Dendro-archeo-ecology in North America and Europe: Re-purposing Historical Materials to Study Ancient Human-Environment Interactions
- 1.2k Downloads
The unique position of dendrochronology at the nexus of archeology, ecology, and climatology allows it to play a pivotal role in the study of past human-environment interactions. Yet, few tree-ring studies in Europe and eastern North America have been used to study pre-industrial land-use changes, forest ecology, and carbon dynamics and thus to constrain the uncertainties surrounding the Early Anthropocene hypothesis (Ruddiman Clim Chang 61:261–293, 2003; Rev Geophys 45(4):RG4001, 2007). Here, we discuss the potential of dendro-archeo-ecology—the use of dendroarcheological material in the study of forest ecology—to document past human land-use and forest alteration, which started in the Neolithic Era (∼12,000–4000 BP) in Europe and after European immigration into eastern North America in the 1620s. In this context, we focus on the dendro-archeo-ecology of (1) Neolithic pile dwellings in the Euro-Mediterranean region and (2) old-growth forest dynamics in eastern North America. We discuss recurring challenges (e.g., low sample depth, short series length) and uncertainties (e.g., species and tree size bias) related to the use of (pre)historic timbers for ecological purposes that need to be carefully addressed. We advocate for a concerted effort to move the use of dendro-archeological material from strictly archeological applications towards exploration of its ecological potential and for a close alliance of dendrochronology with related disciplines that aim to address the same subjects.
KeywordsDendroarcheology Dendroecology Dendro-archeo-ecology Early Anthropocene Europe Forest clearance Land-use change Pile dwellings Neolithic North America
Dendrochronology can play a pivotal role in the study of past human-environment interactions, supported by the annual resolution of tree-ring chronologies, their capacity for absolute dating, and their potential to span time periods relevant to human history (e.g., le Roy Ladurie 1971; Lamb 1995). With many applications in the fields of archeology, ecology, forestry, and paleoclimatology, dendrochronology is uniquely positioned at the nexus of these related fields to study complex feedbacks between humans and their environment (see also Chapter 14). Some of the earliest dendrochronological work (Douglass 1929) was focused on drought as a cause for the sudden abandonment of ancient Puebloan settlements in the American Southwest towards the end of the thirteenth century (Cook et al. 2004). Dendrochronological evidence has also contributed to disentangling the role of drought from other causes in the collapse of the Mayan civilization near the end of the first millennium CE (Stahle et al. 2011) and in the fate of the Roanoke and Jamestown colonies on the mid Atlantic coast of eastern North America in the sixteenth and seventeenth centuries (Stahle et al. 1998). Tree-ring data have been used to show that climate was a contributing factor to the Mongol withdrawal from Hungary in the thirteenth century (Büntgen and Di Cosmo 2016) and to the demise of Greater Angkor, the capital of the Khmer Empire in Cambodia in the fourteenth and fifteenth centuries (Buckley et al. 2010). Tree-ring based climate reconstructions in central Europe (Büntgen et al. 2011) showed exceptional decadal-scale hydroclimate fluctuations over the period 250–550 CE that may be linked to the fall of the western Roman Empire.
Yet, for many (if not all) of these past societal disruptions, climate was not the sole determining cause, but rather an important contributing factor in a complex network of human-environment perturbations in which anthropogenic land-use change often played a significant role (Diaz and Trouet 2014). For instance, extensive deforestation is postulated to have reduced the ecological carrying capacity of the Mayan lowlands (Medina-Elizalde and Rohling 2012) and to have contributed to the eleventh century Meso-American droughts (Cook et al. 2012). Widespread deforestation and over-harvesting of wood are also known to have caused significant environmental stress to the Puebloan (Gumerman 1988; Guiterman et al. 2016) and Roman societies (O’Sullivan et al. 2008).
These interplays between historical land-use changes and climatic effects corroborate Ruddiman’s (2003, 2007) hypothesis that the influence of human activity on the atmosphere’s greenhouse gas composition and thus on climate started long before the onset of the industrial era circa 1850 CE. This ‘Early Anthropocene’ hypothesis postulates that early agriculture and the start of deforestation in Eurasia around 8000 years ago initiated an anomalous increase in the long-term natural atmospheric CO2 trend. In the following millennia, worldwide agricultural intensification and continued widespread forest clearance were responsible for an increase in the amplitude of this positive CO2 trend and associated global temperature anomaly. Ruddiman (2003, 2007) further hypothesized that short-lived CO2 decreases superimposed on the millennial-scale trend were caused by pre-industrial pandemics that caused far-reaching mortality followed by reforestation of abandoned agricultural land (Ruddiman and Carmichael 2007).
Many aspects of the Early Anthropocene hypothesis have been challenged, particularly Ruddiman’s (2003) claim that the pre-industrial CO2 anomaly could be explained solely by anthropogenic land-use changes (e.g., Joos et al. 2004; Ruddiman 2007). However, the paucity of pre-1850 land-use data and inherent uncertainties in estimates of pre-industrial anthropogenic deforestation and carbon emissions (DeFries et al. 1999) hinder a robust quantification of the relative contributions of human activities on atmospheric CO2 (and global temperature) trends over the late Holocene.
To the extent of our knowledge, few tree-ring studies have been used to constrain the uncertainties surrounding the Early Anthropocene hypothesis, even when tree-ring studies have a great potential to contribute to such research. The combination of annual resolution records of climate change with dendroecological data that document changes in pre-industrial forest productivity offers prospects for moving beyond simple temporal correlations towards more detailed analyses of cause and effect.
Currently this potential is limited for deeper time Anthropocene scenarios by a number of factors. Multi-millennial-length tree-ring chronologies in North America and Europe that coincide with the onset and intensification of anthropogenic land-use change and deforestation are rare (e.g, Pilcher et al. 1984; Ferguson and Graybill 1983; Brown et al. 1992; Stahle et al. 1985; Friedrich et al. 2004). Moreover, to develop chronologies that span multiple millennia, tree-ring series from long-lived living trees must be merged with series from remnant wood on the landscape (e.g., Brown et al. 1992), with wood from historical buildings and archeological sites (e.g., Kuniholm et al. 1996), and/or with subfossil wood (e.g., Pilcher et al. 1977). This is especially true in regions such as Europe and North America where climate promotes enough productivity to support high human populations, but rapidly increases the decay of deadwood that has the potential to extend living-tree chronologies. The majority of lowland forests in Europe and North America have been cleared at various stages throughout the course of human history and tree-ring chronologies from living trees are derived from either secondary (post-clearance) forests that reflect only recent forest ecology and management conditions (e.g., Pederson et al. 2014, Druckenbrod et al. 2013), or from remnant stands of old-growth forests that are often located in remote or unproductive sites (e.g., near upper tree line and on steep slopes or shallow soils; Panayotov et al. 2010; Stahle and Chaney 1994; Davis 1996; Therrell and Stahle 1998; Maxwell et al. 2011). Such remaining old-growth stands might not necessarily reflect the ecology and climate sensitivity of the formerly surrounding forests (Babst et al. 2013, 2014a). Whether old-growth forests represent the full, long-term carbon sequestration potential of the landscapes that have been significantly altered by human land-use is difficult to know.
