Longevity, climate sensitivity, and conservation status of wetland trees at Black River, North Carolina

The forested wetlands of Black River preserve one of the great natural areas of eastern North America (figures 1 and 2; 34.49686°N, 78.24265°W). Many living bald cypress (Taxodium distichum) trees at Black River are over 1,000-years old and new research demonstrates that some are over 2,000-years old, making these trees the oldest in eastern North America and as a species the fifth oldest sexually reproducing non-clonal tree taxa on earth. The Black River has been recognized as 'Outstanding Resource Water' by the North Carolina Department of Environment and Natural Resources and is one of the cleanest and high quality waterways in North Carolina (NCDPR 2018). There have been notable attempts to preserve this remarkable stream (Hart 1994), including recognition as a state scenic river, a national scenic river, and a North Carolina State Park. But these proposals for public conservation have yet to succeed and the private Nature Conservancy (TNC) has led preservation efforts in the Black River watershed. TNC now owns 6,400 ha in and adjacent to these old-growth floodplain forests in the Black River Preserve, but thousands of additional hectares with high quality ancient forests remain to be protected. These unprotected ancient forests and the excellent water quality of the Black River are both threatened by logging, water pollution, and development. This article uses dendrochronology and radiocarbon dating to document the extreme longevity of living bald cypress trees at Black River. The discovery of living bald cypress trees well over 2,000-years old provides quantitative evidence for the ecological integrity of this wetland system, extends the proxy paleoclimate record of precipitation variability for the Southeast by over 900 years, and will help advance public and private preservation of the Black River.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Dendrochronology and radiocarbon dating were used to document the age of the living bald cypress trees adjacent to Black River and to develop the proxy tree-ring record of growing season moisture variability for centuries prior to instrumental observations. Non-destructive core samples were all initially collected for paleoclimatic applications (Stahle et al 1988, 2012), but subsequent fieldwork focused on using dendrochronology to document the great age of the trees. Simple age determination of the Black River bald cypress is complicated by the frequency of heart rot, but many ancient individuals are not hollow and can be sampled to determine their minimum age with tree-ring and radiocarbon dating. If the true antiquity of bald cypress can be documented then the very presence of exceptionally old living trees could help promote public and private conservation efforts in the Black River watershed.

To determine the age of individual old trees and to develop a long and well-replicated tree-ring chronology for climate reconstruction, the core sampling of bald cypress at Black River was selective. To document the oldest trees, solid individuals exhibiting the external characteristics of great age were deliberately sampled. These candidate old trees typically include large twisted and low taper stems with flat-topped crowns, a few heavy limbs, canopy die back, burls, and hollow voids (figure 2). The trees were cored above the basal swell, typically at 3 m, to avoid distortion of the annual rings and to provide a date for the minimum age of the individual. The core specimens were mounted, polished, and all rings were dated to the calendar year of formation under the microscope at 10–70 magnifications using the skeleton plot and visual methods of crossdating (Douglass 1941, Black et al 2016). The dated rings were measured with a precision of 0.001 mm and the time series for each core were submitted to correlation analyses of overlapping segments for quality control of dating and measurement (Holmes 1983). The well-dated ring width time series were then detrended and standardized to remove non-climatic growth trend by fitting a spline function to the measurements from each core (the spline functions had a 50% frequency response equal to 67% of the series length). The raw measurements were first power transformed to reduce bias (Cook and Peters 1997) and the ring-width indices were calculated as the ratio of the measurement value by the value of the fitted spline each year. The mean ring width index chronology for the Black River site was then computed as the biweight robust mean of all individually detrended and standardized core series (Fritts 1976, Hoaglin et al 2000, Cook and Krusic 2005).

Correlation analysis was used to document the regional precipitation signal in the mean index chronology from Black River. The chronology was correlated with gridded monthly precipitation totals in the PRISM 4 km resolution monthly precipitation dataset for the eastern United States available from 1895-present (PRISM Group 2004). The highest correlations were noted during the first half of the growing season so the monthly PRISM values were summed to April-June (AMJ) seasonal totals and the Black River chronology was then correlated with these gridded seasonal values to estimate the strength and spatial scale of the moisture signal in these wetland trees. To measure the fidelity of the local to regional precipitation signal in the tree-ring chronology, the PRISM instrumental AMJ precipitation totals from the grid point closest to the Black River collection site were also correlated with the gridded PRISM AMJ precipitation field for the eastern United States. These correlations all extend from 1895–2010, ending in 2010 when the well replicated portion of the Black River chronology also ends. The tree-ring chronology and the nearest instrumental precipitation grid point data were then correlated with the 500 mb geopotential height field from 1948–2010 using NCEP NCAR Reanalysis (Kalnay et al 1998) to document the large-scale atmospheric circulation features important for precipitation and tree growth in southeastern North Carolina.

