

Regional Geology; Soil Science; Erosion; Environmental Management

Introduction

Soil organic carbon (SOC) represents a significant component of the global carbon cycle (Lal, 2004; Smith, 2008); Beyond the utility of soil carbon fluxes for understanding broad scale Earth system processes, SOC dynamics are important to incorporate into land management decisions as conversion and degradation of habitat reduces economically and ecologically valuable soil carbon stocks (Smith, 2008). Additionally, SOC provides numerous benefits to soil health including stabilization of the soil matrix, storage of water, and provision of plant nutrient reservoirs (Lal, 2004; Milne et al., 2015). Quantification of the rates and processes of SOC fluxes is required in order to properly manage coastal systems with respect to ecosystem services.

Erosion of coastal habitats reduces SOC stocks as organic-rich soils are swept into marine and lacustrine environments (Braun et al., 2019; Pendleton et al., 2012; Theuerkauf and Rodriguez, 2017). A significant body of work has addressed the role of coastal erosion in the soil carbon dynamics of marine coasts both in temperate and cold-climate regions (e.g., Coverdale et al., 2014; Jorgenson and Brown, 2005; Vonk et al., 2012). Erosion of permafrost coasts in the Arctic has been shown to be a significant source of carbon to the Arctic seas (Grigoriev and Rachold, 2003; Jorgenson and Brown, 2005; Tanski et al, 2016, Tanski et al., 2019). Fritz et al. (2017) calculated that the contribution of carbon from eroding Arctic coasts is on the same order of magnitude as the flux from all Arctic rivers (Fritz et al., 2017). Saltmarshes contain large stores of soil carbon, but are rapidly losing area due to human activity, climate change, and coastal erosion (Hopkinson et al., 2012). Numerous studies have calculated the carbon budget of saltmarshes and found that rates of carbon loss are disproportionate to the time it took that carbon to accumulate (Coverdale et al., 2014; Sapkota and White, 2019; Theuerkauf et al., 2015; Theuerkauf and Rodriguez, 2017). Likewise, mangroves contain large stocks of carbon that are being lost due to land conversion and sea level rise (Atwood et al., 2017; Donato et al., 2011; Kauffman et al., 2011; Pendleton et al., 2012). In all of these coastal environments, erosion plays a critical role in both exporting once stored carbon and reducing the area of coastal habitats available for carbon storage.

The connection between coastal erosion and carbon budgets of large lacustrine systems such as the Laurentian Great Lakes is poorly understood, although these systems experience high rates of coastal erosion similar to marine and estuarine coasts (Kimball May et al., 1983; Meadows et al., 1997). Coastal erosion is a significant problem in the Great Lakes as it destroys infrastructure and property (Angel, 1995) as well as natural habitat, parks, and recreational areas, many of which contain important carbon-storing habitats such as coastal wetlands. Along Lake Michigan alone, 22% of the shoreline is classified as open-shore wetlands (Office of Coastal Management, 2020, https://www.fisheries.noaa.gov/inport/item/59439), which are vulnerable to erosional loss (e.g. Braun et al., 2019). Great Lakes coastlines provide numerous ecosystem services, including sport fishing, boating, and recreational use (Allan et al., 2015), in addition to the economic and ecologic value of the SOC stored within shoreline habitats. The total quantity of SOC lost due to coastal erosion in the Great Lakes is unknown, however, previous studies in the region suggest that the losses are episodic in response to storm events and high water levels and that carbon is being lost on orders of magnitude shorter timescales than the amount of time it took to accumulate that carbon (Braun et al., 2019). These losses will lead to a permanent reduction in the carbon stock through time if carbon accumulation rates do not increase to balance export or if habitat lost is not balanced by gain in new habitat area.

The episodic nature of SOC budgets in the Great Lakes in response to storms and water level fluctuations necessitates the use of a geomorphic process-driven SOC budget model to explore carbon dynamics through time. While a few studies have examined SOC in Great Lakes coastal wetlands (Bernal and Mitsch, 2008; Braun et al, 2019, 2020), no studies exist exploring SOC dynamics beyond wetland habitats in the region. This represents an important gap in documenting regional SOC budgets as a large portion of the coastal habitats in the Great Lakes region are not wetlands, but rather other SOC-containing environments, such as prairies and savannas. Additionally, these studies do not model carbon storage using a temporally dynamic carbon inventory term nor allow for alongshore variability in the area of SOC-containing habitats through time. In order to address these gaps and improve our understanding of the temporal and spatial dynamics of coastal SOC budgets, we modified the Braun et al. (2019) model to: (1) incorporate spatial variability in carbon storage and areal change in various coastal habitats, (2) include a temporally dynamic carbon inventory term, and (3) enhance the capabilities of the model to realistically capture the dynamics of coastal erosion and SOC export (full model details are presented in the transparent methods, also see Figure S2).

We use this model to evaluate changes in SOC budgets from 1939 to 2019 at a 2.75 km stretch of shoreline along western Lake Michigan that exhibits a high degree of spatial and temporal variability in shoreline retreat rates, SOC storage rates, and coastal habitat types. Illinois Beach State Park (IBSP) is a 1,680-hectare park situation on the Zion Beach-ridge Plain along the shore of Lake Michigan in northeastern Illinois, USA (Figure 1). The Zion Beach-ridge Plain is a dynamic sand body composed of curvilinear beach ridges and swales. This ∼3,700-year-old landscape is dominated by erosion along the northern two-thirds of the park and accretion along the southern third due to the predominantly north to south longshore current (Chrzastowski et al., 1994). Waves and fluctuating water levels are the dominant hydrodynamic processes that generate change and ultimately alter SOC budgets along most coastal environments. In the Great Lakes, water levels fluctuate dramatically across a range of timescales from storm events to millennia (As-Salek and Schwab, 2004; Gronewold and Rood, 2019; Thompson and Baedke, 1997). On timescales relevant to management (seasonal, annual, decadal), Lake Michigan fluctuates in a ∼2 m range, with the long-term average water level at 176.606 m NAVD88 (Figure 2).