As a result, little is known about the forest ecology and carbon dynamics of the forests that existed prior to the intensification of human land-use and forest clearance, which began with early agriculture in the Neolithic Era (∼8000 BP) in Eurasia and after European immigration to eastern North America in the early 1600s. Here, we explore the potential to derive ecological information for these eras from the ample wood evidence that is preserved in the form of archeological and historical building material. Dendroarcheological material has been consolidated to extend tree-ring based climate reconstructions further back in time (e.g., Cook et al. 2004; Büntgen et al. 2006, 2011), but the wealth of information about primary forests that can be provided by wood from early human dwellings and wooden structures (e.g., boats, fence posts) remains largely untapped (but see e.g., Krause 1997). Dendrochronology offers the possibility to explore human alteration of forest ecosystems and to observe these changes over time through diverse wooden remains and contemporary cultural records (e.g., Pausas et al. 2009; Kirby and Watkins 1998; Rackham 1990; Meiggs 1982; Thirgood 1981). We aim to characterize the potential of and challenges related to the use of dendroarcheological material in the study of primary forest ecology—the field of dendro-archeo-ecology—in the context of two different regions and time-scales.
The Euro-Mediterranean region (Section “Dendro-archeo-ecology of Euro-Mediterranean Pile Dwellings from the Neolithic Era”) is arguably the richest on Earth in terms of the preservation of culturally and temporally diverse wooden remains. Humans have occupied the area for well over 500,000 years and throughout this time there is abundant evidence for the use of wood as a resource. 300,000 years ago, hunters in Schöningen (Germany) used wooden spears and tools to kill and butcher horses and to fend off saber-toothed cats (Serangeli et al. 2015; Thieme 1997). Examination of these wooden implements by Schoch et al. (2015) demonstrated that even these early hunter-gatherer societies preferentially selected certain wood types for specific tasks, and worked the material with procedure and skill. 32,000 years ago wood also played a role in survival through the Ice Age, as prehistoric artists who decorated the walls of Chauvet Cave (Ardèche, France) were warmed and lit by wood burning hearths and torches (Cuzange et al. 2007). As the ice retreated ca. 12,000 years ago, a warmer, wetter climate ensued (e.g., Wanner et al. 2008; Mayewski et al. 2004) and afforestation began as tundra vegetation and bare ground transitioned into forest (Huntley 1990). People began to settle in seasonal or permanent locations and to create their own shelter, utilizing wood for a wide range of construction, hunting, fishing, and farming practices (Whittle 1996; Dimbleby 1984). Anthropogenic landscape perturbations and forest clearance thus started in the Early Neolithic (∼8000 BP). For example, the craftsmanship needed for the refined carpentry found by Tegel et al. (2012) in Neolithic (∼7000 BP) water wells in eastern Germany suggests that carpentry—and by extension, forest clearance—in northern Europe developed at the same time as agriculture. Anthropogenic land-use change then intensified in various phases during Roman Times and in the Middle Ages, periods for which ample dendroarcheological samples are available (Tegel et al. 2010). In the section “Dendro-archeo-ecology of Euro-Mediterranean Pile Dwellings from the Neolithic Era”, we focus on the dendro-archeo-ecology of pile dwellings, one of the richest dendroarchaeological resources for prehistoric human-environment interactions in Europe. We consider the dendroecological potential of Neolithic pile-dwelling-derived wood to inform our understanding of prehistoric forest resource management and human response to environmental change.
In eastern North America (Section “Sixteenth to Nineteenth Century Dendro-archeo-ecology in Eastern North America”), coarse-scale deforestation and the construction of wooden buildings started much later. The most intensive changes to the landscape and forests began in the 1620s when Europeans immigrated to the region and continued into the mid-nineteenth century, after which eastward agricultural expansion and industrialization led to a century of natural reforestation (Cronon 1983; Thompson et al. 2013). As a result, old-growth forest stands are rare and patchy throughout the region, which limits our ability to study historic growth patterns and forest processes prior to European immigration. In the section “Sixteenth to Nineteenth Century Dendro-archeo-ecology in Eastern North America”, we posit that historical buildings offer perhaps the only opportunity to enhance our dendrochronological network for the study of the environmental (biotic and abiotic) factors driving tree growth and forest dynamics prior to 1600 CE with substantial replication. In this context, we present two case studies that use dendro-archeological material i) to identify forest disturbances in the Ohio and Hudson River Valleys and ii) to investigate a late 1600s subcontinental forest recruitment pulse (Pederson et al. 2014). The two studies demonstrate the enormous promise of eastern North American timbers in recovering local and regional forest dynamics prior to the nineteenth century. Most promising, these timbers yield data on historical forest dynamics in areas virtually devoid of old-growth forests today. Dendro-archeo-ecology in eastern North America can thus quantify historical forest productivity and reduce the uncertainties surrounding pre-industrial carbon sequestration (DeFries et al. 1999).
16.2 Dendro-archeo-ecology of Euro-Mediterranean Pile Dwellings from the Neolithic Era
Ongoing controversy over Ruddiman’s Early Anthropocene hypothesis is largely centered around the question of whether small, low density populations of pre-modern humans using low-intensity technology could carry out forest clearance, agriculture, and industrial production at levels that could be sufficient to alter global atmospheric chemistry. However, even if early societies did not alter atmospheric chemistry, a rapidly growing range of studies of past human and environmental interaction are beginning to demonstrate a surprisingly early and surprisingly large anthropogenic footprint in a number of locations around the world (Ellis et al. 2013a). Defining and quantifying early human-environmental impact on a range of geographic scales with a precise temporal resolution therefore has a stand-alone significance in helping to understand the numerous complex feedbacks between human and natural coupled systems over the course of our history. In this section we advocate the use of dendro-archaeo-ecology more broadly in terms of establishing a date or possible range of dates for the start of the Anthropocene and in defining the true anthropogenic footprint in the Euro-Mediterranean region around 8000 years ago. We focus in particular on one subset of wooden archaeological remains, which has great potential for in depth ‘re-study’ within this context, namely the prolific and exceptionally well-preserved Euro-Mediterranean pile dwellings.
16.2.1 Pile Dwellings as a Cultural Phenomenon: Geographic and Temporal Spread
The archetypal pile dwellings proliferated across central Europe and around the Alps between 7000–2500 BP (Sherratt 2004; UNESCO 2015, http://sites.palafittes.org/home) and it is in these regions where many of the leading theoretical approaches to the study of such sites have been developed (Pranckėnaitė 2014). The tradition of prehistoric lake settlements, however, expands far across the Euro-Mediterranean region and beyond (e.g. Lithuania—Pranckėnaitė 2014; Slovenia—Čufar et al. 2010, 2015; Northwest Russia—Mazurkevich et al. 2010; Kulkova et al. 2001; Bulgaria—Ivanov 2000; Northern Greece—Karkanas et al. 2011; Chrysostomou 2015; Macedonia—Sherratt 2004). Palaeobotanical evidence suggests that the Neolithic pile-dwellings phenomenon followed a south-north diffusion, starting in the northeast of Spain, central Italy, and Slovenia and spreading towards central Europe and beyond (Menotti 2004). They reached the widest geographical coverage and typological diversity during the European Iron Age (Sherratt 2004) and they represent some of the most commonly preserved archaeological evidence, which (in combination with other contemporary excavations and paleo-records) could potentially be used to reconstruct population density in combination with ecological impact.
The importance of pile dwellings has been internationally recognized and of the 937 archaeologically documented sites around the Alps alone, 111 have been included in the UNESCO World Heritage List (UNESCO 2015). Despite the increasing diversity and geographical spread of these settlements through time, a common feature remains: they all used wood as the main construction material. We will now consider more specifically how the analysis of tree rings from construction timbers, wooden objects (tools, artefacts), and charcoal from pile-dwelling sites has been and can be used (1) to shed light on how our ancestors used, managed, and impacted forest resources following the transition to early settlements and (2) to provide insights into open questions about human responses to environmental changes in prehistory. We will also discuss the potential limitations of this approach.