Radiocarbon dating was used to independently test the age of the two oldest individual trees yet located at Black River. Calibrated AMS radiocarbon dates were obtained for cellulose extracted from small core fragments near the center of the two oldest trees [based on the High Probability Density Range Method (Bronk Ramsey 2009) and INTCAL 13 (Reimer et al 2013) at Beta Analytic, Inc.]. Because the core fragments are only 5 and 12 mm in diameter, they include multiple annual rings in order to provide sufficient cellulose for radiocarbon analysis. The exact annual tree-ring dating range of each radiocarbon specimen was recorded for precise comparison with the AMS ¹⁴C results.

The oldest trees located thus far at Black River are found in a 300 ha section known locally as the Three Sisters Cove, but centuries to millennium old individuals are still present along nearly the entire length of the stream. Dendrochronological analysis of selected tree cores has identified several trees over 1,000-years old and two over 2,000-years old (figure 3(a)). The innermost dated ring on tree BLK232 is 70 BCE and for BLK227 it is 605 BCE, making them at least 2,088- and 2,624-years old in 2018, respectively. Both tree-ring dates represent the minimum age of each tree because they were cored above the basal swell at 3 m above the swamp floor. No specific information is available on the time required for saplings to reach a height of 3 m, but growth rates are very slow at Black River.

Zoom In Zoom Out Reset image size

The tree-ring dating for both of these very old trees has been independently confirmed with radiocarbon dating, within the dating precision of the AMS ¹⁴C estimates. The inner core fragment for BLK227 was tree-ring dated from 548–600 BCE and is within the 95% CI of the AMS ¹⁴C result for this tree (i.e., AMS ¹⁴C = 788–537 BCE; Laboratory Number Beta-497625). The inner core fragment for BLK232 was tree-ring dated from 34–43 BCE, also within the 95% CI for the AMS ¹⁴C result for this ancient tree (i.e., 118 BCE to 26 CE; Laboratory No. Beta-497626).

Heart rot is prevalent in the largest and oldest trees at Black River (e.g., figure 2(a)) but there is evidence suggesting that several other bald cypress trees are over 2,000-years old along this stream. We extracted core samples from four trees with innermost rings that tree-ring dated to at least 200, 235, 395, and 427 CE (figure 3(a)), but these cores did not approach the true center ring due to heart rot or other irregularities. Given the very slow growth rates of these particular trees and all other sampled cypress at Black River, which is only 4.3 cm of radial expansion per century, it is likely that these four trees are also over 2,000-years old [i.e., mean annual ring width = 0.43 mm (std dev = 0.31); n = 179 radii from 110 trees). Because we have cored and dated only 110 living bald cypress at this site, a small fraction of the tens of thousands of trees still present in these wetlands, there could be several additional individual bald cypress over 2,000-years old along the approximately 100 km reach of Black River.

The ancient bald cypress trees at Black River record a unique and otherwise unavailable record of annually resolved climate variability and change in the exactly dated annual growth rings. The long tree-ring chronology derived from the ancient bald cypress exhibits dramatic interannual, decadal, and multidecadal variability (figure 3(b)) and is positively correlated with growing season rainfall totals over the southeastern United States (figure 4(a)), in spite of the frequently flooded conditions that prevail in these forested wetlands. This strong positive precipitation response at Black River, which has been detected in all other available bald cypress chronologies (Stahle and Cleaveland 1992, Stahle et al 2012), reflects large inter-annual variability in water levels, intense evapotranspiration demand during the growing season, root system stratification near the wetland soil surface by frequent anoxic conditions at depth, and the flux of well-oxygenated water through the root zone (e.g., Davidson et al 2006, Stahle et al 2012). Dendroclimatic reconstructions of rainfall based on Black River and other bald cypress chronologies from the region have revealed decade-long droughts and pluvials during colonial and pre-colonial times that exceed any measured during the modern period (Stahle et al 1988), including the severe droughts that impacted the first English attempts to settle in America at the Roanoke Island and Jamestown colonies (Stahle et al 1998).