Figure 1.

Figure 1

Open in a new tab

The three study sites are located along the southwestern shore of Lake Michigan, Illinois, USA

(A) Location of the three study sites along the North Unit of Illinois Beach State Park.

(B) sUAS-derived orthomosaics of the three study sites. Gray boundaries mark the habitat areas. Circles indicate coring locations and are color-coded to show the carbon inventory of each core.

(C) Location of study area in United States of America. Aerial image in panel a downloaded from the Lake County, Illinois Planning, Building, and Development Department.

See also Figure S3 and Table S1.

Figure 2.

Figure 2

Open in a new tab

The greatest loss in carbon stock occurs during intervals of water level rise

Plots depicting the change in carbon stock and change in the water level of Lake Michigan from 1939 to 2019 (left) and the short-term monitoring record from 2018 to 2019 (right). Shaded blue indicates error of the carbon stock, measured using high and low values for the carbon inventory, wetland age, and geospatial analysis. See also Figure S2 and Table S2.

The updated SOC budget model is flexible; it can be applied over large and small spatial and temporal scales to identify trends in SOC, in addition to identifying carbon budget responses to events such as storms or fluctuations in water level. Given this, it can provide important data on SOC dynamics in coastal systems that can be used for management of all SOC-containing shoreline habitats, not just wetlands or those within the Great Lakes region. Additionally, this model allows land managers of any coastal setting to pinpoint areas of high value and vulnerability by identifying hot spots of carbon accumulation and loss.

Results

Coastal habitat area significantly declined over the past 80 years

Carbon storage and export is calculated based on the area of habitat present or eroded. The rate of change in areal extent of habitat for the entire study period and area was −3,569 m2/year, or −1.3 m2/year per meter of shoreline. The rate of habitat area change for the long-term record, 1939 to 2018, was −4,423 m2/year (−1.6 m2/year per meter of shoreline); the rate for the short-term record, 2018–2019, was −13,101 m2/year (−4.8 m2/year per meter of shoreline). The rate of habitat area change in the long-term record was positive only during the record low water levels of the late 1990s and early 2000s, between 1997 and 2000, 2002–2004, and 2007–2008. In the short-term record, the rate of habitat change was positive during the summer growing season of 2019, between April 25 and June 17, 2019, and July 10-26, 2019. During periods of habitat loss, the average rate of loss was −14,098 m2/year; during periods of habitat gain, the average rate was 2,824 m2/year, which highlights the order of magnitude difference between the formation and destruction of habitats and associated carbon storage.

The highest rates of habitat change seen during the study period occurred during the short-term, high-resolution record. This is likely in part due to near record high lake levels, but also due to the increased fidelity of the record. It is likely that during other erosional periods in the past similarly high rates of loss would have been documented. The highest rate of change was −48,890 m2/year (−17.8 m2/year per meter of shoreline), between August 31 and September 11, 2018. Other periods of high magnitude habitat loss include 2017–2018 (−24,336 m2/year; −8.8 m2/year per meter of shoreline), October 24-November 5, 2018 (−36,274 m2/year; −13.2 m2/year per meter of shoreline), November 5-December 18, 2018 (−43,865 m2/year; −16.0 m2/year per meter of shoreline), and March 28-April 25, 2019 (−41,947 m2/year; −15.3 m2/year per meter of shoreline).

Coastal habitats lost an order of magnitude more SOC than was stored in 80 years

All three study sites lost more carbon than they stored over the long-term study period, from 1939 to July 26, 2019. Combined carbon export by erosion at all sites (1939 area: 0.56 km2; 2019 area: 0.20 km2) over the entire study period was 4,420 MgC, while combined carbon storage was 579 MgC. We define carbon export as the product of the amount of soil carbon contained per unit area and area of habitat eroded; carbon storage is the product of the carbon accumulation rate and the total amount of active habitat (see transparent methods for model details). Given the imbalance between carbon stored and carbon exported, these sites collectively are carbon sources on decadal timescales. The carbon stock for the entire study area decreased over the 80-year period, from a total of 7,504 MgC to 3,663 McC, a loss of 51% of the initial stock (Figure 2); this amount is a carbon budget rate of −0.09 kgC/m2/year (Figure 5). Examining the site-specific budgets allows for a more detailed analysis of the variability in SOC dynamics as they relate to geomorphic change and habitat type.

Figure 5.

Figure 5

Open in a new tab

The 80-year carbon budget is negative

Top: carbon storage rate (green), carbon export rate (red), and carbon budget rate (black) for the decadal record, 1939–2019. Error represented by shaded areas. Bottom: water level.

All study sites were carbon sources throughout most of the 80-year record, however, there was substantial temporal and spatial variability in the magnitude of carbon storage and export. Site 1 lost the most carbon of all the sites, 2,322 MgC, which is a carbon budget rate of −0.12 kgC/m2/year. Site 2 lost the highest percentage of its original stock, 88%, which was 1,463 MgC; this site has a carbon budget rate of −0.07 kgC/m2/year. Site 3 lost the least carbon, 635 MgC, which is a carbon budget rate of −0.06 kgC/m2/year. While the sites all lost more carbon than was stored over the entire 80-year study, during some time-steps individual sites stored more carbon than was exported. The 2008–2009 period is the only time when the total carbon budget of all sites combined was positive. The carbon stock increased 0.02% during 2008–2009 as 1.66 MgC was gained between all three sites. Site 1 had a positive carbon budget between: 1974–2002, 2004–2005, and 2006–2009. Site 2 only had a positive carbon budget in 2012–2014. Site 3 had a positive carbon budget between 1939 and 1946, 1993–2000, 2002–2004, and 2008–2010. Even though there were times in this study when a given site had a positive budget, the fact that the other sites had negative carbon budgets during that same time offset any gains when considering the whole study area. For example, combining all of these periods of positive carbon budgets together, these sites had a cumulative carbon budget of 97 MgC, which is an increase of 1% in the carbon stock. During these same periods at the other sites with negative carbon budgets, 1,069 MgC was lost, which is a decrease of 14% in the carbon stock. This offset between carbon gain and loss across the sites can also be observed at discrete time intervals. For example, though site 1 gained 47 MgC between 1974 and 1993, site 2 and site 3 lost 339 MgC in that same period.