16.2.2 Tree-Ring Studies of Pile-Dwelling Sites
Dendrochronology has been systematically implemented in pile-dwelling research since the 1970s (Billamboz 2004). The availability of large amounts of well-preserved timbers at these sites has often resulted in the collection of thousands of samples for tree-ring studies, including timbers from young specimens with few tree rings (Orcel 1980). This has been the case for several sites in Switzerland where 62–83% of piles were derived from trees less than 40 years old (Tercier et al. 1996) and pile dwellings in southwest Germany where the vast majority of investigated timbers had less than 50 rings (Billamboz 2008). Similarly, most of the wood recovered at other pile-dwelling sites in different parts of Europe contained less than 45 rings. For example, at the site of Žemaitiškė 2, in Lithuania, most of the piles were made of trees less than 40 years old (Girininkas 2010), whereas at La Draga—the only pile-dwelling settlement found to date in Spain—numerous posts are from yet younger trees of 10–35 years (Tarrús i Galter 2008). Absolute dates were not obtained for the latter two sites, but dendrochronological research on the timbers from La Draga has provided information about building phases and about the seasonality of construction works at Žemaitiškė 2, which must have taken place during the summer, as inferred from clusters of relative felling dates in the spring and early summer months (Girininkas 2010). For pile-dwellings, where bark is often preserved on the timbers, it is usually presumed that there was little time between felling and construction. The wet working environment rendering a period of storage and drying (during which the cambium would rapidly degrade and the bark be lost) somewhat pointless, as well as perhaps impractical.
In contrast, the inclusion of timbers from young trees (<50 years) in his studies on pile dwellings of Lake Constance allowed Billamboz (2014) to apply dendrotypology—a dendroarcheological approach based on the classification of timber according to tree age, growth patterns, and degree of stem conversion—to infer absolute dates for building activities and identify patterns of woodland management, development, and degradation over the period c. 3900–2400 BCE. His results suggest that clearance of woodlands coincided with phases of human settlement, during which slow growth patterns in timbers reflected long-lived (>100 years) trees from dense forests. These settlement phases were followed by coppicing practices in periods of demographic expansion, which were represented by shorter (<100 years) tree-ring series from faster-growing trees, until the surrounding forest could no longer supply enough timber and large old trees had to be cut and processed into the desired dimensions (indicated by the high level of conversion of timbers from later building phases). Billamboz (2014) thus not only answered primary archaeological questions of dating and seasonality of construction phases, but his approach in defining types of growth patterns including for short-lived trees served also to reconstruct forest dynamics and ancient woodland management practices such as coppicing. Such studies enhance the archaeological record and maximize the potential of applied interdisciplinary tree-ring research.
Yet many studies on wood from pile dwellings focus on more linear outcomes—such as dating—for a variety of reasons, including restrictions on sampling, difficulties of sampling/preservation, and funding for specific purposes. For example, Čufar et al. (2010) describe the materials from seven pile dwellings at Ljubljansko barje in Slovenia—comprising 187 samples from 1039 oak timbers—but point out the limitation of their results, because their research was limited to oak samples with more than 45 rings. This is a potential shortcoming for a number of studies, based perhaps also on the internationally recognized thresholds of a minimum of 100 tree rings for crossdating (Baillie 1982).
Billamboz’s research, as well as other studies from more recent historical periods (e.g., Sass-Klaassen et al. 2008; Domínguez-Delmás et al. 2011), however, have demonstrated that short series with less than 50 tree rings can be confidently crossdated when the bark edge is present in a large set of coeval samples, as Huber (1967) suggested five decades ago. Dendrochronological research of trunks from Žemaitiškė 2 and La Draga has shown that the use of relatively dated series (‘floating chronologies’), not yet anchored in time, can still provide information about occupation periods and building and repair phases (Tarrús i Galter 2008). Furthermore, species suitable for tree-ring studies such as ash (Fraxinus sp.), which are sometimes disregarded (e.g. Čufar et al. 2010; Mazurkevich et al. 2010), have the potential to be crossdated and can provide valuable ecological information (Sass-Klaassen et al. 2004; Sass-Klaassen and Hanraets 2006).
16.2.3 Dendro-archeo-ecology of Pile Dwellings to Inform the Impact of the 8.2K Event
According to Sherratt (2004), the abundance of pile dwellings found and studied during the past 150 years has produced such a bulk of scholarly articles and monographs that it has led to “complacency”: a lack of enthusiasm for these types of settlements. Paradoxically, there are still important open questions to which dendrochronological data from the wooden remains of pile dwellings could hold the answers. For instance, the impact of the 8200 BP event (the 8.2K event) in the development and spread of early farmer communities in the Euro-Mediterranean region is still a subject of much debate (Berger and Guilaine 2009 and references therein). Approximately 8200 years ago, the North Atlantic region suffered a drastic drop in temperatures, possibly due to a sudden flush of >1014 cubic meters of freshwater into the Labrador Sea following the collapse of an ice dam (Barber et al. 1999). This is hypothesized to have induced cold and dry climatic conditions in the Near and Middle East, and cool and wet conditions in Central and Western Europe (Bauer et al. 2004) that disrupted the progression of crop and animal domestication in the eastern Mediterranean and prompted the migration of early farmers into current Greece and Bulgaria (Weninger et al. 2006). In turn, this could have promoted the transfer of knowledge further west into Europe. The archaeological records throughout the northern Mediterranean show a disruption around that time (see Fig. 1 in Berger and Guilaine 2009), but the paleoclimatic proxies (mostly lake and marine sediments) lack the high-resolution temporal and spatial scale needed to effectively assess the impact of this event. Tree rings could offer a potential solution to this, while contributing indirectly to a broader context for the Early Anthropocene hypothesis within coupled natural-human systems: climatic forcing of human systems might in turn lead to human modifications of a new environment, which might then develop to a substantial anthropological footprint leading to new forcing of the natural environment. Tree-ring chronologies (from pile-dwellings and/or paleoenvironmental sites) can thus be used for high-resolution, spatially resolved proxies for detecting the described sudden on-set climatic change. In addition to this, a more developed hypothetical approach might be to use dendrochronology (and/or dendrochronologically facilitated radiocarbon wiggle-match dating) to date occupation phases of sites across the wide geographic region and to explore the spread of this mode of habitation and the spread of technologies and ecological modifications in relation to climate. For example—if the south to north diffusion previously observed is contemporary with changes in climate characterized using a range of proxies, the timing and geographical spread of pile dwellings could be linked to migration driven by cooler conditions proliferating to the North (as in later migration periods). Tree-ring patterns from more or less contemporary pile dwellings could then be used to create maps of ecological impact providing the much needed spatial and temporal resolution to better examine the 8.2K event (e.g., Crutzen and Stoermer 2000; Crutzen 2002; Smith and Zeder 2013) and its place in the wider context of debates on the timing of the transition of the Holocene to the Anthropocene.