Zoom In Zoom Out Reset image size

The regional precipitation signal encoded in the Black River chronology (figure 4(a)) closely follows the region-wide pattern of correlation for the instrumental precipitation data from southeastern North Carolina (figure 4(b)). The correlations between the chronology and gridded precipitation totals are not as strong as the correlation between the gridded instrumental data at adjacent grid points, of course, but the tree-ring chronology can explain approximately half of the variance in nearby instrumental AMJ precipitation totals and has the virtue of spanning the past two millennia. The Black River tree-ring chronology is also correlated with the 500 mb geopotential height field over the Northern Hemisphere during the spring-early summer (figure 4(c)). The spatial pattern of correlation is similar for the co-located PRISM instrumental AMJ precipitation totals and the correlation maps (figures 4(c), (d)) indicate that tree growth and AMJ precipitation both tend to be reduced during the positive phase of the Pacific/North American (PNA) pattern of mid-tropospheric circulation as has been previously described for the eastern United States (Leathers et al 1991). The geopotential height correlations are higher in the Pacific and Atlantic sectors for the tree-ring chronology, but they are higher for the instrumental precipitation totals over the eastern United States (figures 4(c), (d)). These differences may arise from the fact that bald cypress growth is also negatively correlated with temperature (Stahle et al 1988), which is also well correlated with the PNA during the winter-spring.

Living trees over 2,000-years old are extremely rare worldwide. Only eight species have been proven with dendrochronology to live for more than 2,000 years (Brown 1996). Six of these species are located in the western United States, one in Chile, and now bald cypress at Black River, North Carolina. The oldest known Black River bald cypress (BLK227) has an inner ring date of 605 BCE, based on crossdating with other trees at Black River back to as early as 70 BCE, but with four radii from just this one old tree from 70 to 605 BCE. This means that BLK227 is at least 2,624-years old in 2018, moving bald cypress to number five on the worldwide list of the oldest-known continuously living, sexually reproducing, non-clonal tree species based on dendrochronology. Bald cypress is the oldest-known wetland tree species on earth. Only individual trees of Sierra juniper (Juniperus occidentalis) at 2,675-years, giant sequoia (Sequoiadendron giganteum) at 3,266, alerce (Fitzroya cuppressoides) at 3,622, and Great Basin bristlecone pine (Pinus longaeva) at 5,066-years old are known to live longer than Black River bald cypress. The actual difference in the age of the oldest-known bald cypress and Sierra juniper is likely very low because of the extremely slow growth rates at Black River and the time required for trees to reach the 3 m sampling height.

The maximum longevity of wetland tree species has not been well documented. The oldest-known are the millennia-old bald cypress at Black River, but Montezuma bald cypress (T. mucronatum) in the 1,200 to 1,500-year age class have been reported in Mexico (Villanueva-Diaz et al 2007) and Pilgerodendron uviferum at least 859-years old have been noted in southern Chile (Aravena 2007; P. uviferum is found in both upland and bog sites in Valdivian temperate rainforests). The oldest-reported angiosperm species from a wetland settling appears to be a Nyssa silvatica at 679-year old from a bog in New England (Di Filippo et al 2015), and a Nyssa aquatica at least 600-years in the forested wetlands of Bayou DeView, Arkansas (Stahle, unpublished data). Macrolobium accacifolium over 500-years old have been reported from the seasonally flooded forests of the central Amazon based on tree-ring analysis (Schöngart et al 2005).

Certain nutrient-poor wetlands may be consistent with a model of 'longevity under adversity' which Schulman (1954) applied to conifers in semi-arid America, including bristlecone pine (Pinus longeava), the oldest known living trees on earth (Schulman 1958). The adverse low pH, limited nutrients, and frequently flooded conditions at Black River not only exclude most bottomland hardwood tree species, they are also associated with extremely slow growth rates of bald cypress on this black water stream. The differences on tree growth rates between the seasonally flooded forests of the várzea and igapó in the central Amazon have also been attributed to nutrient availability in these white-water and black-water forested wetlands, respectively (Schöngart et al 2005).