The cumulative amounts of carbon exported and stored at each site were plotted against each other to illustrate the relationship between these two fluxes over the decadal record (Figure 3). While the carbon budget was occasionally positive, the cumulative amount of carbon exported was always greater than the cumulative amount of carbon stored, except at site 3 between 1939 and 1961. Habitat at site 3 during this period migrated lakeward due to colonization of the wide sandy beach that had been previously present, allowing the site to accumulate more carbon than was exported during this time. The cumulative amount of carbon exported from all sites during the study period is 7.6 times greater than the cumulative amount of carbon stored.

Figure 3.

Figure 3

Open in a new tab

Cumulative carbon export exceeds carbon storage by seven times

The red line indicates the total amount of carbon exported. The blue line indicates total amount of carbon stored. Shading indicates error of the carbon stored and exported, measured using high and low values for the carbon inventory, wetland age, and geospatial analysis. See also Figure S2 and Table S2.

High water level and increased wave heights align with high carbon export events

Lake Michigan water level data were analyzed for the entire study period and compared with the carbon budget dynamics. The greatest loss in the carbon stock at these sites in the long-term record is seen during periods when water level rises. The largest increase in Lake Michigan water level, +2.1 m, occurs between 1961 and 1974 when the carbon stock dropped by the largest percent, 16%. A similar pattern appears in the data for the most recent rise in water level, between 2014 and 2019: water level gained +1.5 m, while the carbon stock decreased by 14%. The sites lost 8% of the carbon stock in 1946–1961, during a +1.6 m rise in water level. A decrease in the carbon stock during the record highs of the 1980s is seen at sites 2 and 3, but not at site 1, which received sediment nourishment from the construction of North Point Marina (completed in 1989). The carbon stock of all sites decreased 2% during the record low water levels of the early 2000s (2000–2012).

No major increase in carbon storage nor recovery of the carbon stock appears to occur when water level falls. The only time-step we analyzed that fully encompasses a fall in water level is 1997–2000, when water level peaked at 177.30 m IGLD in July 1997, and reached a low of 175.75 m IGLD in December 1999, a fall of 1.55 m. At all sites during this time, the carbon budget is negative; this indicates that although the water level was falling to near record lows, carbon was still being exported more than it was stored. In other time-steps with major decreases in water level (1946–1961, 1961–1974, and 1974–1993), any gain in carbon storage during falling water levels was outweighed by the amount of carbon lost through erosion during subsequent time periods (Figure 3).

The high-resolution 2018–2019 record shows that seasonal trends in carbon loss at the study sites are controlled by water level and wave dynamics. Carbon export is reduced in the winter months and increases during the spring, summer, and fall seasons (Figure 3). At all three sites, carbon export leveled off around October-November 2018, as water level fell ∼0.15 m from the summer 2018 high, and then export began to increase again in March 2019 as water level rose ∼0.55 m.

Wave data for the 2018–2019 short-term record were also compared to the carbon export rate. The average significant wave height for the short-term record was 0.42 m; the 98th percentile of significant wave height was 1.31 m. The time-steps with the greatest percentage of extreme waves (>98th percentile) were August 31, 2018-September 11, 2018 (13%), October 24, 2018-November 5, 2018 (5.4%), November 5, 2018-December 18, 2018 (5.3%), and March 28, 2019-April 25, 2019 (5.0%). Rose diagrams of wave direction were also produced (Figure 4).

Figure 4.

Figure 4

Open in a new tab

The greatest export of carbon occurs during northeasterly extreme wave events

Top: carbon export rate (red) and percentage of waves above the extreme threshold (98th percentile of all waves; in blue), with red shading indicating periods that have wave rose charts depicted below. Middle: carbon export rate (red) and average water level (black), with red shading indicating periods that have wave rose charts depicted below. Bottom: Wave rose charts showing the extreme onshore wave events for each highlighted time bin.

Extreme wave events align with periods of high carbon export and habitat loss. High rates of carbon export and habitat loss between August 31, 2018 and September 11, 2018 occurred during a large wave event. A similar pattern appears between March 28, 2019 and April 25, 2019. While there are few extreme wave events following April 25, 2019, both the carbon export rate and habitat change rate remain elevated, which tracks the seasonal rise in water level during the spring. There is a similar percentage of extreme wave events between October 24, 2018 and December 18, 2018 as during the spring 2019 storm season, but little carbon export occurs during this time. Wave direction differs between these two periods (Figure 4). During the spring stormy period, onshore waves come primarily from the east, with an average direction of extreme waves of 75°. In the fall 2018 storms, wave direction is more southerly, with an average direction of extreme waves of 130° in October 24, 2018 – November 5, 2018, and 115° in November 5, 2018 – December 18, 2018. While the carbon export rate lowered during the fall 2018 storms, the habitat area change rate remained elevated due to enhanced overwash burial of habitat. The sUAS imagery during this period reveals sustained growth of washover fans in response to the fall 2018 storms.

Discussion

Carbon loss through time

Coastal habitats at this study area lost seven times the amount of carbon they stored over the past 80 years. This imbalance between carbon export and carbon storage makes all of these sites carbon sources rather than carbon sinks across decadal timescales. Without considering the coastal geomorphic processes that are at work, the carbon dynamics of these habitats cannot be evaluated accurately and they would likely erroneously be considered carbon sinks given the mere presence of wetlands and other habitats known for their carbon-containing soils. The carbon stock at the three study sites diminished from 7,504 MgC to 3,663 MgC between 1939 and July 2019, a rate of loss of −0.09 MgC/m2/year (Figures 2 and 5).