16.2.4 Euro-Mediterranean Dendro-archeo-ecology in the Context of the Early Anthropocene Hypothesis
Perhaps the foremost evidence in support of dating the Anthropocence prior to the industrial revolution comes from work in China related to the expansion of rice agriculture on methane production c. 5000 years ago (e.g., Li et al. 2009). This study provides strong support for human-induced increases in methane prior to the unprecedented rise of modern times. Moreover, in a review of the archaeological evidence for human modification of the environment of the Yellow River region in China, 8000–2000 BP, Zhuang and Kidder (2014) conclude that humans have been active agents in transforming the physical properties of land, water, and atmosphere in that particular cradle of civilization for much of the Holocene and that the onset of the Anthropocence should not be confined to changes in atmospheric chemistry alone. The same argument can be made for the Euro-Mediterranean region. We propose an approach, which revisits existing dendrochronological data from archaeological sites with dendroecological analysis and then combines this with multiple lines of paleoenvironmental and archaeological data at coarse and fine geographic scales. At the coarse scale these data could be evaluated to assess the progressive timing of the anthropogenic transformation of the environment. The broad-scale precisely dated and temporally resolved ecological change implied populations and associated landscape transformations could then be used to inform model-based approaches for spatially explicit reconstruction of long-term histories of ecosystem transformation, human populations, and land use at a global scale (e.g., Ellis et al. 2013b). Finer scale, site specific data could be used to provide much needed empirical evaluations for existing global scale models. In this way the density of archaeological sites, the abundance of well-preserved wooden remains, and the capacity for tree-ring sequences to combine precision data for ecological/climatic change and human response could perhaps best be used to elucidate the complex interplay of feedbacks superimposed on Early Anthropocence debate.
There seems much potential here, given the large number of archaeological data already collected across this critical time period; however, there are some hurdles to overcome. Many potentially relevant archaeological datasets remain unavailable and/or unpublished for a variety of reasons. Some may have had insufficient sample depth for publication in terms of the primary purpose for which they were collected. Similarly, some may consist of very short sequences that were previously deemed unsuitable for dendrochronological crossdating, or may have been dated by radiocarbon analysis rather than dendrochronology. Some may have a commercial value in terms of dating, others may simply be stored in non-standard legacy digital formats (see Jansma et al. 2010, 2012; Brewer et al. 2011). Work will be needed to collect and consolidate existing archeological data into accessible forms so that new avenues may be explored and dendro-archaeo-ecology will need to be explored as part of ‘big data’ interdisciplinary collaborative efforts.
16.3 Sixteenth to Nineteenth Century Dendro-archeo-ecology in Eastern North America
A catastrophic depopulation of First Nation people of eastern North America began in the sixteenth century (Harriot 1972; Blakely and Detweiler-Blakely 1989; Thornton 1987, 2000) resulting in 100+ years of less intensive forest management and, in some places, afforestation (Cronon 1983). European communities initially sprang up in larger river valleys and then expanded across the region during the early seventeenth century. The growth and expansion of these communities (e.g., for agriculture) led to coarse-scale deforestation across the region. The landscape was significantly transformed with the influx of new people, culture, and land-use philosophies.
Euro-American land use affected the Midwestern United States (U.S.) much later but also resulted in coarse-scale forest clearing. Construction of roads and canals into the Midwestern U.S. during the early 1800s allowed for the utilization of fertile soils for agriculture. The region was soon cleared of forests and old-growth forest was noted to be rare. For example, A.W. Butler (1896) commented on the loss of “tall trees” and “heavy timber” especially in southern Indiana. Florence Hawley (1941) also noted a lack of old forests in the Midwest and Indiana during the 1930s (Senter 1938a, b). As of the late twentieth century, forest cover was 13%, 21%, and 27% respectively for the three Midwestern U.S. states Illinois, Indiana, and Ohio (USDA 2015a, b, c).
Intensive land-use and the paucity of dendroecological collections prior to the sixteenth century limits our understanding of historic eastern North American forest dynamics (McCarthy et al. 2001). Because surviving old-growth forest typically occurs on low-productivity sites (Stahle and Chaney 1994; Therrell and Stahle 1998), there is a bias in our understanding of pre-settlement forest dynamics. For example, the old-growth forest in Illinois, Indiana, and Ohio is extremely small, equaling 0.01% (∼2100 ha), 0.005% (∼500 ha), and 0.005% (∼400 ha) of the total area of each state, respectively (Davis 1993). This is unfortunate as we have little information about forest dynamics on productive sites. Fortunately, the Euro-American doctrine of manifest destiny left important artifacts for dendrochronology preserved across the region: historic timbers from the thousands of structures, water vessels, and wooden objects. These historical timbers give us the opportunity to build a more robust geographical analysis of forest dynamics in the centuries leading up to the Euro-American immigration.
Dendroarchaeological analysis of historic timbers in eastern North America began during the latter portion of the twentieth century (e.g., Stahle 1979; Cook and Callahan 1992). Since then, dendroarchaeological analysis has become a vigorous area of research (e.g., Krusic et al. 2004; Wight and Grissino-Mayer 2004; Grissino-Mayer 2009’ Harley et al. 2011; Baas and Rubino 2012; Martin-Benito et al. 2014; Rubino 2014; Rubino and Baas 2014, New York State and NE North American Dendrochronology Project, https://dendro.cornell.edu/projects/usa.php, the Historic Timbers Project, http://centralapptimbers.weebly.com/blog/introducing-the-historic-timbers-project). The volume and geographic coverage of recent research increases our ability to reconstruct forest dynamics prior to the nineteenth century.
Given the abundance of timber available to early Europeans, construction was routinely performed using timber found on-site: trees were felled and incorporated into buildings as beams, rafters, floorboards, joists, and braces (Senter 1938a; Hutslar 1992; Roberts 1996). This is especially true when large timbers for structures like barns were needed and transport and sawing was impractical or cost prohibitive (Roberts 1996). Therefore, it is possible that individual structures represent a portion of a single forested stand, with specific species often used for particular purposes (Baas and Rubino 2013). As these structures were typically built for agricultural purposes and thus in agriculturally productive areas, it is likely that timbers from these structures represent a more productive portion of the landscape than extant old-growth forests.
Successful crossdating of the timbers found in historic structures has led to centuries-long chronologies from numerous taxa [white oak (Quercus subgenus Lepidobalanus), red oak (Q. subgenus Erythrobalanus), hickory (Carya spp.), ash (Fraxinus spp.), tulip poplar (Liriodendron tulipifera), and American beech (Fagus grandifolia)]. For example, there are eastern U.S. oak chronologies that date to the fourteenth and fifteenth centuries (E.R. Cook, pers. comm.). The taxa of these chronologies are ecologically important species and have promising potential for understanding the structure and dynamics (e.g., disturbance and recruitment patterns) of sixteenth to nineteenth century forests.
Below we present two examples that combine dendroarchaeological and dendroecological analyses of data from historic timbers to yield new information on forest dynamics in eastern North America prior to the nineteenth century.
16.3.1 Forest Dynamics in the Ohio and Hudson River Valleys
Disturbance is integral to the community structure, composition, and development of eastern North American forests (Oliver 1981; White and Pickett 1985). Quantification of abrupt and extended increases in ring width associated with acute disturbances has long been used to elucidate forest history and stand dynamics in eastern North America. These increases in growth, hereafter referred to as growth releases, are the response of trees to increased light and other resources as the result of disturbance-induced tree mortality in dense forests (Lorimer 1985; Nowacki and Abrams 1997) where increased resource availability, primarily light, significantly increases annual growth (e.g., Nowacki and Abrams 1997). Reconstructing centuries of canopy disturbance and tree recruitment gives insight into the timing, frequency, and spatial extent of long-term forest dynamics.
Timbers from four structures (Quercus subgenus Lepidobalanus; n = 114) in central and southern Indiana and eight structures (n = 46) from the mid-Hudson River Valley (New Paltz, New York; Krusic et al. 2004; Pederson et al. 2013) were analyzed to reconstruct historic forest disturbance. The Indiana timbers cover the 1604–1884 CE period while those from the mid-Hudson Valley cover 1449–1805 CE. Growth release events were objectively identified by comparing the growth rates of adjacent 15-year growth segments; a release was defined as an increase in growth of at least 50% (Lorimer 1985), where the median growth rate was calculated for each 15-year segment (Rubino and McCarthy 2004). The number of releases (relativized to a per 100-year basis to facilitate comparison among trees of different ages) and the return interval (years between individual release events) were calculated. There was no clear link between stem age and release—releases and suppressions occurred regardless of an individual stem’s age.