The long tree-ring chronology developed from these ancient bald cypress trees has considerable paleoclimate value. The strong regional scale moisture signal in this chronology has contributed to the success of tree-ring reconstructions of the Palmer drought severity index (PDSI) on a gridded basis for North America (Cook et al 1999, 2007). These PDSI reconstructions in the so-called North American Drought Atlas (NADA) have leveraged the analysis of past droughts and pluvials (Fye et al 2003, Cook et al 2016), the large-scale climate dynamics responsible for moisture variability over the continent (Coats et al 2016, Baek et al 2017), and interactions between climate, ecosystem function (Swetnam et al 2009, Marlon et al 2017), and social impacts over the past two millennia (Stahle and Dean 2011, Burns et al 2014). The recently discovered old trees have extended the existing Black River chronology by 970-years (i.e., from 365 CE to 605 BCE), but the chronology is based on only one tree before 70 BCE. Tree-ring samples are available from only 110 trees out of the tens of thousands still present along this stream and it is likely that the chronology can be improved and perhaps extended with additional sampling of old trees. Old relic logs are sparingly present in shallow submerged context and likely in the thalweg of the main channel on the lower reaches of the Black River. These sources of relic or sub-fossil logs have not been investigated but may help improve and extend the Black River chronology deeper into the Holocene.

The ancient bald cypress forests of Black River survive in the heavily developed agricultural and industrial forest landscape of southeastern North Carolina. Few if any old growth upland hardwood or pine forests of any size are known to still exist in the vicinity of Black River. Exploitation of timber and naval stores began during the colonial era and the Black River was accessible by steamboat from the Cape Fear River during the 19th century (Hart 1994). Many bald cypress and bottomland hardwoods were cut along the Black River, of course, but the old-growth trees that dominate many stands were simply not suited for saw log production (Stahle et al 2012). As a result, extensive tracts of uncut ancient forest remain intact along the Black River.

The survival of ancient bald cypress and other tree species at Black River provides qualitative and quantitative evidence for the ecological integrity of this wetland ecosystem. Qualitatively, an excursion on the Black River where clear tea-colored water flows over white sand and among columnar trees of great age is a unique experience and certainly qualifies this stream as one of the great natural areas of eastern North America (figures 1 and 2). Quantitatively, the minimum age determinations of selected old-growth bald cypress trees are based on dendrochronology (figure 3(a)), the most accurate and precise dating method in geochronology. The habitat value of the Black River is further documented by the high surface water quality (a designated Outstanding Resource Water, NCDPR 2018), the presence of rare and endangered species (NCDPR 2018), and the recent observations of threatened wood storks (Mycteria americana) and swallowtail kites (Elanoides forficatus) in these forested wetlands.

The Nature Conservancy has led the preservation of the Black River, but the state of North Carolina and a few other private landowners have made important conservation contributions as well. A feasibility study to consider the establishment of a state park on the Black River was enacted by the North Carolina General Assembly in 2017 (HB353 in NCDPR 2018) and the Nature Conservancy continues to seek funding to expand their Black River Preserve. Unfortunately, the ancient forests and water quality of the Black River are both imperiled by continued logging, biomass harvesting for wood pellets and garden mulch, and water pollution from confined animal feeding operations and municipal discharge in the watershed. In the longer term, the Three Sisters Cove is only 2 m above mean sea level and is at risk from sea level rise which could occur during the life span of these exceptionally old trees with unmitigated anthropogenic global warming. To counter these threats, the discovery of the oldest known living trees in eastern North America, which are in fact some of the oldest living trees on earth, provides powerful incentive for private, state, and federal conservation of this remarkable waterway.

Funded in part by the National Science Foundation (AGS-1266014), The Nature Conservancy, and private donations to the Ancient Bald Cypress Consortium. We thank John and Maggie DeCuevas, Kemp Burdette, Jordan Burns, Graham Hawks, Hervey McIver, Julie Moore, Zack West, and Katherine P Wolff. We appreciate the helpful editorial suggestions of Neil Pederson and an anonymous reviewer. The tree-ring data described in this article are available from the corresponding author (dstahle@uark.edu) and from the International Tree-Ring Data Bank, NOAA Paleoclimatology Program (https://www.ncdc.noaa.gov/data-access/paleoclimatology-data/datasets/tree-ring).