This study only documented carbon dynamics along a 2.75 km stretch of shoreline, while the Great Lakes as a whole has over 16,000 km of shoreline. Not all sections of the Great Lakes shoreline are composed of the same habitats as IBSP nor do all stretches of coastline erode at the same rates. However, estimates indicate that 85% of the total Great Lakes shoreline is not hardened and that long-term retreat rates are on average −0.7 m/year, which potentially puts coastal habitats at risk of eroding (Kimball May et al., 1983; Schneider et al., 2007). If we combine these data with the carbon inventory of the least carbon-rich habitat at our study sites, mesic sand prairie, we estimate that the Great Lakes region has lost ∼5.1 TgC from coastal habitats over the past 80 years. Given the lack of data on Great Lakes coastal soil carbon content and erosion rates, this estimate is not meant to be taken as an exact value but rather highlights the significant role coastal erosion in the Great Lakes plays in carbon cycling. From the results of our study and others, coastal erosion clearly has the potential to generate significant fluxes of carbon and associated economic and ecologic impacts.

Carbon is lost from coastal habitats through shoreline erosion when coastal processes, such as storm waves or fluctuating water levels, erode these SOC-containing environments. While studies have examined the erosional loss of coastal soil carbon (Ganju et al., 2019; Sapkota and White, 2019) and the global impact of that carbon loss (Pendleton et al., 2012), the pathways and degradation of coastal SOC are complex and not completely quantified (Spivak et al., 2019). Depending on physical and biogeochemical conditions, the carbon may be remineralized and respired as atmospheric CO2 or redeposited in other carbon pools (Hayes et al., 2021, Sapkota and White, 2021, Tranvik et al., 2009) No matter the fate of this eroded carbon, it is no longer stored in the coastal ecosystems where it originated, thus, natural resource managers must critically evaluate whether a site truly is a carbon sink based on the dominant geomorphic processes. The long-term erosional trend at IBSP indicates that coastal ecosystems and associated carbon lost is permanent; over decadal time periods the land and carbon lost is not regained, even during periods of low water levels. Therefore, carbon storage as well as other ecosystem services associated with the eroded habitat are permanently lost during these high water level periods, which necessitates revaluation of whether a given site should be prioritized for shoreline protection due to habitat conservation needs or whether managed retreat should be allowed.

Hydrodynamics

The previous version of this carbon budget model published in Braun et al. (2019) identified short-term impacts of storm events and seasonal water level fluctuations on carbon budgets in coastal wetlands. However, that model did not contain the complexity necessary to evaluate the processes influencing coastal carbon budgets over larger temporal and spatial scales. A broader spatial and temporal context is required to evaluate the role that long-term water level fluctuations play on coastal habitat loss as well as carbon budgets. The expansion of the Braun et al. (2019) model to include a longer timescale, a broader spatial context, a refined carbon export term, and a temporally dynamic carbon inventory term allows us to identify what conditions during the past century led to carbon loss at these sites. This framework can be extrapolated to sites throughout the Great Lakes region as well as globally to assist coastal managers in identifying shoreline areas that are vulnerable to carbon loss on a range of timescales.

Carbon loss at our study sites generally tracks annual and decadal fluctuations in Lake Michigan water levels. Carbon export rates at all sites and time frames (short and long-term) in this study occurred jointly with increasing water level (Figure 5). While saltmarshes tend to experience reduced erosion rates when high water level leads to waves breaking across the marsh platform rather than at the marsh edge (Tonelli et al., 2010, Valentine and Mariotti, 2019), Great Lakes coastal habitats experience increased edge erosion rates during periods of higher water level (Angel, 1995, Meadows et al., 1997, Theuerkauf et al., 2019). Higher water levels amplify the impacts of waves and storms as energy is consistently delivered at higher elevations on the shore profile and therefore can reach larger portions of coastal habitat and soil. Thus, increased export of carbon can occur even during periods with low-magnitude shifts in water level when the base water level is high and wave energy is driven farther up the shoreface.

The high-resolution 2018–2019 data set allows us to examine linkages between hydrodynamics and carbon export on a monthly scale. These data show that carbon is exported during periods with high water level and northeasterly extreme wave events (Figure 4). The short-term record contains three periods with extreme waves, August-September 2018, October-December 2018, and March-April 2019. The October-December 2018 events did not cause similar levels of carbon export as the other two events, even though the late fall storms had maximum wave heights 0.58 m (30%) higher than the other events. These late fall storm events differed from the other storms in two ways: water level was over 0.10 m lower than the other events, and the waves originated from a more southerly direction which generates less erosion than northerly waves given the shoreline geometry of IBSP. It is the combination of these two factors that attenuated the impact of these storms on coastal carbon export. Although the majority of the waves during this period were south to southeasterly, the November 5, 2018 – December 18, 2018 period saw 40% of the extreme waves originate from the northeast to east. These easterly extreme waves, however, did not cause the extensive carbon export that similar waves did during the September 2018 and April 2019 events likely due to the lower water level in the late fall. Wave energy from this event was reaching portions of the shoreface that had already eroded during the event in September 2018. The high-resolution short-term record reveals that large carbon export events occur during extreme wave events when water level is elevated. These findings reiterate the importance of annual fluctuations in water level to coastal habitats and carbon budgets.

Although storm events and anthropogenic modification impact carbon export on shorter timescales, such as the storm around September 7th, 2018 or the construction of North Point Marina in the late 1980s, on longer timescales it is the annual and decadal fluctuations in water level that impact the areal extent of coastal habitats and therefore the carbon budgets. As the extreme wave events in the short-term record revealed, high water level enhances the erosional impact of a storm event. Water level is the driver of major coastal erosion and carbon loss, and this relationship should be taken into account when planning long-term management of Great Lakes coastal sites like IBSP.