Growth release metrics of timbers from four structures in central and southern Indiana and eight structures from the mid-Hudson River Valley (New York); SE = standard error
Hudson River Valley
Releases/century: mean (SE)
Return interval: mean (SE)
Return interval: range
% of structures with ≥1 release event
% of years with release per timber: mean (SE)
1.2 (0.7)–13.5 (1.5)
To better understand the spatial extent of disturbance (stand-wide vs. single tree), the percentage of trees exhibiting a release in an individual year was determined. We focused analysis only on years in which 10 or more timbers were available (Rubino and McCarthy 2004). A stand-wide disturbance was identified if more than 25% of the timbers exhibited a simultaneous release (Nowacki and Abrams 1997; Rubino and McCarthy 2004). Synchronous and asynchronous releases were found in both New York and Indiana structures (Fig. 16.2). The percentage of years in which at least one release was detected in each Indiana structure ranged from 5 to 88%. Periods of synchronous release (> 25% of the trees) were found in only one of the four Indiana structures (the Hanover barn in Fig. 16.2a). In the Hudson Valley samples, a release was detected in 29% of the years analyzed, and one synchronous release period was identified in the late 1500s (Fig. 16.2b).
The mix of synchronous and asynchronous releases suggests classic gap dynamics in the forests from which these structures were built. These dynamics are characterized by patchy disturbances that kill one to a few trees in most years (e.g., Runkle 1990). Periodic, coarser-scale events (ice storms, tornadoes, large wind storms, and drought) likely resulted in synchronous releases at the stand-scale—such as reconstructed from the historic timbers of the Hanover barn—and at the regional scale—such as across the eight structures from the Hudson River Valley (Fig. 16.2). Consequently, we infer that these forests were most likely uneven aged and heterogeneously constructed (e.g., mix of shade-tolerant and—intolerant species). Furthermore, repeated disturbance is also most likely necessary for stems to attain canopy dominance (Runkle 1990; Rubino and McCarthy 2004).
16.3.2 Potential of Timbers for Regional-scale Recruitment Studies
Historic timbers not only have the potential to reveal canopy disturbance prior to the nineteenth century in regions dominated by human land-use, there is potential to acquire a history of regional-scale recruitment. A new hypothesis derived from a range of tree sample types (plots, stumps, historical timbers) indicates a pulse of recruitment in the late 1600s in broadleaf-dominated old-growth forests of eastern North America (Pederson et al. 2014). Of the many questions regarding this finding, an important one is that because many living tree samples were collected after 1980 CE, tree longevity—rather than disturbance—might be driving these observed patterns in recruitment. Given that historical timbers have the potential to reach further back in time, both the age structure of a large collection of timbers as well as the pattern of initial rings can inform the timing of regional-scale recruitment and the conditions under which these trees originated sensu (Lorimer 1985; Lorimer and Frelich 1989). Here, we use a collection of timbers collected between 1938 and 1941 across three states of the lower Midwestern U.S. to conduct an additional test of the late 1600s recruitment pulse hypothesis.
The collection we examined is part of the Hawley-Bell Collection, a collection archived at the University of Arizona (Creasman 2011). Under the direction of Florence Hawley, researchers from the University of Arizona and University of Chicago explored the lower Midwestern U.S. region (Arkansas, Illinois, Kentucky, Missouri, Tennessee) from 1934 to 1941 for old samples from stave mills, structures of the indigenous Mound Builders and European immigrants, and old forests (Bell 1940, 1941). Field notes indicate that one goal was to procure Juniperus, Pinus, and Quercus samples with at least two centuries of “sensitive” rings, where sensitive means strong interannual variability in ring width and the potential of great sensitivity to climate. Investigators often had access to stumps or cross-sections, which eases estimating sample age in the field. Field notes include statements like, “I selected a large sample of the best pieces” (Bell 1941), and “En route from Corning to Little Rock we passed some small mills but the timber was neither old nor sensitive” (Bell 1940).
Inner ring dates from the Hawley-Bell collection are remarkably similar to the Kentucky data and somewhat similar to the eastern broadleaf data (Fig. 16.3). The percent of trees recruited before 1650 from all trees in each collection between 1550 and 1849 equals 8.7%, 2.9%, and 5.5% for the Hawley-Bell, Kentucky, and eastern broadleaf collections, respectively. The percentage of inner ring dates between 1650 and 1699 equals 56.5%, 30%, and 19%, respectively. Unexpectedly, timing and distribution shape of the inner ring dates of the Hawley-Bell collection subset containing the period of increased recruitment (1620–1699) did not significantly differ from those of the targeted trees of the Kentucky collection, whether a sample contained the pith or not (Fig. 16.3; Chi Square test of independence, chi-square = 8.92, DF = 8, prob. = 0.349). The eastern broadleaf collection shows the continual recruitment of trees through time with prominent peaks in the mid 1600s and early and late 1700s.
The continued recruitment of trees through time in the eastern broadleaf data and the lack of continued recruitment in the Hawley-Bell and Kentucky collections can be explained by differences in sampling strategy. The eastern broadleaf collection was developed through sampling plots that were designed to be representative of the forest (Pederson et al. 2014). The people who built historic structures and the dendroclimatologists who collected the Kentucky samples had a significantly different approach. Similar to the objectives of the Hawley-Bell collection, dendroclimatologists targeted and favored the oldest appearing trees in the forest from specific species that were known or hypothesized to be long-lived, cross-datable, and/or sensitive to climatic variation. In addition to this and as discussed above and below, historic builders often targeted trees by size and, when possible, by species. The resulting dendro-archeo-ecology records, therefore, are constrained by economics: some people could afford to buy timbers from elsewhere, whereas other people could only afford to harvest trees on site or repurpose timbers from existing structures on site. Finally, the targeting of trees for dendroclimatology and dendro-archeo-ecology purposes could constrain collections in ways that make them only representative of a specific point in time. Together, these differences give significant uncertainty on the spatial representativeness of dendro-archeo-ecology collections and, in the case of repurposing or buying timbers from elsewhere, representativeness of a certain period.
The Hawley-Bell collection was collected 40–60 years prior to the dendroclimatic collections in Kentucky. Hawley and Bell aimed to select the oldest samples from this collection and were aided in their ability to select the oldest trees by seeing the rings on cross-sections before selecting samples. Dendroclimatologists do not have this luxury. Yet, the Hawley-Bell collection indicates an uptick in recruitment of Quercus in the 1650s and a spike beginning in the 1660s. Recruitment in the 1660–1690s was almost fivefold that of the 1610–1640s (24 vs. 5 samples in each 40-year period) and thus mirrors the 10-fold increase in the Kentucky dendroclimatic collections (72 vs. 7 samples for the same periods; Fig. 16.3). The age structure of the Hawley-Bell collection is also nearly the same as collections made in 1915–16 in a freshly cut old-growth forest in southeastern Kentucky (Haasis 1923) and collections from Missouri, Kentucky, West Virginia, and New York State in the early 1900s (Huntington 1914). If recruitment in broadleaf-dominated forests is simply dominated by the frequent and annual dynamics of forests at fine scales, we would expect the spike in recruitment of the Hawley-Bell collection to occur four to six decades earlier than in the Kentucky collection.