The Braun et al. (2019) carbon budget model showed that wetland carbon budgets at IBSP are controlled primarily by erosional events and that the reduction in carbon storage potential by overwash played a relatively minor role in reducing carbon storage compared to erosion. Over the longer time frame examined in this study we found further evidence of this trend. The amount of carbon exported from shoreline habitats over the 80-year study period was seven times larger than the amount of carbon stored. Any reduction in carbon storage due to overwash is outweighed by the larger impact of erosion and associated carbon export on the carbon stock (Figure 3). Given the influence of erosion on decadal carbon budgets, it also appears that any growth of vegetation on washover deposits has only a minimal positive impact on overall carbon budgets.

While all sites experienced carbon loss during periods of high water level, the impact of the high water levels in the 1980s on the carbon stock is less than other periods of high water level. This response is unexpected given the dramatic erosion that was documented throughout the Lake Michigan basin during this period (Angel, 1995; Meadows et al., 1997). The attenuated impact on carbon stocks during this period is likely due to the effects of updrift human disturbance associated with the construction of North Point Marina in the late 1980s. During and following the construction of North Point Marina, erosion of fill material, as well as the placement of excess sand led to artificial nourishment of the beach area south of the marina (Chrzastowski et al., 1996), which is the northern boundary of site 1. The lack of change in the carbon stock between 1974 and 1993 at site 1 is likely due to the pulse of new sand into the littoral system in the late 1980s, as it would have protected the shoreline during the years of high water level. The period between 1974 and 1993 includes the fall from the high water level of the mid-seventies, the rise to the 1986 record, the fall to a low in 1990, and a gradual rise following that low. Additional aerial imagery of the intervening years between 1974 and 1993 would allow for a more detailed evaluation of the interplay between high water levels, anthropogenic influence, and shoreline habitat change; however, the limited habitat loss and carbon export documented across this entire period suggests that the longer-term impacts of this human disturbance overwhelms any shorter-term erosion that may have occurred during peak water levels.

Carbon budget model updates

The updated model presented in this manuscript evaluates the capacity of a coastal site to store carbon in different habitat types over long temporal scales. The updated carbon export term more realistically reflects the process of carbon loss due to erosional ravinement on the shoreface by using 3D topographic data. The use of spatially averaged carbon accumulation rates and carbon inventories as inputs allows application of the model across larger spatial scales and accounts for heterogeneity within study sites. The temporally dynamic carbon inventory term allows use of the model across decadal timescales. Finally, the parameterization of the model using geospatial data acquired from aerial images reflects the shift toward UAS-data collection in coastal systems (Johnston, 2018) and supports the use of publicly available satellite and aerial imagery data sets to parameterize the model rather than relying on site-specific field data.

The expansion of the model from a wetland-focus to a more holistic landscape focus allows users to examine carbon dynamics across wider areas that are more relevant to management. As this model is suitable for use across broader spatial scales, it can be used to document linkages between coastal processes, landscape impacts, and carbon budgets. For example, the areas in this study site with the most rapid change and greatest loss of carbon align with high-risk areas for coastal erosion defined in Theuerkauf et al. (2019). Site 2 is immediately downdrift of a hardened shoreline and experienced the greatest loss in land over the 80-year study period (17 hectares, 74% of initial area), and in percent of stock lost (88%, 1463 MgC).

Applications for coastal management

The rate at which carbon is exported from these coastal sites is not in balance with the amount of time it took to accumulate these stocks of carbon. The base of the soil organic unit at site 1 was dated to 1934 cal years BP in the north of the site, and 830 cal years BP in the south; sites 2 and 3 were dated to 394 cal years BP. Half of the carbon stored in the three study sites was lost in 80 years, an order of magnitude faster than the shortest amount of time it took to form these carbon stocks. Additional stocks of soil carbon exist in other habitats at IBSP beyond the areal extent of this study, both along the rest of the shoreline at IBSP as well as landward of the study sites. We documented 36 hectares of coastal habitat lost at the three study sites in 1939–2019; this loss represents 2% of the 1680 hectares of IBSP, not taking into account erosion occurring along the 7.75 km of shoreline not included in this study. While coastal erosion and habitat loss is a known concern for coastal managers throughout the Great Lakes region, little has been done historically beyond site-specific shore protection to manage the issue of coastal habitat loss, in part due to the difficulty of deciding where to devote limited resources. Our study provides context on the value of different coastal habitats through the lens of carbon storage, which can be used to prioritize shoreline protection efforts in the context of habitat conservation. The results of our carbon budget show that managers must be accounting for soil carbon loss in coastal systems and cannot assume that these habitats contain sustained carbon pools without examining carbon loss due to erosion.

At our study site, high carbon value areas are located at sites 1 and 3. We define high carbon value as locations with carbon inventories and/or carbon accumulation rates that are high relative to the average values (Figure 6). Habitats with high carbon inventories have large stocks of soil carbon, while habitats with high carbon accumulation rates are able to store carbon faster and therefore recover faster from erosive events. Freshwater marsh and wet sand prairie are the two habitat types in this study with the highest carbon inventories (73,565 and 45,163 gC/m2 respectively). Freshwater marsh, panne, and wet sand prairie have the highest carbon accumulation rates (82, 38, and 27 gC/m2/year respectively). While site 2 contains high value pannes, those pannes lost almost all of their original extent by the end of the study period. Given the minimal area left of panne at this site, as well as the likelihood that the loss is permanent given the trends in historical habitat loss in the area, this site would not be an ideal location for conservation efforts if carbon storage is the management priority. Site 1 contains a large freshwater marsh that began to fill in with washover sediment during the 2018-2019 study period yet remained relatively intact by the end of the study. Site 3 contains large, intact panne habitats along the shoreline that were eroding and filling with washover sediment during the study period. Shoreline protection measures should focus on the high value wet sand prairies and freshwater marsh at site 1 and the pannes at site 3 in order to conserve the carbon stock and storage potential of these habitats.

Figure 6.

Figure 6

Open in a new tab

Eighty years of coastal habitat and carbon loss

Total habitat area lost between 1939 and 2019, with bubbles indicating the loss in carbon stock during that period in each habitat type for each site. The height of the bubble corresponds to the average carbon accumulation rate for that habitat type. Aerial image is from 2017, downloaded from the Lake County, Illinois Planning, Building, and Development Department. See also Tables S4–S10.

Limitations of the study

Study results are limited by the errors and assumptions of the carbon budget model and its inputs. Model error, represented in figures by shading, was determined by running the model with the high and low values for the carbon inventory (standard deviation; Table S4), carbon accumulation rates (radiocarbon date error; Table S4), and geospatial data (see transparent methods). Averaging carbon inventories across habitat types obscures some small-scale (100s m2) variation in carbon inventory in favor of creating a carbon inventory that is acceptable for use across the entire study sites (1000s-10,000s m2). The use of a static carbon accumulation rate assumes that carbon storage in these habitats is linear over decadal timescales, when seasonality is known to cause variations in the carbon accumulation rates in temperate locations. Defining precise boundaries between habitat types over decades introduces error as landscapes evolve and such hard boundaries do not exist in natural settings; this factor was attenuated by manual editing of boundaries when such differences were clearly visible on historical photographs.

Assessing carbon budgets at different scales requires different levels of data. Larger regional budgets make greater assumptions about spatial heterogeneity in carbon stocks in order to track larger-scale patterns; our parameterization of the carbon budget at three relatively small sites allowed us to address the variability in carbon dynamics between different habitats. Site-level studies, such as this one, can capture precise relationships between processes and responses that can later be applied at greater scales; carbon-content data collected for this study can be averaged with other Great Lakes studies to produce regional carbon budgets.

Conclusion

Shoreline erosion can deplete the carbon stock of coastal habitat orders of magnitude faster than the time it took to generate those stocks. This was observed at Illinois Beach State Park where three study sites lost half of the SOC stored within their habitats in the 80 year study period. On decadal timescales, major erosion of carbon occurs during periods of high water level. The high-resolution carbon export data for 2018–2019 showed that major carbon export episodes occur during northeasterly extreme wave events paired with elevated water levels. Low or falling water levels do not lead to increases in the carbon budget, thus these sites are sources of carbon on decadal timescales. While the dominant pattern of the carbon budget at IBSP was loss, the introduction of nourishment sediment appeared to temporarily prevent carbon loss, as seen at site 1 during the record high water levels of the late 1980s. The flux of carbon entering Lake Michigan from these coastal habitats is significant and should be accounted for in regional carbon cycling models. The output of this carbon budget model can assist in the prioritization of conservation efforts by identifying sites with high carbon stocks and carbon accumulation rates.

Resource availability

Lead contact

Further information and requests for resources and materials should be directed to and will be fulfilled by the lead contact, Ethan Theuerkauf (theuerk5@msu.edu).

Materials availability

Direct material requests to lead contact, Ethan Theuerkauf.

Data and code availability

Data presented in this manuscript are publicly available through the Illinois Geospatial Data Clearinghouse under the Coastal category: https://clearinghouse.isgs.illinois.edu/data?field_data_type_value=Coastal. Data presented in this manuscript include "Coastal Habitat Maps along Illinois Beach State Park, derived from aerial and drone imagery, 1939-2020" and "Orthomosaic, digital elevation model, and point cloud derived from unoccupied aerial system (UAS) imagery, Illinois Beach State Park."

Methods

All methods can be found in the accompanying transparent methods supplemental file.

Acknowledgments

We thank Jenny Bueno, Kevin Engelbert, and Cesar Gutierrez for help with fieldwork and processing of sUAS and sediment samples. Thanks to Brandon Curry for help with preparing samples for radiocarbon dating. We also thank Andrew Masterson for his help at the Northwestern Stable Isotope Laboratory. Habitat delineation data was provided by the Illinois Department of Natural Resources, Illinois Nature Preserves Commission, Illinois Endangered Species Protection Board, and the Natural Heritage Database in August 2018. Funding for this project was provided by the Great Lakes Restoration Initiative through a grant from the National Oceanic and Atmospheric Administration. This grant was subawarded to the University of Illinois at Urbana-Champaign through Woolpert.

Author contributions

E.T. procured funding, designed the study, conducted hydrodynamic data processing, and supervised all other data processing and analysis. K.B. created habitat maps and conducted carbon analysis. E.T. and K.B. jointly performed field work, updated model, analyzed data, and wrote the manuscript.

Declaration of interests

The authors declare no competing interests.

Published: May 21, 2021

Footnotes

Supplemental information can be found online at https://doi.org/10.1016/j.isci.2021.102382.

Supplemental information

Document S1. Transparent methods, Figures S1–S3, and Tables S1–S10
mmc1.pdf (1.2MB, pdf)