A spike in age structure in one to a few historical structures built around the same time could be a false positive due to the nearly simultaneous establishment of targeted timbers within a narrow range of sizes. With this combination of timber selection and ecological dynamics, one would expect a flat age structure from a collection drawn from multiple structures over multiple periods (null hypothesis). The ecological process that counters this hypothesis is that the range of shade tolerance and the ability to persist in the understory at the species- and individual tree-level could result in trees of the same size or stature representing trees who were recruited to coring or cutting height in different decades and, sometimes, centuries. A case study of Quercus montana trees—including the oldest-documented Quercus montana—supports this alternative hypothesis. The oldest Quercus montana tree has an inner ring dating to 1578 and appears to have been an understory tree for more than three centuries (Pederson 2010). Three of the four other Quercus montana trees from the same stand that have pith or are likely within 5 years of pith (out of 32 cored trees) date to the 1750s or 1790s (Pederson et al. 2017). Thus, shade tolerance and ecological history of a site could likely make timbers of dendro-archeo-ecology collections relatively diverse in age as builders were targeting trees of specific sizes regardless of age. At minimum, these collections would not likely be monolithic in their representativeness of a specific era.
A re-analysis of the 461 historical timbers used in Pederson et al. (2014) drawn from several collections that cover New Jersey, Pennsylvania, Ohio, Indiana, and Arkansas (but dominated by samples in Ohio) allows us to further investigate this hypothesis. The age structure of this collection results in more trees being recruited prior to 1660 and a relatively flat age structure from 1590 to 1650. However, only 18% of this collection has trees with inner rings between 1530 and 1649 compared to the 15% with inner rings between 1670 and 1689 (the peak in recruitment of this data set), and 35% with inner rings between 1660 and 1699, suggesting that perhaps the increase in recruitment in the mid- to late-1600s from the Pederson et al. (2014) dendro-archeo-ecological collection is not due to a bias driven by targeted collections and site ecology. A flat age structure is not produced by samples drawn from multiple structures built over differing periods and a large space. These findings further suggest that perhaps the relatively narrow period from the mid- to late-1600s was an important event to the structure and function of forests in the central Midwest of the eastern U.S. More collections from a greater range of building periods spread more evenly over a larger region would be an important test of this hypothesis. As it is, our subsample of the Hawley-Bell collection provides additional evidence that episodic disturbance appears to be an important driver of structure of Quercus- and other broadleaf-dominated forests in the eastern U.S.
The ecological potential of tree-ring studies on Euro-Mediterranean pile-dwelling timbers has so far only been explored at a limited number of sites. Replication and expansion of approaches summarized by Billamboz (2014) offer much potential to bring needed spatial and temporal coverage to data relevant in discussions of Anthropocene onset. Revisiting archived wood collections such as the Hawley-Bell collection may lead to integrated dendro-archaeo-ecological studies for the Neolithic and a wide range of other periods. Unbiased sampling strategies that also consider timbers from young trees and that include a wide range of species should then be adopted. These types of sampling strategies are becoming the norm in multidisciplinary studies that involve dendrochronology, such as those on shipwrecks (Haneca and Daly 2014) and that aim not only at dating timbers, but also at gathering information about the organization of the timber supply, the selection and management of species, and the origin of the wood. Such dendroprovenancing studies can add detail to our existing understanding of the intensification of deforestation in Europe and in the Americas. For instance, the lack of suitable wood for the construction of a harbour at the Roman site of Voorburg-Arentsburg, in the current Netherlands, required the transport of timber over more than 600 km (Domínguez-Delmás et al. 2014). Centuries later, following drastic deforestation in the Middle Ages, construction timber in the Low Countries (northern Belgium and the Netherlands) was acquired through importation primarily from the Baltic region, Scandinavia, and Germany (e.g., De Vries and Van der Woude 1997).
We here advocate for the integration of such individual studies in order to optimally inform and calibrate the Early Anthropocene hypothesis and to fill existing gaps in the archeological record. Several initiatives have recently been launched to facilitate integration by addressing common hurdles to data sharing and to new uses for old data. These initiatives include a tree-ring repository for dendro historical/archaeological data in Europe —the Digital Collaboratory for Cultural Dendrochronology, DCCD (Jansma et al. 2012; Jansma 2010), a universal tool for converting non-standard digital tree-ring data to a standard transferable format (Brewer et al. 2011), and a move towards a new tree-ring data standard, TRiDaS (Jansma et al. 2010). For instance, meta-data regarding scars and other aspects of wood anatomy and anthropogenic alteration (Fig. 16.4) can be collected at the time of measurement using the measurement and curation software Tellervo (http://www.tellervo.org/), which is designed to archive multi-user data and metadata according to TRiDaS for future cross-disciplinary applications. The TRiDaS initiative thus presents significant new opportunities for the wider dendrochronological community, in that by standardizing data and metadata collection, it makes data collected for one specific type of study fully accessible for analysis under a different sub-discipline.
This is particularly relevant and helpful for archaeological data from the Neolithic period, and specifically for pile-dwelling remains, as most of the material is preserved as wet wood and so is subject to rapid degradation and decay if effective conservation strategies (e.g., freeze drying) cannot be immediately put in place. For instance, rapid degradation by drying occurs as wood is gradually exposed to the air during archeological excavation (the splitting at the outer edge in Fig. 16.4).
Facilitating the collection of all possibly relevant metadata when processing wet wood is a critical step forward in constructing the platform for a whole range of interdisciplinary future studies. Such concerted efforts are needed to fully exploit the potential of dendrochronology at the archeology—ecology—climate nexus and to move the use of dendro-archeological material from strictly archeological applications (e.g., dating) towards exploration of its ecological potential.
Before we can fully utilize historical timbers for ecological studies, there are more uncertainties to be considered. For example, sixteenth century builders likely biased North American samples because they often chose stems based on size: large diameter stems would be selected for e.g. tie beams, but avoided for other timber structures such as rafters, due to necessary laborious sizing and massive weight. They may have further biased our samples by selecting long, clear boles with little to no branching to avoid delimbing. Such logs would most likely be found in stands with high stem density. Builders also selected species with optimal decay resistance and workability, which potentially limits the number of taxa available for analysis (Hutslar 1992; Baas and Rubino 2013). This is also the case for central and northern Europe, where oak was the preferred species for construction purposes (Haneca et al. 2009). In the eastern U.S., however, we have worked on structures constructed with a larger range of species than expected. Despite the increased activity in dendroarcheology, we are still in the initial growth stage for ecological applications (De Graauw n.d.).
In addition to potential size and species biases, uncertainty might also be introduced by replicate samples when multiple timbers in a structure derive from the same tree. This can be detected through high extremes in inter-series correlations and tree-ring series should therefore be internally crossdated in a first identification step before they are included in further analyses (e.g. Mom et al. 2009; Domínguez-Delmás et al. 2011, 2014). A significant unknown in using building timbers to study forest recruitment is the height at which the tree was cut. The trees from which timbers were cut could be significantly older than the timbers themselves, even when sapwood and/or pith are present in the samples. When it is not possible to collect a complete cross section, core samples may miss the pith adding additional error to recruitment estimates. Finally, when sampling historical buildings, meta-data related to the original forest stand structure (e.g., spatial distribution and canopy class of stems) are missing, which can introduce uncertainty when historic timbers are used for forest ecological research purposes. Together, there is much uncertainty in using dendro-archeological data for ecological studies. However, dendro-archeo timbers may help us recover highly resolved forest dynamics during periods that predate most extant forests. Some recurring challenges such as low sample depth will likely be remedied as this field expands, while other limitations (e.g., short series length) and uncertainties (e.g., species and tree size bias) need to be taken into account and carefully addressed during analysis and interpretation.