References

  1. Allan J.D., Smith S.D.P., McIntyre P.B., Joseph C.A., Dickinson C.E., Marino A.L., Biel R.G., Olson J.C., Doran P.J., Rutherford E.S. Using cultural ecosystem services to inform restoration priorities in the Laurentian Great Lakes. Front. Ecol. Environ. 2015;13:418–424. doi: 10.1890/140328. [DOI] [Google Scholar]
  2. Angel J.R. Large-scale storm damage on the US shores of the Great Lakes. J. Great Lakes Res. 1995;21:287–293. doi: 10.1016/S0380-1330(95)71039-5. [DOI] [Google Scholar]
  3. As-Salek J., Schwab D.J. High-frequency water level fluctuations in Lake Michigan. J. Waterw. Port, Coast. Ocean Eng. 2004;130:45–53. doi: 10.1061/(ASCE)0733-950X(2004)130. [DOI] [Google Scholar]
  4. Atwood T.B., Connolly R.M., Almahasheer H., Carnell P.E., Duarte C.M., Lewis C.J.E., Irigoien X., Kelleway J.J., Lavery P.S., Macreadie P.I. Global patterns in mangrove soil carbon stocks and losses. Nat. Clim. Chang. 2017;7:523–528. doi: 10.1038/nclimate3326. [DOI] [Google Scholar]
  5. Bernal B., Mitsch W.J. A comparison of soil carbon pools and profiles in wetlands in Costa Rica and Ohio. Ecol. Eng. 2008;34:311–323. doi: 10.1016/j.ecoleng.2008.09.005. [DOI] [Google Scholar]
  6. Braun K.N., Theuerkauf E.J., Hurtgen M.T., Masterson A.L., Horton D.E. Loss-on-ignition estimates for soil organic carbon in Great lakes freshwater coastal wetlands. Wetlands. 2020;40:1201–1206. doi: 10.1007/s13157-020-01270-z. [DOI] [Google Scholar]
  7. Braun K.N., Theuerkauf E.J., Masterson A.L., Curry B.B., Horton D.E. Modeling organic carbon loss from a rapidly eroding freshwater coastal wetland. Sci. Rep. 2019;9:1–13. doi: 10.1038/s41598-019-40855-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Chrzastowski M.J., Foyle A.M., Trask C.B. Erosion and Accretion Trends Along the Lake Michigan Shore at North Point Marina and Illinois Beach State Park. Illinois State Geologial Survey Open File Series. 1996 http://hdl.handle.net/2142/45016 1996–1. [Google Scholar]
  9. Chrzastowski M.J., Thompson T.A., Brian Trask C. Coastal geomorphology and littoral cell divisions along the Illinois-Indiana coast of Lake Michigan. J. Great Lakes Res. 1994;20:27–43. doi: 10.1016/S0380-1330(94)71130-8. [DOI] [Google Scholar]
  10. Coverdale T.C., Brisson C.P., Young E.W., Yin S.F., Donnelly J.P., Bertness M.D. Indirect human impacts reverse centuries of carbon sequestration and salt marsh accretion. PLoS One. 2014;9:1–7. doi: 10.1371/journal.pone.0093296. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Donato D.C., Kauffman J.B., Murdiyarso D., Kurnianto S., Stidham M., Kanninen M. Mangroves among the most carbon-rich forests in the tropics. Nat. Geosci. 2011;4:293–297. doi: 10.1038/ngeo1123. [DOI] [Google Scholar]
  12. Fritz Michael, Vonk Jorien, Lantuit Hugues. Collapsing Arctic coastlines. Nature Climate Change. 2017;7:6–7. doi: 10.1038/nclimate3188. [DOI] [Google Scholar]
  13. Ganju N.K., Defne Z., Elsey-Quirk T., Moriarty J.M. Role of tidal wetland Stability in lateral fluxes of particulate organic matter and carbon. J. Geophys. Res. Biogeosci. 2019;124:1265–1277. doi: 10.1029/2018JG004920. [DOI] [Google Scholar]
  14. Grigoriev, M.N., Rachold, V. 2003. The degradation of coastal permafrost and the organic carbon balance of the Laptev and East Siberian Seas, in: Permafrost: Proceedings of the 8th International Conference on Permafrost, 21-25 July 2003, Zurich, Switzerland. pp. 319–324.
  15. Gronewold A.D., Rood R.B. Recent water level changes across Earth’s largest lake system and implications for future variability. J. Great Lakes Res. 2019;45:1–3. doi: 10.1016/j.jglr.2018.10.012. [DOI] [Google Scholar]
  16. Hayes M., Sapkota Y., White J., Cook R. Investigating the impact of in situ soil organic matter degradation through porewater spectroscopic analyses on marsh edge erosion. Chemosphere. 2021;268 doi: 10.1016/j.chemosphere.2020.129266. [DOI] [PubMed] [Google Scholar]
  17. Hopkinson C.S., Cai W.J., Hu X. Carbon sequestration in wetland dominated coastal systems-a global sink of rapidly diminishing magnitude. Curr. Opin. Environ. Sustain. 2012;4:186–194. doi: 10.1016/j.cosust.2012.03.005. [DOI] [Google Scholar]
  18. Johnston D.W. Unoccupied aircraft systems in marine science and conservation. Ann. Rev. Mar. Sci. 2018;11:439–463. doi: 10.1146/annurev-marine-010318-095323. [DOI] [PubMed] [Google Scholar]
  19. Jorgenson M.T., Brown J. Classification of the Alaskan Beaufort Sea Coast and estimation of carbon and sediment inputs from coastal erosion. Geo-Marine Lett. 2005;25:69–80. doi: 10.1007/s00367-004-0188-8. [DOI] [Google Scholar]
  20. Kauffman J.B., Heider C., Cole T.G., Dwire K.A., Donato D.C. Ecosystem carbon stocks of micronesian mangrove forests. Wetlands. 2011;31:343–352. doi: 10.1007/s13157-011-0148-9. [DOI] [Google Scholar]
  21. Kimball May S., Dolan R., Hayden B.P. Erosion of U.S. Shorelines. Eos (Washington. DC) 1983;64:521–523. [Google Scholar]
  22. Lal R. Soil carbon sequestration to mitigate climate change. Geoderma. 2004;123:1–22. doi: 10.1016/j.geoderma.2004.01.032. [DOI] [Google Scholar]
  23. Meadows G.A., Meadows L.A., Wood W.L., Hubertz J.M., Perlin M. The relationship between Great lakes water levels, wave Energies, and shoreline damage. Bull. Am. Meteorol. Soc. 1997;78:675–683. doi: 10.1175/1520-0477(1997)078<0675:TRBGLW>2.0.CO;2. [DOI] [Google Scholar]
  24. Milne E., Banwart S.A., Noellemeyer E., Abson D.J., Ballabio C., Bampa F., Bationo A., Batjes N.H., Bernoux M., Bhattacharyya T. Soil carbon, multiple benefits. Environ. Dev. 2015;13:33–38. doi: 10.1016/j.envdev.2014.11.005. [DOI] [Google Scholar]
  25. Pendleton L., Donato D.C., Murray B.C., Crooks S., Jenkins W.A., Sifleet S., Craft C., Fourqurean J.W., Kauffman J.B., Marbà N. Estimating global “blue carbon” Emissions from conversion and degradation of vegetated coastal ecosystems. PLoS One. 2012;7 doi: 10.1371/journal.pone.0043542. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Sapkota Y., White J.R. Marsh edge erosion and associated carbon dynamics in coastal Louisiana: a proxy for future wetland-dominated coastlines world-wide. Estuar. Coast. Shelf Sci. 2019;226:106289. doi: 10.1016/j.ecss.2019.106289. [DOI] [Google Scholar]
  27. Schneider, J., Heaton, D., Leadlay, H., 2007. Extent of Hardened Shoreline, in: State of the Great Lakes 2007. pp. 1–3.
  28. Sapkota Y., White J. Long-term fate of rapidly eroding carbon stock soil profiles in coastal wetlands. Science of the Total Environment. 2021;753 doi: 10.1016/j.scitotenv.2020.141913. [DOI] [PubMed] [Google Scholar]
  29. Smith P. Land use change and soil organic carbon dynamics. Nutr. Cycl. Agroecosyst. 2008;81:169–178. doi: 10.1007/s10705-007-9138-y. [DOI] [Google Scholar]
  30. Spivak A.C., Sanderman J., Bowen J.L., Canuel E.A., Hopkinson C.S. Global-change controls on soil-carbon accumulation and loss in coastal vegetated ecosystems. Nat. Geosci. 2019;12:685–692. doi: 10.1038/s41561-019-0435-2. [DOI] [Google Scholar]
  31. Tanski G., Couture N., Lantuit H., Eulenburg A., Fritz M. Eroding permafrost coasts release low amounts of dissolved organic carbon (DOC) from ground ice into the nearshore zone of the Arctic Ocean. Glob. Biogeochem. Cycles. 2016;30:1054–1068. doi: 10.1111/1462-2920.13280. [DOI] [Google Scholar]
  32. Tanski G., Wagner D., Knoblauch C., Fritz M., Sachs T., Lantuit H. Rapid CO 2 release from eroding permafrost in seawater. Geophys. Res. Lett. 2019;46:1–17. doi: 10.1029/2019GL084303. [DOI] [Google Scholar]
  33. Theuerkauf E., Braun K., Kaplan M., Vivirito S., Williams J., Nelson D. Coastal geomorphic response to seasonal lake level rise in the Laurentian Great Lakes, USA. Journal of Great Lakes Research. 2019;45(6):1055–1068. doi: 10.1016/j.jglr.2019.09.012. [DOI] [Google Scholar]
  34. Theuerkauf E.J., Rodriguez A.B. Placing barrier-island transgression in a blue-carbon context. Earth’s Future. 2017;5:789–810. doi: 10.1002/2017EF000568. [DOI] [Google Scholar]
  35. Theuerkauf E.J., Stephens J.D., Ridge J.T., Fodrie F.J., Rodriguez A.B. Carbon export from fringing saltmarsh shoreline erosion overwhelms carbon storage across a critical width threshold. Estuar. Coast. Shelf Sci. 2015;164:367–378. doi: 10.1016/j.ecss.2015.08.001. [DOI] [Google Scholar]
  36. Thompson T.A., Baedke S.J. Strand-plain evidence for late Holocene lake-level variations in Lake Michigan. Bull. Geol. Soc. Am. 1997;109:666–682. doi: 10.1130/0016-7606(1997)109<0666:SPEFLH>2.3.CO;2. [DOI] [Google Scholar]
  37. Tonelli M., Fagherazzi S., Petti M. Modeling wave impact on salt marsh boundaries. Journal of Geophysical Research. 2010;115:1–17. doi: 10.1029/2009JC006026. [DOI] [Google Scholar]
  38. Tranvik L.J., Downing J.A., Cotner J.B., Loiselle S.A., Striegl R.G., Ballatore T.J., Dillon P., Finlay K., Fortino K., Knoll L.B. Lakes and reservoirs as regulators of carbon cycling and climate. Limnol. Oceanogr. 2009;54:2298–2314. doi: 10.4319/lo.2009.54.6_part_2.2298. [DOI] [Google Scholar]
  39. Valentine K., Mariotti G. Wind-driven water level fluctuations drive marsh edge erosion variability in microtidal coastal bays. Continental Shelf Research. 2019;176:76–89. doi: 10.1016/j.csr.2019.03.002. [DOI] [Google Scholar]
  40. Vonk J.E., Sanchez-Garca L., Van Dongen B.E., Alling V., Kosmach D., Charkin A., Semiletov I.P., Dudarev O.V., Shakhova N., Roos P. Activation of old carbon by erosion of coastal and subsea permafrost in Arctic Siberia. Nature. 2012;489:137–140. doi: 10.1038/nature11392. [DOI] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Document S1. Transparent methods, Figures S1–S3, and Tables S1–S10
mmc1.pdf (1.2MB, pdf)

Data Availability Statement

Data presented in this manuscript are publicly available through the Illinois Geospatial Data Clearinghouse under the Coastal category: https://clearinghouse.isgs.illinois.edu/data?field_data_type_value=Coastal. Data presented in this manuscript include "Coastal Habitat Maps along Illinois Beach State Park, derived from aerial and drone imagery, 1939-2020" and "Orthomosaic, digital elevation model, and point cloud derived from unoccupied aerial system (UAS) imagery, Illinois Beach State Park."

Articles from iScience are provided here courtesy of Elsevier

RESOURCES