In addition to an improved integration of dendrochronological subdisciplines, the quantification of past land-use changes and their impact on the carbon cycle and on Earth’s climate calls for a close alliance of dendrochronology with related scientific disciplines that aim to address the same subjects (Ellis et al. 2013b). For instance, estimates of eastern North American forest productivity at the time of Euro-American settlement could be complemented by tree composition estimates statistically derived from nineteenth century public land survey records (Paciorek et al. 2016; Goring et al. 2015).
Dendroecology could also be used to calibrate the impact of the indigenous North American depopulation following European contact on historical fire regimes (Liebmann et al. 2016; Taylor et al. 2016), biomass burning, and thus global CO2 levels (Nevle et al. 2011). This work would complement the longer-term context using sedimentary charcoal records (Marlon et al. 2012). Such interdisciplinary collaborations could greatly benefit from the inclusion of model-based approaches, as has been demonstrated in the spatially explicit reconstruction of past land use in Mesoamerica (Cook et al. 2012) and in the Euro-Mediterranean region (Büntgen et al. 2011). Such model-data assimilation exercises are particularly useful when linking past land-use changes and forest productivity estimates to atmospheric CO2 concentrations (Kaplan 2015; Ellis et al. 2013b). Whereas tree-ring data have been used to estimate the carbon sequestration capacity of modern forests (Babst et al. 2014b, c; Dye et al. 2016; Alexander et al. n.d.) and to constrain the global carbon cycle’s sensitivity to climate (Frank et al. 2010), the spatially explicit reconstruction of past forest carbon sequestration capacity will require their pairing with sophisticated and detailed modeling efforts (Ellis et al. 2013b).
At the nexus of archeology, climatology, and ecology, dendrochronology is uniquely positioned to study past human-environment interactions. However, its potential to contribute to the quantification of Holocene-era land-use changes and forest dynamics in Europe and North America is limited because the majority of lowland forests in these regions have been cleared throughout the course of human history. As a result, tree-ring chronologies from living trees are derived from either secondary forests or remnant old-growth stands and might not be representative for forest dynamics and productivity. We advocate for the development and application of dendro-archeo-ecology: deriving ecological information from dendroarcheological collections. Doing so will require a concerted effort to address recurring challenges and uncertainties, adopting unbiased sampling strategies, and revisiting surviving wood collections. This will facilitate the integration of dendrochronology in interdisciplinary collaborations that focus on the quantification of past land-use changes and their impact on the carbon cycle and on Earth’s climate.
The authors thank P. Baker and A. Hessl for very constructive feedback on earlier versions of this manuscript. NP thanks VT for partially subsidizing work conducted on the Hawley-Bell Collection.
- Alexander MR, Rollinson CR, Babst F, Trouet V, Moore DJP (n.d.) Uncertainty in tree-ring based biomass estimates does not alter growth-climate relationships. Trees Struct Funct, in revisionGoogle Scholar
- Baas C, Rubino DL (2012) Using tree-ring growth patterns to date the construction of a beater hay press barn in Allensville, Switzerland County, Indiana. Ind Archaeol 7:35–57Google Scholar
- Baas C, Rubino DL (2013) The most successful press in this or any other country: the material culture of 19th-century beater hay presses in the Mid-Ohio Valley. Mater Cult 45:1–20Google Scholar
- Baillie MGL (1982) Tree-ring dating and archaeology. University of Chicago Press, ChicagoGoogle Scholar
- Bell RE (1940) Informal field notes taken during two months wood collecting in the state of Arkansas—1940. The University of Chicago, Laboratory of Tree-Ring Research, University of Arizona. p 27Google Scholar
- Bell RE (1941) Informal field notes for Dendrochronology Research during the summer of 1941—Illinois, Kentucky, Tennessee, Arkansas and Wisconsin. The University of Chicago, Laboratory of Tree-Ring Research, University of Arizona. p 31Google Scholar
- Billamboz A (2004) Dendrochronology in lake-dwelling research. In: Menotti F (ed) Living on the lake in prehistoric Europe. Routledge Press, New York, pp 117–131Google Scholar
- Brown PM, Hughes MK, Baisan CH, Swetnam TW, Caprio AC (1992) Giant sequoia ring-width chronologies from the central Sierra Nevada, California. Tree Ring Bull 52:1–14Google Scholar
- Butler AW (1896) A century of changes in the aspects of Nature. Proc Indiana Acad Sci 5:3–42Google Scholar
- Chrysostomou P (2015) The Culture of the Four Lakes. Conclusions from the archaeological research in Amindeon basin (2000–2014). In: Chrysostomou P (ed) The Culture of Four Lakes, Wooden artifacts from the Prehistoric lakeside settlements of Amindeon basin. Epikentron, ThessalonikiGoogle Scholar
- Cook ER, Callahan B (1992) The development of a standard tree-ring chronology for dating historical structures in the greater Philadelphia region. Tree-Ring Laboratory, Lamont-Doherty Geological Observatory, Columbia University, New YorkGoogle Scholar
- Cronon W (1983) Changes in the land: Indians, colonists, and the ecology of New England. Hill and Wang, New YorkGoogle Scholar
- Crutzen PJ, Stoermer EF (2000) The Anthropocene. IGBP Newsl 41:12Google Scholar
- Davis MB (1993) Old-growth in the east: a survey. Cenozoic Society, RichmondGoogle Scholar
- Davis MB (1996) Extent and location. Eastern old-growth forests. Island Press, Washington, DCGoogle Scholar
- De Graauw KK (2017) Historic log structures as ecological archives: a case study from eastern North America. Dendrochronologia. doi: 10.1016/j.dendro.2017.06.002
- Domínguez-Delmás M, Benders JF, Kortekaas GLGA (2011) Timber supply in Groningen (north-east Netherlands) during the Early Modern Period (16th–17th centuries). In: Fraiture P (ed) Tree-rings, art and archaeology, proceedings, Brussels, Collection Scientia Artis, vol 7. pp 151–173Google Scholar
- Douglass AE (1929) The secret of the Southwest solved by talkative tree rings. Nat Geogr Mag 56(6):736–770Google Scholar
- Doyen E, Bégeot C, Simonneau A, Millet L, Chapron E, Arnaud F et al (2016) Land use development and environmental responses since the Neolithic around Lake Paladru in the French Pre-alps. J Archaeol Sci 7:48–59Google Scholar
- Girininkas A (2010) The Žematiske 2 pile-dwelling settlement. Archaeol Balt 14:120–135Google Scholar
- Goring S, Mladenoff DJ, Cogbill CV, Record S, Paciorek CJ, Jackson ST, et al (2015) Changes in forest composition, stem density, and biomass from the Settlement Era (1800s) to present in the Upper Midwestern United States. doi:10.1101/026575Google Scholar
- Gumerman GJ (1988) The Anasazi in a changing environment. Cambridge University Press, CambridgeGoogle Scholar
- Haasis FW (1923) Significance of a 255-year age class in an eastern Kentucky forest. J For 21:700–704Google Scholar
- Harriot T (1972) A briefe and true report of the new found land of Virginia (1588). Dover Publications, MineolaGoogle Scholar
- Hawley F (1941) Tree-ring analysis and dating in the Mississippi drainage. Occasional Papers, No. 2. The University of Chicago Press, ChicagoGoogle Scholar
- Huber B (1967) Seeberg, Bürgäschisee-Süd. Dendrochronologie. In: Bandi HG, Hofer P, Jucker H, Strahm H (eds) Seeberg, Bürgäschisee-Süd. Teil 4, Chronologie und Umwelt. Acta Bernensia II Teil 4. Verlag Stämpfli & Cie, Bern. pp 145–156Google Scholar
- Huisman DJ, Klaassen RKWM (2009) Wood. In: Huisman DJ (ed) Degradation of archaeological remains. Sdu Uitgeverij, Den Haag, pp 17–32Google Scholar
- Huntington E (1914) The climatic factor as illustrated in arid America. Carnegie Institution of Washington, Washington, DCGoogle Scholar
- Hutslar DA (1992) Log construction in the Ohio Country, 1750–1850. Ohio University Press, AthensGoogle Scholar
- Ivanov I (2000) Varna and the birth of the European civilization. In: Varna necropolis: the dawn of European civilization. SAGATÓ Publishers, SofiaGoogle Scholar
- Jansma E (2010) Preserving tree-ring data: a repository for the low countries. In: de Groot M, Wittenberg M (eds) Driven by data: exploring the research horizon. Amsterdam University Press/Pallas Publications, Amsterdam, pp 29–33Google Scholar
- Kirby KJ, Watkins C (eds) (1998) The ecological history of European forests. CAB International, OxonGoogle Scholar
- Krusic PJ, Wright WE, Cook ER (2004) Tree-ring dating of the Cahn House New Paltz, New York. Tree-Ring Laboratory, Lamont-Doherty Geological Observatory, Columbia University, New YorkGoogle Scholar
- Kulkova MA, Mazurkevich AN, Dolukhanov PM (2001) Chronology and palaeoclimate of prehistoric sites in western Dvina-Lovat area of northwestern Russia. Geochronometria 20:87–94Google Scholar
- Lamb HH (1995) Climate, history and the modern world. Routledge, LondonGoogle Scholar
- Le Roy Ladurie E (1971) Times of feast, times of famine: a history of climate since the year 1000. Doubleday, LondonGoogle Scholar
- Mazurkevich A, Dolbunova EK, Maigrot Y, Hook D (2010) The results of underwater excavations at Serteya II, and research into pile-dwellings in northwest Russia. Archaeol Balt 14:47–64Google Scholar
- Meiggs R (1982) Trees and timber in the ancient Mediterranean world. Oxford University Press, OxfordGoogle Scholar
- Menotti F (ed) (2004) Living on the lake in prehistoric Europe: 150 years of lake-dwelling research. Routledge, LondonGoogle Scholar
- Mom V, Schultze J, Wrobel S, Eckstein D (2009) Allocating archaeological wood samples to a common source tree and its use for analyzing wooden settlement structures. In: Making History Interactive; Computer Applications to Archaeology, Proceedings of the XXXVII Computer Applications and Quantitative Methods in Archaeology Conference, CAA-Williamsburg, Virginia, pp 234–239Google Scholar
- Nowacki GJ, Abrams MD (1997) Radial-growth averaging criteria for reconstructing disturbance histories from presettlement-origin oaks. Ecol Monogr 67:225–249Google Scholar
- Pederson N, Young A, Stan AB, Ariya U, Martin-Benito D (2017) Low-hanging dendrodynamic ecological fruits regarding disturbance in temperate, mesic forests. In: Amoroso MM, Daniels LD, Baker PJ, Camarero JJ (eds) Dendroecology: tree-ring analyses applied to ecological studies. Springer, ChamGoogle Scholar
- Rackham O (1990) Trees and woodland in the British landscape. JM Dent & Sons Ltd., LondonGoogle Scholar
- Roberts WE (1996) Log buildings of southern Indiana. Trickster Press, BloomingtonGoogle Scholar
- Rubino DL (2014) Tree-ring analysis in Indiana with special attention to dendroarchaeology. Proc Indiana Acad Sci 123:7–13Google Scholar
- Rubino DL, Baas C (2014) Using tree-ring growth patterns to date the construction of a nineteenth century dogtrot in house in Pose County, Indiana. Proc Indiana Acad Sci 123:14–22Google Scholar
- Ruddiman WF, Carmichael A (2007) Pre-industrial depopulation, atmospheric carbon dioxide, and global climate. Interactions between global change and human health. Scripta Varia 106:158–194Google Scholar
- Sass-Klaassen U, Hanraets E (2006) Woodlands of the past—the excavation of wetland woods at Zwolle-Stadshagen (the Netherlands): growth pattern and population dynamics of oak and ash. Neth J Geosci 85(01):61–71Google Scholar
- Sass-Klaassen U, Kooistra M, Kooistra L, Hanraets E, Van Rijn P, Leuschner HH (2004) How did bog oaks grow? Excavation of a past woodland at Zwolle-Stadshagen, The Netherlands. In: Jansma E, Bräuning A, Gärtner H, Schleser G (eds) TRACE—Tree rings in archaeology, climatology and ecology, 2. Schriften des Forschungszentrums Jülich, Reihe Umwelt, vol 44, pp 112–115Google Scholar
- Senter FH (1938a) Can we fix prehistoric dates in the Middle West by tree rings? Indiana Hist Bull 15:118–128Google Scholar
- Senter FH (1938b) Dendrochronology in two Mississippi drainage tree-ring areas. Tree Ring Bull 5:3–6Google Scholar
- Sherratt A (2004) The importance of lake-dwellings in European prehistory. In: Menotti F (ed) Living on the lake in prehistoric Europe: 150 years of lake-dwelling research. Routledge, London, pp 267–276Google Scholar
- Stahle DW (1979) Tree-ring dating of historic buildings in Arkansas. Tree Ring Bull 39:1–28Google Scholar
- Stahle DW, Chaney PL (1994) A predictive model for the location of ancient forests. Nat Areas J 14:151–158Google Scholar
- Stokes MA, Smiley TL (1968) Tree-ring dating. University of Chicago Press, ChicagoGoogle Scholar
- Tarrús i Galter J (2008) La Draga (Banyoles, Catalonia), an Early Neolithic Lakeside Village in Mediterranean Europe. Cat Hist Rev 1:17–33Google Scholar
- Tercier J, Orcel A, Orcel C (1996) Dendrochronological study of prehistoric archaeological sites in Switzerland. In: Dean JS, Meko DM, Swetnam TW (eds) Tree-rings, environment, and humanity: proceedings of the international conference, Tucson, 1994. Radiocarbon, Tucson, AZ, pp 567–582Google Scholar
- Thirgood JV (1981) Man and the mediterranean forest. Academic, LondonGoogle Scholar
- Thornton R (1987) American Indian holocaust and survival: a population history since 1492. University of Oklahoma Press, NormanGoogle Scholar
- Thornton R (2000) Population history of native North Americans. In: Haines MR, Steckel RH (eds) A population history of North America. Cambridge University Press, CambridgeGoogle Scholar
- UNESCO (2015) Prehistoric Pile Dwellings around the Alps, http://sites.palafittes.org. Accessed 30 Jan 2017
- USDA (2015a) USDA Forests of Illinois. https://www.fs.fed.us/nrs/pubs/ru/ru_fs85.pdf. Accessed 30 Jan 2017
- USDA (2015b) USDA Forests of Indiana. https://www.fs.fed.us/nrs/pubs/ru/ru_fs88.pdf. Accessed 30 Jan 2017
- USDA (2015c) USDA Forests of Ohio. https://www.fs.fed.us/nrs/pubs/ru/ru_fs97.pdf. Accessed 30 Jan 2017
- White PS, Pickett STA (1985) Natural disturbance and patch dynamics: an introduction. In: STA P, White PS (eds) The ecology of natural disturbance and patch dynamics. Academic, New York, pp 3–13Google Scholar
- Whittle A (1996) Europe in the Neolithic: the creation of new worlds. Cambridge University Press, CambridgeGoogle Scholar